USF Libraries
USF Digital Collections

Florida's College Placement Test reading scores as an essential indicator for successful completion of the highest colle...

MISSING IMAGE

Material Information

Title:
Florida's College Placement Test reading scores as an essential indicator for successful completion of the highest college preparatory course in reading
Physical Description:
Book
Language:
English
Creator:
Smith, Laura Dandar
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Persistence
Remedial
Developmental
Florida Computerized Placement Test
Mandatory placement testing
Adaptive tests
Logistic regression
Dissertations, Academic -- Higher Education -- Doctoral -- USF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: The purpose of this study was to examine the predictive validity of several variables to determine if the Florida Computerized Placement Test - Reading (CPT-R) score alone, or other variables, could determine whether or not a student would successfully pass the highest level college preparatory reading course. The study examined fall sessions 1997-2004 (n=276,079) reading scores for all forms of the CPT to determine at what standard deviation below the cutoff score of 83 a student could still successfully complete the highest level college preparatory reading course.^ ^According to the College Board, the 83 scaled score, which exempts a student from taking the reading course, equates to approximately a 70% on the paper/pencil version of the test, yet the study revealed that a scaled score of 64 was the average score for fall sessions 1997-2004, which according to previous studies equates to 9/10th reading grade level on the Nelson-Denny Reading Test (Napoli & Raymond, 1998). In addition, the most frequently obtained scaled score was 75 for fall sessions 1997-2004, which equates to an 11th grade reading level on the Nelson-Denny Reading Test; however, the results of this study showed only 61% (49, 281 out of 79,167) of the upper quartile of students (scaled scores >74) passed the highest level college preparatory reading course.^ Although a statistically significant relationship was found between the entry test and successful completion of the highest level college preparatory reading course, the relationship was small, and therefore does not provide very good predictive validity. Interestingly, the study revealed that students who were exempt from the reading course, and still enrolled in the course, did not have higher passing rates in the course. In addition, students with higher placement scores did not have significantly higher passing rates in the reading course than students with lower placement scores. In fact, students with the lowest scaled scores of 11-20 had the highest percentage of successfully completing the highest level college preparatory reading course.The placement test scores in reading indicate a large number of students entering Florida's community colleges are not prepared for college-level courses.^ In addition, the results of this study indicated that the placement test did very little to discriminate between levels of students' actual reading abilities and predict which students will ultimately pass required remedial/developmental reading classes. Although many first-time-in-college students are not recent high school graduates, high schools should be required to include reading as part of the core curriculum, separate and distinct from the language arts courses.Teachers, credentialed in reading, should be teaching reading courses in all four years of high school. Diagnostic testing and year-end testing should occur each year to chart a student's progress for all four years of high school. In addition, Florida's college entrance reading placement test should be revised so that it provides a comprehensive measurement of college-level reading skills.
Thesis:
Dissertation (Ph.D.)--University of South Florida, 2007.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Laura Dandar Smith.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 106 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 - 001925073
oclc - 191544156
usfldc doi - E14-SFE0001892
usfldc handle - e14.1892
System ID:
SFS0026210:00001


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 Ka
controlfield tag 001 001925073
003 fts
005 20080128150825.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 080128s2007 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001892
035
(OCoLC)191544156
040
FHM
c FHM
049
FHMM
090
LB2322 (ONLINE)
1 100
Smith, Laura Dandar.
0 245
Florida's College Placement Test reading scores as an essential indicator for successful completion of the highest college preparatory course in reading
h [electronic resource] /
by Laura Dandar Smith.
260
[Tampa, Fla] :
b University of South Florida,
2007.
3 520
ABSTRACT: The purpose of this study was to examine the predictive validity of several variables to determine if the Florida Computerized Placement Test Reading (CPT-R) score alone, or other variables, could determine whether or not a student would successfully pass the highest level college preparatory reading course. The study examined fall sessions 1997-2004 (n=276,079) reading scores for all forms of the CPT to determine at what standard deviation below the cutoff score of 83 a student could still successfully complete the highest level college preparatory reading course.^ ^According to the College Board, the 83 scaled score, which exempts a student from taking the reading course, equates to approximately a 70% on the paper/pencil version of the test, yet the study revealed that a scaled score of 64 was the average score for fall sessions 1997-2004, which according to previous studies equates to 9/10th reading grade level on the Nelson-Denny Reading Test (Napoli & Raymond, 1998). In addition, the most frequently obtained scaled score was 75 for fall sessions 1997-2004, which equates to an 11th grade reading level on the Nelson-Denny Reading Test; however, the results of this study showed only 61% (49, 281 out of 79,167) of the upper quartile of students (scaled scores >74) passed the highest level college preparatory reading course.^ Although a statistically significant relationship was found between the entry test and successful completion of the highest level college preparatory reading course, the relationship was small, and therefore does not provide very good predictive validity. Interestingly, the study revealed that students who were exempt from the reading course, and still enrolled in the course, did not have higher passing rates in the course. In addition, students with higher placement scores did not have significantly higher passing rates in the reading course than students with lower placement scores. In fact, students with the lowest scaled scores of 11-20 had the highest percentage of successfully completing the highest level college preparatory reading course.The placement test scores in reading indicate a large number of students entering Florida's community colleges are not prepared for college-level courses.^ In addition, the results of this study indicated that the placement test did very little to discriminate between levels of students' actual reading abilities and predict which students will ultimately pass required remedial/developmental reading classes. Although many first-time-in-college students are not recent high school graduates, high schools should be required to include reading as part of the core curriculum, separate and distinct from the language arts courses.Teachers, credentialed in reading, should be teaching reading courses in all four years of high school. Diagnostic testing and year-end testing should occur each year to chart a student's progress for all four years of high school. In addition, Florida's college entrance reading placement test should be revised so that it provides a comprehensive measurement of college-level reading skills.
502
Dissertation (Ph.D.)--University of South Florida, 2007.
504
Includes bibliographical references.
516
Text (Electronic dissertation) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 106 pages.
Includes vita.
590
Adviser: Jan Ignash, Ph.D.
653
Persistence.
Remedial.
Developmental.
Florida Computerized Placement Test.
Mandatory placement testing.
Adaptive tests.
Logistic regression.
690
Dissertations, Academic
z USF
x Higher Education
Doctoral.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1892



PAGE 1

Floridas College Placement Test Reading Scores as an Essential Indicator for Successful Completion of the Highest College Preparat ory Course in Reading by Laura Dandar Smith A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Education Department of Adult, Career, and Higher Education College of Education University of South Florida Major Professor: Jan Ignash, Ph.D. William Benjamin, Ph.D. John Ferron, Ph.D. William Young III, Ed.D. Date of Approval: March 7, 2007 Keywords: persistence, remedial, developmen tal, Florida Computerized Placement Test, mandatory placem ent testing, adaptive tests, logistic regression Copyright 2007, Laura Dandar Smith

PAGE 2

Dedication My dissertation is dedicat ed to my mother, Anne and my father, the late Arthur Dandar, who instilled the belief that an education is the most important gift one can give a child; to my husband Russ, who pr ovided words of encouragement; and to my very dear children, Matthew and Kirstyn, who lovingly remained patient for this day to come.

PAGE 3

Acknowledgments I would like to acknowledge the professi onal assistance of my committee: Dr. Jan Ignash, chair, Dr. John Ferron, Dr. William Young, and Dr. William Benjamin. I am most particularly thankful for the patience and guida nce of Dr. Jan Ignash. I am indebted to her, for I know, without her encouragement, I would not have finishe d. I appreciate the statistical guidance provided by Dr. Ferron and one of his doctoral students, Jesse Coraggio, from the Consulting Office for Res earch in Education Finally, I am appreciative for Dr. Young and Dr. Benjamin s levity at times but, more importantly, their willingness to give their profes sional time and expertise to my study.

PAGE 4

Table of Contents List of Tables................................................................................................................. .iii List of Figures...................................................................................................................v Abstract....................................................................................................................... ....vi I. Introduction..............................................................................................................1 Statement of Problem...............................................................................................3 Significance of the Problem.....................................................................................5 Purpose of the Study................................................................................................7 Research Questions..................................................................................................8 Definition of Terms..................................................................................................9 Limitations/Delimitations......................................................................................12 Summary................................................................................................................14 II. Review of Literature..............................................................................................16 Historical Perspective of Developmental Studies..................................................16 NCES Fall 2000 Study...........................................................................................20 Demographics of Community College Students....................................................22 Academic Background of Developmental Students..............................................23 Indicators of Success fo r Developmental Students................................................24 Mandatory Placement Testing...............................................................................25 NCES Fall 2004 Study...........................................................................................27 Florida Research in Developmental Education......................................................30 Cooling Out Function of Community Colleges.................................................40 Summary................................................................................................................42 III. Methods..................................................................................................................4 4 Research Questions and Hypotheses.....................................................................45 Procedures..............................................................................................................47 Participants/Data Collection..................................................................................47 Variables................................................................................................................47 Instrumentation......................................................................................................48 CPT Predictive Validity.........................................................................................50 Data Analysis.........................................................................................................53 Summary................................................................................................................56 i

PAGE 5

IV. Results.................................................................................................................... 57 Data Collection......................................................................................................59 Data Analysis.........................................................................................................60 Descriptive Statistics..............................................................................................61 Research Question 1..............................................................................................67 Research Question 2..............................................................................................72 Research Question 3..............................................................................................76 Summary................................................................................................................81 V. Summary of Findings, Conclusions and Implications for Theory, Practice and Research............................................................................................83 Method Summary...................................................................................................83 Descriptive Data.....................................................................................................84 Summary of Findings.............................................................................................85 Conclusions............................................................................................................87 Limitations.............................................................................................................89 Implications for Theory.........................................................................................89 Implications for Practice........................................................................................91 Implications for Research......................................................................................94 List of References...........................................................................................................98 Appendices....................................................................................................................105 Appendix A: Comparison of SAS Probability of Passing Rates and Descriptive Statistics (Actual) Passing Rates in Highest Level College Preparatory Reading Course...................................................................106 About the Author................................................................................................End Page ii

PAGE 6

List of Tables Table 1 First Time in College (F TIC) Degree-Seeking Students Taking Entry Level Test: College Prep aratory Success Report by Ethnicity...........34 Table 2 Enrollment of College Preparatory Students by Areas Required and Awards Earned..............................................................................................35 Table 3 Awards (Degrees) Earned by African Americans........................................36 Table 4 Awards (Degrees) Earned by Hispanics.......................................................37 Table 5 Awards (Degrees) Earned by Whites............................................................38 Table 6 Comparison of Awards Earn ed by College Ready and College Preparatory Reading Students.......................................................................39 Table 7 Fall 1997 First Time in College Freshman CPT-Reading Scaled Scores............................................................................................................40 Table 8 FTIC Students Reading Comput erized Placement Test Scores Fall Sessions 1997-2004......................................................................................61 Table 9 Comparison of Upper and Lower Quartiles Reading Placement Scaled Scores With Passing Rates for Students Taking the Highest Level College Preparatory Reading..............................................................63 Table 10 Frequency of Reading Placement Test Scores and Passing Rates in Reading Course for Fall sessions 1997-2004...............................................64 Table11 Computerized Placement Test Reading Scores for Students Required to Enroll in the Reading Course....................................................65 Table 12 Frequency of Grades for Students Required to Enroll in the Reading Course.............................................................................................67 Table 13 Odds Ratios Estimates..................................................................................69 iii

PAGE 7

Table 14 Analysis of Maximum Likelihood Estimates...............................................70 Table 15 Logistic Regression Mode l for Probability of Successful Completion of Reading Course.....................................................................70 Table 16 Association of Predicted Pr obabilities and Observed Responses.................71 Table 17 Frequency of Enrollment for Fall Sessions 1997-2004................................73 Table 18 Odds Ratios Estimates..................................................................................73 Table 19 Analysis of Maximum Likelihood Estimates...............................................74 Table 20 Logistic Regression Mode l for Probability of Successful Completion of Reading Course.....................................................................75 Table 21 Association of Predicted Pr obabilities and Observed Responses.................75 Table 22 Frequency of Grades for Successful Completion of Reading for Fall Sessions 1997-2004...............................................................................77 Table 23 Descriptive Statistics of Grade Point Averages the Session Following FTIC Students Successful Completion of Reading Course Fall Sessions 1997-2004...................................................................78 Table 24 Frequency of Program Levels.......................................................................79 Table 25 Mean Grade Point Averages According to Program of Study......................79 Table 26 Analysis of Program Level to Grade Point Averages...................................80 Table 27 Tukeys Studentized Range (HSD) Test for Grade Point Averages.............81 Table 28 Fall Sessions 1997-2004 Computerized Placement Test-Reading Scaled Score Comparisons............................................................................84 iv

PAGE 8

List of Figures Figure 1. Histogram and boxplot of all FTIC students reading placement test scores......................................................................................................62 Figure 2. Histogram and boxplot of r eading placement test scores of students required to en roll in the reading course..........................................65 Figure 3. Histogram and boxplot of grade point averages the session following successful completion of reading course......................................78 Figure 4. Fall sessions 1997-2004: Comparisons of pe rcentages of students passing reading course and corresponding CPT-R scaled score ranges............................................................................................................82 v

PAGE 9

Floridas College Placement Test Reading Scores as an Essential Indicator for Successful Completion of the Highest College Preparatory Course in Reading Laura Dandar Smith ABSTRACT The purpose of this study was to examin e the predictive validity of several variables to determine if the Florida Com puterized Placement Test Reading (CPT-R) score alone, or other variab les, could determine whether or not a student would successfully pass the highest level college preparatory reading course. The study examined fall sessions 1997-2004 ( n =276,079) reading scores for all forms of the CPT to determine at what standard deviation below the cutoff score of 83 a student could still successfully complete the highest level college preparatory reading course. According to the College Board, the 83 scaled score, whic h exempts a student from taking the reading course, equates to approximately a 70% on the paper/pencil version of the test, yet the study revealed that a scaled score of 64 wa s the average score for fall sessions 1997-2004, which according to previous studies equates to 9/10 th reading grade level on the NelsonDenny Reading Test (Napoli & Raymond, 1998). In addition, the most frequently obtained scaled score was 75 for fall se ssions 1997-2004, which equates to an 11 th grade reading level on the Nelson-Denny Reading Te st; however, the results of this study vi

PAGE 10

showed only 61% (49, 281 out of 79,167) of th e upper quartile of students (scaled scores >74) passed the highest level colle ge preparatory reading course. Although a statistically significant relati onship was found between the entry test and successful completion of the highest leve l college preparatory reading course, the relationship was small, and therefore does not provide very go od predictive validity. Interestingly, the study revealed that students who were exem pt from the reading course, and still enrolled in the course, did not have higher passing ra tes in the course. In addition, students with higher placement scores did not have significantly higher passing rates in the reading course than students with lower placem ent scores. In fact, students with the lowest scaled scores of 11-20 had the highest percen tage of successfully completing the highest level college preparatory reading course. The placement test scores in reading indi cate a large number of students entering Floridas community colleges are not prepared for college-level courses. In addition, the results of this study indicated that the placem ent test did very little to discriminate between levels of students actual reading abilities and predict which students will ultimately pass required remedial/developmenta l reading classes. Implications from the results of this study affect both high school s and colleges. Although many first-time-incollege students are not recent high school gr aduates, high schools should be required to include reading as part of the core curriculum, separate and distinct from the language arts courses. Florida high schools need to impl ement intensive program s of study in reading because students are gravely underprepared for college studies. Teachers, credentialed in reading, should be teaching reading courses in all four years of high school. Diagnostic vii

PAGE 11

testing and year-end testing should occur each year to chart a students progress for all four years of high school. In addition, Flor idas college entrance reading placement test should be revised so that it provides a comprehensive m easurement of college-level reading skills. viii

PAGE 12

Chapter 1 Introduction Since most community colleges in the Unite d States view remediation as part of their mission, it is not surprising that in the fall of 2000, 98% of community colleges offered at least one remedial reading, wr iting, or mathematics course. Cliff Adelman, Senior Research Analyst at the U.S. Depa rtment of Education, reported approximately 63% of the students entering community college s required at least one remedial course (2004). At some community colleges, th is figure approaches 70% (McCabe, 1998). Because community colleges serve the community and can respond quickly to market needs, they have been more successful in attracting nontraditional learners (Miglletti, 1998). Many students arrive at community colleges lacking ba sic skills in math, reading and English. Community colleges respond by offering students who are not eligible to enter four-year institutions an opportunity to remediate their skills and obtain a college education that would otherwise be out of their reach because of poor basic skills (Adelman, 1996). A substantial number who ente r college underprepared are still able to be successful because of developmental education. In Florida, where the proposed study occurred, the Florida Student Database (FSDB) provided information on gender, ethnicity, age and disabilit y. In the school year, 1998-99, the Florida Office of Educational Serv ices and Research reported the typical 1

PAGE 13

community college student was a thirty-one-y ear-old, white female attending part-time, seeking an AA degree and not receiving aid nor having a disabilit y. Whites comprised 65.4%, while Blacks comprised 15.8%, Hispan ics 2.6%, American Indians 0.5% and 0.6% did not report ethnicity or race. Although the average age was thirty-one, 46.5% were twenty-five or younger. Only 2.1% reported a disability; and those reporting a learning disability were the larges t portion of this category. In 2003, the typical Florida college preparatory student was a female between the ages of 26-35. Slightly half of college prepar atory students were 21 years of age or over, about one-third full-time, and two-thirds indi cated an Associate of Arts degree as their educational goal. Finally, almo st 4% of college preparat ory students were disabled (Windham, 2003, p. 2). In Florida, the Division of Comm unity Colleges, examining the 1996 Accountability Report, which focused on succes sful completion of the highest level of college preparatory courses in reading, writing, and mathematics, found a difference among age groups failing the placement test for the first time. When the 24 and younger group are split into even finer age ranges, bei ng out of high school for even one year has a negative impact on the ability to pass the placement test, but even though the older groups failed at least one section of an entry-leve l placement test more often than those 24 and under, they completed the highest level colleg e preparatory courses at a higher rate than the younger students for both reading and mathematics. Tinto (1987) recommended institutions n eed to develop warning systems to identify and track students who may have difficulty completing programs of study. Since 2

