, IIIIIIIIHII IIIIIIII III IIIIIIIII 3 1293 1066 ‘T ' “w ESE-121’ 'W ‘rsa-I“ www.1- 1 was}! gun-M Tun --- ‘ '—--v ,I —— — This is to certify that the dissertation entitled A Test of Tinto's Theoretical Model of College Student Persistence Among Transfer Commuter Students at an Urban University presented by Mary Kathryn Desler has been accepted towards fulfillment ' of the requirements for Ph-D- degree in Jidnnational Administration Major professor Louis C. Stamatakos Date—lune—Z'S; 1985 MS U i: an Affirmative Action/Equal Opportunity Institution 0-12771 MSU LIBRARIES RETURNING MATERIALS: P1ace in book drop to remove this checkout from your record. . be charged if book is FINES wilI returned after the date 0" stamped below. T) I‘Q‘I'WA. A»; IV '1 9 \ “a On . ~ 2‘ J . _‘_ g‘ 2. {J fl {1%} y? m 10 £191: ‘95 €052 €001 APR 2 02003 {Hay 1103Q 62214 A TEST OF TINTO'S THEORETICAL MODEL OF COLLEGE STUDENT PERSISTENCE AMONG TRANSFER COMMUTER STUDENTS AT AN URBAN UNIVERSITY By Mary Kathryn Desler A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Educational Administration 1985 Copyright by Mary Kathryn Desler 1985 ABSTRACT A TEST OF TINTO'S THEORETICAL MODEL OF COLLEGE STUDENT PERSISTENCE AMONG TRANSFER COMMUTER STUDENTS AT AN URBAN UNIVERSITY BY Mary Kathryn Desler This study tested the explanatory power of Tinto's (1975) theoretical model of college student attrition in describing the persistence/withdrawal behaviors of 623 first-time transfer commuter students at an urban university. The investigator also tested the influence of finances on student persistence within the framework of Tinto's model. Data was collected in a two-step longitudinal process; multiple regression and path analysis were used to test whether or not the data was consistent with the model. The findings suggested that Tinto's model is a useful framework for understanding the process of persistence/withdrawal among transfer students in urban university environments. While the amount of variance explained in the dependent variable, persistence, was modest (22%), it was comparable with what is reported in other studies. Academic and social integration accounted for the largest percentage of variance in persistence, which tended to support Tinto's (1975) hypothesis that persistence is predicated, to a great extent, on a student's integration into the academic and social dimensions of the institution. Furthermore, the fact that academic integration had both a significant direct effect on persistence and an important indirect effect on persistence which was transmitted through institutional commitment, suggested that among the transfer commuter students in the study, persistence was defined largely by successful and personally satisfying interaction with the academic dimension of the institution and to a lesser degree by interaction with the social system As expected, the influence of a transfer student's background characteristics on persistence was primarily indirect, with the exception of gender. Controlling for all other variables in the model, men were more likely than women to be enrolled at the subject institution one year later. The more financial help a student received, the fewer hours he/she worked per week. The more number of hours a student worked, on the other hand, the less well integrated the student was in the academic and social dimensions of the university. However, despite this negative influence, the more hours a student worked per week, the greater the likelihood he/she would be enrolled at the subject institution one year later. ACKNOWLEDGEMENTS The completion of a project such as this would not have been possible without the input, guidance, and encouragement of several individuals. I am deeply grateful to them. Paul Duby Dan Gilman Cary North Ernest Pascarella Louis Stamatakos Elfrede Wedam Barbara Zusman ii TABLE OF CONTENTS Chapter 1: Introduction................... Background...... ...... . ..... . .......... Purpose Of StudYOOOOOOOOO......OOOOO... Need for Study.............. ..... 0...... Nature Of Study. 0 O O O O O O I O O O O O O O O O I O O O 0 Definition Of Terms 0 O O O O O O O O O O O O O O O 0 O O 0 Limitations 0 O O O O O O O I O O O O O O O O O O O O O O O O O O O overVieWOOOOOOOI.......COODOOOOOOOOOOOO Chapter 2: Tinto's Theoretical Model Of College Withdrawal........ Chapter 3: Literature Review............... V: Research on College Student Attrition.. Research on Transfer Students.......... Research on Commuter Students.......... Research Based on Tinto's (1975) Model. Summary................................ Chapter 4: Design and Methodology .......... The Path Diagrams...................... Data Collection........................ Sample................................. Measurement of Variables............... Data Analys18...‘00.00.000.00...............OOOOOOOOOOOOOOOOOO Chapter 5: Analysis of Results.................................... Setwise Regression Analysis... ...... ... iii 11 12 12 14 19 21 29 35 42 50 ' 53 53 60 62 65 71 76 76 Direct Effects........ Interaction Effects... sumarYOOOOOOOOOOOOOOO Chapter 6: Conclusions, Implications and Suggestions for Further Research.. OveIView. O O O I O O O O O I I O O ..... O O O O O O O O C I O O I O O O O I O O O O O O O O O O O O O O O O O 0 Presentation and Discussion of Major Findings ........... . ...... COflClUSj-ons O O I O O O O O O O ...... 0 O O O O O O O I O O O O O O O O O O 0 O O O O O O O O O O O O O O O . Implications.......... Suggestions for Further Research......... ...... ... Appendix A: New Student Orientation Survey......................... Appendix B: UIC Experience Survey... Appendix C: Results of Factorially Derived Scales Measuring Academic/Social Integration and Goal/Institu- tional comitmentOO......OOOOOOOOOOOOOOO0......00...... Appendix D: Means, Standard Deviations and Zero—Order Correlations Between Independent and Dependent Variables in Models 1 and 2............................ List of References........ iv 79 89 91 95 95 97 102 110 116 119 123 128 132 134 Table Table Table Table Table Table Table LIST OF TABLES Variables Related to Persistence as Reported in the Literature.000......0.00......00.0.00.........OOOOOCOOOOOO... 51 A Comparison of the Demographic and Academic Characteris- tics of the Total Population of First—Time Transfer Students, the Transfer Students Attending Orientation, and the Sample of Transfer Students in This Study... ........ . 63 R2 Increase Due to Different Variable Sets - Models 1&20..........OOOOOOOOOOOOOOOOO......OOOOOOOOOOOOOOOOOO.... 77 Standardized and Unstandardized Regression Coefficients For All Structural Equations (Model 1)....................... 80 Standardized and Unstandardized Regression Coefficients For All Structural Equations (Model 2)....................... 81 Results of Factorially Derived Scales Measuring Academic/ Social Integration and Goal/Institutional Commitment......... 128 Means, Standard Deviations and Zero-Order Correlations Between Independent and Dependent Variables in Models land ZOOOOOOIOOOOOOOOIOOO. ....... ......OOOOOOIOOOO 0000000000 132 Figure Figure Figure Figure Figure Figure 1: LIST OF FIGURES Tinto's (1975) Conceptual Schema for Dropout From COllegeOO......OOOOOOOOOOOOOOOOOO............OOOOOOOOOOOOOO The Reconceptualization of Tinto's Model (Pascarella, Duby & Iverson, 1983)...................................... Path Diagram of Tinto's (1975) Theoretical Model of College Student Persistence (Model 1)...................... Path Diagram of Tinto's (1975) Theoretical Model of College Student Persistence With Intervening Variable - Hours worked Per week (Medel 2)..........OOOOIOOOOOOOOO.... Significant Standardized Regression Coefficients in Path Diagram Form (MOdel1).........OOOOOOOOOOOOOOOOOO.... Significant Standardized Regression Coefficients in Path Diagram Form (Model 2)............................... vi 15 49 54 55 82 83 CHAPTER 1 INTRODUCTION Background There is little dispute. The past decade has not been an easy one for institutions of higher education. Hard economic times have left both public and private institutions with dwindling financial resources--a situation which is exacerbated by the likelihood of a substantial decline in enrollments as the number of traditionally-aged students (18 - 23 years old) diminishes (Francis, 1980). According to current state funding formulas, which are based on enrollment figures, most public colleges and universities will receive less state appropriations as enrollments decline (Davila, 1982). Similarly5 private institutions, which receive about 50% of their income from tuition and fees, will also face financial problems as a result of lower' enrollments (Carnegie, 1980). It is no wonder, then, that attrition and retention are areas of major concern at most institutions of higher education. College and universities are asking themselves the following questions: 1. What are the strengths and weaknesses of our institution? 2. What kinds of students choose to come to our college or university? 3. What encourages students to stay at our institution once they are enrolled? What are the reasons students cite for leaving this school? 4. Which students are likely to leave our institution during their first year? During the second or third year? Do students who withdraw transfer to other colleges or universities? Would a more accurate description of their leaving behavior be "stopping out," i.e., will they return to our institution in the future? 5. What is it about our institutional environment that either encourages or discourages the positive adjustment and growth of our students? Does the institutional environ- ment contribute to or detract from a student's determi- nation to stay, or is it merely a function of a student's pre—college traits? 