USF Libraries
USF Digital Collections

Beyond academic reputation

MISSING IMAGE

Material Information

Title:
Beyond academic reputation factors that influence the college of first choice for high achieving students
Physical Description:
Book
Language:
English
Creator:
Schoenherr, Holly J
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
CIRP
College choice process
Financial aid
High academic achievement
Selectivity
Dissertations, Academic -- Adult, Career and Higher Education -- Doctoral -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Studies that have investigated college choice factors for high-achieving students repeatedly cite academic reputation as one of the top indicators of choice but have not indicated why some high-achieving students choose to attend universities with a less prestigious reputation than the more highly prestigious options available to them. The purpose of this study was to examine whether differences exist between traditional-aged high achieving students who choose to attend higher-tiered universities and their peers who choose to attend lower-tiered universities.Independent variables were selected based upon Hossler and Gallagher's (1987) three-stage model and previous research findings in the literature and grouped according to: (1) students' individual and family characteristics, including ethnicity, gender, parents' education level, and family income; (2) institutional characteristics, including financial considerations and academic reputation; and (3) the influence of others, including parents, relatives, teachers and counselors. The sample was drawn from the 97 universities which administered the CIRP Freshman Survey in 2004. Data were used for students who were attending their first choice college located more than 100 miles from home. Data were used from students who had received scores at or above 660 on the SAT Verbal, and scores at or above 670 on the SAT Math. For students who did not report scores for both SAT verbal and SAT math, the researcher accepted data from students reporting an ACT composite score of 30 or higher.In addition, in order for their data to be used, students were required to have an A or A+ average in high school. Results were reported as (1) frequencies and descriptive statistics, (2) a correlation matrix, and (3) multiple regression models. The study found the availability of financial aid to be the most important factor in predicting whether students will attend a higher-tiered or lower-tiered university. Although college costs and academic reputation were found to be significant predictors of the tier level of university attended, they were of secondary importance compared with the attention to financial aid by high achieving students.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2009.
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 Holly J. Schoenherr.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 145 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:
aleph - 002029018
oclc - 436765258
usfldc doi - E14-SFE0002842
usfldc handle - e14.2842
System ID:
SFS0027159:00001


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 2200397Ka 4500
controlfield tag 001 002029018
005 20090915140557.0
007 cr mnu|||uuuuu
008 090915s2009 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0002842
035
(OCoLC)436765258
040
FHM
c FHM
049
FHMM
090
LB2322 (Online)
1 100
Schoenherr, Holly J.
0 245
Beyond academic reputation :
b factors that influence the college of first choice for high achieving students
h [electronic resource] /
by Holly J. Schoenherr.
260
[Tampa, Fla] :
University of South Florida,
2009.
500
Title from PDF of title page.
Document formatted into pages; contains 145 pages.
Includes vita.
502
Dissertation (Ph.D.)--University of South Florida, 2009.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
520
ABSTRACT: Studies that have investigated college choice factors for high-achieving students repeatedly cite academic reputation as one of the top indicators of choice but have not indicated why some high-achieving students choose to attend universities with a less prestigious reputation than the more highly prestigious options available to them. The purpose of this study was to examine whether differences exist between traditional-aged high achieving students who choose to attend higher-tiered universities and their peers who choose to attend lower-tiered universities.Independent variables were selected based upon Hossler and Gallagher's (1987) three-stage model and previous research findings in the literature and grouped according to: (1) students' individual and family characteristics, including ethnicity, gender, parents' education level, and family income; (2) institutional characteristics, including financial considerations and academic reputation; and (3) the influence of others, including parents, relatives, teachers and counselors. The sample was drawn from the 97 universities which administered the CIRP Freshman Survey in 2004. Data were used for students who were attending their first choice college located more than 100 miles from home. Data were used from students who had received scores at or above 660 on the SAT Verbal, and scores at or above 670 on the SAT Math. For students who did not report scores for both SAT verbal and SAT math, the researcher accepted data from students reporting an ACT composite score of 30 or higher.In addition, in order for their data to be used, students were required to have an A or A+ average in high school. Results were reported as (1) frequencies and descriptive statistics, (2) a correlation matrix, and (3) multiple regression models. The study found the availability of financial aid to be the most important factor in predicting whether students will attend a higher-tiered or lower-tiered university. Although college costs and academic reputation were found to be significant predictors of the tier level of university attended, they were of secondary importance compared with the attention to financial aid by high achieving students.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
590
Advisor: Donald Dellow, Ed.D.
653
CIRP
College choice process
Financial aid
High academic achievement
Selectivity
690
Dissertations, Academic
z USF
x Adult, Career and Higher Education
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.2842



PAGE 1

Beyond Academic Reputation: Factors that Influence the College of First Choice for High Achieving Students by Holly J. Schoenherr A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Adult, Car eer and Higher Education College of Education University of South Florida Major Professor: Donald Dellow, Ed.D. Alan Balfour, Ph.D. James Eison, Ph.D. W. Robert Sullins, Ed.D. Date of Approval: March 5, 2009 Keywords: CIRP; College Choice Process; Financial Aid; High Academic Achievement; Selectivity Copyright 2009, Holly J. Schoenherr

PAGE 2

ACKNOWLEDGEMENTS There are many individuals who provided me with much-needed support. I would not have considered embarking on this jour ney if not for the unwavering encouragement of my husband, Kevin. His support, in the form of taking on the full-load of parenting and household responsibilities, car ried me through every stage of the program. He is an exceptional husband and father, and I appreciate him more than he may ever understand. My sons, Bryce and Ren, constant sources of energy and motivation, kept me grounded. A project of this magnitude could not be completed without the guidance of a dedicated and competent faculty committee. I especially thank my major professor, Dr. Don Dellow, for his good humor and sage dir ection throughout the process. My sincere thanks are given as well to the remaining me mbers of my committee, Drs. Alan Balfour, Jim Eison, and Bob Sullins, who were always willing to coach and encourage me along the way. It is important to recognize, as we ll, the guidance of Dr. Jan Ignash, who would be named among the members of my doctora l committee if other life opportunities had not presented themselves. She was a source of strength through the defense of my proposal. I want to also acknow ledge the priceless assistan ce of Dr. Roger Boothroyd, who served as an able guide through the treacherous journey of statistical software. Finally, it is my personal relationship with Jesus Christ and His grace that provided me with continua l strength and endurance.

PAGE 3

i TABLE OF CONTENTS List of Tables iii Abstract v Chapter One Introduction 1 College Choice Models 6 Statement of the Problem 10 Purpose of the Study 11 Significance of the Study 13 Limitations 15 Delimitations 16 Assumptions 17 Definitions 18 Chapter Two – Literature Review 19 Academic Reputation 20 College Choice Models 25 College Choices of Hi gh-Achieving Students 30 Student and Family Characteristics 34 Gender and Ethnicity 34 Socioeconomic Status 35 Education Level of Parents 37 Institutional Characteristics 38 Location/Proximity to Home 39 Cost and Availability of Financial Aid 39 Reputation and Prestige 41 Influence of Others 42 The Cooperative Institutional Research Program (CIRP) 44 Summary 46 Chapter Three Methods 47 Research Questions 47 Research Design 48 Secondary Data Considerations 49 Survey Instrument 50 Validity and Reliabilty 52 Sample 53

PAGE 4

ii Data Collection 55 Description of Variables 58 Data Analysis 59 Summary 60 Chapter Four Results 63 Descriptive Statistics 65 Student and Family Characteristics 66 Gender and Ethnicity 66 Socioeconomic Status 68 Education Level of Parents 71 Institutional Characteristics 74 Cost and Availability of Financial Aid 74 Reputation and Prestige 76 Influence of Others 79 Correlations for Independent Variables 82 Student and Family Characteristics 83 Institutional Characteristics 87 Influence of Others 88 Multiple Regression Analysis 90 Summary 94 Chapter Five – Discussion and Conclusions 96 Relationship between Individual and Family Characteristics 97 and College Choice Relationship between Institutional Ch aracteristics and College Choice 101 Relationship of the Influence of Others to College Choice 106 Validation of the College Choice Model 110 Implications for Future Research 112 Implications for Practice 114 Limitations 116 Conclusion 117 References 121 Appendices 136 A: CIRP 2004 Freshman Survey Questionnaire 137 B : CIRP Freshman Survey: Reliability and Validity 141 C: National Universities participa ting in the 2004 Freshman Survey 145 About the Author End Page

PAGE 5

iii LIST OF TABLES Table 2.1 U.S. News and World Report Indicators and Weights 22 for the 2008 College Rankings Table 2.2 Largest Positive Changes in Peer Assessment Scores 24 from 1998-2007 Table 2.3 Largest Positive Changes in 25th Percentile SAT Scores for Lower-Tier Universities from 2002-2007 31 Table 3.1 Independent Variable Constructi on and Coding Scheme 61 Table 4.1 Distribution of Sample SAT and AC T Scores by Tier of 65 Institution Table 4.2 Distribution of Respondent Gender by Tier of Institution 67 Table 4.3 Distribution of Respondent Ethnicity by Tier of Institution 68 Table 4.4 Distribution of Family Income by Tier of Institution 69 Table 4.5 Student Employment Needs by Tier of Institution 70 Table 4.6 Distribution of Father’s Education Level by Tier of Institution 72 Table 4.7 Distribution of Mother’s Education Level by Tier of Institution 73 Table 4.8 Distribution of Parents’ Education Level by Tier of Institution 74 Table 4.9 Importance of College Costs by Tier of Institution 75 Table 4.10 Importance of Financial Aid by Ti er of Institution 76 Table 4.11 Importance of Academic Reputation by Tier of Institution 77 Table 4.12 Importance of Media Rankings by Tier of Institution 78 Table 4.13 Importance of Parental Influence by Tier of Institution 79

PAGE 6

iv Table 4.14 Importance of Relative Influence by Tier of Institution 80 Table 4.15 Importance of Teacher Influence by Tier of Institution 81 Table 4.16 Importance of Counselor Influence by Tier of Institution 82 Table 4.17 Matrix of Correlation Coefficients 85 Table 4.18 Unstandardized Regression Coefficients for Models 91 Table 4.19 Standardized Regression Coeffi cients for Models 93 Table 5.1 Summary of Relationships be tween Independent and 109 Dependent Variables

PAGE 7

v ABSTRACT Studies that have investig ated college choice factors for high-achieving students repeatedly cite academic reputation as one of the top indicators of choice but have not indicated why some high-achieving students ch oose to attend universities with a less prestigious reputation than the more highly prestigious options available to them. The purpose of this study was to examine whether differences exist betw een traditional-aged high achieving students who choose to attend hi gher-tiered universities and their peers who choose to attend lower-tiered universities. Independent variables were selected based upon Hossler and Gallagher’s (1987) three-stage model and previous research fi ndings in the literature and grouped according to: (1) students’ individual and family ch aracteristics, includ ing ethnicity, gender, parents’ education level, and family income; (2) institutional characteristics, including financial considerations and academic reput ation; and (3) the influence of others, including parents, relatives teachers and counselors. The sample was drawn from the 97 unive rsities which administered the CIRP Freshman Survey in 2004. Data were used fo r students who were attending their first choice college located more th an 100 miles from home. Da ta were used from students who had received scores at or above 660 on th e SAT Verbal, and scores at or above 670 on the SAT Math. For students who did not re port scores for both SAT verbal and SAT math, the researcher accepted data from st udents reporting an ACT composite score of 30

PAGE 8

vi or higher. In addition, in order for their data to be used, stude nts were required to have an A or A+ average in high school. Results were reported as (1) frequenc ies and descriptive statistics, (2) a correlation matrix, and (3) multiple regressi on models. The study found the availability of financial aid to be the most important factor in predicting whether students will attend a higher-tiered or lower-tiered university. A lthough college costs a nd academic reputation were found to be significant predictors of th e tier level of university attended, they were of secondary importance compared with the attention to financia l aid by high achieving students.

PAGE 9

1 CHAPTER 1 INTRODUCTION Each year a multitude of high school students complete college admission applications with anticipation of what may be the most significant decision of their young lives. With over 3,500 colleges and universities in the United States, the decision of where to submit applications has become a da unting task for students and parents. It is not surprising that the phenomenon of choosi ng a college would attr act the attention of scholars. Researchers have examined the college choice process with a variety of approaches in an attempt to identify factors that influence the decisions of college-bound high school students. According to Kim ( 2004), “…every student has his or her own preferences about colleges ba sed on institutional type, pres tige, or even a student’s ‘intuitive feelings’ about how his or her personality fits in to a certain college” (p. 47). Consequently, the results of college choice studi es are of particular interest to college administrators who are tasked with shaping th e profile of their entering freshman classes. The college choice process has undergone significant change over the past fifty years. Before 1950, fewer than one out of fi ve high school graduates attended college, and this ratio was even smaller for wome n, students of color and students from lowincome families (Kinzie, Palmer, Hayek, Hossler, Jacob, & Cummings, 2004). The enactment of the Servicemen’s Readjustment Act (more commonly known as the GI Bill) in 1944, and the Supreme Court decision in Brown vs. Board of Education in 1954

PAGE 10

2 opened access to college to an ever-expanding number of students. By 1970 more than fifty percent of high school graduates were going off to college (Kinzi e, et al., 2004). Today there is general agreement that a four-year college degree is essential for future economic success. Several studies have supported anecdot al speculation that college graduates earn significantly more th an their peers with no postsecondary degree (Leslie & Brinkman, 1988; Pascarella & Tere nzini, 1991). Further, the specific college attended may have additional impact on a student ’s future financial status. According to a study conducted by Brewer, Eide & Ehrenberg (1999), there is a significant positive relationship between attendance at an elite private instituti on and future earnings. While economists may debate the extent to which a college education bene fits individuals and society, there is agreement that an edu cated citizenry contributes to economic competitiveness, productivity, government revenues and social equality (Kinzie, et al., 2004). Today’s high school students and their pa rents are generally aware of the longerterm economic benefits of a college educat ion; and students are more likely now than they were fifty years ago to view a college degree as within their grasp. With the influx of greater numbers and greater diversity of students has come increased competition among institutions of higher education for the most talented students. There is pressure on public institutions in particular to mainta in broad access policies; but these pressures often are in conflict with some colleges’ and universities’ desi res to recruit highachieving students to improve academic reputati on and rankings. Four-year institutions in particular have focused greater attention on marketing efforts to meet enrollment goals

PAGE 11

3 (Kinzie, et. al., 2004). Hoyt and Brown (2003) state that, “As a part of its marketing plan, an institution must determine who to contact in an effort to influence student college choice decisions” (p. 1). Access to college and un iversity information through mass media has had a noticeable impact on the manner in which a pplication and admissions processes are approached. By the 1990s students and familie s had much more information that they could realistically use to make educated deci sions regarding the institutions to which they should apply. In response to students becoming more savvy in their decision-making, colleges and universities have adjusted and improved their recruitment and enrollment procedures by incorporating strategies related to fina ncial aid and early admission (Kinzie, et. al., 2004). Within the last twenty-five years, th e competition among colleges and universities to attract students has intensified. Not only are institutions concerned about the number of students they can enroll, but they are partic ularly interested in high-achieving students due to the enhancements that these students can contribute to an in stitution’s reputation. The academic reputation of a university is a ke y factor in the recrui tment of the best and brightest students; but it is also the case th at the recruitment of the best and brightest students is critical for positive developmen t of an institution’s academic reputation. Moreover, with a multitude of colleges and universities vying for the best qualified students, it is a greater challenge for some in stitutions than others to attract the most desirable students to their in stitutions (Geiger, 2002).

PAGE 12

4 Many institutions of higher e ducation in the United Stat es are striving for greater levels of status and prestige. Among res earch-extensive universities, many long for membership in the Association of Amer ican Universities (AAU), which presently includes 62 universities consider ed by many to be the most prestigious in the United States and Canada (AAU web site retrieved August 4, 2008, from http://www.aau.edu/about/default.aspx?id=4020 ). In addition, in the quest for prestige for their universities, admini strators want to achieve an attractive rank in the annual edition of Best Colleges published by U.S. News and World Report ( hereafter USNWR) The Best Colleges report ranks institutions within broader categories of national universities, liberal arts college s, and master’s universities. USNWR defines national universities as those which “offer a full ra nge of undergraduate majors, master’s, and doctoral degrees… [and] are committed to producing groundbreaking research” (USNWR web site, accessed September 21, 2008, from http://colleges.usnews.rankingsa ndreviews.com/ college/national ). The 248 institutions classified as national universitie s are grouped into four tiers, w ith those listed in the first and second tiers receiving i ndividual numerical rankings. An attractive ranking in the USNWR college guide offers nationwide advertising and bragging rights that many institutions co uld never afford to fund from their own budgets. “Prestige is vitally important … becau se it relates so closely to institutional wealth” (Geiger, 2004, p. 83). The wealth to which Geiger refers is not from state appropriations, but from the ability to obt ain additional funds through increased tuition and private contributions. Resear ch has found that a rise in an institution’s ranking may

PAGE 13

5 lead to development success through improve d relationships with proud alumni (Monks & Ehrenberg, 1999; Geiger, 2004). Moreover, universities pay attention to their placement in the rankings because rankings a nd prestige are important to their target student markets who want to attend a prestigious institution (Brooks, 2006). Due to the known positive correlatio n between rankings and recruitment, universities striving for pres tige are likely to dedicate energy and resources into researching the USNWR indicators, which include peer assessment, freshman retention rate, six-year graduation ra te, faculty resources, alum ni giving rate, and student selectivity. With respect to student selectiv ity, the level of achieve ment on standardized tests for an institution’s freshman class is a widely accepted indicator of the quality of the student body. Therefore, the colleges that are most selective tend to garner the greatest levels of prestige. Some universities must wo rk much harder than others to improve on this indicator. Universities such as Harvar d, Princeton and Columbia have a long history of prestige and a solid reput ation for quality, therefore attracting the most qualified students. Well into the future, the names of su ch universities will like ly attract the best and brightest students from around the globe. On the other hand, research-extensive uni versities which find themselves in the third or fourth tier according to USNWR particularly public univ ersities, experience the greatest challenges in making headway with student select ivity indicators. For these lower-tiered institutions, st rategic enrollment planning a nd strong marketing campaigns are necessary to communicate the quality of programs and accomplishments of faculty to students and other key stakeholde rs. Historically, little recruitment effort was required to

PAGE 14

6 attract a sufficient number and quality cohort of students. Howe ver, with the proliferation and effortless access to media rankings gui des, the increased competition between institutions has resulted in constant atte ntion to the success of student enrollment strategies (Kinzie, et. al., 2004). College Choice Models A variety of models have been develope d to provide rationale for college choice behaviors. These models generally fit into one of three types, as identified by Hossler, Braxton, and Coopersmith (1989): econometric, sociological, or combined. Econometric models (Kotler & Fox, 1985; McDonough, 1997) view college attendance as an economic benefit, where students who choos e to attend college do so because the perceived benefits outweigh the benefits of any alternatives. McDonough (1997) proposed that “students maximize perceived co st-benefits in their college choices; have perfect information; and are engaged in a process of rational choice” (p. 3). An econometric model focuses on expected cost s, expected future earnings, student background characteristics, and co llege characteristics as factor s important to the study of college choice (Hossler & Stage, 1992). Some researchers have questioned the applicability of econometric models to studies of college choice, argui ng that students often lack th e ability to adequately and rationally process information affecting matr iculation due to socioeconomic constraints and limited information (Jackson, 1982). Altern atively, sociological th eories, or statusattainment theories as described by Paul sen (1990) and McDonough (1997), focus on the

PAGE 15

7 characteristics that influen ce both social and cultural capi tal, including socioeconomic status and academic ability. A sociological mode l considers the role of certain factors in the attainment of positions or occupations of prestige or status. Two of the most prominent models of college choice are based on an integration of econometric and sociological models. On e commonly referenced model within the related literature for college choice behavi or comes from Chapman and Jackson (1987), whose comprehensive model accounts for a wi de spectrum of vari ables investigated within prior research studies, includi ng “…student characteristics and background, student attitudes, student perc eptions of colleges, college ch aracteristics, money (parental income level, tuition, and financial aid), st udent self-report ed preference s, and actual college choices of students” ( p. 11). Viewing the college choice process as the formation of intermediate summary measures followed by the weight of intermediate constructs, Chapman and Jackson (1987) suggested that coll ege choice is a result of the combination of the following three behaviors: perception formation, preference formation, and choice. The model proposes that students’ perceptions about an institution are synthesized to form a comprehensive evaluation of the inst itution’s value (preference formation), which leads ultimately to observed college choices. According to Chapman and Jackson’ s (1987) model a student’s overall impression of an institution is formed at the perception formation stage. Chapman and Jackson’s (1987) study, which was comprised of surveys and follow-up interviews with over 1,000 high-ability students, supported the premise that early preferences for a particular institution are principally influe nced by perceptions of academic quality,

PAGE 16

8 followed by perceptions of the school’s social climate. Early perceptions of various colleges are formed by a combination of st udents’ individual backgrounds of with students’ previous exposure to the colle ge and the brand that institutions have intentionally or non-inte ntionally promoted. Similar to perception formation, the forma tion of student preferences is believed to be dependent on the intera ctions between the student and the institution, and the influence of the particular college. “Analy sis at the choice phase is based on revealed preference behavior” (Chapman and Jack son, 1987, p. 14). Preferences are largely determined by the combination of early perceptions of the student and special familiarity effects such as whether either parent attended the college. Although the model proposed by Chapman and Jackson (1987) is commonly referenced in college choice studies, the th ree-stage choice model developed by Hossler and Gallagher (1987) has been most widely us ed within the resear ch and was the basis for this study. Hossler, et al (1989) defined the college c hoice experience as a “complex, multi-stage process during which an individual develops aspirations to continue formal education beyond high school, followed later by a decision to attend a specific college, university or institution of advanced vo cational training” ( p. 234). Hossler and Gallagher’s (1987) model outlines three st ages of the college choice process: 1. Predisposition: students’ decisions/aspirations to enroll in postsecondary education. 2. Search: the process of cons idering types of institutions to which to apply. 3. Choice: the selection of an institution to attend.