PAGE 14

the Fall of 1997, Florida community colleges are using the College Boards Computerized Placement Test with cut-offs scores to identify remediation for reading, writing, and mathematics. A scaled score of 83 or higher on the r eading portion exempts a student from having to take remedial re ading and means the student is ready for college-level courses. Many students repe at the remedial courses several times; consequently, Florida enacted a repeat policy, which allows the student to take the remedial course twice at stat e tuition rates; however, upon the third attempt, the student must pay out-of-state tuition and the instru ctor must award a grade. This has not dissuaded developmental students, for many students are persistent, from re-entering college several times in hopes that the second or third try will meet with success. Statement of the Problem The rising costs of attending four-year colleges, the increase in college-bound high school students, and a larger number of nontraditional students have resulted in an increasing number of students enrolling in two-year schools nationwide. Florida community colleges compound the problem of escalating enrollments by not identifying students who are unlikely to ever pass the de velopmental reading cl asses. Consequently, the problem is two-fold: no classroom space for traditional classes and continuous enrollment of students who have serious skills deficiencies and are unlikely to ever graduate. This problem could be alleviated if the placement test was used as a screening tool, not merely for placement into developmental classes. Florida community colleges need to provide counseling that includes info rming a student when hi s skill levels are too deficient to remediate at a community colle ge. Determining a cut-off score in reading 3

PAGE 15

which identifies which students will not be su ccessful in passing, and therefore would not complete a college program of study, would sa ve many students valuable time and money and let them seek other alternatives for career pursuits. Therefore, research needs to be conducted to identify at what point belo w the Computerized Placement Test-Reading (CPT-R) cut-off score of 83, a student in Florida will still successfully pass. In the 1990s, the National Study of De velopmental Education found 77% of developmental students at two-year college s and 98% attending four-year institutions intended to obtain a college degree (B oylan & Bonham, 1992, p.2). Of the 1992, 12 th graders enrolled in postsecondary education and completing coursework within eight years of high school graduation, 69% not need ing remedial coursework earned a specific degree or certificate compared to 30% who needed any remedial reading (Wirt et al., 2004, p. 63). Clifford Adelman claims degr ee completion is the true bottom line for college administrators, state legislators, parents, and most importantly, studentsnot retention to the second year, not persistence without a degr ee, but completion (1999). The NCES Fall 2000 study reported the propor tion of students requiring remedial reading who did not earn postsecondary credentials rose from 57% in 1982 to 70% in 1992 (Adelman, 2004, p. 94). In 1999, one out of eight students took remedial reading courses, and 65% of this group needed to ta ke at least three other remedial courses, including math (Adelman cited in McCusker 1999, p. 1). According to Clifford Adelman, Deficiencies in reading skills are indicators of comprehensive literacy problems, and they significantly lower the odds of a students completing a degree (1996, p. A56). 4

PAGE 16

The Florida Community College System s open door policy is a revolving door policy for many students who leave with nothing more than time and money expended. As more and more nontraditional students apply to community colleges, counselors need to advise students who need remediation, especially in reading, that a two-year degree may not be a realistic option. Community colleges have been criticized for providing a cooling out function, which is nothing more than retaining a student until he finally realizes he will never graduate from the community college. As Adelman (1996, p. A57) has stated, the findings strongl y suggest that we cannot continue to let high-school graduates believe th at they have a good chance of earning a college degree if they leave high school with poor reading skills. Thus, student failure does not come from barriers imposed by the colleges, but from a failure of colleges (especially community colleges) to convey clear inform ation about the preparation th at high school students need in order to have a chance of fini shing a degree (Rosenbaum, 1999). Significance of the Problem Remediation has always been with us, and there is no evidence in the four surveys of remediation conducted in 1983, 1989, 1995 and 2004 by the National Center for Education Statistics of any significant incr ease or decrease in the number of remedial students. According to the Brookings Institute developmental students do not represent a cost-burden; in fact, total public expenditu re is less than 1% of the public higher education budget (Breneman & Haarlow, 1998, p. 5). Yet, many states want accountability for the continuance of remedial programs in community colleges. More states are requiring ou tcome evidence and statewide polic ies governing remedial services. 5

PAGE 17

For example, Florida, Colorado, and South Caro lina prohibit remedial education at fouryear institutions. Virginia, Minnesota, Mary land, Georgia, Nevada, Missouri, New York, and Ohio are also considering similar legisla tion, and some states ar e debating whether to require students to pay back the cost of remediation (Ken tucky Council on Postsecondary Education cited in An Analysis of Devel opmental Education at Michigans Associate Degree-Granting Institutions, 1999). Many states allow remedi al work to count towards institutional credit, for financial aid and fundi ng reasons, but the major ity of states do not permit remedial course work to count towa rds degree or graduation credit (Breneman & Harlow, 1998). The 1998 Reauthorization of the Higher Education Act prompted debate over whether or not developmental and reme dial programs are appropriate at the postsecondary level. Of the nations more th an 12 million undergraduates, about two and one-half million participate in developmental education during any given year (Boylan, 1999, p. 1). The need for remedial reading appe ars to be the most serious barrier to degree completion; in fact, 51.1% of the stud ents needing remedial reading are required to enroll in four or more courses (Wirt et al., 2004, p. 141). The Condition of Education 2004 found 10.6% of all entering college freshmen needed a remedial reading course, and of that group only 7% attained an associ ates degree and 17% attained a bachelors degree (Wirt et al., 2004, p. 63). The number of students being served commands educators to examine this population further. In August, 2000, the Board of Directors of the American Association of Community Colleges recommended one way to improve remedial education was to 6

PAGE 18

evaluate remedial education courses a nd programs regularly to assess student performance, review average time needed for course completion, evaluate student performance in follow-up courses, and comp are graduation rates of students requiring remediation in one or more skills with those who did not (p.2). In a personal communication, Associate Vice Chancellor for Evaluation Dr. Patricia Windham of the Florida Division of Community Colleges and Workforce Education suggested a study to examine at what point below the entry-level placement test cut-off score do students not pass the highest level preparator y course in reading (January, 2004). This information could be us ed to identify which students should not enroll in community college coursework. Furthermore, Adelman (1999) purports that high schools are not providing a rigorous curr iculum, so by identifying which students are not capable of pursing a college degree, many high schools would be challenged to revise high school curriculums which provide the skills students need to successfully matriculate into a college program of study. Adelman further states that students should be advised to either seek another educational provider or receive in tense remediation in a specific time period (1999). Purpose of the Study Reading has been found to be the primar y indicator of successfully completing a college program of study; therefore, the purpose of this study was to examine the predictive validity of several variables to determine if the Florida Computerized Placement Test Reading (CPT-R) score alone or other variables, could determine whether or not a student woul d successfully pass the highest level college preparatory 7

PAGE 19

reading course. The study examined the read ing scores on the CPT to determine at what standard deviation below the cutoff score of 83 a student could still su ccessfully complete the highest level of the college preparatory reading course. Florid a community colleges counseling departments could better serv e students by knowing how many standard deviations below the scaled score of 83 on th e Florida Computerized Placement Test in reading indicates whether a student is likel y to pass the highest level reading college preparatory course. Colleges could then use this information to help make decisions about which students to admit to college programs of study. Research Questions This study focused on what variables determine whether a student can successfully pass the highest level college prep aratory reading course which indicates the student is ready for college-level courses. Th erefore, this study attempted to answer the following questions: 1. Is there a relationship between a students score on the Computerized Placement Test in reading (CPT-R) and success in passing the highest level college preparatory reading course in Florida? 2. Is there a relationship between full-time or part-time enrollment during the semester a student is taking the highest level college preparatory reading course and success in passing the highest level college preparatory reading course in Florida? 3. What are students' GPAs the session follo wing successful completion of the highest level college preparatory reading course acco rding to the program track (Associate of Arts, Associate of Science, A ssociate of Applied Science)? 8

PAGE 20

The independent variable of the study was the placement test score, which was a nominal independent variable. The othe r independent variables were full-time enrollment and part-time enrollment. The de pendent variable was course success in the highest level college-preparato ry reading course, defined as pass or fail, with passing represented as marks of A, B, C, S, or P. In addition, the students GPA following successful completion of the highest level co llege preparatory reading course was the dependent, continuous variable and passing or failing the reading course was the independent variable. Definition of Terms Several definitions were central to the research proposal. First, developmental instruction as defined by Cohen & Brawer (1996) is instruction that pr ovides activities to keep students in school, and helps them improve their basic skills, so they can complete an academic or vocational program satisfactorily. Developmental refers to programs that focus on the whole learner, blending academ ic with the personal strengths and weaknesses students bring to the learning process (Ignash, 1997, p. 3). Others have extended the definition to include activities such as learning skill centers, tutoring, advising, and counselin g (Miller, 1996). The term remedial refers to programs that focus on providing remedies for specific deficiencies in reading, writing and math. Recently, many refer to college preparatory courses as those courses providing remedial/developmental coursework via pre-college courses (i.e. basic skills) based on placement test scores. Thus, a remedial/developmental student is one whose score on the 9

PAGE 21

college placement test requires one or more combinations of classes in preparatory reading, math and writing. Other names incl ude compensatory education and basic skills (NCES 2003, p. 1). Course satisfaction is the final course mark th at is considered passing. For purposes of this study, pass/fail will be used as the variable to represent grades of A, B, C, Satisfactory (S) or Passing (P). Completers are students who have passed the highest level college preparatory course in reading, and non-completers are students who have not achieved course success in the highest level college pr eparatory course in reading. Indicators of success was defined in this study from the literature on developmental education as st udent persistence, developm ental course passing rates, passing grades in college-level courses, grade point averages and/or the ratio of credits attempted to credits earned. Th is study focused on CPT scores and course success in the highest level college prepar atory course in reading. GPA is the acronym for grade point averag e, which is calculated by computing the grades earned in each course with the number of credit hours taken. Only the last attempt of a repeated course is used in computing the grade-point average. A grade of W means a withdrawal from a course and is not computed in the GPA. A grade of W does not override a grade of F. The inst ructor determines an incomplete, and an incomplete (I) received at the end of any term becomes an F if not completed the succeeding fall or spring term. The student ma y not register for another section of the course during the period of the incomplete grade. A grade of N is used only in college 10

PAGE 22

preparatory courses and may be assigned to st udents earning a D or F in a college preparatory course. The grade of N is nonpunitive, indicating progress has been made but not at the level required for successful completion of th e course. College-preparatory courses are not computed in a students GPA. A student who is enrolled in twelve semest er hours in the fall or spring sessions is a full-time student. A student who is enrolled in less than twelve semester hours is a parttime student. Presage variables is another term for predictor va riables such as age, gender, and race/ethnicity. Computerized Placement Test (CPT) is the Florida placement test implemented in July, 1995. The placement test identifies stud ents who need remediation in reading, writing or mathematics prior to entry into college-level classes. The CPT identifies a scaled score of 83 on the reading subtest as exemption from having to take a college preparatory reading course. For purposes of this study, only the read ing subtest scores were examined. College Level Academics Skills Test (CLAST) is used in Florida to determine whether a student will be allowed to gra duate from an accredited community college and/or enroll in upper division courses. The test is usually taken after a student has completed 30 credit hours. The earliest point a student can take the exam is after 18 hours of college credit. The CLAST is classified as a criterion-based test and also a minimum competency test. Scores are recorded as pass/fail. 11

PAGE 23

Florida Basic Skills Exit Te s t is used as the criterion-referenced exit exam for the highest level college pr eparatory course in reading. The te st has different forms with the content developed by Florida reading professors. Successful completion of the highest level college preparatory reading course means the student may enroll in college-level courses. Workforce Development in the Florida Community College System (FCCS) provides training programs for em ployment in industries requ iring technical skills. In addition, FCCS provides continuing educati on and retraining for displaced workers. The Florida Education and Traini ng Placement Information Program (FETPIP) is a data collection system that provides follow-up data on former students, such as employment, military, public assistance pa rticipation, incarceration, and continuing postsecondary education. A major indicator of success for commun ity colleges is the awarding of degrees. Community colleges award various degrees, which include the associate of arts degree (A.A.), the associate of science degree (A.S.) the associate in applied science (A.A.S.), college credit certifi cate, the applied technology diplom a, and the post-secondary adult vocational certificate, which is non-co llege credit for occ upational training. Limitations/Delimitations This study was delimited to developmental programs in the Florida Community College System. Collection of data included the years 1997 to 2005, because these were the years that the Florida College Entry-Level Placement Test (CPT) was implemented as a placement instrument for Florida community colleges with the uniform standard cut-off 12

PAGE 24

score of 83. (Students who have scores on the College Boards SAT-1 or the American College Testing Programs Enhanced ACT test th at meet or exceed the scores in Rule 6A10.0315, Florida Administrative Code may be ex empted from the Florida College EntryLevel Placement Test.) The data from each of the twenty-eight community colleges is submitted electronically and the state compiles the data and forwards it back to each community college for review. Checks and balances ar e in place, for review of the data is continuously evaluated for errors each subse quent semester by the Institutional Research departments of the community colleges and we ll as the states management information data processing. Associate Vice Chancellor fo r Evaluation Dr. Patricia Windham of the Florida Division of Community Colleges and Workforce Education in a personal communication stated the data was accurate due to the checks and balances in place by both her department and the individual co mmunity colleges research departments (September, 2005). Floridas Council of Instructional Affairs, which includes academic administrators from all 28 Florida community colleges, determ ined as of fall semester, 1999, all Florida community colleges were required to administ er the Florida Basic Skills Exit Test, the instrument used as the exit exam for the hi ghest level college preparatory course in reading. In accordance with State Ru le 6A-10.315 Paragraph 19B (Florida Administrative Code Annotated, 1997), the 1997 Florida legislature made passing an exit test a condition for meeting basi c skills requirements. Accordi ng to the law, students must pass both the college preparatory reading cour se and the criterion-referenced test. The 13

PAGE 25

state of Florida does not record student sc ores on the Florida Basic Skills Exit Test; therefore, no predictions were made compari ng student scores on the Florida Basic Skills Exit Test with student scores on the College Placement Test. The comparison was made with the CPT score and successful completion of the highest level college preparatory reading course (pass/fail). It was assumed the CPT provided an accurate assessment of the students ability level and appropriately pl aced students into the college preparatory reading course. Summary Today, a college education is required for many career choices. In reality, many careers are not dependent upon someone having a traditional two-or four-year degree. Students who come to the community college looking for success in their lives depend on educators to counsel them effectively, en suring that decisions in career-planning are sound. Placement tests identify students, who need remediation, and community colleges provide remediation; however, this is not enough. Iden tifying students who cannot effectively complete traditional programs of study means that they need to be told they may be wasting their money, and more impor tantly, their time. Community colleges must go beyond merely placing students into remedial classes and begin to provide counseling to students beyond remediation of basic skills. The Southern Regional Education Board (SREB) (2004) stated only 45% of first-time college freshman attending full-time graduated from 1998-2001, and 32% of students failed to re turn for the second year at community colleges or other higher education institutions (Summers, 2003, p. 64). As of 2004, the SREB reported that with 30% of students graduating, Floridas 14

PAGE 26

community college students were 13 poi nts above the SREB average of 17%. Community colleges can possibly raise graduation rates by identifying those developmental students requiring remedial read ing coursework who are at a very high risk of not benefiting from college level programs. 15

PAGE 27

Chapter 2 Review of the Literature The purpose of remedial/developmental educat ion is "to enable students to gain skills necessary to complete college-level courses and academic programs successfully" (Weissman, Silk & Bulakowski, 1997, p.188); therefore, research to evaluate whether or not remedial/developmental programs are effective focuses on the indicators leading to successful completion of a college program of study which includes student persistence, grade point averages, developmen tal course passing rates, passing grades in college-level courses, and the ratio of credits attempted to credits earned. Many individual colleges and statewide college systems have conducted studies on the various success indicators in developmental education. Historical Perspective of Developmental Studies For almost 200 years, institutions of hi gher learning have been accepting students who may not have met their standards while tr ying to develop ways to meet the needs of diverse learners. One of the most distinctive features of the American educational system is that it gives thousands of worthy stude nts who would otherwis e be excluded a chance to attend higher education (B rint & Karabel, 1989, p. 10). In the early 1800s, education at all levels was provided to Americans; moreover, access to higher education was expanded. Since few opportunities existed for early Americans to obtain prerequisite 16

PAGE 28

skills for college, colleges and universities provided preparatory programs. Because learning an academic language such as Latin wa s not a priority for colonists attempting to survive in a new world, Harvard College ( 1638) was confronted with remediation by providing tutoring in Latin to incoming students. The use of scholar ly books written in Latin, and Latin as the language of instruction, continued into the 18 th century in America (Brubacher & Rudy cited in Boylan, 1987). In the first thirty years of the 19 th century, not enough student s were prepared for college because the development of colleges preceded the development of a widespread secondary school system. Many students who could afford tutoring were instructed by local ministers. However, soon the number of students requiring tutoring was too large. As a result, in the latter part of the 19 th century, many colleges began to offer compensatory education programs that would enable these students to succeed, thus compensating them for their lack of skill s with adequate remedial (very low underprepared) or developmental (average skill, but not at level for college success) programs (Cohen & Brawer, 1996). In 1849, the Universi ty of Wisconsin implemented the first college preparatory department providing reme diation in reading, wr iting and arithmetic (Brier cited in Boylan & White, 1987, p.2). The department became the model for many colleges and universities. During the late 19 th century, women began attendi ng college. Many argued that women were mentally unsuited for education, but, of course, they simply were underprepared. Thus, many of the new womens co lleges provided developmental education. 17

PAGE 29

In the same manner, as more and more black Americans entered colleges, the institutions provided developmental education for them (Boylan & White, 1987). The Morrill Act of 1862, also known as the Land Grant Act, stimulated growth in higher education by requiring institutions to promote higher education for a greater variety of Americans. In 1874, Harvard develo ped a course to remediate deficiencies of freshman, and in 1894, Wellesley College deve loped one of the first remedial courses (Cross, 1971). Harvard also implemented the first composition course, which served as a bridge for entering freshman at the level of competency for the Harvard curriculum (Maxwell cited in Boylan 1988). The Second Morrill Act of 1890 extended la nd grant colleges to the southern states, providing separate but equal school s for black Americans. Colleges were established in almost every state to provide training for merchants and tradesmen as well as engineers and scientists. Since education was not ma ndatory, few people had prior preparation for college; therefore, the major criterion for entry was the ability to pay admission fees. Substantial numbers of stude nts required tutoring, which resulted in tutoring classes outnumbering regular college classes, in some instances (Brier, 1984, p.2). By the turn of the century, more than 80% of U.S. colleges and universities offered college preparatory programs (M axwell cited in Boylan, 1988). The great disparity in admission policies and, in some cases, the lack of any admission policy, led to the establishment of the College Entrance Examination Board in 1890. The National Education Associati on (NEA) Committee on Secondary School Studies, called the Committee of Ten, de vised a secondary school curriculum for 18

