USF Libraries
USF Digital Collections

Toward an applied anthropology of gis :

MISSING IMAGE

Material Information

Title:
Toward an applied anthropology of gis : spatial analysis of adolescent childbearing in hillsborough and pinellas counties, florida
Physical Description:
Book
Language:
English
Creator:
Maes, Kathleen
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Teen births
Spatial analysis
Hot spot analysis
Mapping
Socoieconomic index
Small area analysis
Dissertations, Academic -- Anthropology -- Doctoral -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: This work investigates births to white, African American and Hispanic adolescents in Hillsborough and Pinellas Counties, Florida, from 1992 to 1997 in two age groups - 13 to 17 year-olds and 18 to 19 year-olds - using spatial statistical techniques along with key informant interviews to provide insights into the utility of the research findings. The research developed a method for estimating the adolescent population in inter-census years, which was used to determine denominators for calculating teen birth rates. It also developed a composite deprivation index using socioeconomic indicators at the census block group level. The index provided context for hot and cold spot analysis, areas where expected teen birth rates were statistically higher or lower than expected. The association between socioeconomic deprivation in a neighborhood and rates of teen births was inconclusive, indicating a need for further research. Next steps include investigating individual-level risk and protective factors using multi-level modeling and cluster analysis as alternate analytic methods, and conducting ethnographic investigation to help provide context to the neighborhoods.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2010.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Kathleen Maes.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains X pages.
General Note:
Includes vita.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
usfldc doi - E14-SFE0003466
usfldc handle - e14.3466
System ID:
SFS0027781:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2010 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003466
035
(OCoLC)
040
FHM
c FHM
049
FHMM
090
XX9999 (Online)
1 100
Maes, Kathleen.
0 245
Toward an applied anthropology of gis :
b spatial analysis of adolescent childbearing in hillsborough and pinellas counties, florida
h [electronic resource] /
by Kathleen Maes.
260
[Tampa, Fla] :
University of South Florida,
2010.
500
Title from PDF of title page.
Document formatted into pages; contains X pages.
Includes vita.
502
Dissertation (Ph.D.)--University of South Florida, 2010.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
3 520
ABSTRACT: This work investigates births to white, African American and Hispanic adolescents in Hillsborough and Pinellas Counties, Florida, from 1992 to 1997 in two age groups 13 to 17 year-olds and 18 to 19 year-olds using spatial statistical techniques along with key informant interviews to provide insights into the utility of the research findings. The research developed a method for estimating the adolescent population in inter-census years, which was used to determine denominators for calculating teen birth rates. It also developed a composite deprivation index using socioeconomic indicators at the census block group level. The index provided context for hot and cold spot analysis, areas where expected teen birth rates were statistically higher or lower than expected. The association between socioeconomic deprivation in a neighborhood and rates of teen births was inconclusive, indicating a need for further research. Next steps include investigating individual-level risk and protective factors using multi-level modeling and cluster analysis as alternate analytic methods, and conducting ethnographic investigation to help provide context to the neighborhoods.
590
Advisor: Linda M. Whiteford, Ph.D.
653
Teen births
Spatial analysis
Hot spot analysis
Mapping
Socoieconomic index
Small area analysis
690
Dissertations, Academic
z USF
x Anthropology
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.3466



PAGE 2

ACKNOWLEDGEMENTS I wish to express my most sincere gr atitude to everyone who has supported me over the years in this endeavor. In particul ar, I would like to thank Dr. Linda Whiteford for her continued support and faith in me. I al so want to thank Dr. Steven Reader for the endless hours of patient assist ance and guidance with this re search. A sincere thank you also goes to my committee me mbers, Dr. Susan Greenbaum, Dr. Cecilia Jevitt and Dr. Nancy Romero-Daza, whose work has inspired me. My sincerest gratitude goes to the Children’s Board of Hillsborough County, Florida, for supporting me in my educational as pirations. And to all of my colleagues at the Children’s Board, thank you for your support and encouragement. Last, and most certainly not least, I wa nt to thank my family and friends who never stopped believing in me and who still re main close to me in spite of my absence while writing this dissertation.

PAGE 3

i TABLE OF CONTENTS LIST OF TABLES iv LIST OF FIGURES v ABSTRACT vi CHAPTER ONE: INTRODUCTION TO THE STUDY 1 Introduction to the Study 1 Background 3 Conceptual Framework 5 Purpose of the Study 7 CHAPTER TWO: REVIEW OF RELATED LITERATURE 9 Anthropology and Analysis of Space 9 Anthropology and Geographic Information Systems (GIS) 12 Hot Spot Analysis 15 Multilevel Modeling and Ecosocial Theory 19 Risks and Outcomes of Adolescent Pregna ncy and Childbearing 25 Maternal Risks and Outcomes 26 Neonatal and Child Health Risks and Outcomes 30 Behavioral Risks and Outcomes 35 Social Risks and Outcomes 40 Factors Associated with Adolescent Childbearing 45 Summary of the Literature 55 CHAPTER THREE: RESEARCH DESIGN AND METHODOLOGY 56 Overview of the Research Problem 56 Research Objective 57 Research Design 58 Existing Data Sets Used 58 1990 and 2000 Decennial U.S. Census of Population and Housing 59 U.S. Census Bureau's Annual Population Estimates 61 Florida Community Health Assessment Re source Tool Set 61 Neighborhood Change Database (NCDB) 62 State of Florida Vital Statistics Bi rth Data Files 63 Determining Teen Birth Rates 64 Preliminary Decisions 65 Calculating Denominators 68 Calculating Age by Race and Ethnicity 69

PAGE 4

ii Bridging Race 71 Population Estimates 73 Reconciling 1990 and 2000 Geographic Boundaries 74 Calculating Teen Birth Rate s 75 Hot Spot Analysis 75 Determining Contextual Level Variables 76 Exploratory Interviews 76 Review of the Literature 79 Composite Indices 79 Developing an Index of Socioeconomic Inequality 80 Interviews with Teen Pregnancy Prevention Agencies 82 Summary of Methods 83 CHAPTER FOUR: RESULTS 84 Findings 84 Descriptive Statistics 85 Births to Adolescents in Hillsborough and Pinellas Counties 87 Hot Spot and Cold Spot Analysis 92 Interview Results – Round 1 103 Index of Socioeconomic Inequality 110 Interview Results – Round 2 114 Provider Interview Results 116 Provider Interest 116 Relevance/Usefulness to Providers 116 Perception of Expected and Unexpected Results 117 Funding Agency Interview Results 118 Funder Interest 118 Relevance/Usefulness to Funding Agencies 119 Perception of Expected and Unexpected Results 119 Common Themes Among Provide rs and Funders 120 Summary of Results 121 CHAPTER FIVE: DISCUSSION 122 Discussion of the Study 122 Caveats and Limitations 132 Summary 134 CHAPTER SIX: CONCLUSIONS AND RE COMMENDATIONS 136 Conclusions 137 Recommendations 140 REFERENCES CITED 143 APPENDICES 157 Appendix A: Maps 158

PAGE 5

iii ABOUT THE AUTHOR End Page

PAGE 6

iv LIST OF TABLES Table 2.1 Birth Rates for 15-19 Year Olds in 2000 by Race/Ethnicity 26 Table 3.1 1990 and 2000 U.S. Census Summary File 1 Age by Race Tables 69 Table 3.2 1990 and 2000 U.S. Census Age Categories 70 Table 3.3 Selected Indicators for Index of Socioeconomic Inequality 81 Table 4.1 Births by Age-Group, Race/Ethnicity and by Year for the State of Florida 86 Table 4.2 Datasets with Record Counts Used for Analyses 89 Table 4.3 Average Annual Births 1992-1997 by Age Group and by Race and Ethnicity 92 Table 4.4 Birth Rates (per 1,000 li ve births) 1992-1997 by Age Group and by Race and Ethnicity 93 Table 4.5 Hot Spot Census Tracts in Hills borough and Pinellas Counties 95 Table 4.6 Cold Spot Census Tracts in H illsborough and Pinellas Counties 96 Table 4.7 Factors Influencing Adolescen t Childbearing with Data Availability and Possible Sources 109 Table 4.8 Spearman’s Rho Correlation 113 Table 5.1 Average z-Scores for Census Tract Hot Spots 126 Table 5.2 Average z-Scores for Census Tract Cold Spots 128

PAGE 7

v LIST OF FIGURES Figure 3.1 Hierarchical Relationships of U.S. Census Geography 67 Figure 4.1 Average Annual Numbers of Births 19921997 by Age and Race/Ethnicity in the State of Florida 87 Figure 4.2 Average Annual Number of Births 1992 1997 by Age and Race/Ethnicity in Hillsborough County, Florida 88 Figure 4.3 Average Annual Number of Births 1992 1997 by Age and Race/Ethnicity in Pinellas County, Florida 89

PAGE 8

vi TOWARD AN APPLIED ANTHR OPOLOGY OF GIS: SPATIAL ANALYSIS OF A DOLESCENT CHILDBEARING IN HILLSBOROUGH AND PINE LLAS COUNTIES, FLORIDA Kathleen I. Maes ABSTRACT This work investigates births to wh ite, African American and Hispanic adolescents in Hillsborough and Pinellas Count ies, Florida, from 1992 to 1997 in two age groups – 13 to 17 year-olds and 18 to 19 year-o lds – using spatial statistical techniques along with key informant interviews to provide insights into the uti lity of the research findings. The research developed a method for estimating the adolescent population in inter-census years, which was used to dete rmine denominators for calculating teen birth rates. It also developed a composite deprivation index usi ng socioeconomic indicators at the census block group level. The index provide d context for hot and cold spot analysis, areas where expected teen birth rates were st atistically higher or lo wer than expected. The association between socioeconomic depr ivation in a neighborhood and rates of teen births was inconclusive, indicating a need for further research. Next steps include investigating individual-level risk and pr otective factors using multi-level modeling and cluster analysis as alternat e analytic methods, and conduc ting ethnographic investigation to help provide context to the neighborhoods.

PAGE 9

1 CHAPTER ONE: INTRODUCTION TO THE STUDY Anthropologists are increasingly working in diverse settings outside academia, in sectors such as government, commercial, industr ial and financial as well as in not-forprofit organizations (Hahn 2009). As anthropo logists have expanded their work settings, application of anthropological methods also mu st be expanded. One useful technique, not often employed by cultural anthropologists, is spatial analysis. Spatial analysis, using Geographic Information Systems (GIS), pr ovides a unique analytic perspective by allowing “people to look at data in a w hole new way by seeing all the pieces at once” (Lang 2000:2) as geographically refere nced data in the form of a map. Introduction to the Study This study developed a method for expl oring neighborhood variables within a spatial landscape that may be used to examine a given outcome. Births to adolescents in two Florida counties, Hillsborough and Pinellas are used to test the proposed methods and contribute to an emerging “applied anth ropology of GIS/spatia l analysis” by showing ways in which anthropological methods and pe rspectives can be integrated into a more purely quantitative and sp atial research design. A growing movement toward an “applied anthropology of GIS/sp atial analysis” is underway. At the 101st Annual Meeting of the American Anthropological Association in 2002, Dr. Susan C. Stonich called for the anthropology of rather than anthropology in

PAGE 10

2 spatial analysis/GIS. This new anthropology will require an interdisciplinary approach, integrating a strong theoretical framework with tested methods and pr actical application. The theoretical framework must integrate statistical, geographic cartographic, and anthropological foundations along with the topic being studied (e.g., public health, environmental conservation, etc.) Methods mu st accommodate both spatial and statistical analyses as well as ethnogr aphic techniques. Finally, an “applied an thropology of GIS/spatial analysis” must have practical application and the wo rk must be useful (Stonich 2002). GIS allows the researcher to make visible the abstract concepts of space as applied, for instance, to patter ns of adolescent reproduction. This study uses a mixed method design to provide information from quantitative, spatial, and qualitative perspec tives. Descriptive statistics pr ovide direct insight into the data, helping to clarify distributi ons of adolescent birth rates. Hot spot analysis identifies where rates of adolescent birt hs are significantly higher or lower than expected and presents these hot spots and cold spots spatially on a map. Using area-based socioeconomic measures (ABSMs), an Index of Socioeconomic Inequality was created as part of this research to in vestigate the relationship betw een contextual factors (i.e., neighborhood characteristics) and births to teen. Interviews conducted with county and state-level adolescent pregnancy preventi on service providers and funding agencies provide information regarding individua l-level and neighbor hood-level variables associated with adolescent childbearing, in addition to respondents’ perceptions of the utility and relevance of the information genera ted in this research for the work they do. Aggregated birth data fro m 1992 through 1997 were used in this study. Births to adolescents are presented for two age groups (13 to 17 year-olds and 18 and 19 year-olds)

PAGE 11

3 and by race and ethnicity, including white, Af rican American and Hispanic. Presenting patterns of adolescen t childbearing geographically by age-group and by race/ethnicity provides the opportunity to view these births from a cultural perspective and to consider different cultural norms, mores and customs th at may affect attitudes toward adolescent childbearing. Maps showing the approximate location of teen’s mothers’ home addresses were produced for each of the age and race /ethnic group of adolescents, providing the opportunity to see spatial distri butions and patterns of where these teens live. Maps were also produced to show different levels of socioeconomic inequality/stress in neighborhoods throughout Hillsborough and Pinellas counties, Florida, based on variables associated with adolescent chil dbearing and incorporated into an index developed for this study. These areas, with differing levels of social stress, provide context for hot spot and cold spot nei ghborhoods where adolescent birth rates, by age group and race/ethnicity, are statistically higher or lower than expected. Originally, this study planned to use cluster analysis, a different method to identify hot and cold spots which is not limited by geographic boundaries, and a multilevel modeling approach as an explanatory technique. However, after months of work, it became apparent that this was be yond the scope of a disse rtation project and would require a multi-person team approach to complete, which would be best left for a post-dissertation project. Background I have worked for several years as an applied anthropologist in the Research Department of the Children's Board of H illsborough County, a local Children’s Services

PAGE 12

4 Council in Tampa, Florida. As an independe nt taxing district, the Children’s Board of Hillsborough County uses property tax dollars to strategically invest in primary prevention and early intervention programs and services for children and families in Hillsborough County, Florida. In the past few years, with rising une mployment rates, increasing numbers of home foreclosures, wildly fluctuating gas pr ices, a shortage of affordable housing, and economic recession, both politicians and prac titioners want to ensure wise use of dwindling tax resources resulting from recen t property tax cuts approved by Florida voters. The Children’s Board of Hillsborough County’s Research Department has been charged with helping to determine how th ese tax revenues can be most effectively invested to help ensure the well-being of children and families in the county. Information for strategic decision maki ng comes from many sources and in many forms. Spatial analysis, in the form of ma ps using geographic information systems (GIS) techniques, is one way to examine data. Spa tial analysis is geogra phic in nature and can be a very useful addition to anthropologi sts’ research methods. According to Lang (2002), spatial analysis offers a different pe rspective on data. Spatial analysis, simply put, is the act of comparing the positions of different items or events on a map to identify similarities or relationships. Conditions and relationships no t obvious in a list, table or chart often become readily apparent when displayed on a map. Aldenderfer (1996) notes that GIS/spatial analysis is an important and relatively new method in anthropology. Although mapping and geographical analyses are used by archaeologists (and practitioners in other disc iplines), cultural anthropologists have not fully explored the value and utility of th is technique. Finally, Aldenderfer (1996)

PAGE 13

5 believes that mapping is appropr iate for many anthropological studies. He states that space is an inseparable part of life and societ y and if we want to develop more reliable and robust quantitative ideas to help us unde rstand our ‘place’ in space, GIS [Geographic Information Systems/spatial analysis] has extr aordinary advantages to offer (Aldenderfer 1996). Conceptual Framework The socioeconomic and spatial component s of the research presented here are well-suited to ecosocial theory. Introdu ced by Nancy Krieger in 1994 in conjunction with the Harvard School of Public Health’s Geocoding Project, ecosocial theory uses a multi-level framework to integrate social a nd individual biological characteristics to develop new insights into patterns of hea lth, disease and well-being from dynamic and historic perspectives. According to Krieger (2001), four core concepts define ecosocial theory, including 1. embodiment, 2. pathways to em bodiment, 3. cumulative interplay between exposure, susceptibility, a nd resistance, and 4. accountab ility and agency. The core concept of embodiment refers to how individuals biologi cally incorporate aspects of their material and social environments. Kriege r (2001) contends that all aspects of an individual’s biology must be understood in the context of their history, personal, and cultural ways of living. The second core concept of ecosocial theory, pathways of embodiment is structured simultaneously by society’s arrangements of power and property (patterns of production, consum ption and reproduction) and by human biological constraints and possi bilities (ecological contex t and biological and social

PAGE 14

6 development). The third core concept addresses the cumulative interaction between exposure, susceptibility and resistance These interactions are played out at multiple levels (individual, neighborhood, regional, na tional and international) and in multiple domains (home, work, school, and other public settings, for example) and at multiple scales of time and space. The fourth core concept of ecosocial theory, accountability and agency refers to the way the core concept of embodiment is understood in relation to institutions, communities, households and individuals. Accountability and agency also refers to researchers’ account ability for the theories they use (and choose not to use) as well as explicit consideration of the benefits and limitations of the scale and level of analysis chosen (Krieger 2001). An ecosocial framework is a systematic in tegrated approach that is more than adding individual or biological as pects to a social analysis or adding social factors to a biological study. Strengths of this approach are its capacity to generate new hypotheses (Krieger 2001) and explicitly investigating so cial determinants of population distributions of health, disease, and wellbeing, rather th an treating these determinants as simply background to biomedical phenomena (Krieger 2001). Following an ecosocial framework, the re search conducted for this dissertation investigates adolescent chil dbearing (the biological or i ndividual-level outcome) by examining social determinants (neighbor hood-level variables) using area-based socioeconomic measures (ABSMs). This study us es two main spatial-statistical analyses – hot spot analysis and neighbor hood analysis. Hot spot analys is identifies where rates of adolescent births are statistically significan tly high or low and presents these hot spots and cold spots spatially on a map. Neighborhoodlevel analysis is used to explore small-

PAGE 15

7 area socioeconomic factors and how these f actors geographically a lign to adolescent childbearing. This study also includes a qualitative component. Interviews with teen pregnancy prevention service providers and f unding agencies were conducted to discover factors that may lead to adolescent childbe aring and to discover providers’ opinions on the usefulness and relevance of the results of this type of information for their work. The following research questions are addressed this study: What are the patterns of adolescent ch ildbearing in Hillsborough and Pinellas counties, Florida? Is there a relationship between commun ity-level socioeconomic indicators and adolescent childbearing? Will adolescent pregnancy prevention service providers and funding agencies find this information useful and relevant to the work they do? Purpose of the Study This research contributes to the wo rk some anthropologists have begun in developing an “applied anthropology of GIS/ sp atial analysis.” In her presentation to the 101st Annual Meeting of the American Anthr opological Associati on in 2002, Dr. Susan C. Stonich called for the anthropology of rather than anthropology in spatial analysis/GIS. In other words, Stonich envisioned an anthropology of spatial approaches. Stonich (2002) called for a problem-drive n or hypothesis-driven approach, using theories, concepts, and methods that help to clarify links between anthropology and GIS/spatial analysis. She called for the expansion of “qualitative” approaches and integration of qualitative data with spatial an alysis to investigate, for example, the distribution of health risks or disease rather than having to rely on the perception of the distribution of health risks.

PAGE 16

8 This research begins to fill gaps that Stonich (2002) indicates currently exist between anthropology and spatial analysis. Firs t, this research is problem-driven, rather than data-driven. It uses GIS and spatial-stat istical analyses to i nvestigate patterns of adolescent childbearing in Hillsborough and Pi nellas counties, Florida, as well as the neighborhood-level patterns of socioeconomic inequality in the communities where adolescents live. In additi on, this research contributes to an “applied anthropology of GIS/spatial analysis” by demonstrating how anthropologists who work on more regional scales can now use readily-available and fam iliar technology as part of their toolkits. Finally, interviews with adolescent pregna ncy prevention service providers and funding agencies conducted as part of this dissertati on’s research serve to link the more regional spatial analyses with a qualitative approac h, as Stonich (2002) suggests. Maps produced as part of this study allow service provi ders and funders to view information on adolescent childbearing in a di fferent way which can contribu te to data-driven decisions rather than decisions made on perceptions. Inte rviews also served to provide insights on the usefulness and relevance of this research to service providers and funders and served as a bridge between larger-s cale analytic methods and th e ethnographic technique more familiar to anthropologists.

PAGE 17

9 CHAPTER TWO: REVIEW OF RELATED LITERATURE There are several aspects of my study that can be informed by work already done in the field of anthropology as well as in othe r disciplines. One of the newer techniques to anthropology is the use of Geographic Info rmation Systems (GIS) which has been used by archaeologists to a much gr eater degree than cultural an thropologists. Advances in GIS technology have facilitated new methods in data collection and analysis. This study can also be informed by work done primarily in the disciplines of geography and public health where hot spot analysis and small ar ea analysis (using area-based socioeconomic measures or ABSMs) have been used in a wide variety of wa ys that can inform anthropological investigations. Finally, an enormous amount of literature has been devoted to the topic of adolescent pregnanc y and childbearing. Because this research uses births to adolescents to demonstrat e the utility of the analytic method being developed which can contribute to an “appl ied anthropology of GIS/spatial analysis,” a review of this literature can illuminate and in form the work done for this dissertation in addition to providing ideas for future research. Anthropology and Analysis of Space The study and analysis of space, although most often associated with geography, also plays an important role in anthr opology. In the past, as well as today, the anthropological literature de monstrates consideration of geography, space, and place.

PAGE 18

10 Intermittently throughout its history, anthropol ogical inquiry has explored the relevance of space. Space has held importance in anth ropological theory, cultural perspectives, and ethnographic methods of data collection and re cording. As Aldenderf er (1996) notes, the concept of diffusion and perspectives of eco logical anthropology are two anthropological theories relying on spatial thought. Beginning with a historical examinati on of the role of space in anthropology, Aldenderfer (1996) notes that around the turn of the cen tury much anthropological thought was focused on the notion of diffu sion. Scholars attempted to identify geographic centers of diffused material culture, kinship syst ems, social organization, or political systems, among others, using the concept of spatial proximity to explain similarities and difference between cultures. In the United States, for example, Alfred Kroeber (1939) defined large culture areas such as California, the Eastern Woodlands, and the Great Plains in his book Cultural and Natural Areas of Native North America in an attempt to map out ethnic groups defined by language or similarities in material culture. Interestingly, Aldenderfer (1996) points out that with this approach, the size of the region could be scaled up or down de pending on the problem being investigated, resulting in relatively coarse sets of data for large areas, down to very fine-grained, detailed lists of overlapping traits within small areas. By the 1930s and 1940s, the discipline of anthropology had rejected the notion of diffusion in favor of re-emergent ideas of evol utionary explanations for observed patterns of cultural similarities and differences. Despite the potential for explanatory improvement during this period, Aldenderfer ( 1996:6) contends spat ial thinking “turned

PAGE 19

11 inward” using history, place and region as the primary means of understand cultural diversity, resulting in space becoming “passive and sterile as an analytical concept.” Nevertheless, spatial thinking one again gained importance in anthropology with the growth of ecological theory within the discipline (Aldenderfer 1996). Julian Steward, the first and leading proponent of the cultural ecological pe rspective, demonstrated the relationship between culture and environmen t of a given area. St eward’s 1938 study of the Great Basin Shoshone drew from an ecosystems concept, where geographically referenced human activities varied, at least in part, as a conseq uence of spatial and temporal variations in ener gy availability and its flow through the system. Steward’s ecological approach was subsequently adapted and modified to include other regions and problems. In addition to cultural anthropology’s cons iderations of space, Aldenderfer (1996) argues that archaeologists of th e first half of the twentieth century engaged in similar perspectives. Archaeologists of the time were engaged in determining patterns of diffused cultural traits based on materi al remains found in excavated sites. Methodologically, artifacts such as spear point s were examined, categorized stylistically and temporally, and their spatial distributi on was determined. Th e regional distribution of artifact types established the spatial/temporal cultural boun daries and thus, the cultural history of a region. Aldenderf er (1996) further contends that although this method of defining spatiotemporal cultural boundaries has b een criticized, it is stil l used in the same manner today. During this early period of anthropology, the tools used to delineate space and manipulate data were fairly simple, consisting primarily of maps, map overlays and tables

PAGE 20

12 (Aldenderfer 1996). Data such as culture elem ent distribution lists a nd cultural traits, for example, were tabulated by hand and pl otted on maps. However, by the 1980s, significant technological innovations in Geogr aphic Information Systems had emerged, setting the stage for inquiry on la rger spatial scales and the ab ility to do so in a practical manner. Since then, many archaeologist and some cultural anthropologists have taken advantage of this new technology by incorporati ng spatial data in their research efforts. Anthropology and Geographic Information Systems (GIS) Anthropology’s historic a nd continuing use of mapping and quantitative methods set the stage for the use of GIS and spatial technologies. Cultural anthropologists frequently employ ethnographic techniques in their studies, which can entail mapping spatial relationships and community features as part of the res earch (Greenbaum 1998, Whiteford 2000). Cultural anthropologists also readily employ quantitative data collection and analyses, and test qualitativ e hypotheses using quant itative techniques (Greenbaum 1998). As far back as thirty years ago, arch aeologists began usi ng the relatively new remote sensing (RS), geographic informa tion systems (GIS) and global positioning systems (GPS) technologies as part of thei r systems of data collection and analyses. These emerging techniques were facilitated by developments in computer hardware and software that made the technologies more wi dely available (Conant 1994). In the early years, some anthropologists feared the valu e of fieldwork would be diminished by these emerging technologies. On the contrary, the opposite has been true and fieldwork has actually been encouraged (Conant 1994). However, Stonich (2002) notes that while

PAGE 21

13 anthropologists, in general, have embraced these new technologies’ data collection and analytic methods, most spatial analyses have been done by archaeologists. Geographic Information Systems (GIS) a nd Remote Sensing (RS) applications have become widely used tools in archaeol ogy. Archaeologists, such as Eric E. Jones (2006), have studied settlement patterning at a regional scale. Jones (2006) used viewshed analysis, a GIS-based method, to determine how the natural and political landscapes affected choices in settlement location among the Late Woodland and early historic Onondaga Iroquois, showing how both pr oductive soils and settlement defensibility entered into choice of settlement locations. His work demonstrated how GIS can surpass statistica l analyses in helping to unders tand behavior by using spatial modeling. Other archaeologists have used GIS tec hniques on a more micro-scale. For example, Abe et al. (2002) examined cutmar k patterns from assemblages with differing levels of fragmentation within a small ge ographic area by analyzi ng the frequency of cutmarks over the observed geographic area usin g the image analysis capability of GIS. To a lesser degree than archaeologists, cu ltural anthropologists have also used GIS techniques to investigate contemporary is sues and phenomena. For example, Silltoe (2002) investigated the reasons for failure of development efforts in Bangladesh by integrating indigenous knowledge with a ne wly-developed comput erized model using databases and geographic information sy stems (GIS) to integrate quantitative environmental information. Other recent anthropological i nvestigations of the effect s of space and distance in modern cultural interactions have also us ed GIS spatial technologies. Futemma and

PAGE 22

14 Brondzio (2003) used remote sensing data and GIS mapping technologies to understand changes in land use in the Lower Amazon over time and the implication for agricultural intensification and forest conservation. C oppolillo (2000) investig ated the ecological impacts of and on pastoral grazing by determ ining associations of space and distance on cattle productivity and herding practices in East Africa, including the effects of traveling further from home, maintaining large herds, and the impact of high settlement densities on herding practices. Although the use of GIS/spatia l analysis had been increa singly used in cultural anthropology in the years since its emergence, several challenges and gaps still exist. Loker (2005) contends that although the comple x challenge of descript ion and analysis of human-environmental interactions has been aided by newer spatial technologies, concepts and methodological innovations, this task is st ill difficult, for two primary reasons. First, although computer-assisted GIS an alysis is a powerful tool, it has not yet been able to accurately characterize social and natural systems, which are extremely complex and variable in both space and time. Also, Loker (2003) notes that envi ronmental data tend to be continuous in nature while social data tend to be point data, such as individuals, households, villages, etc., ma king linkages a difficult task. In addition to the challenges pointed out by Loker (2003), Stonich (2002) notes that most work done by cultu ral anthropologists has been narrowly focused on land use and the effects of climate change. Stonich ( 2002) also contends that work tends to be data driven rather than problem or hypothesis driven. The notable exception to this is the work done by Romero-Daza (2004) where she us ed cluster analysis and GIS technology to locate geographic areas with statistically significant high levels of low birth weight

PAGE 23

15 births in Hillsborough County, Fl orida. Finally, Stonich (2 002) asserts that although cultural anthropology currently uses spatial analysis/G IS techniques and technologies, the discipline has not yet fully developed an “appl ied anthropology of spatial analysis/GIS.” Although relatively little work has been published by cultural anthropologists on their use of GIS/spatial technologies, ot her disciplines such as public health, epidemiology and medical geography have more readily embraced these newer techniques, including hot spot analysis and mu ltilevel modeling using GIS technologies. Hot Spot Analysis Hot spot analysis gained increased atten tion in the 1980’s as a result of growing concern about adverse environmental effect s on population health (Lawson 2001). Since then, spatial epidemiology began to use special statistical methods to identify clusters of chronic diseases and conditions. Clustering, according to Lawson (2001:104) is “any area within the study region of significant elevated risk.” He con tinues by noting that this definition makes no assumption about shape or extent, but would qua lify as a cluster provided the area meets some sta tistical criteria. This defini tion, often referred to as hot spot clustering, is a way of cl assifying or grouping where cl usters are groups of highly similar entities (Brimicombe 2005). Closely aligned with the research conduc ted in this dissertation, the California Department of Health used GIS in a problem -driven approach aimed at more effective delivery of public health services and resour ce allocation. The project used hot spot analysis to locate census tracts where teen births were clustered. The authors used a Poisson distribution test (for areas with less than 100 births) and a Chi square test (for

PAGE 24

16 more than100 births) to calcul ate significance. The maps that were produced were used to justify allocation of funds to areas of greatest need for t een pregnancy services (Taylor and Chavez 2002). The authors note that, “the use of GIS to improve the ability to more effectively deliver public health services is in its infancy” (Taylor and Chavez 2002:33) and offer several experiences with their project that may have partic ular relevance to the work being done in this dissertation. First, ZI P Codes were not a reliable geographic level because ZIP Code boundaries changed, therefor e precluding aggregation of several years of birth data. Next, although census tracts worked well for hot spot analysis, public health staff often had trouble situating themselves in relation to the census tracts. To overcome this problem, landmarks such as schools and major roads were included on maps. Finally, when analyzing births for th e entire State of California, it was often difficult to present enough detail in maps (ro ads, schools, program site locations, etc.) without the map appearing cl uttered and hard to read. Another study, which used similar analys is techniques to those employed in this dissertation, was conducted in California by Goul d et al. (1998) to identify areas of the state in need of pregnancy prevention and pr enatal care programs. Using birthrates for 15 to 17 year-old adolescents by ZIP Code fo r 1992 to 1994, the authors identified ZIP Code hot spots where birthrates to teens were si gnificantly higher than the state mean. Working with individuals identified by the Ca lifornia State Office of Family Planning to be part of a state-wide advi sory board, hot spots were aggr egated into potential project areas based on the number of births and clini cal experience of sta ff within each area. Finally, potential proj ect areas were described by fact ors that would affect program

PAGE 25

17 design and implementation, including raci al and ethnic composition, socioeconomic characteristics (education, median income, poverty, female-headed households and unemployment) and characteristics related to teenage pregnancy (repeat adolescent births, late or no prenatal care, low birth weight births, Medicaid-pai d births, and fathers younger than 24 years old). Gould et al.’s (1998) study identified 210 hot spots, and based on input from local area providers and professionals, these hot s pots were divided into 82 potential project areas. To assist in adolescent pregnancy-re lated program planning, the potential project areas were categorized as high or low prio rity based on birth rates, demographic and socioeconomic characteristics of the areas and staff case loads. The authors state that the results of this study will help to ensure the most judicious use of limited resources. A number of studies have used cluster analysis, another method of detecting hot spots which does not rely on predetermined areas such as census tracts or ZIP Codes. For example, Publiatti et al. (2002) conducted a spatial analysis of the prevalence of multiple sclerosis (MS) to discover hot spots of the disease in the province of Sass ari in northern Sardinia. Results showed a clustering patte rn in the southwestern communes where an MS epidemic is suspected to have previ ously occurred, leading the authors conclude that because MS is not a single-source infectio us disease. Their study may help test the hypothesis that a widely and evenly-spread environmental agent may produce disease in subgroups of genetically more susceptible individuals. Similar to hot spot technique s used in the study above by Publiatti et al. (2002), English et al. (2003) investigated changes in low bi rth weight births to test whether rapid population growth, economic pressure and ne ighborhood instability in the communities