6. What effect are our programs or services having on persis— tence/withdrawal decisions? Some of the information that is necessary to answer these questions is readily available to institutional decision-makers in the form of enrollment data (number, percentages, or trends). But ‘many of the questions can only be answered after research studies are conducted that explore students' interactions with the collegiate environment and whether or not such interaction affects subsequent decisions to stay or withdraw. A relatively new phenomenon in the field of higher education is the emergence of a number of large urban universities. According to Klotsche (1966), Goodall (1970), and Spaights (1980), the overriding characteristic of these universities is their urban nature and their willingness to respond to the needs of their communities. Davila (1982) expands this definition by citing four characteristics which urban universities share and four ways in which urban—oriented campuses can be dissimilar. In_ this particular report, she ‘notes that urban—oriented campuses share the following characteristics: 1. they are located in the inner city of major metropolitan areas; 2. they are fundamentally community and service—oriented; 3. they are largely non-residential; 4. they have non-selective admissions policies and a hetero- geneous student body (Davila, 1982, p. 1). In other ways, however, she states that they can be very dissimilar: 1. they can be two-year community colleges, four—year colleges or universities; 2. they can be either public or private; 3. they can have a long history of serving a largely local clientele or can be newly involved with this mission; 4. their student bodies may differ widely, reflecting both the traditions of the institution and the characteristics of surrounding neighborhoods (Davila, 1982, p. 1). Finally, Gusfield, Kronus & Mark (1970) take a somewhat different approach when they define urban universities; they picture the students who attend urban universities and the campus life as the distinguishing features of these institutions. 1. Urban campuses are characterized by an increasing propor- tion of minority students and first—generation college students, that is, those who are the first of their fami- lies to attend college. 2. The "campus life" of an urban university is quite different from that of a more traditional institution. Specifically, there may be none, at least not in the traditional sense. With faculty and students alike commuting to the institution, and the university itself competing with the attractions of the city, many informal, after-hours activities are forfeited. 3. Students at urban universities do not necessarily go to school, full-time; they' drift. in, they' drift out. They' are involved in a network of other kinds of relationships. College education in the city is something less central, less engrossing, and less of a discernible life cycle stage than we have come to think of it as represented in the traditional colleges. Despite its uniqueness, the urban university faces many of the same problems with which the older residential institution must grapple. The Division of Urban Affairs of the National Association of State Universities and Land Grant Colleges (NASULGC) reports that although enrollment trends are more encouraging for public urban institutions than for other categories of post-secondary institutions, their relative growth rates have also slowed dramatically in the past few years (Rudnick, 1983). ‘Moreover, their enrollment is (”3 a type (part-time, non-traditional, and often non-credit) that generates the least financial support from budgets based on full-time enrollment figures. Thus, urban universities are equally concerned about attrition and retention. Since the urban university is a relatively new phenomenon in the United States, less research has been conducted at these institutions. Therefore, urban university decision—makers have less information upon which to base program decisions which could impact attrition and retention (Davila, 1982). Since the 1960's, a second phenomenon has been quietly sweeping through the field of higher education, and it is particularly evident at urban universities: a rise in the number of transfer students who have attended at least one college or university before enrolling in the school from which they intend to receive a bachelor's degree (Goodale & Sandeen, 1971; Moore, 1981). For example, much like the other urban institutions in the United States, the University of Illinois at Chicago (UIC) depends, to a great extent, on transfer students from nearby community or city colleges and other universities to make up its beginning class of new students each year. In fact, 43% of the new students who enrolled at UIC for the first time in the fall of 1983 were first-time transfer students. However, this is not a new trend at UIC. A research memorandum issued by the Office of School and College Relations in February, 1983, indicates that during the ten year period from the fall of 1970 through the fall of 1979, approximately 402 of the new undergraduates at UIC entered as transfer students, while 60% entered as beginning freshmen (Anderson, 1983). However, Anderson (1983) goes on to report some rather disturbing figures; two years after approximately 1800 transfer students enrolled at UIC for the first time in the fall of 1979, only 382 of the community college transfers and 39% of the four-year transfers were retained. The comparable figure was 767. for students who entered UIC as beginning freshmen. Similar retention rates among transfer students at other urban universities are reported in literature (Avakian, MacKinney & Allen, 1982). Yet despite the fact that the number of students who transfer to urban universities contributes in a substantial way to the total enrollment fugures at these institutions, little is known about which transfer students are more likely to stay or leave urban university environments, why some students choose to stay or leave, and what can be done, if anything, to retain them. Purpose of the Study Within this context, the purpose of this study was to examine the process of persistence/withdrawal of first-time transfer commuter students at an urban university. There were three main {objectives)of this study: 1) to __’5E§t thewexp’lanatory‘”(power“of“ammodel developed by VincenthintQAinml975 in describing the persistence/withdrawal behavior of first-time transfer commuter students at an urban university; 2) to ascertain the relative influence of several variables or variable sets.in explaining persistence among first-time_transf¢r commuter students at an urban institution; and 3) to test the influence of finances on student persistence/ withdrawal behavior within the framework of Tinto's (1975) model. In keeping with Tinto's (1975) model, the following hypotheses were developed and tested in this study: 1. The integration variables, as a set, will account for the largest amount of variance in the dependent variable, persistence. 2. The influence of the pre-enrollment characteristics (including financial status) on persistence will be mediated by the student's commitments and first year experiences at the subject institution. More specifically, none of the pre—enrollment characteristics ‘will have a significant direct effect on persistence. 3. Initial commitments to the goal of graduation and to the subject institution will directly affect academic and social integration, respectively. 4. The influence of academic and social integration on per- sistence will be indirect, transmitted through subsequent measures of goal and institutional commitment, respectively. 5. Goal commitment and institutional commitment (measured one year after matriculation) will be the only two variables that will have significant direct effects on persistence. As such, they will be the best predictors of persistence. 6. The number of hours a student worked per week will indirectly influence persistence through several intervening variables. More specifically, the more hours a student works per week, the less integrated he/she will be in the academic and social dimensions of the university. Furthermore, the less well integrated the student is, the more likely it will be that the student will not be enrolled at the subject institution one year later. In addition to hypothesis testing, this study provided some answers to the following questions: 1. Does the environment of an urban university contribute to or detract from a transfer student's determination to stay, or is it merely a function of a student's pre-enrollment characteristics? 2° “8‘ features °f the£3355..-YeéF,..:_¥.PS.§iS9¢e of transfer students in urban university environments are most closely associated with persistence/withdrawal and which, if any, may be amenable to institutional program or policy action? 