PAGE 17

9 In this model of college choice, the three processes typically do not occur concurrently but rather simultaneously, ofte n overlapping one another. The first stage of predisposition is define d as the phase in which students decide whether or not to pursue formal education after high school. Several factors that have been found to predispose students toward coll ege include socioeconom ic status, students’ academic achievement, parents’ education le vels, ethnicity, gender, encouragement from high school counselors and teacher s, support from peers, and parental expectations and encouragement (Hossler & Stage, 1992). During the search stage, students access information on specific colleges to further ex amine the opportunities and benefits. It is within this phase that students are most likely to consider external and institutional information sources. Factors that may be c onsidered by students at this second phase include cost of attendance, av ailability and offers of fina ncial assistance, and academic reputation. The third stage of college choice is the application of the predisposition factors combined with the information gath ered during the search phase (Hossler & Gallagher, 1987). Hossler and Gallagher’s model was the basis for the current study. This study examined how significant differences among hi gh achieving students in each of the first two stages may impact the level of academic reputation, measured by the USNWR assigned tier, of the college of first choice. Predisposition-related factors to be included as independent variables were grouped with in the categories of student and family characteristics (gender, ethnicity, parents’ education levels, and family income) and the influence of others (parents, relatives, t eachers, and counselors). Search-related factors

PAGE 18

10 considered for this study were grouped as inst itutional characteristic s (costs, financial aid, and academic reputation). Statement of the Problem There are many postsecondary institutions in the United States that provide quality education. However, due to the diffi culty in quantifying the value of a degree from any individual institution, many college s rely on external re cognition, including media-based rankings, to validate assertions of quality. One of the most commonly cited benchmarks for quality is the composition of an institution’s entering freshman class. The charge of enrollment planning officers at ambitious lower-tiered research universities is to be acutely aware of the factors that are important to the high-achieving students they are trying to attract. Studies that have investig ated college choice factors for high-achieving students repeatedly cite academic reputation as one of the top indicators of choice (Chapman & Jackson, 1987; Goenner & Snaith, 2004; Ma nski & Wise, 1983). These results fail to provide an indication as to why some high-achieving student s choose to attend universities with a less presti gious reputation than the more highly prestigious options available to them. The literature on college choi ce is vast and investigates many factors, in addition to institutional reputation, that students consider when choosing to enroll at a particular university. Some of the other factors include educ ation level of the parents, cost and financial aid packages, availability of certain programs, location of the campus, and the influence of parents and others.

PAGE 19

11 There is some consensus among research ers that institu tional prestige and academic reputation are of primary importan ce to high ability students when choosing a college. According to Manski and Wise (1983 ), students tend to choose a college where the mean SAT score of their student class is within 100 points of their own scores. However, the literature in this area offe rs little guidance to enrollment management professionals at lower-tier universities. Many high-achieving stude nts are choosing to attend less prestigious universi ties. For those students, to what extent do individual characteristics, family circumstances or inst itutional attributes play a role in swaying them away from a more selective university? Purpose of the Study The purpose of the study was to examine whether differences exist between traditional-aged high achieving students who c hoose to attend higher-tiered universities and their peers who choose to attend lower-tiered universities. Specifically, the researcher applied a causal-comparative research desi gn using multiple regression to identify whether significant differences exist betw een high-achieving students who chose to attend a higher-tiered univers ity and those who chose to a ttend a lower-tiered university. The independent variables were selected based upon Hossler a nd Gallagher’s (1987) three-stage model and previous research fi ndings in the literature and grouped according to: (1) students’ individual and family characte ristics, such as ethni city, gender, parents’ education level, and family income; (2) inst itutional characteristics, such as financial

PAGE 20

12 considerations and academic reputation; a nd (3) the influence of others, including parents, relatives, teachers and counselors. The data for the study was gathered from the Cooperative Institutional Research Program (CIRP) Freshman Survey for 2004. CIRP has been conducting national longitudinal studies of American college st udents since 1966 and has surveyed over eight million students. The CIRP Freshman Survey, managed by the Higher Education Research Institute (HERI) at th e University of California at Los Angeles, is administered annually to over 400,000 entering freshmen at approximately 700 two-year and four-year colleges and universities (Higher Educati on Research Institut e webpage, March 2008). The survey gathers information about (a) established behaviors in high school, (b) academic preparedness, (c) admissions deci sions, (d) expectations for college, (e) interactions with peers and f aculty, (f) student values and goals, (g) student demographic characteristics, and (h) concerns about fina ncing college (HERI web site, retrieved June 2, 2008, from http://www.gseis.ucla.edu/her i/cirp.php). The 40-question survey is attached at the end of this research proposal as Appendix A. A review of the research that has examin ed the college choice of high achieving students in U.S. postsecondary institutions prov ided the basis for the research questions addressed in this study. The following re search questions guided the study: 1. To what extent do students’ indivi dual characteristics (e.g. gender and ethnicity) relate to college c hoice for high achieving students?

PAGE 21

13 2. To what extent do students’ family ch aracteristics (e.g. pa rents’ education level and family income) relate to college choice for high achieving students? 3. To what extent do financial considerati ons associated with college (e.g. cost and financial aid) relate to college choice for high achieving students? 4. To what extent does academic reputati on of the institution relate to college choice for high achieving students? 5. To what extent does the influence of significant others (e.g. parents, relatives, teachers, and counselors) relate to college choice for high achieving students? Significance of the Study An exploration of the factors related to the individual characteristics and institutional preferences of high ability students who choose to enroll in a non-selective university is not only an intere sting research question but also an issue of relevance to state policymakers and college administrato rs. The present study adds to the body of literature related to college choice by e xploring differences between high achieving students who attend higher-tiered universiti es and high achieving students who attend lower-tiered universities. The existing literature on the subject of college choice and high achieving students demonstrates that high achieving students differ from the general student population as far as the manner in which they approach the college choice process and

PAGE 22

14 the factors that are most important to them (Bradshaw, Espinosa & Hausman, 2001). There is also some agreement within th e literature on college choice that, although financial factors are considered important to high-achieving students, the criterion that typically grabs the top spot is college qua lity (Chapman & Jackson, 1987). The literature is limited in providing a broa d and comprehensive understand ing of the college choice decisions of high-ability stude nts who choose to attend lowe r-tiered institutions. The present study addressed these gaps within th e literature. The results of this study should be of particular interest to lower-tiered universities. It is apparent that high achieving students who choose to attend lowe r-tiered universities are eith er not giving preference to the factor of college quality or are viewi ng college quality differently than how it is commonly defined by the media. This study can be differentiated from pr evious research on student choice in several ways. First, this study explored a nd proposed, using regression techniques, a prediction model of high achieving student enro llment probability towards either highertiered or lower-tiered research universities Second, the study used data received from high achieving students enrolled at 97 national re search universities that participated in the 2004 Freshman Survey; whereas earlier li terature on the subj ect is by and large limited to high achieving students at only a ha ndful of institutions, or the studies do not focus on the unique characteristics and pr iorities of high-ac hieving students.

PAGE 23

15 Limitations Limitations refer to “limiting conditions or restrictive weaknesses” (Locke, Spiruduso & Silverman, 2007, p. 16). All resear ch studies have limitations, possibly related to the difficulty of controlling variable s within the research design or related to the limited types of data that can be gathered due to ethics or feasibility. This study, as well, has its limitations. First, the process i nvolved with college choice is inherently difficult to study due to the complex, longitudi nal, interactional and cumulative issues involved with selecting a college (Hossler, et al., 1989). This study did not allow for the exploration of longitudinal perceptions and cumulative influences on the process. Secondly, there are limitations concerni ng the reliability and validity of the Cooperative Institutional Research Program (C IRP) Freshman Survey. To address these issues, the Higher Education Research Inst itute (HERI) has addressed these questions through a document posted on its website enti tled “CIRP Freshman Survey: Reliability and Validity” (Higher Education Research Institute, retrieve d July 27, 2008, from http://www.gseis.ucla.edu/heri/PDF s/CIRP_Reliability_Validity.PDF ). However, the document lacks evidence for either reliability or validity except to mention that item values remain consistent over time for th e same respondent. In addition, HERI offers neither content validation evidence nor a ny indication of relationships with other measures. Thirdly, using secondary data precludes the possibility of exploring some factors that may differentiate the matriculation of high achieving students. For example, some research has supported the hypothesis that so me students are particularly drawn to

PAGE 24

16 institutions with a strong reput ation for athletics; and addition al research has explored the influence of peers, particular ly romantic relationships, and a significant factor in college choice. The 2004 Freshman Survey does not ga ther such information from students, and therefore the factors of athlet ic reputation and peer influen ce were not explored in the current study. Finally, the decision to use data gath ered through the 2004 Freshman Survey limited the number of institutions from whic h student data were obtained, because the survey is offered and administered by colle ges and universities on a voluntary basis. Although 97 (39%) of 249 national research universities partic ipated in the survey, the lack of full inclusion limited data available to address the research questions explored for this study. Delimitations Delimitations refer to the external valid ity and generalizabilit y of the study based upon the research design. The author acknowledged two deli mitations pertaining to the current study. First, this study considered only students who ar e attending national research universities as de fined by the 2003 edition of Best Colleges by U.S. News and World Report The results of the study cannot be ge neralized to student s attending liberal arts colleges or regional master’s universities. Second, the data were taken from a convenience sample of students attending uni versities which participated in the 2004 CIRP Freshman Survey.

PAGE 25

17 Assumptions As with all research studies, this researcher makes some assumptions in the design and interpretation of results for this study. First, the study is grounded in Hossler and Gallagher’s (1987) model, which outlines th ree stages of the college choice process: 1. Predisposition: students’ decisions/aspirations to enroll in postsecondary education. 2. Search: the process of cons idering types of institutions to which to apply. 3. Choice: the selection of an institution to attend. In this model of college choice, the three processes typically do not occur concurrently but rather simultaneously, often overlapping one another. However, the variables that were used in the study were assigned to either the predisposition stag e or search stage and therefore assumed to fit nea tly within only one stage. A second assumption is related to the use of secondary data as the source for this study. Researchers must be cautious when d eciding to use secondary data sources, and note the disadvantages of such a choice. For example, the researcher may make incorrect assumptions about the intended definition of certain terms used in the instrument. “In some cases, secondary analysts are able to ch ange their concepts’ definitions to match the original ones and still be faithful to their theo retical framework. In othe r cases, this is not possible—the concepts are too different and fo rcing the fit is not appropriate” (Moriarty, et. al, 1999, p. 148). The current study assumed that the respondents’ understanding of the survey questions reflects th e researcher’s understanding.

PAGE 26

18 Definitions To facilitate understanding of the author’s intended meaning of certain terms, the following definitions are provided. Traditional-age student refers to a student who enroll ed at the university the year following graduation from high school. High achieving student is defined as a student who (1 ) received scores at or above 660 on the critical reading portion of the SAT and scor es at or above 670 on the mathematics portion of the SAT, or scores 30 or above on the ACT, and (2) had at least an “A” average in high school. National research university refers to any college or university listed as a National University in the 2003 edition of Best Colleges by U.S. News and World Report Higher-tiered university refers to a university whic h was ranked in Tier One or Tier Two in the 2003 edition of Best Colleges by U.S. News and World Report excluding any university which had been placed in Tiers Three or Four within the past five years. Lower-tiered university refers to a university which was ranked in Tier Three or Tier Four in the 2003 edition of Best Colleges by U.S. News and World Report excluding any university which had been placed in Tier s One or Two within the past five years.

PAGE 27

19 CHAPTER 2 REVIEW OF THE LITERATURE For generations students generally have believed that college attendance has a positive impact on their future success, a notion promoted by higher education institutions. In addition, st udents also have recognized that a degree from some institutions is more valuable than a degree from others. Clearly, students and parents in the twenty-first century continue to put si gnificant effort into selecting the “right” college, and institutions like wise dedicate significant reso urces toward recruiting the “right” students. Since the early twentieth cen tury, several research studies have been conducted in an effort to understand the va rious factors which are most important to students and their families when making th e choice of which college to attend. To appreciate and more fully understand th e complexity of the college choice process, various topics must be examined. Firs t, the literature revi ew will explore issues from the perspective of the institutions, namely the strategies and resources that are dedicated to improving institutional academic reputation. Second, the review will discuss college choice models and human capital development theory as the conceptual frameworks referenced in related college choi ce literature. Finally, there will be a review of the college choice literature, including a re view of college choice models that examine the relationship between college choice a nd student characteristics, institutional characteristics and external influences.

PAGE 28

20 Academic Reputation American college bound students have a c hoice of thousands of colleges, and prestige undeniably affects that choice, espe cially for those students who have excelled academically and for those students who come from families with abundant financial means. Fried (2005) observed that, “As the marketplace for stude nts has expanded from regional to national to internat ional, and as the number of institutions offering degrees has increased, the importance of reputation ha s grown significantly” (p. 21). Similarly, Sevier (1994) posited five observations pe rtinent to the image of higher education institutions, namely that (1) people are mo re influenced by prior knowledge than new knowledge; (2) image has a tremendous and often underappreciated effect on college choice; (3) institutions with strong images ar e able to recruit better faculty, and faculty are more likely to stay longer; (4) institutions with strong images tend to have a greater percentage of annual fund participation; and (5) image-building is seen as a legitimate pre-recruiting function at a ha ndful, but growing number of market-oriented institutions (pp. 60-61). In a financial sense, a univ ersity’s prestige ha s very real impacts. According to Geiger (2002), there are two primary fact ors which have an impact on reputation for research universities, namely selectivity of the freshman class and the scholarly productivity of its faculty. Selectivity refers to the percentage of applying students who are admitted to the institution. Generally, privat e institutions are more selective than their public counterparts. This difference can be attr ibuted to the size of many institutions and the expectation for the public universities to serve a broad popul ation of students.

PAGE 29

21 Admissions directors at na tional universities have observed the impact of rankings, whereas an institution’s decline in the ranking is followed by a decline in applications submitted by high achieving stude nts (Espeland, 2007). The recruitment of high achieving students among research extensiv e universities is highly competitive due to the national and internati onal recognition that these stude nts bring. In addition, a high achieving student body will help the instit ution attract world class faculty and researchers, further strengthening the univers ity’s image. Thus, just as top students are attracted to schools with outst anding faculty, so are top faculty attracted to schools with outstanding students, creating a winwin for the institution (Brooks, 2006). For prestige-oriented universities a high or improved ranking in USNWR is noteworthy and likely to be mentioned frequent ly in institutional literature which will gain the attention of targeted students, faculty and donors. Pr ivate and public institutions alike are driven toward reports published by USNWR which ranks American universities using a self-developed formula. The first edition of Best Colleges was published by USNWR in 1983. The ranking included 76 institutions and was based solely on a reputation survey completed by nearly 1,300 pr esidents of four-year colleges (Machung, 1998). The original report gained instant popularity among prospect ive undergraduates and their parents because it was the first time that such information was available in a comprehensive format (Webster, 1992). Although USNWR has made some adjustments since the initial Best Colleges edition in 1983, the methodology generally cons ists of quantitative information at the undergraduate level such as freshmen reten tion and graduation ra tes, test scores

PAGE 30

22 (SAT/ACT) for first-time students, the per centage of classes with fewer than 20 or greater than 50 students, per centage of full-time faculty, f aculty-to-student ratios and alumni-giving rate. These factors are combined with the reputational scores derived from the survey of university presidents and provosts The indicators are th en standardized and weighted to produce the overall score that is used for rank-ordering (Clarke, 2002). A complete description of categ ories and indicators used by USNWR in the most recent edition of Best Colleges is provided in Table 2.1. Table 2.1 U.S. News and World Report Indicators and Weights for the 2008 College Rankingsa Ranking Category Category Weight Indicator Indicator Weight Academic Reputation 25% Academic Reputation Survey 100% Student Selectivity 15% Acceptance Rate High School Standing Top 10% SAT/ACT Scores 10% 40% 50% Faculty Resources 20% Faculty Compensation Faculty With Top Terminal Degree Percent Full-time Faculty Student/Faculty Ratio Class Size, 1-19 Students Class Size, 50+ Students 35% 15% 5% 5% 30% 10% Graduation and Retention Rate 20% Average Graduation Rate Average Freshmen Retention Rate 80% 20% Financial Resources 10% Educational Expenditures Per Student 100% Alumni Giving 5% Alumni Giving Rate 100% Graduation Rate Performance 5% Graduation Rate Performance 100% aThese indicators and weights are for the national liberal arts and national university rankings only.

PAGE 31

23 For the schools historically ranked at the top (i.e. Princeton, Ha rvard, Yale), it is unlikely that the annually reported reputational scores for these institutions will deviate in the foreseeable future. According to Sheehan (1996), reputation is hi ghly correlated with resources, and the two characteristics “tend to feed on each other” (p. 18). In fact, a review of the change in reputation score fo r national universities over a ten-year period (1998-2007) demonstrates the static nature of th e survey results. This researcher’s review of the data revealed that no university in th e last ten years has ma de significant headway with its reputation score. USNWR began reporting the five-point peer assessment score (PAS) with its 1998 Best Colleges issue. A review of the historical tr ends reveals that in the years between 1998 and 2007 (inclusive), only 30 of the 248 national universities experienced an absolute change of 0.2 points or more, and only one college had a change as high as 0.5 points (see Table 2.2). Those postsecondary institutions with aspirations of improving their USNWR rankings develop strategies based upon the we ight given to various indicators; however, few indicators are under the control of institu tions. As demonstrated in the previous paragraph and in Table 2.1, the indicator with the greatest weight (academic reputation score) has very little to do with efforts made by individual institutions. Because student selectivity is one of the few indicators considered among the ranking criteria over which institutions have some amount of control, th e universities that have made prestige a priority have made strategic changes to th eir admissions criteria. Public universities, which historically have a reputation for access and open admission, are now turning away a larger and larger proportion of their app licants in the name of increased quality.

PAGE 32

24 Table 2.2 Largest Positive Changes in Peer Assessment Scores from 1998-2007 School Name Public/ Private 2007 PAS 1998 PAS MAXa 19982007 MINb 19982007 1998-2007 Change University of Alabama Public 3.1 2.6 3.1 2.6 0.5 University of Arkansas Public 2.9 2.5 2.9 2.5 0.4 Northeastern University Private 3.1 2.8 3.1 2.8 0.3 Andrews University Private 2.1 1.8 2.1 1.8 0.3 Nova Southeastern University Private 2.0 1.7 2.0 1.7 0.3 University of San Francisco Private 3.0 2.7 3.0 2.7 0.3 University of Miami Private 3.2 3.0 3.2 3.0 0.2 Montana State University--Bozeman Public 2.6 2.4 2.6 2.4 0.2 Pepperdine University Private 3.1 2.9 3.1 2.9 0.2 George Washington University Private 3.5 3.3 3.5 3.3 0.2 Idaho State University Public 2.5 2.3 2.5 2.3 0.2 Middle Tennessee State Univ. Public 2.1 1.9 2.1 1.9 0.2 University of Central Florida Public 2.5 2.3 2.5 2.3 0.2 University of La Verne Private 2.1 1.9 2.1 1.9 0.2 University of Alaska--Fairbanks Public 2.6 2.4 2.6 2.4 0.2 University of Alabama--Huntsville Public 2.6 2.4 2.6 2.4 0.2 University of South Dakota Public 2.5 2.3 2.5 2.3 0.2 Biola University Private 2.0 1.8 2.0 1.8 0.2 University of Southern California Private 3.9 3.7 3.9 3.7 0.2 Howard University Private 2.9 2.7 2.9 2.7 0.2 New York University Private 3.8 3.6 3.8 3.6 0.2 San Diego State University Public 2.8 2.6 2.8 2.6 0.2 University of San Diego Private 2.8 2.6 2.8 2.6 0.2 University of Colorado--Denver Public 2.9 2.7 2.9 2.7 0.2 University of Montana Public 2.8 2.6 2.8 2.6 0.2 aReflects the highest score received between 1998 and 2007 (inclusive) bReflects the lowest score received between 1998 and 2007 (inclusive) All universities need sufficient enrollment to operate; however, institutions would like to be in the position of having a sufficient pool of applicants so that they can select the students who will shape the ideal class profile and demonstrate a lower acceptance rate. An improved ranking tends to lead to an increase of appli cations from qualified students (Monks & Ehrenberg, 1999). A fall in th e rankings can put even greater pressure on the admissions office (Mufson, 1999). Desp ite the limitations in the ranking

PAGE 33

25 methodology, movement in the rankings appears to have some correlational relationships with future selectivity, with evidence that application rates will decrease following a drop in the rankings (Hoxby, 1997). Research-extensive universities placed in the third and fourth tiers have to work particularly hard to attract top quality stud ents away from their top-tier competit ors (Geiger, 2004). College Choice Models Effective enrollment management begins with an understanding of the college choice process, including the timing of vari ous stages and knowledge regarding factors which are considered most important to the recruitment pool (DesJardins, et al., 1999). Institutions that understand the effects that various factors have on the tendency of students to prefer one type of institution over another are armed with information that may be helpful in the development of effective marketing strategies. A variety of models have been develope d to provide rationale for college choice behaviors. These models generally fit into one of three types, as identified by Hossler, Braxton, and Coopersmith (1989): econometric, sociological, or combined. Econometric models (Kotler & Fox, 1985; McDonough, 1997) view college attendance as an economic benefit, where students who choos e to attend college do so because the perceived benefits outweigh the benefits of any alternatives. McDonough (1997) proposed that “students maximize perceived co st-benefits in their college choices; have perfect information; and are engaged in a process of rational choice” (p. 3). An econometric model focuses on expected cost s, expected future earnings, student

PAGE 34

26 background characteristics, and co llege characteristics as factor s important to the study of college choice (Hossler & Stage, 1992). Some researchers have questioned the applicability of econometric models to studies of college choice, argui ng that students often lack th e ability to adequately and rationally process information affecting matr iculation due to socioeconomic constraints and limited information (Jackson, 1982). Altern atively, sociological th eories, or statusattainment theories as described by Paul sen (1990) and McDonough (1997), focus on the characteristics that influen ce both social and cultural capi tal, including socioeconomic status and academic ability. A sociological mode l considers the role of certain factors in the attainment of positions or occupations of prestige or status. Two of the most prominent models of college choice are based on an integration of econometric and sociological models. On e commonly referenced model within the related literature for college choice behavi or comes from Chapman and Jackson (1987), whose comprehensive model accounts for a wi de spectrum of vari ables investigated within prior research studies, includi ng “…student characteristics and background, student attitudes, student perc eptions of colleges, college ch aracteristics, money (parental income level, tuition, and financial aid), st udent self-report ed preference s, and actual college choices of students” ( p. 11). Viewing the college choice process as the formation of intermediate summary measures followed by the weight of intermediate constructs, Chapman and Jackson (1987) suggested that coll ege choice is a result of the combination of the following three behaviors: perception formation, preference formation, and choice. The model proposes that students’ perceptions about an institution are synthesized to

PAGE 35

27 form a comprehensive evaluation of the inst itution’s value (preference formation), which leads ultimately to observed college choices. According to Chapman and Jackson’ s (1987) model a student’s overall impression of an institution is formed at the perception formation stage. Chapman and Jackson’s (1987) study, which was comprised of surveys and follow-up interviews with over 1,000 high-ability students, supported the premise that early preferences for a particular institution are principally influe nced by perceptions of academic quality, followed by perceptions of the school’s social climate. Early perceptions of various colleges are formed by a combination of stude nts’ individual bac kgrounds with students’ previous exposure to the colleg e and the brand that institutio ns have intentionally or nonintentionally promoted. Similar to perception formation, the forma tion of student preferences is believed to be dependent on the intera ctions between the student and the institution, and the influence of the particular college. “Analy sis at the choice phase is based on revealed preference behavior” (Chapman and Jack son, 1987, p. 14). Preferences are largely determined by the combination of early perceptions of the student and special familiarity effects such as whether either parent attended the college. Although the model proposed by Chapman and Jackson (1987) is commonly referenced in college choice studies, the th ree-stage choice model developed by Hossler and Gallagher (1987) has been most widely used within the research and will be the basis for this study. Hossler, et al (1989) defined the college c hoice experience as a “complex, multi-stage process during which an individual develops aspirations to continue formal

PAGE 36

28 education beyond high school, followed later by a decision to attend a specific college, university or institution of advanced vo cational training” ( p. 234). Hossler and Gallagher’s (1987) model outlines three st ages of the college choice process: 1. Predisposition: students’ decisions/aspirations to enroll in postsecondary education. 2. Search: the process of cons idering types of institutions to which to apply. 3. Choice: the selection of an institution to attend. In this model of college choice, the three processes typically do not occur concurrently but rather simultaneously, ofte n overlapping one another. The first stage of predisposition is define d as the phase in which students decide whether or not to pursue formal education af ter high school. According to the Hossler, et al.’s (1989) model of college choice, the pred isposition stage is a “developmental phase in which students determine whether or not they would like to continue their education beyond high school” (p. 209). The predisposition stage coincides with the transition from middle school to high school during which time students tend to be open to the positive influences of significant others at home a nd school. When these adolescents receive positive messages and encouragement from parents and other significant individuals in the area of academic development, there is a positive effect on future college success. According to this model, the predispositi on stage is a “longitudinal development phase involving the initial formation and subseque nt reassessment of college aspirations” (Brasier, 2008, p. 22). Several factors that ha ve been found to pred ispose students toward college include socioeconomic status, st udents’ academic achievement, parents’

PAGE 37

29 education levels, ethnicity, gender, encour agement from high school counselors and teachers, support from peers, and parental expectations and encouragement (Hossler & Stage, 1992). During the search stage, students engage in accessing information on specific colleges to further examine the opportunities a nd benefits. It is within this phase that students are most likely to consider external and institutional information sources. Factors that may be considered by students at this second phase include cost of attendance, availability and offers of fi nancial assistance, and academic reputation. The third stage of college choice is the application of the predisposition factors combined with the information gathered during the search phase (Hossler & Gallagher, 1987). Regardless of the efficiency with which students move thr ough the three-step process, it is during the third stage that students choose one institution over another (Kim, 2004). Hossler and Gallagher’s model will be th e basis for the current study. This study will examine how significant differences am ong high achieving students in each of the first two stages may impact the importance of the academic reputation, measured by the USNWR -assigned tier, of the college of first choice. Predisposition-related factors to be included as independent variab les will be grouped within th e categories of student and family characteristics (gender, ethnicity, pa rents’ education levels, and family income) and the influence of others (parents, relative s, teachers, and counselors). Search-related factors considered for this study will be grouped as institu tional characteri stics (costs, financial aid, and academic reputation).