PAGE 26

18 near the US/Mexico border aff ected reproductive health. This study (English et al. 2003) found geographic hot spots that sh owed statistically significant increases in full-term and pre-term low birth weight rates, providing an illustration of how hot spot cluster analysis can be used as a method to explor e incidence patterns in a region. Hot spot analysis has also been used to study births to adolescents in Washington D.C. (Johnson-Clarke 2000) to track changes in childbearing over time, as well as a wide variety of other issues, including environmen tal issues such as carbon monoxide levels (Meng and Niemeier 1998), biodiversity (P odolsky 2003), crime (Eck et al. 2005), and housing (Wang and Varady 2005). The wide va riety of applications using hot spot analysis demonstrates the fl exibility of this approach. Hot spot analysis has several strengths and the information generated can be used in a number of ways. First, maps of hot s pots convey spatial information to visually identify high risk groups in crit ical areas and to determine shifts in the distribution of teen births (Johnson-Clarke 2006). Hot spot analysis has also been used to target interventions and for resources allocation (Gould et al 1998, Johnson-Clarke 2006, Taylor and Chavez 2002). Johnson-Clarke (2006) also notes that hot spot analysis and GIS techniques can be used for public health surveillance a nd monitoring, health-related policymaking, and for tracking racial/ethnic or economic disparitie s in health. Gould et al. (1998) state that hot spot analysis has potentia l for measuring specific outcomes of program interventions over time and for comparing communities that are testing the efficacy of different types of interventions (Gould et al. 1998). Finally, Johnson-Clarke (2006) notes that hot spot analysis is particularly useful in genera ting research hypotheses. As Rushton (2003:65) explains, “Exploratory analyses investigate alternative under standings of the pattern of

PAGE 27

19 disease in any region with the aim of selecting more appropri ate subsequent analyses that might resolve the ambiguities that typically aris e in the early states of investigations.” The idea of hot spot analysis as an exploratory device lead s to a discussion multilevel modeling, an explanatory method that can be used in conjunction with hot spot analysis. Multilevel Modeling and Ecosocial Theory As discussed in the previous chapter, ecosocial theory was introduced by Nancy Krieger in 1994 in conjunction with the Ha rvard School of Public Health’s Geocoding Project. Ecosocial theory uses a multi-level fr amework to integrate social and individual biological characteristics to develop new insights into patte rns of health, disease and well-being from dynamic and historic perspe ctives. Ecosocial theory emphasizes the social aspects and production of health (or disease) and well-being and, rather than focusing solely on individuals, it positi ons health and well-being as collective phenomena. By examining both individual and group-level processes, this theory encourages multi-level analysis by consider ing the combined effects the neighborhood and the individual (Krieger 1994). Ecosocial theory grew from what Kriege r (1994) saw as a major inadequacy in the “web of causation” metaphor that was (and still is) popular in epidemiology. The web of causation metaphor was developed by MacMahon et al. in 1960 and uses the image of a spider’s web with multiple inters ections in the web representing specific risk factors or outcomes and the strands in the web symbolizi ng diverse causal pathways. And although this metaphor encourages inves tigation of the interaction between multiple

PAGE 28

20 causes and multiple effects, the model omits di scussion of the origins of multiple causes, what Krieger (1994:891) calls a “spiderless web.” It also fails to differentiate between the causes of cases (individua l risk) and the cause s of incidence (population risk). Krieger (1994) states that the ‘web’ sans ‘spider’ discour ages consideration of why population patterns of disease and well-being exist and why these patters persist or change over time. In other words, the web of causation views group patterns of disease/well-being as merely an aggregati on of individuals’ tra its and choices. Ecosocial theory, according to Krieger (1994:896), strives to embrace “population-level thinking and rejects th e underlying assumptions of biomedical individualism without discar ding biology.” She envisions ecosocial theory as “an evolving bush of life intertwine d at every scale, micro and ma cro, with the scaffolding of society that different core so cial groups daily reinforce or seek to alter” (Krieger 2001:671). For anthropologists, ecosocial th eory is appealing because it facilitates consideration of culture, vi ewing culture as indivisible from political, economic, religious, biological and social aspects of daily life, as well as allowing for cross-cultural comparison of disease/well-being. This multi-dimensional perspective of disease and well-being is especially wellsuited to multi-level modeling techniques. Multilevel modeling disentangles the two levels of variation (individua l-level and neighborhood-level) to distinguish variations due to characteristics of an area from characteris tics of individuals who live in these areas (Krieger et al., N.d.). Multile vel modeling, according to H ox (1998:147), is a family of statistical procedures “specifica lly geared toward the statistical analysis of data that have a hierarchical or clustered structure.” Jones and Duncan (1998:95) also note that

PAGE 29

21 multilevel models serve as a useful tool for “quantitative analysis when the problem under investigation has a multilevel structure, when a process is thought to operate at more than one level or scale, or when the researcher is particularly interested in variability and heterogeneity and no t just overall average values.” Data are conceptually cons idered to occupy different hierarchical levels for purposes of analysis. In other words, multilevel approaches use data sets of individuallevel variables nested within geographic ar eas, such as neighborhoods, with associated neighborhood-level (contextual) variables (D iez-Roux 2001). For example, individuallevel data is available in bi rth records, including demogra phic data (e.g., age of mother, race/ethnicity, marital status, address, prenatal care utilizat ion, parity, etc.), whereas neighborhood-level variables usually in cludes aggregated information (e.g., demographic, housing, economic data, etc.). Diez-Roux (2001; 2003) continues, noting that multilevel regression equations with individuals as the units of analysis can simultaneously include both individual-level and neighborhood-level predictors, allowing examination of neighborhood or area effects afte r individual-level c onfounders have been controlled. Multilevel analysis also allows ex amination of individual-level characteristics as modifiers of the neighborhood effect, and vice versa. Finally, multilevel analysis permits simultaneous examination of with in and between neighborhood variability in outcomes, as well as the extent to which between-neighborhood vari ability is explained by both individual-level and neighborhood-level factors. O’Campo et al. (1997) assert that multilevel models have several advantages. First, multilevel analytic methods are more consistent with social theories’ more traditional methods of analysis because they explicitly accommodate multiple levels of

PAGE 30

22 data. Next, multilevel models allow for inclusion of macro-level factors in current explanatory models, which can increase our understanding of how neighborhood factors affect differences in individual risk. Add itionally, multilevel analysis helps to eliminate potential confounding of individual-level e xplanatory models due to omission of neighborhood factors. Finally, O’Campo et al (1997) contend that multilevel approaches can inform design of effective intervention strategies by improving our understanding of how neighborhood factors influenc e individual health outcom es. Diez-Roux (2003) adds that multilevel approaches provide alternativ es to the individual/population dichotomy. The strength of multilevel models lies in their ability to integrate individual and ecological study designs while avoiding the li mitations of each of these designs. The first step in generating a multilevel model is to determine individual and neighborhood variables relevant to the probl em or phenomenon being investigated. While vital statistics data contains individual level data (e.g., age, race/ethnicity, cause of death, gender, marital status, etc.), most vi tal records lack socioeconomic data, making it difficult, if not impossible, to investigate the role that social position or economic deprivation, for example, play s in individual-level outcomes (Krieger et al. 2003a). The solution to this paucity of neighb orhood-level socioeconomic data, according to Krieger et al. (N.d.), is to geocode the vital statistics or public health su rveillance data and then use U.S. Census area-based socioeconomic measures (ABSMs) to ascertain both the cases and the population in a given area. In multilevel models, appropriate ABSMs, are usually determined through literature on th e issue being investig ated, and are applied both independently and together in the form of an index. An index is a set of indicators, combined in a standardized way, that summarizes complex or multi-dimensional

PAGE 31

23 characteristics of a geographical area or high light what is happening there. An index provides the big picture, making indicators easier to interpret th an trying to find a trend in many separate indicators (Saisana and Ta rantola 2002). Some commonly used public health indices measure material depriva tion (e.g., Carstairs Index, Townsend Index), income inequality (e.g., Robin Hood Index) and access to material resources (e.g., U.S. CDC Index of Local Economic Resour ces) (Krieger et al. N.d.). Area-based socioeconomic measures (A BSMs) are calculated within small geographic areas, and as mentioned, are used in dependently and together in the form of an index. The advantage of cal culating measures on smaller scales is that populations within small areas tend to be more homogene ous and therefore a wi der range of social and economic variation between areas can be distinguished. Usi ng data on population and housing characteristics from the U.S. Cens us, such as census tracts, block groups or ZIP Codes, facilitates establishm ent of the relationships between these small-area variables and the health or vital st atistics data (Carstairs 1981). A lot of discussion has been generated around appropriate development and use of area-based socioeconomic measures (ABSMs), which are the neighborhood or socioeconomic-level aspect of ecosocial studi es. According to Elliott and Wartenberg (2004), analysis of data at a local or small-area scale pr esents some unique challenges that must be considered. First, the quality of the data can present problems that must be addressed. The effects of missing variables, re liability and validity of the data become magnified at a small-area level of analysis. Another set of considerations directly relates to the scale of analysis. Sm all areas or groups of areas narrow the investigation, which reduces the size of the popul ation at risk within give n boundaries, leading to small

PAGE 32

24 numbers of events or unstable risk estima tion. Also, studies with small populations are more sensitive to errors and local variations in the quality of both the health (numerator) data and the population (denominator) data than studies conducted ove r larger areas. If undetected, local variations in data quality coul d lead to serious bias es. Next, small area studies are susceptible to confounding, whic h can result in false exposure-disease associations. Finally, data can be misinterpr eted as individual-lev el socioeconomic data, rather than being used as complimentary ne ighborhood-level data th at can be analyzed together with individual level data in multilevel models (Elliott and Wartenberg 2004, Krieger et al. 2003a, Krieger et al 2005, Taylor and Chavez 2002). In spite of the limitations, Krieger et al. (2003a) stat e that area-based socioeconomic measures (ABSMs) have severa l strengths. First, ABSMS can be used with any database that includes addresses, as is the case for of the birth data used in my study. Next, ABSMs provide data for determ ining neighborhood (contextual) effects on health that goes beyond effects that are due to individual-level socioeconomic position. Finally, ABSMs are equally applicable to all people, regardless of factors such as age, gender, and employment status (Krieger et al. 2003a). The ecosocial framework used in this dissertation’s res earch design views biological/individual f actors and social/ neighborhood fact ors as integrated. Area-based socioeconomic measures (ABSMs) provide the context and the Florida Vital Statistics Birth Records provide individua l-level data on where births to adolescents, by race and ethnicity, occurred. Although not part of th is dissertation’s research, information regarding birth outcomes, to both the adolescent mother and her baby, can provide

PAGE 33

25 insights into the causal relationships between individual factors and possible neighborhood risk factors associated with adolescent childbearing. Risks and Outcomes of Adoles cent Pregnancy and Childbearing To get an idea of the magnitude of teen childbearing, the Robin Hood Foundation’s (Maynard 1997) repor t on adolescent childbearing, Kids Having Kids estimates that nearly one million teenagers in the United States become pregnant each year. This is approximately ten percent of a ll girls ages 15 to 19 years old. An estimated 48 percent of these pregnancies are not carried to term, but rather are terminated or end in miscarriage. Over half (52 percent – about half a million teens) will bear children each year, according the report (M aynard 1997). More than 175,000 of these new mothers in the United States are under 18 years old. In 2000, the teen birth rate for adolescents ages 15 through 19 in the United States was at 47.7 per 1,000 females, down from 60.3 per 1,000 in 1990 (Guttmacher Institute 2004). This decline was observed in all racial and ethnic groups. Birth rates for African American adolescents declin ed from 112.9 (per 1,000 females ages 15 to 19) in 1990 to 77.4 in 2000. Similar decreases in teen birt h rates are found among white teens where birth rates fell from 51.2 to 43.2 (per 1,000 fe males ages 15-19) and among Hispanic adolescents where birth rates declined from 99.5 to 87.1 during the period from 1990 to 2000 (Guttmacher Institute 2004). Nevertheless, th e teen birth rate in the United States is the highest of any industrialized nation, n early twice as high as the United Kingdom, which has the next highest te en birth rate (Maynard 1997).

PAGE 34

26 In Florida, the birth rate was 51.0 per 1,000 for all births among 15-19 year olds, a ranking of number 16 out of the 50 states in 2000. The birth rates in Florida for white, African American and Hispanic teens in 2000 were 37.0, 82.0 and 58.0 respectively (Guttmacher Institute 2004). Table 2.1 belo w compares national and State of Florida birth rates among adolescents ages 15 -19 in 2000 by race and ethnicity. Table 2.1 Birth Rates for 15-19 Year Olds in 2000 by Race/Ethnicity White African American Hispanic Total Florida 37.0 82.0 58.0 51.0 United States 43.2 77.4 87.1 47.7 Sources: Guttmacher Institute 2004, Florida Department of Health 2005 As the table shows, Florida had a lower birth rate among white and Hispanic teens than the nation as a whole, while Florida’s African American and ove rall birth rates were higher than the United States in 2000 among 15-19 year-olds. To put these statistics in context, we must look at the meaning of the numbers in terms of maternal and child outcomes. A dolescent mothers under age 20 are more likely to experience adverse medical outcomes, both during pregnancy and later in life. And the younger the mother, the greater the risk that she and her baby will experience health complications, primarily due to inadequate prenatal care, poor nutrition, and other lifestyle factors (March of Dimes 2004). Maternal Risks and Outcomes Like their more mature counterparts, pr egnant adolescents ar e at risk of poor perinatal obstetric outcomes, including pre gnancy-induced hypertension, anemia, and pre-term labor and delivery. However, there is disagreement in current literature about the extent to which the age of the adolescen t mother versus pre-pregnancy and pregnancy

PAGE 35

27 behaviors and risk contribute to poor childbearing outcomes ( Ehrlich and Vega-Matos 2000). For example, Ananth et al. (2001) assert that age is at least partia lly linked to biological factors that increase risk of pret erm delivery at younger ages, while Geronimus (2003) contends there is no difference in birth outcomes between adolescents and nonadolescent mothers when outcomes are compared within racial and ethnic groups. One series of related disorders that can a ffect pregnant adolescents are gestational hypertensive disorders. One type of ge stational hypertension is preeclampsia, a temporary condition in which high blood pressu re develops in women with previously normal blood pressure after the 20th week of pregnancy and returns to normal after delivery. Typically, preeclampsia affects at leas t five to eight percent of all pregnancies, increasing the risk of preterm delivery (Ana nth et al. 2001). St udies by Galvez-Myles and Myles (2005) and Eure et al. (2002) found younger teens were significantly more likely to develop preeclampsia than ol der teens and women over 20 years old. Pregnant adolescents are also at risk of developing gestational diabetes, characterized by high glucose levels during pregnancy among wome n without a history of diabetes (American Diabetes Association 2006). A retrospective review of pregnancies at the Milwauk ee Regional Medical Complex showed that gestational diabetes disproportionately a ffected African American wo men, occurring in about one percent of pregnant African American adoles cents and almost five percent of pregnant African American women over age 30 years old (Lemen et al. 1998). There is a strong association between obesity and diabetes me llitus in pregnant women (Nuthalapaty and Rouse 2004) and women with gestational diabet es are at high risk of developing type-2 diabetes in the future (Turok et al. 2003). Unlike the small for ge stational age infants

PAGE 36

28 associated with preeclampsia, infants born to mothers with gestational diabetes are often large for gestational ag e (Turok et al. 2003). According to the U.S. National Library of Medicine and the National Institutes of Health (2009), gestational ag e is a measurement of fetal development. Small for gestational age means a fetus or infant is smaller than normal for the baby's gender and gestational age, the period of time between conception and birth during which the fetus grows and develops inside the mother's womb. Large for gestational age means a fetus or infant is larger than normal based on gender and gestational age. Gestational age is the number of weeks measured from the first da y of a woman’s last menstrual cycle to the current date. Normal pregnancies range from 38 to 42 weeks. Also called fetal age, gestational age can be determined with ultr a sound before that baby is born by measuring the head size, abdomen and thigh bone. After the baby is born, developmental gestational age is determined by measur ing the infant’s weight, length, head circumference, hair and skin condition, reflex es, muscle tone, posture, and vital signs. A baby born small for gestational age is small in size and is at higher risk of low body temperature, low blood sugar and increased red blood cells, while a baby born large for gestational age is at higher risk of birth injury and complications from low blood sugar. (U.S. National Library of Medicine and the National In stitutes of Health 2009). Pregnant women are especially vulnerabl e to anemia during pregnancy due to increased iron demands. A ten-year retrospe ctive chart review of pregnant African American adolescents at an i nner-city clinic by Ch ang et al. (2003) showed that African American women are at high risk, and African American adolescents have the highest risk of developing anemia during late pregna ncy. Their research also showed that the

PAGE 37

29 adolescents in the study were at risk for several adverse birt h outcomes. Teens with low hemoglobin concentrations (diagnosed with anemia) in the third trimester of their pregnancy had increased risk of preterm de livery and delivering a low birth weight baby (Chang et al. 2003). A low birth weight infant wei ghs less than 2,500 grams (5 pounds, 8 ounces) at birth and is at in creased risk of infant mo rtality, Sudden Infant Death Syndrome (SIDS), blindness, deafness, respirat ory difficulties, mental illness, retardation, cerebral palsy, dyslexia, a nd hyperactivity (National Ca mpaign to Prevent Teen Pregnancy 2007). Another pregnancy-related risk to adoles cents is delivery by ces arean section, that is, delivery of a baby through surgical inci sion in the mother’s lower abdomen wall and uterus, which poses risks for both mother a nd baby. Several conditions may necessitate birth by cesarean section, in cluding prolonged or ineffec tive labor, placenta previa, placenta abruption, abnormal presentation, fetal distress, medical problems and multiple births, among others (American College of Surgeons 2006). In addition to a longer healing period, mothers who have had surgical deliveries are also at risk of increased bleeding, infection, endometriosis, reaction to anesthesia, injuries to the bladder and blood clots. Infants are at risk of prematurity, reactions to th e effects of anesthesia, fetal injury and are more like to have breathi ng problems than babies delivered vaginally (Mayo Clinic 2004). While Amini et al. (1996) found cesarean deliveries for young teens ages 12 to15 years old was significantly higher than for older teens (16-19 years old) and adults (age 20 and older), studies by Eure et al. (2002) and Galvez-Myles and Myles (2005) found that teens were significantly less likely than adult women to deliver by cesarean section.

PAGE 38

30 Although most teen mothers only have one child during their teenage years, some adolescent mothers go on to b ear another child. Nationally, about 17 percent of teen mothers have a second baby within three years after the birth of thei r first baby (National Center for Health Statistics 2001). From a bi ological perspective, Smith and Pell (2001), in their population-based retrospective cohort study, found that second teenage births are associated with an almost threefold increase in risk of preterm delivery and stillbirth. There are several factors that place teenag e mothers at risk of delivering a second child within two years of her first birth, including not using lo ng-acting contraception (Depo Provera, Norplant or IUD) within the first three months of delivery, plans to have another baby within five years of the first child, not being in school within three months after delivery, experiencing intimate partner vi olence, not being in a relationship with the father of the first child, the father of the firs t child being more than three years older than the mother, and having many friends who were also adolescent parents (Raneri 2006). Jevitt (1983) found longer-term participation in a teen service progr am was a protective factor against repeat pregnancy. In her study of the Teen Service Program at Grady Memorial Hospital she found that 79 percent of teens did not become pregnant while enrolled in the program. In addition, program participants had lower neonatal death rates after participating the Teen Services Program. Neonatal and Child Health Risks and Outcomes In addition to adolescent mothers fa cing increased risk of poor pregnancy outcomes, their offspring are also at risk of poor outcomes, including preterm birth, low birth weight, and still births. One compre hensive study by Amini et al. (1996) clearly

PAGE 39

31 illustrates the high rate of poor birth outcomes for teen mothers. Although this research was conducted more than a decade ago, it is on e of the few that investigated several biological and psychosocial birth outcome s and also conducted a comparison of outcomes between age-groups. In a 19-year prospective study of pregnancy in adolescent women in Cleveland, Ohio, Amini et al. ( 1996) compared the obstetric outcomes of young teens (12-15 years old), older teens ( 16-19 years old) and adults (age 20 and older). Of the 69, 069 births in the Metro He alth Medical Center from January 1, 1975 to December 31, 1993, 1,875 (2.7%) were to young teens and 17,359 (25.3%) were to older teens. To test the hypothesis that teens have different characteristics and obstetric outcomes, the researchers examined variable s including age, race and ethnicity, insurance status, mode of delivery, gravidity, prenatal visits, admission gesta tional age and use of alcohol, tobacco and narcotics (Amini et. al .1996). Outcomes included infant status, birth weight, and APGAR scores. Using parametric analyses and analyses of covariance for continuous variables, X2 tests for categorical variables to model multi-level associations, and regression and time series analyses for testing and modeling tr ends, the authors found that young teens fared worse in maternal and child outcomes than di d older teens and adu lts, and that African Americans of any age had worse outcomes than either white or Hispanic mothers and babies (Amini et. al.1996). In sum, this study found: Cesarean deliveries for young teens ages 12 to15 years old we re significantly higher than for older teens and adults (11.6%, 9.4%, and 10.2% respectively). Overall, 17.3% of young teens delivered early, a significantly higher rate of preterm deliveries (<37 weeks gestation) than older teens and adults.

PAGE 40

32 For all deliveries, both y ounger and older teen mothers were 1.2 times more likely than adults to deliver a very low birth weight infant (1500 grams or less) after adjusting for delivery year, race, parity, pa tients’ health care insurance status, and marital status. The proportion of infants with one-minute APGAR scores of less than 5 was highest among young teens (14.1%) compared to the proportion of older teens (12.6%) and adults (13.3%). There was no statistically significant di fference in the number of still births between the three groups (1.7 per1,000 birt hs). However, the antepartum fetal death rate and the in-hospital neonata l mortality rates were higher among young teens 12 to 15 years old (4.4 per 1,000 birt hs) compared to older teens (3.5 per 1,000 births) and adults (6.2 per 1,000 births). Smoking, alcohol and narcotic use among young teens was significantly lower than the other two groups. Smoking was most prevalent among older teens while alcohol and narcotic use was signi ficantly greater among adults. More than 95% of teenage mothers did not have private health insurance. This is significantly higher than in the adu lt population (81.6%), indicating a lower socioeconomic status among teen mothers than the overall patient population. After adjusting for insurance status an d year of delivery, African American females were 1.2 times more likely to have a teenage birth than white females. In addition, infants born to African American mothers of any age had significantly longer hospital stays than either white or Hispanic infants. Other research supporting Amini et al .’s (1996) study is discussed below. One of the most frequent, and potentially harmful, outcomes among newborns of teen mothers is a low birth weight infa nt, less than 2,500 grams at birth (5 pounds, 8 ounces). Infants born with low birth weight, or born prematurely, have an increased risk of infant mortality, Sudden Infant De ath Syndrome (SIDS), blindness, deafness, respiratory difficulties, mental illness, reta rdation, and cerebral palsy. Additionally, the risk of later diagnosis of dyslexia and hyperactivity is nearly twice as high among low birth weight infants (National Camp aign to Prevent Teen Pregnancy 2007).

PAGE 41

33 According to the Maternal and Child Hea lth Bureau (2004), low birth weight is one of the leading causes of neonatal mortal ity (within 28 days of birth). Low birth weight infants are more likely to experience l ong-term disability or to die during the first year of life than are infants of normal weight. Vital statistics reports on birth outcomes in 2000 shows that as a mother’s age increases the frequency of low birth weigh infants decreases (Martin et al. 2002). Other studies have shown that a baby’s bi rth weight can influence aspects of their future life, including obesity, diabetes and in telligence. Low birth weight babies and babies born large for gestational age are at increased risk of obesity in later life (Parsons et al. 2001, Singhal et al. 2003). Additionally, lo w birth weight babies are at increased risk of developing Type 2 diabetes (Ric h-Edwards 1999), and studies have found an association between normal birth weight and higher intelligence (Matte 2001, Reichman 2005). A retrospective cohort study by Phipps et al (2002) showed that infants born to young teens, ages 15 years and younge r, are at increased risk of death within the first year (post neonatal mortality) compared to infants born to older mother, ages 23 to 29 years old. These findings were consistent acr oss all racial and ethnic groups. The literature also shows that infants of adolescent mothers are at risk of preterm births. Preterm birth, according to Akinbami et al. (2000), accounts for most low birth weight births in the United States and there is a clear relationship to infant morbidity and mortality. Preterm births (births occu rring before the end of the 37th week of pregnancy) account for over 60 percent of low birth weight babies. Teenage mothers, as an ageperiod-cohort study by Ananth et al. (2001) show ed, are at high risk of preterm birth. In

PAGE 42

34 addition, the study showed African American adolescents had higher rates of preterm births (about 17 percent of teen births) than their white c ohort (about 10 percent of teen births). In their population-based retrospectiv e cohort study, Smith and Pell (2001) found that the risk of preterm birt h and still birth almost tripled in second births to mothers between the ages of 15 and 19 years old. Si milarly, Akinbami et al.’s cross-sectional analysis of U.S. Natality Files for the year s 1990 to 1996 investigat ed risk of preterm birth to multiparous mothers from 10 to 20 years old. This study found a decreasing adjusted odds ratio as age of the adolescent in creases. However, clear racial disparities exist. African American multiparous teens had nearly twice the percentage of preterm births as white and Hispanic multiparous teen mothers. In fact, 25 year-old African American mothers had about the same percenta ge of preterm births as 15 to 17 year old white and Hispanic mothers (Akinbami et al. 2000). Teen mothers are also at risk of a still birth. In a nationwide inpatient sample for the years 1995 through 2002 with 5,874,203 deliver ies identified for analysis, Bateman and Simpson (2006) found that women who were under 20 years old were more likely to have a pregnancy outcome of stillbirth (odds ratio, 1.11; 95% CI, 1.08–1.14), as were women who were 35 to 39 years old (odds ratio, 1.28; 95% CI, 1.24–1.32). Extremes of maternal reproductive age predicted a higher ri sk of stillbirth and this effect persisted even after adjustment for several maternal, placental, and fetal risk factors (Hughes and Riches 2003). Being born to a teenaged mother puts child ren at risk of several adverse outcomes later in life. Children of t eenage mothers are more likely to be poor, abused, or neglected

PAGE 43

35 than those of women who dela y childbearing, and th ey are less likely to receive proper nutrition, health care, and cognitive and so cial stimulation (Annie E. Casey Foundation 1998). Maynard’s (1997) study showed that ch ildren born to teen mothers are also at greater risk of lower inte llectual and academic achievem ent and social behavioral problems. Children of teenage mothers are almost three times more likely to be incarcerated during their adolescence or early 20s than the children of older mothers. These children also are less likely to gra duate from high school, more likely to be unemployed and to become teenage parents th emselves than children born to women who delay childbearing (Maynard 1997). Behavioral Risks and Outcomes There are several behaviors that research ers agree affect pregnancy outcomes. These include obtaining adequate prenatal care, proper nutrition, alcohol, drug and tobacco use during pregnancy and risky sexual behavior. Early and adequate prenatal care is critical to improving maternal perinatal outcomes. Delaying prenatal care places pre gnant women at risk of timely preventive care that can detect complica tions of pregnancy which resu lt in maternal and/or fetal morbidity and mortality. Women who delay pr enatal care (entry af ter the first 12 weeks of gestation) or receive no prenatal care ar e three times more likely to have a low birth weight infant than women receiving adequa te prenatal care (An achbe and Sutton 2003). Of all the maternal age groups, pregnant adolescents are the least likely to get early and adequate prenatal care. In 2002, almost seven percent of pregnant teens between 15 and 19 years old received late or no prenatal care. This compares to less than

PAGE 44

36 four percent of pregnant wo men for all other ages (Guttm acher Institute 2003, March of Dimes 2004). Akinbami et al.’s (2000) na tional cross-sectional analysis found that multiparous adolescents were more likely to ob tain late, inadequate or no prenatal care when compared to 25-year-old multiparous women. Another behavioral factor associated with birth outcomes is proper nutrition. Recent literature has devoted a considerable amount of discussion to the topic of nutrition. The American College of Obstet ricians and Gynecologists (ACOG) (2008) recommends that pregnant women eat a well balanced diet and increase the number of servings of a variety of foods from each of the six basic food groups. For most women, this is about 300 extra calories a day (American College of Obstetricians and Gynecologists 2007). Although ACOG reco mmends taking folic acid supplements several years prior to conception, vitamin a nd mineral supplements during pregnancy are not recommended unless advised by the doctor. Guidelines for pregnancy weight gain have been highly controversial in the United States over the past 50 years. In th e 1960s, women were encouraged to gain only about 15 pounds during their pregnancy. Ho wever, by the 1970, obstetricians began challenging these guidelines after they came to realize that the practice of severely restricting weight gain during pregnancy was associated with increased risk of preterm birth and low birth weight (Abrams et al. 2000, Barwarsky et al. 2005, Howie et al. 2003). This recognition led to a report by th e Institute of Medicine (IOM) in 1990, and later in 2009, recommending pregnancy wei ght gain based on pre-pregnancy body mass index rather than just weight gain. Subseque nt studies have shown that pregnancy weight gain within the IOM ranges generally l eads to better outcomes for both mothers and

PAGE 45

37 infants, and maternal weight gain outside th e IOM parameters are a ssociated with twice as many poor pregnancy outcomes (Abrams et al. 2000), including gestational diabetes (Turok et al. 2003), cesarean deliveries (Galti er-Dereure 2000), shoulder dystocia (Jevitt et al. 2008), low Apgar scores, macrosomia, and neural tube def ects (Galtier-Dereure 2000). In addition, mothers who are overweight are less likely to breast feed, and more likely to have delayed milk production and to early cessation of breast feeding (Jevitt et al. 2007). In one national study comparing adoles cent and adult maternal weight gain, Howie et al (2003) found teens were more likel y to gain an excessi ve amount of weight during their pregnancy than older non-a dolescent women. The authors found that excessive weight gain is most likely attributable to the a dolescent’s need for nutrition to meet her own growth needs as well as the n eeds of her growing fetus. The study also found that younger mothers were most likely to gain excessive weight during pregnancy, and the risk of high maternal weight gain decreased as maternal age increased. There is evidence that these teens tend to retain the excess weight postpartum and are more likely to retain some of this weight and continue to gain weight with each subsequent pregnancy putting them at even greater risk and at an even younger age of hypertension, heart disease, diabetes and some t ypes of cancers. Excessive weig ht gain was most often found among non-Hispanic teens. The investigators sp eculate that this is due to inadequate prenatal care and nutritional advice (Howie et al. 2003). As Jevitt (2005) notes, annual gynecologic visits are often the only tim e young women see a health care provider and she recommends that gynecologic and family -planning care providers monitor weight

PAGE 46

38 gain and offer their patients counseling and guidance on weight maintenance or reduction at annual examinations. Studies also recognize that th ere are other biolog ical and social factors that affect birth outcomes, and inadequate or excessive we ight gain can serve as useful markers of risk. Barwarsky et al. (2005) investigated th e interrelationship of pre-pregnancy factors, pregnancy factors and social i ssues related to excessive or inadequate gestational weight gain. The investigators found that women with high pre-pregnancy BMI were most likely to have excessive we ight gain during pregnancy while women with a low prepregnancy BMI were most likely to have inadeq uate weight gain. Of significance in this study, of all the food groups, low dairy consum ption was found to be associated with inadequate weight gain. In addition, women who experienced high stress during pregnancy were more likely to have inadequa te weight gain compared to women with low stress levels. This study suggests prepregnancy education a nd interventions are more appropriate for women at risk of exce ssive weight gain while interventions during pregnancy may be more successf ul with women at risk of inadequate weight gain (Barwarsky et al. 2005). The use of alcohol, tobacco and drugs during pregnancy is associated with associated with pregnancy complications, in trauterine growth re striction, low birth weight, and infant mortality. According to the Centers for Disease Control and Prevention (2007), pregnant adolescents are mo re likely to smoke than pregnant women over age 25. In 2004, about 17 percent of pre gnant teens under age 19 smoked during the last three months of their pr egnancies, compared to an overall rate of 13 percent among all women.