3. If transfer students are considered to be more of a hetero— geneous rather than.:1'homogeneous group as the literature suggests, what within the urban university environment encourages different groups of students to persist, or is it merely a function of the background characteristics they bring with them? Need for the Study In recent years the research literature on college student attrition has multiplied rapidly. However, as Bean (1982a), Spady (1970), and Tinto (1975) argue, much of the research has been largely descriptive in nature rather than theory-based, concentrating on characteristics of the institution and whether they are related to dropout or persistence. Few studies examine the interaction. of pre—college traits and students' experiences during the first few months, and few are longitudinal in design-—two ingredients which are considered essential if one is to fully understand the dropout process (Bean, 19823; Spady, 1970; Tinto, 1975). As a result, although there is a wealth of studies reported in the literature which relate statistically reliable associations between students with certain characteristics and attrition or persistence, few offer a complete explanation of the dropout process (Terenzini, Lorang & Pascarella, 1981). Furthermore, most of the research dealing 'with college student attrition has been conducted with samples of freshman students at residential colleges and universities; little is known about the persistence/withdrawal behaviors of transfer commuter students, especially in urban university environments. Therefore, this study was beneficial for two reasons. First, this study contributed to the literature on college student attrition by a) using the theoretical. model of student attrition. developed ‘by 'Tinto (1975) to explore the process of persistence/withdrawal among a population of students about whom little is known——transfer commuter students who attend urban universities; and b) fine-tuning Tinto's (1975) model to examine the role finances play in persistence. Secondly, in practical terms, this study identified some of the underlying causes of first-time transfer commuter student attrition at the subject institution. Since improving retention rates is not only a matter of academic interest but also of institutional survival, the results of this study may have direct implications for program or policy action and, therefore, might be helpful to university decision-makers. Nature of the Study This study was empirical in nature and was conducted at the University of Illinois at Chicago (UIC) which is an. urban commuter institution with an undergraduate population of approximately 20,000 and a graduate population of nearly 5,000. The sample was drawn from 1855 transfer students who enrolled at UIC for the first time in the fall of 1983. This investigator used the explanatory model developed by Vincent Tinto in 1975 to examine the process of persistence and withdrawal among first-time transfer commuter students at [NIL A complete test of the 10 model was conducted. In brief, Tinto's model is both longitudinal and complex. Persistence and withdrawal behaviors are considered to be largely an outcome of student's interactions with the academic and social systems of an educational institution. Although the model takes into account the attributes, skills, abilities, and commitments of an entering student, the model does not focus directly on these characteristics other than as they interface with the academic and social systems of the institution. A more complete description of the model can be found in Chapter 2. Tinto's model is based on the theory that a cause and effect relationship exists between a complex array of variables. As such, Tinto's model is a causal model and that determines, among other things, the type of data to be collected and the method by which they are to be analyzed (Kenny, 1979; Pedhazur, 1982). In keeping with Tinto's model, data was collected in a two-step longitudinal process. A survey assessing pre-college characteristics and initial commitments was distributed to first-time transfer commuter students attending a New Student Orientation Program during the summer of 1983 just prior to matriculation. A second survey seeking information about the student's actual experiences and subsequent commitments was mailed in May, 1984, after three academic quarters. Other information was ascertained through registration information provided by the Office of Admissions and Records (UIC) in the fall quarter 1984 after late registration. In order to adequately test whether or not the causal model was consistent with the data, two statistical techniques 'were employed: multiple regression and path analysis. Use of these two techniques to 11 quantify and interpret causal theory (models) has received a great deal of support in the literature (Alwin & Hauser, 1975; Kenny, 1979; Pedhazur, 1982). Definition of Terms In this dissertation study, several terms were used which need definition. The words attrition, withdrawal, and/or drop out were used interchangeably and were defined as the cessation of individual student membership in a single institution. This included students who voluntarily withdrew and those who were dismissed for academic reasons. As Bean (1982b) suggested, Students who are dismissed for academic reasons are also considered dropouts, despite the fact that they are not voluntary dropouts. They are included because students dismissed for academic reasons represent failure of the socialization process more than mental deficiencies and because excluding students who flunk out of school requires the arbitrary exclusion of extremely low values on grades(p. 292). The terms persistence and retention were used to describe a student's continued enrollment one year later at the subject institution. The phrase first-time transfer was used in this study to describe a student who had completed 18 or more quarter hours (15 or more semester hours) at another institution, but who was a new student at the subject institution. Finally, the term commuter was defined in its broadest sense in this study, i.e., a student who did not live on the campus. No attempt was made to determine whether the student lived at home, with his/her parents, or had his/her own apartment. 12 Limitations The fact that this study was conducted at a single institution from a single—year sample was both a limitation and a delimitation. A single institutional study from a single—year sample limits the generalizability of the findings to other groups of first-time transfer commuter students or to other institutions. Trends or generalizations remain suggestive until the model is tested with transfer commuter students via multiple institutional studies or is replicated at other single institutional sites in similar studies. On the other hand, conducting this study at a single institution from a single-year sample eliminated possible confounding variables based on group or institutional differences. Secondly, since this investigator operationally defined the dependent variable as first-year to second-year persistence versus withdrawal (voluntary and involuntary), it was impossible to distinguish permanent withdrawals from students who "stop out." Therefore, it is possible that the definition of withdrawal used in this study may be capturing some stop out behavior as well. Overview This dissertation study is discussed in five remaining chapters. Chapter 2 contains a complete discussion of the model developed by Vincent Tinto in 1975. In Chapter 3 the existing literature in four areas is reviewed: research (n1 college student attrition, research (n1 transfer students, research on commuter students, and research based on Tinto's model. This literature review suggested the variables which. were later used to operationally define the various elements in Tinto's model. The design and methodology of the study is described in Chapter 4. 13 Included in this chapter is a description of the sample on which the study was conducted, the operational definitions of the variables and how they were measured, and a discussion of the two statistical techniques used to analyze the data. Chapter 5 includes a presentation of the findings derived from the analysis of the data collected. Finally, Chapter 6 contains a summary of the study, a presentation of the major findings, conclusions, implications, and suggestions for further research. CHAPTER 2 TINTO'S THEORETICAL MODEL OF COLLEGE WITHDRAWAL Among the major reviews of the literature related to student attrition, the lack of theoretical models has created difficulties both in constructing and in comparing empirical studies (Bean, 1982a; Spady, . ,J , 1970; Tinto, 1975). The need for theory in the study of student M attrition was stated clearly by John Bean (1981) when he wrote, The advantage of using a theoretical perspective in designing research problems is well known. Theory guides research, and prevents either the reinvention of the wheel or the analyses of variables which show little potential for explaining the attrition process. Theory can potentially bring order to a confusing array of variables, each of which, in someone's study or other, has had a significant zero-order correlation with attrition (p. 4). Although a search of the literature revealed the existence of several other models of student attrition (Bean, 1981; Spady, 1970), the most widely cited and the most widely tested theoretical model of student attritidn was developed by Vincent Tinto (1975). A description of the model,presented in Figure 1, follows. Tinto (1975) hypothesizes that a student comes to :1 particular collegiate institution with a variety of ”family background characteristics (e.g., socioeconomic status), _individual attributes (e.g., gender, ability, ethnic background), and pre-college educational experiences (e.g., high school grade point average, high school rank, level. of high school involvement) whichgqlead" to Dthe ‘development of .. . u'{'\g.ro. M.” * féfidcational expectations And commitment5§ both to the goal of graduation ‘.fiJII—‘/'- : 14 15 Commitment. Academic System Commitments r’ _ '7 _ —l h Grade I dorm-nee. . ____ ,» - L» Academic _ _____ 13-th \ r _I I Intellectual | “'an \ l— -' ‘1 Goal PM < Development J \ Goal : J, I :Commitmen I L\__T_ '_l I Commitment I 1:22:33 L-Al '[———-_4*-——— ———— -{ 1+ 3.2m " I Institutional ' '— - —-_l ' : /V' -mmitment L‘ I lPeer-Group | ’J ' ' nterectione Soc. L______. ‘1 , »-~....:.:..x ----.J | acuity I L Interectoom J 32.] 3.2.