PAGE 38

30 College Choices of High Achieving Students Given the breadth of the literature on college choice, the remainder of this chapter will focus on a review of prior studies that ha ve applied college choice frameworks to the educational choices of high achieving students. A review of the lite rature suggests that the characteristics of students (e.g. gende r, ethnicity, and family income), the characteristics of institutions (e.g. cost location, reputation, and programs), and the influence of others (e.g. pare nts, teachers, and counselor s) together influence the matriculation decisions of students (DesJardin s, et al., 1999). Speci fically, the research has supported the hypothesis that students of high socioeconomic status with high educational aspirations, high academic abilit y, and highly educated parents are more likely to choose institutions that cost more, are furthe r from home, and are highly selective (Hossler, et al., 1989; Paulsen, 1990). Moreover, high-achieving students are more likely to attend selective universities and out-of-state universities than students with low or average achievement levels (Braxton, 1990). Although the existing research related to college choi ce and matriculation is considerable, and despite the importance of such information, little research has been done to consider differences in the factors th at are considered most important to students choosing to attend a highly selective (or highe r-tiered) university and those whose first choice school is a less selective institution. Th e literature su pports the notion that highachieving high school students consider academ ic reputation to be among the most important when deciding where to go to coll ege. However, when one reviews the trends in admissions of high-achieving students to Tier three and Tier four in stitutions, it is clear

PAGE 39

31 that these lower-tiered universities are c onsistently attracting greater numbers of academically attractive students. Of the 54 lower-tiered universities for which SAT scores were reported in the 2002 and 2007 editions of Best Colleges 41 experienced an increase in the 25th percentile for incoming students, and the freshman class for 24 of those institutions increased SAT scores by 30 points or more, as reflected in Table 2.3. For those universities that are setting strate gic goals to improve their position in the Table 2.3 Largest Positive Changes in 25th Percentile SAT Scores for Lower-Tier Universities from 2002-2007 School Name 2007 25th Percentile 2002 25th Percentile 2007 25th 2002 25th Temple University 1000 920 80 Texas Tech University 1040 970 70 University of Texas--Dallas 1120 1060 60 Rutgers--Newark 1020 960 60 Seton Hall University 1010 950 60 University of La Verne 930 870 60 University of South Florida 1030 970 60 Hofstra University 1060 1010 50 St. John's University 940 890 50 Georgia State University 990 940 50 Old Dominion University 960 910 50 Virginia Commonwealth University 960 920 40 Indiana U.-Purdue U.--Indianapolis 880 840 40 San Diego State University 980 940 40 University of Rhode Island 1020 980 40 Univ of Massachusetts--Boston 960 920 40 Northern Arizona University 960 920 40 Portland State University 930 890 40 University of North Texas 1000 960 40 Univ of Maryland--Baltimore County 1110 1080 30 Adelphi University 1000 970 30 University of Houston 950 920 30 Texas Woman's University 850 820 30 George Mason University 1000 970 30

PAGE 40

32 USNWR rankings, a study with a specific focu s on the factors that influence the academically-talented students to less presti gious institutions would be considerably helpful. According to economists, students who pay more to attend a se lective college are making sound economic decisions, as every 100 -point increase in a college’s average SAT is associated with 3 to 7 percent highe r earnings for its graduates (Kane, 1998). Dale and Krueger (1999) added that the payoff is greatest for students from disadvantaged family backgrounds, although those students are le ss likely to make th e initial investment. Hoxby (1997) noted that students who invest in prestige ea rn their investments back several times over, but some researchers ques tion the cause-and-effect of these results, arguing that the higher earnings may be correlated more with the traits and drive of highachieving students and less to do with their college alma mater. Moreover, Avery and Hoxby (2003) found that high-ability students were likely to be more analytic and long-sighted regardi ng their college investme nt; and advised that students are better off to refuse a full ride at a lower-ranked co llege and spend their money on the higher-ranked school. There are ot her non-tangible benef its of attending a higher-tiered school which are not calculated as easily, such as de veloping professional and social networks (Behrman, Kletzer, McPherson & Schapiro, 1998). Although highachieving students have been found to behave as rational human capit al investors, Avery and Hoxby (2003) identified three circumst ances that impact rational investment behavior, namely:

PAGE 41

33 1. Credit constraints – The family income level is too high to qualify for needbased aid, and the family is unwilling to pay for a highly-selective college; 2. Misinformation – The student is nave about the various levels of financial resources and subsequently chooses a college at which he accumulates less human capital than what could have been possible; 3. Lack of concern – The student simply is not concerned about maximizing his lifetime utility when choosing a college. Existing literature in the area of college choice behavior can be categorized in a number of ways. The broadest category sepa rates the literature into explorations of whether students attend college (access) and where students attend college (choice) (Hu & Hossler, 2000). Hossler, Braxton and Coopers mith (1989) further narrowed the field of college choice studies by students’ decisions to (1) attend any type of higher education, (2) attend a vocational school, two-year institut ion or four-year institution, and (3) attend a specific institution over other reasonable options. In addition, researchers have investigated students’ choi ces between (4) private vers us public institutions (Hu & Hossler, 2000), (5) expensive versus less expe nsive institutions (Orf ield, 1992), (6) firstchoice versus lower-choice institutions (Chapman & Jackson, 1987), and (7) highly selective versus less selec tive institutions (Hearn, 1991). Relative to the factors that tend to influe nce the types of decisions outlined in the previous paragraph, additional studies reveal further categorization of the college choice literature. Studies indicate that matriculation decisions are related to (1) students’ individual and family charac teristics (Brewer, et al., 1999; Hearn, 1987; Manski & Wise,

PAGE 42

34 1983; Paulsen, 1990); (2) institutional characteris tics, such as financ ial considerations and academic reputation (Fuller, Manski, & Wise, 1982; Hossler & Gallagher, 1987; Weiler, 1994); and (3) the influences of signi ficant others, including parents, relatives, teachers and counselors (Bradshaw, et al., 2001; Hossler, Schmit, & Vesper, 1999; Lillard & Gerner, 1999). Student and Family Characteristics When considering the relationship betw een the individual ch aracteristics of students and college choice, the history of the literature demonstrates that race, gender and social class have the stronge st relationship with educationa l attainment (Kinzie, et al., 2004). According to McDonough (1997), “Afr ican-Americans, women, and low-SES students are especially likely to attend less se lective institutions even if their ability and achievements are high” (p. 5). Not surprisingl y, Brewer, et al. (1999) found that students from high socioeconomic backgrounds and stude nts who are academically talented are more likely to attend elite institutions. Gender and Ethnicity Prior research indicates that African American and Hispanic students are more sensitive than their white peer s to the costs of higher educ ation, are more responsive to grants and scholarships (Johnson, Stewart & Eberly, 1991; Hoyt & Brown, 2003), and “African Americans are more sensitive than other students to changes in tuition and financial aid, even after controlling for soci oeconomic status and academic ability” (Kim,

PAGE 43

35 2004, p.45). In addition, African American stude nts are less likely to attend selective institutions than are white students (Hearn, 1987) and they are significantly less likely to attend their first choice institution (Kim 2004). Gender differences also exist among African American students, where the quality of social life and participation in athletics tends to be more important to males than to females (Briggs, 2006; Hubbard, 1999), and the economic benefits of a college education is more important to females than males (Hubbard, 1999). In addition, whereas male students have b een found in the past to have higher college aspirations than females, recent studies (Chenowith & Ga llagher, 2004; Reynolds & Pemberton, 2001) have reported evidence to the contrary. With the exception of Hispanic females, the literatu re indicates that females have stronger academic goals than males; although Asian American males have been found to possess significantly higher college aspirations than females and all other ethnic male groups (Mau, 1995). The gender-ethnic group that appears to have the lowest college aspirations is the Native American male group. For both male and female students, Hispanics and Native Americans have demonstrated lower educat ional aspirations than white and African American students (Mau, 1995). Socioeconomic Status Socioeconomic status is a nother common category that re searchers use to segment students in college choice st udies. A number of research studies demonstrated the disparity between low and middle income stude nts and high income students, with high

PAGE 44

36 income students being more likely to attend in stitutions which are more costly and more selective (Brewer, et al., 1999; Hearn, 1987; Manski & Wise, 1983; Paulsen, 1990). Prior research indicates that low-income and first-generation students are comparatively disadvantaged against their more affluent pe ers when it comes to the variety of colleges from which they are able to choose (Kinzie, et al., 2004). The basis of this argument comes from a number of sources, one bei ng the increasing institutional and federal reliance on granting loans rather than grants. In a study grounded in the status attainment perspective, Hossler and Stage (1992) hypothesized that family socioeconomic status had a direct relati onship with parental encouragement and students’ academic achie vement. According to Sewell, Haller, and Portes (1969), who introduced the Wisconsin status attainment model, a basic question raised by status attainment research is “By wh at mechanisms are social origins translated into attainment outcomes?” (p. 83). Subseque ntly, Hossler and Stage (1992) argued that parental encouragement and expectations, along with high school experiences, directly influence college aspirations in student s, regardless of ge nder, ethnicity and socioeconomic backgrounds. Further, they sugg ested that socioeconomic status has an indirect impact on a student’ s predisposition to attend college, as there is a positive relationship between socioeconomic status, students’ academic success, and students’ perceptions of the educational expect ations that others have for them.

PAGE 45

37 Education Level of Parents Being raised by parents who lack awareness of the college experience may put students at a disadvantage when it comes to making decisions about where to go to college and how to be successful once enroll ed. Further, first-generation students have been found to receive less encouragement and support from their families than multigeneration students when it comes to college attendance (Arredondo, 1999). These firstgeneration students may grow up assuming that co llege is not a good fit for them or is not a realistic dream. However, students appear to have a higher likelihood of viewing college as realistic when their parents stress the importance of educational success (Ceja, 2004). Research findings differ when reporting on behaviors of first-generation students in the college application process. McDonough (1994) report ed that, compared with students who are raised by colle ge graduates, first-generatio n students are more likely to limit the number of institutions to whic h they apply and to apply to nonselective institutions. However, a study of colleg e-bound high school students in New Hampshire revealed no significant differences in the type or quality of college under consideration between students whose parents possesse d postsecondary degrees and those whose parents had not completed a college edu cation (Toutkoushian, 2001). In fact, first generation students were found to be equally likely as those w ith college-educated parents to consider atte nding a selective school.

PAGE 46

38 Institutional Characteristics The relationship between students’ preferen ces and institutional characteristics is a significant determinant of where students ul timately decide to attend college (Weiler, 1994). Hoyt and Brown (2003) reviewed twenty-t wo studies related to college choice in order to identify institutional factors that we re most frequently cited as important to students. The views of over 30,000 students in 18 states were represented in the comprehensive review. Among the 22 studies exam ined, nine factors were identified that took first place as far as level of importance to students. Those nine factors, in order of frequency, were (1) academic reputation, (2 ) location, (3) quality of instruction, (4) availability of programs, (5 ) quality of faculty, (6) cost s, (7) reputable program, (8) financial aid, and (9) job outcomes. Other variables which were included in the studies, but did not make the number one spot include (1 0) variety of courses offered, (11) size of the institution, (12) surrounding community, (13) availability of gra duate programs, (14) student employment opportunities, (15) quality of social lif e, (16) class size, (17) graduate school outcomes, (18) extracurricu lar programs, (19) friendly/personal service, (20) affiliation, (21) admission requirements, a nd (22) attractiveness of campus facilities (p. 5). It is important to note that the factors identified by Hoyt and Brown (2003) above were a result of the review of perceptions of students from a variety of segments, including high school students with a full range of academic abilities community college students, non-traditional-age university transf er students, and even non-attendees of any college. Therefore, one may infer that the relative importance of the factors in the

PAGE 47

39 preceding paragraph will likely vary by specific market segment. The factors that appear to be important to high-achieving students include academic reputation, quality of the student body, and scholarship awards (Bra dshaw, et al., 2001; Hoyt & Brown, 2003; Litten, 1982). Location/Proximity to Home The current generation of college-bound hi gh school students is much more likely to attend college out-of-state than were previous generati ons. Whereas 93% of undergraduates attended college in their home states in 1949, that percentage dropped to 75% by 1995 (Hoxby, 1997). Students are more likely to attend college outside of their local market area when they are male, when they belong to a higher socioeconomic status, when their parents have higher e ducation levels, and when they have high academic abilities and educational aspira tions (Hoyt & Brown, 2003; Paulsen, 1990). Cost and Availability of Financial Aid There appears to be an ever-widening gap between the costs of higher education and the family and external resources availa ble. For understandable reasons, the financial realities of a college educati on are likely to influence a student’s choice of where to attend college; and the subject has drawn a great deal of attention from researchers (Braunstein, McGrath & Pescatrice, 1999; DesJardins, Ahlburg & McCall, 2006; Ehrenberg & Sherman, 1984; Hossler & Gallag her, 1987; Hossler, et al., 1999; Kim, 2004; McPherson & Schapiro, 1991; Parker & Summers, 1993). Much of the existing

PAGE 48

40 research supports the notion that, regardi ng students’ interests in developing human capital, students consider the trade-offs betw een current costs and future expectations of financial and non-financial benefits (Hill, 2008). As a strategy to recruit greater numbers of high-achieving students, institutions may increase levels of educational spending pe r student. This is the case particularly at private institutions that can more easily rais e tuition to address financial needs (Hoxby, 1997). For some Ivy League institutions, fo r example, annual educational spending exceeds $45,000 per student (Geiger, 2002). At elite private institutions, students carry much of the financial burden. Understanding th at tuition increases may result in deterring the students they are trying to attract, ma ny institutions accompany tuition increases with increased allocations for both need-based and merit-based financial aid. “The [institutions’] objectives are diverse – from a pur ely altruistic desire to relax constraints facing the needy to a college’s self-interested desire to enroll high aptitude students who raise its profile or improve education fo r other students on campus” (Avery & Hoxby, 2003, p. 3). For high-ability students, assessing the best combination of multiple offers of financial assistance can be a daunting task, as they may qua lify for both need-based and merit-based aid, both state-f unded and privately-funded scholarships, federal work-study programs, and aid packages from each of th e colleges in which they are interested. However, if a significant proporti on of financial aid is in the form of loans, some of the most desirable institutions realistically may be out of reach for many high-achieving students.

PAGE 49

41 There is some agreement w ithin the literature that, wh ile the availability of financial aid is considered important by most college-bound students, the impact of cost and financial aid decrease as students’ income level and academic ability increase (Kim, 2004; Manski & Wise, 1983; Paulsen & St. John, 2002). This financial gap often discourages or prohibits lowincome students from attending higher-tiered institutions, even when controlling for academic ability (Hearn, 1991). As financial considerations would appear to be an obvious factor likely to influence enrollment decisions, the author considers this in the design of the study by including in the samp le only those students who are attending the university that was their first choice am ong all of the universities to which they applied. Reputation and Prestige Since the 1950’s, when institutions began to geographically broaden their recruitment to a national market, high-ach ieving students have been drawn to elite institutions (Geiger, 2002). The general conc lusion of the existing literature exploring college choice for high-achieving students is that academic reputation is consistently the primary factor in the college choice deci sion. Manski and Wise (1983) concluded from their study that students tend to select a co llege where the average SAT score is within 100 points of their own scores. In their st udy of students’ decision to attend the University of North Dakota, which is listed in the third tier by U.S. News & World Report in the 2008 America’s Best Colleges edition, Goenner and Snaith (2004) found that academic reputation was the most important factor. Social life on the campus came in

PAGE 50

42 second in order of importance to students, except for out-of-stat e students. Similarly, results of a study of college freshmen at a large Midwestern university by Johnson, et al. (1991) indicated that academic reputation and quality of programs were the most important factors affecti ng the decision to attend. Although some research on the importance of media rankings has been conducted, little is still known about th e population of students which most heavily value such indices. Goenner and Snaith (2004), fo r one, found that for students attending the University of North Dakota, national media rankings did not play a major role in the college choice process for those students na tive to the region, but did seem to be important to students who came from out-ofstate. Hossler and Foley (1995) hypothesize that rankings do not have an impact on the general college-bound student population, and that this non-interest is especially the case with non-traditional-age, low-income and high income students, suggesting that rankings ma y have some influence with middle-income students and those attend ing regional campuses. Influence of Others The choice of where to go to college is ar guably one of the bi ggest decisions of a young adult’s life. For high school students c onsidering a college career, guidance from trusted loved ones and respected role models is needed to think through all of the considerations. Among those having some influe nce with students’ choice of college are parents, other relatives, high school counselors and teachers.

PAGE 51

43 Several scholars (Levine & Nidiffer, 1996; Cabrera & LaNasa, 2000; Tierney & Venegas, 2006) have found parental influence to be a significant predictor of student matriculation. In Levine and Nidiffer’s (1996) study of matriculat ion behaviors of lowincome students, the researchers found that students who attended prestigious universities were more likely to receive motivational me ssages from parents than from counselors, peers and other educational role models. In addition, Cabrera a nd LaNasa (2000) found parental influence to have a direct and positive relationship with the formation and maintenance of college aspirations. Finally, according to a 2007 report by the National Postsecondary Education Cooperative (MacAllum, Glover, Queen & Riggs, 2007), “Regardless of socioeconomic status (SES) or ethnic and racial category, parents play the strongest role in the college choice and decision-making processes for traditional-aged students” (p. iii). Despite the strong influence from pare nts, many students consider high school counselors to be an important source of info rmation (Bradshaw, et al., 2001; Gonzalez, et al., 2003). The advice of high school counselors is more influential with students whose parents had little formal education and who came from lower SES backgrounds (MacAllum, et al., 2007). Lillard and Gerner ( 1999) explored the impact that a disrupted family has on the likelihood of students applyi ng to and attending four-year colleges and selective four-year colleges a nd found that a disruption alone is not a significant indicator of the likelihood of students atte nding a particular type of institution. Rather, there was a relationship between the levels of resources available to the family and type of college choice, regardless of whether or not the parental unit was intact in the family.

PAGE 52

44 The Cooperative Institutional Research Program (CIRP) The data for the study will be gathered from the Cooperative Institutional Research Program (CIRP) Freshman Surv ey for 2004. The CIRP Freshman Survey, managed by the Higher Education Research Institute (HERI) at the University of California at Los Angeles, is administered annually to over 400,000 entering freshmen at approximately 700 two-year and four-year co lleges and universitie s (Higher Education Research Institute webpage, March 2008). Th e survey gathers information about (a) established behaviors in high school, (b) academic pr eparedness, (c) admissions decisions, (d) expectations for college, (e) inte ractions with peers a nd faculty, (f) student values and goals, (g) student demographic characteristics, and (h) concerns about financing college (HERI web s ite, retrieved June 2, 2008, from http://www.gseis.ucla.edu/heri/cirp.php ). The 40-question survey is attached at the end of this research proposal as Appendix A. The CIRP was selected because it has been identified as the most comprehensive of several broad-based instruments which surv ey freshmen on issues related to college choice. Other available surveys include the ACT Profile with six factors related to college choice, the Admitted Student Questionnaire (A SQ) Plus which includes 13 factors, and the National Center for Education Statistics ’ (NCES) National Edu cational Longitudinal Study of 1988 (NELS) offering 15 choice f actors (Hoyt & Brown, 2003). The CIRP details 21 factors related to college choice. Another strength of the CIRP is its high response rate relative to other su rveys. This can be attributed to many of the participating institutions asking students to complete the survey during a freshman orientation

PAGE 53

45 program. The strong level of responses on the CIRP minimizes issues of unrepresentativeness of respondents, as the CIRP may more closely resemble a census than a survey (Porter & Whitcomb, 2005). Due to the strengths of comprehensiv eness and representativeness, many researchers have turned to the CIRP to answer a variety of questions related to postsecondary education. In addition, the seri es of surveys offered by CIRP, including the Freshman Survey, Your First College Year, and the Senior Survey, provide opportunities for longitudinal studies. Admini stered at the point of college entry, the Freshman Survey gathers baseline data; while Your First College Year, given to students at the end of their freshman year, gathers information about institutional characte ristics and student experiences in the college environment (Keup, 2004). For example, a study on college student engagement and retention is most effectively assessed if the rese archer can compare a college senior’s responses with her responses as a new freshman, and conclusi ons can be made regarding whether the student’s level of engagement is attributable to institutional policies and practices or to the characteristics of the student (Asti n, 2005-2006). Further, Keup (2004) noted that multivariate analyses of data from the Fres hman Survey and Your First College Year may provide important information about potenti al causal connections between variables. Research conducted by Astin and Lee (2003), for which CIRP survey data was used, indicated that 86 percen t of the variance in student outco mes could be explained solely on entering student characteristics.