PAGE 47

39 Akinbami et al.’s (2000) national cross-sec tional analysis found that white mothers have the highest prevalence of smoking and the prevalence of tobacco use in this group peaks between the ages of 15 and 19 years. African American teens also have an increasing prevalence of tobacco use as maternal age rises. Hispanic teens have a relatively low prevalence of matern al tobacco use for all age groups. In addition, appropriate wei ght gain is affected by cigarette smoking. In a study among Medicaid eligible women ages 17 and olde r, Furuno et al. (2004) examined for the first time, the association between low maternal weight gain, as defined by the Institute of Medicine (IOM 1990) recommendations and ciga rette smoking. This study showed that smokers were 1.34 times more likely than non-smokers to gain less than the IOM recommended weight (Furuno et al. 2004). In addition, the odds of inadequate maternal weight gain consistently grew with an increasing number of cigarettes smoked per day. In the United States, young people use alcohol and other drugs at high rates, and are more likely to engage in high risk behaviors, such as unprotected sex, when they are under the influence of alcohol or drugs. Accordi ng to the Centers for Disease Control and Prevention (2004), 23 percent of high school students who had sexua l intercourse during the past three months used alcohol or drugs beforehand. Galvez-Myles and Myles’ (2005) researc h, however, contradicts some of the adolescent tobacco and drug-use findings. In their retrospective cohort study in rural Texas, the authors found that teens did not ha ve significantly higher frequencies of drug or tobacco use. However, teens in this study did have a higher incidence of sexually transmitted diseases than did the teens in Eure et al.’s (2002) retrospective cohort study in

PAGE 48

40 an inner-city setting, suggesting possible geog raphic differences or differences between rural and urban adolescents. Finally, risky sexual behavior can put teens at risk. According to the Centers for Disease Control and Prevention (2004), thre e million teens in the United States are affected by sexually transmitted diseases annua lly. Teens account for about one-fourth of all the cases reported. Sexually transmitted diseases include chlamydia and gonorrhea (which can cause sterility and infant bli ndness), syphilis (which can cause infant blindness and death as well as maternal deat h) and HIV (the virus which causes AIDS, which may be fatal to the mother and infant). Social Risks and Outcomes Studies show that teen parenthood tends to result in lack of social and economic resources. When compared to peers who delay childbearing, teenage mothers are more likely to be poor, receive welfare, and ar e less educated. Estimates show that approximately two-thirds of families be gun by teen mothers are poor and about onequarter of adolescent mothers will go on welfar e within three years of the child’s birth (The National Campaign to Prevent Teen Pr egnancy 2010). In addition, only about 38 percent of adolescents having children be fore they are 18 earn a high school diploma (Perper et al. 2010) and less th an two percent will complete college by the time they are 30 (Hoffman 2006). Although adolescent pregnancy can be found in all soci oeconomic groups and all racial groups, large variations are fo und between different socioeconomic and racial/ethnic groups. The Penn Study of Teenage Pregnancy examined several

PAGE 49

41 psychosocial factors associated with adolescent pregnanc y from the perspective of 326 African American teens between the ages of 13 and 17 years w ho were enrolled in family planning and obstetric services at the Hospital of the University of Pennsylvania in the early 1980s. The study followed teen participants for two years to investigate changes in initial pregnancy status as well as change s in perceptions and behavior over time (Freeman and Rickels 1993). The Penn Study revealed changes in psyc hological and emotional factors. The delivery group reported deterioration in es teem in the family relations dimension, suggesting that childbearing di d not enhance their status in the family, but instead resulted in additional responsibilities and di fficulties. The delivery group also exhibited the least future orientation as rated by the in terviewers at every a ssessment point, whereas the never-pregnant and abortion groups were more likely to recognize that postponing motherhood was important for achieving other educational and caree r goals (Freeman and Rickels 1993:37). One of the more interesting findings in the Penn Study concerns teens’ desires to become pregnant. Although study participan ts reported that they never strongly “wanted” a pregnancy, the study showed that those who became pregnant did so because pregnancy was not sufficiently unwanted enough to prevent them from becoming pregnant. In other words, study participan ts did not believe the consequences of pregnancy would negatively affect their lives The study also showed that personal and familial factors associated with making pr egnancy unwanted enough to actively avoid it included strong educational and career goals that teens fe lt would be impeded by a baby, and family attitudes that oppose childbeari ng at a young age. The opposite was true for

PAGE 50

42 study participants who gave birth. They reported less support from their families for education and career goals and nearly two-thirds thought thei r mothers would be happy if they had a baby (Freeman and Rickels 1993). A great majority of the teens in the Pe nn Study perceived a generalized level of social and familial acceptance of early ch ildbearing. Twenty-eight percent of the participants had a sister who gave birth as a teen and over three-fourths said their close girlfriends had babies. And most of the study participants believed that their family would provide financial, materi al and social support if they had a baby. Interestingly, most teens felt that their family and frie nds would not support them in a decision to terminate their pregnancy. Although interviews with family members and friends was not part of this study, perception appears to guide behavior of these teens (Freeman and Rickels 1993). The Penn Study also briefly investigated repeat pregnancies during the study period. Results showed that adolescents w ho had a second pregnancy fell further behind in school and were less likely to have set occupational goals by the two-year follow-up. In addition, the study found that thos e who had a repeat birth were: less likely to be using contraception more likely to be younger (ages 13-15) at first birth four times more likely to be below their school grade level twice as likely to have fallen behind in school less likely to believe they coul d achieve their oc cupational goals more likely to have higher depression levels as measured by the SCL-90 Freeman and Rickels (1993:117) concluded that “Childbearing brings personal satisfaction but also has negative social and economic outcomes that young teens neither

PAGE 51

43 know nor understand.” In a ten-year followup to the Penn Study, researchers found that 35 percent of the teens without children duri ng the study still had no children after almost ten years. Those with children during the study had an average of 1.7 children with a range from 1 to 4. Overall, about 83 per cent had completed high school. All but one of those not completing high school had a child during the study. Wome n with children had lower incomes than those without children ($12,000 compared to $18,000) in their midtwenties, and this was not changed by when they had the child, whether it was before or after completing high school, nor by the number of children they had. The researchers concluded that long-term ec onomic disadvantage was rela ted to teen childbearing (Freeman and Rickels 1993:117). In a particularly compelling nationa lly-conducted retrospective cohort study among 9,159 women, Hillis et al. (2004) found that the impact of cumulative exposure to childhood abuse and family dysfunction increa sed the likelihood of adolescent pregnancy and long-term psychosocial consequences. A ssigning one point to each of eight adverse childhood experiences (ACEs), the research ers found that as the number of ACEs increased, so did the likelihood that an adolescent would be come pregnant. In addition, as the number of ACEs increased, women w ho gave birth as an adolescent were incrementally more likely to experience longterm psychosocial consequences, including family problems, job problems, financial pr oblems, difficulty cont rolling anger and high levels of stress as adults. Unlike the studies discussed above that attempt to show teenage childbearing places women at risk of undesira ble psychosocial outcomes as adults, Hillis et al. (2004) concluded that exposure to adverse chil dhood experiences results in adolescent childbearing and undesirable psychosocial outcomes in adulthood.

PAGE 52

44 From a more anthropological perspectiv e, viewing the issue of teenage child bearing historically and cross-culturally pr ovides insights into modern-day social concerns. Lawlor and Shaw (2002) note the acceptable age for chil dbearing has shifted over time. The authors observe that with the rise of western biomedicine in the 18th century, came the medicalization of pregnanc y. This was the beginning of a “shift of power relations by which women’s bodies and th e reproductive process came to be seen as legitimate subjects for social control” (Lawlor and Shaw 2002:552). By the end of the 20th century, many developed countries of the world deemed teenage childbearing to be a “national public health problem requiring targeted intervention,” regardless of the country’s teen birthrate (Lawlor and Shaw 2002:552). Although most developed countries view teen pregnancy and child bearing as socially, culturally and economically unacceptable, in so me cultures (such as in an ultra-orthodox Jewish community in Jerusalem and in the country of Nepal) childbearing at a young age is encouraged (Lawlor and Shaw 2002a, Smith 2001). Research has found no association between age of the mother and advers e birth outcomes in these cultures. The literature also supports the notion that othe r factors beside age affect birth outcomes. Cunnington (2001:40) states, “The risks associated with young age (OR ranging from 1.2-2.7) are modest compared to those for the social, behavioural and economic risk factors.” The author asserts th e literature clearly demonstrates that the increased risk of adverse birth outcom es, such as anemia, pregnancy-induced hypertension, low birth weight, prematurity, intrauterine grow th retardation and neonatal mortality, were predominantly caused by the so cial, economic and behavioral factors that predispose some young women to pregnancy.

PAGE 53

45 Although there is compelling evidence that teen pregnancy causes biopsychosocial harm, Hoffman (1998) cautions against confusing correlation with causation. He contends that many studies merely show correlations or associations which put adolescent mothers and their children at increased risk of certain undesirable outcomes. For example, Hoffman (1998) cont ends that adolescent childbearing does not cause low educational attainment or poverty, and there is no evidence that changing a woman’s age at first birth would dramatically change conditions in her life. This idea is support by Hillis et al.’s (2004) study of a dverse childhood experiences. Still, whether adolescent pregnancy leads to adverse mate rnal and child outcomes or whether these outcomes are consequences of the same adve rsity that led to a dolescent pregnancy continues to be debated. Factors Associated with Adolescent Childbearing Although teen birth indicator s tell us about existing c onditions and the outcomes of childbearing, they do little to inform us about why teens are getting pregnant. There are dozens of hypotheses on the causes and factor s leading to adolesce nt childbearing and an extensive review of all th ese hypotheses is beyond the scope of this literature review. However, four studies in particular exemp lify the various perspec tives and hypotheses of why some teenagers get pregnant an d bear children. These hypotheses are anthropological in nature and use perspectives that fit well with socioeconomic theory used in this dissertation’s rese arch. This selected literatur e on teenage childbearing that attempts to explain why adolesce nt girls get pregnant and bear children include biological interpretations (Geronimus’ 2003 Weathering Hypothesis ), political economic

PAGE 54

46 considerations (Luker’s 1996 asse rtion that teen childbearing is a result, not a cause, of poverty), individual explanations (Kaplan’ s 1997 notion of “relationship poor”), and social perspectives (Stack’s 1974 ideas of support systems). While these authors do not believe early childbearing is necessarily a dvantageous to adoles cents, these authors do provide possible reasons fo r why it is occurring. Geronimus’ (2003) Weathering Hypothesis explores the biological effects of social and economic inequality on African Am ericans. Geronimus (2003) contends that the difference in fertility timing between low-income African Americans and more advantaged whites is tied to health consider ations, social support and future educational and career opportunities. She notes that among African Americans in high-poverty urban areas, early childbearing mitigates severe health risks by reducing rates of infant mortality and the chance of the child bei ng orphaned. In addition, members of the extended family, which often form the new mo ther’s support network, are more likely to be in better health and can therefore assi st with child rearing and will be less likely, themselves, to need care due to poor health. In contrast, the nuclear family ideal found in more advantaged, white populations, “calls for pa rents to be self-sufficient in the care of their children,” which is best achieved by delaying child bearing. In general, members of more advantaged groups “can expect access to high-quality and advanced education as well as opportunities for financial secur ity, rewarding careers and long lifetimes” (Geronimus 2003:886). Geronimus (2003) states that the messa ge often promoted by organizations and programs addressing teen pregnancy conte nd that teen childbearing has negative consequences for teen mothers, their childre n, and society, and that this “well-publicized

PAGE 55

47 conventional wisdom continues to hold teen ch ildbearing to be, in al l cases and in every aspect, an antisocial act and an important public health problem, especially when practiced by urban African Americans” (G eronimus 2003:882). Geronimus’ (2003) also challenges the perceived association between teenage childbearing and poor maternal, infant and social outcomes, asserting this no tion is not supported by empirical evidence. In fact, a significant body of reputable scie ntific evidence exists contradicting this perception (Geronimus 2003). Contrary to the “dominant culture’s beliefs,” studies have shown that “Among African Americans, rates of low birth weight and infant mortality are lowest for babies whose mothers are in their mid to late teens.” Discussions of unmarried young people having babies persisted for years. Using a social-construction model for her analysis Luker (1996) argues th at public perception of teenage mothers has come to represent ch allenges faced in modern society, including societal perceptions involving race, age, gender and poverty. Luker (1996) presents the voices of the young mothers to tell their stories of motherhood, the challenges they face and thei r attempts to find meaning in motherhood. She also presents a multi-causal explanation fo r adolescent childbearing, stating that teen pregnancy and childbearing can be attrib uted to poverty, limited life choices, ineffectiveness with respect to contracep tion, and difficult negotiations around sex with their male partners. According to Luker (1996), three majo r issues childbearing by unmarried women, the proper age to begin childbearing, and who is fit for parenthood have been at the forefront throughout US hist ory. She investigates the e volution of public perceptions about teenage pregnancy during the twentieth century, and argu es that teenage pregnancy

PAGE 56

48 should be recognized not as a cause of povert y, but rather as a result of poverty. She concludes that teen mothers are not pathol ogical, but rather th ey are young mothers whose problems must be understood in th e context of the larger society. Kaplan’s (1997) ethnographic work with black teenage mothers examines the strategies these girls adopt for survival. Her observations reveal that although these strategies make sense to thes e young girls within their social environment, they turn out to be inadequate. Kaplan empha sizes that the lives of thes e young mothers is not deviant, but rather that pregnancy at a young age is intentional. Th is, she concludes, raises the possibility that socioeconomic status is deep ly intertwined with the psychological growth of adolescent girls. Her work found that the black community does not now, nor ever did, condone teenage pregnancy. “Pregnant teenage girls were considered deviants in the past and are still considered so toda y by many in the black community” (Kaplan 1997:12). Kaplan discusses the social and econo mic changes that have eroded black families’ base of social support. First, community life of the 1950s and 1960s where two-parent families lived in stable neighborhoods and everyone knew each other, no longer exists. In addition, economic shifts during this time had a dramatic impact on poor black neighborhoods. Jobs and small busin esses disappeared as the service sector overtook the industrial base, and small apartmen t buildings were replaced with densely populated housing projects. This resulted in a rise in poverty and single-parent households, setting the stage for ga ngs, drugs and family disruption. Kaplan studied 32 teenage mothers from areas of Oakland and Richmond, California. She found that black middle-class fl ight to the suburbs left urban teens with

PAGE 57

49 drug dealers, sports figures and show business entertainers as models for success. School courses focused on skills needed for low-payi ng jobs, such as cosmetology or secretarial positions. Often girls lacked relationships teach ers or other responsible adults in school, which would have enabled them to see beyond gender ideology and stereotypes about black girls. In addition, the girls often had a strained relationship with their mothers, a history of childhood sexual abuse, a lack of unders tanding of their sexuality, no positive role models for relationships with men, and little kno wledge of birth control. From the girls’ perspectives, they believed they gave birt h to receive the love (from their child) and support (from the baby’s fathe r) that they did not get from their mothers. Kaplan found that two assumptions from earlier literature regarding reactions to teen childbearing no longer apply. First, adult black mothers do not generally support and encourage their daughters to keep and raise the babies, and second, support of teen mothers and their babies is no longer linked to the existence of an extended family support system. From the pregnant teens’ perspective, thei r mothers viewed them as deviant. All but one teen in this study repor ted that their mothers insisted that they have an abortion. After the baby was born, the te ens reported that they felt their mothers were punishing them because they were not mothered as they had been before the baby was born. Additionally, their mothers made them feel incapable of caring for their babies by usurping responsibility for the children or hars hly criticizing the new mothers’ child care practices. Teen mothers all also reported confusion about their mother’s wishes regarding marriage, place of re sidence and financial support. Finally, most of the teens

PAGE 58

50 said they primarily relied on one or two frie nds for support, the babies’ fathers and other kin, as well as acceptance and assistance from their churches, were absent from their support system. Regarding adult mothers, Kaplan found a strong belief in what she calls a “mothering mandate,” a belief that all mother s are expected to mother their children. This is consistent with the theory of the patriarchal family stru cture, contending that fathers provide economic support and authority while mothers are responsible for raising children with the proper values and behavior All mothers, Kaplan found, were deeply disappointed in their daughters when they b ecame pregnant, despite the aspirations they held for their daughters. However, lower class mothers felt their daughters had failed them, while middle class mothers felt ch eated of their image of middle-class respectability. Kaplan’s study also reveals complex and widely variable re lationships between the teen mother and the baby’s father. The father contributed very little, if at all, to the financial support of the child, or to the emotional support of the child’s mother. Nevertheless, most teen mothers felt that the fathers loved the children. A majority of the new mothers knew very little about their babies' fathers' persona l histories or their current lives. And, most relationships ended abruptly when the father found out the girl was pregnant, in spite of the lengt h of the relationship. Contact with the child was limited and sporadic. For the most part, these patterns reflected the teens’ relationships with their own fathers and provided a model for the mean ing of the mother/father relationship and family life.

PAGE 59

51 At one time or another, all the teen moth ers in Kaplan’s study were on some form of public assistance. Their e xperiences with the welfare sy stem and the public perception of welfare recipients also de monstrate a lack of support. In response, Kaplan found that teen mothers developed strategies to help th em handle the stigma associated with welfare use. Yet, due to the structure of the system these teen mothers regarded public assistance programs as hindering the possibility of becomi ng self-sufficient. Ka plan reports that a majority of teen mothers she interviewed regretted having a baby at such a young age. Working from a feminist perspective, Kaplan contends that childbearing among African American teens should be analyzed by studying the inte ractive effects of gender, race and class. Kaplan (1997:68) concludes, “B lack teenage girls’ experiences are rarely understood as part of the larger economic and social shifts in the lives of teenagers, women, and Americans in general.” One classic work, Stack’s 1974 ethnogr aphic study of poor, black families explores cooperation within social networks as an adaptive strategy for dealing with urban poverty and racism. She presents a picture of complex pa tterns of exchange interactions among kin and non-kin in The Fl ats, a poor black, urban community. The cooperative life style, built upon exchange a nd reciprocity, enables community members to respond to poverty, unemploy ment and scarce resources. Using the ethnographic method, Stack inves tigates how people manage to survive despite great economic hardship. She found people in The Flats employed a number of survival strategies that involved family and extended kin. For example, Stack found that membership in a household is fluid in Th e Flats, yet fluctuations in household composition did not significantly affect the durable kin networ ks and cooperative familial

PAGE 60

52 arrangements. It is the resulting collectiv e power that kept people from going hungry and helped to keep a roof over their heads. A co mplex network of reciprocal exchange with a kin network also plays a large pa rt, as does attracting others to participate in these social networks. Marriage is one example of this, and potential mates must be chosen carefully on their ability to fulfill kinship obligations. Regarding the subject of motherhood, St ack (1974:46) states that, “Men and women in The Flats regard child-begetti ng and childbearing as a natural and highly desirable phenomenon.” She cont ends that restrictions on age, marital status, and number of children are all but non-ex istent, and in fact, “very few women in The Flats are married before they have given birth to one or more children (Stack 1974:50). When a young woman becomes pregnant with her first chil d, it is unlikely that she and the father will set up a separate household. They will more likely remain living in the homes of the kin who raised them. In the case of teen ch ildbearing, the first chil d is often raised by a close female relative, although a majority of mothers in The Flats are the natural mothers of the children they are raising. Stack demonstrates that the negative feat ures often attributed to poor African American families (fatherless, matrifocal, disorganized, unstable) were not characteristic of the families she studied. Instead she found “the Black urban family, embedded in cooperative domestic exchange, proves to be an organized tenacious, active, lifelong network” (Stack 1974:124). She found that poor blacks share the same dreams and aspirations as mainstream society. However, the immediate need for survival is met in creative ways through strong loyalties to kin, internal sanctions, and a complex network

PAGE 61

53 of exchanges and obligations. Stack’s ( 1974:129) study presents “a powerful challenge to the notions of a self-per petuating culture of poverty.” Smith (2002) questions whethe r this public health “probl em” lies with the age of the mother, with other factors affecting the he alth of the mother and baby or with the attitudes towards women’s repr oductive lives. The literature supports the notion that the young age of the mother presents a social or moral concern rather than an actual public health problem. Cunnington (2001:40), for ex ample states, “It ma kes little biological sense for young women to be able to reproduce at an age that puts thei r children at risk.” In fact, Cunnington’s (2001) ex tensive review of the lite rature shows most teenage pregnancies are low risk. Additionally, wo men having babies their 30s and 40s and women who receive infertility treatment are not considered a public health problem in spite of the increased risk to both moth er and infant(s) (Lawlor and Shaw 2002). The literature also supports the notion that othe r factors beside age affect birth outcomes. Cunnington (2002:40) states, “The risks associated with young age (OR ranging from 1.2-2.7) are modest compared to those for the social, behavioural and economic risk factors.” The author asserts th e literature clearly demonstrates that the increased risk of adverse birth outcom es, such as anemia, pregnancy-induced hypertension, low birth weight, prematurity, intrauterine grow th retardation and neonatal mortality, were predominantly caused by the so cial, economic and behavioral factors that predispose some young women to pregnancy. Finally, the literature supports the noti on of teen childbearing being socially unacceptable in the dominant culture of the Un ited States. Lawlor and Shaw (2002a:558) argue that “the underlying problem lies in society’s attitudes towards young people and

PAGE 62

54 specifically in attitudes towards women’s reproductive lives.” Although it has been alleged that teenage childbearing is associ ated with increased risk of poor social, economic and health outcomes for both the mo ther and child, the authors argue that if women in the United States could begin thei r families in their teens and return to education or a career in thei r twenties without prejudice a nd with proper support, there would be no problem. Supporting this idea, among cultures where ear ly childbearing is encouraged, social and community support appe ar to serve as prot ective factors against the new mother’s inexperience and economic hardship (Lawlor and Shaw 2002, Smith 2001). Solinger (1992) also explores the hist oric relationship between race and unwed motherhood, including the time in which St ack conducted her ethnogr aphy. The findings of the two authors are similar. Solinge r’s exploration of attitudes toward unwed pregnancy prior to the US Supreme Courts 1973 decision legalizing abortion in the case of Roe v. Wade reports similar cultural, racial and economic reasons why black, single, pregnant women were not generally rejected by their families or turned away from their communities. Solinger, like Stack, notes that black families, for the most part, accepted unwed pregnancies and made a place for the ne w mother and child in the family and the community. The response of the black co mmunity to out-of-wedlock pregnancy and childbearing was to organize itself to accommoda te the mother and child. In contrast, the white community was unwilling and unable to do so, and simply reorganized itself by expelling the mother and child, usually by labeling the mother as ‘psychologically impaired,’ ‘deeply neurotic,’ or at the very least, an unfit mother and by placing the child with a nice middle-class couple who could provide the baby with a proper family.

PAGE 63

55 Solinger (1992) concludes that, “Race, in the end, was the most accurate predictor of an unwed mother’s parents’ res ponse to her pregnancy; of society’s reaction to her plight; of where and how she would spend the months of her pregnancy; and most important, the most accurate predictor of what she would do with the “fatherless’ child she bore, and of how being moth er to such a child would a ffect the rest of her life” (Solinger 1992:18). In other words, Soli nger contends that in post World War II American society, race was emerging as a vi tal pressing social issue and it was the distinction between races that ecl ipsed issues such as social cl ass, regional differences in mores, and the age of the new mother. Summary of the Literature The literature reviewed in this chapter provided direction and guidance, as well as insights and ideas for relevant technique s to inform a newly emerging “applied anthropology of GIS/spatial analysis.” Sp atial studies using GIS techniques conducted by anthropologists and archaeologists, as we ll as authors in other disciplines, have provided a basis on which to begin to build th is dissertation’s research. The GIS methods of hot spot analysis, cluster analysis and multilevel models have guided the methods used and provided future directions for research. While ecosocial theory was used to frame my research, this represents only one way to frame spatial studies. Finally, literature on maternal and child health outcomes, re sulting from adolescent childbearing, and hypotheses regarding the causes of teen mother hood present information that was directly used in my analyses.

PAGE 64

56 CHAPTER THREE: RESEARCH DESIGN AND METHODOLOGY This research presents an analysis that takes the first steps toward the development of an “applied anthropology of spatial analysis/GIS” by using software and techniques that are more readily available and familiar to many anthropologists for new and challenging applications. This research identifies the distribution of teen births in Hillsborough and Pinellas Counties and considers the neighborhood factors that contribute to adolescent childbearing in these areas using spatial analysis/GIS techniques. A mixed method design provides information from quantitative, sp atial, and qualitative perspectives. Overview of the Research Problem Florida is a large, heterogeneous state with considerable diversity in population density, demographic characterist ics and cultural variation in di fferent areas of the state. And like Florida, the counties of Hillsborough an d Pinellas are also diverse. During the years between the 1990 and 2000 U.S. Decennial Censuses, these two counties, also known as part of the greater Tampa Bay area, have seen an increase in minority residents as well as a growing population of youth under 20 years of age (Florida Legislature 2009). This increase in population and ethnic divers ity has not occurred at the same rate in all parts of the Tampa Bay area, nor has it occurred at a steady ra te over time. Looking at county-level data obscures the sometime s dramatic differences at the sub-county

PAGE 65

57 geographic level. In other words, dem ographic and socioeconomic variations in communities and neighborhoods are not detectable when analysis is conducted on a wider geographic scale, such as c ounties (Aiken et al. 1991, Krie ger et al. 2004, Reader 2000). With rising demand and dwindling resour ces, local health agencies, communitybased organizations and policymakers need local data for program planning, program evaluation and resource allocati on (Jia et al. 2004). Using sm all area analysis (also called area-based socioeconomic measures or ABSMs) this research focuses on contributing to the development of an “applied anthropology of spatial analysis/GIS” by identifying neighborhood variables, that is, informati on on neighborhood-level conditions, that can help to provide a different perspective on the ways in which patterns of teenage childbearing align with conditions in the ne ighborhoods where these teen mothers live. Research Objective The work described here builds on spatia l analytic methods described by Steven Reader (2000), Nancy Krieger (2004) and Ta ylor and Chavez (2002). Reader’s (2000) method analyzes spatial varia tion of low birth weight black infants in Florida by first identifying statistically significant high rate cl usters and statistically significant low rate clusters of low birth weight bl ack births and then comparing these clusters with the sociodemographic characteristics of their respec tive census tracts. Similarly, Taylor and Chavez (2002) investigated teen birth hot spot s in California at the census tract level. Krieger (2003b) measures socioeconomic ine quality and its relation to public health issues. Again, the unit of analysis for Krieger’ s work is the census tract. All of these

PAGE 66

58 works were conducted at a sub-county level b ecause, as the authors state, spatially aggregated outcome data can hide important geographic variability. Research Design This study incorporates three aspects of applied anthropology – framework, fieldwork, and practical applic ation (Hopper 2000). It is comp rised of two main spatialstatistical analyses that provide a framework for this research – hot spot analysis and neighborhood analysis. Hot spot analysis iden tifies where rates of adolescent births are statistically significantly high or low and presents these ho t spots and cold spots spatially on a map. Neighborhood-level analysis is used to geographically explore socioeconomic factors and how these factors ge ographically align to teenage ch ildbearing. In addition, a series of two interviews with teen pregnanc y prevention service providers and individuals from funding agencies were conducted to discov er factors that they articulated as perhaps leading to adolescent childbear ing and to discover providers ’ opinions on the usefulness and relevance of the results of this type of research for the work they do. Existing Data Sets Used Several existing data sets were used in the development of this model. These data sets include: U.S. Census Bureau, American FactFinder Datasets o 1990 U.S. Census of Population and Housing Summary File 1 (1990 SF1) o 1990 U.S. Census of Population and Housing Summary File 3 (1990 SF3) o 2000 U.S. Census of Population and Housing Summary File 1 (2000 SF1) o 2000 U.S. Census of Population and Housing Summary File 3 (2000 SF3)

PAGE 67

59 U.S. Census Bureau's Annual Population Estimates Florida Community Health Assessm ent Resource Tool Set (CHARTS) Neighborhood Change Database (NCDB) State of Florida Vital Statistics Birth data files from 1992 through 1997 All of these data sets contain only de-identified data. In other words, these files do not contain any information that could identify an individual (e.g., name, address, Social Security Number, insurance number, etc.) A description of each of these datasets and a discussion how it is used in this dissertation’s research follows. 1990 and 2000 Decennial U.S. Census of Population and Housing Every ten years, in years ending in zero, the United States conducts a decennial census to count the population and housing units for the entire United States. Although the primary purpose of the decennial censu s is to provide population counts for determining how seats in the U.S. House of Representatives are apportioned, census data are used in many other ways that are well suite d to this research. According to the U.S. Census Bureau (1994), census data are used to determine the dist ribution of government program funding (e.g., Medicaid), in planning the right locati ons for schools, roads, and other public facilities, for helping real estate agents and potential residents learn about a neighborhood, and in identifying trends over time that can help predict future needs. Investigating trends in adol escent childbearing, as demonstr ated in this research, can assist in focusing efforts in high need ar eas as well as program planning and funding allocations.

PAGE 68

60 Two versions of questionnaires are used to collect data in each decennial U.S. census. The first, referred to as the Short Form Questionnaire, asks a limited number of basic questions of each member of each household. The other version of the questionnaire, referred to as the Long Form Questionnaire, asks all of the questions included on the Short Form Questionnaire pl us additional questions regarding population and housing characteristics. The Long Form Questionnaire polls a sample of households, resulting in data sets that ar e statistically weighted to repr esent the entire population. The Short Form contains data from the 100 percen t count on age, race, sex, marital status, and Hispanic origin, as well as a limite d amount of housing-related information (U.S. Census Bureau 2002) Data collected from the census Short Form Questionnaire are presented in Summary Files 1 and 2 (SF1 and SF2), wh ile Summary Files 3 and 4 (SF3 and SF4) contain data from the Long Form Questionnai re. According to the U.S. Census Bureau (2002), Summary File 1 provides numbers for the exact data collected, even for very small groups and areas, whereas, Summary F ile 3 gives estimates for small groups and areas (U.S. Census Bureau 2002) Data from the 1990 and 2000 U.S. Census of Population and Housing Summary File 1 were used in determining denominat ors for the model being developed in this research. Although SF1 files contain fewer data elements than SF 3 files, there is more specificity in ages which better lends itself to calculating the datasets necessary to determine denominators for birth rates in the model. Data from the 1990 and 2000 U.S. Census of Population and Housing Summar y File 3 were used in the neighborhood analysis. Detailed application of sp ecific data is discussed below.

PAGE 69

61 U.S. Census Bureau's Annual Population Estimates In addition to conducting decennial censu s counts, the U.S. Census Bureau’s Population Estimates Program produces population numbers between censuses. Estimates are produced for each year be tween past censuses, while population projections are for future y ears. In general, populati on estimates are calculated using data for births, deaths and migrati on collected from various sources and are used to determine Federal funding allocations and in monitoring demographic changes, among other things. Data from the U.S. Census Bureau’s Population Estimates Program were used (in conjunction with the 1990 and 2000 decennial census data) in determining denominators for the birth rates in the mode l being developed in this research. The details of this process are described below (U.S. Census Bureau N.d.(a)). Florida Community Health Assessment Resource Tool Set The Florida Department of Health’s Community Health Assessment Resource Tool Set (CHARTS) includes data related to health statistics, such as births, deaths, disease morbidity, population and behavioral risk factors. Data in the Florida CHARTS interactive statisti cal database also includes population estimates obtained in 5-year age-groups (e .g., 0-4, 5-9, 10-14 etc.) from the Florida Legislature’s Office of Economic and Demographic Research (EDR). Population estimates on the Florida CHARTS website for individual ye ar of age populations are calculated by dividing the 5-year age-group totals by 5. Data for the population under one year of age are obtained from birth and infant d eath data, rather than by dividing the 0-4

PAGE 70

62 year-old population by 5. Data for the 1-4 year-old population are determined by subtracting the under one year-old age popul ation based on birth and infant death data from the 0-4 year-old population pr ovided by EDR. The difference is then divided by 4 to produce estimates for the indivi dual ages of 1 year-olds, 2 year-olds, 3 year-olds, and 4 year-olds. In addition to age specific data, Fl orida CHARTS also provides population data and estimates by race and by Hispanic origin. Although the U.S. Census Bureau provides population estimates for inter-decennial census years, the data provided by the EDR was updated in 2009. The decision was made to use this more recent data for population estimates for wh ite and African American adolescents. Population estimates for Hispanic and N on-Hispanic populations are not available from the EDR prior to 2004, so these estimates were obtained from the U.S. Census Bureau’s Population Estimates. Single ag e by race data were used to determine denominators for the birth rates in the model being developed and to calculate expected adolescent birth rates for the clus ter analysis discussed in further detail below. Neighborhood Change Database (NCDB) The Neighborhood Change Database (NCCD) pr ovides selected U.S. Census data elements from the 1970, 1980, 1990 and 2000 censuses. Funded by a grant from the Rockefeller Foundation, The Urban Institute pa rtnered with GeoLytics, Inc., a private firm specializing in the development of de mographic and geographic data products, to

PAGE 71

63 develop a method of normalizing earlier censu s data to the 2000 census tract data elements and geographic boundaries, allowing for comparison over time. The NCDB was used in this research to “bridge” the racial categories between the 1990 and 2000 U.S. Censuses. The method used by the NCDB to “bridge” race was developed by Jeffrey Passel of the Urban In stitute’s Population Studies Center. This method, according to Tatian (2003:4-7), “assi gns multiracial groups to single races according to the rules below, in descending order of priority: 1) Black + any other race, assign to Black, otherwise 2) Asian + any other race, assign to Asian, otherwise 3) Native Hawaiian/Other Pacific Islande r (NH/OPI) + any other race, assign to NH/OPI, otherwise 4) White + any other race, assign to White, otherwise 5) American Indian/Alaskan Native (AI/ AN) + any other race assign to AI/AN, otherwise 6) Assign to “Some other race” (only peopl e selecting this al one are assigned to that bridging category)” Although the NCDB uses th e census tract as the pr imary geographic unit of analysis, this database was also used to verify the method used to “normalize” 1990 to 2000 census block groups. This method is described in furthe r detail below. State of Florida Vital Statistics Birth Data Files Vital statistics records, which include bi rths, deaths and marriages, are collected and maintained by the Florida Department of Health. Vital statistics provide the foundation upon which many parts of a public health program are constructed and are regarded as an indispensable tool for the proper planning, management, and evaluation of many health programs (Florida Department of Health 2007).