PAGE 30

college matriculation. The College Entr ance Examination Boards major objectives were to standardize admission procedures, raise academic standards, and eliminate college preparatory programs. With the adve nt of admission testing and the inception of junior colleges at the turn of the century, colleges and universities began to phase out college preparatory programs. However, because of the various types of colleges, it was impossible to have a uniform admission stan dard. Consequently, there was never a standard of admission for all colleges in the United States. By the 1940s, junior colleges and special divisions within universities provided college preparatory programs. The community colleges ways of dealing with the underprepared took many forms, but primarily, all pr ovided some type of alternate instruction either as a separate course program or an integral program of study. The Veterans Adjustment Act of 1944, providing educatio nal monies for returning World War II veterans, created a new resurgence in pr oviding preparatory programs in colleges and universities. However, it wa snt until the 1960s that remedi al education finally became a larger component of higher education as increasing numbers of students enrolled in higher education. During the twentieth centu ry, junior colleges became the predominant provider of remedial education, although most four-year schools kept vestigial programs. The Higher Education Act of 1965 establis hed a philosophy of open admissions by providing financial aid, special services, a nd incentives for minority recruitment, resulting in increased nu mbers of underprepared students. By 1977, over 80% of colleges and universities offered some sort of college preparatory program (Roueche & Snow cited in Boylan, 1988, p. 3). Throughout American postsecondary education 19

PAGE 31

history, a consistent 80% of colleges and universities have met the needs of underprepared students. In the 1980s, legislatively mandated assessments began. Most states found about 30% of entering students were deficient in at least one basic skill. By 1985, over 90% of community colleges used placement tests; few were used as barriers to entry (Cohen & Brawer, 1996). In the 1990s, approximately 42% of high school students enrolled in college, and of that figure, 29% were enrolled in at least one remedi al course. At public 2year institutions, 41% of first-time freshmen enrolled in one or more remedial courses (National Center for Education Statistics, 1995, p. 3). NCES Fall 2000 Study The latest study conducted through NCES Postseconda ry Education Quick Information System (PEQIS) provided freshman enrollment statistics for Fall 2000 in Title IV degree-granting institutions as well as a ny changes in remediation from the Fall 1995 study. Between 1995-2000, the study did not find any significant change in enrollment of entering freshman in at least one remedial course (2003, p. iv). The National Center for Education Statistics (2003, p.18) also reported in fall 2000: 28% of entering freshman of all ages in all types of degree-gr anting institutions enrolled in remedial coursework. 22% of entering freshman of all ages in all types of degree-g ranting institutions enrolled in remedial mathematics, 14% in writing, and 11% in reading. 20

PAGE 32

Among two-year colleges, 20% of entering fr eshman of all ages were enrolled in remedial reading courses at public institutions compared to 9% at private institutions. Among four-year colleges, 6% of entering fr eshman of all ages were enrolled in remedial reading courses at public institutions compared to only 5% at private institutions. 98% of public two-year colleges offered co llege-level remedial courses compared to 59% to 80% of other types of institutions. Public 4-year institutions were significant providers of remedial education (80% vs 59%) compared to private 4-year institutions. It should be noted that for reporting purposes, private-for-profit institutions are included in the data for private not-for profit instituti ons since there are few private for-profit institutions in the sample. Comparison of the NCES remediation studies for Fall 1995 and Fall 2000 indicates no difference in the proportion of freshman enrolled in at least one remedial course; however, there was an increase in re strictions colleges have placed upon students in taking regular courses while enrolled in remedial courses. This type of policy limits access to federal financial aid because the Higher Education Act of 1965 was amended so that students may not be eligible for financial aid if they are solely enrolled in remedial courses or if remediation exceeds one year (NCES, 2003, p. iv). The proportion of institutions reporting more than one year in remedial courses increased from 33 to 40% in the Fall 1995 and Fall 2000 studies, respectively (NCES, 2003, p. iv). 21

PAGE 33

Demographics of Community College Students The public community college open admi ssion policies have resulted in more diverse demographic student popul ations. Students of color ar e provided access to higher education because of remedial programs. Th e majority of devel opmental students are white (Boylan et al. cited in Boylan, 1999). Less than one-third are minorities with African-Americans representing the largest gr oup followed by Hispanics. Between 52% to 57% are women; moreover, over 80% are U.S. citizens. Non-citi zens participate in developmental reading and writing to attain the skills required to become citizens (Knopp cited in Boylan, 1999). One of five is married and two out of five receive some form of financial aid and almost one in ten is a veteran (Knopp, 1996, p. 3). Also one in three works 35 hours or more per week. According to the National Study of Developmental Education, almost three in five are 24 year s old or younger with age ranges from 16 to 60 years old (Boylan et al. c ited in Boylan, 1999, p. 3). As of 2004, the average age of a community college student in the United States dropped to 29.7. The fastest growing categorie s were students less than 25, increasing overall by 25% in five years. From 1998 to 2004, the American Indian and white students decreased, 13.9 % and 3.3% respectively, while African-American and Hispanic increased, 25.7% and 47.9% respectively. The la rgest percentage increase was in the not reported category (580%) because upgraded soft ware allows students to self-select and the current race/ethnicity options do not pe rmit mixed backgrounds. Another remarkable trend is a greater number of student s not reporting gender (Armstrong, 1999). 22

PAGE 34

Academic Background of Developmental Students Research has also been conducted on the academic background of developmental students. Students are identified as underp repared for college by SAT and ACT tests as well as institutional assessment instruments. Nationally, developmental students fall into the bottom half of the score distributions (Boylan, 1999). The American Council on Education states there are exceptions to this finding because 18% of those taking remedial courses have SAT scores around 1000, while 5% have scores above 1200 (Knopp, 1996. p 4). The mean cumulative high school grade point average (GPA) for entering remedial students into community colleges was 2.40 and upon completion from a two-year institution, developmental students attain ed a grade point average of 2.28 (Saxon & Boylan, 1999, p. 6). The NCES (2003) six-y ear study found a cumulative GPA of 2.42 for women who enrolled in developmental courses compared to 2.84 for those not enrolled. In Florida, the Community College System tracks th e performance of its students in the State University System and the mean cumulative GPA has remained stable for both former community college Associate of Arts degree students and state university natives. Florida community college student GPAs have ranged from 2.86 in 1994-95 to 2.97 in 2002-03, and state university student GPAs impr oved from 2.92 in 1994-95 to 3.03 in 2002-03 (p. 2). National retention rates for first time enrolled students in developmental courses or programs were higher than the population as a whole. Females are slightly overrepresented in the group successfully comp leting all remedial/developmental courses 23

PAGE 35

attempted. At two-year instit utions, 74% of students remained in school at least one year; 67% remained in school at four-year institutions (Boyl an, Bonham & Bliss, 1994). Surprisingly, the majority did not leave due to academic standing but for personal reasons. Indicators of Success for Developmental Students Several studies comparing persis tence rates and grade point averages of developmental and non-developmental student s have been conducted. For example, Sinclair Community College (1994) found developmental stude nts had higher persistence rates and slightly lower GPAs than non-deve lopmental students. Persistence rates might be higher due to students having to stay in college longer since college preparatory courses delay them from taking college-level courses or pers istence may be coupled with feeling more prepared to con tinue (Walleri, 1987). Kulik, Kulik, and Shwalb (1983) also found remedial programs were related to impr oved persistence and grade point average; in other words, students placed in remedial tracks persisted longer than students who were not placed in remedial tr acks, but as they stayed in school longer trying to overcome basic skill deficiencies, grad e point averages declined. As a group, developmental students attempt courses, persist longer but have lower GPAs and fail/withdraw at a higher rate than college-level students (Weissmann et al., 1997). The Illinois Community College Board reported remedial students ha d higher persistence rates, yet the more remedial courses that a student was required to enroll in, the lower his completion rate (ICCB, 1998). Minnesota community colleges f ound that persistence rates were higher for developmental course takers than stude nts who failed to enroll in developmental education. Furthermore, persistence rates of developmental course takers were higher 24

PAGE 36

than, or not that different from, those of college-prepared studen ts (Shoenecker, 1996). Michigans State Department of Education, analyzing deve lopmental education at its associate degree-granting institutions, found developmental students remained enrolled longer, but non-developmental students had higher GPAs and completed more of the credits attempted (1999). The Michig an study also found students who passed developmental courses were more likely to pa ss college-level courses. An NCES study (2004) reported 45% of remedial/developmenta l students were identified as persisters, compared to a little over one-third of non -remedial/developmental students, who were identified as persisters. A si gnificant finding in relation to pe rsistence rates was that 9.3% of the remedial/developmental students were still in school at the end of the study compared to only 3.9% who needed remedi ation and did not seek remedial help-indicating the need for remediation doe s have a negative impact on time to degree and a positive one with persistence. Successful remedial/developmental students were more likely to graduate (4.7%) th an those who did not complete all remedial/developmental courses attempted. Mandatory Placement Testing Persistence rates and GPA correlations to persistence rates are not the only kinds of studies conducted to measure successful developmental programs. Several studies focused on mandatory placement and the seque nce of developmental courses prior to enrollment in college-level classes. Fo r instance, the Minnesota study, which included the entire population enrolled in its comm unity colleges, found students who completed the developmental course sequence achieved si gnificantly higher ratios of credits earned 25

PAGE 37

to attempted, higher cumulative GPAs, and hi gher persistence rates than developmental students who did not take the recommended sequence. The placement test was the College Board's Descriptive Tests of Language and Mathematics Skills. Students who earned a "D" or who failed a developmental course were excluded from the study since it was assumed treatment had not been administ ered. More than 22% of the community college students needed developmental readi ng, yet only 17.4% were pl aced into a course at those colleges offering reading program s (only 13 out of 22 Minnesota community colleges had reading programs in 1988). Moreover, 2.7 % of the developmental students were permitted to choose between a developmental and a college reading course. The strongest indicator of success in the study was in the reading content area, and therefore, it was recommended that efforts to increase co mpliance in reading course enrollment be undertaken as well as mandatory early comple tion of all developmental course work. In fact, developmental educati on improved the success of underprepared students, so their performance was indistinguishable from that of college-prepared students. The Academic Senate for California Community Colleges conducted a basic skills survey in 1998 and concurred with the Mi nnesota findings that many students are not retained after assessment for basic skills in struction because they never enroll in the developmental course sequence; more than half of Californias community college freshman needed basic skills courses, but only 29% actually enrolled in basic skills courses (2000). According to th e California Chancellors Office Fact Book (p.46) less than 25% of basic skill students showed any improvement in basic skills in a three-year 26

PAGE 38

period (1995-1998). California community co lleges have no research on why this percentage is so low. Studies conducted by Johns on County Community College in Overland Park, Kansas, showed a high correlation between ma ndatory placement testing in reading and English, and overall student academic succe ss as well (Amey & Long, 1998). Mandatory placement testing, followed by requiring developmental students to complete development coursework prior to enrolling in college-level wor k, is now required by many community colleges The National Center for Education Statistics examined the high school Class of 1982 college transcri pts with degrees earned by 1993. The academic careers of 2.45 million students in more than 2500 institutions were analyzed. Of the students who had earned more than a semester of college credit by 1993, 55% who did not take any remedial courses, and 47% who took only one remedial course, earned bachelor's degrees (Adelman cited in McCusker, 1999, p. 1). However, only 24% who took three or more remedial cour ses earned bachelor's degrees. NCES Fall 2004 Study The latest NCES (2004) study reports the proportion of students requiring remedial reading who earned no postsecondary credentials rose from 57% to 70%; whereas, the proportion of students re quiring remedial math who earned no postsecondary credentials rose from 49% to 58% (NCES, 2004, p. 94). Furthermore, 16.6% of those needing only remedial reading obtained a bachelors degree. Reading deficiency is an indicator of lower odds in completing any degree (Adelman cited in McCusker, 1999, p. 1). 27

PAGE 39

These studies provide suffi cient evidence to suggest that students should be required to remediate, they should not delay th e basic skills courses, and those deficient in two or three basic skill areas should not be permitted to take college-level coursework (Weissman, Silk & Bulakowski, 1997). Weis sman, Silk and Bulakowski had good, hard data to support the fact that developmental education coursework needs to be mandatory and those with serious skill improvement needs should be enrolled in developmental coursework their first semester. This fi nding, based on their data, was the strongest policy recommendation from their study. Colle ges need to regularly evaluate the educational attainment of remedial stude nts to improve policies and programs to maximize student success since reme diation is a growth industry. Illich and McCallister (2004) conduc ted a study in Texas at McLennan Community College to examine the practice of allowing students to concurrently enroll in remedial and college-level courses. Their fi ndings showed students concurrently enrolled in remedial and college-level courses under-per form in the college-lev el classes compared to students who are only enrolled in college-lev el courses. This effect, however, is limited to only those students who do not successfully co mplete their remedial courses. Students concurrently enrolled who successfully passed their remedial studies performed as well in their college-level courses as did students who only enrolled in coll ege-level courses (p. 448). In a national survey on remedial educa tion in community colleges, Lewis and Farris (1996) concluded that only 2% of the instit utions did not permit st udents taking collegelevel courses concurrently with remedial clas ses (cited in Illich & McCallister, p. 437). 28

PAGE 40

Due to a lawsuit settlement, the State Chancellors Office of the California Community Colleges implemented a series of regulations governing the use of placement tests. Colleges were required to gather and report on the predictive validity of the tests. Essentially, colleges had to prove that using the tests to group and place students led to their likelihood of success in a course. Ar mstrong from San Diego Community College produced a model to explain th e variance in course outcomes using test scores, student background data, and instructor differences in grading practices (2000). Armstrong found student dispositional characteristics explai n the high proportion of variance in the dependent variables and instructor gradi ng practices make accurate placement more difficult. Dispositional factors included affec tive, behavioral, and cognitive traits, such as past experiences or performance in school, i nvolvement in school activities, high school GPA, high school preparation, and percei ved importance of attending school. Not surprising, the key dispositional factors that were most significant were high school GPA, course load in math and English, and grade in last high school math or English class. A statistically significant relationship existed be tween course grade and the placement test but not enough to have practical significance. Among full-time instructors, placement test scores were not significantly predictive for fi nal grades. However, entering the instructors characteristics (grading policy) accounted for the greatest amount of variance in final grade% to 20% (Armstrong, p. 690). Dr. Edward Behrman at National University contends the amount of variance in the cour se grade accounted for by scores on contentgeneral reading testsmay be too low to wa rrant the continued use of these tests to predict success in a particular course ( 2006, p.42), or at the very least, refining the 29

PAGE 41

placement test so that it becomes a better pred ictor of success is a better choice than no placement test. Linda Suciu examined CPT scores and information from questionnaires at Trident Community College in South Carolina to pr edict success in an introductory mathematics course (1991). By choosing a score to maxi mize a correct predictiona score above the cut-off for passing, and a cut-off score be low for failingher prediction of student success ranged from 55% to 72%. Suciu then selected a range of scores which improved the chance of correctly predicting success by leaving only 5.2% to 16.6% of the students in various mathematics courses incorrectly predicted. One notable finding was that completion of assignments and amount of prac tice correlated positively with success in all the courses. In addition, stude nts 25 years old and older were more frequently successful in developmental mathematics courses than those students under 25. According to Suciu, Cut-off scores should be chosen in such a way that those predicted to succeed do succeed, while those who are predicted to fail actually fail(Suciu, 1991, p. 6). Florida Research in Developmental Education A number of studies focusing on compone nts contributing to student success in developmental studies, particularly mathematics, have also been conducted in Florida, the site of this proposed study, ye t none of the studies addressed looking at particular cut-off scores on placement tests in reading as an essential indicator for successful completion of the highest level college prep aratory reading course. Margaret Cran (1998) examined the correla tions between studen t presage variables and performance on the mathematics subtest of the Florida College Level Academic Skills 30

PAGE 42

Test (CLAST). Cran used data which incl uded the CLAST mathematics subtest scores, entry level mathematics subtest scores, gender, racial-ethnic, and si x entry-level test variables (including the CPT) for 4,139 first-time CLAST examinees from the State of Florida database. Cran hypothesized if student presage data could indicate success or failure on the CLAST, the community coll eges could identify at-risk students immediately. The strongest rela tionship was between entry level mathematics scores and CLAST successa coefficient of determin ation was 30%. Negative correlations were found for gender and race/ethnicity (p. viii). Wendy Bush (2001) also examined the rela tionship of student characteristics to determine if they affected the prediction of st udent failure in the first college preparatory mathematics courses at a community college The six factors were high school GPA, gender, ethnicity, CPT scores, enrollment st atus, and financial aid status. The only significant factor was gender for the pre-algebra course, with females being less likely to fail. And high school GPA and ethnicity were significant predictive variables of failure for Elementary Algebra. Linda Clemons conducted a study to predict community college student performance on the Florida Basic Skills Exit Test (FBSET) in elementary algebra in a collaborative instruction environment. Clem ons explored whether the predictive qualities of the following variables: math anxiety sc ores, perceived usefulness of mathematics, college placement test scores (CPT), and pa ssing elementary alge bra during collaborative instruction could predict eligibility to take th e FBSET. Then the study sought to ascertain if any of the variables or a combination of the variables could predict the FBSET score. 31

PAGE 43

Finally, she also wanted to determine if ge nder had a role in determining the outcome on the exit test. The study involved only one Florida community college examining 238 students who had CPT scores and enrolled fo r the first time in elementary algebra in spring 2001. Clemons data revealed only the CPT score predicted exit test scores; the average CPT scores of completers of elemen tary algebra were higher than those who did not successfully complete elementary algebra (p. viii). Pat Smittle designed a study in 1995 to identify predictors of academic performance at Santa Fe Community College in Florida. College academic performance was determined by college GPA at the end of the first college year. Smittle found the strongest relationship with college GPA was the high school GPA (.52). Data revealed a difference in CPT scores, overall high school GPA, and senior year absences for students with higher college GPAs and students with lower colleg e GPAs. Only 23% of the students with high school GPAs below 2.0 earned college GPAs of 2.0 or higher. High school GPA accounted for 13% of the variance; whereas, senior year absences accounted for 15% of the college GPA variance (p. 4) Neither race nor ge nder was a significant predictor of college GPA. Transcripts of Florida high school students were analyzed by Jeffrey Roth (2001) to determine if course choice, course load, grades in math and English, overall GPA, the tenth grade standardized test score (GTAT) in math and reading, race, and gender affected performance on the CPT upon entry to commun ity colleges in the fall of 1994. Roth created a High School Performance variable fo r math and English classes to account for the differences in the number of courses comp leted, their difficulty level, and course final 32