}. Figure 1. Tinto's (1975) conceptual model for dropout from college (Copyright, American Educational Research Association, Washington D.C. Used by permission) from college and to the institution. According to Tinto (1975), theseugoal and institutional commitments are not only a reflection of a multidimensional process of interaction between individuals, their families, and their prior experiences in schooling, but also serve to influence how well individuals perform in college and how they will interact with, and subsequently become fiintegrated into an institution's_social and academic system. The degree to which individuals integrate into the academic system is measured by grade performance and intellectual development, while social integration) is *measured by peer group interactions, informal interaction with faculty, and involvement in extracurricular activities. Tinto (1975) indicates that "other things being equal, the higher the degree of integration into the college systems, the greater will be the student's commitmentmto the specific institution and to the goal of 16 geollegeucompletionf (p. 96). In the end, it is the interplay between the individual'sq_comm1tment to the goal of college ‘completion and the individual's commitment to the institution that determines whether or not the student decides to stay or drop outof college. Presumably, either low goal commitment or low institutional commitment can lead to dropout. It is important to note that Tinto's model is primarily concerned with accounting for the persistence/withdrawal _behavior within institutions (Tinto, 1975, 1982). Although it takes into account the attributes, pre—college experiences, and commitments of entering students, the model does not focus directly on these characteristics other than as they interface with the academic and social systems of the institution. A second observation is that Tinto's (1975) concept of academic and .ESEEEI, integration is analogous to Astin's (1984) theory’ of *student “involvement. .According to Astin (1984), "Student involvement refers to the amount of physical and psychological energy that the student devotes to the academic experience" (p. 297). Thus, a highly involved (or well integrated) student could be characterized as one who, "devotes considerable energy to studying, spends time on Tcampus, participates actively in student organizations and interacts frequently with faculty members and other students" (Astin, 1984, p. 297). Conversely, an uninvolved student (or one who is not integrated into the the academic and social dimensions of the institution), "spends little time on campus, abstains from extracurricular activities, and has infrequent contact with faculty members or other students" (Astin, 1984, pp. 297—298). The act of dropping out can be viewed as the ultimate form of noninvolvement. According to Astin (1984) the amount of student learning, personal 17 development, and/or rate of persistence associated with any educational program is directly proportional to the quality and quantity of student involvement in that program. A third observation is that Tinto includes ‘goal. commitment fiend“ institutional commitment twice in the model. One type of goal commitment, for example, seems to be the _product of pre-college characteristics and the second seems to be the “product of _academic integration. At any particular point in a student's career, however, the student will have a single notion of goal commitment, which is expected to be the product of both prematriculation characteristics and academic and social integration. Although Tinto does not elaborate on this point, one would expect that the latter set of goal commitments would be the best predictors of dropout decisions, not the initial set. Although Tinto (1982) claims that his model was developed to explain certain, not all, modes or facets of dropout behavior that may occur, he acknowledges that his model does not give sufficient emphasis to the role of finances in student decisions concerning persistence/withdrawal. In discussing this point, Tinto (1982) observes that much of the impact of finances occurs at the point of entry into the higher education system, i.e., it may induce students to enter particular institutions (institutional commitment) and may influence the initial weighting of the costs and benefits of higher education attendance “(goal commitment). Furthermore, it may also be the case that "financial need ”will have a greater imp'arctwupon“dropout.,§arly..in the educational career when the degree goal is still quite distant. At any rate, Tinto (1982) urges researchers to fine-tune their models of college student attrition to account for the way in which finances may influence persistence. 18 In summary, Tinto's theoretical model of 'collegehgwithdrawal, is composed of the fellowing basic elements: pre-college characteristics ...-H,‘ - -./ ru»,,, nice . . ‘— (including family background, individial attributes, and past educational experiences); goal cgmmitment, (measured. both before and after ~ ...--.‘4 “We-7 matriculation); institutional commitment (measured both before and after matriculation); academic _ .-..‘an. intergration; social integration; and persistence/withdrawalwdecisions. ”WTintO's chief support for the inclusion of these elements and the hypothesized relationships believed to exist between them in the model, comes from the literature on college student attrition. This body of literature, as well as a review of the studies which have been conducted to test the explanatory power of Tinto's model, are the subject of Chapter 3. CHAPTER 3 LITERATURE REVIEW Research in four separate areas is reviewed in this chapter: research (n1 college student attrition, research (n1 transfer students, research on commuter students, and research based on 'Tinto's ‘model. The research on college student attrition was reviewed in order to more clearly define the basic elements in Tinto's model and to generate a list of variables which were later operationally defined and measured in this study. No research studies were located that directly related to the persistence/withdrawal behaviors of transfer commuter students at either residential or urban institutions. Therefore, what could be concluded about the persistence/withdrawal tendencies of transfer commuter students at urban universities had to extrapolated from research in two separate areas: research on transfer students and research on commuter students. The review of this literature indicated some additional variables which should be included when testing the explanatory power of Tinto's model in explaining the persistence/withdrawal behaviors of transfer commuter students. Tinto's model has been the focus of a number of studies which are reported in the literature. A review of these studies provided the investigator with the information needed to replicate a test of the model, e.g., the operational definitions of the variables, the methods used to collect the data, and the statistical techniques used to analyze 19 20 the data. There were several characteristics of the research in these four areas which limited its usefulness to this study. First, most of the research on college student. attrition. lacked :1 multivariate research approach which seems essential if one is to understand the simultaneous effects of a complex array of factors which contribute to a student's decision to dropout of college. ‘Using any technique other than a multivariate one, obscures important details of student withdrawal, i.e., it is difficult to isolate the independent effects of various factors on college dropout. Secondly, what we know about transfer students from the literature is based primarily on comparisons of academic achievement of junior college students (measured by grade point average) with native students (students who began and stayed at a particular institution). Very little is known about students who transfer from one four—year institution to another. Furthermore, most of these studies were conducted at traditional residential institutions making it difficult to generalize their findings to an urban commuter institution. Thirdly, much of what we purportedly know about commuter students comes from research studies which compare commuter students with a known quantity——residential students. Most of the studies reviewed in this chapter are demographic surveys conducted at primarily residential colleges and universities which also have a population, however large or small, of commuter students. The results of these studies are hard to argue with-~commuter students who attend residential colleges or universities (k) differ substantially from students who live on campus. In fact, although controversial, such comparisons have left many /. V 21 educators with. a ‘negative view of commuter students. For example, Chickering (1974) writes that the differences between residential and commuter freshmen are the differences between the "haves" and the "have nots." Recently, however, a few researchers (Foster, Sedlacek & Hardwick, 1978; Pettyway, 1968; Rue & Stewart, 1982) have begun to treat commuter students as a heterogeneous rather than a homogeneous group. Although this perspective will likely lead to a less negative view of commuter students, most of the research reported in this chapter does not treat commuter students as a diverse group. Similarly, although Tinto's model has been the focus of a number of studies which are reported in the literature, they all have one common characteristic which limits their usefulness to this study: they are based on data collected from samples of freshman students. Research on College Student Attrition It seemed impractical to report all of the studies dealing with college student attrition. Therefore, in addition to Astin's work (1975, 1977), five articles from the literature which review the attrition-related research between 1925 and 1975 (Cope, 1978; Pantages & Creedon, 1978; Spady, 1970; Summerskill, 1962; Tinto, 1975) encapsulate the literature review. These particular articles were selected for review because of their comprehensive and thorough approach to attrition—related research. Pre-College Characteristics. Pre-college characteristics represent facts about students before entering college. These variables precede the student ' 8 interaction with the academic and social systems of the institution. However, it is expected that these variables will contribute little to the explained variance of dropout when information 22 is known about a student's academic and social integration. The most important of these variables is probably pre-college schooling (high school grades and transfer GPA). According to Pantages & Creedon (1978) and Summerskill (1962), the general conclusion they drew from the literature was that