PAGE 72

64 Birth records provide information on the number of births and birth rates for geographical areas and various population groups Data are used to estimate and predict family size and population growth. Health o fficials use birth data in planning and evaluating a wide range of health progra ms, including maternal and child health programs. Economists use birth statistics to estimate the size of the future labor force, and businesses use the data to estimate the fu ture market demand for their product. In addition to informing public programs, a birt h record is the fundamental document where proof is required of age, citizenship, or family relati onship is required (Florida Department of Health 2007). Florida Vital Statistics birth files contain records of all live birt hs to residents of Florida, regardless of where the baby was born. Cooperativ e agreements exist between states for sharing vital statistics records. Wh en a Florida resident gives birth in another state, that state will send a copy of the birt h certificate to the Vital Statistics Office in Florida (Florida Depart ment of Health 2007). Birth data provide the numerators for the hot spot analysis described below. Birth data were also used for descriptive statisti cs which provide contex t and insights into the distribution of variables. Determining Teen Birth Rates To begin the hot spot analysis process, t een birth rates needed to be determined. Rates serve to standardize the number of birt hs to adolescents by geographic area, thus allowing for comparison across areas with vastly different numbers of teens. For this study, birth rates are defined as the number of live births per 1,000 teens (for each age

PAGE 73

65 from 13 through 19 years old) in each racial category and for Hispanic origin. The formula is as follows: Number of births to [age] year-old [ra ce/ethnic group] females in a given geographic area x 1,000 Total number of [age] year-o ld [race/ethnic group] female s in a given geographic area The numerator in this formula was derived fr om State of Florida Vital Statistics Birth data files and the denominator was determ ined by using U.S. Census data. While obtaining numbers for the numerator was relativ ely straight-forward, several steps were needed to determine denominators. Preliminary Decisions The rate calculation formula above require d four preliminary decisions to be made, including the lowest age limit to be includ ed for analysis, whether all, or just some, of the racial categories would be used fo r analysis, whether th e ethnic category of Hispanic or Latino would be used, and what geographic level would be most appropriate. Based on review of data from the Flor ida Department of Health, there were roughly 2,800 births to females under age 20, with 1,900 of these births to adolescents in Hillsborough County and 900 births to teens in Pinellas County for each year during the 1990s. As may be expected with births to teen agers, the numbers of births increase as the ages of teenaged females increase. In cons ideration of the low nu mber of births to females age 12 years and younger, a decision was made to use age 13 years as the youngest age for analytic purposes. Also entering into this decision is the fact that a very different set of circumstances are usually pr esents with births to females under age 12

PAGE 74

66 years. Therefore, this study will regard “teen ager” in the strictest sense by investigating births to females, ages 13 to 19 years old. Next, a decision had to be made rega rding which racial categories would be analyzed and whether or not the ethnic ca tegory of Hispanic/Latino would be used. Florida Department of Health, vital statisti cs data show that there are fewer than 350 births per year to females under age 20 in th e entire State of Fl orida who identified themselves as races other than white or black /African American. Due to small numbers, analysis proceeded with the white and black/A frican American racial categories only for Hillsborough and Pinellas Counties, Florida. Although anthropology recognize s race as a social cons truct with no biological basis, information on individuals’ race is us ed in many datasets, including the U.S. Census and Florida Vital Statistics birth data. For the purpose of this dissertation, race is used as a proxy for cultural va riation, denoting perspectives such as differing worldviews, concepts of kinship, and customs, for exampl e, that may frame mores and attitudes of adolescent childbearing. Third, it was determined that the ethnic category of Hispanic/Latino would also be included in analyses. Data from the Flor ida Department of Health show there were a growing number of births to Hispanic adoles cents under 20 years of age each year in the 1990s. Increasing to over 4,500 births in 1997 in Florida, there were most likely sufficient numbers for analysis. Finally, a decision had to be made regard ing the geographic level of analysis. Knowing the analysis would take place at a ge ographic level within, and smaller than, the county, the U.S. Census data was again used in the decision process. Figure 3.1 below

PAGE 75

67 shows the ways in which U.S. Census tabul ates data geographically, beginning with census blocks and moving up to the entire United States. Figure 3.1 Hierarchical Relationships of U.S. Census Geography Source: U.S. Census Bureau, Standard Hierarchy of Census Geographic Entities. The lines represent “nesti ng” relationships where geographies connected by lines do not cross the boundaries of the next larger geographic unit. For example, the line joining counties and census tracts means that census tracts are completely contained within a given county and do not cross the coun ty line. Therefore, census tracts are a subdivision of a county. To ensure an adequate sample size for populations in small areas, the geographic units considered for analysis included census blocks, block groups and census tracts. The “nesting” relationship of these geographies a llow for aggregation to a higher geographic level should small population number necessita te aggregation to ensure statistical reliability.

PAGE 76

68 The census block usually contains less than 100 people and is the smallest geographic unit for which the Census Bureau tabulates data. Data at the census block level is only reported in Summ ary File 1. U.S. Census bl ock geography may or may not correspond to city blocks bounded by streets. For example, in rural areas, blocks may cover many square miles and have boundaries that are not streets. The next largest geographic area, the block group, is an aggregation of census blocks which generally contain about 1,000 people. The final geographi c area under consideration for analysis is the census tract, relatively perm anent statistical subdivision of a county. Census tracts are composed of block groups and an optimum size of 4,000 people (U.S. Census 2000a). The decision was made to initially conduct da ta analysis at the block group level. Although the smallest geographic area (the bloc k-level in this case) provides the most specificity regarding population characteristi cs, U.S. Census geography includes what are termed “water blocks.” Water blocks have zero population associated with them and therefore are unsuitable for deriving denominat ors. Block groups, the next smallest suitable geographic area, provide the desired specificity of population characteristics. Should population or sub-populati on numbers be too small, the data from block groups will readily aggregate into census tracts, the next appropriate level of geography. Calculating Denominators After decisions were made regarding ages, race and ethnicity, and geographic level, the next step in de termining birth rates involved deriving a denominator. Numerators were obtained directly from 1992 through 1997 Florida Vital Statistics Birth data for white, African American and Hi spanic adolescent fema les ages 13 through 19

PAGE 77

69 years old. As mentioned pr eviously, however, determini ng the denominator for this model is not quite so simple. To calculate a denominator at the bl ock group level, four processes needed to occur, including “bridg ing races” for the 2000 U.S. Census, deriving single ages from age-group U.S. Census data, reconciling 1990 and 2000 geographic boundaries and estimating the si ngle age adolescent populatio n. These procedures were necessary because data for ag e and race are reported differently in the 1990 and the 2000 U.S. Censuses. In addition, census geogra phies changed as population increased or decreased in certain areas. Finally, because censuses with block group level data are only conducted every ten years, population estimates had to be used to calculate population denominators for inter-decennial census years between 1990 and 2000. Calculating Age by Race and Ethnicity Table 3.1 below presents the 1990 and 2000 U.S. Census of Population and Housing Summary File 1 tables that c ontain age, race and ethnicity data. Table 3.1 1990 and 2000 U.S. Census Summary File 1 Age by Race Tables 1990 U.S. Census Tables 2000 U.S. Census Tables P12B AGE -Universe: White females P12A SEX BY AGE (WHITE ALONE) -Universe: People who are White alone P12D. AGE -Universe: Black females P12B SEX BY AGE (BLACK OR AFRICAN AMERICAN ALONE) -Universe: People who are Black or African American alone P12F. AGE -Universe: American Indian, Eskimo, Aleut females P12C SEX BY AGE (AMERICAN INDIAN and ALASKA NATIVE) -Universe: People who are AIAN alone P12H. AGE -Universe: Asian or Pacific Islander females P12D SEX BY AGE (ASIAN) -Universe: People who are Asian alone P12J. AGE -Universe: Other Race females P12E SEX BY AGE (NATIVE HAWAIIAN and ACIFICISLANDER) -Universe: People who are NHPI alone P12F SEX BY AGE (SOME OTHER RACE) -Universe: People who Some Other Race alone P12G SEX BY AGE (TWO OR MORE RACES) -Universe: People who are Two or more races P13B. AGE -Universe: Females of Hispanic origin P12H. SEX BY AGE (HISPANIC OR LATINO) -Universe: People who are Hispanic or Latino Source: U.S. Census Bureau, Standard Hierarchy of Census Geographic Entities.

PAGE 78

70 Tables used for calculating single year of age by race and ethnicity from the 1990 U.S. Census include P12B (White Females), P 12D (Black Females) and P13B (Hispanic Females). Tables from the 2000 U.S. Census used to calculate age by race and ethnicity include P12A (White Alone), P12B (Black or African American Alone), P12G (Two or more races) and P12H (Hispanic or Latino). Table 3.2 below shows how the 1990 and 2000 U.S. Census reported ages in Summary File 1 data. Table 3.2 1990 and 2000 U.S. Census Age Categories Source: U.S. Census Bureau, Standard Hierarchy of Census Geographic Entities. Data from the 2000 U.S. Census were more a ggregated, resulting in fewer reported single years of ages. Individual ages, by race and ethnicity, were determined using the same method used by the Florida CHARTS database discussed previously in this chapter. Single year of age populations are calculated by dividing the age-group totals by the number of ages in the group. To determine single ages for white, African American and Hispanic females between 13 and 19 years old in 1990, the number of females in the 12 and 13 years 1990 US Census Age Categories 2000 US Census Age Categories Under 1 year Under 5 years 1 and 2 years 5 to 9 years 3 and 4 years 10 to 14 years 5 years 15 to 17 years 6 years 18 and 19 years 7 to 9 years 10 and 11 years 12 and 13 years 14 years 15 years 16 years 17 years 18 years 19 years

PAGE 79

71 category was divided by 2 to come up with 13 year old females in each race category. The U.S. Census presents each of the other ages by single age and by race. For the 2000 data, all ages were grouped and therefore n eeded to be divided by the number of single years of age in the group for white, African American and Hispanic females 13 to 19 years of age. Bridging Race As Table 3.1 above shows, the two decenni al censuses collected and reported information on race in different ways. Asid e from some terminology changes in the 2000 U.S. Census (for example, Eskimo or Aleu t from the 1990 U.S. Census was changed to Native Alaskan in 2000) and disaggregati ng the 1990 category of Asian or Pacific Islander (into two categories of Asian and Native Hawaiian/Other Pacific Islander in 2000), the most dramatic change to the questi on of race in the 2000 U.S. Census is that respondents were allowed to identify more than one race (U.S. Census 2000b). While previous Decennial Censuses allowed responde nts to select only one race, the 2000 U.S. Census allowed respondents to select up to six races. Nationally, only about 2.4 percent of respondents selected more than one r ace in the 2000 Census, but this proportion was much higher in some census tracts (Tatian 2003). In addition to the race question, a separa te question regarding ethnicity asked respondents whether they consider themselves to be Hispanic or Latino. This question was asked in a similar way in earlier years, so no special method was needed to compare these data across the censuses (Tatian 2003).

PAGE 80

72 The U.S. Census Bureau provided counts for all of the 63 possible combinations of the six racial groups collected in the Ce nsus 2000 Short Form Questionnaire. To allow comparison of 2000 Census race data with the 1990 census categories, the Neighborhood Change Database (NCDB) took the counts of all the multiracial categories in the 2000 U.S. Census and reapportioned them into si ngle racial groups, both in terms of population numbers and proportions. Indivi duals self-identifying as “wh ite” plus one or more other races were assigned to the “white only” categor y, those identifying themselves as “black or African American” plus one or more othe r races were assigned to the “black only” category, and so on. The first step in “bridging” race for the 2000 U.S. Census, so that the races are comparable to the 1990 U.S. Census, was to di vide the “Two or More Races” age groups in the 2000 U.S. Census Summary File 1 tabl es into single ages by census block group. Age groups were split evenly by the number of ages within each group as described above. Then, using the Neighborhood Change Database (NCD) Census CD, the proportion of individuals in each census trac t from the Two or More Races category who were assigned to the white racial categor y was calculated by multiplying the number of individuals in the “Two or More Races” category in each block group in a given census tract by the proportion provided by the NCD. These numbers were then added to the corresponding individual age in the "White Only" category by single age. This process was repeated to calculate the “black or Af rican American” populati on by single age. As noted above, there is no need to “bridge” th e Latino or Hispanic ethnic group because data were collected in the same way for both the 1990 and 2000 censuses.

PAGE 81

73 Population Estimates The final step in determining denominat ors for this study involved finding the incremental change in the female population, by age and race, in the intervening years between the 1990 and 2000 U.S. Censuses. Although annual population estimates are provided by the U.S. Census Bureau (U.S. Ce nsus Bureau N.d (a)) and from the Florida Office of Economic and Demographic Resear ch (Florida Legislature 2009) on a county level, annual population estimates at the censu s tract or census bloc k group level are not available. The Florida Office of Economic and Demogr aphic Research (Florida Legislature 2009) establishes population estimates by using the number of births, deaths, immigrants and emigrants by county each year. Because da ta are not available for the number of births, deaths, and individuals who move into or leave a census tr act or census block group, an alternate method of determ ining the population was employed. Using the number of females for each single age, 13 through 19 years old, by race and ethnicity for 1990 and for 2000 by census block group, which was determined in the previous steps, an annual growth rate formul a was applied to each single age for white, African American and Hispanic female s for each year between 1990 and 2000. The formulas used to determine the incr emental change in the female population, by age and race, in the intervening years between the 1990 and 2000 U.S. Censuses are: i = (FV/PV)1/n – 1 where the Annual Growth Rate i = (2000 population/1990 population) 1/10 – 1 and FV = PV(1+i)n where the Final Value = 1990 popula tion (1+Annual Growth Rate)number of years past 1990

PAGE 82

74 This analysis resulted in the denominator s that were used to run the Hot Spot analysis. Reconciling 1990 and 2000 Geographic Boundaries Much of the work described so far focused on calculating and bridging 1990 and 2000 U.S. Census age and race data so they are comparable. In addition to differences in reporting age and race data, geographic boundaries also change d between the two decennial censuses. Census geographic boundari es are intended to remain stable over time to facilitate comparisons between censu ses, but significant population increases or decreases over time can nece ssitate adding, splitting or merging geographies such as block groups or census tracts. Because of these boundaries changes between the 1990 and 2000 censuses, a methodology was developed by Steven Reader during this study to remap 1990 block groups and their associated data into 2000 block group geographic bound aries. The steps involved are quite complicated, but the basi c procedure used geographic information system (GIS) software to overlay the 2000 block group boundaries onto the 1990 block group boundaries. Then, 1990 block-level data was used to determine the proportion of persons in 1990 block groups that contri buted to the new 2000 block group. For example, if a 1990 block group split into two block groups for 2000, the population may not have been distributed evenly. Reader’s me thod allows an exact weight to be allocated to each portion of the two new block groups. The population weights were then applied to block level population c ounts to convert the data to 2000 block group boundaries. Proportions (such as the pr oportion of Hispanic persons) were remapped by first

PAGE 83

75 converting the numerator and denominator valu es (Hispanic persons / total persons) and then recalculating the proportion. This was don e for the entire state of Florida, although only data for Hillsborough and Pinellas Counties were used for this study. Calculating Teen Birth Rates Using the population denominators calculated in th e previous steps, teen birth rates were calculated for each census block group and each census tract in Hillsborough and Pinellas Counties using Florida Vital Stat istics birth data for the years 1992 through 1997. Because there was concern about very small numbers of adolescent females in many of the census block groups, single ages of the adolescent mothers giving birth in each census block group (numerator) were a ggregated by 13 to 17 year-olds and by 18 and 19 year-olds. Similarly, the population of females in each census block group (denominator) was also aggregated to th ese age groups for the years 1992 through 1997. Hot Spot Analysis Once birth rates by age and by race/ethni city were estimated for each census block group in the State of Flor ida, a hot spot analysis was conducted. Hot spot analysis is a statistical method of exploratory data anal ysis, and is especially well-suited to large multivariate datasets, such as the datasets being used in this research. This technique can reveal the underlying structure of the dataset, natural sub-classes, interesting or unusual patterns and potential ou tliers (Gordon 1999). In this study, hot spot analysis was used to explore incidence patterns of births to adol escents in various regions of Hillsborough and Pinellas Counties by identifying locations wher e births to adolescents are statistically

PAGE 84

76 higher than if they had occurred by chance alon e. Additionally, a cold spot analysis was conducted to determine where births to adolesce nts are statistically lo wer than if they had occurred by chance alone. Using the population denominators calculat ed in the previous steps, an overall birthrate for white, for African American and for Hispanic females 13 through 17 years old and for females 18 and 19 years old was determined for Hillsborough and Pinellas Counties. Due to small numbers of births, th e single age groups were aggregated into the two age groups above. Then, a Chi square test was used to determine significance in teen birth rates at a level of .05. Determining Contextual Level Variables To begin a neighborhood contextual analys is, the relevance and availability of block group-level variables needed to be de termined. First, characteristics that are relevant to adolescent childbe aring were determined. Several sources of data for the neighborhood-level variables were explored, including intervie ws with service providers, a review of the literatu re and composite indices. Exploratory Interviews Short, open-ended survey interviews we re conducted with a convenience sample of five service providers and individuals from funding agencies, all of whom work directly or indirectly with at -risk or pregnant adolescents by providing direct services or funding for programs. Agencies were located in the 211 database (211atyourfingertips.org) and contacted for possible participati on. Several programs

PAGE 85

77 were identified that offered services to pregna nt and parenting teens, but only three were found that focused specifically on pregnancy prevention activities the Girls Drop-In Program, the Child Abuse Council, Positive SPi N. Due to the small number of programs that specifically offer teen pregnancy preven tion services, a decision was made to include a program that provides repeat pregnancy pr evention services. I was familiar with the services at Alpha House, so this agency wa s contacted and agreed to participate. Surveys interviews were conducted with staff from Alpha House, the Girls DropIn Program, the Child Abuse Council, Posi tive SPiN and the Children’s Board of Hillsborough County, all with individuals w hom I am acquainted. These programs all have different service delivery strategies wh ich offered the possibility that respondents may have different ideas about factors asso ciated with adolescent childbearing. All agency staff currently work directly with th eir program participants, or in the case of Alpha House and the Children’s Board, responde nts have worked directly with program participants within the past fi ve to six years. Alpha House works primarily with pregnant and parenting teens, but has a strong repeat pregnanc y prevention component in their program. The Girls Drop-In Program and the Child Abuse Council offer pregnancy prevention education and youth development activities, and Positive SPiN provides counseling to at-risk adolescents and their fa milies. The Children’s Board respondent is considered to be the agency’s content expert on healthy births and previously worked in direct service. Four respondents were fe male and one was male; one respondent was African American and four were white. The purpose of the survey interviews was three-fold. First, these surveys were conducted to help verify neighborhood-level variables found in the literature to be

PAGE 86

78 associated with adolescent childbearing that could feed into the Index of Socioeconomic Inequality developed as part of this dissertation research. Second, I sought to discover any neighborhood-level variable s that providers thought may contribute to adolescent childbearing, but are not found in the literature. And fi nally, I wanted to clearly establish, before beginning the second set of interviews, the fact that my dissertation research was not connected in any way to my role with the Children’s Board. This was an ethical decision based on the considerat ion that the Children’s Board is a funding agency, one of the very few that focus on funding prevention activit ies, and currently funds these programs. Providers were asked, based on thei r experience and expertise, which demographic factors they regard as most likel y to place an adolescent female at risk of becoming pregnant. All survey interviews we re conducted face-to-face at the site of the program. The responses were tape recorded and transcribed. Respondents answered the following two open-ended questions: 1. Based on your expertise and experience wi th teenage child-bearing, what do you think are the most significant demographic factors that place adolescent girls at risk of getting pregnant? 2. Which of these factors do you think are most significant? The data were analyzed using a compone ntial analysis, that is, the analysis looked for specific responses to the questions asked. Interviews were transcribed and responses were sorted and coded manually. Regarding validity of the data, all of the interview respondents represent individuals whose work focuses specifically on teen pregnancy prevention activities. Although the number of interviews is small, each of these respondents is professionally engaged in teen pregnancy prevention programming

PAGE 87

79 and activities on a daily basis. They represent the best key informants in the Tampa Bay area. The interview responses were categori zed according to individual-level and neighborhood-level (demographic) indicators and then by the type of services the agency or program provides. On the first sort, the responses were divided into individual-level (personal) factors and neighborhood-level (d emographic) factors that respondents believed had an influence on adolescent childbearing. Review of the Literature Drawing on the literature re lated to adolescen t childbearing, vari ables associated with teenage births was explored. The large amount of literature on this topic provided insights into many neighborhood-level variable s. For example, studies (which are discussed in more detail in the Review of the Literature) found that conditions such as poverty and living in single parent house holds were associated with adolescent childbearing. Composite Indices One option to address neighborhood-level variables is a composite index. A composite index combines a number of indica tors that include a range of economic, social and housing issues, for example, into a single score. Two widely-used existing composite indices, the Townsend Index and the Carstairs Index, were examined for availability of the variables, and relevance and alignment to the conceptual framework within this research model. In addition to examining existing composite indices,

PAGE 88

80 literature on variables associated with adolescent childbearin g afforded the opportunity to develop a new index using Summar y File 3 U.S. Census data. Because this model’s development was gene rative in nature, data from the first round of interviews had to be analyzed to determine whether or not the results yielded indicators where data were available and suitable fo r neighborhood analysis. After determining that data for many of the indica tors identified during the interviews was not readily available, a decision was made to us e the variables identified in the literature, relying primarily on selected indicators used in Harvard’s Public Health Disparities Geocoding Project (Krieger et al. 2004). Variables were derived from the 1990 U.S. Census Summary File 3 at the census block gr oup level, the smallest geographic level for which data are available, and were used to explore neighborhood cont extual effects. Developing an Index of Socioeconomic Inequality Before the neighborhood variables could be mapped, a composite socioeconomic index for all census block groups in the Hillsborough and Pinellas Counties was created using the variables in Table 3.3 below. An i ndex is a set of indicators, combined in a standardized way, that summa rize complex or multi-dimensional characteristics of a geographical area or highlight what is ha ppening there. An index provides the big picture, making indicators easier to interpret th an trying to find a trend in many separate indicators (Saisana and Tarantola 2002). To develop the socioeconomic index for my research, the z-score was calculated for each variable in Table 3, by census block group, and the sum of a census block

PAGE 89

81 group’s z-scores is a block gr oup’s index of deprivation. Th e larger the score the more deprived an area is assumed to be (Krieger et al. 2004). Table 3.3 Selected Indicators for Index of Socioeconomic Inequality Construct Operational definition 1990 Census variable 2000 Census variable A) Occupational class 1) Working class % of persons employed in predominantly working class occupations, i.e., as non-superv isory employees, operationalized as % of persons employed in the following 8 of 13 census-based occupational groups: administrative support; sales; private household service; other service (except protective); precision production, craft, repair; machine operators, assemblers, inspectors; transportation and material moving; handlers, equipment cleaners, laborers. P78 P50 2) Unemployment % of persons age 16 and older in the labor force who are unemployed (and actively seeking work) P71 150A, 150B, 150H B) Income 3) Median household income Median household income in year prior to the decennial census (for US in 1989 = $30,056) P80A P53 4) Low income % of households with income < 50% of US median household income (i.e., < $15,000) P80 P52 5) High income % of households with incomes > 400% of the US median household income (i.e., > $150,000) P80 P52 C) Poverty 6) Below poverty % of persons below the federallydefined poverty line, a threshold which varies by size and age composition of the household; in 1989, it equaled $12,647 for a family of 4. P117 P87 D) Wealth 7) Expensive homes % of owner-occupied homes worth > $300,000 (400% of the median value of owned homes: 1989) H61 H74 E) Education 8) Low: < high school Percent of persons, age 25 and older, with less than a 12th grade education P57 P37 9) High: > 4 yrs college Percent of persons, age 25 and older, with at least 4 years of college P57 P37 F) Crowding 10) Crowded households Percent of households with > 1 person per room H69, H49 H20 G) Other 11) Single Parent Households Percent of households with single male or single female with children <18 P19 P12 Source: Krieger et al. 2004, U.S. Cens us Bureau American FactFinder, SF3, The scores were mapped, by census bloc k group, and presented by quartiles on a color-coded map, with black re presenting areas with the highest inequality index score and white representing the lowest. A comparison was then made between the socioeconomic inequality index score and the hot spot analysis.

PAGE 90

82 Interviews with Teen Pregnancy Prevention Agencies A second semi-structured intervie w was conducted after mapping the neighborhood-level index data w ith some of the same responde nts who participated in the first interview. Two provider s, representing the Child A buse Council and the Girls DropIn Program, work directly with adolescen t females, offering pregnancy prevention services. One provider, Alpha House, wo rks primarily with teens who are already pregnant or parenting, but ha s a strong educational componen t in its program aimed at repeat pregnancy prevention for the adolescent s it serves. In addi tion, individuals from two public agencies that fund pregnancy pr evention services, Florida Department of Health and the Children’s Board of Hillsborough County, were interviewed. Results of the hot spot and the neigh borhood-level analyses were presented to these service providers to obtain their reactions and feedback on any expected and unexpected results in these analyses. Respondents were also asked about their perceptions of the util ity of this method to the work they do. As in the first set of interviews, respondents were audio-taped and th eir responses were aggregated to ensure anonymity. Data collected during these interv iews were coded and analyzed by interest in the data shown on the ma ps, how useful or relevant respondents perceived the information would be for the work they do, a nd by respondents’ perceptions of what they perceived as expected or unexpected results. Responses were also analyzed by common themes that emerged among respondents.

PAGE 91

83 Summary of Methods The methods used in this research show where teen births occurred in Hillsborough and Pinellas Counties, where teen birth rates are statistically higher or lower than would be expected, how well-off people living in these areas are, and how useful providers think this type of analysis might be for their work in adolescent pregnancy prevention. The quant itative and spatial analysis answers the questions, “how much?” and “where?” while the qualitative interviews provides in sights into the utility of the method to service providers. Together, these methods are inte nded to contribute to the work being done on developing an “applied anthropology of GIS/spatial analysis.”

PAGE 92

84 CHAPTER FOUR: RESULTS This research presents an analysis that takes the first steps toward an “anthropology of spatial analys is/GIS” by using software and techniques that are more readily available and familiar to many anthropologists, and by integrating small-scale and personal techniques of traditi onal anthropology with larger-scale, more regional methods as recommended by Aldenderfer (1996). The results identify the distribution of teen births in Hillsborough and Pinellas Counties in relation to the nei ghborhood factors that have been identified as associated with adolescent childbearing in these areas using spatial analysis/GIS techniques. Findings This study included all live births to white, African American or black and Hispanic adolescents between 13 and 19 years old in Hillsborough County and Pinellas County, Florida from 1992 to 1997. A Chi-Square test was used to determine where birth rates, by age-groups and by race/ethnicity, were statistically higher (h ot spots) or lower (cold spots) than would be expected. This research also investigated contextual, or neighborhood-level, variables in relation to the teen mothers’ homes by developing an index using U.S. Census variab les to indicate the level of socioeconomic inequality by census block groups. This helped to esta blish the relationship between areas with

PAGE 93

85 statistically higher or lowe r than expected birth rate s and the level of neighborhood socioeconomic inequality. Finally, interviews with five teen pr egnancy prevention program providers and individuals from funding agencies were conduc ted to help determin e important factors that influence teenage chil dbearing. In addition, a sec ond set of provider and funding agency interviews were conducted to elic it opinions on how us eful the type of information presented in this research might be to public health pregnancy prevention programs, and how the technique in this research might be used. Descriptive Statistics An examination of birth statistics for th e State of Florida can provide context for birth data in Hillsborough and Pinellas Coun ties. Table 4.1 below shows the number of births by age for white, black/African American and Hispanic females in Florida for the years 1992-1997. For these years, there wa s an annual statewide average of 191,773 births to women of all ages. By far, the la rgest number of births were to white women, accounting for about 75 percent of all births annually, recognizi ng white individuals make up almost 80 percent of the Florida’s pop ulation. Similarly, individuals of Hispanic descent comprise almost 17 percent of Fl orida’s population and account for about 15 percent of births. On the other hand, Afri can Americans comprise about 17 percent of Florida’s population and account for almost 23 percent of the births (US Census 2000). When viewed within racial and ethnic gr oups, dramatic differences can be seen in the percentage of adolescent bi rths. As Table 4.2 below shows, the percent of births to white, African American and Hispanic adoles cents account for approximately 10 percent,

PAGE 94

86 23 percent and 15 percent of all births, respec tively, within each ra cial and ethnic group. Births to teens of all races made up about 13.3 percent of all births from 1992 to 1997 in the State of Florida. Table 4.1 Births by Age-Group, Race/Ethnicity and by Year for the State of Florida Race / Ethnic Age 1992 1993 1994 1995 1996 1997 Ave # Births / Yr % of Births 13-17 5,103 5,241 5,695 5,857 5,560 5,623 5,513 3.83% 18-19 9,460 9,506 9,591 9,639 9,783 9,990 9,661 6.72% White All ages 144,138 145,302 143,535 142,733 143,287 144,178 143,862 100% 13-17 4,896 4,971 4,888 4,581 4,344 4,248 4,655 10.69% 18-19 5,854 5,499 5,317 5,118 5,207 5,460 5,409 12.43% Black All ages 45,161 44,617 43,207 42,257 42,347 43,594 43,530 100% 13-17 10,075 10,297 10,674 10,565 10,006 9,983 10,267 5.35% 18-19 15,479 15,173 15,100 14,953 15,181 15,664 15,258 7.96% All Races All ages 192,876 193,887 191,021 189,636 190,385 192,832 191,773 100% 13-17 1,343 1,489 1,798 1,893 1,821 1,865 1,701 6.04% 18-19 2,066 2,226 2,478 2,680 2,825 2,847 2,520 8.95% Hispanic* (Ethnicity) All ages 29,420 31,610 32,899 1,893 35,738 37,337 28,149 100% *Hispanic is an ethnicity; Hispanic individuals may be of any race Figure 4.1 below graphs the average annua l number of births (1992-1997), by age and race/ethnicity, to mothers in Florida, providing a differe nt perspective on the ages of childbearing. The numbers were derived by adding births to white women, black women and Hispanic women, by age, and dividing by six, the number of years of data. The chart shows that the largest number s of babies are born to white women who are approximately age 30 years, while black births are clearly skewed toward women around 20 years of age. Hispanic births fo llow a nearly a normal curve with the most babies being born to women who are between 25 and 30 years of age.