PAGE 44

grade. The Math High School Performance variable had a larger positive effect on passing the CPT math subtest than GPA or tent h grade scores; whereas, tenth grade scores had the larger effect on CPT reading and writ ing subtests. Finding that Math High School Performance variable has the larger eff ect on passing the CPT suggests high school students need to take more challenging math cour ses, even at the risk of lowering GPAs. In addition, finding that tenth grade scores on the GTAT is the strongest predictor of success in passing the CPT reading and writing s ubtests may indicate it can be used to predict unpreparedness. However, when cont rolling for English High School Performance variable, tenth grade scores, and GPA, Blacks and Hispanics did not pass the CPT reading and writing subtests at the sa me rate as whites. Students taking similar math coursework revealed no racial differences in passing the CPT math subtest. A recent developmental education survey conducted by the Florida Division of Community Colleges and Workforce Educati on revealed that there is no consensus among Florida community colleges on current CPT cut-off score ranges. In addition, there is currently no consensus on policies or practices regarding the college-preparatory exit exam. Neither the administration of the ex it exam or cut-off scores for passing are standardized. However, sixteen of the twenty -eight community colleges stated students needed a C or better to sit for the fi nal exam. Furthermore, only eight community colleges examined subsequent college-level course success and six community colleges are currently initiating or examining tracking mechanisms. The Florida Division of Community College s and Workforce Education Office of Student and Academic Success has provided statistics for a program review of 33

PAGE 45

developmental education using the stude nt database for Fall 2000-01 through 2002-03 (see Table 1). Blacks failed the CPT at a higher percentage (8 4.11%) than Hispanics (72.24%) and Whites (58.66%); moreover, Bl acks had lower passing rates (69.29%) in the highest level college-pre paratory course in reading than Hispanics and Whites (70.56% and 75.96%, respectively). However, th e three major ethnic groups combined still revealed a failure rate of approximately 28% in the highest level college preparatory reading course. The cohort Other had the best success rate, which may suggest that the ethnic choices on student applications need to be more refined and updated to represent current demographic trends, so th at more accurate analysis of student data can be applied. Table 1 First Time in College (FTIC) Degree-Seeking Students Taking Entry Level Test: College Preparatory Success Report by Ethnicity Cohort by Ethnicity Total Cohort Number Failed Entry Level Test % Failed Entry Level Test Failed Reading Subtest Enrolled Any Level Reading Passed Highest Level % Passed Highest Level Reading Blacks Non-Hispanic 6,778 5,701 84.11 4,416 3,579 2,480 69.29 Hispanics 6,818 4,925 72.24 3,253 2,490 1,757 70.56 Whites 24,869 14,588 58.66 7,421 5,467 4,153 75.96 Asian/Pacific Islands 1,121 746 66.55 574 402 314 78.11 American Indian/Alaskan 165 108 65.45 62 47 32 68.09 Other 450 316 70.22 223 160 132 82.50 Total Cohort 40,201 26,384 65.63 15,949 12,145 8,868 73.02 Note. Grades of A, B, C, S, and P are cons idered passing for Fall 2000-2001 and Fall 2002-2003 Florida database. 34

PAGE 46

A major indicator of success for community colleges is the awarding of degrees. Community colleges award various degrees, which include the associate degree, the associate in applied science, college credit certificate, the app lied technology diploma, and the postsecondary adult voc ational certif icate, which is non-college credit for occupational training. The Florida student database revealed that only 34.6% (466 out of 1, 346) of students who needed only college-preparato ry reading courses obtained a degree (see Table 2), compared to 40.7% (4,119 out of 10,114) of students who were college-ready. Students needing all three areas of remedia tion had the lowest percentage (9.9%) in obtaining a community college degree. Table 2 Enrollment of College Preparatory Student s by Areas Required and Awards Earned System Original Cohort Awards Total Number Percent Number Earned Rate College Ready 10,114 28.20 4,119 40.70 Need only Math 7,726 21.60 1,623 21.00 Reading 1,346 3.80 466 34.60 Writing 541 1.50 167 30.90 Need Reading and Math 4,114 11.50 639 15.50 Reading and Writing 1,318 3.70 322 24.40 Math and Writing 1,735 4.80 274 15.80 Need all three areas 8,930 24.90 884 9.90 Needing any remediation 25,710 71.80 4,375 17.00 Total Cohort 35,824 100.00 8,494 23.70 Note. First time in college (FTIC) with Complete Placement Scores and Florida database 19992000 through 2003-2004. 35

PAGE 47

The following tables provide inform ation on the success of obtaining a community college degree by ethnicity. In 1999, only 12.9% of FTIC African Americans who had complete placement scores were college ready; whereas, 87.1% needed some type of remediation (see Table 3). Only 6.8% of college ready African Americans obtained a community college degree, and onl y 3.9% needing any type of remediation obtained a degree. African Americans only n eeding remediation in reading were the highest percentage of degree earners (11.8%), but the per centages are misleading when the original numbers are cons idered (e.g. 11.8% represents on ly 27 students). In addition, those needing all three areas of remediation had the lowest percentage for obtaining a degree (2.1%). Table 3 Awards (Degrees) Earned by African Americans Remediation Original Cohort Awards Earned Number Percent Number Rate College Ready 795 12.90 54 6.80 Need only Math 825 13.30 49 5.90 Reading 229 3.70 27 11.80 Writing 76 1.20 7 9.20 Need Reading and Math 900 14.60 36 4.00 Reading and Writing 284 4.60 22 7.70 Math and Writing 263 4.30 10 3.80 Need all three areas 2,810 45.50 59 2.10 Needing any remediation 5,387 87.10 210 3.90 Total Cohort 6,182 100.00 264 4.30 Note. Fall 1999 First time in college (F TIC) with Complete Placement Scores and Florida database 19992000 through 2003-2004. 36

PAGE 48

A larger percentage (22.5%) of FTIC Hispanics who had complete placement scores were college ready as compared to African Americans, but this percentage represents a student populat ion which is almost double that of African Americans. Hispanics needing any type of remediat ion (77.5%) resulted in only 5.9% obtaining community college degrees (see Table 4). On ly 11.2% of the college ready Hispanics obtained a community college degree, and onl y 5.9% needing any form of remediation obtained a degree. Hispanics only needing re mediation in reading were the highest percentage of degree earners (12.1%). In addition, those needing all three areas of remediation had the lowest percenta ge for obtaining degrees (3.3%). Table 4 Awards (Degrees) Earned by Hispanics Remediation Original Cohort Awards Earned Number Percent Number Rate College Ready 1,390 22.50 156 11.20 Need only Math 1,164 18.80 90 7.70 Reading 280 4.50 34 12.10 Writing 99 1.60 11 11.00 Need Reading and Math 804 13.00 48 6.00 Reading and Writing 298 4.80 23 7.70 Math and Writing 262 4.20 12 4.60 Need all three areas 1,893 30.60 63 3.30 Needing any remediation 4,800 77.50 281 5.90 Total Cohort 6,190 100.00 437 7.10 Note. Fall 1999 First time in college with complete placement scores and database 1999-2000 through 2003-2004. 37

PAGE 49

In Table 5, Whites needing any type of remediation (65.7%) resulted in only 5.7% obtaining community college degrees, which was similar to the percentage of Hispanics (5.9%). Whites needing only remediation in reading resulted in 10.1% receiving awards, while those needing both read ing and mathematics resulted in 5.9% obtaining a degree. In addition, those needing all three areas of remediation ha d the lowest percentage for obtaining degrees (3.5%). Table 5 Awards (Degrees) Earned by Whites Remediation Original Cohort Awards Earned Number Percent Number Rate College Ready 7,523 34.30 712 9.50 Need only Math 5,512 25.10 355 6.40 Reading 724 3.30 73 10.10 Writing 334 1.50 36 10.80 Need Reading and Math 2,262 10.30 133 5.90 Reading and Writing 572 2.60 37 6.50 Math and Writing 1,161 5.30 50 4.30 Need all three areas 3,832 17.50 133 3.5 Needing any remediation 14,397 65.70 817 5.70 Total Cohort 21,920 100.00 1,529 7.00 Note. Fall 1999 First time in college (FTIC) with Complete Place ment Scores and Florida database 1999-2000 through 2003-2004. In Table 6, comparisons are made with college preparatory students by ethnicity that only needed reading and obtained a co mmunity college degree with college ready students by ethnicity who obtai ned a degree. Table 6 was cr eated by combining Tables 3, 4, and 5 to provide information on the tota l of the three ethnic groups, which was not provided by The Division of Community Colleges and Work force Education Office of 38

PAGE 50

Student and Academic Success. Examining the results when all three ethnic groups are combined reveals very few students who only need college preparat ory reading earn a community college degree. Table 6 Comparison of Awards Earned by College Ready and College Preparatory Reading Students FTIC degree seeking taking Entry Level Test College-Ready Reading Only Awards Earned With Reading CollegeReady Awards Earned Black/ NonHispanic 795 (12.90%) 229 (3.70%) 27 (11.80%) 54 (6.80%) Hispanic 1,390 (22.50%) 280 (4.50%) 34 (12.10%) 156 (11.20%) White 7,523 (34.30%) 724 (3.30%) 73 (10.10%) 712 (9.50%) Total of 3 ethnic groups 9,708 (28.00%) 1,233 (3.50%) 134 (10.80%) 922 (9.40%) Total Cohort 34,292 (~100%) Associate Vice Chancellor for Evaluation Dr. Patricia Windham of the Division of Community Colleges and Workforce Educ ation provided analysis of the degrees earned by college preparatory students within various ranges of CPT reading scaled scores. The CPT reading scaled score of 83 means a student is exempt from taking a college-preparatory course in reading. Co mbining the scaled scores from 83 to 120 (11,601 out of 27, 626 students), the percentage of students passing the CPT-R is 41.9%. The remaining two ranges are students (58%) who are required to take one or more courses in college-preparato ry reading. And within thos e two ranges, only 14.8% of 39

PAGE 51

students obtained a degree. Students who were exempt from taking reading had the highest percentage (22.4%) of earning a degree. Entering calcu lations for the total cohort finds the overall percentage of students at taining community college degrees was 18% while 27% were still enrolled at a univers ity and 40.3% were employed, which is defined by the Florida Department of Education as academically successful (see Table 7). The success rates are calculated through a form ula which includes the total number of students, the number of stude nts who graduated, were still enrolled in good standing, or left in good standing. Table 7 Fall 1997 First Time in College Freshman CPT-Reading Scaled Scores Scaled Number Awards Earned Tr ansfer to Awards or Still Scores Students SUS Transfer Enrolled Success % # % # % # % # % # % 20-50 3,513 12.7 347 9.9 219 6.2 428 12.2 967 27.5 1,192 33.9 51-82 12,512 45.3 2,030 16.2 1,370 10.9 2,389 19.1 3,551 28.4 4,991 39.9 83100 8,714 31.5 1,908 21.9 1,264 14.5 2,177 25.0 2,259 25.9 3,728 42.8 101120 2,887 10.5 691 23.9 420 14.5 777 26.9 696 24.1 Note. First Time in College (FTIC) Student Database for Fall 1997-1998 through 2001-2002 and various Performance Based Program Budgeting files created from the State University System (SUS) St udent Data Course File. Success is measured by employment, based on fall 2001 follow-up of the Original Cohor t by Florida Education and Training Placement Information Program. 1,244 43.1 Total 27,626 100.0 4,976 18.0 3,273 11.8 5,771 20.9 7,473 27.0 11,155 40.3 Cooling-Out Function of Community Colleges Burton Clarks institutional case study of San Jose Junior College, The Open Door College, was instrumental in promp ting researchers to examine the long-term educational attainments of community colleg e entrants (Diel, 2001) Clark (1960) stated that the junior colleges provided a cooli ng-out function in wh ich counseling, testing, 40

PAGE 52

and other policies are devices to subtly convince incompete nt students who wish to transfer to give up their or iginal goals and pursue an alternative terminal vocational program; thus, the community college served as a screening device, a gatekeeper, in effect, for 4-year institutions. Brint and Ka rabel (1989) concurred with Clark by detailing the ways occupational and vocational programs have expanded in community colleges to hinder the transfer function of the community college and subsequently encourage students to opt out of a baccalaureate transfer program of study. Several studies in the 1970s and 1980s revealed that merely being at a community college ra ther than a 4-year institution reduces the probability that a st udent will obtain a bachelors degree (Deil, 2001). However, many community colleges today are not barriers to student success. Deils (2001) research findings indicate a warming-up pattern among students attending community colleges. Faculty who are committed to the transfer mission of the college plus support systems such as tutoring and small class size provide positive support towards attainment of a bachelors degree. Deil claimed the community college thus provides pockets of oppor tunity since the priority is transfer; consequently, the student defines success as the attainment of a bachelors degree (Deil, 2001, p. 7).Thus, Deil recommended a study should be conducted to evaluate placement test scores as indicators for successful completion of college preparatory courses to ascertain whether or not low scores are a fundamental barrier to a student pursuing a two-year or four-year degree. 41

PAGE 53

Summary In 1998-99, there were 76,960 co mmunity college students enrolled as first-timein-college (FTIC). By 2003-04, FTIC student s increased to 102,201an increase of 33% (Armstrong, 2004, p.6). In addition, more st udents are in the 17-24 years old range, which is an increase of 25% in the last five years (Armstrong, 2004, p. 1). Although the exact percentages vary, slightly one-third of FTIC are co llege-ready, another one-third need one remedial course, and the final th ird need two or more remedial courses (Armstrong, 1999 p. 1). The NCES (2004) st udy indicated only 16% of students who were assigned remedial reading courses comp leted bachelors degr ees, compared to 58% of students who were not required to take any remedial courses. However, Florida students who were assigned remedial reading courses completed bachelors degrees at a higher percentage (17.5%) (Armstrong, 2005) Furthermore, Florida reported a significantly lower percentage (25.4%) of stude nts obtaining bachelors degrees than the national percentage reported in the NCES study (58%). The Florida Department of Educatio n released information on the 2000-2001 cohort of FTIC degree-seeking students who failed the entry test in reading, writing or mathematics. A majority (65.63%) of student s failed at least one entry level test. Comparing the success rates for remedial reading stude nts in attaining a two-year or fouryear degree to those not needing remediation in reading suggests that the placement test should not be used only as crit eria for placement into developmental courses, but also as a screening device to permit admission counselor s to advise students with very low basic skill levels of other career choices that do not require traditional college programs of 42

PAGE 54

study. Studies of persistence rates and course completion rates provide only a snapshot of community college students who were succes sful. This study examined variables at the beginning of a program of study to identify wh ether or not a student should even seek a college degree. 43

PAGE 55

Chapter 3 Methods With student enrollments in higher e ducation increasing throughout Florida, traditional four-year institutions are turning away students who are college level ready, resulting in community colle ges facing unprecedented enrollments not only in traditional college classes, but also remedial instruction. However, an ethical dilemma exists should community colleges continue to plac e underprepared students into remediation knowing the negative effects on retention rate s and matriculation while turning away students who are college-ready? Should they welcome students who are seriously underprepared and not inform them that th eir time and money may be expended with very little possibility of ever graduating? As described in chapter two, reading has been found to be the primary indicator of successfully completing a colleg e program of study; therefore, the purpose of this study was to examine the predictive validity of seve ral variables to determine if the Florida Computerized Placement Test Reading (CPT-R ) score alone, or other variables, could determine whether or not a student would successfully pass the highest level college preparatory reading course. The study examined only the reading scores on the CPT to determine at what standard deviation below the cutoff score of 83 a student could still successfully complete the highest level college preparatory reading course. Smittle 44

PAGE 56

(1993) studied both the concurrent and predic tive validity of the CPT-R and was able to suggest a cut-off score of 83 on the CPT-R, representing college-reading ability, the score at which a student would be exempt from taking the highest preparat ory reading course. Concurrent research suggests the score of 83 equates to a 12 th grade reading level on the Nelson Denny Reading Test as well as th e Directed Reading Program (Napoli & Raymond, 1998, p. 3). For the community college s open-door policy to be effective, reliable placement and diagnostic procedur es need to be employed to identify and determine student needs. Research Questions and Hypotheses Therefore, this study attempted to answer the following questions with the hypotheses tested at the 05 level of significance. 1. Is there a relationship between a student s score on the Computerized Placement Test in reading (CPT-R) and success in passing the highest level college preparatory reading course in Florida? 2. Is there a relationship between full-ti me or part-time enrollment during the semester a student is taking the highest level college preparatory reading course and success in passing the highest level co llege preparatory reading course in Florida? 3. What are students' GPAs the session following successful completion of the highest level college preparatory readi ng course according to the program track (Associate of Arts, Associate of Scien ce, Associate of Applied Science)? 45

PAGE 57

The null hypothesis of this research study for question one is that there was no relationship between the CPT-R score and a stud ent successfully pass ing the highest level college preparatory reading c ourse. Likewise, the null hypoth eses for the variables fulltime enrollment and part-time enrollment, was there was no relationship between these variables and a student successfully passing th e highest level college preparatory reading course. A student who was enrolled in twelve semester hours in the fa ll or spring sessions was a full-time student. A student who was enrolled in less than twelve semester hours was a part-time student. The null hypothe sis for question three was there was no relationship between the GPA and the students successful completion of the highest level college preparatory course in reading, according to program track. Successful completion, or cour se satisfaction, of the college preparatory reading course is the final course mark that is considered passing. For purposes of this study, pass/fail will be used as the variable representing passing grades of A, B, C, S, or P. A grade of W means a withdrawal from a course and is not computed in the GPA. A grade of W does not override a grad e of F. The instruct or may also enter an I, which is an incomplete for a course, and an I received at the end of any term becomes an F if not completed the succeedi ng fall or spring term. The student may not register for another section of the course dur ing the period of the incomplete grade. A grade of N is used only in college prepar atory courses and may be assigned to students earning a D or F in a colle ge preparatory course. The grade of N is non-punitive, indicating progress has been made but not at the level required for successful completion of the course. College-preparatory course s are not computed in a students GPA. 46

PAGE 58

Procedures Analysis of archived student scores for fall sessions 1997-2004, approximately 35,000 scores per year ( n = 276,079) for all forms of the CPT-R were carried out to determine the success of students whose exit sc ores (i.e. 81, 82) were clustered around the 83 cut-off score as well as how many standard deviations belo w the cut-off score a student could successfully pass the highest level co llege reading preparatory course. Successful completion of the highest level college preparatory reading course included students who passed during the fall session of each year in the study. Other variables such as full-time or part-time enrollment and the student s GPA the semester following successful completion of the highest level college-preparatory reading course were evaluated to determine if any relationship exists with pa ssing the highest leve l preparatory reading course. Participants/Data Collection The sample for the study was first-ti me enrolled Florida community college freshman. Variables included CPT-R scores enrollment status, the students GPA the semester following successful completion of th e highest level college preparatory reading course, and the program tracks A.A. (Associate of Arts), A.S. (Associate of Science) and A.A.S. (Associate of Applied Science) of th e students passing the highest level college preparatory reading course from the Florid a Student Database, fa ll sessions 1997-2004. Variables In the first analysis, CPT-R scores were the independent, con tinuous variable and the final grade in the highest level college pr eparatory reading course was the dependent, 47