PAGE 54

46 Summary Research has revealed that graduate s from higher-tiered and private 4-year institutions generally earn higher salaries th an graduates from other types of colleges, even when controlling for other characteri stics (Leslie & Brinkm an, 1988). Therefore, it would seem advantageous for a student to enroll in and graduate from a higher-tiered university if possible to do so. The literature review demonstrates that, for traditional-age high-achieving students, several factors aff ect the ultimate choice of where students choose to attend. There are individual characteristics of students and their families, such as gender, ethnicity and socioeconomic st atus, which appear to indicate some matriculation tendencies. Likewise, students develop perceptions and preferences about institutions from experiences, marketing efforts of the institutions, and media publications. Those institutional characteris tics may include academic reputation, costs and financial aid, and social climate. Finally, college choi ce decisions of high-ability students are impacted by the influences of significant people in their lives, including parents, counselors and teachers.

PAGE 55

47 CHAPTER 3 METHOD This chapter describes the method used by the researcher to address the research questions. The section addresses the resear ch design, including a description of the sample, the survey instrument and data collec tion plan; and it also describes the plan for data analysis using multiple regression. The purpose of the study was to examine whether differences exist between traditional-aged hi gh achieving students who choose to attend higher-tiered universities a nd their peers who choose a lo wer-tiered university. The research questions and predictor variable s for the study were chosen based on prior research that has been conducted relating to college choice. The researcher employed variables that have been identified in th e literature as factors which high achieving students prioritize during the college choice experience. Research Questions A review of the research that has examin ed the college choice of high achieving students in U.S. postsecondary institutions prov ided the basis for the research questions addressed in this study. The following re search questions guided the study: 1. To what extent do students’ individual characteristics (e.g. gender and ethnicity) relate to college choice for high achieving students?

PAGE 56

48 2. To what extent do students’ family char acteristics (e.g. parent s’ education level and family income) relate to colleg e choice for high achieving students? 3. To what extent do financial considerations associated with college (e.g. cost and financial aid) relate to college choice for high ach ieving students? 4. To what extent does academic reputati on of the institution relate to college choice for high achieving students? 5. To what extent does the influence of si gnificant others (e.g. parents, relatives, teachers, and counselors) relate to co llege choice for high achieving students? Research Design To answer this study’s research ques tions, a causal-comparative research design was utilized. Causal-comparative methodology allows for the exploration of possible causes for the phenomenon being studied by comparing subjects for whom a characteristic is present (e.g. attendance at a higher-tiered university) with similar subjects for whom the characteristic is absent or present to a lesser degree (e.g. attendance at a lower-tiered university). The study applied a quantitative research design incorporating secondary analys is of data gathered by the Higher Education Research Institute (HERI) through the Cooperative Institutional Research Program (CIRP) Freshman Survey for 2004. Secondary analysis of the CIRP data was selected because the methodology provides an efficient and reliable means of obtaining data. The research methods chosen for this study are consiste nt with previous li terature on factors influencing college choice. Among the areas disc ussed in this secti on are considerations

PAGE 57

49 when using secondary data, information about the CIRP Freshman Survey, and discussion regarding the da ta collection process. Secondary Data Considerations Secondary databases can serve as an excel lent source of larg e sample sets of student data, but there are considerations c oncerning advantages and disadvantages that should be made by the researcher prior to co mmitting to the use of secondary data. One of the principal advantages of using secondary data is the savings of time, costs, and resources. Data collection tends to be the most expensive aspect of a research project, and the use of secondary data allows the research er to devote more attention to other issues related to the study (Moriarty, H. J., Deatri ck, J. A., Mahon, M. M., Feetham, S. L., Carroll, R. M., Shepard, M. P., & Orsi, A. J., 1999). Researchers w ho use secondary data have the opportunity to eliminate several tim e-consuming steps in the research process such as developing the instrument, obtai ning the sample, collecting the data, and preparing the data for analysis. Other advant ages of conducting research using secondary data is the ability to study larger samples, to study more representative samples, and to include more variables than can be done in many studies that are based on primary data (Moriarty, et. al, 1999). Finally, when th e study involves a nati onal population, large sample sets provided by national databases can provide the power needed to make generalizations of the findings (Hilton, 1992). Researchers must be cautious, however, when deciding to use secondary data sources, and note the disadvantages of such a choice. The first drawback has to do with

PAGE 58

50 the lack of intimate knowledge that the researcher has of the data. Information on the instrument and procedures used to collect th e data may not be read ily available, raising questions of validity and reliability. In addi tion, the secondary dataset may not be a good fit for the purpose of addre ssing a new research question. For example, the researcher may make incorrect assumptions about the inte nded definition of certain terms used in the instrument. “In some cases, secondary anal ysts are able to change their concepts’ definitions to match the original ones and sti ll be faithful to their theoretical framework. In other cases, this is not possi ble—the concepts are too differe nt and forcing the fit is not appropriate” (Moriarty, et. al, 1999, p. 148). Researchers should also safeguard agains t forcing a match between the research study at hand and the identified secondary da tabase (Kiecolt & Nathan, 1985). Moreover, the use of secondary data makes it more difficu lt to detect bias in the study, because the researcher did not participate in either the development and testing of the instrument or the identification of th e sample (Moriarty, et. al, 1999). Fi nally, there may be issues of timeliness between the collection of the da ta and the completion of the secondary analysis, particularly if significant events occurred between the tw o processes that may impact the relevance and gene ralizability of the findings sa mple (Moriarty, et. al, 1999). Survey Instrument The data for the study were gathered from the Cooperative Institutional Research Program (CIRP) Freshman Survey for 2004. CIRP has been conducting national longitudinal studies of American college st udents since 1966 and has surveyed over eight

PAGE 59

51 million students. The CIRP Freshman Survey, managed by the Higher Education Research Institute (HERI) at th e University of California at Los Angeles, is administered annually to over 400,000 entering freshmen at approximately 700 two-year and four-year colleges and universities (Higher Educati on Research Institut e webpage, March 2008). The survey gathers information about (a) established behaviors in high school, (b) academic preparedness, (c) admissions deci sions, (d) expectations for college, (e) interactions with peers and f aculty, (f) student values and goals, (g) student demographic characteristics, and (h) concerns about fina ncing college (HERI web site, retrieved June 2, 2008, from http://www.gseis.ucla.edu/her i/cirp.php). The 40-question survey is attached at the end of this research proposal as Appendix A. The CIRP was selected because it has been identified as the most comprehensive of several broad-based instruments which surv ey freshmen on issues related to college choice. Other available surveys include the ACT Profile with six factors related to college choice, the Admitted Student Questionnaire (A SQ) Plus which includes 13 factors, and the National Center for Education Statistics ’ (NCES) National Edu cational Longitudinal Study of 1988 (NELS) offering 15 choice f actors (Hoyt & Brown, 2003). The CIRP details 21 factors related to college choice. Another strength of the CIRP is its high response rate relative to other su rveys. This can be attributed to many of the participating institutions asking students to complete the survey during a freshman orientation program. The strong level of responses on the CIRP minimizes issues of unrepresentativeness of respondents, as the CIRP may more closely resemble a census than a survey (Porter & Whitcomb, 2005).

PAGE 60

52 Validity and Reliability As mentioned in the previous section, one of the disadvantages of using secondary data is that the researcher does not have first-hand information about the steps taken by the administrators of the database to maximize validity and reliability. The credibility of research studies depends greatly on the validity and reliability of the measures. For studies using surveys or que stionnaires as the measurement instrument, validity refers to the accuracy of the infere nces or interpretations one makes from the responses, and reliability refers to the cons istency or accuracy of the responses. HERI has addressed these questions through a document posted on its website entitled “CIRP Freshman Survey: Reliability and Va lidity” (retrieved Ju ly 27, 2008, from http://www.gseis.ucla.edu/heri/PDF s/CIRP_Reliability_Validity.PDF ). The document is attached as Appendix B. A survey question is considered reliabl e if similar results are yielded when repeatedly administered to similar samples. HERI has made the assertion that the majority of questions in the CIRP Freshman Survey have exhibited a “great deal of stability” over the nearly forty years that the survey has been administered and that observed exceptions to this stability have been “linked to temporal tr ends or to real and meaningful exogenous shocks” (p. 1). In addi tion, HERI states that nearly 90 percent of the participating institutions are repeat participants, which helps to ensure sample consistency over time. Validity refers to the interpretation of survey responses and the degree to which the interpretation is supporte d by evidence and theory (Gall, Gall & Borg, 2007). HERI admitted that it has not conduct ed factor analysis for all survey items,

PAGE 61

53 but referred researchers to lit erature where the validity of th e CIRP has been investigated (Astin, 1991, 1992; Luo & Jamieson-Drake, 2005). Sample Institutional participation of the CIRP Freshman Survey is voluntary. The scope of this study was limited to public and privat e national research unive rsities as identified by USNWR As not all national unive rsities participated in the CIRP, the sample was drawn from the 97 universities which admini stered the survey in 2004, as listed in Appendix C. Participation in the 2004 Fres hman Survey included 32 (33%) Tier one universities, 32 (33%) Tier tw o universities, 19 (20%) Tier three universities, and 14 (14%) Tier four universities. A glaring issue th at the researcher addresses as a limitation is the disparity in participation between hi gher-tiered and lower-tie red universities. The researcher has failed to find any explanati on for the gap in institutional participation, particularly between the highest-ranked and lowest-ranked universities. Because this study is focused on the college choice behaviors of high achieving students, data were used from only those st udents who indicated that they had received scores at or above 660 on the critical reading portion of the SAT, and scores at or above 670 on the mathematics portion of the SAT. For students that did not report scores for both SAT verbal and SAT math, the research accepted data from students reporting an ACT composite score of 30 or greater. The scores of 660 and 670 were used as benchmarks because they represent the poi nt at which students scored in the 90th percentile for the reading and math portions of the test (College Board website, accessed

PAGE 62

54 March 29, 2008). In addition, for their data to be used, students were required to have an A or A+ average in high school. Descriptive st atistics for the students and institutions that were selected for inclusion in this study are reported in Chapter 4. The researcher attempted to control for va riables such as issues that arise for students who attended a university that was not their first choi ce. Student data were used only for students who indicated that they were attending their first choice college. This is important to maintain integrity in the examination of the factors which influence students to attend a lower-tiered university, because students truly are choosing to attend a university if the university is their first choice If a student enrolled at a university that was his or her second, third, or further choice, the data would not be truly reflective of the institutional attributes that are most important to the student. While some important research has been done exploring the reasons that students do not enroll at their firstchoice institution (Chapman & Jackson, 1987), this line of inquiry is outside the scope of the present study. Finally, the study limited the sample to st udents enrolled full-time and those who are attending an institution located more th an 100 miles from their home. There are unique issues and extraneous variables asso ciated with students who choose to attend college part-time and similarly for those who choose to attend a school close to home. Many students may settle for a university that is within a short commute from home. Those students are not of interest for the purpos es of this study, beca use issues related to convenience are unrelated to the nature of the questions this st udy seeks to address.

PAGE 63

55 Data Collection The data for the study was gathered from the Higher Education Research Institute (HERI) at the Universi ty of California at Los Angeles, via the Coopera tive Institutional Research Program (CIRP) Freshman Survey for 2004. The CIRP Freshman Survey is administered annually to over 400,000 entering freshmen at approximately 700 two-year and four-year colleges and uni versities (Higher Education Re search Institute website, March 2008). Each year, HERI invites re gionally accredited institutions of higher education (excluding proprietary, special vocat ional and semi-professional institutions) to participate in the CIRP survey. The national population for the survey is all baccalaureate degree-granting institutions which admit first-time freshmen. Participants represent public and private institutions, historically Black colleges and universities, and both religious and non-sectarian institutions. Institutional contribution to the Integrat ed Postsecondary Education Data System (IPEDS) and a full-time freshman class size of 25 students are required for eligibility. Although institutional participation in the survey varies from year-to year, most of the postsecondary institutions that pa rticipate in the survey are repeat customers and typically ask students to complete the surv ey during freshman orientation ( http://www.gseis.ucla.edu/ heri/cirpoverview.php ). Institutions may administer the survey in any of the following ways: 1. Proctored setting with paper questio nnaires – This is the recommended method as it results in the highest response rate. 2. Mail-out survey with paper questionnaire

PAGE 64

56 3. Email notification of the web-su rvey option – New for 2008 survey 4. A combination of paper and web-base d questionnaires – New for 2008 survey The survey gathers information about: (a) established behaviors in high school, (b) academic preparedness, (c) admissions deci sions, (d) expectations for college, (e) interactions with peers and f aculty, (f) student values and goals, (g) student demographic characteristics, and (h) concerns about fi nancing college (Highe r Education Research Institute, 2008). The 40-question survey is at tached at the end of this research proposal as Appendix A. The CIRP Freshman Survey was selected because it was identified as the most comprehensive of the existing and availabl e broad-based instruments that survey freshmen on issues related to college choi ce. Other available surveys include the ACT Profile with six factors related to colleg e choice, the Admitted Student Questionnaire (ASQ) Plus which includes 13 factors, and th e National Center for Education Statistics’ (NCES) National Educational Longitudina l Study of 1988 (NELS) offering 15 choice factors (Hoyt & Brown, 2003). The CIRP details 21 factors related to college choice. HERI publicizes procedures on its website for requesting information from its databases (http://www.gseis.uc la.edu/heri/gainaccess.php). Fo llowing is a list of items which are evaluated by HERI staff in determini ng whether to provide data for a particular study: 1. HERI data adequately matches the proposed research project;

PAGE 65

57 2. The study design is adequate to answer the questions being asked, theoretical grounding is evident, and the proposal provides sufficient detail about dependent and independent variables; 3. The proposal details the process by wh ich the investigator will acquire appropriate institutional review board approval; 4. The intended plan specified by the investigator in volves advancing scholarship; and 5. The research is conducted in a manner that minimizes conflicts with other research conducted by HERI staff or other investigators under previously approved projects. National universities for which data ar e available from the 2004 CIRP were categorized into four ordinal gr oups: (1) Tier one, (2) Tier two, (3) Tier three, or (4) Tier four, according to their assignment by USNWR The 2003 issue of Best Colleges placed a total of 248 universitie s into the “national universities” category, of which the top 51 were assigned a numerical rank and placed into Tier one; another 78 were placed unranked into Tier two; 65 were placed unranke d into Tier three; and 55 institutions were unranked and placed into Tier four. HERI makes available on its website a participation history for each of the surveys that it administers. USNWR ranking statuses from the 2003 issue of “America’s Best Colleges” were assigned by the researcher to institutions participating in the 2004 Freshman Survey, and the information were provided to HERI. The researcher received a SPSS-formatted data file from HERI that includes 2004 Freshman Survey responses

PAGE 66

58 from (1) students who indicated that they had received scores at or above 660 on the critical reading portion of the SAT, and sc ores at or above 670 on the mathematics portion of the SAT, or the equi valent ACT score; (2) studen ts who reported an A or A+ average in high school, (3) stude nts who reported attending thei r college of first choice, (4) students who are enrolled full-time, and (5) students who are a ttending an institution located more than 100 miles from their home. Description of Variables Guided by theory and relevant existing literature, a limited nu mber of variables from the CIRP 2004 Freshman Survey data base were used to operationalize the constructs referenced within the research que stions. This section will further describe the variables selected for this study, beginning with the ordinal outcome variable and concluding with a discussion of th e various independent variables. The outcome (or dependent) variable fo r this study is the tier level of the university at which a student is enroll ed. The data received from the 2004 CIRP Freshman Survey included responses from st udents attending institut ions classified by CIRP as public or private research univers ities. Each participating institution was assigned a tier level of one, two, thre e, or four, based upon its assignment by USNWR in its 2003 Best Colleges edition. The independent variables selected for th is study are listed in Table 3.1 and are grouped according to major categories within the college choice literature. Studies indicate that students’ enrollment decisions are related to: (1) students’ individual and

PAGE 67

59 family characteristics (Brewer, et al., 1999; Hearn, 1987; Hearn, 1991; Manski & Wise, 1983; Paulsen, 1990); (2) students’ preferences about the colle ges they are considering (Fuller, Manski, & Wise, 1982; Hossler & Ga llagher, 1987; Weiler, 1994); and (3) the influences of significant others, including parents, relatives, teachers and counselors (Bradshaw, et al., 2001; Hossler, Schmit, & Vesper, 1999; Lillard & Gerner, 1999). Coding of the variables is based on the st ructure of options available to students responding to the 2004 CIRP Freshman Survey. Data Analysis To provide some initial understanding of the differences between high achieving students enrolled at each of the four tiers of national universities, frequencies and descriptive statistics are presen ted to gain an understanding of the distribution of the data. In addition, a correlation matrix of all independent variables in the study is presented to demonstrate the resulting relationships between variables. Multivariate analyses involving multiple regression models were conduc ted to examine the predictive ability of the independent variables, while controlling fo r other variables in th e model, in relation to choice of college for high-achieving student s. Because the dependent variable is a set of ordinal outcomes ( USNWR tier assignment), multiple regression is the preferred statistical method for understa nding the relationship between the independent variables and students’ matriculation behaviors. The outcome variable is the student’s choice of college, with four possible outcomes according to the tier category to wh ich the university was assigned. Regression

PAGE 68

60 has commonly been used in studies of co llege choice research (Hu & Hossler, 2000). First, using SPSS statistical so ftware, a series of regressi on analyses were conducted to test the significance of observed differences among traditional-age high-achieving college freshmen in terms of (1) the students’ individual characterist ics, (2) the students’ preferences about the co lleges they are considering, and (3 ) the influences of others. The CIRP 2004 Freshman Survey includes severa l questions pertaining to the research questions that guide the study at hand. The researcher select ed a group of variables from the questions included in the Freshman Survey, as listed in Table 3.1, and developed a plan for measuring the variables consisting of descriptive statistics and multiple regression. Summary The purpose of the study was to examine whether differences exist between traditional-aged high-achieving students who choose to attend higher-tiered universities and their peers who choose a lower-tiered university experience. The researcher has proposed to explore the stated research questions by engagi ng in a causal-comparative research design that uses secondary data fr om 87 public and private research universities participating in the 2004 Freshman Survey administered by the Higher Education Research Institute. The researcher analyzed the data using descriptive statistics and multiple regression. The proposed methods are cons istent with prior research in the area of college choice.

PAGE 69

61 Table 3.1 Independent Variable Construction and Coding Scheme Variable name Operational Definition Student and Family Characteristics Gender Female Male Ethnicity White African American Asian American Hispanic Other ethnicity Parents Education Father’s education High school graduate or less Some postsecondary education College degree Some graduate school Graduate degree Mother’s education High school graduate or less Some postsecondary education College degree Some graduate school Graduate degree Family income Less than $50K $50K-100K $100K-150K Greater than $150K Student Employment Very little or no chance Some chance Very good chance Q1: “Your sex” Reference group A dummy equal to 1 if the student is male Q25: “Please indicate your ethnic background.” Reference group A dummy equal to 1 if the student is an African American; 0 = other A dummy equal to 1 if the student is an Asian American; 0 = other A dummy equal to 1 if the student is a Hispanic; 0 = other A dummy equal to 1 if the student indicated a group not mentioned above; 0 = other Q28: “What is the highest level of formal education obtained by your parents?” Reference group Equal to 1 if the father attended college or other postsecondary school Equal to 2 if the father has a college degree Equal to 3 if the father attended graduate school Equal to 4 if the father has a graduate degree Reference group Equal to 1 if the mother attended college or other postsecondary school Equal to 2 if the mother has a college degree Equal to 3 if the mother attended graduate school Equal to 4 if the mother has a graduate degree Q22: “What is your best estimate of your parents’ total income last year? Consider income from all sources before taxes.” Reference group Equal to 1 if family income is between $50,000 and $99,999 Equal to 2 if family income is between $100,000 and $149,999 Equal to 3 if family income is greater than $150,000 Q40: “What is your best guess as to the chances that you will…Get a job to help pay for college expenses?” Reference group Equal to 1 if “some chance” marked Equal to 2 if “very good chance” marked

PAGE 70

62 Table 3.1 (continued) Independent Vari able Construction and Coding Scheme Variable name Operational Definition Institutional Characteristics Institutional costs Not important Somewhat important Very important Financial aid Not important Somewhat important Very important Academic reputation Not important Somewhat important Very important Media rankings Not important Somewhat important Very important Q37: “How important was [the cost of attending this college] in your decision to come here?” Reference group Equal to 1 if cost of attendance was somewhat important Equal to 2 if cost of attendance was very important Q37: “How important was [the offer of financial assistance by the college] in your decision to come here?” Reference group Equal to 1 if financial assistance was somewhat important Equal to 2 if financial assistance was very important Q37: “How important was [the academic reputation of the college] in your decision to come here?” Reference group Equal to 1 if academic reputation was somewhat important Equal to 2 if academic reputation was very important Q37: “How important was [rankings in national magazines] in your decision to come here?” Reference group Equal to 1 if media rankings were somewhat important Equal to 2 if media rankings were very important Influence of Others Parental influence Not important Somewhat important Very important Relative influence Not important Somewhat important Very important Teacher influence Not important Somewhat important Very important Counselor influence Not important Somewhat important Very important Q29: “In deciding to go to college, how important to you was [your parents wanting you to go]?” Reference group Equal to 1 if parental influence was somewhat important Equal to 2 if parental influence was very important Q37: “How important was [your relatives wanting you to come here] in your decision to come here?” Reference group Equal to 1 if the influence of relatives was somewhat important Equal to 2 if the influence of relatives was very important Q37: “How important was [advice from a teacher] in your decision to come here?” Reference group Equal to 1 if teacher advice was somewhat important Equal to 2 if teacher advice was very important Q37: “How important was [advice from a high school counselor] in your decision to come here?” Reference group Equal to 1 if counselor advice was somewhat important Equal to 2 if counselor advice was very important

PAGE 71

63 CHAPTER 4 RESULTS The study applied a quantita tive research design incor porating secondary analysis of data gathered by the Hi gher Education Research In stitute (HERI) through the Cooperative Institutional Research Program (CIRP) Freshman Survey for 2004. Because both the researcher and the Higher Educati on Research Institut e (HERI) wanted to maintain anonymity for both student and in stitutional responses, neither student nor institutional identifiers were provided by HERI to the researcher. To link the responses for the independent variables with the outco me variable of tier group, the researcher provided HERI with the tiers assigned to each institution according to the 2003 edition of America’s Best Colleges; and HERI then adde d the outcome variable to the dataset prior to distributing the data to the researcher. The dataset was sent as an SPSS file to the researcher via email. The scope of this study was limited to public and private national research universities as identified by USNWR As not all national universities participated in the CIRP, the sample was drawn from the 97 unive rsities that administered the survey in 2004, as listed in Appendix C. Participation in the 2004 Freshman Survey included 32 (33%) Tier One universities, 32 (33%) Tier Two universities, 19 (20%) Tier Three universities, and 14 (14%) Ti er Four universities. As expected, due to the large proportion of participating Tier One institutions in the 2004 Freshman Survey, there was

PAGE 72

64 a large disparity in the number of eligib le respondents among tie r groups. The resulting dataset for the study included responses fr om 6,889 students. Se venty-seven percent (n=5,335) of the respondents were from Ti er One institutions, compared to 16.7% (n=1,149) from Tier Two, 4.7% (n=324) from Ti er Three, and 1.2% (n=81) from Tier Four institutions. Student data were used only for student s who indicated that they were attending their first choice college and th at their selected institutions were located more than 100 miles from their homes. In addition, because this study focused on the college choice behaviors of high-achieving students, data were used only from those students who indicated that they had received scores at or above 660 on the Critical Reading portion of the SAT, and scores at or above 670 on the mathematics portion of the SAT. For students that did not report scores fo r both SAT verbal and SAT mat h, the researcher accepted data from students reporting an ACT composite score of 30 or greater. The scores of 660 and 670 were used as benchmarks because th ey represent the point at which students scored in the 90th percentile for the reading and math portions of the test (College Board website, accessed March 29, 2008). The ACT composite score of 30 was used because ACT and the College Board have identified a score of 30 on the ACT as comparable to a score of 1330-1350 on the combination of the SAT verbal score and SAT math score (ACT website, accessed December 4, 2008). In addition, for their data to be used, students were required to have an A or A+ av erage in high school. Descriptive statistics for the distribution of SAT verbal, SAT mat h, and ACT composite scores for each tier group are reported in Table 4.1.