PAGE 95

87 Figure 4.1 Average Annual Numbers of Births 1992 1997 by Age and Race/Ethnicity in the State of Florida 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 1015202530354045 Age of MotherAverage Annual Number of Births White Black Hispanic The table and chart above for births in the State of Florida can serve as a basis for comparison with Hillsborough and Pinellas counties, Florida. Births to Adolescents in Hills borough and Pinellas Counties Florida Vital Statistics Birth data for the years 1992 through 1997 were used in this analysis. To begin, the data we re cleaned and sorted as follows. First, each year’s birth dataset was sorted by county code using column “COUNTYFP00” (Hillsborough = 057, Pinellas = 103). There were a total of 82,111 birth records for Hillsborough County and 56,124 birth records for Pinellas County for the combined years of 1992-1997. The aggregated birth records for 1992-1997 were then sorted by column “AGE_MOM” for each county. In this six-year period, 17 of the 82,111 birth records in Hillsborough County and 3 of the 56,124 birth records in Pinellas

PAGE 96

88 County contained missing age data (AGE _MOM = 99). Missing data comprised a negligible percentage of each county’s dataset. Next, the six years of aggregated birt h records were sorted by “RACE_MOM” and “ETHNIC_M,” and selected by white, blac k and Hispanic. Using the “AGE_MOM” column for each county, data were sorted by race and by ethnicity, and the average annual number of births for Hillsborough Count y and Pinellas County were calculated. The average annual number of births was deri ved by adding births to white women, black women and Hispanic women, by ag e, and dividing by six, the nu mber of years of data. Figure 4.2 Average Annual Number of Births 1992 1997 by Age and Race/Ethnicity in Hillsborough County, Florida 0 100 200 300 400 500 600 700 1015202530354045 Age of MotherAverage Annual Number of Births White Black Hispanic Figures 4.2 and 4.3 show the average annual number of births (1992-1997), by age and race/ethnicity, to mothers in Hillsborough and Pinellas Counties, Florid a. The patterns in these two county-level graphs are very similar to the State of Florida births as seen in Figure 4.1

PAGE 97

89Figure 4.3 Average Annual Number of Births 19 92 1997 by Age and Race/Ethnicity in Pinellas County, Florida 0 100 200 300 400 500 600 700 1015202530354045 Age of MotherAverage Annual Number of Births White Black Hispanic Next, using the “AGE_MOM” column for each county, data were selected and aggregated by age-groups (ages 13 to 17 and 18 to 19) by white, black and Hispanic. Table 4.2 Datasets with Record Counts Used for Analyses County Age Group by Race/Ethnicity Beginning Record Count Not Geocoded Error in Geocoding Geocoded (Records Used in Analyses) Hillsborough White 13-17 2,820 211 15 2,594 Black 13-17 2,191 131 3 2,057 Hispanic 13-17 1,142 94 6 1,042 White 18-19 4,600 324 39 4,237 Black 18-19 2,351 120 9 2,222 Hispanic 18-19 1,494 101 7 1,386 Pinellas White 13-17 1,351 101 3 1,247 Black 13-17 1,259 61 0 1,198 Hispanic 13-17 112 18 0 94 White 18-19 2,584 193 0 2,391 Black 18-19 1,357 68 0 1,289 Hispanic 18-19 212 28 0 184 All Adolescents 21,247 1,239 67 19,941 This step created 12 separate datasets show n in Table 4.2 above. There were no missing data on race and ethnicity in Pinellas Count y, and only four births missing data on race and two missing data on ethnic ity in Hillsborough County.

PAGE 98

90 The final step in cleaning and sorting the birth data involved displaying each county’s birth data by race/et hnicity and age-group on a map (see Maps in Appendix A) using the latitude and longitude columns in the birth data sets Rather than using a birth mother’s home address for geographic loca tion, the Florida Department of Health geocodes (geographically references) these a ddresses, that is, they provide a set of coordinates (latitude and longitude) which rela ted to the birth mother’s home address, thus allowing the data to be displayed on a map. To ensure accurac y, the birth data from Hillsborough and Pinellas Counties were cleane d geographically and birth records that were not geocoded (no data in columns “L ATIT” or “LONGI”) or were geocoded to a location outside of Hillsborough or Pinellas Counties, were removed from the datasets. The maps in Appendix A, displaying bi rths to white teens between 13 and 17 years old from 1992 to 1997, shows higher inci dents of births in the more populated areas, as would be expected. In Hillsborough County, higher numbers of births to white adolescents in this age range are found prim arily along the Interstate Highways in the cities of Tampa and Plant City as well as in the more populous communities of Brandon, Palm River, Gibsonton and Ruskin in the south and central part of th e county and in the community of Town 'n Country in the west-central part of the county. In Pinellas County, the incidence of births to white 13 to 17 year-olds are highest primarily in the largest cites of St. Petersburg and Clearwater. Because Pinellas County is mo re densely populated than Hillsborough County, births are more evenly distributed throughout Pinellas County whereas Hillsborough County, births clearl y cluster in more densely populated communities.

PAGE 99

91 The spatial distribution of births to white teens betw een 18 and 19 years old from 1992 to 1997 (see Appendix A) clearly aligns with the incidence of bi rths to the younger white adolescents described above. There are, ho wever, more births to 18 and 19 year-old white adolescents. The spatial distribution of births to black teens betw een 13 and 17 years old from 1992 to 1997 looks very different than the distribution of birt hs to their white counterparts, as the maps in the Appendi x shows. In Hillsborough County, births to younger black adolescents are clearly clustere d in the urbanized areas north of Tampa between Interstates 275 and 75. There is also a small cluster of births in Plant City. In Pinellas County, the incidence of births to white 13 to 17 ye ar-olds are highest in south St. Petersburg. There is also a cluster of births in the City of Clearwater. As with the incidents of births to 18 a nd 19 year-old white adolescents, maps in Appendix A show the spatial di stribution of births to blac k teens between 18 and 19 years old from 1992 to 1997 clearly aligns with bi rths to the younger black adolescents as described above. And again, there are more births to 18 and 19 year-old black adolescents. Births to Hispanic teens in both age groups (maps in Appendix A) show a pattern similar to African American births. However, there are more births to Hispanic teens in rural areas of Hillsborough County than white or African American adolescent births. In Pinellas County, births to Hispanic teens a ppear fairly evenly distributed throughout the county, with no clear visual clus ters appearing on the maps. Table 4.2 below shows the age-groups by race and ethnicity for each county with the final record count used for determining birth rates and in hot spot analyses shown in the

PAGE 100

92 last column labeled “Geocoded (Records Us ed in Analyses).” Between four and eight percent of the birth records in each cate gory contained missing geocoding data or contained an error in geocoding with the exception of Hispanic 13-17 year-olds (16%) and Hispanic 18-19 year-olds (13%) in Pinellas County. Tables 4.3 and 4.4 below provide an ove rview of births in Hillsborough and Pinellas Counties. Table 4.3 presents th e average annual number births by race and ethnicity for 13 to 17 year-olds, for 18 and 19 year-olds and for women of all ages in these two Florida counties. The numbers were derived by adding the births for each race/ethnicity by age group and dividing by six, the number of years of data. Table 4.3 Average Annual Births 1992-1997 by Age Group and by Race and Ethnicity White Black Hispanic 13 -17 555 403 176 Hillsborough Co. 18-19 2,273 1,162 570 All ages (10-58) 61,410 16,788 11,218 13 -17 270 252 22 Pinellas Co. 18-19 1,299 679 106 All ages (11-56) 45,160 9,251 2,518 As table 4.3 shows, Hillsborough County had a higher average annual number of births than Pinellas County. The tabl e also shows the largest numbe r of births were to white women, with more births to olde r teens than younger teens. Hot Spot and Cold Spot Analysis As mentioned in the previous chapter, a hot spot or cold spot analysis is an exploratory technique that iden tifies where adolescent birth rates are statistically higher or statistically lower than expected. The first step in conducting these analyses was to

PAGE 101

93 determine birth rates. Birth rates were cal culated using the geocode d birth records (the last column of Table 4.2) as the numerator and the age-group race/ethnicity female population of each census tract in H illsborough and Pinellas Counties. The geocoded births by age-group and r ace/ethnicity were calculated using MapInfo (8.0) to determine the number of births in each census tract. This became the numerator for the birth rates. The denomina tor (number of females in each census tract by age-group and by race/ethnicity) was calcu lated using the method described in the previous chapter. The births were then ma tched to the age-group by race/ethnicity female population in each census tract and the births were divided by the population to produce birth rates. As an overview, Table 4.4 shows adoles cent birth rates, presented by age group and by race/ethnicity in Hillsborough a nd Pinellas counties for the years 1992-1997. Table 4.4 Birth Rates (per 1,000 live births) 1992-1997 by Age Group and by Race and Ethnicity County Hillsborough Pinellas Age-Group 13 -17 Year-Olds 18-19 Year-Olds 13 -17 Year-Olds 18-19 Year-Olds White Births 2,5944,2371,2472,391 Population 117,88553,196101,36340,964 Birth Rate 22.079.612.358.4 Black Births 2,0572,2221,1981,289 Population 29,25311,83017,4736,409 Birth Rate 70.3187.868.6201.1 Hispanic Births 1,0421,38694184 Population 24,68410,5603,2311,304 Birth Rate 42.2131.329.1141.1 Table 4.4 also includes the number of births (numerator) and the population (denominator), used to calculate birth rates, by age-group and race/ethnicity, for adolescents in Hillsborough and Pinellas counties.

PAGE 102

94 The next step in the hot/cold spot anal yses used the birth rates matched to census tract data for Hillsborough and Pinellas Counties. The birth data are not normally distributed, meaning the data do not fall in to a bell-shaped curve. In a normal distribution, half of the data fall above the mean and half below the mean, and 68 percent of the data fall within one standard deviat ion from the mean, 95 percent fall within two standard deviations, and almost all of the da ta, 99.7 percent, fall wi thin three standard deviations from the mean. The birth data be ing used in this research are skewed by outliers, that is, birth rates th at are much higher or much lower than the rest of the data, resulting in a non-normal distribu tion. For this analysis, it was these outliers that were of interest. Because the data are not normally distributed, a Chi Square analysis, using the CHIDIST command in MS Excel (2003), was us ed to determine which, if any, of the outliers at the high end of the distribution were statistically significantly different from the rest of the data. This is the met hodology followed by Taylor and Chavez (2002) in their study of adolescent childbearing in Califor nia. The results ar e shown in Table 4.5 below. The table shows the census tract of the hot spot, the birth rate and the significance level of the Chi Square analys is by county, race/ethnicity an d age group. Most hot spots are significant at the .0001 level. The Chi S quare test only analyzed census tracts where the age-group population was over 20 females. The most hot spots were found among old white adolescents in Hillsborough County, and among older African American teens in both Hillsborough and Pinellas counties. The most hot spots for Hispanic adolescents were found among 13 to 17 yearolds in Hillsborough County.

PAGE 103

95Table 4.5 Hot Spot Census Tracts in Hillsborough and Pinellas Counties Hillsborough Pinellas 13 -17 Year-Olds 18-19 Year-Olds 13 -17 Year-Olds 18-19 Year-Olds Hot Spot Census Tract Birth Rate (/1,000 births) Significance Hot Spot Census Tract Birth Rate (/1,000 births) Significance Hot Spot Census Tract Birth Rate (/1,000 births) Significance Hot Spot Census Tract Birth Rate (/1,000 births) Significance 8 409 >.0001 10 441 >.0001 208 433 >.0001 9 686 >.0001 26 496 >.0001 218 310 >.0001 30 455 >.0001 31 427 >.0001 32 341 >.0001 124.02 456 >.0001 129 290 >.0001 138.04 305 >.01 139.07 894 >.0001 White 141.09 538 >.0001 11 310 >.01 214 403 >.0001 202.04 248 >.05 12 329 >.0001 233 269 >.01 209 601 >.0001 29 298 >.05 270 247 >.05 210 279 >.0001 30 364 >.0001 213 350 >.0001 31 479 >.0001 218 341 >.0001 33 407 >.0001 234 313 >.0001 49 433 >.0001 263 356 >.0001 112.06 499 >.0001 268.04 329 >.0001 Black 120.02 338 >.0001 268.16 249 >.05 19 312 >.01 10 332 >.0001 247 251 >.05 229.02 375 >.0001 37 488 >.0001 49 446 >.0001 259.02 340 >.0001 248.02 322 >.0001 39 463 >.0001 108.07 307 >.0001 259.02 260 >.01 43 306 >.01 127.02 404 >.0001 264 242 >.05 130.01 611 >.0001 267.02 245 >.05 132.06 556 >.0001 133.07 332 >.0001 138.04 784 >.0001 139.07 437 >.0001 Hispanic 141.09 333 >.0001 Finally, a cold spot analysis was conduc ted by again using the birth rates matched to census tract data for Hillsborough and Pinell as counties. Cold spots are areas where birth rates are significantly lo wer than would be expected. Again, because the data ar e not normally distributed, the mean and standard deviation are no longer relevant. Of interest are data at th e “tails” of the distribution. However, areas with no births, and therefore a birth rate of zero, ar e problematic in that they can disguise important variations in the data. For example, an area with an

PAGE 104

96 adolescent population of 2,000 and a birth rate of zero is very different from an area with a zero birth rate, but the popul ation is only 20 teens. To identify potential cold spots, a ‘ what if...’ scenario was applied. In other words, the question asked was, “If there was one birth, rather zero birt hs, in this census tract, what would the birth rate be? Would this birth rate still fall at the bottom of the tail?” To run this scenario, in areas where zero births occurred, one birth was substituted and then the birth rate was cal culated. Data were then sorted by birth rates. For the purpose of this analysis, up to the lowest one percent of the census tracts (two or three census tracts, depending on how many census tr acts had low population) were considered for possible cold spots. To help account fo r possible errors in population calculations, census tracts with a population of less than 20 adolescent fe males were not used. Table 4.6 shows the census tracts considered to be cold spots. Population and number of expected births were rounded to the nearest whole number. Table 4.6 Cold Spot Census Tracts in Hillsborough and Pinellas Counties Hillsborough Pinellas 13 -17 Year-Olds 18-19 Year-Olds 13 -17 Year-Olds 18-19 Year-Olds Cold Spot Census Tract Census Tract Population Actual # of Births Birth Rate (/1,000 births) Using what if Scenario Cold Spot Census Tract Census Tract Population Actual # of Births Birth Rate (/1,000 births) Using what if Scenario Cold Spot Census Tract Census Tract Population Actual # of Births Birth Rate (/1,000 births) Using what if Scenario Cold Spot Census Tract Census Tract Population Actual # of Births Birth Rate (/1,000 births) Using what if Scenario 139.05 993 0 0 50 1,565 1 .64* 273.12 1,151 1 .8* 201.05 1,406 1* .71 110.05 365 0 0 109 3,958 5 1.3* 268.09 519 0 .19 white 109 364 0 0 black none 109 1,068 0 .9 none none 114.14 261 1 3.8* 109 501 1 2.0* none none Hispanic 50 153 0 6.5 denotes the actual birth rate of the census tract and not a birth rate derived by using the ‘ what if…’ scenario.

PAGE 105

97 Births to white 13 to 17 year-old adoles cents, in Table 4.6 above, will be used to better illustrate the what if… scenario. There were 14 census tracts in this age/race group with zero births. One birth was substituted in each of these census tracts and the birth rate was calculated using one bi rth. Data were then sorted by birth rate and the three census tracts with the lowest birth rates were considered for potential cold spots. Census tracts with the lowest data points were consid ered if the data were not close to the next lowest point. In Hillsborough County, Florida, Table 4.6 s hows cold spots for most age-groups. For white 13-17 year-olds, birt h rates in three census tracts were still zero in spite of adding one birth using the ‘what if…’ scenario. This is well below the county average of 22 births per thousand adolescent females and well below the next birth rate in the tail of the distribution. Two census tracts showed births to white adolescents, ages 18 to 19 years old, to be below the county average of 79.6 births per thousand. The expected birth rate for black 13 to 17 year-old teens in Hillsborough County was 70.3 births per thousand, and for this population, the lowest birth rates all appear ed about the same. However, for 18 to 19 year-old African American teens, the ‘what if…’ scenario yielded a birth rate of less than one bi rth per thousand population in one census tract. For Hispanic adolescents, 13-17 years old, one census tract had an actual bi rth rate of 3.8 births per thousand, and was well below the next lowest bi rth rate in the distribution. For 18 to 19 year-old Hispanic adolescents in Hillsbor ough County, two census tracts showed rates well below the expect birth rate of 131.3 bi rths per thousand and well below the next lowest birth rate in th e tail of the distribution.

PAGE 106

98 In Pinellas County, Florida, Table 4.6 only shows cold spots for the two white age-groups. There was one census tract for wh ite 13-17 year-old teens with an actual birth rate well below the average birth rate of 12.3 births per t housand. Similarly, two census tracts for white 18-19 year-olds had act ual birth rates below the 58.4 births per thousand average for Pinellas County and well belo w the next lowest data point in the tail of the distribution. The distribution of birt h rates for African American and Hispanic adolescents showed the lowest birth rates to be very close to each other. No studies were found that investigate cold spots. To better u nderstand cold spots and the neighborhood settings were they are foun d, three cold spot areas were selected for ethnographic observation, including Ce nsus Tracts 50 and 109 in Hillsborough County, Florida, and Census Tract 268.08 in Pi nellas County, Florida. Observation of these neighborhoods provided some informa tion about the accuracy of the Index of Socioeconomic Inequality, developed as part of this research, as well as current conditions in these neighborhoods. Although th e Index of Socioeconomic Inequality used 1990 census data to coincide with the bi rth data used in this research, neighborhood observation can also provide in sights about the amount of cha nge that has taken place in an area over time. The Hillsborough County census tracts were selected because Census Tract 50 has a positive z-score (+0.615) in the Index of Socioeconomic Inequality, indicating a higher level of socioeconomic stress, while Census Tract 109 has a negative z-score (-0.913), indicating a lower level of socioeconomic stress. A ll cold spot census tracts in Pinellas County have negative zscores (lower socioeconomic st ress). Because two of the cold spot census tracts are located near the c ity of Safety Harbor, I arbitrarily selected one

PAGE 107

99 of these, Census Tract 268.08, which has an Index of Socioeconomic Inequality z-score of -1.252. Census Tract 50 is located near downtown Tampa. The cold spot analysis shows this area has a low rate of births for 18 and 19 year-old white and Hi spanic adolescents. Census tract boundaries include Interstate 275 on the north, the Hillsborough River on the east, the Crosstown Expressway on the sout h and North Rome Avenue on the west. Overall, the area is a mix of single and mu lti-family homes, businesses, light industry, schools, and churches. Three streets in this area have high or relatively high traffic volume, including Kennedy Boulevard, which runs east/west in th e southern part of this census tract, Cypress Boulevard, which also runs east/west in the northern part of this census tract, and North Boulevard, which runs north /south in the eastern part of this census tract. There are several bus stops locate along these streets. One of the most striking features of this area is the University of Tampa, located on the east side of this census tract along the Hillsborough River. In addition, Tampa Preparatory School, an inde pendent secondary school with a college-preparatory curriculum, is also located on the Hillsborough River, just north of the University of Tampa. In sharp contrast, across the stre et from Tampa Prepar atory School on North Boulevard, is Oakhurst Section 8 Housing. Most of the single-family homes are one-s tory block construction and appear to be about 30 to 40 years old. The homes are re latively small and located fairly close to each other, yet are set back from the st reets so they have a front yard.

PAGE 108

100 Several homes had small bicycles and toys in the yard, indicating younger children may live in this neighborhood. In addition, I saw two homes with signs indicating they are licensed hom e-childcare facilities. There were also indications that adolescents may live in this area. A Boys and Girls Club is located adjacent to Tampa Preparatory School and two c hurches, Beulah Baptist Church and New Salem Baptist Church, had signs advertising th eir after school programs. The Index of Socioeconomic Inequality shows that Census Tract 50 has an average z-score of +0 .615, indica ting a higher level of socioeconomic stress in this area. If factors such as poverty, low median house hold income and low educational attainment are risk factors for adolescent childbearing, we would expect to se e higher adolescent birth rates in this area. Inst ead, analysis shows that Census Tract 50 is a cold spot where teen birth rates are lower than expected. Th ere are two possible reasons that birth rates are lower than expected in this high stress area. First, there may be a large collegestudent population living in th is area. Although these stude nts have not completed their education (low education level) and most likely do not have hi gh-paying jobs (low median income, poverty), they plan to atta in a college degree some day and therefore may be delaying childbearing. Another reason may be the availability of after school programs in this area for neighborhood children. During the first round of interviews that I conducted, several respondents indicated that they felt out-o f-school time programs help to reduce the risk of adolescent childbearing. The observations of Census Tract 50 de scribe current-day conditions in the neighborhood. To discover what conditions were like 15 to 20 years ago, and which would coincide with the birth data I use d, an ethnographic inve stigation using key

PAGE 109

101 informant interviews along w ith archival research coul d help to clarify neighborhood conditions. Census Tract 109, also located in Hills borough County, Florida, is a cold spot area where adolescent birth rate s for white, black and Hispanic adolescents are lower than expected. Beginning at the corner of Ea st Fowler Avenue and Bruce B. Downs Boulevard, the eastern boundary of this cen sus tract runs north along Bruce B. Downs Boulevard to East Fletcher Av enue, then west to North 46th Street, then north to the Hillsborough River. Following the Hillsbor ough River southeast to East Fletcher Avenue, the boundary then proceeds west to North 50th Street, then sout h to East Fowler Avenue and west to Bruce B. Downs Boul evard. Census Tract 109 encompasses the entire campus of the University of South Fl orida. The portion of Census Tract 109 which is north of East Fletcher Avenue is The Claw at USF Golf C ourse and undeveloped conservation area. There are currently two areas of on-cam pus housing located just south of East Fletcher Avenue near the north west part of campus and one in the southeastern area of campus on Alumni Drive. The 1990 U.S. Ce nsus shows there were 3,062 individuals living in this census tract, of which 1,465 we re ages 18 and 19 years old, while the 2000 U.S. Census shows a total of 2,598 individuals living in this censu s tract, of which 1,487 were 18 to 19 year-olds. We would expect to see lower levels of adolescent childbearing in census tracts with a high z-score in the I ndex of Socioeconomic Inequality. In Census Tract 109, the Index shows a relatively low zscore of -0.913. As with Cens us Tract 50, the university setting of Census Tract 109 ma y suggest that educational aspirations may contribute to

PAGE 110

102 delayed adolescent childbearing. Archival research and key informant interviews with long-time faculty and staff, as well as alum ni, would most likely be the best method of discovering what the University of S outh Florida was like 15 to 20 years ago. Finally, I conducted a drive-by observati on of Census Tract 268.08 in Pinellas County, Florida, located just south of the city of Safety Harbor. The boundaries of this census tract include the Cour tney Campbell Causeway/Gulf to Bay Boulevard on the south, South McMullen Booth Road on the we st, the Ream Wilson Train on the north (just south of State Highwa y 590) and Tampa Bay on the east. Bayshore Boulevard runs along Tampa Bay in this census tract and ther e are no homes between this road and the bay until the far northern pa rt of this census tract. Two streets, the Courtney Campbell Caus eway/Gulf to Bay Boulevard and South McMullen Booth Road, are major thoroughfares yet they remain almost exclusively residential with the exception of a gas station, restaurant a nd small strip mall. Homes are set far back from McMullen Booth Road and vi sually separated by larg e trees and fences. The southern part of Census Tract 268.08 has a mix of smaller homes, condominiums, apartments, and a small area of manufactured homes. Heading north on Bayshore Drive, appro aching Drew Street, the land becomes higher inland and the homes become larger. With the exception of Ruth Eckerd Hall near the northwestern part of this census tract, the northern part of this census tract is exclusively residential. Homes are large, perh aps three or four bedrooms, and appear to be build in the 1970s and 1980s. Homes are set back several feet from the street, yards are well-kept, and I could see many pool cages. There were no bus st ops in this census tract.

PAGE 111

103 All homes in the residential neighbor hoods in Census Tract 50 near downtown Tampa and Census Tract 268.08 near Safety Har bor, Florida, were built about the same time in the respective neighborhoods and no ne w residential construction was observed. U.S. Census data also shows the population d ecreased slightly in Census Tract 109 at the University of South Florida. However, what can not be determined from these observations and from the maps in Appendix A, are what the changes in the demographic composition of these neighborhoods mean to the residents who live there. The effects of neighborhood demographic change, and the meaning of this change to neighborhood residents, are illustrated in the ethnogra phic work done by Ashley Spaulding (2008) in Census Tract 105. Also known to residents as the Greenwood area, Census Tract 105 is located to the northeast of Hillsborough Avenue and North 56th Street, Spaulding’s (2008) work describes the de mographic changes that occurred in the Greenwood area of Hillsborough C ounty, Florida, as a result of public housing residents’ relocation to a this neighborhood and the re sulting social dynam ics between long-time neighborhood residents and those who were relo cated to this area. Spaulding’s work clearly illustrates that the hot/cold spot analysis and the Index of Socioeconomic Inequality developed as part of my resear ch are only preliminary investigations and clearly need to be grounded in ethnography. Interview Results – Round 1 Short, open-ended interview surveys were conducted with staff from five agencies that provide services focused explicitly on preventing adolescent childbearing, including Alpha House, the Girls Drop-In Program, the Child Abuse Council, Positive SPiN and

PAGE 112

104 the Children’s Board of Hillsborough County. Providers were asked, based on their experience and expertise, which demographic fact ors they regard as most likely to place an adolescent female at risk of becoming pregnant. These surv ey interviews were relatively short, lasting from 6 minutes to 12 minutes, and were c onducted face-to-face at the site of the program. Responses were taped and later transcribed. The first round of survey responses were categorized according to individual-level and neighborhood-level (commun ity) indicators and then by th e type of services the agency or program provides. To preserve respondent anonymity, the position held by respondents in their respective agency is not discussed because the agencies are all relatively small and it is very possible to identi fy an individual by her or his job position. On the first sort, the responses were di vided into individual -level (personal) factors and neighborhood-level (demographic) factors that respondent s believed had an influence on adolescent childbearing. Th ese are paraphrased below and accompanied by examples and quotes where appropriate to illu strate responses. Additionally, factors that respondents felt were most influential have been noted. Individual-Level Factors that responden ts believed may lead to adolescent childbearing included: Lack of a good male role model in their life ( Most influential – one respondent ) One of the service provider responden ts began the survey interview by stating that she felt that a girl would be at risk of becoming pregnant if she does not have a good male role model in her life. She said, “The most significant [factor] is not having a good male role model in their life.” She continued by adding, “Teens, because of what they are going through, they need a male role model to get their self-confidence. Th ey have to have someone who’s interested, sensitive, caring, tr usting, respectful.” She later confirmed th at she believed that lack of a good male role mode l in a teen’s life was the most important factor that put a teen girl at risk of becoming pregnant.

PAGE 113

105 Cultural values regarding pregnancy and childbearing ( Most influential – one respondent ) Another provider said that she believe d there were cultural differences in the meaning of adolescents having babies. She stated, “I think some ethnic groups sometimes tend to have children and raise them because of their cultural group beliefs, like not having abortions and accepting the baby as a member of their family.” She believed that the cultural valu es of the teens and their parents represented one of the most influential factors affecting a dolescent childbearing. Supportive family and kinship support system (Most influential – one respondent) Yet another respondent thought that th e cultural norms or beliefs of a girl’s family influence whether or not she was at risk of becoming a teen mother. She speculated, “I tend to think that African American sometimes have closer family ties – and Hispanics – and they te nd to have children, because they have more family ties and the fam ily helps bring up the child instead of a the young girl bringing up the child alone.” Getting pregnant vers us having a baby The same respondent as immediately a bove stated later in the conversation that, “Some girls get pregnant, but do not always have the baby, they terminate the pregnancy. You know, more rich girls get pregnant than you think.” Getting pregnant is an illustration of risk taking behavior. Rape or molestation Another individual-level risk fact or for teen pregnancy noted by a respondent included rape or molesta tion. The respondent stated that, “…a girl being raped or molested while growi ng up and thinking that is all men are looking for and that’s a way to get their attention by acting out sexually. Then they get pregnant.” Lack of self-worth and self-confidence One of the service providers felt that issues of self-esteem can influence whether or not a girl is at risk of getting pregnant. “Some teens have feelings of acceptance, I mean, seeing acceptance as given through having sex at a young age. They need this because they are going through a rough adolescence. And they think the boy loves them and this builds their self -worth and selfconfidence.” The respondent was from one of the youth development programs and her response reflects the focus of that program. Desire to get pregnant Two respondents believed that sometimes a girl wants to have a baby. The first respondent thought this is because, “they want someone to love and someone to love them back.” The other respondent suggested that a girl may get

PAGE 114

106 pregnant because, “Sometimes having a baby makes them feel older, more mature.” Not married but can live with relatives One respondent believed the consequen ces of having a baby have different impacts on different girls, noting, “Some girls have it easy be cause of their family who wants the baby. And they can live at hom e with their family and don’t have to worry about anything, you know, like food or rent or a babys itter, things like that.” The male respondent offered the follow ing three factors th at may influence adolescent childbearing. Involvement of male partner during pr egnancy (could be desirable or not) The responded stated that he believed, “Another issue, it’s more the lack of involvement of the male partner in the life the pr egnant teen and her soontobe offspring. He, well, sometimes there’s not a desire to have that male involved because that girl is living with her pare nts or because, for whatever reason. And sometimes it’s better if he’s not involve d, for example, if he is bad-news or a bad influence.” Weathering effect [the physic al consequences of social inequality, such as poor health status and poor birth outcomes, among others] (African Americans) caused by racism, stress, issues of poverty, lack of educational opportunities o Effects self-worth, obesity, overea ting, hypertension, high blood pressure o Impact birth outcomes of teen s (and women later in life) This respondent noted that there wa s considerable discussion about the weathering effect in a recent meeting he attended. Here he de scribes Geronimus’ weathering hypothesis, which contends that African American adolescents tend to have children at a younger age while they are still healthy en ough to have children and will be healthy enough to raise them. Getting pregnant by men older than them (rather than boys their own age) Finally, he offered the following idea, stating, “There’s also one factor that was talked about some time ago – I haven’t heard this come up in a couple of year – is around the rumor that most of these girls were not getting pregnant by other teens, they were getting pre gnant by men who were older than them. And connotating some issues of statutory rape and some of those things. I don’t know that that’s really the case. I th ink there were blips on the radar screen.” Although the respondent cited th is as a possible factor c ontributing to adolescent childbearing, there appears to be some que stion as to whether or not he believes this is true.

PAGE 115

107 Neighborhood-Level Factors that respondent s believed may lead to adolescent childbearing included: Diminishing programs and supports in the community: ( Most influential – one respondent ) o Youth development activities o Teen pregnancy prevention o Gender-specific programs for boys and girls o Last 5 – 6 years teen pregnancy increased since funding for programs decreased The respondent from the Children’s Board, stated, “I think diminishing programs and supports in the community. There was a major de-funding of youth development activities, teen pregnancy prevention activities, gender specific programs for boys and girls. Those have di minished through several entities in the county, both at the county, the Ch ildren’s Board. And Work Force Wages began changing their strategies, their st rategic plan, which impacted some of those programs and supports and we have seen the rates of teen pregnancy increasing in the last 4 to 6 years since those funds have been withdrawn.” This may reflect the nature of the type of work the respondent does. It is worth noting that this respondent considers boys, as we ll as girls, in the context of teen pregnancy. Programs as a diversion from sex ( Most influential 2 respondents ) Two direct service provided believed th at keeping kids bus y with supervised activities was a way to prevent adoles cent childbearing. One respondent said, “Kids have a lot of time, like after school, when they are unsupervised. If the parent is working, kids need somewhere to go so they won’t get in trouble. There is a lot of peer pressure and kids, te ens, still need gui dance so they don’t get pregnant or in trouble with th e law, things like that.” Similarly, the other respondent believed that, “It’s easy for teens to get bored. They need something to do, something like sports or hobbies to fill their time.” Cultural changes – ages of acceptable childbearing has gotten older over the decades ( Most influential – one respondent ) The respondent noted that with longer life expect ancy, modern society’s idea of the acceptable age of childbe aring has increased. She stated, “There’s the cultural aspect of what today’s cultur e considers to be a problem in teen pregnancy, whereas 50 to 100 years it was not atypical for a girl of 16 or 17 or 18 to get married and have children.” Availability of different options o influenced by affordable health car e, birth control, abortion options, religious beliefs, ideas of family

PAGE 116

108 This respondent believed there is a comb ination of factors th at influence teen childbearing. She noted, “Lots of things influence if a girl gets pregnant, like if her parents are working and have a good job, they might have good health insurance. So she can get good health care and birth control, or she can afford to have an abortion, unless there are religious beliefs about that. Also, what the family thinks about keeping the baby.” Teen motherhood less stigmatized now than 20 years ago From one provider’s perspective, motherhood at a young age is more accepted by society than it used to be. Parent(s) working (lack of superv ise when girl is out of school) This respondent believed that lack of supervision puts a girl at risk of getting pregnant. She stated, That’s another issue, too. With both parents working, there is no one to serve in a s upervisory capacity in the home when that girl is out of school. And most parents will tell you that the hours between 2:00 and 4:00 are deadly in terms of maybe wher e some of these i ssues of pregnancy are coming from.” Nevertheless, this respondent wa s quick to point out that girls whose parent or parents are not working also get pregnant. Lack of sex education in schools – both boys and girls This respondent reflected what may be considered a personal belief about the responsibilities of the educational system. “The other big issue is that Hillsborough County is a very conservative c ounty in terms of the school system. And until we open up a more comprehensiv e sexual education component for both boys and girls, we are going to be dealing with an uphill battle. We need comprehensive sexuality educ ation in the school system and the best place for that to start is at the middle sc hool level. Our teens, unfort unately in this county, are not seeing a lot of this stuff beyond basi c health education until high school.” This respondent considered educating boys, as well as girls, as a way to prevent teen childbearing. Next, because an Index of Socioeconomic In equality was developed as part of this research, interview responses were placed into one of three categories – Social, Economic, and Other. In addition, indicator data availability is noted in Table 4.7 below, along with the possible data source based on wh ether or not geographi c data are available for the indicator.