PAGE 59

categorical variable. In the second anal ysis, enrollment status was an independent, categorical variable and the final grade in the highest level college preparatory reading course was the dependent, categorical vari able. Enrollment status was coded as a dichotomous variable using a for full-time and a for part-time enrollment. Passing or failing the highest level college preparatory course in reading was a dependent, categorical variable; passing th e course was coded as and failing the course was coded as . In the third anal ysis, the GPA the semester following successful completion of the highest level college prep aratory reading course was an dependent, continuous variable, and passing or failing the highest level college preparatory reading course was an independent categorical variable along with the three program tracks of A.A. (Associate of Arts), A.S. (Associate of Science) and A.A.S. (Associate of Applied Science). Instrumentation The NCES Fall 2000 study revealed that 57 to 61% of all postsecondary institutions administer placement tests. In July, 1993, the Florida Department of Education solicited proposals for testing products for a common placement testing program. The College Entrance Examination Board won the contract in December, 1993. The initial contract required core placem ent tests in reading, writing and elementary algebra as well as additional tests for lower and higher level mathematics. The College Entrance Examination Board and the Educa tional Testing Service proposed using the Computerized Placement Test TM (CPT) which was part of the ACCUPLACER system. The computerized adaptive testing technique customizes tests according to each students 48

PAGE 60

ability. Each student is presented with a series of test questions at an appropriate level of difficulty for the students abilities, knowledge, and background. Easy and difficult questions are avoided, and accurate results ar e obtained with fewer questions and no time limit. Institutions that do not have computer testing labs are provided with written versions. The Florida postsecondary in stitutions implemented the test in July 1995, with permission to delay full-scale implementation fo r one year. The tests are used primarily in Floridas public community colleges. The State Board of Education established minimum passing scores for each subtest, permitting individual institutions to set higher passing scores. However, by June 30, 1997, all community colleges were required to adopt uniform standards. The standards incl ude a reading comprehension standard score of 83 or higher, which exempts a student from taking a developmental reading course. Ranges also exist within each subsection for placement into different levels of developmental reading, writing and mathema tics. Furthermore, in 1996, the Florida Legislature amended Section 240.117 of the Florida Statues to permit the common placement test to be administered to high school tenth-grade students. The results of the common placement tests are not reported the same way as statewide tests. An annual report is prepared to describe the number of students who are placed into developmental studies by instit ution and each students test scores are recorded in the database maintained by the Division of Community Colleges. In the fall of 1998, 30,063students who took the Florida CPT and subsequently enrolled in the highest level college pr eparatory reading course had an average CPT reading score of 63. 49

PAGE 61

CPT Predictive Validity The College Boards Computerized Pl acement Test (CPT) has high levels of statistical reliability as we ll as content and construct validity (Napoli & Raymond, 1998). Internal reliability (alpha=.90) and test-ret est reliability (r=.90) are both high (CEEB cited in Napoli & Raymond, 1998, p. 3). Clemons reported a .92 reliability estimate for the CPT (2002, p. 47). Content and construct validity means that the test contains a representative sample of items of what it purports to measurereading comprehension skills. CPT reading scores (CPT-R) have been found to be accurate in repeated tests and consistent across items (CEEB cited in Napoli & Raymond, 1998). The items on the reading subset were selected by reading specialists from a larger group of items presented by an advisory committee of expe rts in reading. The specialists defined the chosen set of items as representative of colle ge-level skills in reading, re sulting in the reading subtest having content and construct validity. The CPT-R is used nationw ide at 350 colleges and universities; however, the criterion-related validity has not been thoroughly examined. Criterion-related validity consists of concurrent validitythe degree to which scores on two or more subtests measure the same thingand predictive valid itythe degree to which scores predict performance (Anastasi cited in Napoli & Ra ymond, 1998). In this case, the CPT-Rs concurrent validity is measured by the degree to which scores on the test correlate with other tests measuring reading skills, a nd predictive validity is measured by how accurately the test predicts future reading pe rformance. Either concurrent validity or predictive validity (or both) determines the level of criterion-related validity. Criterion50

PAGE 62

related validity permits the test to be used as a reliable assessment tool for placement and curriculum decisions. Napoli examined the predictive-validity of the CPT-R by using overall college grade point average and performance in introductory psychology classes, which were used as criterion variables. Si gnificant correlations existed between CPT-R scores and course grades (r =.52) and between CPT-R scor es and overall grade point average (r = .41). Furthermore, the study wa s successful in identifying specific cutoffs on the CPT-R distribution as predictive of successful and unsuccessful academic outcomes (Napoli, 1998, p. 2). However, nor m-referenced tests which includes the CPTR, reveals little more than the relative positi on of each test-taker on the score distribution. In 1993, Pat Smittle studied both predictive and concurrent validity of the CPT subtests in reading comprehension, sentence skills, arithmetic and algebra against the ACT to establish the criterion-validity of each subtest. Smittle found the CPT tests were better predictors of ove rall academic performance in coll ege than the ACT tests. The CPT reading subtest was more discriminating among levels of reading competency than the ACTs composite reading placement test, thus establishing the CPT reading subtests concurrent validity with another norm-referen ced test. According to the College Entrance Examination Board the primary function of the CPT is to determine which course placements and whether or not students need remedial studies (CEEB cited in Smittle, 1995, p. 2). Smittle also was able to suggest a cut-off score on the reading subtest (83) which represented college-level reading ab ility and placed the same percentage of students at each course level as those pr eviously placed at those levels using the traditional paper and pencil tests (c ited in Napoli & Raymond, 1998). 51

PAGE 63

In 1995, Murphy examined the construct and predictive validity of the CPT reading subtest to the three subscores of the Nelson Denny Reading tests. Using a sample size of 663 college students, si gnificant correlations were found between the CPT-R and the Nelson Denny Vocabulary section (r = .69). The Nelson Denny provides gradeequivalents to reading scores. However, the grade-level score assignments still needed validation. Napoli and Raymond continued the assessmen t of the criterion-related validity of the CPT by examining the concurrent validity of the CPT-R and the Degrees of Reading Power (DRP). The studys goal was to create a grade-level equivalency table, allowing for the conversion of CPT-R scores to valid reading grade levels. The DRP test, through extensive studies, has demonstrated high leve ls of reliability (K R-20=.95) and construct validity and criterion-related validity (K oslin cited in Napoli & Raymond, 1998). The DRP scores are converted into grade specific readability levels from 4 th grade through 12 th grade and first-year college levels. Resu lts from the study f ound that a substantial correlation exists between the DRP and the CPT-R. The CPT-R has a high degree of reliability and validity to identify basic read ing proficiency skills necessary for first-year college-level textbooks. The study not only affirms Smittles previous college-level cut point of 83, but also equates specific CP T-R scores with expected grade-level performance. The Standard Error of Estimate predic ting DRP grade level performance from CPT-R is equal to [+-] 1.27. In addition, the study also replicated the CPT-R grade-level equivalencies produced in Murphys 1995 analys is of the Nelson Denny Reading Tests. 52

PAGE 64

Since generalizations resulting from this st udy are limited by the nature of the sample community college students scoresfurth er studies should be conducted before assuming these results can be used for other types of institutions Data Analysis Using the Statistical Analysis System (SAS) for this study, descriptive and inferential statistics were obtained. Descri ptive statistics included central tendency, (mean, mode, and median) variability (sta ndard deviation, variance, and range) and distribution (skewness and kurto sis) of the incoming freshman CPT-R scores, full-time or part-time enrollment, subsequent GPAs with success in passing the highest level collegepreparatory reading course. Relationships were examined using correlations (interval level or higher) for statisti cal significance between the i ndependent variable of CPT-R score, enrollment status and GPAs with the variable passing/failing the highest level college preparatory reading course. The cons ideration of how fina l grades are reported was not a concern since in order to pass the highest level preparat ory reading course, a student must not only pass the course based on instructors evaluations, but must also pass the Florida Basic Skills Exit Test; therefore, this dependent variable was entered as pass/fail. The statistical software progra m, SAS was used to generate frequency distributions and calculate means and standard errors for all quantitative variables. Passing or failing the highest level colle ge preparatory reading course based on placement test scores and enrollment status creates a binary response. The validity of interpretation of the results depended on the de sign of the study; th erefore, a logistic 53

PAGE 65

regression analysis was used. The odds of pa ssing the highest level college-preparatory reading course was expressed as: P (passing) odds = 1 P (not passing) The simple logistic regression equation with the independent variable X (full/part-time) is: logit (P) = a + bX The dependent logistic transformation of the odds, known as logit, is the dependent variable of passing or failing th e highest level college -preparatory reading course. The assumption is the relation between the logit (P) and X is linear. Similar to a simple linear regression, b is the expected change of log it (P) with a unit change in X. Therefore when b is positive, increases in X me ans increases in logits. When b is negative, increases in X means decreases in logits. There are three acceptable data formats fo r logistic regression; however, the raw data format using LOGISTIC procedure in SA S yielded the richest information for this study (Peng & So, 2002). Evaluations of the logistic regre ssion model included the overall model evaluations, statistical test s of individual predictors, goodness-of-fit statistics, and validations of predicted probabilities. The inferential sta tistics included the likelihood ratio, distributed as chi-square with degrees of freedom equal to the number of predictors ( df = 1). The descriptive stat istics included Akaike Info rmation Criterion and Schwarz Criterion to compare two differe nt models from the same sa mple. Statistical tests of 54

PAGE 66

individual predictors prov iding individual parameter estimates was tested by the likelihood ratio. The goodness-of -fit statistics assessed the fit of the logistic model against the data. The validations of predicte d probabilities determined to what degree predicted probabilities match w ith actual outcomes, using meas ures of association and/or a classification table. The measure of associ ation for this study was Somers D statistic and the c statistic. The classification table wa s a two-way classification table, which minimized the bias of using same observa tions in both model-fitting and predicting probabilities. A logistic regression model was used to test the null hypothesis because the criterion variable (passing or failing the highest level colleg e-preparatory reading course) was dichotomous instead of continuous. Furthe rmore, a logistic regression discerned the relationship between the criterion variable and multiple predictor variables (CPT-R scores, full-time enrollment and part-tim e enrollment), taken independently. The students GPA is a continuous interval variable. A logistic regression is valid with retrospective data -college placement scores. Therefore, analysis was run with the full l ogistic regression model. Predictor variables included CPT-R scores, full-ti me and part-time status, to determine if a significant relationship existed with the criterion variable. Before concluding that the null hypothes is was not rejected, those predictor variables with a p-value less than 0.05 were selected to explai n the data. If a variable had a p-value less than 0.05, then a logistic regres sion model was used to test the relationship 55

PAGE 67

between the variable and the criterion variab le. If a positive slope was obtained then a relationship existed between the selected variable and the criterion variable. Descriptive and inferential statistics we re also obtained. Descriptive statistics included central tendency (mean, mode, and me dian), variability (standard deviation, variance, and range) and distri bution (skewness and kurtosis) of the GPA with respect to which program track the student was in with success in passing or not passing the highest level college preparatory reading course. Rela tionships were examined using correlations (interval level or higher) for statistical si gnificance between passi ng the highest level college preparatory reading course and the GPA. A one-way analysis of variance (ANOVA) for the third question was used to test the null hypothesis that there was no relationship between a student passing the highest level college preparatory reading course, th e independent, categoric al variable and the students GPA (dependent, continuous vari able) the following sessi on. The program track was also included in this analysis of the GPAs testing for differences in the means of the dependent variable broken down by the le vels of the independent variable. Summary Chapter 3 outlined the methods used to examine the research questions. The study included the Florida Student Data base from Fall Sessions 1997-2004. The researcher tabulated the results of the data to determine relationships between the variables. 56

PAGE 68

Chapter 4 Results The rising cost of attending four-year co lleges, the increase in college-bound high school students, and a larger number of nontra ditional students have resulted in an increasing number of students enrolling in community colleges nationwide. Cliff Adelman, Senior Research Analyst at the U.S. Department of Education, reported approximately 63% of the students entering community colleges require at least one remedial course (2004). According to Ad elman, Deficiencies in reading skills are indicators of comprehensive literacy problems, and they significantly lower the odds of a students completing a degree (1996, p. A56). The National Center for Education Statistics Fall 2000 study repor ted the students requiring remediation in reading and who did not earn postsecondary cred entials rose from 57% in 1982 to 70% in 1992. This problem could be alleviated if the placement te st was used as a screen ing tool, rather than just for placement purposes into developmental classes because we cannot let students believe they have a good chance of earning a college degree if th ey leave high school with poor reading skills (Adelman,1996, p.A57). The purpose of this study was to examin e the predictive validity of several variables to determine if the Florida Com puterized Placement Test Reading (CPT-R) 57

PAGE 69

score alone, or other variab les, could determine whether or not a student would successfully pass the highest level co llege preparatory reading course. The study examined the reading scores on the CPT to determine at what standard deviation below the cutoff score of 83 (exe mpt from reading) a student could still successfully complete the highest level college preparatory reading course. A scaled score of 83 means the student has at tained a 70% on the r eading portion of the placement test. Concurrent research s uggests the score of 83 equates to a 12 th grade and college-level reading level on the Nelson Denny Reading Test as well as the Directed Reading Program (Napoli & Raymond, 1998, p.3) Community college counselors could use this information to make decisions about which students to admit to college programs of study. Since the Fall of 1997, Florida community colleges have used the College Boards Computerized Placement Test; therefore, the data included first time in college (FTIC) Florida community college freshman reading scores ( n = 276,079) for Fall 1997 through Fall 2004. The study attempted to answer the following questions with the hypotheses tested at the 05 level of significance. 1. Is there a relationship between a student s score on the Computerized Placement Test in reading (CPT-R) and success in passing the highest level college preparatory reading course in Florida? 2. Is there a relationship between full-ti me or part-time enrollment during the semester a student is taking the highest level college preparatory reading course 58

PAGE 70

and success in passing the highest level co llege preparatory reading course in Florida? 3. What are students' GPAs the session following successful completion of the highest level college preparatory readi ng course according to the program track (Associate of Arts, Associate of Scien ce, Associate of Applied Science)? Chapter 4 will discuss: (a) data collection, (b) data anal ysis, (c) descriptive statistics (d) findings related to question one that investigat ed whether or not there was a relationship between reading pl acement test scores and su ccessful completion in the highest level college preparat ory reading course, (e) findi ngs related to question two which investigated with full or part time enrollment was related to whether or not a student successfully complete d the reading course, (f) findi ngs related to question three which investigated whether or not there was a relationship of successfully completing the reading course and the grade point average the following session, according to program track, and (g) chapter summary. Data Collection The data for this study was obtained from Associate Vice Chancellor for Evaluation Dr. Patricia Windham of the Fl orida Division of Community Colleges and Workforce Education in March 2006. Variab les included Computerized Placement Test (CPT-R) scores in reading fo r Fall 1997 through Fall 2004, enrollm ent status (full or part time), the students GPA the semester follo wing successful completion of the highest level college preparatory reading course, acco rding to program tracks A.A. (Associate of Arts), A.S. (Associate of Science) and A. A.S. (Associate of Applied Science) and 59

PAGE 71

students final marks in the reading course. Only students who took the placement test in the fall and then enrolled in the highest le vel college preparatory reading course were used in the study to ensure the performance on the placement test had no intervening variable such as other courses requiring reading where a student may have had extensive tutoring to improve reading performance that would inadvertently affect the level of course performance in the reading course. Identity variables were not included, so students could not be individually identified. The researcher adhered to the University of South Floridas Institutiona l Review Board (IRB) policies and procedures for the protection of human subjects. The Statistic al Analysis System SAS version 9.1 (SAS, 2002) was used to analyze the data. Using an alpha level of .05, logistic regression was conducted for questions 1 and 2, and an anal ysis of variance ANO VA for question 3. The review of the literature re vealed no previous studies have examined these research questions. Data Analysis Analysis of archived student sc ores for the years 1997-2005, which was approximately 35,000 scores per year (n = 276,079) for all forms of the CPT-R, were carried out to determine the success of st udents whose exit scores (i.e. 81, 82) were clustered around the state of Floridas 83 cut-o ff score. In addition, the researcher also examined how many standard deviations below the cut-off score a student could successfully pass the highest level in the co llege reading preparatory sequence. Using SAS, descriptive and inferential statistics were obtained. Descriptive statistics included 60

PAGE 72

central tendency (mean, mode, and median), va riability (standard deviation, variance, and range), and distribution (skewn ess and kurtosis) of the incoming freshman CPT-R scores. Descriptive Statistics The Computerized Placement Test in readi ng yields scaled scores ranging from 0 to 120. A scaled score is a statistical convers ion of raw scores, the actual items missed, on the placement test. Scaled scores report comparable results when different test forms are used over time; thus, scaled scor es provide performance standards. For this study, descriptive statistics were obtained first using the entire sample and then using only students who were required to take the reading course. A scaled score of 83 means the student has at tained a 70% on the reading po rtion of the placement test and is exempt from the reading course; howev er, the mean score for all students taking the reading placement test and entering Fall 1997 through Fall 2004 ( n = 276,079) was 64, with a standard deviation of 13.6. The mode the most frequently obtained score, was 75 which equates to approximately 62% of the items correct on the CPT-R (Table 8); therefore, students are entering college with inadequate reading skills. Table 8 FTIC Students Reading Computerized Pla cement Test Scores Fall Sessions 1997-2004 N Mean Median Mode Standard Deviation Variance Range Kurtosis Skewness 276,079 64 67 75 13.60 185.29 119 +0.62 -0.84 Note: Florida First Time in College (FTIC) Student Database ( n=276,079) for Fall sessions 19972004 using SAS. 61

PAGE 73

The kurtosis was +.62 and the skewness was -0.84, re sulting in an approximately normal distribution. Five students had perfect scores of 120.The .25 quartile revealed scaled scores of 57 or lower and the .75 quartile had scaled scores of 74 or higher (Figure 1). Figure 1 Histogram and boxplot of all FTIC st udents reading placement test scores. # 122.5+* 8 0 .* 18 0 .* 35 0 .* 137 0 .* 259 0 .* 617 | .* 834 | .** 1293 | .************************ 22229 | .************************************************ 45175 +-----+ .************************************************ 45348 | | 62.5+*************************************** 36833 | + | .************************** 23661 +-----+ .******************* 17575 | .************** 12845 | .********** 9243 | .******** 6920 | Reading Scaled Scores .******* 6291 | .**** 3367 0 .* 121 0 .* 832 0 .* 281 0 .* 46 0 2.5+* 8 0 ----+----+----+----+----+----+----+----+----+--* may represent up to 945 students Note: Florida First Time in College (FTIC) Student Database ( n=276,079) for Fall Sessions 1997-2004 using SAS. The researcher created a frequency chart (T able 9) to calculate the percentage of students who passed the reading course in both quartiles, and found both groups performed somewhat the same. 62