PAGE 73

65 To provide some initial understanding of the differences between high achieving students enrolled at each of the four tiers of national universities, frequencies and descriptive statistics are presen ted to gain an understanding of the distribution of the data. In addition, a correlation matrix of all independent variables in the study is presented to demonstrate the resulting relationships between variables. Multivariate analyses involving multiple regression models were conduc ted to examine the predictive ability of the independent variables, while controlling fo r other variables in th e model, in relation to choice of college for high-achieving students. Table 4.1 – Distribution of Sample SAT a nd ACT Scores by Tier of Institution Tier SAT/ACT Scores 1 2 3 4 SAT Verbal n=4,808 n=671 n=103 n=18 Mean 723 711 716 704 Median 720 700 700 690 Mode 800 700 700 680 Standard Dev 43 39 42 36 Range 660-800 660-800 660-800 660-770 SAT Math n=4,808 n=671 n=103 n=18 Mean 736 718 718 711 Median 730 710 710 700 Mode 800 700 720 700 Standard Dev 42 37 36 38 Range 670-800 670-800 670-800 670-800 ACT Composite n=527 n=478 n=221 n=63 Mean 32 31 31 31 Median 32 31 31 31 Mode 32 30 30 30 Standard Dev 1 1 1 1 Range 30-36 30-35 30-36 30-34

PAGE 74

66 Descriptive Statistics A review of the research that has examin ed the college choice of high achieving students in U.S. postsecondary institutions prov ided the basis for the research questions addressed in this study. The following rese arch questions guided this study: 1. To what extent do students’ individual characteristics (e.g. gender and ethnicity) relate to college choice for high achieving students? 2. To what extent do students’ family char acteristics (e.g. parent s’ education level and family income) relate to colleg e choice for high ac hieving students? 3. To what extent do financial considerations associated with college (e.g. cost and financial aid) relate to college choice for high ach ieving students? 4. To what extent does academic reputati on of the institution relate to college choice for high achieving students? 5. To what extent does the influence of si gnificant others (e.g. parents, relatives, teachers, and counselors) relate to co llege choice for high achieving students? Student and Family Characteristics Gender and Ethnicity Respondents included slightly more males (n=3,596, 52.2%) than females (n=3,286, 47.7%). Table 4.2 provides a summ ary of the number and percentage of female and male respondents by each of th e four institutional tiers. The greatest differences observed were in Tier Two ins titutions, with 25% more males (n=638) than

PAGE 75

67 females (n=511), and in Tier Four institutions, with 31% more females (n=46) than males (n=35). Table 4.2 – Distribution of Respondent Gender by Tier of Institution The ethnic diversity of the sample in th e current study was less than optimal. The sample of respondents was overwhelmingl y comprised of white students (n=5,571, 80.9%). Table 4.3 provides a summary of the num ber and percentage of students within each ethnic group by each of the four inst itutional tiers. Respondents from Tier One institutions were more diverse than any other tier, with white students accounting for three out of four (n=4,112, 77.1%) respondents. Minorities comprised less than 7% of any other tier group. Representati on of Black students was especially low, with only five (0.4%) in Tier Two, one (0.3%) in Tier Three, and zero in Tier Four. Compared with the lower-tiered institutions the Ti er One institutions received responses from a strikingly higher percentage of Asian students. Nearly 15% (n=797) of the respondents from Tier Tier Gender 1 2 3 4 Female 2563 511 166 46 % of Tier 48.1% 44.5% 51.2% 56.8% Male 2765 638 158 35 % of Tier 51.8% 55.5% 48.8% 43.2% No Response7 0 0 0 % of Tier 0.1% Total 5335 1149 324 81 Note: X2 (3, N=6,882) = 9.442, p=.024

PAGE 76

68 One institutions identified themselves as Asian, compared to 3.2% (n=37) from Tier Two, 3.1% (n=10) from Tier Three, and 2.5% (n=2) from Tier Four institutions. Table 4.3 – Distribution of Respondent Ethnicity by Tier of Institution Socioeconomic Status The frequency distribution indicated some distinct differences in family income among tier groups. Table 4.4 demonstrates an in verse relationship between income level and tier group. That is, the pr oportion of students indicati ng a family income of $150,000 Tier Ethnicity 1 2 3 4 White 4112 1075 308 76 % of Tier 77.1% 93.6% 95.1% 93.8% Blacka 76 5 1 0 % of Tier 1.4% 0.4% 0.3% 0% Asian b 797 37 10 2 % of Tier 14.9% 3.2% 3.1% 2.5% Hispanicc 223 25 2 3 % of Tier 4.2% 2.2% 0.6% 3.7% Other d 127 7 3 0 % of Tier 2.4% 0.6% 0.9% 0% No Response0 0 0 0 % of Tier Total 5335 1149 324 81 aNote: X2 (3, N=6,889) = 11.177, p=.011 bNote: X2 (3, N=6,889) = 155.255, p<.001 cNote: X2 (3, N=6,889) = 19.708, p<.001 dNote: X2 (3, N=6,889) = 18.941, p<.001

PAGE 77

69 per year or greater is highest for Tier One st udents, and declines at each step of the tier ladder. Students attending Tier Four institutio ns were twice as likely to have a family income of $100,000 or less than were students attending Tier One institutions; with more than one in five (n=18, 22.2%) Tier Four students reporting a family income of $50,000 or less, compared with only one of ev ery ten (n=550, 10.3%) st udents at Tier One institutions. There is a similar disparity wh en examining the other end of the income spectrum. Students attending Tier One inst itutions reported a fam ily income of over $150,000 at nearly twice the ra te (n=1,789, 33.5%) of Tier Tw o students (n=198, 17.2%), three times the rate of Tier Three students (n=37, 11.4%), and nearly seven times the rate of Tier Four students (n=4, 4.9%). Table 4.4 – Distribution of Family Income by Tier of Institution Tier Family Income 1 2 3 4 Less than $50K 550 138 64 18 % of Tier 10.3% 12.0% 19.8% 22.2% $50K-$100K 1348 419 139 38 % of Tier 25.3% 36.5% 42.9% 46.9% $100K-$150K 1146 276 60 9 % of Tier 21.5% 24.0% 18.5% 11.1% Over $150K 1789 198 37 4 % of Tier 33.5% 17.2% 11.4% 4.9% No Response502 118 24 12 % of Tier 9.4% 10.3% 7.4% 14.8% Total 5335 1149 324 81 Note: X2 (9, N=6,233) = 263.626, p<.001

PAGE 78

70 Relative to the response ra te obtained from the sample for other independent variables, the response rate from students re garding family income was noticeably low. Nearly one of ten (n=656, 9.5%) students in the sample failed to respond to the question in the 2004 Freshman Survey regarding fam ily income. The relatively large number of missing data may limit analyses and conclu sions regarding family income and its relationship with the tier level of univers ity that a student chooses to attend. As conveyed in Table 4.5, the majority of students among all tier groups indicated that there was at least some chance that they would need to seek employment to help pay for college expenses, with students in Tier Four institutions indicating a stronger need than the students in other tier groups. Only three (3.7%) of the 81 students enrolled at Table 4.5 – Student Employment Needs by Tier of Institution Tier Need for Student Employment 1 2 3 4 No/Little Chance 1325 225 97 3 % of Tier 24.8% 19.6% 29.9% 3.7% Some Chance 1625 387 96 30 % of Tier 30.5% 33.7% 29.6% 37.0% Very Good Chance2221 519 121 48 % of Tier 41.6% 45.2% 37.4% 59.3% No Response 154 18 10 0 % of Tier 2.9% 1.2% 3.1% 0% Total 5335 1149 324 81 Note: X2 (6, N=6,707) = 42.866, p<.001

PAGE 79

71 Tier Four institutions responded that there was little to no chance that they would need to get a job to pay for their education. This lo w response rate from Tier Four students, in contrast to the responses from Tiers 1 (n=1,325, 24.8%), 2 (n=225, 19.6%), and 3 (n=97, 29.9%), indicates that the Tier Four students in the sample rarely perceived themselves as having a financial status that would allow th em to study without wo rking at least part time. Education Level of Parents As expected, students atte nding Tier One institutions reported the highest levels of education for their fathers, with over half of respondents (n=2,952, 55.3%) reporting that their fathers possessed a graduate degree. Further, the fathers of Tier One students were least likely to lack any college experi ence. Less than 6% (n=300) of fathers of students attending Tier One in stitutions lacked a college education; however, the percentage rises over 11% for fathers of students in Tiers 2 (n=131), 3 (n=45), and 4 (n=9). A summary of the education level for fathers of respondents by each of the four institutional tiers is provided in Table 4.6. Similar to the results regarding the edu cation level of father s, students attending Tier One institutions reported the highest levels of education for their mothers, with over one-third of respondents (n=2,024, 37.9%) re porting that their mothers possessed a graduate degree. Contrary to the results for th e fathers, the mothers of Tier Four students were least likely to lack a college educati on. For both fathers and mothers, Tier Three students had the highest percen tage of parents with no colle ge education. A summary of

PAGE 80

72 the education level for mothers of respondents by each of the four in stitutional tiers is provided in Table 4.7. Table 4.6 – Distribution of Father’s E ducation Level by Tier of Institution Tier Father’s Education 1 2 3 4 No College 300 131 45 9 % of Tier 5.6% 11.4% 13.9% 11.1% Some College 407 156 55 14 % of Tier 7.6% 13.6% 17.0% 17.3% College Degree1638 458 130 42 % of Tier 30.7% 39.9% 40.1% 51.9% Grad Degree 2952 397 93 16 % of Tier 55.3% 34.6% 28.7% 19.8% No Response38 7 1 0 % of Tier 0.7% 0.6% 0.3% 0% Total 5335 1149 324 81 Note: X2 (12, N=6,843) = 317.756, p<.001

PAGE 81

73 Table 4.7 – Distribution of Mother’s E ducation Level by Tier of Institution Eighty-six percent (n=4,590, 86%) of Tier On e students reported that their fathers had earned some type of college degree. In comparison, 74.4% (n=855) of Tier Two students, 68.8% (n=223) of Ti er Three students, and 71.7% (n =58) of Tier Four students reported having fathers with college degrees. The results for the mothers were similar, with 82% (n=4,399) of Tier One students repo rting that their mothers had earned some type of college degree. In comparison, 73.8% (n=848) of Tier Two students, 65.1% (n=211) of Tier Three student s, and 70.4% (n=57) of Tier Four students reported having mothers with college degrees. These result s are summarized in Table 4.8, which also shows the distribution for students, by tie r group, with both parents earning college degrees or with both parents lacking a college education. Tier Mother’s Education1 2 3 4 No College 325 114 40 4 % of Tier 6.1% 9.9% 12.4% 4.9% Some College 576 182 72 20 % of Tier 10.8% 15.8% 22.2% 24.7% College Degree2375 584 161 44 % of Tier 44.5% 50.8% 49.7% 54.3% Grad Degree 2024 264 50 13 % of Tier 37.9% 23.0% 15.4% 16.1% No Response35 5 1 0 % of Tier 0.7% 0.4% 0.3% 0% Total 5335 1149 324 81 Note: X2 (12, N=6,848) = 216.871, p<.001

PAGE 82

74 Table 4.8 – Distribution of Parents’ Education Level by Tier of Institution Institutional Characteristics Cost and Availability of Financial Aid Table 4.9 summarizes the responses from high achieving students regarding the level of importance they placed on the cost s of college when choosing to attend. One observation from a review of the distribution of responses is that the majority of students (n=3,002, 56.3%) enrolled at Tier One institu tions found the costs of attendance to be unimportant regarding their matriculation deci sions. The proportion of Tier One students Tier Parents’ Education1 2 3 4 No College Degree Father 707 287 100 23 % of Tier 13.3% 25.0% 30.1% 28.4% Mother 901 296 112 24 % of Tier 16.9% 25.8% 34.6% 29.6% Both 385 162 57 15 % of Tier 7.2% 14.1% 17.6% 18.5% College/Graduate Degree Father 4590 855 223 58 % of Tier 86.0% 74.4% 68.8% 71.7% Mother 4399 848 211 57 % of Tier 82.5% 73.8% 65.1% 70.4% Both 4074 721 168 49 % of Tier 76.4% 62.8% 51.9% 60.5%

PAGE 83

75 finding costs unimportant is considerably highe r than for students from Tier Two (n=302, 26.3%), Tier Three (n=63, 19.4%), and Tier Four (n=12, 14.8%) institutions who reported costs as not factoring into their ultimate college choice. Similarly, students attending Tier Four institutions indicated costs to be very im portant at nearly three times the rate (n=36, 44.4%) of st udents attending Tier One in stitutions (n=834, 15.6%). Table 4.9 – Importance of College Costs by Tier of Institution Similar to the results of student response s to the importance of college costs in their matriculation decisions, th e responses regarding the import ance of financial aid also demonstrate that students attending the lowe r-tiered institutions were much more conscious of financial aid awards than were students attending the highest-tiered institutions. A summary of th e responses from high achieving students regarding the level of importance they placed on the financial aid when choosing to attend is provided in Tier College Costs 1 2 3 4 Not Important 3002 302 63 12 % of Tier 56.3% 26.3% 19.4% 14.8% Somewhat Important 1438 472 128 33 % of Tier 27.0% 41.1% 39.5% 40.7% Very Important 834 363 129 36 % of Tier 15.6% 31.6% 39.8% 44.4% No Response 61 12 4 0 % of Tier 1.1% 1.0% 1.2% 0% Total 5335 1149 324 81 Note: X2 (6, N=6,812) = 549.261, p<.001

PAGE 84

76 Table 4.10. Interestingly, Tier Three student s placed the greatest emphasis on financial aid, with three out of every four students (n=246, 75.9%) indicating that offers of financial aid were very importa nt in their matriculation deci sions. Similar to the results regarding college costs, most students attending Tier One institutions (n=2,808, 52.6%) indicated that financial aid awar ds were not considered in se lecting the college to attend. Table 4.10 – Importance of Financial Aid by Tier of Institution Reputation and Prestige When it comes to the importance of academic reputation to high achieving students when selecting a coll ege, the results of this study support the existing research that asserts that academic reputation is the most important factor. A description of the responses from students regarding the level of importance placed on academic reputation Tier Financial Aid 1 2 3 4 Not Important 2808 217 17 8 % of Tier 52.6% 18.9% 5.3% 9.9% Somewhat Important 917 318 56 30 % of Tier 17.2% 27.7% 17.3% 37.0% Very Important 1543 601 246 43 % of Tier 28.9% 52.3% 75.9% 53.1% No Response 67 13 5 0 % of Tier 1.3% 1.1% 1.5% 0% Total 5335 1149 324 81 Note: X2 (6, N=6,804) = 767.177, p<.001

PAGE 85

77 may be found in Table 4.11. The majority of st udents in all tier gr oups indicated that the academic reputation of students’ college of c hoice was a very important factor in their decision to attend, although there are observable differe nces between tier groups. Students enrolled at the higher-tiered instit utions were most likely to rate academic reputation as very important, with 86.9% (n=4,638) of Tier One students and 73.4% (n=843) of Tier Two students responding accordingly. The proportion of students who considered academic reputation as very impor tant then drops to 53.7% (n=174) of Tier Three students and 55.6% (n=45) of students attending a Tier Four institution. Virtually none (less than 1%) of the students at Tier One universities responded that academic reputation was not at all important in their college choice decision. Table 4.11 – Importance of Academic Reputation by Tier of Institution Tier Academic Reputation1 2 3 4 Not Important 42 19 13 6 % of Tier 0.8% 1.7% 4.0% 7.4% Somewhat Important 614 277 135 30 % of Tier 11.5% 24.1% 41.7% 37.0% Very Important 4638 843 174 45 % of Tier 86.9% 73.4% 53.7% 55.6% No Response 41 10 2 0 % of Tier 0.8% 0.9% 0.6% 0% Total 5335 1149 324 81 Note: X2 (6, N=6,836) = 401.892, p<.001

PAGE 86

78 In addition to gathering information re garding academic reputation, the research also collected student responses to th e importance of media rankings in their matriculation decisions. A su mmary of the responses is contained in Table 4.12. The greatest degree of importance on media rankings was indicated by the students enrolled at Tier One institutions, with over 80% (n=4,297) of respondents in that group responding that media rankings were at least somewhat im portant in their decision to enroll at the particular university. Students in the lower-tiered groups indi cated the least interest and placement of importance on the media’s ranki ng of postsecondary institutions, with twothirds (n=53, 66.3%) of students attending Tier Four institutions indicating that these rankings were not at all important. Table 4.12 – Importance of Media Ra nkings by Tier of Institution Tier Media Rankings 1 2 3 4 Not Important 979 422 183 53 % of Tier 18.6% 37.2% 57.2% 66.3% Somewhat Important 2420 527 108 21 % of Tier 45.9% 46.5% 33.8% 26.3% Very Important 1877 185 29 6 % of Tier 35.6% 16.3% 9.1% 7.5% No Response 59 15 4 1 % of Tier 1.1% 1.3% 1.2% 1.2% Total 5335 1149 324 81 Note: X2 (6, N=6,810) = 549.593, p<.001

PAGE 87

79 Influence of Others Following student and family characteristics and institutional characteristics, the influence of others was explored as a f actor affecting the co llege choices of high achieving students. There are four groups of “o thers” that were investigated, including parents, relatives, teachers, and counse lors. The descriptive data in Table 4.13 demonstrate smaller differences among the tier groups than with other independent variables that have been discussed, with a nywhere from 31% to 40% of the students in each tier group indicating that the influence of their parents was very important in their college decision. Table 4.13 – Importance of Parental Influence by Tier of Institution Tier Parental Influence1 2 3 4 Not Important 1459 286 75 13 % of Tier 27.4% 24.9% 23.2% 16.1% Somewhat Important 2104 470 145 34 % of Tier 39.4% 40.9% 44.8% 42.0% Very Important 1730 388 103 33 % of Tier 32.4% 33.8% 31.8% 40.7% No Response 42 5 1 1 % of Tier 0.8% 0.4% 0.3% 1.2% Total 5335 1149 324 81 Note: X2 (6, N=6,840) = 11.981, p=.062

PAGE 88

80 The descriptive data contained in Tabl e 4.14 convey the responses from students regarding the importance of th e influence of relatives in their decision to attend their college of first choice. A comparison of th e responses among tier gr oups indicates less variance on this vari able than for most of the other vari ables in this study. No more than 6% of students within any of the tier groups indicated that th e influence of relatives was very important, while approximately two-thirds of students within any given tier group responded that the influence of relatives was not a factor that infl uence their choice of which college to attend. Table 4.14 – Importance of Relative Influence by Tier of Institution The influence of teachers and counselors, according to the high achieving students in the sample, appears to be no more important than the influence of relatives. Tables 4.15 and 4.16 summarize the students’ response s to the importance of teachers and Tier Relative Influence1 2 3 4 Not Important 3373 787 210 55 % of Tier 63.2% 68.5% 64.8% 67.9% Somewhat Important 1593 312 95 22 % of Tier 29.9% 27.2% 29.3% 27.2% Very Important 318 37 17 3 % of Tier 6.0% 3.2% 5.3% 3.7% No Response 51 13 2 1 % of Tier 1.0% 1.1% 0.6% 1.2% Total 5335 1149 324 81 Note: X2 (6, N=6,822) = 20.319, p=.002

PAGE 89

81 counselors respectively in their matriculat ion decisions. No more than 5% of the respondents considered the influence of teacher s to be very important; and no more than 7% indicated similar levels of importan ce for counselors. Although the majority of students in all tier groups responded that the in fluence of teachers and counselors was not important in their college enrollment d ecisions, the highest proportion of students indicating such sentiments for both “other ” came from the students at Tier Four institutions. Students atte nding Tier One institutions on the other hand, responded slightly more frequently than the students in other tier groups th at the influence of counselors and teachers were somewhat or very important in their college choice process. Table 4.15 – Importance of Teacher In fluence by Tier of Institution Tier Teacher Influence1 2 3 4 Not Important 3509 859 254 65 % of Tier 65.8% 74.8% 78.4% 80.3% Somewhat Important 1507 257 63 15 % of Tier 28.3% 22.4% 19.4% 18.5% Very Important 258 20 4 0 % of Tier 4.8% 1.7% 1.2% 0% No Response 61 13 3 1 % of Tier 1.1% 1.1% 0.9% 1.2% Total 5335 1149 324 81 Note: X2 (6, N=6,811) = 72.264, p<.001