PAGE 117

109Table 4.7 Factors Influencing Adolescent Childbear ing with Data Availability and Possible Sources Social Influences Geographic Data Economic Influences Geographic Data Other Influences Geographic Data Friends and Family Family financial resources -Poverty -Income level Role model or mentor N/A Available health care choices Accessible health care Religious beliefs State reporting laws for minors N/A N/A N/A Insurance from parent’s employer Medicaid eligibility -Working class -Poverty -Unemployed Pregnancy prevention programs Available Accessible Appropriate N/A Involvement in activities during free time Ability to pay for activity N/A Encouragement from parent(s) Friends’ opinions N/A N/A Working parent(s) Supervision N/A Type of work Job benefits -Working class -Income level Low education Sex education State and local policies N/A Cultural Norms and Values Live with relatives Crowded conditions Acceptability of early childbearing N/A N/A = data not available geographically or on a sub-county geographic level Variables identified by the se rvice providers during interviews, and that have data available geographically, included neighbor hood levels of poverty, income levels (median income, high income and low inco me), working class, and unemployment. Individual-level (personal) factors have been included in the interview analysis, even though they are not part of the larger analysis undertaken at this time. The Index of Socioeconomic Inequality, which is informed partly by these interview results, uses only

PAGE 118

110 neighborhood-level variables. However, a mu ltilevel model, which is the next logical step to the research conducted here, uses both neighborhood-level and individual-level variables that may place an adolescent girl at risk of having a baby. The individual-level variables discovered during thes e interviews can be used as a preliminary investigation for a multi-level model analysis. In addition, all variables whether they are found in the literature or not, can be used as part of a multilevel model. The multilevel modeling technique will separate variables that help explain adolescent childbearing from those that do not. Index of Socioeconomic Inequality A significant contribution of this re search to an “applied anthropology of GIS/spatial analysis” involved the development of an Index of Socioeconomic Inequality. An index essentially combines several variables into one, thus eliminating the need to analyze each variable separatel y. It provides a single area-b ased measure by statistically combining variables with different units of analysis (e.g., individuals, households, dollars, education level, and job status), allowing block group s in Hillsborough and Pinellas Counties to be compared in relati on to each other. While several indices are often used to measure socioeconomic de privation and disadv antage for spatial epidemiological research (e.g., the Townsend Index, the Carstairs Index, and the AreaBased Socioeconomic Measures (ABSMs) used in the Harvard School of Public Health’s Geocoding Project), this research develope d an index that aligned with Harvard’s ABSMs which include contextual (neighborhood-level) variables that relate specifically to adolescent childbearing.

PAGE 119

111 In their Area-Based Socioeconomic M easures Index, Kriege r et al. (2003c) provide variables demonstrating socioeconomic position. However, the authors state that variables used in any index must be meaningf ul, and at times it will be necessary to add relevant variables to any index that is developed. In other words, although socioeconomic position must be considered wh en investigating social and public health issues, any index being developed must al so include variables specific to the investigation. For example, in their study of the relations hip between low birth weight and lead poisoning, the authors added the pe rcentage of housing un its built before 1950 (a time when lead-base paint was still being used in homes) to their Area-Based Socioeconomic Measures Index. Based on the social and economic i ndicators cited during interviews, and indicators found in the literatur e to be associated with adol escent childbearing (Freeman and Rickels 1993, Guttmacher Institute 2004, National Campaign to Prevent Teen Pregnancy 2004), and using variables from Harvard’s ABSMs, a composite Index of Socioeconomic Inequality was developed based on relevant ne ighborhood-level, or community, variables. Variables used include low educational level (less than 12th grade education), unemployment, working class jobs, low income/poverty (median household income, income less than 50 percent of poverty level, poverty level), and crowded conditions (greater than or equa l to one person per room in re sidence). Measures that are negatively associated with adolescent childbearing (Freeman and Rickels 1993, Guttmacher Institute 2004, National Campaign to Prevent Teen Pregnancy 2004) include greater than fours years of college, high income (over 400 percent of median income) and expensive homes (as a measure of wealth). The variable “singl e parent households,”

PAGE 120

112 found in the literature and duri ng interviews as a contribu ting factor to adolescent childbearing, was added to the socioeconomic variables used in the Harvard Geocoding Project. Because birth data used in this research range from 1992 to 1997, neighborhoodlevel data from the 1990 U.S. Census we re used to create this index. A correlation analysis was conduc ted using SPSS (v13.0) to test the appropriateness of the indicators selected fo r inclusion in the Index of Socioeconomic Inequality. As discussed above, the data are not normally distri buted, and therefore a Spearman rho correlation was used. Eleven neighborhood-level variables were used in addition to birth rates for the six race/ethnicity and age-groups. The results of the Spearman rho correlati ons are shown in Table 4.8 below. The significance level for all variab les tested was p = .000 in a 2-tailed test. Although the strength of the correlations varies across the race/ethnic ity and age-groups, there are fairly strong correlations w ith all variables except two: Income over 400% of Median Income and Expensive Homes. Both of these variables show minimal negative relationship between and adolescent child bearing across all r ace/ethnicity and agegroups. The analysis shows that as the levels of single parent househ olds, low educational attainment, poverty, unemployment, blue co llar employment, low income households, poverty status, and crowding increase, levels of adolescent childbear ing also increase. And as levels of education, income and wealth (as measured by expensive homes) increases, adolescent childbearing decreases.

PAGE 121

113Table 4.8 Spearman’s Rho Correlation W13_17 rates B13_17 rates H13_17 rates W18_19 rates B18_19 rates H18_19 rates Single HHolds Correlation Coefficient .341 .418 .249 .364 .440 .227 Sig. (2-tailed) .000 .000 .000 .000 .000 .000 N 1338 1338 1338 1338 1338 1338 <12 grade Correlation Coefficient .404 .388 .307 .441 .366 .265 Sig. (2-tailed) .000 .000 .000 .000 .000 .000 N 1338 1338 1338 1338 1338 1338 >4 yrs college Correlation Coefficient -.340 -.298 -.250 -.402 -.279 -.199 Sig. (2-tailed) .000 .000 .000 .000 .000 .000 N 1338 1338 1338 1338 1338 1338 Unemployed Correlation Coefficient .274 .316 .187 .230 .318 .192 Sig. (2-tailed) .000 .000 .000 .000 .000 .000 N 1338 1338 1338 1338 1338 1338 Working Class Correlation Coefficient .386 .343 .278 .437 .307 .259 Sig. (2-tailed) .000 .000 .000 .000 .000 .000 N 1338 1338 1338 1338 1338 1338 Income<50% Correlation Coefficient .311 .373 .194 .322 .351 .143 Sig. (2-tailed) .000 .000 .000 .000 .000 .000 N 1338 1338 1338 1338 1338 1338 Income>400% Correlation Coefficient -.246 -.230 -.144 -.299 -.212 -.155 Sig. (2-tailed) .000 .000 .000 .000 .000 .000 N 1338 1338 1338 1338 1338 1338 Med HH Inc Correlation Coefficient -.335 -.384 -.238 -.343 -.371 -.206 Sig. (2-tailed) .000 .000 .000 .000 .000 .000 N 1338 1338 1338 1338 1338 1338
PAGE 122

114 areas with very low, low, high and very high levels of socioeconomic inequality. Hot spots and cold spots, by race/ethnicity and age-group were overlaid on the Index thematic map (see Appendix A). In general, hot spots align with areas of high or low levels of socioeconomic inequality, with a few exceptions. However, co ld spots aligned with all four levels of socioeconomic inequality. As the maps in Appendix A show, some of the hot spots for births to African American 13-17 year-olds a nd 18-19 year-olds, as well as hot spots for births to18-19 year-old Hispanic adoles cent are in areas w ith a low level of socioeconomic inequality. In addition, the ma ps in Appendix A show hot spots and cold spots are found in urban, suburban and rural ar eas of Hillsborough and Pinellas counties. Interview Results – Round 2 A second round of semi-struc tured interviews was conduc ted with five service providers and professionals, all of whom work directly with at-risk or pregnant adolescents or provide funding for these serv ices. Although I had hoped to interview the providers from the first round of interviews Positive SPiN was unable to participate. However, an individual from the Florida De partment of Health in Tallahassee was going to be in Tampa and arrangements were made for an interview. Four respondents were female and one was male; one respondent was African American and four were white. Two providers, representing the Child Abus e Council and the Girl s Drop-In Program, work directly with adolescent females, o ffering pregnancy prevention services. One provider, Alpha House, works primarily wi th teens who are already pregnant or parenting, but has a strong educational com ponent in their program aimed at repeat

PAGE 123

115 pregnancy prevention for the adolescents they serve. In addition, individuals from two public agencies that fund pregnancy preventi on services, Florida Department of Health and the Children’s Board of Hillsborough County, were interviewed. The purpose of these semi-structured interviews was to obtain providers’ reactions and feedback on expected/unexpected results of this research as well as the utility of this method to the work they do. First, they were shown six maps with birth locations for white, black and Hispanic adoles cents in the 13 to 15 year-old and 18 to 19 year-old age categories (see Appendix A). Next they were shown the hot spot/cold spot maps (see Appendix A). Finally, they were introduced to the thematic Index of Socioeconomic Inequality map (also found on the maps in Appendix A). These interviews lasted from 45 minutes to 80 minut es, with only one interview lasting less than one hour. Respondents answered the following three questions: 1. Are the results of this analysis consiste nt with your knowledge and experience with teenage child-bearing? Which factors did you expect to see? Wh ich factors did you not expect to see? 2. How reliable do you think the in formation from this analysis is? 3. Do you think this type of information would be useful to you in your work? Data gathered from the interviews were transcribed and analyzed using a componential analysis, that is, responses that specifically address the questions asked. In addition, other comments made during these in terviews were also analyzed. Based on these questions, the interview responses were categorized and then sorted and coded manually by respondents’ interest in the info rmation presented to them, by how relevant

PAGE 124

116 or useful they think this type of informati on would be for the work they do, and by their perception of the level of reliability of the information presented based on their experiences and expertise. Additionally, common themes that emerged in all five interviews were analyzed. Provider Interview Results Provider Interest. Without exception, all providers expressed interest in the maps shown to them during the interviews. One pr ovider noted, “Maps are always interesting. You get a perspective of data th at’s different.” Another said, “I like it because I get to see the bird’s eye view.” And yet another provider stated, “It’s interesting because you can see multiple factors.” However, there was also some indication th at maps were not appropriate for all of the respondents. Although all of the provider respondents stat ed that the maps were of interest to them, it became apparent early in one interview that the respondent was having difficulty getting oriented to locations on the maps and understanding the relationship between hot spots and cold spots, and the so cioeconomics of different communities. Relevance/Usefulness to Providers. All providers stated that information, such as adolescent birth distribu tions, hot spots and the so cioeconomic well-being of communities would be useful information. Th ey indicated several ways in which this type of information would be useful to them. Referring to the thematic map of the Socioeconomic Inequality Index, one provider stated, “It would be helpful to see wh at area has more of a need than another.”

PAGE 125

117 They also indicated the socioeconomic indicat ors would be useful for service planning, “…looking at where you can do the most in your community.” In addition to identifying need and targ eting prevention services in communities, providers also noted that the information and the maps would be helpful in securing funding for their existing programs or possibl e new initiatives. On e respondent stated, “This would be great if we ever apply for a grant or additional funding or want to do additional activities. Things like that.” A nother noted, “Something like this easily could be put into some kind of grant.” However, not all comments were positive. One provided stated that maps are not always the best way to present information, noting, “Sometimes they show small populations, and with small populations, it’s hard to justify services.” Another provider, when discussing the map with the Socioec onomic Inequality Index which was overlaid with hot/cold spots, stated, “This is too dist racting for me.” The same respondent, when reflecting on the data presented by race and by age-group, also offered, “I think dealing with race from a data persp ective can get ve ry confusing.” Perception of Expected and Unexpected Results. Many of the maps (see Appendix A) displayed conditions or circumstances that provide rs expected to see. For example, one respondent noted, “[The bi rths] seem to cluster around where the populations of the races are. I would expect that.” Also speaking of the birth maps, one provided noted, “…generally I would have expected it to look that way.” There were, however, several things that providers did not expect to see. Speaking of the map showing births to 13 to 17 year-old African American adolescents,

PAGE 126

118 one respondent stated, “I’m amazed at how low these are compared to some of the data I was seeing from like 2006-2007. [The number of births] has truly increased in this population. I believe it has increased significantly throughout the state.” There was considerable discussion about the hot/cold spot maps. One provider simply stated, “Wow, that is interesting. I ne ver would have thought th at was a hot spot.” Another offered, “In the Sulphur Springs area, ther e is a lot of work al ready in that area. But it’s the white adolescents in the Sulphur Springs area that are showing up [as a hot spot]. Yet Plant City is not that bad of a hot spot.” The Index of Socioeconomic Inequality also generated some discussion. One provider stated, “I would ha ve expected darker [higher levels of socioeconomic inequality] in the Plant City area. Interes ting.” Another offered this observation, “With hot spots, a neighborhood is not always that bad, but it’s next door to the bad areas. Hot spots don’t really correspond to the high inequality areas. But are they affected or impacted by that? By living in what one might call a blighted area, is that causing these hot spots that surrounded this area of high inequality?” Funding Agency Interview Results Funder Interest. Both individuals from the two f unding agencies expressed great interest in the information presented in th e maps. One respondent simply noted, “…this is data we would look at.” Th e other stated, “it’s definitely something interesting to see,” and added, “It would be interesting to see if we are spending our money where we can make an impact, and if not, what can we do in another part of the county where we haven’t looked at before.”

PAGE 127

119 Relevance/Usefulness to Funding Agencies. As with providers, respondents from the funding agencies indicated that information about the adolescent birth distributions, hot/cold spots and the socioec onomic well-being of communities would be useful information for they agency. They i ndicated several ways in which this type of information could be used, such as “Funding for special projects,” “…to pinpoint areas that we can make an impact, certainly that is where we will want to put the money,” or to, “…concentrate funds or implement a special project in an area.” One respondent stated, “…it would be essential if you were to do a needs assessment to see where the need is in the community. I can imagine someone who is not a data guru taking a look at this and seeing where they can make a difference. Certainly this would be helpful for providers to use the data.” On the other hand, according to one responde nt, the utility of presenting data by race and ethnic group, although interesting, may not be useful to that particular funding agency. “To be politically correct, we probably would put it all together. In a grant, it probably wouldn’t be politically corr ect to single out a population.” Perception of Expected and Unexpected Results. Almost all of the discussion focused on the hot/cold spot maps. One re spondent observed that, “Hot spots don’t necessarily align with where the most births are.” The focus was also primarily on unexpect ed results. For example, when looking at the hot/cold spot maps of the Hispanic populations, one respondent stated, “I’m not surprised at this considering what’s happening in that [the Hispanic] community. I would imagine that data from 08-09 is the same thing.” The other respondent commented,

PAGE 128

120 “Interesting – why is a hot spot for 18 to 19 year-old Hispanic teens showing up in Safety Harbor?” One funder also remarked on the Hi spanic 13-17 year-old hot spots in the area of Brandon and Valrico, stati ng, “I can totally understand Wimauma, Ruskin, Gibsonton. It floors me – Brandon and Valrico. It’s further enough away. It’s really odd.” Common Themes Among Providers and Funders Three common themes emerged among individuals from both provider and funding agencies, including dw indling economic resources, the need for more current birth data, and an interest in Hispanic adolescents. Each respondent mentioned lack of funding due to the current economic crisis, often on several occasions during their interview. One provider stated, “Money is so very tight right now and continuing to shrink, so you’re going to want to put your dollars where they have the most effect.” Anothe r noted, “Considering th e shortage of funding now-a-days, it is hard to know whether to offe r services where there are the most births or where there is a smaller, but higher risk population.” Yet anot her provider offered, “When you don’t have the dollars you had a few years ago, well, it’s tough now.” And a respondent from a funding agency commented, “I certainly think this would be useful for an agency and they could use this, especi ally now what’s happening with the shrinking budget.” Respondents from both service providing agencies and funding agencies also expressed the desire for more current data, no ting that the data they get is often outdated. One respondent stated, “The data we use is so metimes unfortunately several years old, so it wasn’t up to date.”

PAGE 129

121 All respondents spent most of their ti me looking at the maps of Hispanic adolescent hot spots. One respondent brought up the topic of Hispanic teens more often than white or African American. Comments included, “[Florida] has had such a large influx of Hispanic population,” and “The Hisp anic population has increased dramatically [in recent years].” Summary of Results This chapter presented this study’s results showing where teen births occurred in Hillsborough and Pinellas Counties, where teen birth rates are statistically higher or lower than would be expected, the socio-ec onomic well-being of people living in these areas, and how useful providers and funding agencies think this type of analysis might be for their work in adolescent pregnancy prevention. Results are presented by race or ethnic group (white, black and Hispanic) and each of these groups is presented by age-gr oup (13 to 17 year-olds and 18 to 19 yearolds). Descriptive birth statistics for the St ate of Florida as well as for Hillsborough and Pinellas Counties were presented. Maps were created to show the spatial distribution of births in Hillsborough and Pinellas Counties. Maps were also produced to show hot spots and cold spots, areas where adolescent childbearing is statistically higher or lower than would be expected. Finally, an Index of Socioeconomic Inequa lity was generated using 1990 U.S. Census data.

PAGE 130

122 CHAPTER FIVE: DISCUSSION This research offers data analysis me thods that contribute to an “applied anthropology of GIS/spatial analysis.” The study included quan titative, spatial and qualitative techniques and used an iterative design which integrated data on births and birth rates, small area indicators of soci oeconomic well-being, and key informant interviews to investigate adolescent chil dbearing in Hillsborough and Pinellas counties, Florida. Further work is needed in deve loping and testing sound hypotheses and applying rigorous study methodologies to advance the field. Discussion of the Study The research conducted in this study was iterative in nature. It began with calculating adolescent birth rates and determin ing hot and cold spots (where birth rates were higher or lower than e xpected). Next interviews were conducted with teen pregnancy prevention service providers and funders to determine neighborhood or demographic factors that contribute to adol escent childbearing. The results of these interviews helped to inform the Index of So cioeconomic Inequality, developed to provide context to where these adolescents live. A second set of interviews asked these service providers if they thought this type of inform ation would be helpful in the work they do. During the analysis of the first round of survey interviews, when sorting responses by individual-level and neighborhood -level (community) indicators and then

PAGE 131

123 by the type of program services several patterns emerged. First, it was interesting to note that the funding agency and the two yout h development programs noted out-of-school time activities as the most important factor infl uencing whether or not a girl is at risk of getting pregnant. This may reflect the focus of their work as well as personal beliefs. Second, all of the respondents, at some point in the interview, identified solutions to the risk factor they identified. For example, one respondent talked about teens needing something to when they are not in school and offered suggestions such as participation in sports or hobbies. This may indi cate that providers have an in terest in protective factors associated with delayed childbearing. Fi nally, although respondents were asked to identify what they felt was the most influe ntial factor affecting adolescent childbearing, two respondents identified two f actors as being equally influe ntial. This suggests that, while there are several factors that place an adolescent girl at risk of getting pregnant, there is not consensus about the most important factor. During the second round of interviews, all respondents indicated they were interested in the type of information provide d in the maps of adolescent births, hot and cold spots, and the Index of Socioeconomic In equality. Respondents also indicated they thought this type of information would be us eful in the work they do, and each person listed at least two ways in which they c ould possibly use the information. What was quite interesting, however was the fact that the cold spots on the maps generated almost no discussion from providers, even though disc ussion during the first round of interviews, which investigated risk factors for adol escent childbearing, suggested funders and providers were interested in factors and strate gies that protect and pr event teen girls from having babies.

PAGE 132

124 There are two observations regarding th e second, longer interview process worth noting. First, a study by Taylor and Chavez (2002) which investigated adolescent childbearing in California, found that many service provide rs had difficulty orienting themselves to maps. This was not the case with the providers I interviewed. In fact, just the opposite was true and respondents clearly r ecognized the different areas of the county by correctly referencing diffe rent areas by name. The s econd observation was the body language of the respondents and their level of engagement with the maps. All of the respondents took their time and carefully looked over various parts of the counties. And all respondents touched the maps, using their ha nds to point out or ex plore different parts of the counties. In addition, a ll respondents spent at least ha lf of the interview standing up and bending over the maps on the table or de sk. While exploring the maps, they made very little eye contact with me when they spoke. Finally, I had to remind each respondent at least once during the interview that they were looking at data from the 1990s and not current data. This reflects providers’ desire for current da ta as noted above. A major focus of this study investigated the relationship between socioeconomic well-being and adolescent childbearing. The I ndex of Socioeconomic Inequality, created as part of this research, is shown in the ma ps in Appendix A with hot spots and cold spots overlaid. Geographic areas contai ning hot spots and cold spots are outlined on the maps. A Spearman’s rho correlation tested the streng th and significance each of the variables considered for use in the Index. These elev en variables, which are indicators of socioeconomic well-being, were then standardiz ed using z-scores. Gradients on the maps show four levels of socioeconomic well-bei ng, ranging from very high to very low levels of socioeconomic inequality.

PAGE 133

125 A z-score standardizes a distribution in terms of the number of standard deviations that a score is from the mean of the distribution. A negative z-score means that the original score was below the mean and a positive z-score means that the original score was above the mean. The actual value of the z-score corresponds to the number of standard deviations the score is from the m ean and in what direc tion (above or below). For example, if each of the eleven indicator s shown in Table 4.8 had z-score of two (two standard deviations above the mean) for a given block group, by adding the eleven zscores together, the final z-score for th at block group would be 22.0. A score of 22.0 would indicate a very high level of socio economic inequality, meaning the population is experiencing greater levels of socioeconomic deprivation. In all of Hillsborough County, Florida, th e z-scores ranged from +23.1 to -10.3 and in all of Pinellas County, Florida, zscores ranged from +20.0 to -8.0. Higher zscores indicate greater levels of socioeconomic inequality. The z-scores for the hot spots in Hillsborough and Pinellas counties by race/e thnicity and age-group are shown in Table 5.1. Z-scores for the Index of Socioeconomic In equality were calculated at the block group level to show differences that exist w ithin census tracts. Because the Index was formulated by block group, in order to compar e the level of socioeconomic inequality in an area to hot spots and cold spots presente d at the census tract level, average z-score were calculated for hot spot and cold spot areas by adding the block group z-scores within each census tract-level hot/cold spot and dividing by the number of block groups within each hot/cold spot census tract.

PAGE 134

126Table 5.1 Average z-Scores for Census Tract Hot Spots Hillsborough Pinellas 13 -17 Year-Olds 18-19 Year-Olds 13 -17 Year-Olds 18-19 Year-Olds “Hot Spot” Census Tract Index Average z-Score “Hot Spot” Census Tract Index Average z-Score “Hot Spot” Census Tract Index Average z-Score “Hot Spot” Census Tract Index Average z-Score 8 0.728 10 2.546 208 4.244 9 -0.537 26 4.650 218 4.244 30 7.841 31 7.041 32 6.641 124 2.852 129 4.822 138 1.707 139.02 2.996 White 141.03 3.804 11 -1.205 214 4.418 202.04 -0.716 12 3.588 233 0.225 209 8.831 29 2.782 270 -1.130 210 7.269 30 7.841 213 2.310 31 7.041 218 4.244 33 10.60 234 4.483 49 2.310 263 -1.345 112.06 -0.256 268.04 -1.418 Black 120.02 1.897 268.07 -1.423 19 6.359 10 2.546 247 2.176 229.02 -0.072 37 3.431 49 2.546 259.02 4.533 248.02 -0.068 39 7.247 108.07 2.769 259.02 4.533 43 16.24 127 0.706 264 -0.046 130 1.200 267.02 -2.978 132.02 -4.080 133.01 1.707 138 1.707 139.02 2.996 Hispanic 141.03 3.804 If there is indeed an a ssociation between the neighbor hood-level variables used to calculate the Index of Socioeconomic In equality and high rates of adolescent childbearing as shown by the associations in the Spearman’s Rho correlation, we would expect to see high rates of a dolescent childbearing in areas wi th higher z-scores. In most

PAGE 135

127 cases this is true. The exceptions in Hillsborough County, where z-scores were low (negative numbers) in hot spot areas are found in the Carrollwood, Old Seminole Heights, and Valrico areas of the county. In Pinellas County, over one-thi rd (9 of the 21 hot spots) have negative z-scores, showing lower levels of socioeconomic inequality in these communities. These low zscores are associated with hot spots for both younger (13 to 17 year-old) and older (18 to 19 year-old) African American adolescents and for older Hispanic adolescents. Hot spot areas and the level of socioec onomic inequality associated with these areas raises two very different questions. Firs t, in hot spot areas (areas of higher than expected rates of adolescent childbearing) wh ere there are high levels of socioeconomic stress, the question then becomes, “If birth ra tes to adolescents are expected to be high where there are high levels of socioeconomic in equality, then why are the birth rates even higher than expected?” Second, hot spot areas where there are lower levels of socioeconomic stress raises th e question, “Why are adolescent birth rates so high in areas where they should be lower?” All of the areas shown in Table 5.1 above offer opportunities for further investig ation. Ecosocial theory stat es that neighborhood factors alone can never fully explain adolescent chil dbearing. And, as noted, hot spot analysis is an exploratory technique (Eng lish et al. 2003) which requires further investigation in order to understand why something is happe ning and why conditions look like they do. Examination of individual-level factors, perhaps using ethnographic techniques or a multilevel modeling approach, may help to explai n the higher than expected birth rates to adolescents in these geographic areas.

PAGE 136

128 While several hot spots were found in H illsborough and Pinellas Counties, there far few cold spots found, that is, areas where th e birth rate was much lower than would be expected. Again, if there is an associati on between the neighbor hood variables used to calculate the Index of Socioeconomic Inequa lity, as demonstrated by the Spearman’s Rho correlation, and very low rates of adolescent childbe aring, we would expect to see lower z-scores in cold spot areas where there are lower levels of socioeconomic stress. Table 5.2 below shows the average z-scores for th e cold spots in Hillsborough and Pinellas Counties by race/ethnicity and age-group. Table 5.2 Average z-Scores for Census Tract Cold Spots Hillsborough Pinellas 13 -17 Year-Olds 18-19 Year-Olds 13 -17 Year-Olds 18-19 Year-Olds Cold Spot Census Tract Index Average z-Score Cold Spot Census Tract Index Average z-Score Cold Spot Census Tract Index Average z-Score Cold Spot Census Tract Index Average z-Score 139.04 -0.144 50 4.430 273.07 -0.171 201.05 -1.368 110.01 -4.528 109 3.678 268.09 -1.251 white 109 3.678 black --109 3.678 ----114.02 -3.495 109 3.678 Hispanic 50 4.430 --Just two cold spot areas show higher le vels of socioeconomic inequality, both in Hillsborough County. One is located near the University of South Florida and the other is near downtown Tampa in the vicinity of the University of Tampa, perhaps suggesting that student status may play a role in the higher le vels of socioeconomic disparity in these

PAGE 137

129 areas. As with the hot spot areas, the questions that arise are, “If we expect low rates of adolescent childbearing in ar eas that are socioeconomically better-off, why are these areas so much lower than expected?” and “W hy are adolescent birt h rates not higher in more socioeconomically stressed areas?” As wi th hot spot areas, these cold spots call for further investigation. While there is some evidence of an asso ciation between adolescent childbearing and neighborhood socioeconomic well-being based on the results of the Index of Socioeconomic Inequality, clearly other f actors influence childbearing. We cannot ignore individual-level experience, social processes and state le vel factors, as well as the interaction between these aspects, that can al so affect adolescent childbearing and birth outcomes (Colen et al. 2006, Diez-Roux 2001, Hox 1998, Jones and Duncan 1998, Krieger et al. 2005, O’Campo 1997). It is also important to note here that the Index of Socioeconomic Inequality is an area-based population measure and is not intended to be substituted for individual-level measures. These area-based measures provide context while individual level measures, such as ri sk taking behavior, values, mores, social relationships, religious beliefs, and other cultu ral factors can all pr ovide explanation and meaning to adolescent childbearing. Results of the Index of Socioeconomic In equality are consiste nt with Leventhal and Brooks-Gunn’s (2008) research which s hows that socioeconomic conditions of neighborhoods where adolescents live are associ ated with their well-being. Their study found that high socioeconomic status (SES ) neighborhoods were associated with adolescent’s educational achievement wh ile low-SES neighborhoods were associated

PAGE 138

130 with low social well-being and high sexual and fertility outcomes. Yet anomalies exist in the results in Hillsborough and Pinellas Counties, Florida, An unusual aspect of this study is the inve stigation of cold spots as well as hot spots. While research conducted by English et al. (2003) Gould et al. (1998), JohnsonClarke (2000) Romero-Daza (2004), and Taylor and Chavez (2002) discussed in Chapter Two look for hot spots, the research in this di ssertation also investigated cold spots where rates of adolescent childbearing were lower than expected. These cold spots offer opportunities for investigation of possible prot ective factors associated with youth in a given area. It also offers a bridge for ant hologists between the la rger-scale quantitative study done for this dissertation and opportunity to investigate the role of culture using more traditional anthropologi cal ethnographic techniques as discussed by Greenbaum (1998). This may be one of the first steps to ward an “applied anthropology of GIS/spatial analysis.” Another potential contribution of this dissertation’s research, which has remained relatively unnoticed to this point, was the creation of age and race/ethnicity specific denominators for each block group’s female a dolescent population for each year of the corresponding birth data. The number of step s involved – calculati ng individual ages by race and ethnicity, bridging the racial cat egories of the 1990 and 2000 U.S. Censuses, incorporating inter-census population estim ates, reconciling Census boundaries, and calculating each census block group’s fema le population, for whites, blacks, and Hispanics for each single age for each year between the 1990 and 2000 Censuses – was time consuming, yet critical to determining de nominators used in calculating birth rates. This lengthy procedure helped to ensure grea ter data validity by eliminating the need to

PAGE 139

131 use either 1990 or 2000 U.S. Census data when the birth data used in this research fall between these two censuses. This method of calculating age and race/et hnicity specific denominators by year has great potential for future analyses. The 23rd decennial U.S. Census takes place on April 1, 2010, and current population and housin g statistics will most likely be very different than the previous U.S. Census in 2000, just as the 1990 U.S. Census differed from the 2000 U.S. Census. Each decennial census presents a “snapshot” in time of population and housing conditions, and demogr aphic and socioeconomic conditions will vary over the ten-year period leading to th e subsequent census enumeration. Until such time as the U.S. Census Bureau begins provi ding age, gender and race specific data by year at different geographic levels, this method of calculating these data will help to provide greater data validity. Last, but not least, the interviews co nducted with service providers were an important part of this study. The inte rviews found that service providers and professionals, who work directly with at-ris k or pregnant adolescen ts or provide funding for these services, considered the results of th is research and analysis would be useful in their provision and funding of services. Res pondents considered the information in the maps to be potentially useful for serv ice planning, identifying need, targeting communities where prevention services are needed, and in applying for grants. In addition to addressing the applied anthropol ogical goal of making our work useful, the interviews also served as a bridge from the larger scale region al hot/cold spot and socioeconomic analyses to a sm aller-scale inquiry more cons istent with anthropological techniques.

PAGE 140

132 Caveats and Limitations This research relied primarily on two data sources – the U.S. Census and Florida Vital Statistics birth data. Each has its strengths and weak points. As Kirby (1996) notes, censuses do not provide a total enumer ation of a population, although they strive to. The same is true for vita l statistics data se ts, which also tend to be undercounted. Nationally, the 1990 U.S. Census missed almost two percent of the population (National Public Radio 2009). In Hillsbor ough County, Florida, for children under age 18, the U.S. Census Bureau estimates there was an undercount of: 4.1 percent white (6,561) 6.5 percent black (2,720) 6.3 percent Hispanic (1,922) For children under 18 years old in Pinellas County, Florida, the U.S. Census Bureau (1999) estimates an undercount of: 4.3 percent white (6,825) 6.3 percent black (1,528) 5.5 percent Hispanic (295) The 1990 U.S. Census population underc ount affects denominators used to determine teen birth rates. However, the 2000 U.S. Census introduced a method for estimating populations which will be used in subsequent census enumeration and helping to provide a more accurate population count. Additionally, the method for determining single age by gender by race by year, developed as part of this dissertation, helped to partially account for this undercount by using an annual growth rate calculation to move the population forward toward the more accurate 2000 U.S. Census.