age, per s3, did not affect attrition. In short, the same factors which cause some students to delay entrance into college may continue and cause students to drop out. Although gender did not appear to be a significant variable in determining persistence, it did become significant when scholastic, environmental, institutional, and longitudinal factors were taken into consideration. For example, men seemed to finish college in greater proportions than women, and women tended to voluntarily withdraw as opposed to being dismissed for academic reasons (Pantages & Creedon, 1978; Tinto, 1975). Astin (1975) reported that women were more likely than men to complete a degree in four years. He also found that among freshman women, those who were married or had marriage plans were most likely to dropout, although among male freshmen, getting married at the time of college entrance was positively related to persistence. The evidence was not clear regarding the relationship between socioeconomic status (as measured by parents' level of education, parents' income level, or father's occupation) and attrition. Astin (1975), Cope (1978), and Tinto (1975) indicated that the family's socioeconomic status appeared to be inversely related to dropout. Specifically, students from lower status families exhibited higher dropout rates than did students from higher status families even when intelligence had been taken into account. But Spady (1970) cautioned 23 against drawing too many conclusions from such studies because he claimed that the independent influences of these factors on attrition were not well documented. Numerous, more specific findings seemed to exist that cited particular family attributes as being related to students' persistence in college. The most important of these factors appeared to be the quality of relationships, the central values expressed within the family, and the interest and expectations parents had for their child's education (Spady, 1970; Tinto, 1975). College persisters tended to come from families whose parents enjoyed more open, democratic, supportive, and less conflicting relationships with their children. Furthermore, persisters seemed not only to get ‘more parental advice, praise, and expressed interest in their college experiences, but they also had parents who expressed greater expectations for their child's further education (Tinto, 1975). Pantages & Creedon (1978) reported ambiguous results regarding the relationship of ethnicity to persistence. They concluded that ethnicity alone did not seem to be a major factor when academic factors were held constant. However, Astin (1982) contended that whites were much more likely to complete a bachelor's degree than were blacks, Hispanics, or American indians. All of the reviewers agreed that high school grade point average, high school rank, and scholastic aptitude were the best predicators of persistence/withdrawal, and they accounted for half of the ‘variance between those who persisted and those who withdrew (Astin, 1975; Cope, 1978; Pantages & Ckeedon, 1978; Spady, 1970; Summerskill, 1962; Tinto, 1975). Of the three, Tinto (1975) indicated that past grade performance 24 was the best predictor of persistence. Undergraduates usually pay their college expenses through one or a combination of different sources of aid: family (parent or spouse), scholarships, loans, and work. Astin (1975) is one of the few researchers who has studied the impact that finances have on a student's ability to stay in college. Like Tinto (1982) hypothesized, Astin (1975) seemed to feel that the impact of finances was felt most acutely prior to matriculation. Astin (1975) included several variables related to a student's financial status in his research. He summarized his findings as follows: 1. Receiving support from parents or spouse enhanced a student's ability to complete college. 2. Scholarships or grants were associated with small increases in student persistence rates. The amount of aid seemed to be the most important factor related to persistence. 3. Reliance on loans was associated with decreased persistence among men, the effects were highly variable among women, and were associated with increased persistence among blacks. 4. Participation in federal work-study programs seemed to enhance student persistence, particularly among women and blacks. 5. Reliance on savings or other assets appeared to decrease a student's chances of finishing college. Astin (1975) also examined the relationship between student employ- ment. and. persistence. In. this context, he concluded. the following: 1. Having a part-time job usually increased the student's 25 chances of persisting. 2. For most students who worked full-time, the positive effects of employment on persistence were not only lost, but actually reversed. 3. On-campus jobs ‘were preferable to off-campus jobs when related to persistence. 4. Students with off-campus jobs were more likely to drop out if their work was related to their career goals. Finally, Astin (1975) indentified several other pre—college characteristics which were significant predictors of persistence/ withdrawali They’ were religion, high school size, citizenship, and marital status. Pre-College Commitments. According to Tinto (1975), most researchers seemed to concur that once an individual's ability was taken into account, it was commitment to the goal of college completion that was most influential in determining college persistence. In fact, the higher the degree aspiration, the more likely the individual was to remain in college. Spady (1970), however, reviewed several studies suggesting that lofty goals positively influenced graduation only when they were clear and realistic. In order :3: determine what was "clear and realistic," researchers combined measures of academic competence and college commitment (Pantages & Creedon, 1978). These results indicated that students with high academic competence and moderate to high college commitment were most likely to persist. Students with high competence but moderate to low commitments tended to transfer to other colleges, or dropout and re-enroll at a later time. Students with low competence but 26 with moderate to high commitment tended to persist in college until they were dismissed for academic reasons. Finally, students with both low commitment and low competence were likely to dropout and were unlikely to ever re-enroll at any college. Further, the nature and strength of college goals were differentially linked to certain outcomes depending on the gender of the student, i.e., female dropouts seemed to 'have lower levels of goal commitments relative to persisters than did male dropouts (Spady, 1970; Tinto, 1975). Spady (1970) claimed that the substance of student's educational and curricular interests was also related to attrition. For example, students who were dismissed for academic reasons typically failed to see their education as :1 process involving intellectual growth. Instead, Spady (1970) reported, they were extremely utilitarian in outlook. Academic and Social Integration. Persistence in college is not merely the outcome of individual characteristics, prior experiences, or prior commitments, however. According to Tinto (1975), persistence in college is the outcome of a longitudinal process of interactions between the individual and the institution; the institutional environment being composed of both an academic dimension and a social dimension. Academic integration involves a student's performance in college (grade point average) and his/her intellectual development. Social integration is measured by a student's relationship to his/her peers, informal interaction with faculty, and involvement in extracurricular activities. Astin (1975) and Summerskill (1962) both reported that a highly significant relationship existed between attrition and first semester (:ollege grades. But Summerskill (1962) cautioned that poor grades were a 27 far more stable predictor of attrition than were good grades a predictor of retention. Therefore, he felt it was important to distinguish between dropouts who were academically dismissed from those who were voluntary withdrawals. In fact, Tinto (1975) wrote that if students with unsatisfactory grades were excluded from research samples, those students who withdraw "generally show both higher grade performance and higher levels of intellectual development than do the average persisters" (p. 117). When gender was taken into account, male students were more often academic dismissals than were females. According to Pantages & Creedon (1978), study habits, while not a powerful predictor of attrition, certainly played some role in determining the likelihood of persistence or withdrawal. Intellectual development was also related to persistence in college (Tinto, 1975). As was reported in the section on commitment, several studies indicated that students who valued education as a process in gaining knowledge, rather than a mere process of vocational development, tended to be those who persisted. Spady (1970) reported in his review of the literature that intellectual development was more related to persistence among females than among, males, probably reflecting the traditional. societal. expectations. Both. Summerskill (1962) and 'Tinto (1975) included the concept of person-environment fit or congruency when they wrote about intellectual development. They concluded that it was not simply the absence or presence of intellectual development which was important, but the degree and the prevailing intellectual climate of the institution. Voluntary withdrawal, then, became a means of "coping" with the lack of congruency between the intellectual development of the individual and the climate of the academic system (Tinto, 1975). 