PAGE 90

82 Table 4.16 – Importance of Counselor Influence by Tier of Institution Correlations for Independent Variables A review of the correlation coefficients of the variables indicates that there are numerous relationships that are significant at the 0.01 level. One w ill note that there are correlation coefficients as small as 0.031 that are marked as statistically significant. The large sample size used for this study (n=6,889) produced many statistically significant correlations that account for so little variance that they are of little practical use. Table 4.17 exhibits the correlation coefficients for all variables examined in the study. There were only two independent variable s that did not show a statistically significant relationship with the dependent variable of tier group, namely gender ( r =.003) and influence of a parent ( r =.029). Of all of the dependent variables, the institutional characteris tics of financial aid ( r =.304), rankings ( r =-.268), costs ( r =.266), Tier Counselor Influence1 2 3 4 Not Important 3394 839 222 68 % of Tier 63.6% 73.0% 68.5% 84.0% Somewhat Important 1524 255 85 9 % of Tier 28.6% 22.2% 26.2% 11.1% Very Important 350 40 12 3 % of Tier 6.7% 3.5% 3.7% 3.7% No Response 67 15 5 1 % of Tier 1.3% 1.3% 1.5% 1.2% Total 5335 1149 324 81 Note: X2 (6, N=6,801) = 58.195, p<.001

PAGE 91

83 and academic reputation ( r =-.234) were most strongly corr elated with tier group. These results indicate a significant relationship between student s who responded that college costs and financial aid awards were very im portant and their attendance at a lower-tiered (i.e. Tier Three or Four) uni versity. Conversely, the correlat ion coefficients indicate a significant relationship between students w ho responded that an institution’s academic reputation and placement in the rankings were very important and their attendance at a higher-tiered (i.e. Tier One or Two) university. Student and Family Characteristics Gender was found to have significant relati onships with three of the independent variables. The negative correlation ( r =-.104) implies a significan t relationship between being male and responding affirmatively of th e likelihood of having to work to pay for college. In addition, as it rela tes to their matriculation decision of high achieving students, a significant relati onship was found between being male and the importance placed on the influence of parents ( r =-.046) and the influence of teachers ( r =-.035). The results indicated several statistic ally significant relationships between parents’ education level and other independent variables. Father’s education level was found to be positively relate d to Asian ethnicity ( r =.103) but negatively related to Hispanic ethnicity ( r =-.074). In other words, there is a significant relationship between being Asian and having a father with a relativ ely high level of educat ion; and there is a significant relationship between being Hispanic and having a father with a relatively low level of education. No significant relations hip was found between Asian ethnicity and

PAGE 92

84 mother’s level of education, but, similar to the fathers, there was a negative relationship between Hispanic ethnicity a nd mother’s education level ( r =-.047). Not surprisingly, the educa tion level for both fathers and mothers showed strong positive correlations with family income (FatherEd=.389; MotherEd= .295), which validates the notion that hi gher education levels yield higher income levels. The correlation coefficient matrix further indicates other statistically significant relationships involving family income. In addition, family in come was found to be positively related to the importance of an institu tion’s academic reputation ( r =.067) and placement in media rankings ( r =.109), as well as with th e influence of parents ( r =.049) and relatives ( r =.049) There is a significant negative, albeit w eak, relationship between family income and the ethnic categories of Black ( r =-.045), Asian ( r =-.069), and Hispanic ( r =-.044), which indicates that identification with any of those three ethnic groups is negatively related to income. Family income was also f ound to be negatively re lated to the chances that students would have to work to pay for college ( r =-.303); and the importance of college costs ( r =-.264) and financial aid ( r =-.457) in the choice of where to enroll. No relationship of significance was found be tween family income and gender ( r =.013), ethnicity other than white, Black, Asian, or Hispanic ( r =-.010), or the influence of teachers ( r =-.011) or counselors ( r =.014).

PAGE 93

85 Table 4.17 Matrix of Co rrelation Coefficients Variable 1 2 3 4 5 6 7 8 9 1 Tier 1.00 2 Gender -.003 1.00 3 EthBlack -.038** -.016 1.00 4 EthAsian -.136** -.034** -.017 1.00 5 EthHisp -.047** -.005 .043** -.045** 1.00 6 EthOther -.047** -.003 .080** -.022 .033** 1.00 7 FatherEd -.204** -.003 -.008 .103** -.074** .029* 1.00 8 MotherEd -.164** -.009 .004 .014 -.047** .014 .460** 1.00 9 Income -.187** .013 -.045** -.069** -.044** -.010 .389** .295** 1.00 10 Employ .031** -.104** .018 -.006 .023 -.017 -.157** -.106** -.303** 11 Costs .266** -.020 .027* .012 -.003 -.009 -.146** -.115** -.264** 12 FinAid .304** -.008 .048** -.031** .041** .007 -.260** -.204** -.457** 13 AcadRep -.234** -.024* -.022 -.026* .005 -.011 .037** .030* .067** 14 Rankings -.268** -.005 .012 -.026* .020 .025* .069** .033** .109** 15 InfParent .029* -.046** -.005 .049** -.022 -.024 .049** .016 .049** 16 InfRel -.038** -.016 .014 .047** .002 .024* .056** .039** .049** 17 InfTeach -.099** -.035** .001 .041** .022 .021 -.017 -.017 -.011 18 InfCouns -.079** .019 .003 .022 .031* -.020 -.010 -.013 .014 Correlation is significant at the 0.05 level ** Correlation is significant at the 0.01 level

PAGE 94

86 Table 4.17 Matrix of Correla tion Coefficients (cont.) Variable 10 11 12 13 14 15 16 17 18 1 Tier 2 Gender 3 EthBlack 4 EthAsian 5 EthHisp 6 EthOther 7 FatherEd 8 MotherEd 9 Income 10 Employ 1.00 11 Costs .116** 1.00 12 FinAid .209** .493** 1.00 13 AcadRep .027* -.048** -.050** 1.00 14 Rankings -.028* -.038** -.088** .292** 1.00 15 InfParent .004 .044** .003 .058** .100** 1.00 16 InfRel -.043** .044** -.012 .066** .121** .316** 1.00 17 InfTeach .007 .054** .047** .078** .110** .125** .311** 1.00 18 InfCouns -.026* .100** .064** .060** .119** .118** .168** .459** 1.00 Correlation is significant at the 0.05 level ** Correlation is significant at the 0.01 level

PAGE 95

87 Institutional Characteristics The importance of cost of attending th e institution of choi ce was found to be positively correlated with the following variab les: the likelihood that the student will need to work to pay for school ( r =.116), and the influences of parents ( r =.044), relatives ( r =.044), teachers ( r =.054), and counselors ( r =.100). The importance of cost of attending the institution of choice was found to be negatively correlated with the following variables: the education level of fathers ( r =-.146), the education level of mothers ( r =.115), family income ( r =-.264), the importance of academic reputation ( r =-.048), and the importance of media rankings ( r =-.038). No statistically significant relationship was found between the costs of attendance and ethnicity. The directional relationshi ps for the importance of fi nancial aid are generally reflective of those relationships involving the importance of cost of attendance, as the relationship between the two responses was very positive ( r =.493). One exception to this observation has to do with the two variable s’ relationships with ethnicity. Although no statistically significant rela tionship was found between the costs of attendance and ethnicity, there were significant positive co rrelations reported between the importance of financial aid and the et hnic groups of Black ( r =.048) and Hispanic ( r =.041). However, there was a negative relations hip between the importance of financial aid and being Asian. Another exception to the similarity in correlation coefficients for costs and financial aid has to do with their relationshi p with the importance of the influences of others. There was no relationship between the importance of financial aid and the importance of the influences of parents ( r =.003) and relatives ( r =-.012), although

PAGE 96

88 significant positive relationships were found be tween those two variables and the cost of college. There was a significant positive correlation between the importance of academic reputation and the importan ce of media rankings ( r =.292). Therefore, as one might expect, the relationships between the two vari ables and other variable s in this study have many similarities. One exception to this has to do with the variable of Asian ethnicity. Although no statistically signifi cant relationship (at the p= .01 level) was found between being Asian and the importance of academ ic reputation, Asian ethnicity holds a significant positive relationship with the importance of media rankings ( r =.095). Except for the Asian ethnic group, no relationship was found between ethnicity and academic reputation or ethnicit y and media rankings. Influence of Others Male students were more likely than fema le students to indicate a high level of importance place on the influence of others in their decision of which college to attend. Although no significant relationshi ps were found for females, male students indicated a statistically significant level of importan ce placed on the influence of parents ( r =-.046) and the influence of teachers ( r =-.035). Except for the Asian ethnic group, no relationship was found between ethnicity and the influen ce of others. However, significant positive relationships were found between Asian iden tification and the influence of parents ( r =.049), the influence of relatives ( r =.047), and the influence of teachers ( r =.041). The importance of the influence of parents was found to be positively correlated

PAGE 97

89 with the following independent va riables: father’s education ( r =.049), family income ( r =.049), the importance of institutional costs ( r =.044), the importance of academic reputation ( r =.058), and the importance of media rankings ( r =.100). There was no relationship between the influence of parents and mother’s education ( r =.016) or the importance of financial aid ( r =.003). The importance of the influence of relatives was found to be positively correlated with the following independent va riables: father’s education ( r =.056), mother’s education ( r =.039), family income ( r =.049), the importance of institutional costs ( r =.044), the importance of academic reputation ( r =.066), and the importance of media rankings ( r =.121). There was a negative relationship betw een the influence of relatives and the likelihood of student employment ( r =-.043). Neither the variable of the influence of teachers nor the variable of the influence of counselors held a signifi cant relationship with the educ ation levels of fathers or mothers, family income, or the likelihood of student employment. The importance of the influence of teachers was found to be positively correlated with the following independent variables: the impor tance of institutional costs ( r =.054), the importance of financial aid ( r =.047), the importance of academic reputation ( r =.058), and the importance of media rankings ( r =.100). Similarly, the impor tance of the influence of counselors was found to be positively corr elated with the following independent variables: the importance of institutional costs ( r =.100), the importance of financial aid ( r =.064), the importance of academic reputation ( r =.060), and the importance of media rankings ( r =.119).

PAGE 98

90 Multiple Regression Analysis All 17 of the independent variables within the broader groups of student and family characteristics, institutional characteristics, and influence of others, were regressed on the dependent variable of institutional tier group. The results of the regression associating tier of first choice university for high achieving students from the predictor variables related to student and family characteristics, institutional characteristics, and the influence of others, is presented in Table 4.18. The sample for the regression analysis consisted of 6,889 high achieving students who scored in the top 10% of SAT test-takers, or the equivalent ACT result, and graduated from high school with an A average. Variables were included into the stepwise regression equation in order of the proportion of variance added by the variable. Betas for variables that are in the model, as well as those variables that did not enter the equation, were examined as each new variable entered the equation. Fourteen variables entered the equation, including ethnicity (4 variables), father’s education, mother’s edu cation, student employment, institutional costs, financial aid, academic reputation, media rankings, parental influence, teacher influence, and counselor influence. All fourteen variables in the model were significant predictors at the p<.001 level. Three variables did not enter the regression equation, namely gender, family income, and influence of a relative. The obtained R2 value was .245, suggesting that nearly 25% of the variability in tier level was accountable by the set of independent variables. The adjusted R2 value was .243. Cohen’s (1992) effect size was computed to be 32, which can be interpreted as a large effect using Cohen’s guidelines, where .02=small, .15=medium, and .35=large.

PAGE 99

91 Table 4.18 – Unstandardized Regres sion Coefficients for Models Unstandardized Regression Coefficient ( b ) at Step: Model Variable Constant ( ) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1 FinAid 1.110 .210 .195 .192 .140 .137 .119 .121 .121 .123 .119 .120 .121 .126 .127 2 Rankings 1.341 -.204 -.165 -.167 -.157 -.153 -.147 -.152 -.150 -.150 -.149 -.146 -.146 -.145 3 AcadRep 1.747 -.244 -.238 -.248 -.245 -. 238 -.241 -.242 -.241 -.243 -.243 -.239 -.241 4 Costs 1.700 .120 .124 .121 .124 .121 .119 .119 .119 .122 .123 .122 5 EthAsian 1.735 -.225 -.207 -.202 -.205 -.210 -.215 -.217 -.217 -.216 -.218 6 FatherEd 1.887 -.051 -.052 -. 053 -.055 -.042 -.042 -.042 -.044 -.043 7 InfTeach 1.896 -.091 -.099 -.097 -.097 -.097 -.072 -.071 -.070 8 InfParent 1.856 .056 .055 .055 .055 .057 .057 .057 9 EthHisp 1.858 -. 201 -.203 -.197 -.193 -.193 -.189 10 MotherEd 1.914 -.032 -.032 -.032 -.032 -.032 11 EthBlack 1.917 -.293 -.292 -.291 -.275 12 InfCouns 1.920 -.054 -.057 -.058 13 EthOther 1.957 -.034 -.034 14 Employ 1.961 -.179 Note: All statistics are si g nificant at p <.001

PAGE 100

92 The unstandardized regression coefficient ( b ) may be defined as the expected change in the dependent variable ( Y ) associated with a unit ch ange in the independent variable ( X ), as demonstrated in the following simple linear equation: Y = + b X + e where Y equals the raw score on the depende nt variable (i.e. tier level); equals the intercept, or constant; b equals the regression coefficient; X equals the raw score on the independent variable; and e equals the error, or residual. For the model in this study, the standard error was .531, which indi cates that predictions of tier level of institution tended to be off by about one half of a level. To get a further sense of the contributi on of each independent variable to the prediction of tier level of institution attended, standardized regression coefficients were calculated. If the scores for the dependent ( Y ) and independent ( X ) variables were standardized to z scores, one would use a standa rdized regression coefficient ( ). As in the case of b is interpreted as the expected change in Y associated with a unit change in X Further, a unit change in X when it has been standardize d, refers to a change of one standard deviation in X Standardized regression coeffici ents for each of the fourteen independent variables in the model are listed in Table 4.19. The regression coefficients for all fourteen variables in the model were found to be statistically significant at p<.001. As demonstrated in Table 4.19, the importance of financial aid uniquely accounted for the largest proportion of variability in the model. The R2 for this variable was .092, which indicates that 9% of the vari ability of institutiona l tier level can be attributed to the importance of financial aid. As the entire model of 14 variables had an

PAGE 101

93 Table 4.19 Standardized Regressi on Coefficients for Models Standardized Regression Coefficient ( ) at Step: Model Variable Model R2 R2 Change 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1 FinAid .092 .092 .304 .282 .278 .205 .199 172 .176 .175 .178 .173 .175 .176 .177 .183 2 Rankings .148 .056 -.239 -.189 -.192 -.179 -.175 -.167 -.172 -. 170 -.170 -.168 -.165 -.164 -.164 3 AcadRep .176 .028 -.174 -.169 -.175 -.173 -.169 -.170 -.171 -.170 -.172 -.171 -.172 -.170 4 Costs .193 .017 .150 .154 .151 154 .151 .149 .148 .149 .152 .152 .153 5 EthAsian .207 .014 -.119 -.109 -. 106 -.108 -.111 -.113 -.114 -.115 -.116 -.115 6 FatherEd .218 .011 -.111 -.113 -.116 -.120 -.095 -.095 -.095 -.093 -.097 7 InfTeach .225 .007 -.085 -.092 -.091 -.091 -.090 -.068 -.066 -.066 8 InfParent .229 .005 068 .067 .067 .067 .070 .069 .069 9 EthHisp .233 .004 -.062 -.063 -.061 -.060 -.058 -.058 10 MotherEd .236 .003 -.058 -.057 -.057 -.058 -.058 11 EthBlack .238 .003 -.051 -.051 -.048 -.047 12 InfCouns .240 .002 -.052 -.054 -.056 13 EthOther .241 .001 -.038 -.039 14 Employ .243 .001 -.040 Note: All statistics are significant at p<.001

PAGE 102

94 R2 of .243, the importance of financial aid uniquely accounted for over one-third of the variability for the entire model. Further, the four variables with the largest changes in R2 to the model make up nearly 80% of the total variance of the m odel in predicting tier level of institution. Specifically, over 19% of the variance ( R2=.193) in tier level of first choice institution was accountable by the four vari ables related to institutional characteristics, namely the importance of financial aid ( R2 change=.092), the importa nce of media rankings ( R2 change=.056), the importance of academic reputation ( R2 change=.028), and the importance of costs ( R2 change=.017). Ten additional vari ables were added to the model based on the statistical significance of th e F scores; however, those ten variables accounted for only 5% of the variance of the out come variable, namely tier level of first choice institution. Summary The scope of this study was limited to public and private national research universities as identified by USNWR The sample was drawn from the 97 universities which administered the survey CIRP Freshman Survey in 2004, as listed in Appendix C. To provide some initial unders tanding of the differences be tween high achieving students enrolled at each of the four tiers of national universities, re sults were reported in three ways, including (1) frequencies and descriptive statistics, (2) a correlation matrix, and (3) multiple regression models. As expected, due to the large proportion of participating Tier One institutions, there was a large disparity in the number of eligible respondents among tier groups. The

PAGE 103

95 resulting dataset for the study included re sponses from 6,889 students. Seventy-seven percent (n=5,335) of the responde nts were from Tier One inst itutions, compared to 16.7% (n=1,149) from Tier Two, 4.7% (n=324) from Ti er Three, and 1.2% (n=81) from Tier Four institutions. Although the gender representation for the sample resembled what the researcher expected, the ethic diversity of the sample in th e current study was less than optimal. The sample of respondents was overw helmingly comprised of white students (n=5571, 80.9%). All 17 of the independent variables with in the broader groups of student and family characteristics, institutional characteristics, and influence of others, were regressed on the dependent variable of institutional ti er group. Variables were included into the stepwise regression equation in order of the proportion of va riance added by the variable. Fourteen variables entered th e equation, including ethnicity (4 variables), father’s education, mother’s education, student employm ent, institutional costs, financial aid, academic reputation, media rankings, parental in fluence, teacher influence, and counselor influence. Three variables di d not enter the regression equa tion, namely gender, family income, and influence of a relative. The importance of financial aid accounted for the largest proportion of variability in the model. Further, over 19% of the variance ( R2=.193) in tier level of first choice institution was accountable by th e four variables related to institutional characteristics, namely the im portance of financial aid, the importance of media rankings, the importance of academic reputation, and the importance of costs. Further analyses and discussion of th e results are included in Chapter 5.

PAGE 104

96 CHAPTER 5 DISCUSSION AND CONCLUSIONS This chapter discusses the major findings of this multivariate research study. Particular attention is given to the interpre tation of research ques tions posed for the study and how the results relate to the colleg e choice models examined in Chapter 2. Implications for future educational research and implications for policy and practice will be discussed. The chapter will close with concluding remarks from the researcher. The first sections of this chapter address the results of the study as they relate to the review of the literature and specifically how they address the re search questions that have guided this study: 1. To what extent do students’ individual characteristics (e.g. gender and ethnicity) relate to college choice for high achieving students? 2. To what extent do students’ family char acteristics (e.g. parent s’ education level and family income) relate to colleg e choice for high achieving students? 3. To what extent do financial considerations associated with college (e.g. cost and financial aid) relate to college choice for high ach ieving students? 4. To what extent does academic reputati on of the institution relate to college choice for high achieving students? 5. To what extent does the influence of si gnificant others (e.g. parents, relatives, teachers, and counselors) relate to co llege choice for high achieving students?

PAGE 105

97 Relationship between Individual and Fam ily Characteristics and College Choice The first research objective was to e xplore the relationship between students’ individual characteristics (e.g. gender and ethnicity) and college choice for high achieving students, as follows: Research Question #1: To what extent do students’ individual characteristics (e.g. gender and ethnicity) relate to college choice for high achieving students? When considering the relationship betw een the individual ch aracteristics of students and college choice, the literature suppo rts the notion that r ace, gender and social class have a strong relationship with edu cational attainment. Although gender was not identified as a statistically significant fact or for the regression model, differences in gender representation were observed in the frequency distribution (see Table 4.2). The greatest differences observed were in Tier Two institutions, with 25% more males (n=638) than females (n=511), and in Tier Four institutions, with 31% more females (n=46) than males (n=35). The results of frequency distributi on are in line with McDonough’s (1997) findings that, regardless of academic ability and achievements, women are less likely to attend highly selective institutions. Prior research suggests that African-Ame rican students tend to enroll in less selective institution and that Hispanic students have demonstrated lower educational aspirations than African-American students (Mau, 1995). In addition, African-American and Hispanic students have been found to be mo re sensitive than thei r white peers to the costs of higher education, and, therefore, ar e more responsive to grants and scholarships (Johnson, Stewart & Eberly, 1991; Hoyt & Brown, 2003). With the exception of Hispanic

PAGE 106

98 females, the literature indicates that female s have stronger academic goals than males; although Asian American males have been found to possess signifi cantly higher college aspirations than females and all other ethnic male groups (Mau, 1995). The strikingly low representation of Black and Hispanic students in the sample for the current study made it difficult to draw strong relationships between a student’s identification as Black or Hispanic and atte ndance at a selectiv e institution. Frequency distributions displayed in Table 4.3 show lo w representation of students from these two ethnic groups across all four tiers, with no Bl ack students represented in the sample for Tier Four institutions. However, both factors were statistically significant when regressed against the outcome variable of tier institution and, therefore, were included in the final regression model. Compared with the lower-tiered instituti ons the Tier One institutions received responses from a noticeably higher percentage of Asian students. N early 15% (n=797) of the respondents from Tier One institutions id entified themselves as Asian, compared to 3.2% (n=37) from Tier Two, 3.1% (n=10) from Tier Three, and 2.5% (n=2) from Tier Four institutions. Except for the Asian et hnic group, no relationship was found between ethnicity and the influence of others. However, significant positive relationships were found between Asian identification and the influence of parents ( r =.049), the influence of relatives ( r =.047), and the influence of teachers ( r =.041). The second research question related to the relationship between family characteristics and college choice fo r high achieving students, as follows:

PAGE 107

99 Research Question #2: To what extent do students’ family characteristics (e.g. parents’ educati on level and family income) relate to college choice for hi gh achieving students? Socioeconomic status is a common factor that researchers have identified to segment students in college choice studies. A number of research studies have supported the premise that students from high socio economic backgrounds and students who are academically talented are more likely to at tend elite institutions (Brewer, et al., 1999; Hearn, 1987; Manski & Wise, 1983; Paulsen, 1990) and that low-income and firstgeneration students are compara tively disadvantaged against their more affluent peers when it comes to the variety of colleges from which they are able to choose (Kinzie, et al., 2004). Consistent with other college choice re search findings, the frequency distribution for the current study indicates an inverse rela tionship between income and tier level (see Table 4.4). That is, as the increments of fam ily income increase, the tier level decreases. Students attending Tier F our institutions were twice as likely to report a family income of $100,000 or less than were students attending Ti er One institutions; with more than one in five (22.2%) Tier Four students repor ting a family income of $50,000 or less, compared with only one of every ten (10.3%) stud ents at Tier One institutions. There is a similar disparity when examining the othe r end of the income spectrum. Students attending Tier One institutions reported a fa mily income of over $150,000 at nearly twice the rate (33.5%) of Tier Two students (17.2%), three times the rate of Tier Three students (11.4%), and nearly seven times the ra te of Tier Four students (4.9%).