PAGE 141

133 In addition to undercounts, the 1990 U.S. Census presented another challenge related to race and ethnicit y. Although the 2000 U.S. Cens us reported number of nonHispanic individuals by race, the 1990 U.S. Cens us did not. Therefore, in this research white and black adolescents may also be Hispanic. In other words, the Hispanic adolescents in the population denominators in this research are also included in the denominators for white and African American adolescents. Therefore, it is important to consider that the maps showing the distributi on of births to Hispan ic adolescents most likely duplicates births to wh ite and African American adol escents. This has cultural, and perhaps linguistic, implicat ions for service providers. While ages by race and ethnic ity were calculated for each inter-census year for the adolescent population, data were not available to calcul ate inter-census indicators used in the Index of Socioeconomic Inequalit y. As the U.S. Census Bureau’s American Community Survey data becomes available for more indicators and on a smaller geographic scale, inter-censu s estimation of neighborhood-le vel indicators may become possible. As shown, there is a heavy reliance on seconda ry data sets in this research and, in fact, any research involving GI S or spatial analysis. A lthough the data are publicly available there are still ethical considerat ions in the way the data are used and disseminated. Anthropologists ha ve ethical responsibilities to the people we study and communities affected by our work, as well as our colleagues, students, our employers and to society as a whole. When considering a newly emerging anthropology of GIS, ethical considerations must be examined. Dependi ng on the research being conducted, ethical considerations must include ensuring confid entiality while sharing data and study results

PAGE 142

134 appropriately. Appropriate dissemination of data and research results would consider how the people being studied are represented as well as sharing how to (and how not to) interpret study results. It would also consider the conse quences of research and take steps to ensure local political powers are not reinforced and politically sensitive areas are protected (American Anthropological Association 1998, Brondizio 2002, Cassell and Jacobs 1987, Society for Applied Anth ropology 1983, Whiteford and Trotter 2008). Reflecting on the interview process, pe rhaps only one interview should have been conducted, rather than two. Maps presente d during the second, longe r interview engaged the respondents and provided a visual repr esentation of how factors contributing to adolescent childbearing were used in my study. I believe soliciting respondents ideas about factors affecting adoles cent childbearing (from the firs t, short interview survey) could have been accomplished more effectively during the interview with the maps (my second interview), when there was some c ontext for how these variables were used. Finally, the methods described in this research are exploratory and can not be substituted for ethnographic inve stigations. While useful in helping to highlight what may be happening and where, quantitative a nd spatial methods alone can not answer the questions why or how. This is not meant to be a stand-alone method, but rather a place to begin further investigations. Summary This study has shown some association be tween hot spot and cold spot areas of Hillsborough and Pinellas Counties where adolescen t childbearing is higher or lower than

PAGE 143

135 would be expected by chance. While this is not always the case, th ere is enough evidence for further investigation. The Index of Socioeconomic Inequality, wh ich was developed as part of this study, showed inconsistency in relation to so cioeconomic stress and teen births. The incompatible index scores with hot/cold s pots were found primarily in Pinellas County, Florida, where the adolescent fema le population was relatively small. Interviews with individuals who pr ovide adolescent pr egnancy prevention services or who work at agencies that f und pregnancy prevention programs and activities bridged the spatial analytic portion of this research with more traditional anthropological ethnographic techniques. Respondents expressed interest in the resu lts of this study and identified several ways in which they thought the type of information provided would be useful for their program or for possible future projects.

PAGE 144

136 CHAPTER SIX: CONCLUS IONS AND RECOMMENDATIONS To contribute to an “applied anthropol ogy of GIS/spatial analysis,” this study must be more than the geography of health or an epidemiological study. As Aldenderfer (1996) points out, one of the challenges of working toward an “applied anthropology of GIS” is to integrate small-scale and personal techniques of traditional anthropology with larger-scale, more regi onal methods. This research has combined spatial analysis within two Florida counties with key informant inte rviews, a more widely used technique among cultural anthropologists. In addition to contributing to an “applie d anthropology of GIS/ spatial analysis,” this study makes several contributions to the di scipline of anthropology as a whole. First, this research presents the opportunity to view adolescent childbearing from a holistic perspective by presenting racial and ethnic analys is that can be used to consider race in the context of culture. Becau se there is no biological basis for race, there is an opportunity to explore race and et hnicity in the context of cult ure. Second, this research presents a multi-method approach, showing one example of how to bridge spatial and quantitative inquiry with more ethnographic investigation. Third, this research is informed by the contentious debates on cause and effect of race/ethnicity and influences on adolescent childbearing. By investigati ng neighborhood-level fact ors that influence adolescent childbearing, the focus shifts from the individual to the influences of outside factors. Fourth, the focus of this resear ch was to develop a method that can inform

PAGE 145

137 practitioners. As applied anthropologists, we want our work to be useful. Finally, there is interest within the disc ipline of anthropology regarding spatial analysis. Dr. Susan Stonich presented her ideas on how to move toward an anthropology of rather than anthropology in spatial analysis/GIS to the 101s t Annual Meeting of the American Anthropological Association in 2002 Conclusions While archaeology has embraced spatial anal ytic techniques on larger geographic scales often using sophisticated technol ogy, as discussed in Chapter 2, cultural anthropologists employ similar techniques, but usually on a smaller s cale. According to Greenbaum (1998), ethnographic approaches ar e often used in anth ropological inquiry and employ several technique s for studying small populations, including key informant interviewing, mapping spatial re lationships and the use of qua ntitative data, all of which were used in this research. Just as anth ropologists have been using ethnography, as a way “to understand cultural and social di fferences within and among communities” (Greenbaum 1998:120), this research has used newer technologies on a larger scale which provide a different and somewhat more holistic perspective on local phenomena. The research undertaken in this disserta tion set out to address three questions: What are the patterns of adolescent ch ildbearing in Hillsborough and Pinellas counties, Florida? Is there a relationship between commun ity-level socioeconomic indicators and adolescent childbearing? Will adolescent pregnancy prevention service providers and funding agencies find this information useful and relevant to the work they do?

PAGE 146

138 First, this study has investigated pa tterns of adolescen t childbearing in Hillsborough and Pinellas Counties, Florida, in several ways. Descriptive statistics show patterns in childbearing among 13 to 17 year-o ld and 18 and 19 year-old white, African American and Hispanic teens. Although white teens have the larg est number of births, birth rates are highest for African Americ an teens in both age groups. In addition, hot/cold spot analysis has shown neighbor hoods where births to adolescents are statistically higher or lower th an would be expected. While there were no hot spots for white 13 to 17 year-olds in Pinellas County and no hot spots for black 13 to 17 year-olds in Hillsborough County, there were a large num ber of hot spots for 13 to 17 year-old Hispanic teens in Hillsborough County. Alt hough hot spot analysis is a descriptive technique which is exploratory in nature, the maps provide a way to view the hot spots spatially and offer the opportunity fo r more targeted research studies. Regarding the second question and rela tionships between community-level socioeconomic indicators and adolescent child bearing, there were mixed results when these hot/cold spots were overlaid on a ma p showing the Index of Socioeconomic Inequality. Hot and cold spots were locat ed in neighborhoods experiencing both high and low socioeconomic stress. Although the Sp earman’s rho analysis showed correlation between the selected indicators used in the index and adolescent childbearing, there are clearly other factors at work which need to be explored. Addressing the third research question rega rding the usefulness and relevance of spatial investigation of adol escent childbearing, interviews with adolescent pregnancy prevention services providers and funders high lighted the utility of this study’s results. Respondents believed the information provided in the maps could be used for service

PAGE 147

139 planning, to obtain additional f unding, and to help target pr evention efforts for maximum impact. In fact, there is also evidence that the methods used in this study are applicable to service providers other than those a ddressing teen pregnancy prevention. I was recently asked by the Hispanic Services Counc il for a map that displays the Index of Socioeconomic Inequality and areas of Hillsborough County where there is a high concentration of Hispanic families. The Hisp anic Services Council plans to use this map in their strategic planning process. This dissertation’s resear ch demonstrates the feasibility of the mapping techniques, hot spot analysis and development of an index that were used to investigate adolescent childbearing in this research. The software (MS Excel and GIS mapping software) used in this study is readily available and increasingly being used by anthropologists. From a cultural perspective, strati fying female adolescents by race/ethnicity and age-groups helped provide a more detailed pers pective of adolescent childbearing. In addition, cal culating birth rates by agegroup and race/ethnicity to correspond to each year’s birth data helped to ensure a higher level of validity for population denominators. The use of publicly -available data allows this methodology to be replicated anywhere in the United States and the flexibility of this publicly-available neighborhood data means it can be used with a wide variety of outcomes or events and can track these outcomes or events over time. This research can contribute to the field of applied anthropology and other social sciences by demonstrating the f easibility and utility of this problem-driven approach and by providing a greater understa nding of how contextual, or neighborhood-level, factors can be analyzed. Although this study ha s yielded important information about the

PAGE 148

140 relationship between adolescent childbear ing and the socioeconomics of the neighborhoods where these girls live, further work is needed in developing and testing hypotheses and applying rigorous research methodologies to advance the emerging field of an “applied anthropology of GIS/spatial analysis.” Recommendations Bearing in mind that this study is just one small step toward an “applied anthropology of GIS,” there are many things that can be done to build on this research. These include alternate ways to assess hot and cold spots, different ways to conceptualize neighborhood-level variables through multileve l modeling techniques, and working to build a conceptual framework for anthropological small area analyses. While useful, the hot spot analysis us ed in this study is limited by geographic boundaries. As Kirby (1996) point s out, these boundaries are arti ficial and social groups are not confined to ZIP Codes or Census Tract s, but rather are influenced by factors from a wide variety of sources. Hot spot assessmen t using cluster analysis is not dependent on these artificial political boundari es and may prove to be anot her way to view hot spots, especially on a larger scale. The idea of hot spot cluster analysis as an exploratory device leads to a discussion of another meaningful method of analysis. Recently, hierarchical models (multilevel models) have emerged as a methodology to ha ndle the interplay of individual-level variables and neighborhood variab les. This method would be th e next logical step in the research presented here, building on the Inde x of Socioeconomic Inequality which was

PAGE 149

141 developed using neighborhood-level variables. In addition, multilevel methods provide a bridge between statistical modeling and de scriptive mapping (Krieger et al. 2003). Multilevel modeling is an important tool that allows simultaneous study of individual-level and neighbor hood-level factors, and provide s information on how these two levels of risk interact. Analyses of i ndividual-level and neighbor hood-level variables, in addition to their interaction, is useful for developing better e xplanatory models of adolescent pregnancy, childbe aring and birth outcomes. Examining individual-level variables and neighborhood-level risk factors, as well as interacti on between the two, will help to increase understanding of the many factors responsible for adolescent birth rates, birth outcomes and risks and protective factor s. Multilevel analysis can also help to identify racial differences in birth rates and di sparities in birth outcomes. In general, this will lead to a better understanding of the complex causes of adolescent pregnancy and adverse birth outcomes. In addition further analysis, the informa tion provided by this type of study has potential practical applications. As indicated in interviews, providers can use information on where to focus their prevention activities a nd for writing grants for expanding services or implementing special projects. Funding agen cies can use the type of information in this research to make funding decisions and to track the impact of the services they fund. There are also several possible ways in which anthropologists can use information from this type of research. First, anthropologists can use this research as a beginning point for further investigation of cold spots to better unders tand protective factors that allow adolescents to delay ch ildbearing. In addition, anth ropologists can work with service providers and funding agencies to assi st them in interpreting the data regarding

PAGE 150

142 areas where teen births rates are higher than expected or where th ere are a large number of births to adolescents. Fi nally, anthropologists can assist teen pregnancy prevention service providers in designing and impleme nting new and creative ways to deliver prevention services and to provide these se rvices in a culturally sensitive manner. Perhaps the greatest challe nge facing an emerging “applied anthropology of GIS/spatial analysis” is inte grating statistical, spatial and anthropol ogical theory with analytic applications. As Kirby (1996:1860) asserts, without critic al attention to the aspects of small area analysis conducted in this research, “the gul f between theory and practice will widen into an ocean.” Here is where anthropological theory, perspectives and methods can begin to contribute to th is newly emerging “applied anthropology of GIS/spatial analysis” by provi ding a way in which phenomena can be viewed, interpreted and understood.

PAGE 151

143 REFERENCES CITED Abe, Yoshiko. Curtis W. Marean, Peter J. Nilssen, Zelalem Assefa, and Elizabeth C. Stone 2002 The Analysis of Cutmarks on Archaeofauna: A Review and Critique of Quantification Procedur es, and a New Image-Analysis GIS Approach. American Antiquity 67(4):643-663. Abrams, Barbara, Sarah L. Altman, and Kate E Pickett 2000 Pregnancy Weight Gain: Still Controve rsial. American Journal of Clinical Nutrition 71 (supplemental);1233S-1241S. Aickin, Mikel, Clara N. Dunn, and Timothy J. Flood 1991 Estimation of Population Denominators fo r Public Health Studies at the Tract, Gender, and Age-Specific Level. Am erican Journal of Pub lic Health 81(7):918920. Akinbami, Lara J., Kenneth C. Schoendorf, and John L. Kiely 2000 Risk of Preterm Birth in Multiparous Teenagers. Pediatrics and Adolescent Medicine 154(11):1101-1107. Aldenderfer, Mark 1996 Introduction. In Anthropology, Space and Geographic Information Systems. Mark Aldenderfer and Herbert D.G. Maschner, eds. Pp. 3-18. New York, Oxford University Press. American College of Obstetrics and Gynecology (ACOG) 2003 Nutrition During Pregnancy: ACOG Technical Bulletin Number 179 – April 1993. International Journa l of Gynecology and Obstetrics 43(1):67-74. 2007 Especially for Teens: Having a Baby. Electronic document, http://www.acog.org/publicatio ns/patient_education/bp103.cfm accessed March 30, 2010. 2008 Nutrition During Pregnancy. Electronic document, http://www.acog.org/publicatio ns/patient_education/bp001.cfm accessed March 30, 2010. American Diabetes Association 2006 Gestational Diabetes. Electronic document, http://www.diabetes.org/ge stational-diabetes.jsp accessed April 2, 2006.

PAGE 152

144 American Anthropological Association 1998 AAA Statement on Race. http://www.aaanet.org/issues/policy-advocacy/AAAStatement-on-Race.cfm accessed September 6, 2009. American College of Surgeons 2006 About Cesarean Chil dbirth. Electronic document, http://www.facs.org/public_info/ operation/brochures/cesarean.pdf accessed May 3, 2006. Amini, Saeid B., Patrick M. Catalan o, LeRoy J. Dierker, and Leon I. Mann 1996 Births to Teenagers: Trends and Obstetric Outcomes. Obstetrics and Gynecology 87(5):668-674. Anachbe, N.F. and M.Y. Sutton 2003 Racial Disparities in Reproductive H ealth Outcomes. American Journal of Obstetrics Gynecology 2003: 188(4): S37-42. Ananth, Cande V., Dawn P. Misra, Kitaw Demissie, and John C. Smulian 2001 Rates of Preterm Delivery Among Bl ack Women and White Women in the United States over Two Decades: An Age-Period-Cohort Analysis. American Journal of Epidemiology 154(7):657-665. Annie E. Casey Foundation 1998 Kids Count Special Report: When Teens Have Sex: Issues and Trends. Baltimore, MD: Annie E. Casey Foundation. Anselin, Luc 2002 Mapping and Analysis for Spatial Social Science. Workshop on Spatial Analysis in Anthropology at th e 101st Annual Meeting of the American Anthropological Association, New Orleans, LA, November 21, 2002. Bateman, Brian T., and Lynn L. Simpson, Lynn L. 2006 Higher Rate of Stillbirth at the Ex tremes of Reproductive Age: A Large Nationwide Sample of Deliverie s in the United States. American Journal of Obstetrics and Gynecology 194(3):840-845. Brawarsky, P., N.E. Stotland, R.A. Jackson, E. Furentes-Afflick, G.J. Esocbar, N. Rubashkin, and J.S. Haas. 2005 Pre-Pregancy and Pregnancy-Rela ted Factors and the Risk of Excessive or Inadequate Gestational Weight Gain. International Journal of Obstetrics and Gynecology 91:125-131.

PAGE 153

145 Brimicombe, Alan J. 2005 Cluster Detection in Point Even t Data having Tendency Towards Spatially Repetitive Events. Paper presented at 8th International Conference on GeoComputation, Ann Arbor, August 1-3. Brondizio, Eduardo S. 2002 Enhancing Ethnographic Research Usi ng Remote Sensing Data. Workshop on Spatial Analysis in Anthropology at the 101st Annual Meeting of the American Anthropological Association, New Orleans, LA, November 21, 2002. Carstairs, Vera 1981 Small Area Analysis and Health Service Resear ch. Journal of Public Health 3(2):131-139. Cassell, Joan and Sue-Ellen Jacobs 1987 Handbook on Ethical Issues in An thropology. Special publication number 23. Joan Cassell and Sue-Ellen Jacobs, eds. Washington, D.C.: American Anthropological Association. Centers for Disease Control and Prevention 2004 Florida Youth Risk Behavior Survey. Tallahassee: Florida Coordinated School Health Program. Electronic document, http://www.doh.state.fl.us/ disease_ ctr l/epi/Chronic_Disease/YRBS/Intro.htm accessed April 22, 2006. 2007 Preventing Smoking and Exposure to Secondhand Smoke Before, During, and After Pregnancy Electronic document, http://www.cdc.gov/chronicdisease/re sources/publications/fact_sheets/ pdf/smoking.pd, accessed March 30, 2010. Chang, Shih-Chen, Kimberly O. O’Brien, Maur een Schculman Nathanson, Jeri Mancini, and Frank R. Witter 2003 Hemoglobin Concentrations Influence Birth Outcomes in Pregnant AfricanAmerican Adolesce nts. Journal of Nutrition 133:2348-2355. Colen, Cynthia G., Arline T. Geronimus, and Maureen G. Phipps. 2006 Getting a piece of the pie? The Economic Boom of the 1990s and Declining Teen Birth Rates in the United States. Social Science and Medi cine 63 (6): 15311545. Conant, Francis Paine 1994 Human Ecology and Space Age Technology: Some Predictions. Human Ecology 22 (3): 404-413. Coppolillo, Peter B. 2000 The Landscape Ecology of Pastoral Herdin g: Spatial Analysis of Land Use and Livestock Production in East Africa. Human Ecology 28 (4):527-560.

PAGE 154

146 Cunnington, Aubrey J. 2001 What’s So Bad About Teenage Pregnanc y? The Journal of Family Planning and Reproducti ve Health Care 27(1):36-41. Diez-Roux, Ana V. 2001 Investigating Neighborhood and Area Effects on Health. American Journal of Public Health 91(11):1783-1789. 2003 Potentialities and Limita tions of Multilevel Analysis in Public Health and Epidemiology. In Methodology and Epistemology of Multilevel Analysis, Daniel Courgean, ed. Pp. 93-119. Bo ston: Kluwer Academic Publishers. Eck, John E., Spencer Chainey, James G. Ca meron, Michael Leitner, and Ronald E. Wilson 2005 Mapping Crime: Understanding Hot Spots. SPECIAL REPORT AUG. 05. Washington, D.C: Nati onal Institute of Justice. Ehrlich, Ginny and Carlos A. Vega-Matos 2000 The Impact of Adolescent Pre gnancy and Parenthood on Educational Achievement: A Blueprint for Education Policymakers’ Involvement in Prevention Efforts. Alexandria VA: National Association of State Boards of Education. Elliott, Paul and Daniel Wartenberg 2004 Spatial Epidemiology: Current Appr oaches and Future Challenges. Environmental Health Perspectives 112(9):998-1006. English, Paul B., Martin Kharrazi, Stephani e Davies, Rusty Scalf, Lance Waller, and Raymond Neutra 2003 Changes in the Spatial Pattern of Low Birth Weight in a Southern California County: The Role of Individual and Neighborhood Level Factors. Social Science and Medicine 56:2073-2088. Eure, Chineta R., Michael K. Lindsay, and William L. Graves 2002 Risk of Adverse Pregnancy Outc omes in Young Adolescent Parturients in and Inner-City Hospital. American Journal of Obstetrics and Gynecology 186(5): 918920. Florida Department of Health’s Community Health Assessment Resource Tool Set N.d. (CHARTS). Florida Department of Health, Office of Health Statistics & Assessment. http://www.floridacharts.com /charts/BirthQuery.aspx

PAGE 155

147 Florida Department of Health 2005 Florida Youth Risk Behavior Su rvey. Electronic document, http://www.doh.state.fl.us/disease_ctrl/ epi/Chronic_Disease/YRBS/Intro.htm accessed April 22, 2006. 2007 Vital Records Registration Handbook. Office of Vital Statistics. Electronic document, http://www.doh.state.fl.us/Planning_eval/Vital_Statistics/ VR_HBK_2006_Rev.pdf accessed July 17, 2009. Florida Legislature 2009 Florida Demographic Summar y. Office of Economic and Demographic Research. Florida Legislature, Tallahassee, FL. Electronic document, http://edr.state.fl.us /population/popsummary.pdf ), accessed July 25, 2009. Freeman, Ellen W. and Karl Rickels 1993 Early Childbearing: Perspectives of Black Adolescents on Pregnancy, Abortion, and Contraception. Newbury Park, CA: Sage Publications. Furuno, Jon P., Lisa Galli cchio, and Mary Sexton 2004 Cigarette Smoking and Low Maternal Weight Gain in Medicaid-Eligible Pregnant Women. Journal of Women’s Health 13(7):770-777. Futemma, Clia and Eduardo S. Brondzio 2003 Land Reform and Land-Use Changes in the Lower Amazon: Implications for Agricultural Intensific ation. Human Ecology 31(3):369-402. Galtier-Dereure, Florence, Catherine Boegner and Jacques Bringer 2000 Obesity and Pregnancy: Complications and Cost. American Journal of Clinical Nutrition 71(5):1242-1248. Galvez-Myles, Rosa, and Thomas D. Myles 2005 Teenage Pregnancy in the Texas Pa nhandle. Journal of Rural Health 21(3):259262. Geronimus, Arline T. 2003 Damned If You Do: Culture, Id entity, Privilege and Teenage Childbearing in the United States. Social Science and Medicine 57:881-893. Gordon, A.D. 1999 Classification. New York: Chapman and Hall Gould, Jeffery B., Beate Herrehen, Tanya Ph am, Stephan Bera and Claire Brindlis 1998 Small-Area Analysis: Targeting High-Ri sk Areas for Adolescent Pregnancy Prevention Programs. Fam ily Planning Perspectives 30(4):173-176.

PAGE 156

148 Greenbaum, Susan D. 1998 The Role of Ethnography in Creating Linkages with Communities: Identifying and Assessing Neighborhood Needs and Strengths. In Promoting Cultural Competence in Children’s Mental Hea lth Services, Mario Hernandez and Mareasa R. Isaacs, eds. Baltimore, MD: Paul H. Brookes Publishing Company. Guttmacher Institute 2003 Youngest Mothers' Infants Have Greatly Elevated Risk of Dying by Age One. Perspectives on Sexual and Reproduc tive Health 35(1). Electronic document, http://www.guttmacher.org/ pubs/journals/3505203.html accessed March 30, 2006. 2004 U.S. Teenage Pregnancy Statis tics With Comparative Statistics For Women Aged 20-24. Electronic document, http://www.guttmacher.org /pubs/teen_stats.html accessed, March 21, 2006. Hahn, Robert 2009 Going Where No Anthropologist Has Gone Before. Anthropology News 50 (3):31. Hillis, Susan D., Robert F. Anda, Shata R. D ube, Vincent J. Felitti, Polly A. Marchbanks, and James S. Marks 2004 Association Between Advers e Childhood Experiences and Adolescent Pregnancy, Long-Term Psychosocial Cons equences, and Fetal De ath. Pediatrics 113(2): 320-327. Hoffman, Saul D. 1998 Teenage Childbearing Is Not So Bad Af ter All…Or Is It? A Review of the New Literature. Family Planning Perspectives 30(5):236-239. 2006 By the Numbers: The Public Co sts of Adolescent Childbearing. The National Campaign to Prevent Teen Pregna ncy Washington, DC. Electronic document, http://www.thenationalcampaign.org/co sts/pdf/report/BTN_ National_Report.pdf accessed March 30, 2010. Hopper, Kim 2000 An Anthropological Perspective on the City and Mental He alth. International Journal of Mental Health 28(4):30-40. Howie, LaJeana D., Jennifer D. Parker, and Kenneth C. Schoendorf 2003 Excessive Maternal Weight Gain Patterns in Adolescents. Journal of the American Di etetic Association 103(12):1653-1657. Hox, Joop 1998 Multilevel Modeling: When and Why. In Classification, Data Analysis, and Data Highways, Ingo Balderjahn, Rudol f Mathar, and Martin Schader, eds. Pp. 147-154. New York: Springer Verlag.

PAGE 157

149 Hughes, Patricia, and Samantha Riches 2003 Psychological Aspects of Perinata l Loss. Current Opinion in Obstetrics and Gynecology 5:107–111. Institute of Medicine, Nati onal Academy of Sciences. 1990 Nutrition During Pregnancy. Wa shington (DC): National Academy Press; 1990. 2009 Weight Gain During Pregnancy: Reexamining the Guidelines. Washington (DC): National Academy Press. http://www.nap.edu/catalog/12584.html Jevitt, Cecilia M. 1983 A Study of Repeated Pregnancy Among Adolescents Enrolling in a Teen Services Program. 2005 Weight Management in Gynecologic Ca re. Journal of Mi dwifery & Women’s Health 50:427-430. Jevitt, Cecilia, Ivonne Hern andez, and Maureen Gror 2007 Lactation Complicated by Overweight and Obesity: Supporting the Mother and Newborn. Journal of Midw ifery & Women’s Health 52(6):606-613. Jevitt, Cecilia M., Shannon Morse, and Yong Sue O’Donnell 2008 Shoulder Dystocia: Nursing Prevention and Posttrauma Care. Journal of Perinatal & Neonatal Nursing 22(1):14–20. Jia, Haomiao, Peter Muennig, and Elaine Borawski 2004 Comparison of Small-Are Analys is Techniques for Estimating County-Level Outcomes. American Journa l of Preventive Medicine 26(5):453-460. Johnson-Clarke, Fern 2006 Identifying Teen-Birth “HotSpots” Using GIS: District of Co lumbia, 2000-2001. http://dchealth.dc.gov /DOH/frames.asp?doc=/doh/lib/doh/ id_teen_birth_hot_spots2000_2001.pdf Jones, Eric E. 2006 Using Viewshed Analysis to Expl ore Settlement Choice: A Case Study of the Onondaga Iroquois. American Antiquity 71 (3):523-538. Jones, Kelvyn and Craig Duncan 1998 Modelling Context and Hete rogeneity: Applying Mu ltilevel Models. In Research Strategies in the Soci al Sciences: A Guide to New Approaches, Elinor Scarbrough and Etic Tanenbaum, eds. Pp. 95 -123. Oxford: Oxford University Press.

PAGE 158

150 Kaplan, Elaine Bell 1997 Our Kind of Girl: Unraveling the Myths of Black Teenage Motherhood. Berkeley, CA: University of California Press. Krieger, Nancy 1994 Epidemiology and the Web of Causation: Has Anyone Seen the Spider? Social Science and Medicine 39:887-903. 2001 Theories for Social Epidemiology in the 21st Century: An Ecosocial Perspective. International Journal of Epidemiology 30:668-677. Krieger, Nancy, Jarvis T. Chen, Da vid H. Rehkopf, and S.V. Subramanian 2003a Race/Ethnicity, Gender, and Monitori ng Socioeconomic Gradients in Health: A Comparison of Area-Based Socio economic Measures – The Public Health Disparities Geocoding Project. Amer ican Journal of Public Health 93(10):16551671. Krieger, Nancy, Jarvis T. Chen, Pamela D. Waterman, M. J Soobader, S. V. Subramanian, and R. Carson 2003b Choosing Area Based Socioeconomic Measures to Monitor So cial Inequalities in Low Birth Weight and Childhood L ead Poisoning: The Public Health Disparities Geocoding Project (US). Journa l of Epidemiology and Community Health 57: 186-199. Krieger, Nancy, Sally Zierler, Joseph W. H ogan, Pamela Waterman, Jarvis Chen, Kerry Lemieux, and Annie Gjelsvik 2003c Geocoding and Measurement of Neighborhood Socioeconomic Position: A U.S. Perspective. In Neighborhoods and Health, Ichiro Kaachi and Lisa F. Berkman, eds. New York, NY: Oxford University Press. Krieger, Nancy, Pamela D. Waterman, Jarv is Chen, David Rehkopf and SV Subramanian 2004 Geocoding and Monitoring US Socioecono mic Inequalities in Health: An Introduction to Using Area-Based Socio economic Measures -The Public Health Disparities Geocoding Project monogr aph. Boston, MA: Harvard School of Public Health. Electronic document http://www.hsph.harvard.e du/thegeocodingproject/ ), accessed on July 29, 2009. Krieger, Nancy, Jarvis T. Chen, Pamela D. Waterman, David Rehkopf, and S.V. Subramanian 2005 Painting a Truer Picture of US Socioeconomic and Racial/Ethnic Health Inequalities: The Public Health Dispari ties Geocoding Project. American Journal of Public Health 95: 312-323.

PAGE 159

151 Krieger, Nancy, Pamela D. Waterman, Jarv is T. Chen, David H. Rehkopf, and S.V. Subramanian N.d. Geocoding and Monitoring U.S. Socioeconomic Inequalities in Health: An Introduction to Using Area-Bas ed Socioeconomic Measures. The Public Health Disparities Geocoding Projec t monograph. Boston, MA: Harvard School of Public Health. http://www.hsph.harvard.edu/thegeocodingproject/ accessed January 11, 2010. Kroeber, Alfred L. 1939 Cultural and Natural Areas of Native Nort h America. Berkeley : University of California Press. Lang, Laura 2000 GIS for Health Organizations. Redlands, CA: ESRI Press. Lawlor, Debbie A., and Mary Shaw 2002 Too Much Too Young? Teenage Preg nancy is Not a Public Health Problem. International Journal of Epidemiology 31:552-554. 2002a What a Difference a Year Makes? Too Little Too Late. Teenage Pregnancy is Not a Public Health Problem. International Journal of Epidemiology 31:558-559. Lawson, Andrew B. 2001 Statistical Methods in Spatial Epidemiology. New York: John Wiley & Sons. Leadbeter, Bonnie J. and Niobe Way 2001 Growing Up Fast: Transitions to Early Adulthood of Inner-City Adolescent Mothers. Mahwah, NJ: Lawrence Erlbaum Associates, Publishers. Lemen, Paul M., Thomas R. Wigton, Amy J. Miller-CcCarthey, and Dwight P. Cruikshank 1998 Screening for Gestational Diabet es Mellitus in Adolescent Pregnancies. American Journal of Obstetrics and Gynecology 178(6):1251-1256. Leventhal, Tama and Jeanne Brooks-Gunn 2008 Neighborhood Residence and Youth Deve lopment: Empirical Findings and Theoretical Models. Th e Prevention Researcher 15(2):3-6. Loker, William M. 2005 The Rise and Fall of Flue-Cured Tobacco in the Copn Valley and Its Environmental and Social C onsequences. Huma n Ecology 33(3):299-327. Luker, Kristin 1996 Dubious Conceptions: The Politics of Teenage Pregnancy. Cambridge, MA: Harvard University Press.

PAGE 160

152 MacMahon, Brian, Thomas F. Pugh, and Johannes Ipsen 1960 Epidemiologic Methods. Boston: Little Brown and Company. March of Dimes 2004 Teenage Pregnancy: Facts You Should Know. http://www.marchofdimes.com/professionals/681_1159.asp accessed April 2, 2006. Martin, Joyce A., Bradly E. Hamilton, Stepha nie J. Ventura, Fay Menacker, and Melissa M. Park 2002 Births: Final Data for 2000. CDC National Vital Statisti cs Reports 50(5):1-102. Maternal and Child Health Bureau (MCHB) 2004 Child Health USA 2004. Electronic document, http://www.mchb.hrsa.gov/mchirc/c husa_04/pages/0200introduction.htm accessed March 21, 2006. Matte, Thomas D., Michae line Bresnahan, Melissa D. Begg, and Ezra Susser. 2001 Influence of Variation in Birt h Weight Within Normal Range and Within Sibships on IQ at Age 7 Year s: Cohort Study. British Medical Journal 323:310314. Maynard, Rebecca A. 1997 Kids Having Kids: Economic Co sts and Social Consequences of Teen Pregnancy. Washington DC: The Urban Institute Press. Mayo Clinic 2004 Cesarean Birth and the Road to Recovery. Electronic document, http://www.mayoclinic.com /health/c-section/PR00101 accessed May 3, 2006. Meng, Yu and Debbie A. Niemeier 1998 Project Level Carbon Monoxide “H ot-Spot” Analysis for Level of Service D Intersections. Transportation Research Record (1641), 73 – 80. National Campaign to Prevent Teen Pregnancy 2004 General Facts and Stats. Electronic document, http://www.teenpregnancy.org/ resources/data/genlfact.asp accessed April 27, 2006. 2007 Why It Matters: T een Pregnancy and Other Health Issues. Electronic document, http://www.thenationalcampaign.org/why-it-matters/pdf/health.pdf accessed March 30, 2010. 2010 Why It Matters: The Costs of Teen Childbearing. Electronic document, http://www.thenationalcampaign.org/why-it-matters/pdf/costs.pdf accessed March 30, 2010.