28 Integration into the social system of the college involves interaction with peer groups, involvement in extracurricular activities, and informal interaction with faculty. Cope (1978), Tinto (1975) and Pantages & Creedon (1978) all cited several studies which seemed to support the notion that peer support was associated with persistence. However, Spady (1970) claimed that researchers found this was true only if friendship ties existed with people having strong academic orientations. Other studies seemed to lend support to the obvious fact that patterns of excessive and largely superficial socializing were more descriptive of dropouts than of persisters. The absence of supportive groups was, in turn, more often associated with voluntary withdrawal than it was with academic dismissal, particularly for females (Spady, 1970). The majority of the research findings indicated that participation in extracurricular activities for both genders was positively related to persistence (Astin, 1975, 1977; Pantages & Creedon, 1978; Spady, 1970; Summerskill, 1962; Tinto, 1975). But Pantages & Creedon (1978) described one study which found that students who dropped out were more likely to come from the two extremes of the spectrum: they either participated to a very great degree, or not at all. Nonetheless, Tinto (1975) considered that participation in extracurricular activities heightened a student's commitment to the institution and, therefore, reduced the probability of that person dropping out of college. Finally, research findings also indicated that informal interaction with faculty was related to persistence in college (Astin, 1977; Pantages & Creedon, 1978; Tinto, 1975), especially when the interaction focused on discussions of intellectual or course-related matters (Pascarella, Duby, Terenzini & Iverson, 1983; Pascarella & Terenzini, 1977). 29 Summary} The research on college student attrition indicated that many variables affect whether students decide to stay or leave various collegiate environments. Some of these are linked to the characteristics the students bring with them, while others are more related to students' interaction with the academic and social systems of the institution. In order to determine if there were additional 'variables ‘which should be included when testing the explanatory power of Tinto's model among transfer commuter students, a review of the literature relating to transfer students and commuter students was conducted. Research on Transfer Students Pre-College Characteristics. ‘When compared. with. native students (students who began and stayed at a particular institution) enrolled at four-year institutions of higher education, transfer students seemed to exhibit the following pre-college characteristics: 1. They scored lower on tests of academic ability and performed less well academically in high school (Cross, 1969; Goodale & Sandeen, 1971); 2. They came from families of lower economic status and with less formal education (Cross, 1969; Goodale & Sandeen, 1971; Knoell, 1965); 3. They tended to be less independent, reflective, tolerant, and self-confident and were more conventional and pragmatic (Cross, 1969; Goodale & Sandeen, 1971: Laudicina, 1974); 4. They tended to spend more time working while attending college (Goodale & Sandeen, 1971); 5. They were motivated to attend college primarily by the promise of economic rewards and upward mobility (Laudicina, 30 1974). More recently, Moore (1981) wrote that despite the research available, it was difficult to accurately describe a transfer student. Rather than assume they all shared the same characteristics, she described transfer students as :1 diverse group including students who 1. transferred from a two-year to a four-year college or from a four-year college to another four-year college, or from :3 four-year college to a two year college; 2. brought credits two months old, two year old, or twenty years old; 3. were as diverse as military personnel and displaced home— makers. In short, it appears as though transfer students are demographically a heterogeneous rather than a homogeneous group. Pre-College Commitments. According to Goodale & Sandeen (1971), transfer students who entered four-year institutions did not aspire to educational and occupational goals as high as those aspired to by native students. Furthermore, Buckley (1971) and Donato (1973) concluded that prior to matriculation, transfer students' expectations of the academic environment at the four-year institutions were highly incongruent with the actual situation as perceived by native upper class students. According to these researchers, transfer students exaggerated the intellectual and nonintellectual climate of the university and, although these expectations did not impede later academic performace, they were a source of irritation and required the student's adjustment. Academic and Social Integration. We seem to know very little about the transfer students' actual college experiences. Knoell & Medsker 31 (1965) found that many transfer students who had been student leaders at their previous institutions became almost totally inactive after transfer when they failed to find assignments for which they did not have to compete with freshmen. And, when compared with native students, Knoell & Medsker (1965) claimed transfer students tended to have much less identity with the four-year college. The majority of the studies conducted on transfer students attempted to measure transfer students' academic success in terms of grade point average. While there were variations in the results from study to study, there were several strikingly consistent trends in the academic performance of transfer students at senior institutions: 1. There was a decline in average grades immediately after transfer (Cross, 1969; Hills, 1965; Hood, 1967; Knoell & Medsker, 1965; Laudicina, 1974; Webb, 1971). 2. The average grades of transfer students improved in sub— sequent quarters (Hills, 1965; Knoell. & Medsker, 1965; Laudicina, 1974; Webb, 1971). 3. Students transferring from four—year institutions performed better academically their first term than did students who transferred from two-year colleges (Goodale & Sadeen, 1971; Hills, 1965; Young, 1964). 4. Native students tended to earn higher grades than transfer students throughout their careers (Goodale & Sandeen, 1971; Hills, 1965; Webb, 1971). 5. Community college grades were the best predictor of academic success at four-year institutions (Nickens, 1970; Phlegar, Andres & McLaughlin, 1981; Wray & Leischuck, 32 1971). Persistence/Withdrawal Decisions. There were few studies which attempted to explore the persistence/withdrawal process among transfer students. Knoell. & ‘Medsker (1965) compared the attrition rates of transfer students with those of native students in ten states and concluded that, overall, transfer students were more likely to drop out, took longer to graduate, and did so in smaller proportions. Although Lindsay, Marks & Hamel (1966) found much the same phenomenon at Pennsylvania State University, they concluded that "non-ability factors played an important, if not the most important role in the achievement and drop out of native and transfer students" (p. 12). Lara (1981) identified some of these non-ability factors which seemed related to retention in a study completed at UCLA. He found that when successful and unsuccessful community college transfer students were compared, success could only be predicted by the student's grade point average, resolution of financial need, relationships with faculty members, community college writing experience, and a major in either the fine arts or social sciences. Goodale & Sandeen (1971) and Wessel, Engle & Smidchens (1978) identified two additional nonability factors which seemed to separate those transfer students who persisted from those that withdrew: financial problems and indecision regarding curricular or career goals, respectively. Only three major studies included all three of the parameters pertinent to this study: transfer students, urban universities, and persistence/withdrawal. In a study conducted at the University' of Missouri-St. Louis, Avakian, MacKinney & Allen (1982) compared first-time freshmen and transfer students from the fall of 1975 through the fall of 1979. race, 33 and grade point average; the dependent variable persistence/withdrawal. The results revealed the following: 1. The retention rate of first—time freshmen was higher than the retention rate of transfer students, although the cumulative percentage of graduates among transfer students by the fourth. year ‘was about double the percentage of graduates among first-time freshmen; Men and white students had higher retention rates than did the women and the black students, respectively. Grade point average was strongly and positively related to retention in both first-time freshmen and transfers. The independent variables that were included in study were gender, was These findings are supported by a study completed at the University of Illinois at Chicago. Anderson (1983) analyzed the academic progress of community college transfers, senior college transfers, and continuing sophomores and juniors (natives) through eight terms, years, 1. or two academic after transfer. The results of this study are summarized below. The retention rate of first—time freshmen (76%) was higher than the retention rate for either two-year college transfers (38%) or four-year college transfers (39%). Six percent of the two-year college group and 8% of the four- year group had graduated two years after transfer. At the same time, almost 462 of the native sophomores and juniors had graduated. Both two-year and four-year transfer students experienced a first term drop in mean grade point average. The two-year college transfers did not recover their pre-transfer grade 34 point averages during the two years included in this study; four-year transfers exceeded their pre-transfer grade point average during the sixth and seventh terms. 