PAGE 108

100 Previous research has found that firstgeneration students tend to receive less encouragement and support from their families than multi-generation students when it comes to college attendance (Arredondo, 1999) Students appear to have a higher likelihood of viewing college as realistic wh en their parents stre ss the importance of educational success (Ceja, 2004). Research fi ndings are inconsiste nt when reporting on behaviors of first-generation students in the college applica tion process. McDonough (1994) reported that, compared with students who are raised by college graduates, firstgeneration students are more likely to limit the number of institutions to which they apply and to apply to nonselective institutions. Ho wever, a study of college-bound high school students in New Hampshire reveal ed no significant differences in the type or quality of college under consideration between student s whose parents possessed postsecondary degrees and those whose pare nts had not completed a coll ege education (Toutkoushian, 2001). In fact, first generation students were found to be equally likely as those with college-educated parents to consid er attending a se lective school. As expected from the sample data for the current study, students attending Tier One institutions reported the hi ghest levels of education for their parents, with over half of respondents (55.3%) reporting that their fathers possessed a graduate degree and nearly 38% reporting graduate degree atta inment for their mothers (see Table 4.7). Further, the fathers of Tier One students were least likely to lack any college experience. Less than 6% of fathers of students attendi ng Tier One institutions lacked a college education; however, the percentage rises to mo re than 11% for fathers of students in Tiers 2, 3, and 4. Contrary to the results for the fath ers, the mothers of Tier Four students were

PAGE 109

101 least likely to lack a college education. For bot h fathers and mothers, Tier Three students had the highest percentage of pa rents with no college education. An examination of the correlation coefficient matrix (see Table 4.17) indicates several statistically significant relationships between parents’ education level and other dependent variables. Not surprisingly, the education level for both fathers ( r =.389) and mothers ( r =.295) showed strong positive correlations with family income, which supports the notion that higher educati on levels yield higher income levels. Somewhat unexpected, however, was that the influence of parents in choosing a college was found to have a significant positive correlation with father’s education level ( r =.049) but not with mother’s education level ( r =.016). Relationship between Institutional Characteristics and College Choice The third research objective was to expl ore the relationship between the cost of college and the importance of financial aid awards and college choice for high achieving students, as follows: Research Question #3: To what ext ent do financial considerations associated with college (e.g. cost and financial aid) relate to college choice for high achi eving students? The financial realities of a college education are likely to influence a student’s choice of where to attend college; and much of the existing research supports the notion that students consider the trad e-offs between current costs and future expectations of financial and non-financial benefits. As a st rategy to recruit gr eater numbers of high achieving students, institutions may increase le vels of educational spending per student.

PAGE 110

102 This is the case particularly at private institutions that can more easily raise tuition to address financial needs (Hoxby, 1997). Understa nding that tuition increases may result in deterring the students they are trying to attract, many institutions accompany tuition increases with increased allocations for both need-based and merit-based financial aid. For high-ability students, assessing the best combination of multiple offers of financial assistance can be a daunting task, as they may qualify for both need-based and meritbased aid, both state-funded and privatel y-funded scholarships, federal work-study programs, and aid packages from each of the colleges in which they are interested. The current study investigated the extent th at the costs of college and offers of financial aid influenced the tier level of attendance for high achieving students. Regardless of tier group, the majority of students in the sample among all tier groups indicated that there was at least some chance th at they would need to seek employment to help pay for college expenses, with students in Tier Four institutions indicating a stronger need than the students in ot her tier groups (see Table 4.4). No t surprisingly, there is a significant positive correlation between the need for student employment to pay for college and both the importance of college costs and the importance of financial aid. Only three (3.7%) of the 81 students enrolled at Tier Four institutions responded that there was little to no chance that they would n eed to get a job to pa y for their education. Compared with the other independent va riables explored in this study, costs ( r =.266) and financial aid ( r =.304) were strongly correlated with tier group, both in a positive direction. These results would tend to support the claim that students who responded that college costs and financial aid awards were very important were likely to attend a lower-tiered (i.e. Tier Three or Four) university. As demonstrated in Table 4.19,

PAGE 111

103 the variable of importance of financial aid uniquely accounted for the largest proportion of variability in the model. Specifically, 9% of the variability of institutional tier level can be attributed to the importance of financia l aid. As the entire mode l of 14 variables had an R2 of .243, the importance of financial aid uni quely accounted for over one-third of the variability for the entire model. The variable of importance of college costs accounted for 1.7% of the variability in predicting tier group. This study affirms the results of previous studies on college choice, but fills a gap in understanding the matriculation decisions of high achieving college-bound students. There is some agreement among scholars that, wh ile the availability of financial aid is considered important by most college-bound studen ts, the impact of co st and financial aid decrease as students’ income level and academic ability increase and that this financial gap often discourages or prohibits low-in come students from a ttending higher-tiered institutions, even when controlling for acad emic ability. The current study examined the responses only of students with high levels of academic ability and found the availability of financial aid to be the single most importa nt factor in predicting whether students will attend a higher-tiered or lo wer-tiered university. The im portance of financial aid accounted for over five times the variability of the importance of college costs. Therefore, although college costs were found to be a significan t predictor of the tier level of university attended, it was of secondary importance compared with the attention to financial aid by high achieving students. Students were more likely to view financ ial aid awards as a key matriculation factor if they were Black or Hispanic, had parents who possess relati vely low levels of postsecondary education, and came from a relatively lower income family. These

PAGE 112

104 students were also likely to respond that the influence of a teacher or counselor was important in their choice of college. On the ot her hand, they were less likely to belong to an Asian ethnic group or to view the acad emic reputation or media rankings for the school as important in th eir decision to attend. The fourth research question went to the heart of the study to explore the importance of academic reputation to high achieving students and whether differences exist among students who attend higher-tiere d versus lower-tiered universities, as follows: Research Question #4: To what exten t does academic reputation of the institution relate to college c hoice for high achieving students? Access to college and un iversity information through mass media has had a noticeable impact on the manner in which a pplication and admissions processes are approached. Not only are instit utions concerned about the nu mber of students they can enroll, but they are particularly interest ed in high achieving students due to the enhancements that these students can c ontribute to an inst itution’s reputation. Recruitment of the best and brightest student s is critical for positive development of an institution’s academic reputation. Moreover, un iversities pay attenti on to their placement in the rankings because rankings and prestige are important to academically attractive students who want to attend prestigious in stitutions. There is some consensus among researchers that institutional prestige and academic reputation are of primary importance to high ability students when choosing a colleg e. However, the literature in this area offers little guidance to enrollment management professionals at lower-tiered universities

PAGE 113

105 as to the factors that persuade some hi gh achieving students to attend lower-tiered universities. The results of this study support the existing research that asserts that academic reputation is an important factor of ma triculation for high ach ieving students. The majority of students in the sample, in all tier groups, indicated that a cademic reputation of the students’ college of choice was a very im portant factor in their decision to attend, although there were noticeable differences between tier groups (see Table 4.11). Students enrolled at the higher-tiered institutions we re most likely to rate academic reputation as very important, with 86.9% of Tier On e students and 73.4% of Tier Two students responding accordingly. The proportion of stude nts who considered academic reputation as very important then drops to 53.7% of Tier Three students and 55.6% of students attending a Tier Four institution. Virtually none (less than 1%) of the students at Tier One universities responded that academic reputation was not at all important in their college choice decision. In addition to gathering information re garding academic reputation, the study also collected student responses to the importanc e of media rankings in their matriculation decisions. Although research exists on the im portance of media rank ings, little is known about the population of students which most heavily value such indices. Researchers that have studied the influence of media ranki ngs on matriculation have concluded that students are most likely to find them important if they are of traditional college age, from middle income families, and are planning to attend a school outside of their region (Goenner and Snaith, 2004; Hossler and Foley, 1995).

PAGE 114

106 Four out of five students attending Tier One institu tions responded that media rankings were at least somewhat important in their decision to enroll at the particular university (see Table 4.12). Stude nts in the lower-tiered groups indicated the least interest and placement of importance on the media’s ra nking of postsecondary institutions, with two-thirds (66.3%) of students attending Tier Four institutions i ndicating that these rankings were not at all important. Simila rly, according to the correlation coefficient matrix, there was a significant relationship between students who considered placement in the rankings to be very important and st udents who attended higher-tiered (i.e. Tier One or Two) universities. Also, it is wort h noting that a signif icant relationship was revealed between Asian students and th e importance of rankings. No significant relationship was identified for any other ethnic group. Relationship of the Influence of Others to College Choice Following student and family characteristics and institutional characteristics, the influence of others was explored as a factor aff ecting the college choices of high achieving students. There are four groups of “oth ers” that were investigated in the current study, including parents, relativ es, teachers, and counselors. Research Question #5: To what exten t does the influence of significant others (e.g. parents, relatives teachers, and counselors) relate to college choice for high achi eving students? The choice of where to go to college is arguably one of the most important decisions of a young adult’s life. For high sc hool students consideri ng a college career, guidance from trusted loved ones and respected role models is needed to think through all

PAGE 115

107 of the considerations. Prior studies have concluded that parental encouragement and expectations influence college aspirations in students, regardless of gender, ethnicity and socioeconomic backgrounds (Hossler and Stag e, 1992), that parental influence is a significant predictor of student matriculation, and that st udents who attend prestigious universities are more likely to receive motivational messages from parents than from counselors, peers and other educational ro le models (Levine and Nidiffer, 1996). In addition to the strong influence from parents and relatives, some scholars have found that a number of students consider high school couns elors and teachers to be an important source of information (Bradshaw, et al., 2001; Gonzalez, et al., 2003) particularly for students from lower SES backgrounds and w hose parents had little formal education (MacAllum, et al., 2007). Compared with the outcomes related to student and family characteristics and institutional characteristics the frequency distribution for th e current study reflects little variation among student responses in each of the four tier groups related to the importance of the influence of others The frequency of students who indicated that the influence of their parents was very important in their college decision ranged from 31% to 40% among tier groups. No more than 6% of students within a ny of the tier groups indicated that the influence of relatives was very important, while approximately twothirds of students within any given tier gr oup responded that the influence of relatives was not a factor that influen ce their choice of which colleg e to attend. The influence of teachers and counselors, according to the hi gh achieving students in the sample, the influence of relatives seems to be minor (see Tables 4.15 and 4.16).

PAGE 116

108 When the four variables were regressed ag ainst the outcome vari able of tier level, three of the variables were f ound to have a significant rela tionship and were therefore included in the model. The variables of influence of teachers ( R2=.007), influence of parents ( R2=.005), and influence of counselors ( R2=.002), collectively accounted for 1.4% of the variability in the model for pred icting tier level of a ttended school for high achieving students. The variable of influence of relatives was not included in the model. A review of the standardized regression coeffi cients for the three variables in the model reflects inconsistencies in the direction with which the variables influence the outcome of institutional tier level. There is a negative relationship between the influence of teachers ( = -.066) and tier le vel and, similarly, a negative relationship between the influence of counselors ( = -.056) and institutional tier group. Conversely, the influence of parents ( = .069) has a positive relationship w ith institutional tier level. These results indicate that those students who were most influenced by teachers and counselors tended to enroll at a higher-tier ed university. This finding is inconsistent with previous research that linked the infl uence of teachers and counselors with students of low SES backgrounds and attendance at lowe r-tiered universities For students whose parents have had little or no experien ce with postsecondary education, it is understandable that teachers and counselors would become a repl acement advocate and role model for higher education. In additi on, these professionals may help students navigate through the admission and enrollment process if parents lack the ability or willingness to take on those responsibilities. Table 5.1 provides a summary of the resu lts of the study, indicating the factors that were determined to possess significant relationships with the outcome variable for

PAGE 117

109 tier level. As observed from the table, the results from this study support prior research that found relationships between high ach ieving students and certain individual characteristics, family characteristics, inst itutional characteristics, and the influences of others. The purpose of this study was to iden tify differences among the students attending the higher-tiered universities and their peers who chose to attend a lower-tiered university. The strongest predicto r of enrollment at a lowertiered university was whether the student considered th e availability of financial aid to be very important in choosing a college. The importance of financial aid wa s followed by the importance of costs of college. The final predictor was the influence of a parent. Table 5.1 – Summary of Relationships betw een Independent and Dependent Variables Variable Higher-Tiered Lower-Tiered Student/Family Characteristics Gender Ethnicity Asian Parents Education High Low Income High Low Institutional Characteristics Costs Very important Financial Aid Very important Academic Reputation Very important Media Rankings Very important Influence of Others Parents Important Relatives Teachers Important Counselors Important

PAGE 118

110 Validation of the College Choice Model The three-stage choice model developed by Hossler and Gallagher (1987) was the basis for this study. Hossler, et al. (1989) defined the college choice experience as a “complex, multi-stage process during which an individual develops aspirations to continue formal education beyond high school followed later by a decision to attend a specific college, university or institution of advanced vocational training” (p. 234). Hossler and Gallagher’s (1987) model outlines th ree stages of the college choice process: 1. Predisposition: students’ decisions/aspirations to enroll in postsecondary education. 2. Search: the process of cons idering types of institutions to which to apply. 3. Choice: the selection of an institution to attend. The first stage of predisposition is defined as the phase in which students decide whether or not to pursue formal education af ter high school. Factors that have been found to predispose students toward college include socioeconomic status, students’ academic achievement, parents’ education levels, et hnicity, gender, encouragement from high school counselors and teachers, and parental expectations and encouragement (Hossler & Stage, 1992). During the search stage, st udents engage in ac cessing information on specific colleges in order to further examin e the opportunities and benefits. It is within this phase that students are most likely to c onsider external and institutional information sources. Factors that may be considered by st udents at this second pha se include cost of attendance, availability and offers of fina ncial assistance, and academic reputation. The third stage of college choice is the application of the predisposition factors combined with the information gathered during the search phase (Hossler & Gallagher, 1987).

PAGE 119

111 This study examined how differences am ong high achieving students in each of the first two stages may impact the level of prestige, measured by the USNWR -assigned tier, of the college of first choice. The de sign of the study included predisposition-related factors of student and family characteristics (gender, ethnicit y, parents’ education levels, and family income) and the influence of others (parents, relatives, teachers, and counselors). Search-related fact ors considered for this study were grouped as institutional characteristics (costs, financial aid, and academic reputation). Results of the current st udy indicate that for high achieving students the second stage in the model tends to be a better predictor than the first stage in predicting the outcome of college choice. The four variables associated with institutional characteristics were found, through both correlation and regr ession analyses, to be more significant predictors of college choice than any of the ot her variables that were included as part of this study. Specifically, over 19% of the variance ( R2=.193) in tier level of first choice institution was accountable by the four variable s related to institutional characteristics, namely the importance of financial aid ( R2 change=.092), the importance of media rankings ( R2 change=.056), the importan ce of academic reputation ( R2 change=.028), and the importance of costs ( R2 change=.017). This finding is significant as Hossler and Gallagher’s (1987) model has not been appl ied specifically to the matriculation phenomenon for high achieving students in terested in attending national research universities.

PAGE 120

112 Implications for Future Research Prior to this study, the specifi c scope of factors related to the tier level of college of first choice for high achieving studen ts were unknown and unmeasured. The CIRP 2004 Freshman Survey, the survey instrument used for this study, provided the source of secondary data to address factors related to student and family characteristics, institutional characteristics, and the influence of others. Wh ile the factors selected for inclusion in this study were grounded in the lit erature, there are othe r factors that likely contribute to the outcome of college choice th at were not possible to include, as not all factors of interest were captured by the CIRP. This researcher does not necessarily r ecommend that changes be made to the Freshman Survey to include an endless arra y of college choice factors. Rather, it is suggested that future research related to the relationship be tween the college choices of high achieving students and st udent and family characteristics, institutional characteristics, and the influence of others, not rely solely on the data which can be provided by the Freshman Survey to answ er these research questions. The CIRP Freshman Survey has been an effective tool for providing useful information for researchers interested in th e factors related to matricul ation; however, the 2004 survey did not provide any way to capture data rela ted to the importance of college athletic programs or the influence of peers. The litera ture suggests that thes e two factors, among others, may assist in the explanation of the relationship between college choice and institutional characteristics and the influence of others. The CIRP Freshman Survey gathers so me intriguing information that was not related to the scope of the present study. Furt her investigation into some of these factors

PAGE 121

113 is recommended. A complete copy of the 2004 Freshman Survey may be found in Appendix A. Some examples of survey questions that may lend themselves to future research include: 1. What is the highest academic de gree that you intend to obtain? 2. For the activities below, indicate which ones you di d during the past year. (Response choices include but are not limite d to: attended a religious service; was bored in class; participated in organized demonstrations; smoked cigarettes; drank beer; felt overwhelmed by all I had to do; felt depressed; and performed volunteer work.) 3. Rate yourself on each of the following tra its as compared with the average person your age. (Response choices include but are not limited to: academic ability; artistic ability; compassion; courage; drive to achieve; generosity; and time management). 4. During your last year in high school, how much time did you spend during a typical week doing the follo wing activities? (Response c hoices include but are not limited to: studying/homework; socializing with friends ; talking with teachers outside of class; exercise or sports; and partying.) 5. Please indicate the importance to yo u personally of each of the following. (Response choices include but are not lim ited to: becoming an authority in my field; influencing the political structure; raising a family; being very well off financially; helping to promote racial unde rstanding; and working to find a cure to a health problem.)

PAGE 122

114 The questions summarized above hold potential to address research questions which are related to other theoretica l constructs outside the sc ope of this study. However, researchers who are interested in other aspe cts of college choice behaviors may find the CIRP data useful. Future research should build upon the inve stigation of the fact ors that influence specific institutional enrollment decisions of academically talented students. One possibility is to explore behavioral and persona lity characteristics of these bright students in relation to their choice of college. It woul d also be interesting to build a study that investigates self-perceptions of high achievi ng students, and how those self-perceptions impact their matriculation decisions. Implications for Practice The phenomenon of choosing a college con tinues to attract the attention of scholars. Consequently, the results of college choice studies are of pa rticular interest to college and university administrators tasked with shaping the profile of their entering freshman classes. There is pressure on public institutions in particular to maintain broad access policies; but these pressures are ofte n in conflict with some colleges’ and universities’ desires to rest rict access to high-achieving st udents in order to improve academic reputation and rankings. Because of th e attention given to academic reputation, the recruitment of high achieving students continues to be a ch allenge for national universities that consistently find themselves in the third or fourth tier according to the rankings of USNWR’s annual edition of Best Colleges

PAGE 123

115 The existing literature on the subject of college choice and high achieving students demonstrates that high achieving students differ from the general student population as far as the manner in which they approach the college choice process and the factors that are most important to them (Bradshaw, Espinosa & Hausman, 2001). There is also some agreement w ithin the literature on college choice that the criterion that typically grabs the top spot is college qua lity (Chapman & Jackson, 1987). The literature has been limited in providing a broad and comprehensive understanding of the college choice decisions of high-ability students who choose to attend lowe r-tiered institutions. The present study addressed thes e gaps within the literature. The resu lts of this study should be of particular interest to lower-tiered universities. Because student selectivity is one of the few indicators considered among the ranking criteria over which institutions have so me amount of control, the universities that have made prestige a priority have made st rategic changes to their admissions criteria. Some colleges and universities have adju sted and improved their recruitment and enrollment procedures by incorporating stra tegies related to financial aid and early admission. Public universities, which historic ally have a reputation for access and open admission, are now turning away a larger and la rger proportion of thei r applicants in the name of increased quality. The results of this study should bring encouragement to enrollment management professionals at lower-tiered uni versities, as the strongest pred ictors of college choice are factors within the control of the institutions The four institutional variables of financial aid awards, media rankings, academic reputati on, and college costs, were respectively found to account for the strongest levels of variability within th e regression model.

PAGE 124

116 Although ten other variables associated with student characteristics, family characteristics, and the influence of others were significant enough to enter the model, the four variables associated with institutional characteristics rose to the top. The implication of this finding is that colleges and universities may attract more high achieving students if they can offer attractive financial aid p ackages and keep the costs of attendance competitive with other national research universities. A final strategy would be to increase outreach efforts to high school counsel ors and teachers, as they may serve as an advocate for the institution. Limitations Although the results of this study have sh ed some light on the differences among high achieving students who choose to attend coll eges categorized in various tier levels, there are some limitations of the study that should be acknowle dged when interpreting the data and drawing conclusions with the findings. First, the disparity of the number of cases per tier group limits the extent to which conclusions can be drawn. As expected, due to the large proportion of participating Tier One institutions in the 2004 Freshman Surve y, there was a large disparity in the number of eligible respondents among tier groups. In addition, the proportion of students who met the standardized test score criteria notic eably decreased with ea ch change in tier group. The resulting dataset for the study included responses from 6,889 students. Seventy-seven percent (n=5,335) of the respond ents were from Tier One institutions, compared to 16.7% (n=1,149) from Tier Two, 4.7% (n=324) from Tier Three, and 1.2% (n=81) from Tier Four institutions. With the c onstraints that were placed on the eligible

PAGE 125

117 sample, such as distance from home and atte ndance at the students’ first choice school, it is not surprising to yield 81 cases from th e 14 participating Tier Four institutions. A second limitation to the findings of this study is the amount of missing data for some of the independent variables, particularly related to the variable of family income. Relative to the response rate ob tained from the sample for other independent variables, the response rate from students regarding fa mily income was noticeably low. Nearly one of ten (n=656, 9.5%) students in the sample failed to respond to th e question in the 2004 Freshman Survey regarding family income. The relatively large number of missing data may limit analyses and conclusions regarding family income and its relationship with the tier level of univers ity that a student chooses to attend. Conclusion Studies that have investig ated college choice factors for high-achieving students repeatedly cite academic reputation as one of the top indicators of choice. These results fail to provide an indication as to why so me high-achieving stude nts choose to attend universities with a less presti gious reputation than the more highly prestigious options available to them. An explora tion of the factors related to the individual characteristics and institutional preferences of high ability students w ho choose to enroll in a nonselective university is not only an interest ing research question but also an issue of relevance to state policymakers and college ad ministrators. The pres ent study adds to the body of literature related to college choi ce by exploring differences between high achieving students who attend higher-tiered uni versities and high ac hieving students who attend lower-tiered universities.