PAGE 161

153 National Center for Health Statistics 2001 Births to Teenagers in the Unite d States, 1940-2000. National Vital Statistics Reports, September 25, 2001. 2002 Births: Final Data for 2002. Na tional Vital Statistics Reports, December 17, 2003. National Public Radio 2009 Census Stirs Debate In Washington. Weekend All Things Considered, April 19, 2009 Nuthalapaty, Francis S., and Dwight J. Rouse 2004 The Impact of Obesity on Obstetrica l Practice and Outcome. Clinical Obstetrics and Gynecology 47(4): 898-913. O’Campo, Patricia, Xiaonan Xue, MeiCheng Wang, and Margaret O’Brien Caughy 1997 Neighborhood risk Factor for Lo w Birthweight in Baltimore: A Multilevel Analysis. American Journal of Public Health 87(7):1113-1118. Parsons, Tessa J., Chris Power, and Orly Manor 2001 Fetal and Early Life Growth and Body Mass Index from Birth to Early Adulthood in 1958 British Cohort: Longitudinal Study. British Medical Journal 323:1331-1335. Perper, Kate, Kristen Peterson, and Jennifer Manlove 2010 Diploma Attachment Among Teen Mothers. Child Trends, Fact Sheet: Washington, DC. Electronic document, http://www.childtrends.org/Files//Child_Trends2010_01_22_FS_DiplomaAttainment.pdf accessed March 30, 2010. Phipps, Maureen Glennon, Jeffrey D. Blume, and Sonya M. DeMonner 2002 Young Maternal Age Associated with Increased Risk of Postneonatal Death. Obstetrics and Gynecology 100(3):481-486 Podolsky, Richard 2003 Biodiversity Hotspot Analysis Using Diversidad Soft ware: A Background Document In Preparation of the Laguna Merin Workshop. Publiatti, M., G. Solinas, S. Sotgiu, P. Astiglia, and G. Rosati 2000 Multiple Sclerosis Distribution in Northern Sardinia: Spatial Cluster Analysis of Prevalence. Neurology 59(5):277-282. Raneri, Leslie G. 2006 Which Adolescent Mothers are Most at Risk of repeat Pr egnancy? Abstracts 38:95-96.

PAGE 162

154 Reader, Steven 2000 Analyzing the Sp atial Variability of Low Birth Weight Incidence Among the Black Population of Florida. In Papers and Proceedings of the Applied Geography Conferences, Volume 23, pp. 166-174. Tampa, Florida: Applied Geography Conferences, Inc. Reichman, Nancy E. 2005 Low Birth Weight and School Read iness. School Readines s: Closing Racial and Ethni c Gaps. Volume 15(1):91-116. Rich-Edwards, Janet W., Graham A. Colditz, Meir J. Stampfer, Walter C. Willett, Mathew W. Gillman, Charles H. Hennekens, Frank E. Speizer, and JoAnn E. Manson 1999 Birthweight and The Risk for T ype 2 Diabetes Mellitus in Adult Women. Annals of Internal Medicine 130:322-324. Romero-Daza, Nancy 2004 "Low Birth Weight in Hillsborough County: Combining GIS and Anthropology.” Presentation to Anthropology Club, University of South Florida, April 12, 2004. Rushton, Gerard 2003 Improving the Geographic Basi s of Health Surveillance using GIS. In GIS and Health, Anthony C. Gatrell and Markku Loytonen, eds. Pp. 63-80. Electronic book, http://www.netlibrary.com. ezproxy.lib.usf.edu/Reader/ accessed passim June 2006. Saisana, Michaela and Stefano Tarantola 2002 State-of-the-art Report on Current Met hodologies and Practic es for Composite Indicator Development. Jo int Research Center. http://compositeindicators.jrc.ec.europa.e u/Document/state-of-the-art_EUR20408.pdf accessed October 23, 2008. Sillitoe, Paul 2000 Let Them Eat Cake: Indigenous Know ledge, Science and the 'Poorest of the Poor.' Anthropology Today 16(6):3-7. Singhal, Atul, Jonathan Wells, Tim J. Cole, Mary Fewtrell, and Alan Lucas 2003 Programming of Lean Body Mass: A Link Between Birth Weight, Obesity, and Cardiovascular Disease? Am erican Journal of Clin ical Nutrition 77:726-730. Smith, Gordon C.S., and Jill P. Pell 2001 Teenage Pregnancy and Risk of A dverse Perinatal Outcomes Associated with the First and Second Births: P opulation Based Retrospective Cohort Study. British Medical Journal (233): 476-481.

PAGE 163

155 Smith, Sally 2002 Too Much Too Young? In Nepa l More a Case of Too Little, Too Young. International Journal of Epidemiology 31:557-558. Society for Applied Anthropology 1983 Ethical and Professional Responsib ilities. Electronic document, http://www.sfaa.net/sfaaethic.html accessed February 3, 2010. Solinger, Rickie 1992 Wake Up Little Susie: Single Pregnancy and Race Before Roe v. Wade. New York, NY: Routledge. Stack, Carol B. 1974 All Our Kin: Strategies for Survival in a Black Community. New York, NY: Harper & Row. Steward, Julian H. 1938 Basin-Plateau Aboriginal Sociopolit ical Groups. Bureau of American Ethnology, Bulletin 120. Washington, D.C. Stonich, Susan C. 2002 The Future of Spatial Analysis in Anthropology. Workshop on Spatial Analysis in Anthropology at the 101st Annual Meeting of the American Anthropological Association, New Orleans, LA, November 21, 2002. Tatian, Peter A. 2003 Census CD Neighborhood Change Database (NCDB) Data Users’ Guide Long Form Release. Washington, D.C.: The Urban Institute. Taylor, Don and Gilberto Chavez 2002 Small Area Analysis on a Large Scale – The California Experience in Mapping Teenage Birth “Hot Spots” for Res ource Allocation. Journal of Public Health Management Practice 8(2):33-45. Turok, David K., Stephen D. Ratc liffe, and Elizabeth G. Baxley 2003 Management of Gestational Diabet es Mellitus. American Family Physician 68(9): 1767-1780.

PAGE 164

156 U.S. Census Bureau N.d. American FactFinder Datasets. http://factfinder.census.gov/servlet/Da tasetMainPageServlet?_program=DEC&_ta bId=DEC2&_submenuId=datasets_1&_lang=en&_ts=266574506897 N.d.(a) Population Estimates. http://www.census.gov/popes t/datasets.html#cntyinter N.d.(b) Standard Hierarchy of Census Geographic Entities. http://www.census.gov/geo/www/geodiagram.pdf 1994 Geographical Areas Reference Manual. Washington, DC: U.S. Department of Commerce. Electronic document, http://www.census.gov/geo/www/garm.html ), accessed February 5, 2004. 1999 Net Undercount and Undercount Ra te for U.S. and States (1990), http://www.census.gov/dmd/www/pdf/understate.pdf accessed January 17, 2010. 2000a Census 2000: Census 2000 Ge ographic Terms and Concepts, Appendix A Electronic document, http://www.census.gov/geo/www/tiger/glossry2.pdf ), Accessed on February 5, 2004 at 2000b Race Data: Racial and Ethnic Classifications Used in Census 2000 and Beyond. http://www.census.gov/population/www/ socdemo/race/racefactcb.html 2002 Comparing SF 3 Estimates with Corresponding Values in SF 1 and SF 2. Last Revised: August 08, 2002 at 12:00:02 PM. Electronic document, http://www.census.gov/Press-Re lease/www/2002/sf3compnote.html ), accessed February 5, 2004 U.S. National Library of Medicine and the National Institutes of Health 2009 Medline Plus. Electronic document, http://www.nlm.nih.gov/medlineplus/ accessed March 29, 2010. Wang, Xinhao and David P. Varady 2005 Using Hot-Spot Analysis to Study the Clustering of Section 8 Housing Voucher Families. Housing Studies 20 (1):29-48. West, Kirsten and J. Gregory Robinson 1999 What Do We Know About The U ndercount of Children? Population Division Working Paper No. 39, August 1999. Wash ington, D.C.: U.S. Census Bureau, Population Division. Accessed Sept 12, 2009, http://www.census.gov/population/www/documentation/twps0039/twps0039.html Whiteford, Linda M. 2000 Local Identity, Gl obalization, and Health in Cuba and the Dominican Republic. In Global Health Policy, Local Realitie s: The Fallacy of the Level Playing Field, Linda M. Whiteford and Lenore Mande rson, eds. Boulder, CO: Lynne Rienner Publishers. Whiteford, Linda M. and Robert T. Trotter II 2008 Ethics for Anthropological Research and Practice. Long Grove, IL: Waveland Press.

PAGE 165

157 APPENDICES

PAGE 166

158 0 2.5 5 miles Appendix A: Maps 2000 Census Tracts, Hillsborough County, Florida 106 104.01 107.02 104.02 9 109 110.10 50 49 43 51.02 40 41 42 61 140.04 138.02 138.03 136.01 138.01 139.08 141.06 141.05 132.07 131 134.09 139.09 139.10 134.08 134.05 134.06 135.03 135.05 133.14 133.05 132.05 133.12 123.04 122.08 122.07 134.07 132.06 132.04 124.03 141.07 141.08 139.07 37 102.04 102.03 103.04 125.01 125.02 126 127.02 127.01 130.02 114.11 111.07 115.15 115.16 103.03 102.07 102.08 102.06 110.09 8 10 111.08 11 5 14 116.05 23 114.12 115.08 117.05 117.07 117.08 117.03 47 65 71 72 140.02 140.03 1 2 3 4.01 4.02 6 7 12 13 151617 18 19 20 21 22 25 26 29 30 31 3233 34 35 36 38 39 44 45 48 57 58 60 62 63 66 68.01 70 101.03 101.05 101.06 101.07 101.08 102.05 103.05 105 107.01 108.03 108.04 108.05108.06 108.07 108.08 110.03 110.05 110.06 110.07 110.08 110.11 111.03 111.04 111.05 111.06 112.03 112.04 112.05112.06 113.01 113.02 114.06 114.07 114.08 114.09 114.10 114.13 114.14114.15 114.16 115.04 115.05 115.06 115.07 115.09 115.10 115.11 115.12 115.13 115.14 116.03 116.06 116.07 116.08 116.09 116.10116.11 116.12 116.13117.06 118.02 118.03 118.04 119.01 119.02 119.03 120.01 120.02 121.03 121.04 121.05 121.06 122.04 122.05 122.06 123.01123.03 124.01 124.02 128 129 130.01 130.03 130.04 132.03 132.08 133.06 133.07 133.08133.09 133.10 133.11 133.13 134.04 135.04 136.02 138.05 139.03 139.11 139.12 140.06 46 64 59 73 68.02 135.01 53 67 69 54 55 137.01 24 137.02 138.04 27 140.05 141.04 28 51.01 141.09

PAGE 167

0 2.5 5 miles Appendix A (Continued) 159 2000 Census Tracts, Pinellas County, Florida 268.16 268.09 201.03 201.05 202.04 285 284.01 284.02 202.02 203.01 229.02 229.01 210 213 215 237 281.02 282 224.02 251.06 246.02 230 250.13 250.16 245.05 251.10 277.01 276.01 251.08 253.01 251.15 251.07 251.09 251.21 251.12 250.11 250.10 253.05 253.06 252.06 250.15 250.14 240.05 240.04 249.02 245.06 245.07 245.02 253.03 252.03 254.07 261 262 254.01 257 255.04 276.02 263 265 271.03 268.15 268.14 269.11 269.10 269.07 270 271.05 269.09 268.1 269.04 268.11 273.20 273.17 273.18 273.15 273.16 273.13 272.02 272.04 274.01 275.01 228.02 226.02 250.12 228.01 227 220 221 226.01 225.03 248.01 250.04 269.05 267.02 266.02 254.10 254.11 250.09 249.04 249.06 249.01 247 231 219 208 283 222 225.01 225.02 248.02 251.13 251.16 255.03 255.01 267.01 268.12 268.17 268.13 268.04 267.03 266.01 254.09 254.08 249.05 246.01 233 234 218 209 206 207 201.01 223.01 223.02 251.20 251.19 252.07 253.04 256.01 256.02 258 264 269.08 272.08 268.08 243.02 242 241 232 245.03 243.01 244.03 235 216 212 205 202.01 274.02 272.07 272.06 272.05 273.14 273.12 273.11 274.03 279.02 271.01 254.05 202.05 260.02 260.01 236 204 280.02 278 275.02 259.02 273.10 280.01 279.01 281.01 252.04 259.01 272.01 214 244.05 203.02 250.07 273.19 254.04 244.06 240.02 238 224.01 251.18 271.04 273.08 244.07 245.08 239 244.04 252.05 273.09 277.02 240.01 251.11 251.14 250.01

PAGE 168

Births to 18-19 Year-Old White Adolescents (1992-1997) 0510milesAppendix A (Continued) 0510miles 0510miles160 Births to 18-19 Year-Old White Adolescents (1992-1997) TARPO N SPRINGS TARPO N SPRINGS TARPO N SPRINGS TARPO N SPRING S TARPO N SPRING S TARPON SPRING S TARPON SPRING S TARPON SPRING S TARPO N SPRING S SAFETY HARBOR O LDSMAR O LDSMAR O LDSMAR O LDSMAR O LDSMAR OLDSMAR OLDSMAR OLDSMAR O LDSMAR LARGO PINELLAS PARK BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH SEMINOLE SEMINOLE SEMINOLE SEMINOLE SEMINOLE SEMINO LE SEMINO LE SEMINO LE SEMINOLE MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACHST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ CARROLLWO OD RIVERVIEW SUMMERFIELD BALM/WIMAUMA BRANDON APOLLO BEACH APOLLO BEACH APOLLO BEACH APO LLO BEACH APO LLO BEACH APO LLO BEACH APO LLO BEACH APO LLO BEACH APO LLO BEACH G IBSO NTON G IBSO NTON G IBSO NTON GIBSO NTO N GIBSO NTO N GIBSONTO N GIBSONTO N GIBSONTO N GIBSO NTO N PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER THO NO TOSASSA NEW TAMPA RUSKIN TOWN-NCOUNTRY VALRICO CITRUS PARK SEFFNER PLANT CITY SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SPRING S SPRINGS SPRING S SPRING S SPRING S SPRINGS SPRINGS SPRINGS SPRING SI 2 7 5U S 3 0 1I 2 7 5U S 3 0 1I 7 5U S 4 1G a n d y B lv d THONOTOSASSA SEFFNER LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ CARRO LLWOO D PLANT CITY BRANDON VALRICO RIVERVIEW SUMMERFIELD APO LLO BEACH APO LLO BEACH APO LLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH G IBSO NTON G IBSO NTON G IBSO NTON GIBSO NTO N GIBSO NTO N GIBSONTO N GIBSONTO N GIBSONTO N GIBSO NTO N NEW TAMPA BALM/WIMAUMA PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER RUSKIN CITRUS PARK TO WN-NCOUNTRY LARG O PINELLAS PARK INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN RO CKS BEACH BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR SEMINOLE SEMINO LE SEMINO LE SEMINOLE SEMINOLE SEMINOLE SEMINOLE SEMINOLE SEMINOLE CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBO R SAFETY HARBO R SAFETY HARBO R SAFETY HARBOR TARPO N SPRING S TARPO N SPRING S TARPO N SPRING S TARPON SPRING S TARPON SPRING S TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRING S OLDSMAR OLDSMAR OLDSMAR O LDSMAR O LDSMAR OLDSMAR OLDSMAR OLDSMAR O LDSMAR SULPHUR SPRINGSAlt US 19I 2 7 5U S 1 9I 2 7 5I 4I 7 5U S 4 1G a n d y B lv dU S 3 0 1U S 3 0 1

PAGE 169

Births to 13-17 Year-Old Black Adolescents (1992-1997) Births to 18-19 Year-Old Black Adolescents (1992-1997) 0510miles 0510miles161 Appendix A (Continued) CITRUS PARK APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APO LLO BEACH APO LLO BEACH APO LLO BEACH APOLLO BEACH BALM/WIMAUMA SUMMERFIELD RIVERVIEW G IBSO NTON G IBSO NTON G IBSO NTON GIBSONTO N GIBSONTO N GIBSONTO N GIBSONTO N GIBSONTO N GIBSONTO N RUSKIN LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ PLANT CITY CARRO LLWOO D TO WN-NCO UNTRY PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER BRANDON SEFFNER VALRICO THO NO TO SASSA NEW TAMPA DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN SAFETY HARBO R SAFETY HARBO R SAFETY HARBO R SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMARO LDSMAR O LDSMAR O LDSMAR OLDSMAR TARPON SPRING S TARPON SPRING S TARPON SPRING S TARPON SPRINGS TARPON SPRINGS TARPO N SPRINGS TARPO N SPRINGS TARPO N SPRINGS TARPON SPRINGS CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR PINELLAS PARK SEMINOLE SEMINOLE SEMINOLE SEMINO LE SEMINO LE SEMINOLE SEMINOLE SEMINOLE SEMINO LE MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN RO CKS BEACH LARG O ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG SULPHUR SPRING SI 2 7 5U S 1 9G a n d y B lv dA l t U S 1 9I 4I 7 5U S 4 1U S 3 0 1I 2 7 5 NEW TAMPA THONOTO SASSA SEFFNER BALM/WIMAUMA PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER BR ANDO N VALRICO RIVERVIEW SUMMERFIELD CARROLLWO OD LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ APO LLO BEACH APO LLO BEACH APO LLO BEACH APO LLO BEACH APO LLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APO LLO BEACH RUSKIN CITRUS PARK TO WN-NCO UNTRY G IBSONTON G IBSONTON G IBSONTON G IBSO NTON G IBSO NTON GIBSONTO N GIBSONTO N GIBSONTO N G IBSO NTON PLANT CITY LARGO INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR SEMINOLE SEMINOLE SEMINOLE SEMINO LESEMINO LE SEMINO LE SEMINO LE SEMINO LE SEMINO LE PINELLAS PARK ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN SAFETY HARBO R SAFETY HARBO R SAFETY HARBO R SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR TARPON SPR INGS TARPON SPR INGS TARPON SPR INGS TARPON SPRINGS TARPON SPRINGS TARPO N SPRING S TARPO N SPRING S TARPO N SPRING S TARPON SPRINGS OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR O LDSMAR O LDSMAR O LDSMAR OLDSMAR SULPHUR SPRINGSU S 1 9Alt US 19I 2 7 5G a n d y B lv dU S 3 0 1I 7 5U S 4 1I 4I 2 7 5U S 3 0 1

PAGE 170

Births to 13-17 Year-Old Hispanic Adolescents (1992-1997) Births to 18-19 Year-Old Hispanic Adolescents (1992-1997) 162 Appendix A (Continued) 0510miles 0510miles NEW TAMPA THONOTOSASSA SEFFNER BALM/WIMAUMA PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER BRANDO N VALRICO RIVERVIEW SUMMERFIELD CARRO LLWOO D LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ APOLLO BEACH APOLLO BEACH APOLLO BEACH APO LLO BEACH APO LLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APO LLO BEACH RUSKIN CITRUS PARK TOWN-NCO UNTRY GIBSONTO N GIBSONTO N GIBSONTO N G IBSO NTON G IBSO NTON G IBSO NTON G IBSO NTON G IBSO NTON G IBSO NTON PLANT CITY LARG O INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN RO CKS BEACH BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR SEMINOLE SEMINOLE SEMINOLE SEMINOLE SEMINOLE SEMINO LESEMINO LE SEMINO LE SEMINOLE PINELLAS PARK ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBO R SAFETY HARBO R SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBO R TARPO N SPRING S TARPO N SPRING S TARPO N SPRING S TARPON SPRING S TARPON SPRING S TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRING S O LDSMAR O LDSMAR O LDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR SULPHUR SPRINGSU S 1 9Alt US 19I 2 7 5G a n d y B lv dU S 3 0 1I 7 5U S 4 1I 4I 2 7 5U S 3 0 1 SUMMERFIELD APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APO LLO BEACH APO LLO BEACH APO LLO BEACH APOLLO BEACH NEW TAMPA THO NO TO SASSA SEFFNER BALM/WIMAUMA PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER BRANDO N VALRICO RIVERVIEW CARROLLWOO D LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ RUSKIN CITRUS PARK TO WN-NCO UNTR Y GIBSO NTON GIBSO NTON GIBSO NTON GIBSONTO N GIBSONTO N G IBSONTO N G IBSONTO N G IBSONTO N GIBSONTO N PLANT CITY TARPO N SPRINGS TARPO N SPRINGS TARPO N SPRINGS TARPON SPRING S TARPON SPRING S TARPON SPRING S TARPON SPRING S TARPON SPRING S TARPON SPRING S SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBO R SAFETY HARBO R SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR O LDSMAR O LDSMAR O LDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR BELLEAIR LARG O INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN ROCKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH INDIAN RO CKS BEACH PINELLAS PARK SEMINO LE SEMINO LE SEMINO LE SEMINOLE SEMINOLE SEMINOLE SEMINOLE SEMINOLE SEMINOLE ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEAC H MADEIRA BEAC H MADEIRA BEAC H MADEIRA BEAC H MADEIRA BEAC H MADEIRA BEAC H DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN SULPHUR SPRINGSU S 1 9Alt US 19I 2 7 5I 7 5U S 4 1G a n d y B lv dU S 3 0 1I 4I 2 7 5U S 3 0 1

PAGE 171

Legend Hot Spot Cold Spot Major Road 0510milesAppendix A (Continued) 0510milesHot and Cold Spots: 18-19 Year-Old White Adolescents Hot and Cold Spots: 13-17 Year-Old White Adolescents 163 Legend Hot Spot Cold Spot Major Road OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGSTARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGSU S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v dI 2 7 5 I 2 7 5 I 2 7 5 I -2 7 5 I 2 7 5 I -2 7 5 I 2 7 5 I 2 7 5 I 2 7 5A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 19 A l t U S 1 9I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARKTOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER TARPON SPRINGS TARPON SPRINGS TARPON SPRINGSTARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGSG a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v dA l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 Al t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9I -2 7 5 I -2 7 5 I -2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARKCITRUS PARK CITRUS PARK TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKINRUSKIN RUSKIN RUSKIN THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA

PAGE 172

Hot and Cold Spots: 13-17 Year-Old Black Adolescents Appendix A (Continued) Hot and Cold Spots: 18-19 Year-Old Black Adolescents 164 0510miles Legend Hot Spot Cold Spot Major Road 0510miles Legend Hot Spot Cold Spot Major Road TEMPLE TERRACE TEMPLE TERRACE TEMPLE TERRACE TEMPLE TERRACE TEMPLE TERRACE TEMPLE TERRACE TEMPLE TERRACE TEMPLE TERRACE TEMPLE TERRACE CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT SAFETY HARBORSAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACHG a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v dI 2 75 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 19 A l t U S 1 9 A l t U S 19 A l t U S 19 A l t U S 1 9 A l t U S 19U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK EAST LAKE EAST LAKE EAST LAKE EAST LAKE EAST LAKE EAST LAKE EAST LAKE EAST LAKE EAST LAKE TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER RIVERVIEWRIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGSTARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGSI 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v dU S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA CITRUS PARKCITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKINRUSKIN RUSKIN RUSKIN THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA

PAGE 173

Appendix A (Continued) Hot and Cold Spots: 18-19 Year-Old Hispanic Adolescents Hot and Cold Spots: 13-17 Year-Old Hispanic Adolescents 165 0510miles Legend Hot Spot Cold Spot Major Road 0510miles Legend Hot Spot Cold Spot Major Road G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v dI 2 75 I 2 75 I 2 7 5 I 2 7 5 I 2 7 5 I -2 7 5 I 2 7 5 I 2 7 5 I 2 7 5U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9Al t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH INDIAN ROCKSINDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE LUTZ LUTZ LUTZ LUTZ LUTZ LUTZLUTZ LUTZ LUTZ KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACHAPOLLO BEACH APOLLO BEACH APOLLO BEACH RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH A N ROCKS A N ROCKS A N ROCKS A N ROCKS A N ROCKS IA N ROCK S IA N ROCK S IA N ROCK S A N ROCKS B EACH B EACH B EACH B EACH B EACH BEACH BEACH BEACH B EACH LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGSU S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v dI 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 19 A l t U S 19 A l t U S 1 9I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARKCITRUS PARK CITRUS PARK TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN THONOTOSASSATHONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA

PAGE 174

Index of Socio-Economic Inequality Very High Inequality High Inequality Low Inequality Very Low Inequality Legend Hot Spot Cold Spot Major Road 0510milesAppendix A (Continued) Index of Socio-Economic Inequality Very High Inequality High Inequality Low Inequality Very Low Inequality Legend Hot Spot Cold Spot Major Road 0510milesHot and Cold Spots: 18-19 Year-Old White Adolescents with Socio-Economic Inequality Index Hot and Cold Spots: 13-17 Year-Old White Adolescents with Socio-Economic Inequality Index 166 ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACHST PETE BEACH TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGSG a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v dI 2 7 5 I 2 7 5 I 2 7 5 I -2 7 5 I 2 7 5 I -2 7 5 I 2 7 5 I 2 7 5 I 2 7 5U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 19 A l t U S 19 A l t U S 1 9I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA CITRUS PARKCITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGSTARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGSG a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v dA l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARKCITRUS PARK TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA

PAGE 175

Index of Socio-Economic Inequality Very High Inequality High Inequality Low Inequality Very Low Inequality Legend Hot Spot Cold Spot Major Road 0510miles Index of Socio-Economic Inequality Very High Inequality High Inequality Low Inequality Very Low Inequality Legend Hot Spot Cold Spot Major Road 0510milesHot and Cold Spots: 13-17 Year-Old Black Adolescents with Socio-Economic Inequality Index Appendix A (Continued) Hot and Cold Spots: 18-19 Year-Old Black Adolescents with Socio-Economic Inequality Index 167 TEMPLE TERRA CE TEMPLE TERRA CE TEMPLE TERRA CE TEM PLE TERRACE TEM PLE TERRACE TEMPLE TERRACE TEMPLE TERRACE TEMPLE TERRACE TEM PLE TERRACE CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD THONOTOSAS SA THONOTOSAS SA THONOTOSAS SA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINE LLAS PARK PINE LLAS PARK PINE LLAS PARK PINELLAS PARK ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN GULFPORT GULFPORT GULFPORT GULFPORT GULFPORTGULFPORT GULFPORT GULFPORT GULFPORT SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFE TY HARBOR SAFE TY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFE TY HARBOR INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS BEACH BEACH BEACH BEACH BEACH B EACH B EACH B EACH BEACH LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRI NGS TARPON SPRI NGS TARPON SPRI NGS TARPON SPRINGS ST PETE BE ACH ST PETE BE ACH ST PETE BE ACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACHG a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v dI 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS SULPHUR SPRINGS KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH K NIGHTS/ANTIOCH K NIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH K NIGHTS/ANTIOCH NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK EAST LAKE EAST LAKE EAST LAKE EAST LAKE EAST LAKE EAST LAKE EAST LAKE EAST LAKE EAST LAKE TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO PA LM RIVER PA LM RIVER PA LM RIVER PALM RIVER PALM RIVER P ALM RIVER P ALM RIVER P ALM RIVER PALM RIVER RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD APOLLO BE ACH APOLLO BE ACH APOLLO BE ACH APOLLO BE ACH APOLLO BE ACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BE ACH RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGSTARPON SPRINGS TARPON SPRINGSI 2 7 5 I 2 7 5 I 2 7 5 I 27 5 I 27 5 I 2 75 I 2 75 I 2 75 I 27 5G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v dU S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD CARROLLWOOD PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA CITRUS PARKCITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH RUSKIN RUSKIN RUSKIN RUSKIN RUSKINRUSKIN RUSKIN RUSKIN RUSKIN THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA

PAGE 176

Appendix A (Continued) 0510miles Index of Socio-Economic Inequality Very High Inequality High Inequality Low Inequality Very Low Inequality Legend Hot Spot Cold Spot Major Road Hot and Cold Spots: 18-19 Year-Old Hispanic Adolescents with Socio-Economic Inequality Index Index of Socio-Economic Inequality Very High Inequality High Inequality Low Inequality Very Low Inequality Legend “Hot Spot” “Cold Spot” Major Road 0510milesHot and Cold Spots: 13-17 Year-Old Hispanic Adolescents with Socio-Economic Inequality Index 168 G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v dI 2 7 5 I 2 75 I 2 75 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER CLEARWATER OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR OLDSMAR ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACHMADEIRA BEACH MADEIRA BEACH MADEIRA BEACH INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH BEACH LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO LARGO ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BEACH TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA CITY OF TAMPA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONELUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACHAPOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG ST PETERSBURG S T PETERSBURG S T PETERSBURG S T PETERSBURG ST PETERSBURG CLEARWATER CLEARWATER CLEARWATER CLEARWATE R CLEARWATE R CLE ARWATER CLE ARWATER CLE ARWATER CLEARWATE R SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR SAFETY HARBOR DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN DUNEDIN OLDSMAR OLDSMAR OLDSMAR OLDS MAR OLDS MAR OLDSMAR OLDSMAR OLDSMAR OLDS MAR GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT GULFPORT PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK PINELLAS PARK P INELLAS PARK P INELLAS PARK P INELLAS PARK PINELLAS PARK MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BEACH MADEIRA BE ACH MADEIRA BE ACH MADEIRA BE ACH MADEIRA BEACH INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS INDIAN ROCKS BEACH BEACH BEACH BEACH BEACH BEACHBEACH BEACH BEACH LARGO LARGO LARGO LARGO LARGO LA RGO LA RGO LA RGO LARGO ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BE ACH ST PETE BE ACH ST PETE BEACH ST PETE BEACH ST PETE BEACH ST PETE BE ACH TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARP ON SPRINGS TARP ON SPRINGS TARPON SPRINGS TARPON SPRINGS TARPON SPRINGS TARP ON SPRINGSG a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v d G a n d y B l v dI 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9 U S 1 9A l t U S 19 A l t U S 19 A l t U S 1 9 A l t U S 19 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 1 9 A l t U S 19I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5 I 7 5U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1 U S 3 0 1U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1 U S 4 1I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4 I 4I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5 I 2 7 5PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY PLANT CITY BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA BALM/WIMAUMA SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SUN CITY CENTER SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SULPHUR SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS SPRINGS FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST FOREST HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS HILLS KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE KEYSTONE LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ LUTZ KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH KNIGHTS/ANTIOCH NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA NEW TAMPA CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK CITRUS PARK TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRYTOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY TOWN-N-COUNTRY SEFFNE R SEFFNE R SEFFNE R SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER SEFFNER BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON BRANDON VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO VALRICO PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER PALM RIVER RIVE RVIEW RIVE RVIEW RIVE RVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW RIVERVIEW B LOOMINGDALE B LOOMINGDALE B LOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE BLOOMINGDALE LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA LITHIA GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON GIBSONTON SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD SUMMERFIELD APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH APOLLO BEACH RUSKIN RUSKINRUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN RUSKIN THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSASSA THONOTOSAS SA THONOTOSAS SA THONOTOSAS SA THONOTOSASSA

PAGE 177

ABOUT THE AUTHOR Kathleen Maes moved to Tampa in 1993 from Green Bay, Wisconsin. In 1996, she completed her Master’s Degree in Applie d Anthropology at the University of South Florida, and shortly after, decided to pursue her Ph.D. Over the past 17 years, Kathleen has b een employed as a Research Associate at the Lawton and Rhea Chiles Center for Healthy Mothers and Babies at the University of South Florida and at the Juveni le Welfare Board in Pinellas County, Florida. Dr. Maes currently works in the Research and Evalua tion Department of the Children’s Board of Hillsborough County.