3. By the end of the second year following transfer, 282 of the community college transfers and 17% of the four-year college transfers had been dropped or left while on aca- demic difficulty. Finally, when Zusman (1984) studied undergraduate transfer student retention at the University of Illinois at Chicago, (UIC), she noted some striking differences in demographic and academic characteristics when new undergraduate transfers who withdrew after one quarter were compared with all new undergraduate transfers and all undergraduates. This comparison revealed that transfers who dropped out after only one quarter were, on the average, two years older than all new transfers, and three years older than all undergraduates. The percentage of non—degree students was much higher in the group of transfers who dropped out (18.9%), than in the groups of all new transfers (8.9%) and all undergraduates (3.1%). In addition, transfers who dropped out were more likely to have enrolled in the college of liberal arts and sciences (71.4%) than all new transfers (63.72) or all undergraduates (53.1%). Furthermore, while over half of the transfers who dropped out were enrolled part-time (50.72), one-fourth of the new undergraduate transfers (26.7%), and only one-fifth of all undergraduates (20.1%) were part-time students. Based on her findings, Zusman (1984) concluded that those new undergraduate transfer students who dropped out after only one quarter at UIC appeared to have many of the characteristics of the so—called nontraditional student. 'They' were older, more likely to ‘be taking 35 classes part-time, and more likely to be classified as non-degree students. Summary. The review of the literature on transfer students suggested several additional variables which ought to be measured in this study. They were: initial college of enrollment, part—time/full-time status, transfer institution (two—year/four-year), transfer grade point average, and transfer hours. In addition to identifying these variables, the literature seemed to suggest that transfer students were more of a heterogeneous rather than a homogeneous group. Therefore, it seemed as though any study which sought to explore the underlying causes of attrition among transfer students should look at the differences between various types of transfer students. Research on Commuter Students Pre-College Characteristics. Demographic studies which compared commuter students with residential students generally pictured commuter students as follows: 1. they tended to be older (Chickering & Kuper, 1971); 2. they came from lower socioeconomic backgrounds (Drasgow, 1958; Dressel & Nisula, 1966; Fenske & Scott, 1972; George, 1971; Glass & Hodgin, 1977); 3. a large percentage were married (Glass & Hodgin, 1977); 4. a larger percentage of men commuted to school than women (Glass & Hodgin, 1977). In terms of their past educational experiences, commuter students tended to be less able academically than their residential peers. Researchers have also found that commuter students seemed to have lower 36 high school grades and ACT scores, and. most commuter students have participated in fewer than the average number of high school extracurricular activities (Chickering, 1974; Fenske & Scott, 1972; Palm, 1981). Other investigators (George, 1971; Kysar, 1964; Minkevich, Rickey & Marshall, 1972; Schuchman, 1974; Stark, 1965) who studied the mental health, personalities, and psychological differences 'betweenT commuter students and residential students placed the commuter student at the lower end of the "have not" continuum. For example, Schuchman (1974) and Kysar (1964) concluded that commuter students were beset with feelings of inferiority and inadequacy, often making them unable to break the ties with home and to move in a more self-directed pattern. In another study, Minkevich, Rickey & Marshall (1972) administered the Edwards Personal Preference Schedule to a group of commuter students who were attending either a two or a four-year college. They found that when compared with commuter students attending four-year colleges, commuter students attending two-year colleges tended to be less autonomous, more dependent, more likely to do what was expected, and to let others make decisions for them. They also tended to conform to custom and to avoid the unconventional. But, commuter students attending four-year colleges were more inclined to form strong friendships, to do things with friends rather than alone, and to participate in friendly groups. Furthermore, Bishop & Snyder (1976) and Dressel & Nisula (1966) noted that the fact that commuter students worked more and held a wider variety of jobs seemed to indicate they were more independent and self-reliant than their residential counterparts. Pre-College Commitments. In terms of pre-college commitments to 37 either the goal of graduation or to the particular institution they were attending, commuter students were depicted in the literature as having lower degree aspirations than residential students (Burnett, 1982; Dressel & Nisula, 1966; Fenske & Scott, 1972), and as choosing an institution more on the basis of proximity and cost than on their commitment to the particular college or university (Dressel & Nisula, 1966; Fenske & Scott, 1972; Trivett, 1974). Furthermore, other studies concluded that prior to matriculation, commuter students indicated less interest than residential students in participating in extracurricular or social activities at the institution in which they were enrolling and, in general, reported that vocational preparation or improving their present social and economic status was their primary purpose in attending college (Arthur, 1977; Burnett, 1972; Chickering, 1974; Dressel & Nisula, 1966; Flanagan, 1976; Glass & Hodgin, 1977; Palm, 1981). Academic and Social Integration. What we know about commuter students" actual experiences in college is based primarily on studies comparing commuter students with residential students at predominantly residential campuses. It seems self—evident that the experience of attending college is much different for a commuter student than it is for a residential student. After studying the commuter student at Wayne State University (Detroit), Ward & Kurz (1969) described the educational experience of a commuter student as that of the "divided life." The "divided life" they described involved students' academic pursuits, their family and home obligations, and their work responsibilities. It comes as no surprise, Ward & Kurz (1969) concluded, that many commuter students arranged their 38 academic schedules so as to minimize their on-campus time. This fragmented life seemed to affect the degree to which commuter students became integrated into the academic and social systems of the college or Indyersity. In fact, Astin (1977) wrote, "Apparently the impact of the college experience is diluted for commuter students and for students attending public institutions" (p. 71). Many researchers (Dressel & Nisula, 1966; Eddy, 1959; Flanagan, 1976; Glass & Hodgin, 1977) concluded that this lack of involvement made it difficult for commuter students to subsequently identify with the institution they were attending. On the basis of his longitudinal study, Astin (1977) went even further to suggest that this lack of involvement affected persistence. But Dressel & Nisula (1966) wrote that this may be something over which the institution has little control because, "this detachment seems to be as much a part or more the result of the student's preference or nature than it is the result of lack of attention to the commuter students on the part of the institution ‘he (she) attends" (p. 45). There seemed to be little difference between commuter students and residential students when it came to measuring intellectual growth in terms of academic achievement (Call, 1974; Dressel & Nisula, 1966; Graff & Cooley, 1970). In fact, it was reported that commuter students seemed more likely than their residential counterparts to identify with the academic community. Moreover, a commuter's classroom experience and informal contact with faculty seemed to make the greatest positive impact on both their academic achievements and satisfaction with the college experience in general (Davis & Caldwell, 1971; Iverson, 1982; Keller, 1980; Liu & Jung, 1980; Pascarella, Duby, Terenzini & Iverson, 1983). 39 There were numerous attempts to ‘measure the commuter student's college experiences and how that experience (coupled with their pre-college characteristics) influenced subsequent intellectual and personal growth. The results, however, seemed inconclusive. .Although some researchers reported that when compared. with their residential counterparts, commuter students developed more slowly in areas of personal growth (Demos, 1966; Schuchman, 1974; Ward & Kurz, 1969; Welty, 1976), Chickering & ther (1971) asserted that the slower change rate found in commuter students might have made the changes which did occur more permanent. On the other hand, in a study conducted at Columbia University Teacher's College, Le Moal (1980) found that commuter students and residential students did not necessarily differ initially and grew to the same extent on selected measures of personal growth in their freshman year. Academically, the picture was brighter for the commuter student than it was socially; Numerous studies indicated that commuter students tended to be much less involved in social activities than did residential students either because of their "divided lives" or due to lack of interest (Astin, 1975, 1977; Baird, 1969; Burnett, 1982; Chickering, 1974; Chickering & Kuper, 1971; Christian, 1973; Davis & Caldwell, 1977; Demos, 1966; Dressel & Nisula, 1966; Johnson, 1981; Trivett, 1974; Ward & Kurz, 1969). Astin (1975) claimed that this reduction in the socializing effects of the educational experience may have been of minor importance to many non-traditional students-~those who ‘were 'married, older, or attending part-time. But it was the traditional-aged student who paid the larger price for the decreased involvement in the educational process. 