PAGE 74

Table 9 Comparison of Upper and Lower Quartiles Re ading Placement Scaled Scores with Passing Rates for Students Taking the Highe st Level College Preparatory Reading CPT R Scaled Scores Number of Students Students Passing Reading Course % Passing Upper Quartile: >=74 79,167 49,281 61% Lower Quartile: <=57 72,289 40,891 59% Note: Florida First Time in College (FTIC) Student Database ( n=276,079) for Fall Sessions 1997-2004 using SAS. Furthermore, students whose scaled scores ranged from 11-20 may include students who had adequate reading skills, but decided while taking the placement test to refrain from completing it and just enroll in the reading course, which might help explain the 74% passing rate (Table 10). In addition, only 50% of the 3,845 students who obtained a scaled score of 83 or higher on the CPT-R test, which does not require a reading course, passed the highest college preparatory readi ng course. Students who are exempt from taking the reading course still ta ke the course for a variety of reasons, but usually the primary reason is to refresh their reading skills. And yet this does not explain why the passing rates are so low and also incongruous, suggesting extraneous factors other than a reading placement test score contribute to a students success in the course. 63

PAGE 75

Table 10 Frequency of Reading Placement Test Scor es and Passing Rates in Reading Course for Fall Sessions 1997 Scores Frequency Passed % Passed Failed <10 88 61 69% 27 11-20 1112 831 74% 281 21-30 4649 2036 43% 2605 31-40 13,896 7148 51% 6748 41-50 22,683 13,108 57% 9575 51-60 46,437 27,580 59% 18,857 61-70 79,219 48,628 61% 30,591 71-80 90,652 56,797 62% 33,855 81 6981 4450 63% 6049 82 6515 4051 62% 2531 83 402 218 52% 184 84-90 1694 826 48% 870 91-100 1368 686 50% 682 101-110 328 178 54% 150 111-120 53 35 66% 18 Total 276,079 166,633 60% 109,446 Note: First time in college (FTIC) with Complete Placement Scores ( n = 276, 079) and Florida database Fall sessions 1997-2004 using SAS. For this study, however, since scores of 83 or higher did not require enrollment in a reading course, those scores were not include d in future analysis; resulting in a sample size of 272,232 students. Because the number of students whose scores were 83 or higher was relatively small, the mean score did not change. The mode, the most frequently obtained score, was 75 (Table 11). 64

PAGE 76

Table 11 Computerized Placement Test Reading Scores for Students Required to Enroll in the Reading Course N Mean Median Mode Standard Deviation Variance Range Note: First Time in College (FTIC) Student Database ( n=272,232 which does not includ e students scoring 83 or higher) for Fall sessions 1997-2004 using SAS. Kurtosis Skewness 272,232 64 67 75 13.30 177.08 81 0.57 -0.97 The kurtosis was 0.57 and the skewness wa s -0.97, resulting in a negatively skewed distribution. The extreme observations were five students who scored an 82. The .25 quartile revealed scaled scores of 57 or lower and the .75 quartile contained scaled scores of 74 or higher (Figure 2). Figure 2. Histogram and boxplot of reading placement test scores of st udents required to enroll in the reading course. # 82.5+*********************** 21583 | .************************************************ 45175 | .************************************************ 45348 +-----+ .********************************************* 42103 *-----* .*************************************** 36833 | + | .************************** 23661 +-----+ .******************* 17575 | .************** 12845 | 42.5+********** 9243 | .******** 6920 | .******* 6291 0 .**** 3367 0 .* 121 0 .* 832 0 .* 281 0 .* 46 0 Reading Scaled Scores 2.5+* 8 ----+----+----+----+----+----+----+----+----+--* may represent up to 945 students Note: First Time in College (FTIC) Student Database ( n =272, 232) for Fall 1997-2004 using SAS. 65

PAGE 77

Successful completion of the highest level college preparatory course in reading means the student actually took the course and was awarded a grade of A, B, C, P (pass) or S (satisfactory). A grade of WP (wit hdrawal pass) was not included as successful completion of the course since the student te sted out of the course prior to taking the course. As shown in Table 12, 272,232 students who scored 82 or below placed in the highest level college prepar atory course in reading a nd 164,690 students (60.50%) passed with grades of A, B, C, S, or P, whic h means that 107,542 students (39.50%) did not pass the reading course. Even though this course is not averaged into the grade point average, 41.5% of community college instructors preferred to enter a grade of A, B, or C, revealing the actual level of student performance in the course. The designation of satisfactory performance (S) was 17.01% or 46,298 students. Students who obtained a grade entered as P for passi ng was 1.74% or 4,642 students. 66

PAGE 78

Table 12 Frequency of Grades for Students Requi red to Enroll in the Reading Course Grades Frequency Percent Cumulative Frequency Cumulative Percent A 29,155 10.71 29,155 10.71 B 51,169 18.80 80,324 29.51 C 33,326 12.24 113,650 41.75 P (pass) 4,742 1.74 118.392 43.49 S (satisfactory) 46,298 17.01 164,690 60.50 D 6,161 2.26 170,851 62.76 F 14,898 5.47 185,749 68.23 I (incomplete) 1,369 0.50 187,118 68.73 PR (progress need to re-enroll in course) 11,545 4.24 198,663 72.97 U (unsatisfactory) 8,571 3.15 207,234 76.12 W (withdrawal) 29,517 10.84 236,751 86.96 WF (withdraw fail) 912 0.34 237,663 87.30 WP (withdraw pass) 335 0.12 237,998 87.42 X (no institutional grade awarded) 34,208 12.57 272,206 99.99 Z (audit, no credit) 26 0.01 272,232 100.00 Total passing 164,690 60.50 Note: FTIC students from Florida database Fall sessions 1997-2004 (n=272,232) using SAS. Research Question 1 Is there a relationship between a students score on the Computerized Placement Test in reading (CPT-R) and success in passing the highest level college preparatory reading course in Florida? The null hypothesis for question one is th at there is no relationship between the CPT-R score and a student successfully pass ing the highest level college preparatory reading course. The validity of interpretati on of the results depended on the design of the study; therefore, a logistic regression analysis was used for research question one because the independent variable, test scores, was continuous and the dependent variable, 67

PAGE 79

successful completion of the reading course was coded as pass or fail, a dichotomous variable. The odds of passing the highest le vel college-preparatory reading course was expressed as: P (passing) odds = 1 P (not passing) Results for a logistic regression are interprete d like a regression since the researcher is questioning whether there is a relationship be tween placement test scores and success in completing the reading course; t hus, logistic regression fits an intercept/slope model. The dependent logistic transformati on of the odds, known as logit is the dependent variable of passing or failing the highest level college-preparatory r eading course. The assumption is that the relation between the logit (P) and x is linear. Similar to a simple linear regression, b is the expected change of logit (P) w ith a unit change in x. Therefore, when b is positive, increases in x affects increases in logits. The odds ratio for question one is 1.009, m eaning for every one point increase on the reading portion of the placement test, th e log odds of passing the reading course increases by .00907 (Table 13). An odds rati o close to 1.0 suggests that there is no change due to the predictor variable. The Confidence Interval (CI) for the proportional odds ratio lies between 1.009 and 1.010 and sinc e it does not include 1, the researcher must reject the null hypothesis (p< 0.05) that there is no relationship between student scores and successful comple tion of the reading course. 68

PAGE 80

Table 13 Odds Ratio Estimates Effect Point Estimate 95% Wald Confidence Limits Reading Test Scores 1.009 1.009 1.010 Note: FTIC students from Florida database Fa ll sessions 1997-2004 (n=272,232) using SAS. The analysis of maximum likelihood estimates (Table 14) provides detailed analysis on the variable of placement test scores. The intercept and the slope (Table 14) for the simple logistic regression (chisquare = 64.59, p <.0001) with the independent variable x (scaled reading scores on the placement test) was: logit (P) = a + bx Log odds = -.1534 + .00907*(x) if x=1 then Log odds = -.1443 Odds = e -.14433 = .85 Probability = _odds_ 1+ odds 69

PAGE 81

Table 14 Analysis of Maximu m Likelihood Estimates Parameter Estimate Standard Error Wald Chi-Square P Intercept (b 0 ) -0.1534 0.0191 64.5949 <.0001 Placement Test (b 1 ) 0.00907 0.000293 958.6788 <.0001 Odds e -.14433 Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS Logistic. For example, if a student scores a 54 on the reading placement test, the probability of passing the reading course is 58%; whereas, if a student scores an 82 on the reading placement test, the probability of passing th e reading course is 64%. The x in the above equation would be replaced by 54 or 82, respectively (Table 15). Table 15 Logistic Regression Model for Probability of Successful Comp letion of Reading Course Placement Score Log Odds Odds Probability 54 .336 1.40 .58 82 .590 1.80 .64 Note: FTIC students from Florida database Fall sessions 1997-2004 ( n =272, 232) using SAS. The validations of predicted probabilitie s determined to what degree predicted probabilities match with actual outcomes, using measures of association. The measure of association for this study wa s Somers D statistic and the c statistic, which assesses the quality of the model based on sample size and the independent variable (Table 16). Somers D is used to determine the strength and direction of th e relation between the pairs of variables. Values range from -1.0 (where all pairs disagr ee) to 1.0 (where all 70

PAGE 82

pairs agree). Therefore it equals the differe nce between the percent concordant and the percent discordant divided by 100. The Conc ordant was approximately 52%. The higher the percent means the better the predictive power of the model. The model is statistically significant. Table 16 Association of Predicted Pr obabilities and Observed Responses Percent Concordant Percent Discordant Somers D C 51.9 45.6 .063 Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS. .532 For a logistic regression with high pred ictive accuracy, the receiver operator characteristic (ROC) curve should rise quickly so that the area under the curve is large for a model with high predictive accuracy. Th e ROC curve is a traditional method for showing the relationship between sensitivity and the false positive rate. In other words, if the ROC curve rises slowly and has smalle r area under the curve, then the logistic regression model has low predictive accuracy. The c test, which provides an estimate of the area under the ROC curve was only .53 (perfect association is 1.0). The null hypothesis for research question one is that there was no relationship between the CPT-R score and a student succe ssfully passing the hi ghest level college preparatory reading course. This hypothesis can be reject ed (p<.0001); in other words, there is a statistically significant relationship between the students placement test reading scores and their successfully completing the highest level college preparatory reading course; however, the effect is very small. The placement test score is not an 71

PAGE 83

essential indicator of whether or not a student will successfully complete the highest level college preparatory reading course. Research Question 2 Is there a relationship between full-time or part-time enrollment during the semester a student is taking the highest le vel college preparatory reading course and success in passing the highest level co llege preparatory reading course in Florida? The null hypothesis for the variable full-time enrollment (twelve or more semester hours) and part-time enrollment was there was no relationship between this variable and a student successfully passing th e highest level college preparat ory reading course. In terms of student numbers, there was no remarkable difference in the number of students who attended college on a full-time basis compared to those who attended on a part-time basis. There were 7.71% more part time students than full time students attending (Table 17).The descriptive statistics for enrollment st atus, which was coded as a for full-time and for part-time, does not lend itself to interpretation because it was a nominal variable and therefore has no mean, mode, or median. 72

PAGE 84

Table 17 Frequency of Enrollment for Fall Sessions 1997 2004 Frequency Percent Cumulative Frequency Cumulative Percent Full-time 106,790 39.23 106,790 39.23 Part-time 127,799 46.94 234,589 86.17 S* 37,620 13.82 272.209 99.99 Z* 23 0.01 272,232 100.00 Total 272,232 Students categorized as S were enrolled in the summer term and students categorized as Z are not enrolled for the fall session; both cate gories were not part of the analysis. Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS. The odds ratio for enrollment status to successf ul completion of the highest level college preparatory reading course was 1.059 with Wald 95% confidence intervals for the odds ratios of 1.042 1.076. Since is not incl uded in the confidence interval, enrollment status is associated with success in the course (Table 18), but the association is very small. Table 18 Odds Ratio Estimates Effect Point Estimate 95% Wald Confidence Limits Reading Test Scores 1.059 1.042 1.076 Note: FTIC students from Florida database Fall sessions 1997-2004 ( n =272, 232) using SAS. 73

PAGE 85

The simple logistic regression equation w ith the independent variable x (full or part-time enrollment) was: Log odds = +.4037 + .0573*x The dependent logistic transformation of the odds, known as logit, is the dependent variable of passing or failing the highest level college-pr eparatory reading course. Similar to a simple linear regression, b is the expected change of logit (P) which is either full-time or part-time enrollment (Table 19). Since full time was coded as then the logistic model becomes .4037 + .0573*(1), whic h equates to .4610 (Table 19). If a student is part time then the value of x b ecomes and the equati on then equals .4037. Table 19 Analysis of Maximum Likelihood Estimates Parameter Estimate Standard Error Wald Chi-Square p Intercept ( b 0 ) .4037 .00502 6473.1961 <.0001 Full or Parttime (b 1 ) .0573 .00804 50.8253 <.0001 Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS. Odds e .4610 If a student is full time then the probabili ty of passing the r eading course is 61%, whereas, a part time student has a 60% probabi lity of passing the course (Table 20). Odds = e +.4610 = 1.585 Probability = 1.585 2.585 = .61 74

PAGE 86

Table 20 Logistic Regression Model fo r Probability of Successful Completion of Reading Course Enrollment Status Log Odds Odds Probability Full time .46 1.585 61% Part time .40 1.497 60% Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS. The validations of predicted probabilitie s determined to what degree predicted probabilities match with actual outcomes, using measures of association. The measure of association for this study wa s Somers D statistic and the c statistic, which assesses the quality of the model based on sample size and the independent variab le (Table 21). The Concordant was approximately 25%. The higher the percent means the better the predictive power of Wald ( p<.0001) which was statistically significant. The model is statistically significant a nd may be attributed to whether a student is full-time or part-time since the area under the RO C curve is significant. Table 21 Association of Predicted Probab ilities and Observed Responses c* Percent Concordant Percent Discordant Somers D 24.5 23.1 .014 *Area under the receiver operator characteristic curve ROC curve .507 Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS. 75

PAGE 87

Research Question 3 The final research question examined if there was a difference in grade point averages of students who successfully comple ted the highest level college preparatory reading course and the progr am of study they chose. What are students' GPAs the session following successful completion of the highest level college preparatory reading course according to the program track (Associate of Arts, Associate of Scien ce, Associate of Applied Science)? The null hypothesis for question three wa s there was no relationship between the grade point average (GPA), according to pr ogram track the subsequent session, and the students successful completion of the highest level college pr eparatory course in reading. An ANOVA was performed followed by a Tuke y test (alpha=.05) which indicated a significant difference in the group means of th e three program tracks, meaning there is some effect of successfully completing the reading course and obtaining a higher GPA in an Associate of Applied Science program trac k, rather than an Associate of Science or Associate of Arts program track. The GPA aver ages were somewhat different, but there was not a marked difference. Apparently, th e highest GPA was in the A.A.S. program (2.40), yet most students declared an A.A. pr ogram of study with a slightly lower GPA of 2.33. Students self report a program of study on the college application for admission. Students either select a speci fic program of study or indicate they are undecided. For the purposes of this study, only students who decl ared a major are represented. The State of 76

PAGE 88

Florida codes the majors as follows: A.A. = 0, A.S. = 1 and A.A.S. = A. Furthermore, the data set only contained students who completed th e course with an A, B, C, or S. Descriptive statistics were obtained for reading grades, grade point averages and programs of study for the 35,102 students who identified a program of study. The frequency of grade assignment (Table 22) re vealed approximately 39% of students with a declared program of study were assigned a grade of B. Table 22 Frequency of Grades for Successful Comple tion of Reading for Fall Sessions 1997-2004 Grades Frequency Percent Cumulative Frequency Cumulative Percent A 8,436 24 8,436 24 B 13,574 39 22,010 63 C 7,493 21 29,503 84 S 5,599 16 35,102 100 Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS. Actual semester grade point averages for the session following successful completion of the reading course were availa ble from the Florida student database. All grade point averages were according to a 4.0 gr ade scale (Table 23). The researcher only included students earning an A, B, C, or S. Descriptive statistics included central tendency, (mean, mode, and median) variabilit y (standard deviation, variance, and range) and distribution shape (skewness and kurtosi s). The kurtosis was -0.17 and the skewness was -0.58, resulting in an approxi mately normal distribution. 77

PAGE 89

Table 23 Descriptive Statistics of Grade Point A verages the Session Following FTIC Students Successful Completion of Reading Course Fall Sessions 1997-2004 N Mean Median Mode Standard Deviation Variance Range Kurtosis Skewness 34,896 2.33 2.50 3.00 1.03 1.05 4.00 -0.17 -0.58 Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS. Figure 3 Histogram and boxpl ot of grade point averages the session following successful completion of reading course. # 4.1+******************* 1880 | .** 144 | 3.7+************ 1186 | .******************* 1888 | 3.3+******************** 1954 | .************************************************ 4874 +-----+ 2.9+********* 845 | | .**************************** 2833 | | 2.5+**************************** 2809 *-----* .************************* 2508 | + | 2.1+********************************************* 4528 | | .****** 596 | | 1.7+*************** 1516 +-----+ .************* 1310 | 1.3+********** 1012 | .**************** 1606 | Grade Point Averages 0.9+*** 274 | .**** 384 | 0.5+**** 354 | .** 193 | 0.1+********************** 2202 | ----+----+----+----+----+----+----+----+----+--* may represent up to 102 counts and missing 206 Observations Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS. The .25 quartile revealed a grade poin t average of 1.75 or lower and the .75 quartile contained grade point averag es of 3.0 or higher (Figure 3). Only 35,102 students out of 272, 232 student s declared a major, which means only 13% of first time in college freshman declared a major in the Fall sessions of 19972004 (Table 24). There were 206 missing grade point averages, so the total grade point averages reported for program levels was 34, 896. 78

PAGE 90

Table 24 Frequency of Program Levels Program-Level Frequency Percent Cumulative Frequency Cumulative Percent A.A. 27,328 77.85 27,328 77.85 A.S. 5,887 16.77 33,215 94.62 A.A.S. 1,887 5.38 35,102 100.00 Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS. Most students chose the Associate of Arts pr ogram track and the grade point averages of the three programs of study were not remarkably different (Table 25). Table 25 Mean Grade Point Averages According to Program of Study Program Level n Mean Mode Skewness Standard Deviation Variance Kurtosis A.A.S. 1,880 2.40 3.00 -0.62 1.03 1.07 -0.07 A.A. 27,154 2.33 3.00 -0.58 1.00 1.00 -0.10 A.S. 5,862 2.27 3.00 -0.53 1.12 1.26 -0.50 Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS. A one-way analysis of variance ( ANOVA) was used to test the null hypothesis that there was no relationship between the students GPA (dependent, continuous variable) the following session, and a stude nt passing the highest level college preparatory reading course, the independent, categorical variable, according to program of study. ANOVA is the method for comparison of three or more groups and has the advantage of testing whether or not a difference occurs between the groups. The 79