PAGE 126

118 The three-stage choice model developed by Hossler and Gallagher (1987) was the basis for this study. Hossler and Gallagher’s (1987) model outlines three stages of the college choice process: 1. Predisposition: students’ decisions/aspirations to enroll in postsecondary education. 2. Search: the process of cons idering types of institutions to which to apply. 3. Choice: the selection of an institution to attend. Results of the current study indicate that for high achievi ng students the second stage in the model has more influence than the first stage in predicting the outcome of college choice. The four variables associated w ith institutional char acteristics were found, through both correlation and regression analyses to be more significant predictors of college choice than any of the other variables that were incl uded as part of this study. This finding is significant as Hossler and Gallagher’s (1987) model has not been applied specifically to the matriculation phenomenon for high achieving students interested in attending national res earch universities. The results of this study support the existing research that asserts that academic reputation is an important factor of ma triculation for high ach ieving students. The majority of students in the sample, in all tier groups, indicated that ac ademic reputation of the students’ college of choice was a very impor tant factor in their decision to attend; however, students enrolled at the higher-tiered institutions most frequently indicated that academic reputation was very important. The re gression analysis confirmed a significant relationship between a student’s attitude to ward the importance of academic reputation and the tier level of his first choice college. Specifically, the results of the study indicated

PAGE 127

119 a significant relationship between students who considered academic reputation to be very important and students who attende d higher-tiered (i.e. Tier One or Two) universities. In addition to gathering information re garding academic reputation, the study also collected student responses regarding th e importance of media rankings in their matriculation decisions. Research ers that have studied the in fluence of media rankings on matriculation have concluded that students ar e most likely to find them important if they are of traditional college age, from middle in come families, and are planning to attend a school outside of their region (Goenner a nd Snaith, 2004; Hossler and Foley, 1995). The results of the study indicated a significant re lationship between students who considered placement in the rankings to be very impor tant and students who attended higher-tiered (i.e. Tier One or Tw o) universities. This study found the availability of financial aid to be th e most important factor in predicting whether students will attend a higher-tiered or lower-tiered university. Students who consider the availa bility of financial aid to be very important tend to attend lower-tiered universities. The importance of financial aid accounted for over five times the variability of the importance of college costs. Therefore, a lthough college costs and academic reputation were found to be significant predictors of the tie r level of university attended, they were of secondary importance co mpared with the attention to financial aid awards by high achieving students. This study affirms the results of previous studies on college choice, but fills a gap in understanding the matriculation decisions of high achieving college-bound students. There is some agreement among scholars that, wh ile the availability of financial aid is

PAGE 128

120 considered important by most college-bound studen ts, the impact of co st and financial aid decrease as students’ income level and academic ability increase and that this financial gap often discourages or prohibits low-in come students from a ttending higher-tiered institutions, even when cont rolling for academic ability.

PAGE 129

121 REFERENCES The ACT. ACT-SAT Concordance Website retrieved December 4, 2008, from http://www.act.org/aap/concordance/ Arredondo, P. (1999). Multicultura l counseling competencies as tools to address oppression and racism. Journal of Counseling & Development 77(1), 102. Association of Ameri can Universities. AAU Members Website retrieved August 4, 2008, from http://www.aau.edu/about/default.aspx?id=4020 Astin, A. W. (2005-2006). Making sens e out of degree completion rates. Journal of College Student Retention: Research, Theory & Practice 7(1-2), 5-17. Astin, A. W., & Lee, J. J. (2003). How risky are one-shot cross-sec tional assessments of undergraduate students? Research in Higher Education 44(6), 657-672. Avery, C., & Hoxby, C. (2003). Do and should financial aid pa ckages affect students’ college choices? (Working Paper No. 9482) NBER. Becker, G.S. (1964). Human capital: A theoretical and em pirical analysis, with special reference to education. New York: National Bureau of Economic Research. Becker, G.S. (1971). Economic theory New York: A. Knopf.

PAGE 130

122 Behrman, J., Kletzer, L., McPherson, M., & Schapiro, M. O. (1998). Microeconomics of college choice, careers, and wages. The Annals of the American Academy of Political and Social Sciences, 559 (1), 12-23. Best Colleges 2009. U.S. News & World Report web site. Accessed September 21, 2008, from http://colleges.usnews.rankingsa ndreviews.com/college/national Bradshaw, G., Espinoza, S., & Hausman, C. (2001). The college decision-making of high achieving students. College and University, 77 (2), 15. Brasier, T. G. (2008). The effects of parental involvem ent on students' eighth and tenth grade college aspirations: A comparative analysis. Unpublished Dissertation North Carolina State University. Braunstein, A., McGrath, M., & Pescatrice, D. (1999). Measuring the impact of income and financial aid offers on college enrollment decisions. Research in Higher Education, 40 (3), 247-259. Braxton, J. M. (1990). How st udents choose colleges. In D. Hossler, & Bean, J. & Associates (Eds.), The strategic management of college enrollments (pp. 57-67). San Francisco, CA: Jossey-Bass.

PAGE 131

123 Brewer, D. J., Eide, E. R., & Ehrenberg, R. G. (1999). Does it pay to attend an elite private college? Cross-cohor t evidence on the effects of college type on earnings. The Journal of Human Resources 34(1), 104-123. Briggs, S. (2006). An exploratory study of the factors influencing undergraduate student choice: The case of higher education in Scotland. Studies in Higher Education, 31 (6), 705-722. Brooks, R. C. (2006). Factors that influen ce traditional-age, high-achieving students to enroll at a research-extensive university in the Southern region of the United States. (Doctor of Philosophy, Louisiana State University). Cabrera, A.F. & LaNasa, S.M. (2000). Over coming the tasks on the path to college for America’s disadvantaged. New directi ons for institutional research, 107, 31-43. Cabrera, A. F. (1994). Logist ic regression analysis in hi gher education: An applied perspective. In J. C. Smart (Ed.), Higher Education: Handbook of Theory and Research (pp. 225-256). New York, NY: Agathon Press. Ceja, M. (2004). Chicano college aspirati ons and the role of parents: developing educational resiliency. Journal of Hispanic Higher Education 3(4), 338-362.

PAGE 132

124 Chapman, R. G., & Jackson, R. (1987). College choices of acade mically able students: The influence of no-need financial aid and other factors. Research Monograph No. 10. New York, NY: College Board Publications. Chenoweth, E., & Galliher, R.V. (2004, Oct ober 15). Factors influencing college aspirations of rural West Virginia high school students. Journal of Research in Rural Education, 19(2). Retrieved August 16, 2008, from http://www.jrre.psu.e du/articles/19-2.pdf Clark, C. R., & Hossler, D. (1990). Marketi ng in nonprofit organizations. In D. Hossler, & Bean, J. & Associates (Eds.), The Strategic Management of College Enrollments (pp. 68-85). San Francisc o, CA: Jossey-Bass. Clarke, M. (2002). Quantifying quality: What can the U.S. news and world report rankings tell us about the qua lity of higher education? Educational Policy Analysis Archives, 10 (16). Cohen, J. (1992). A power primer. Psychological Bulletin 112, 155-159. College Board. SAT percentile ranks. Web site retrieved March 29, 2008, from http://www.collegeboard.com/prod_downloads /highered/ra/sat/S AT_percentile_ran ks.pdf

PAGE 133

125 Dale, S. & Krueger, A. (1999). Estimating the payoff to attending a more selective college: An application of selection on observables and unobservables. NBER Working Paper 7322. DesJardins, S. L., Ahlburg, D. A., & McCall, B. P. (2006). An integrated model of application, admission, enrollment, and financial aid. Journal of Higher Education, 77 (3), 381-429. DesJardins, S. L., Dundar, H., & Hendel, D. D. (1999). Modeling the college application decision process in a la nd-grant university. Economics of Education Review, 18 (1), 117-132. Douglass, D. (1997). Economic returns on inve stments in higher education. In H. Bowen (Ed.), Investment in learning: The individual and social value of American higher education (pp. 359-379). Baltimore, MD: Johns Hopkins University Press. Ehrenberg, R. G., & Sherman, D. R. (1984). Op timal financial aid policies for a selective university. The Journal of Human Resources, 19 (2), 202. Espeland, W. M., & Sauder, M. (2007). Ranki ngs and reactivity: How public measures recreate social worlds. American Journal of Sociology, 113 (1), 1-40. Freid, L. (2005). Reputation and prestige in American research universities: An exploration of the history of rankings a nd the increasing importance of student

PAGE 134

126 selectivity in perceptions of quality in higher education. (Ed.D., University of Pennsylvania). Fuller, W. C., Manski, C. F., & Wise, D. A. (1982). New evidence on the economic determinants of postsecondary schooling choices. The Journal of Human Resources, 17 (4), 477-498. Gall, M. D., Gall, J. P., & Borg, W. R. (2007). Educational research: An introduction Boston, MA: Pearson Education, Inc. Geiger, R. L. (2004). Knowledge and money: Research universities and the paradox of the marketplace Stanford, CA: Stanford University Press. Geiger, R. L. (2002). The competition for high-ability students: Universities in a key marketplace. In S. Brint (Ed.), The future of the city of the intellect (pp. 82-106). Stanford: Stanford University Press. Goenner, C., & Snaith, S. (2004). Assessing th e effects of increased admission standards. College and University, 80 (1), 29-34. Gonzalez, K., Stoner, C. & Jovel, J. (2003). Ex amining the role of so cial capital in access to college for Latinos: Toward a college opportunity framework. Journal of Hispanic Higher Education 2(1), 146-170.

PAGE 135

127 Hearn, J. C. (1991). Academic and nonacademic influences on the college destinations of 1980 high school graduates. Sociology of Education, 64 (3), 158-171. Hearn, J. C. (1987). Impacts of undergraduate experiences on aspira tions and plans for graduate and professional education. Research in Higher Education, 27 119-141. Higher Education Research Institute. CIRP Freshman Survey: Reliability and Validity Retrieved July 27, 2008, from http://www.gseis.ucla.edu/heri/PDFs/ CIRP_Reliability_Validity.PDF ). Higher Education Resear ch Institute. (2008). The Freshman Survey. Retrieved March 29, 2008, from http://www.gseis.ucla.e du/heri/cirpoverview.php Hill, J.L. (2008). The ambiguous effects of undergraduate debt: Extending the human capital model of graduate school enrollment. Dissertation Hilton, T. L. (1992). Using national data bases in educational research Hilsdale, NJ: Lawrence Earlbaum Associates. Hossler, D., Braxton, J. M., & Coopersmith, G. (1989). Understanding student college choice. In J. Smart (Ed.), Higher Education Handbook of Theory and Research New York, NY: Agathon Press. Hossler, D., & Foley, E. M. ( 1995). Reducing the noise in the college choice process: The use of college guidebooks and ratings. In R. D. Walleri, & M. K. Moss (Eds.),

PAGE 136

128 Evaluating and Responding to College Guidebooks and Rankings (New Directions for Institutional Research Series ed., pp. 21-30). San Francisco, CA: Jossey-Bass. Hossler, D., & Gallagher, L. (1987). Studying college choice: A three-phase model and the implication for policy makers. College and University, 2 207-221. Hossler, D., Schmit, J. L., & Vesper, N. (1999). Going to college: How Social, Economic, And Educational Factors Influen ce the Decisions Students Make Baltimore, Md.: Johns Hopkins University Press. Hossler, D., & Stage, F. K. (1992). Family and high school experience influences on the postsecondary educational plan s of ninth-grade students. American Educational Research Journal 29(2), 425-451. Hoxby, C. M. (1997). How the Changing Market Structure of US Higher Education Explains College Tuition NBER Working Paper N o. W6323). Cambridge, MA: National Bureau of Economic Research. Hoyt, J. E., & Brown, A. B. (2003). Marketing UVSC: How prosp ective students view the college. Unpublished manuscript. Hu, S., & Hossler, D. (2000). Willingness to pay and preference for private institutions. Research in Higher Education, 41 (6), 685-701.

PAGE 137

129 Hubbard, L. (1999). College aspirations among low-income African American high school students: Gendered strategies for success. Anthropology & Education Quarterly, 30(3), 363-383. Jackson, G. (1982). Public efficiency and private choice in higher education. Educational Evaluation and Method Analysis 4, 237-247. Johnson, R. G., Stewart, N. R., & Eberly, C. G. (1991). Counselor impact on college choice. The School Counselor, 39 (2), 84-91. Kane, T. (1998). Racial and ethnic preferences in college admission. In C. Jencks, & M. Phillips (Eds.), The black-white test score gap Washington, D.C.: Brookings Institution. Keup, J. R. (2004, Mar-Apr). The cooperative institutional research program Freshman Survey and Your First College Year: using longit udinal data to assess to first year of college. Assessment Update 16(2), 8-10. Kiecolt, K. J., Nathan, L. E. (1985). Secondary analysis of survey data Beverly Hills: Sage Publications, Inc. Kim, D. (2004). The effect of financial ai d on students' college choice: Differences by racial groups. Research in Higher Education, 45 (1), 43-70.

PAGE 138

130 Kinzie, J., Palmer, M., Hayek, J., Hossler, D., Jacob, S. A., & Cummings, H. (2004). Fifty years of college choice: Social, polit ical, and institutional influences on the decision-making process. Lumina Foundation for Education, 5 (3), 40-44. Kotler, P., & Fox, K. (1985). Strategic Marketing for Educational Institutions. Englewood Cliffs, NJ: Prentice-Hall. Leslie, L., & Brinkman, P. (1988). The Economic Value of Higher Education New York, NY: American Council on Education and Macmillan. Levine, A. & Nidiffer, J. (1996). Beating the odds: How the poor get to college San Francisco: Jossey-Bass. Lillard, D., & Gerner, J. (1999). Getting to the Ivy League: How family composition affects college choice. Journal of Higher Education, 70 (6), 706-730. Litten, L. H. (1982). Different strokes in the applicant pool: Some refinements in a model of student college choice. Journal of Higher Education, 53 (4), 383-402. Locke, L. F., Spirduso, W. W., & Silverman, S. J. (2007). Proposals that Work: A Guide for Planning Dissertations and Grant Proposals (5th ed.). Thousand Oaks, CA: Sage Publications.

PAGE 139

131 MacAllum, K., Glover, D. M., Queen, B., & Riggs, A. (2007). Deciding on postsecondary education: Final report ; NPEC 2008-850. Washington, DC: National Postsecondary Education Cooperative. Machung, A. (1998). What's at stake: College rankings and the new media. Change, July/August 13-16. Manski, C. F., & Wise, D. A. (1983). College Choice in America Cambridge, MA: Harvard University Press. Mau, W. (1995). Educational planning a nd academic achievement of middle school students: A racial a nd cultural comparison. Journal of Counseling & Development 73(5), 518-535. McDonough, P. M. (1997). Choosing Colleges: How Socia l Class and Schools Structure Opportunity New York, NY: SUNY Press. McDonough, P.M. (1994). Buying and selling higher education: The social construction of the college applicant. Journal of Higher Education 65, 427-446. McPherson, M. S., & Schapiro, M. O. (1991). Do es student aid affect college enrollment? New evidence on a persistent controversy. The American Economic Review, 81 (1), 309-318.

PAGE 140

132 Michael, R.T. & Becker, G.S. (1973). On the new theory of consumer behavior. Swedish Journal of Economics 75(4), 378-396. Monks, J., & Ehrenberg, R. G. (1999). U.S. News & World Report's college rankings. Change, 31 (6), 42. Moriarty, H. J., Deatrick, J. A., Mahon, M. M., Feetham, S. L., Carroll, R. M., Shepard, M. P., & Orsi, A. J. (1999). Issues to consider when choosing and using large national databases for research of families. Western Journal of Nursing Research 21(2)m 143-153. Mufson, S. (1999), Rankings all-importan t to GWU. [Electronic version]. Washington Post, pp. A01. Orfield, G. (1992). Money, equity, and college access. Harvard Educational Review, 62 (3), 337-372. Parker, J., & Summers, J. ( 1993). Tuition and enrollment yiel d at selective liberal arts colleges. Economics of Education Review, 12 (4), 311-324. Pascarella, E. T., & Terenzini, P. T. (1991). How College Affects Students: Findings and Insights from Twenty Years of Research San Francisco, CA: Jossey-Bass Publishers.

PAGE 141

133 Paulsen, M. B. (1990). College Choice: Understanding Student Enrollment Behavior Washington, DC: ERIC Cleari nghouse on Higher Education. Paulsen, M. B., & St. John, E. P. (2002). Soci al class and college costs: Examining the financial nexus between college choice and persistence. Journal of Higher Education, 73 (2), 189-236. Porter, S. R., & Whitcomb, M. E., (2005). Non -response in student su rveys: the role of demographics, engagement and personality. Research in Higher Education 46(2), 127-152. Reynolds, J. & Pemberton, J. (2001). Rising college expectations among youth in the United States: A comparison of the 1979 and 1997 NLSY. Journal of Human Resources 36(4), 703-724. Ruby, C. (2007). Coming to the United States: An examination of the factors related to international students’ gr aduate school choice. Dissertation Abstract International 68(04), (UMI No. AAT3259708). Sevier, R. A. (1994). Image is everything -strategies for measuring, changing, and maintaining your institution's image. College and University, 69 (2), 60-75. Sewell, W. H., Haller, A. O., & Portes, A. (1969). The educational and early occupational attainment process. American Sociological Review 34, 82-92.

PAGE 142

134 Sheehan, G. R. Assessment of the quality of institutes of higher education: Correlational analysis of U.S. news and world repo rt's university and college rankings and institution size, age, sele ctivity, graduation rate, facu lty salaries, wealth, and private/public status. (Ed.D., Univer sity of Southern California). Sweetland, S.R. (1996). Human capital theory : Foundations of a field of inquiry. Review of Educational Research 66(3), 341-359. Tierney, W.G. & Venegas, K.M. (2006). Fictiv e kin and social capita l: The role of peer groups in applying and paying for college. American Behavioral Scientist 49(12), 1687-1702. Toutkoushian, R. K. (2001). Do parental inco me and educational at tainment affect the initial choices of New Hamp shire's college-bound students? Economics of Education Review, 20 (3), 245-262. Undergraduate Ranking Criteria and Weights. U.S. News & World Report, America’s Best Colleges 2008 Posted 8/17/2007. Retrieved July 13, 2008, from http://colleges.usnews.rankingsandrevie ws.com/usnews/edu/college/rankings/about /weight_brief.php Webster, D. S. (1992). Rankings of undergraduate education in the U.S. News & World Report and Money : Are they any good? Change, 24 (March/April), 18-30.

PAGE 143

135 Weiler, W. C. (1994). Transition from consider ation of college to the decision to apply. Research in Higher Education, 35 (6), 631-646.

PAGE 144

136 APPENDICES

PAGE 145

137 Appendix A 2004 CIRP Freshman Survey

PAGE 146

138 Appendix A 2004 CIRP Freshman Survey (cont).

PAGE 147

139 Appendix A 2004 CIRP Freshman Survey (cont).

PAGE 148

140 Appendix A 2004 CIRP Freshman Survey (cont).

PAGE 149

141 Appendix B: CIRP Freshman Survey: Reliability and Validity

PAGE 150

142 Appendix B: CIRP Freshman Survey : Reliability and Validity (cont.)

PAGE 151

143 Appendix B: CIRP Freshman Survey : Reliability and Validity (cont.)

PAGE 152

144 Appendix B: CIRP Freshman Survey : Reliability and Validity (cont.)

PAGE 153

145 Appendix C National Universities part icipating in the 2004 Freshman Surveya aUniversities participating in the CIRP according to the Higher Education Research Institute website, accessed March 30, 2008, from http://www.gseis.ucla.edu/heri/PDFs/vdeck.pdf bThe tier assigned to each institution is based upon the 2003 Best Colleges edition by U.S. News & World Report. Tier 1 Tier 2 Tier 3 Tier 4 Boston College American University Duques ne University Adelphi University Brandeis University Baylor University Hofstra University Biola University Brown University Catholic University of America Indiana U of Pennsylvani a Cleveland State University California Inst of Tech Clarkson University Mississippi State Univer sity Georgia State University Carnegie-Mellon Univ Colorado State University Northeastern University Idaho State University Case Western Reserve U Florida State Universi ty Oklahoma State U North Dakota State Univ Cornell University Fordham Univ ersity Oregon State University Northern Arizona University Duke University Iowa State University Saint John's Univ-Queens Oakland University Emory University Loyola University of Chicago South Dakota State University Texas A&M Univ-Kingsville Georgia Inst of Tech Marquette University S outhern Illinois Univ Texas Woman's University Johns Hopkins University Miami Un iversity Texas Tech University Univ of Arkansas-Little Rock Massachusetts Inst of Tech Michigan Tech University Univof Arkansas-Fayetteville Univ of Louisiana at Lafayette New York University Ohio State University University of Idaho Univ of Mass-Boston Northwestern University Purdue University Univ ersity of Illinois-Chicago University of Toledo Rensselaer Polytechnic Inst Rutgers U-New Brunswick Un iversity of New Mexico Rice University Rutgers University-Newark University of North Dakota Tulane University Seton Hall University Univ of WisconsinMilwaukee Univ of California-Davis Southern Methodist University Utah State University Univ of California-Irvine SUNY-Bi nghamton Wayne State University Univ of California-LA SUNY-Stony Brook Univ of California-San Diego SUNY-University at Buffalo Univ of Calif-Santa Barbara Texas A & M University University of Chicago Texas Christian University University of Michigan University of Alabama Univ of N Carolina-Chapel Hill Univ of California-Riverside University of Notre Dame Univof California-Santa Cruz University of Pennsylvania University of Kentucky University of Rochester Univ of Mass-Amherst University of Southern Cal University of Pittsburgh University of Virginia University of San Diego Vanderbilt University University of Vermont Wake Forest University Virginia Tech

PAGE 154

ABOUT THE AUTHOR Holly J. Schoenherr received a Bachelor’s degree in Criminal Justice from Kent State University and a Master’s degree in Bu siness Administration from the University of South Florida. Ms. Schoenherr began her prof essional career as a social worker, serving inner-city youth and families in Columbus, Ohio. With two of her colleagues, she ventured into the creation of The Leadership and Challenge Center an adventure-based training program serving both non-profit and for-profit organizati ons in areas of leadership, communication, and st rategic planning. She has se rved as a human resources executive for a professional employment orga nization, an internati onal corporation, and a public research university. Currently serving as Special Assistant to the Provost at the University of South Florida, Ms. Schoenhe rr’s portfolio includes communications, faculty and executive searches, human resources, facilities planning and space allocations, and service on several univ ersity-wide committees and workgroups.