PAGE 91

hypothesis was that all thr ee program tracks have the same population mean; no difference existed between the three groups GPAs. The F statistic and p-value rejected the null ( F =13.65, p <.0001), indicating differences in the means between the three groups (Table 26). Table 26 Analysis of Program Level to Grade Point Averages Source Df Sum of Squares Mean Square F Value Model 2 28.67 14.33 13.65 Error 34,893 36,663.77 1.05 Corrected Total 34,895 36,692.45 Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS The follow up test, the Tukey test (alpha = .05) indicated a signi ficant difference in the group means (Table 27) of all three program tracks (p <.05). The ANOVA was robust to the assumption of homogeneity of variance. Even though the groups were not equal in size, the variances among the three groups were somewhat the same. In addition, the assumption for the independence of observations may not have been met due to instructors current grading practices. Profe ssors each have individual biases on how they evaluate their courses and how students gr ades are assigned. Based on the descriptive statistics in Table 25 and wh at is known about the robustness of ANOVA (Cody & Smith, 1997) there appears to be no s ubstantial violation to the no rmality or equal variance 80

PAGE 92

assumption. There is likely some relationship between a student succ essfully passing the reading course in a specific program tr ack and the GPA the following session. Table 27 Tukey's Studentized Range (HSD) Test for Grade Point Averages Program Level Difference Between Means 95% Confidence Limits A.A.S. A.A. 0.063090 0.005794 0.120380 A.A.S. A.S. 0.125673 0.061994 0.189352 A.A. A.S. 0.062583 0.027981 0.097184 Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS Although the GPA averages were somewhat different, there was not a marked difference. The highest GPA wa s in the A.A.S. program, yet most students declared an A.A. program of study with a slightly lower GPA of 2.33. Summary The study revealed there wa s a statistically significant relationship between students scores on the reading component of the CPT and successful completion of the highest level college preparat ory reading course. However, the research does not identify any one particular scaled score which woul d provide information on how many standard deviations below the scaled score of 83, wh ich is the cut-off sc ore not requiring the reading course, a student could still successf ully complete the highest level college preparatory course in reading (F igure 4.) It appears that stud ents can still pass the highest level college preparatory reading regardless of the scaled score on the placement test. 81

PAGE 93

Figure 4. Fall sessions 1997-2004: Co mparison of percentage of students passing reading course and corresponding CPT-R scaled score ranges. 0%10%20%30%40%50%60%70%80% <10 21 to 30 41 to 50 61 to 70 81 to 90 101 to 110Percentage PassingScaled Score Ranges Scaled Score Ranges Percentage Passing Note: Florida database for FTIC Fall sessions 1997-2004 ( n =272, 232) using SAS. Furthermore, results of the study revealed students who successfully complete the course are attaining grade point averages the following session that meet the requirements for maintaining academic standing, an indicati on that many of the students may stay in school and complete a program of study. 82

PAGE 94

Chapter 5 Summary of Findings, Conclusions, and Implicat ions for Theory, Practice, and Research Community colleges are committed to welcoming all students to participate; however, planning a program of study so student s are successful requires a placement test to identify deficiencies. The purpose of the st udy was to investigate whether there was a significant correlation between student test scores on the reading component of the Computerized Placement Test (CPT-R) and su ccessful completion of the highest level college preparatory reading course, offered by co mmunity colleges in the state of Florida. Furthermore, this study examined whether or not full-time or part-time status had a relationship to a student successfully comple ting the highest level college preparatory reading course. In addition, this study exam ined the association between successfully completing the highest level college prepar atory reading course and the grade point average (GPA) in college studies the fo llowing session, according to program track (Associate of Arts, Associate of Scien ce, Associate of Applied Science). Method Summary The data for this study was obtaine d from Associate Vice Chancellor for Evaluation Dr. Patricia Windham of the Fl orida Division of Community Colleges and Workforce Education in March 2006. Various statistical techniques including logistic regression and ANOVA were used to study the data and analyze the results. 83

PAGE 95

Descriptive Data The sample for the study included students who enrolled for the first time at one of the twenty-eight Flor ida community colleges during th e Fall terms of 1997-2004, scored 82 or lower on the reading portion of the comput erized placement test and enrolled in the highest level college prep aratory reading course ( n = 272,232). Table 28 depicts the number of participants who enrolled in the read ing course during the fall sessions with the central tendencies of mean, median and mode. Table 28 Fall Sessions 1997-2004 Computerized Placement Test-Reading Scaled Score Comparisons 1997 1998 1999 2000 2001 2002 2003 2004 Number of Students 27,883 30,063 33,782 30,892 35,211 44,037 35,434 38,777 Mean 62 63 64 65 65 62 65 65 Median 65 66 67 68 68 66 69 68 Mode 68 75 75 75 75 69 73 71 Note: Fall 1997-2004 first time in college (FTIC) with Reading Placement Scores ( n =272, 232). Students self-report a program of study on the college application for admission. Students either select a speci fic program of study or indicate they are undecided. For the purpose of this study only students who declar ed a major are represented. The State of Florida codes the majors as follows: A.A. = 0, A.S. = 1, and A = A.A.S. Completion with a grade of A, B, or C was considered successf ul completion of the c ourse since the State 84

PAGE 96

of Florida considers only thes e letter grades as passing and was coded as for passing, and for not passing. Actual semester grad e point averages were available from the Florida student records database. All grade point averages were computed using a 4.0 grade scale. Summary of Findings Using quantitative analysis tec hniques, this study explored three research questions, each of which is presented below with a summary of the findings for each question. Research Question 1 Is there a relationship between a students score on the Computerized Placement Test in reading (CPT-R) and success in passing the highest level college preparatory reading course in Florida? Using SAS Logistic regression, th e researcher determined a relationship did exist between scores on the CPT in reading and successful completion of the reading course. Students with higher reading placement scores had greater odds of passing the reading course. For every point scored on the read ing placement test, the log odds of passing the course increased by .00907. The likelihood ratio te st associated with logistic regression provided the evaluation of the st atistical significance of the relationship of the placement test in reading and success in the reading course. However, even though the odds ratio indicated a student had a better probability of passing the read ing course as scaled scores increased, the change in probability was very sm all. The reality is that, between the years 1997-2004, 40% of students, regardless of their placement test score, failed the reading course. Even students identified as passing th e placement test and therefore not required 85

PAGE 97

to enroll in the reading course failed the reading course. Studen ts with scaled scores below 10 passed the reading course. The placement test scores, therefore, are not indicative of whether or not a student will successfully complete the reading course. Determining how many standard deviations below the cut-off sc ore of 83 cannot be determined from this study. Research Question 2 Is there a relationship between full-time or part-tim e enrollment during the semester a student is taking the highest level college preparatory reading course and success in passing the hi ghest level college preparatory reading course in Florida? The results demonstrated that students w ho are identified need ing reading may be more successful with a full-time program of st udies. If a student is full time, then the probability of passing the reading course is 61% whereas a part time student has a 60% probability of passing the course. While the difference may be statistically significant, the difference is very small; the actual comparison suggests part time students are just as likely to pass the course as full time students. Research Question 3 What are students' GPAs the session following successful completion of the highest level colle ge preparatory reading course according to the program track (A.A., A.S., A.A.S.)? An ANOVA procedure followed by a Tukey Studentized Test looked for differences among the three program tracks. The assumptions of normality, homogeneity 86

PAGE 98

of variance, and independence of observations were met. The quantitative analysis showed differences among the means of the three programs of study were statistically significant. However, the differenc es reflected reveal only sli ght variations in grade point averages (GPAs). Students in an A.A.S. program had slightly higher GPAs than students in an A.A. or A.S. program. Only 35,102 students who successfully completed the reading course had declared a program of study and had GPAs. Of that aggregate, 39% of those students were assigned a grade of B in the highest level college preparatory reading course. Twenty-four percent rece ived As, 21% obtained Cs and 16% received an S, satisfactory completion. Conclusions The study was conducted to determine whether or not a placement test could be an essential indicator of student success in th e highest level college preparatory reading course. Although studies have been conducted on placement testing and its relationship to developmental mathematics courses, the res earcher found no studies have been done to determine if the CPT reading test had a relationship to the reading course. The sample (n = 272,232) consisted of first-ti me-in-college Florida community college students who were required to take the highest level college preparatory reading course. Statistically significant relationsh ips were found between the entry test and successful completion of the reading course. St udents declaring an Associate of Applied Sciences program of study achieved GPAs somewhat higher than students declaring an Associate of Science or Associate of Arts program of study. 87

PAGE 99

The study revealed that students who we re exempt from the reading course because of their placement test scores, and still took the course, did not have higher passing rates in the course. Also students with higher placement scores did not have particularly higher passing rates in the reading course than students with lower placement scores. In fact, descriptive results revealed students who obtained CPT reading scores in the 11-20 scaled score range had a 74% passing rate in the reading course, which was 24 points higher than the 91-100 scaled score range (50%), a range not requiring enrollment in the reading course. According to the College Board, the 83 s caled score, which exempts a student from taking the reading course, equates to 70% on the paper/pe ncil version of the test, yet the study revealed that the s caled score 64 was the average score for Fall sessions 19972004, which according to previous studies equates to 9/10 th reading grade level on the Nelson-Denny Reading Test (Napoli & Raymon d, 1998). In addition, the most frequently obtained scaled score was 75 from Fall se ssions 1997-2004, which equates to an 11 th grade reading level on the Nelson-Denny Reading Te st; however, the results of this study showed only 61% (49, 281 out of 79,167) of th e upper quartile of students (scaled scores >74) passed the highest level college preparatory reading c ourse. It is essential to conclude Florida high schools need to implem ent intensive programs of study in reading because students are gravely underprepared for college studies. Results of the study suggest passing the reading course is significant toward maintaining good academic standing, which ensure s a student generally can continue in a program of study. The placement test may only be one of several essential indicators that 88

PAGE 100

would determine if a student would be successful in the reading course. More importantly, however, successfully completing the reading co urse is an indication the student may well finish a program of study since grade point averages the session following successful completion of the reading course indicated students were able to on average obtain grade point averages permitting continuance in their program of studies. Limitations The study was delimited to developmental programs in the Florida Community College System. Only students who took the Florida College Entry-Level Placement Test (CPT) in the fall session and subsequently enrolled in the highest level college reading preparatory course were included in the st udy. The state of Florida considers assigned grades of A, B, or C as passing grades in the reading course. It was assumed the CPT provides an accurate assessment of the students reading ability. Implications for Theory According to the results of this study, placement tests scores have a significant relationship to student success in the readi ng course. Other essential indicators (e.g. persistence, high school c ourses) including nonacademic variables (e.g. economic background, single parent), not measured in this study, may have a more definitive impact on whether or not a student is successful in th e reading course. The placement test merely reveals that the student has reading defici encies, whereas, performance in the reading course includes the ability of the student to apply good study skill habits as well as the ability of the instructor to provide teaching methods that pr omote successful completion of the course. 89

PAGE 101

Former studies have found that success in college could be attributed to the high school grade point average (Cohen, 1989); how ever, recent studies by Cliff Adelman, Senior Research Analyst at the U.S. De partment of Education, (1999) suggest high schools need to provide more rigorous curricul a for students, so that more students can successfully matriculate into a college program of study. Th e quality of the high school curriculum is a better predictor of college success than scores obtained on the computerized placement test in reading. Dr. Edward Behrman at National Univer sity contends using content-general reading tests for placement into developmenta l or credit-level courses lacks content, criterion, and construct validity for placem ent purposes. Behrman (2006) recommends using content-specific reading tests. In other words, better placement testing may be the answer, rather than eliminating placement tes ting. The key may be to evaluate whether a student needs learning assistance in a particul ar credit-level course. Although there are many academic and nonacademic reasons why a student may not have been successful in the reading course, Behrman believes the amount of variance in the course grade accounted for by scores on content-general read ing tests may be too low to warrant the continued use of these tests to predict su ccess in a particular course (2006, p.42 ). Behrman and Street (2005) found a content-sp ecific reading test for an introductory anatomy course was a significant predictor of course grades, but a content-general test was not. Behrman claims, Perhaps one of the more perplexing issues in placement testing is how to achieve a more accurate pr ediction that takes into account the various academic (and perhaps nonacademic) variable s that affect academic performance (2006, 90

PAGE 102

p. 42). Behrman concludes that refining the placement test so that it becomes a better predictor of success is a better choice than no placement test. Developmental educations major focus ha s not been in theoretical frameworks, but in classroom practices. The majority of developmental educators do not seek out theories to help their students, but turn to best practices in the fi eld. Since many reading instructors do not steep themselves in theo ry-based instruction, teaching practices do not change dramatically, but evolve over time. Th e majority of reading instructors still use mastery learning for instruction, which in itse lf, creates a personal theory of teaching for individual instructors. Strategic learning which is supported by many researchers in the field of reading suggests that instructors need to begi n to understand the complexity of the relationship between le arning and studying so that st udents apply strategies and various processes to different types of content. Implications for Practice The results of this study lead to several implications for practice in Floridas community colleges. 1. Continue to use placement test scores to place students in reading courses to promote success in future courses requiring reading even though the results of the study reveal the placement test scores cannot suggest successful completion of the reading course, at the ve ry least, students are made aware that they do have deficiencies in reading which may be correct ed by enrolling in the reading course. 2. To better evaluate the successful completion of the highest level college preparatory reading course in Florida community colleges, the 91

PAGE 103

State needs a standard-setting committee to establish a cut-off score for the Florida Basic Skills Exit Test, in addition to standardizing test administrative practices. 3. Scrutinize program tracks of study for st udents who need developmental reading to determine if students should be counseled into specific majors. 4. College counselors and advisors should provide extensive career exploration in areas which may not always require a colle ge degree, or even a certificate, to students who have made little progress, especially after three attempts (on the third attempt, students must pay full tuition), in successful completion of identified courses since the results of the study revealed approximately 40% of FTIC students fail the readi ng course the first time. 5. Developmental instructors need to ex plore innovative delivery methods coupled with student learning styles and learning communities in developmental reading courses given that national research s uggested the reading course may be the main indicator of future success in college courses, and this current study revealed the placement test merely rec ognized deficiencies for remediation. 6. Results of the study suggest that othe r factors may contribute to a student successfully completing the reading cour se. One factor may be the various teaching styles of community college pr ofessors. Therefore, community colleges and the state educational agencies sh ould provide developmental educators training workshops which provide and promot e current research trends as well as instructional techniques, so developmen tal educators can address the needs and 92

PAGE 104

challenges of the increased numbers of underprepared students who must meet the demand of our nations workforce at a ll levels. The National Center for Public Policy and Higher Education President Patr ick Callan claims the United States will have a competitive disadvantage if public policy makers do not address the students who are unprepared for college, re sulting in outsourci ng high-level jobs of the future. 7. The study revealed that a 75 scaled scor e was the most frequent score on the reading portion of the CPT for Fall sessions 1997-2004, revealing no improvement in reading ability for FTIC students; therefore, Florida high school English courses should include regular tes ting of reading comprehension so that students reading skills impr ove prior to college entry. 8. Since the study revealed no improvement in high school students scores on the reading portion of the CPT from Fall se ssions 1997-2004, the reading skills set on the CPT should be incorporated into a high school elective and become part of the core curriculum. 9. Since 1996, the CPT may be administered to evaluate Florida tenth graders; however, since it is voluntary, few students take the test. The test should be mandatory, to identify students who need remediation prior to graduation. Students who are identified as remedial should be required to take mandatory reading classes in the summer following tenth grade. 10. Since the study suggests othe r factors may contribute to whether or not a student will be successful in the highest level preparatory reading course, namely, class 93

PAGE 105

instruction, the Florida Department of Education needs to provide ongoing professional workshops at the State leve l, bringing together both community college reading instructors and high school instructors. 11. Create a new placement test that is more a ligned with exit test standards in the highest level college preparatory reading course, one that measures Grade 13 college-level reading skills, rather th an the ACCUPLACER test currently used. Implications for Research In response to the academically underprepared, the U. S. Education Department has recently established a national research cen ter to address topics such as remediation and learning communities for unprepared students (Lederman, 2006). The National Research and Development Center on Postsecondary Education will be located at the Community College Research Center at Columbia Universitys Teacher College. The focus is to improve access to higher educati on as well as improve the rates of earning degrees. Based on the findings of this study and the limited research conducted in the field of reading placement tests, and the relationships to success in the reading courses, future research in Florida should examine the re lationship between successfully completing the reading course and the subsequent success in other courses requiring reading. For example, research should be conducted to determine if there is a relationship between passing the reading course and successful co mpletion of courses which require college reading skills, included, but not limite d to, composition, humanities, sociology or government. The ultimate success of the studen t is not the score on the placement test, but 94

PAGE 106

whether or not successful completion of the reading course predicts future success in college level courses. The results of this study suggest seve ral other areas for future research: 1. Develop credible course-specific placement tests to determine if, as current research suggests, placement testing should transcend placement in just developmental reading courses, but extend to placement in specific college-level classes. 2. Verify Cliff Adelmans recommendation that high schools should provide a rigorous curriculum, by examining high sc hool transcripts for courses which may render a predictive quality, and whether or not there is a rela tionship to passing the CPT and/or passing the reading course, and more importantly earning a degree. 3. A recent study revealed that students should not work more than 15 hours a week if they are to be successful in college studies. Conduct studies at the community college level to see if a certain number of hours of employment should restrict the number of hours of course work. 4. Examine the relationship of placement sc ores in the other two developmental disciplines of mathematics and English, and subsequent success in the highest level preparatory courses in mathematics and English. 5. Develop a survey which would determine the characteristics of students who are successful in the reading course, specifically, study habits or others nonacademic factors which contribute to a students success. 95

PAGE 107

6. Current research suggests students should form learning communities for support in the learning process. Colleges sh ould implement learning communities and track students to see if this provides suppor t in the first year and subsequently a positive impact in developmental studies and/or future studies. Students comfortable in learning communities may pr ove to be an integral part of any workforce in the most dynamic business environments. 7. A longitudinal study including all Flor ida community colleges should be conducted to determine the best practic es which contribute to successfully completing the highest level college preparatory reading course. 8. A follow-up study should be conducted to see if the students whose scores are clustered around 83 complete a program of study. 9. A future study should look at results by age and/or age and program tracks because A.A.S. degree students are often older. Th e current study revealed that students in an A.A.S. program track averaged a higher grade point average th an students in the A.A. or A.S. program tracks, the session following successful completion of the highest level college prep aratory reading course. Students who are underprepared for colle ge-level courses due to reading deficiencies would be better served, if at th e very least, high schools returned to teaching reading skills in their core curriculum. Cont ent area courses, English courses and reading electives having a prescribed set of reading sk ills would be the first step in ameliorating the influx of students requiring reading c ourses when they enter college. Continuing reading programs beyond middl e school for all students should be implemented because 96

PAGE 108

school administrators should not be content that students are passing a reading test on the tenth grade reading level, but should be prom oting reading achievement which assures all students are indeed ready for college. 97

PAGE 109

List of References Access and Quality: Improving the performan ce of Massachusetts Community College developmental education programs. (1998). Massachusetts Community College Developmental Education Committee. (ERI C Document Reproduction Service No.ED 428 785) Adelman, C. (1996). The truth a bout remedial work: It's more complex than windy rhetoric and simple solutions suggest The Chronicle of Higher Education, 43 (6), A56. Adelman, C. (1999). Answers in the tool box: Academic intensity, attendance patterns, and bachelor's degree attainment Washington, DC: U.S. De partment of Education. Adelman, C. (2004). Principal indicators of student ac ademic histories in postsecondary education, 1972-2000. Washington, DC: U.S. Departme nt of Education, Institute of Educational Science. Adelman, C. (2004). The empirical curriculum: Change s in postsecondary course-taking, 1972-2000. Washington, DC: U.S. Department of Education. Allan, B. Bartek, B. Melton. F. Runk, P. & Schaffer, L. (1995). Exemplary Practices: CostEffectiveness in developmental education 1992-1994 Grant Project, Results of the Survey of developmental education practices (report No. JC 950 500) East Lansing, MI: National Center for Research on Teacher Learning, (ERIC Document Reproduction service No. ED 387 195). American Association of Community Colleges (2000). AACC Position Statement on Remedial Education. NW Washington, DC. Amey, M., & Long, P. (1998). Developmental course work and early placement: Success strategies for underprepared community college students. Community College Journal of Research and Practice, 22 (1), 3-10. An Analysis of Developmental Education at Michigans Associate Degree-Granting Institutions. (1999). Developmental Educa tion Steering Committee. (ERIC Document Reproduction service No ED 445 741) 98

PAGE 110

Armstrong, J.David. (2000). Who is the typi cal FCCS student? Fast Facts, Florida Community College System. FF-32, Fl orida Department of Education. Armstrong, J. David. (2004). Analysis of stude nt characteristics, 1998-99 Compared to 2003-04. Data Trend #29, Florida Department of Education www.flboe.or/cc Armstrong, J. David. (May, 2005). Developmen tal education in Florida community colleges. Florida Department of Education Armstrong, William B. (2000). The association among student success in courses, placement test scores, student background da ta, and instructor grading practices. Community College Journal of Research and Practice, Taylor & Francis 24, 681-695. Baron, W. (1997). The problem of student retention: The Bronx Community College solution---the freshman year initiative pr ogram. (ERIC Document Reproduction Service No. ED 409 971). Behrman, E.H. (2006). Can content-specific reading test help identify students in need of learning assistance Journal of Devel opmental Education, 29(3), 42. Behrman, E. H. & Street, C. (2005) The validity of using a conten t-specific reading test for college placement. Journal of College Reading and Learning, 35 (2), 5-21. Boylan, H. (1983). Is developmental education wo rking? An analysis of research. National Association for Remedial/Developmental Studi es in Post-Secondary Education (Serial Number 2). Boylan, H. (1988). The historical roots of developmental education. Research in Developmental Education 5 (3). Boylan, H. (1995). The scope of developmenta l education: Some basic information on the field. Research in Developmental Education, 12(4), 1-4. Boylan, H. (1999). Developmental Education: Demographics, Outcomes, and Activities. Harvard Symposium 2000: Developmental Education. Journal of Developmental Education Volume 23, Issue 2, Winter, 1999. Boylan, H., Bonham, B., & Bliss, L. (1992). Characteristics components of developmental programs. Research in Developmental Education. 11(1). Boylan, H., Bonham, B., & Bliss, L. (1994). Report of the National Study of Developmental Education. Presented at the National Conference on Research in Developmental Education, Charlotte, NC, November. 99

PAGE 111

Boylan, H. R., Bonham, B. S., & Bliss, L. B. (1994). Who are the developmental students? Research in Developmental Education, 11 (2). Boylan & White. (1987). Educating all the nations people: The historical roots of developmental education. Research in Developmental Education, 4 (4). Breneman, D.W. (1998). Remediation in higher education: Its extent and cost, in Brookings papers on educational policy 359-383. Washington, D.C: The Brookings Institute. Breneman, D. W. & Haarlow W. N. (1998) Remediation in higher education a symposium featuring remedial education costs a nd consequences. Thomas Fordham Foundation, www.fordhamfoundation.org/library/remed.html Breneman, D. W. & Haarlow W. N. (1999). William N. Esta blishing the Real Value of Remedial Education. Chronicle of Higher Education 45, (31). Brier, E. (1984). Bridging the academic preparation gap: An historical view. Journal of Developmental Education. 8(1), 2-5. Brint.S, & Karabel, J. (1989). The Diverted Dream: Community Colleges and the Promise of Educational Opportun ity in America, 1900-1985 New York: Oxford University Press. Brubacher, J.S. & Rudy, W. (1976). Higher educational transition: A history of American colleges and universities 1636-1976. New York: Harper-Collins. Bush, Wendy. (2001). A Transcript Analysis of the Char acteristics of First-Time InCollege Students Who Fail Their First Preparatory Mathematics Course in a Community College ( a UCF dissertation) UM I publication AAT 301 3901). Cainous F. (1987). Remedial /Developmental Student Characteristics Survey. Casazza, M. & Silverman, S. (1996). Learning assistance and developmental education. San Francisco: Jossey-Bass. Clark, Burton R. (1960). The Open Door College: A Case Study. New York: McGraw-Hill. Clemens, Linda. (2002). Predicting community college student performance on the Florida Basic Skills Exit Test in el ementary algebra in a collabor ative instruction environment (a UF dissertation) UMI publication AAT 3056728. Cody R & Smith.J., (1997). Applied Statistics and the SAS Programming Language Upper Saddle River: Prentice Hall. 100

PAGE 112

Cohen, A., & Brawer, F. (1989). The American Community College. San Francisco: JosseyBass. Cohen, A.M. & Brawer, F. B. (1996). The American Community College (3 rd ed.) San Francisco: Jossey-Bass Publishers. College Entrance Examination Board. (1990). College board computerized placement tests coordinators guide. Princeton, NJ: Author. Conlin, S. (1989). Percentage of Southwestern Community College students by passing remedial work and the effects on subsequent grades and dropout rate. (ERIC Document Reproduction Service No. ED 37 240) Congos, D.& Schoeps, Nancy. (1997). [Electro nic Version] A Model for Evaluating Retention Programs. Journal of Developmental Education. 21 (2). Cran, Margaret R. (1998). Correlations between student presage variables and performances on the mathematics subtest of the Florida College Level Academic Skills Test (CLAST) (a USF dissert ation) UMI publication AAT 9842145. Cross, K. (1971). Beyond the open door. San Francisco: Jossey-Bass. Deil, Regina. ( 2001). Cooling-out or WarmingUp? Identity, Integration and Aspirations Among Community College Students. [Submitted Proposal] Elifson, Joan M. & Pounds, M. Linda. (1995) [Electronic Version] Planning for and Assessment of Developmental Programs. Journal of Developmental Education 19 (1), 2. Farmer, V. & Barham, W. (2001). Selected models of developmental education programs in higher education. New York: University Press of America. Ignash, Jan. (1997). New Directions for Community Colleges no. 100, Winter 1997. San Francisco: Jossey-Bass Publishers. Illich, Paul A., Hagan, Cathy, & McCallister, Leslie. (June, 2004). Performance in collegelevel courses Among Students Concurrently Enrolled in Remedial Courses: Policy Implications. Community college Journal of Research and Practice. V. 28(5) p. 435-53. Knopp, L. (1996). Remedial education: An undergraduate student profile. ACE Research Briefs. 6 (8), 1-11. Kuehner, Alison. (1999) The effects of com puter-based vs. text-based instruction on remedial college readers. Journal of Adolescen t & Adult Literacy v. 43, 2, p. 160. 101

PAGE 113

Kulik, C. Kulik, J. & Shwalb, B. (1983). Co llege programs for high risk and disadvantage students: A meta-analysis of findings Review of Educational Research 53, no. 3: 397414. Lederman, D. (2006). Research on higher ed gets a boost. Inside Higher Ed. Retrieved October 15, 2006, from http://insidehighered com/news/2006/07/ii/edresear. Livingston, A., and Wirt, J. (2004). The Condition of Education 2004 in Brief (NCES 2004076). U.S. Department of Educati on, National Center for Education Statistics.Washington, DC:US. Government Printing Office. Maxwell, M. (1991). Evaluating academic s upport systems: A sourcebook. Kensington, MD: M.M. Associates. Maxwell, M. (1997). Improving Student Learning Skills, Clearwater, FL: H & H Publishing. McCabe, R. & Day, P. (1998). Developmental Edu cation: A twenty-first century social and economic imperative. Mission Viejo, CA: League for Innovation. McCusker, M. (1999). ERIC review: effectiv e elements of developmental reading and writing programs. Community College Review. 27 (2).93. Miglletti, Cynthia, Strange & Carney ( 1998) Learning styles, classroom environment preferences, teaching styles, and remedial co urse outcomes for underprepared adults at a two-year college. Community College Review v.26 no 1 p. 1-19. Napoli, A. & Raymond, Lanette. (1998) [E lectronic Version] The CPT Reading comprehension test: a validity study. Journal of Developmental Education. 22 (1). National Center for Education St atistics (1995). College-level re medial courses in the falloff 1995. Washington, DC: U.S. Office of Educatio n, Office of Educational Research and Improvement. National Center for Education Statistics (1995). Remedial ed ucation at higher education institutions in fall 1995. Washington, DC : U.S. Office of Education, Office of Educational Research and Improvement. National Center for Education Statistics (2003) remedial e ducation of degree-granting postsecondary institutions in fall 2000. Washington, DC: U.S. Office of Education, Office of Educational Research and Improvement. Peng, C. & So. T. (2002). Logistic regres sion analysis and reporting: A primer. Understanding Statistics I(1), 31-70. 102

PAGE 114

Performance Tracking of Former FCCS and SUS Students Between 1994-95 and 2002-03. Fast Fact #76. Florida Department of Education www.fldoe.org/cc Rosenbaum, James. (1999). Unrealistic Plans and Misdirec ted Efforts: Are Community Colleges Getting the Right Messa ge to High School Students? Community College Research Center Institute on Education and the Economy Columbia University. Roth, J. et al. (2001). [Electronic Version] Effect of high school course-taking and grades on passing a college placement test. High School Journal. 84 (2). Rouche, J. E. & Rouche, S. D. (1999). Keepi ng the promise: Remedial education revisited. Community College Journal 69, no.5: 12-18. Saxon, P.&Boylan, H. (1999). Characteristics of community college remedial students prepared for the league of innovation in the community college. National Center for Developmental Education. Schoenecker, C. et al.(1996, May). Developmental education outcomes at Minnesota community colleges. Paper presented at the Annual Forum of the Association for Institutional Research. (ERIC Document Reproducti on Service NO. ED 397 712). Sinclair Community College. (1994). The impact of developmental education on student progress: A three-year longitudinal analysis, Dayton, OH: Sinclair Community College, Institutional Planning and Research (ERIC Document Reproduction Service No. 383 382). Smittle, Pat. (1993). Computer ad aptive testing: A New Era. Journal of Developmental Education. 17 (1), 8. Smittle, Pat. (1995).Academic performance pr edictors for community college student assessment. Community College Review. 23 (2). No. 383-382. Suciu, Linda. (1991). Prediction of success in introductory college mathematics courses at Trident Technical College. (a USC dissertation) UMI publication AAT 9214977. Sugarman, R. P. & Kelly, P.J. (1998, May). An analysis of remedial education at Kentucky's public universities and community colleges. Paper presented at the Association for Institutional Research: 38 th Annual Forum, Minneapolis, Minn. Summers, Michael D. (Spring, 2003). Attr ition research at community colleges. Community College Review. V.30(4) p. 64. 103

PAGE 115

The State of Basic Skills Instruction in Ca lifornia Community Colle ges.(2000). Academic Senate for California Community Colleges, Sacramento. (ERIC Document Reproduction Service No Ed 447 876). Tinto V. (1987). Leaving College: Rethinking the courses and cures of student attrition. Chicago: University of Chicago Press. U.S. Department of Education, Nationa l Center for Education Statistics, Remedial Education of Degree-Granting Postsecondary Institutions in Fall 2000, NCES 2004010, by Basmet Parsad and Laurie Lewis. Pr oject Officer: Bernar d Greene. Washington, DC: 2003. Walleri, R.D. (1987, May). A longitudinal st udy of "guided studies" students. Paper presented at the annual forum of the associa tion of institutional research, Kansas City, MO( ERIC Document Reproducti on Service No. ED 293 432). Weissman, J. Silk, E., & Bulakowski, C. (1997). Assessing developmental education policies. Research in Higher Education 38: 187-200. Windham, Patricia. (1996). College preparatory success rate report. Data Trend. Windham, Patricia. (2003). A comparison of community college student demographics by program areas. Florida Divisi on of Community Colleges. Data Trend. Wirt, J., Choy, S., Rooney, P., Provasnik, S.,and Tobin, R. (2004). The Condition of Education 2004 (NCES 2004-077), U.S. Departme nt of Education, Na tional Center of Education Statistics. Washington, DC: U.S. Government Printing Office. 104

PAGE 116

Appendices 105

PAGE 117

Appendix A Comparison of SAS Probability of Passing Ra tes and Descriptive Statistics (Actual) Passing Rates in Highest Level Co llege Preparatory Reading Course Comparison of SAS % Probability and Descriptive Statistics (Actual) Passing % in ReadingScaledLogSAS %ScaledLogSAS %ScaledLog SAS % ScoreoddsProbabilit y PassingScoreoddsProbabilit y Passin g ScoreoddsProbabilit y Passing 0-0.15340.8580.46 410.2181.24420.55 810.58131.7883080.6418163% 1-0.14430.8660.46 420.2281.25550.56 820.59031.8046020.6438262% 2-0.13530.8730.47 430.2371.26690.56 830.59941.8210440.6468352% 3-0.12620.8810.47 440.2461.27850.56 840.60851.8376360.648 4-0.11710.8890.47 450.2551.29010.56 850.61761.8543790.65 5-0.10810.8980.47 460.2641.30190.57 860.62661.8712750.652 6-0.0990.9060.48 470.2731.31380.57 870.63571.8883250.65484-9048% 7-0.08990.9140.48 480.2821.32570.57 880.64481.905530.656 8-0.08080.9220.48 490.2911.33780.57 890.65381.9228910.658 9-0.07180.9310.48 500.31.350.5741-5057%900.66291.9404110.66 10-0.06270.9390.480-1069%510.3091.36230.58 910.6721.9580910.662 11-0.05360.9480.49 520.3181.37470.58 920.6811.9759320.664 12-0.04460.9560.49 530.3271.38720.58 930.69011.9939350.666 13-0.03550.9650.49 540.3361.39990.58 940.69922.0121020.668 14-0.02640.9740.49 550.3451.41260.59 950.70832.0304350.67 15-0.01740.9830.50 560.3551.42550.59 960.71732.0489350.672 16-0.00830.9920.50 570.3641.43850.59 970.72642.0676030.674 170.000791.0010.50 580.3731.45160.59 980.73552.0864420.676 180.009861.010.50 590.3821.46480.59 990.74452.1054520.678 190.018931.0190.50 600.3911.47820.6051-6059%1000.75362.1246350.68 9 1-10 0 50% 200.0281.0280.5111-2074%610.41.49160.60 1010.76272.1439930.682 210.037071.0380.51 620.4091.50520.60 1020.77172.1635280.684 220.046141.0470.51 630.4181.51890.60 1030.78082.183240.686 230.055211.0570.51 640.4271.53280.61 1040.78992.2031320.688 240.064281.0660.52 650.4361.54670.61 1050.7992.2232050.69 250.073351.0760.52 660.4451.56080.61 1060.8082.2434620.692 260.082421.0860.52 670.4541.57510.61 1070.81712.2639020.694 270.091491.0960.52 680.4631.58940.61 1080.82622.2845290.70 280.100561.1060.53 690.4721.60390.62 1090.83522.3053440.70 290.109631.1160.53 700.4821.61850.6261-7061%1100.84432.3263490.7001-11 0 54% 300.11871.1260.5321-3043%710.4911.63320.62 1110.85342.3475450.70 310.127771.1360.53 720.51.64810.62 1120.86242.3689340.70 320.136841.1470.53 730.5091.66310.62 1130.87152.3905180.705 330.145911.1570.54 740.5181.67830.63 1140.88062.4122980.707 340.154981.1680.54 750.5271.69360.63 1150.88972.4342780.709 350.164051.1780.54 760.5361.7090.63 1160.89872.4564570.711 360.173121.1890.54 770.5451.72460.63 1170.90782.4788380.713 370.182191.200.55 780.5541.74030.64 1180.91692.5014240.714 380.191261.2110.55 790.5631.75620.64 1190.92592.5242150.716 390.200331.2220.55 800.5721.77220.6471-8062%1200.9352.5472130.71811-12 0 66% 400.20941.2330.5531-4051% Note:SAS LOGISTIC Database fall sessions 1997-2004 FTIC Flor ida Community College Students Computerized Placement Test. 106

PAGE 118

About the Author Although originally from Pittsburgh, Pennsylvania, Laura Smith graduated from Naples High School in Naples, Florida. Laura Smith received a bachelors degree in Early Childhood/Elementary Education from the University of South Florida in 1972, and a masters degree of science in Reading and Language Arts from Duquesne University in 1977. She began her teaching career while in the masters program. She was the first full-time developmental reading in structor at the then St. Petersburg Junior College in 1979 and established one of the first learning su pport centers in Florida at the Tarpon Campus. While at St. Petersburg College, Professor Laura Smith has taught developmental courses in mathematics, wr iting and reading. She entered the Ed.D program at the University of South Florida in 2001. Mrs. Smith has authored a reading test preparation booklet for the Florida CLAST, as well as a phonics guide for parents. Laura Smith has been a member of the Florida Developmental Education Association since the second year of its inception.