40 Because Astin (1977) found :1 strong, positive association between involvement and persistence, he suggested, "any programs that involved commuters more in campus life will have a positive effect on persistence" (p. 133). Counelis & Dolan (1974) made an interesting observation, however. Based on a study completed at the University of San Francisco, they postulated that perhaps commuter students were involved, but that the setting for their involvement was off—campus rather than on-campus and, therefore, that it was difficult to measure. Persistence/Withdrawal Decisions. After Burnett (1982) reviewed the literature on commuter students, he concluded that when compared with residential students, commuters had higher attrition rates and a greater tendency to take longer than four years to complete a degree. The results of Astin's (1977) research seemed to support this finding. Astin (1977) went on, however, to identify the variables most closely related to the persistence/withdrawal tendencies of commuter students—-those which are designed to measure academic and social integration. In a study conducted at the University of Illinois at Chicago (UIC) (then the University of Illinois at Chicago Circle), Zaccaria & Creaser (1971) investigated the differences in ability, personality characteristics, and social status between commuter students who withdrew (either voluntarily or involuntarily) and those who graduated within five years of matriculation. They concluded that a major factor contributing to attrition in commuter students at UIC was unsatisfactory grade performance. However, when scholastic standing *was controlled, the effects of several pre-college characteristics could be seen more clearly. 1. Males from lower socioeconomic backgrounds were more likely 41 to drop out when confronted with the prospect of academic failure and dismissal than were higher status males with comparable ability. 2. Commuter students who continued until graduation had different personal needs than students of similar ability who chose to withdraw. For example, commuter students who withdrew in good academic standing appeared to be less conforming to rules, regulations, and expectations of others. The male withdrawals seemed to be more assertive and the females seemed to have greater heterosexual concerns in comparison to the male and female commuter students who graduated. For commuter students, life off the campus is likely to be as important, if not more important, that their experiences with the academic and social life on campus (Eddy, 1959). Glass & Hodgin (1977) cited five reasons why some students commute to college; these seemed to indirectly identify some specific variables that could affect persistence/withdrawal decisions: limited financial resources, a desire to retain their identity with their local community, the need to continue their present job, parental influence, and institutional proximity. Many of these same characteristics were identified by other researchers (Arthur, 1977; Bishop & Snyder, 1976; Chickering, 1974; Dressel & Nisula, 1966; Eddy, 1959; Stark, 1965; Ward & Kurz, 1969) as being particularly problemmatic for commuter students, but no attempt was made to specifi- cally link these with attrition or retention. Summary. Although the review of the literature on commuter students did not reveal any new variables that should be included in this study, 42 it did underscore the importance of several variables already identified in explaining the persistence/withdrawal behaviors of commuter students--age, gender, ethnic background, financial factors, peer group interaction, informal interaction with faculty, involvement in extracurricular activities, and first semester (quarter, year) grade point average. Research Based on Tinto's Model Tinto's model has been the focus of a number of studies which are reported in the literature. All of these studies bore one common characteristic: they' were based on data collected from samples of freshman students. In other ways, however, the studies differed. Some were conducted at residential institutions, others at two and four—year urban commuter institutions. Some involved data collected from single institutional samples, while others were multi-institutional nature. With the exception of the study done by Pascarella, Duby, Miller & Rasher (1981), all of the researchers defined the dependent variable as being voluntary withdrawal. Finally, the statistical procedures used to analyze the data varied from study to study. By far the majority made use of multivariate techniques, while some employed path analysis which allowed the researcher to study the direct and indirect effects and the magnitude of certain variables on persistence/withdrawal. Although the results of the studies conducted at both residential and urban commuter institutions generally supported a number of Tinto's theoretical expectations, some interesting variations in patterns of influence were seen in the different kinds of institutions. Residential Institutions. As Tinto (1975) predicted, in each of the studies conducted at residential institutions (Baumgart. & Johnstone, 43 1977; Braddock, 1981; Munro, 1981; Pascarella & Chapman, 1983; Pascarella & Terenzini, 1977, 1979; Terenzini & Pascarella, 1978, 1980) the characteristics students brought with them to college did not account for a significant amount of the variance found between those students who persisted and those students ‘who 'voluntarily' 'withdrew. More specifically, when Munro (1981) and Pascarella & Chapman (1983) conducted a path analysis of data collected from multi-institutional samples, they found that the effects of background characteristics were mainly indirect, transmitted through college experience variables. However, Terenzini. & Pascarella (1980) reported that background traits (particularly gender) seemed tn) ‘be influential in their interactions with students' experiences following enrollment. These results led these two researchers to conclude that at residential campuses, different features of the institutional environment may have different influences on different kinds of students. Nonetheless, as far as the persistence/withdrawal process was concerned, what happened to students during their freshman year seemed more important than the students' pre—college characteristics. Tinto (1975) argues that both goal and institutional commitment will have direct effects on persistence/withdrawal decisions. Research conducted at residential institutions seemed to support this hypothesis. In the studies involving multi—institutional samples (Munro, 1981; Pascarella & Chapman, 1983; Terenzini, Lorang & Pascarella, 1981) the institutional and goal commitment variable sets made the largest contribution to group discrimination, and they were found to have direct effects on persistence. Interestingly, in her sample drawn from the National Longitudinal Study of the High School Class of 1972, Munro 44 (1981) reported while goal commitment had the strongest effect on persistence, the two variables which contributed to that effect were educational aspirations of the students and the educational aspirations of parents. Central to Tinto's model is the concept of academic and social integration, and just as Tinto (1975) predicted, in all of the studies of the overall model that were conducted at residential institutions (Baumgart & Johnston, 1977; Braddock, 1981; Munro, 1981; Pascarella & Chapman, 1983; Pascarella & Terenzini, 1979, 1980; Terenzini, Lorang & Pascarella, 1981; Terenzini & Pascarella, 1978) the variable sets measuring the concepts of academic and social integration both accounted for a significant amount of the variance found between those students who persisted and those that voluntarily withdrew. Terenzini. & Pascarella (1978) reported. that academic. and social variables continued to contribute to the amount of variance found even when pre-college characteristics and pre-college expectations were accounted for. This added credence to the hypothesis that academic and social integration influence persistence/withdrawal. Attempts to discern which of the two (academic integration or social integration) were more important have yielded ambiguous results. In one study (Terenzini & Pascarella, 1977) social integration seemed to be more important than academic factors in explaining the variance between those who persisted and those who withdrew; in other studies, the order was reversed (Munro, 1981; Terenzini & Pascarella, 1978). As was mentioned previously, there was some indication that the college experience may have different effects on different students. For example, when they analyzed the data separately by gender, Pascarella & 45 Terenzini (1979) noted that the results indicated that academic integration. was more important for men and that social integration (particularly relationship with peers) was more important for women. When Braddock (1981) used Tinto's model to study the dropout behavior of black students attending traditionally' white and traditionally' black institutions in Florida, he found that the academic and social integration variables accounted for a significant proportion of the variance in black student drop out in traditionally white, but not in traditionally black college environments. In numerous studies conducted on residential campuses, one individual variable repeatedly made a statistically significant contribution ix: the explanation of variance between those students who persisted and those who withdrew-~informal contact with faculty (Pascarella & Terenzini, 1977, 1979, 1980; Terenzini & Pascarella, 1977, 1978). But as several researchers (Pascarella & Terenzini, 1977, 1979, 1980; Terenzini & Pascarella, 1978, 1980) discovered, not all types of informal contact 'were of equal importance in. fostering; academic. and social integration, nor was the effect of this contact equally strong for men and women. The single most positive and important type of contact was that involving discussions of intellectual and course-related matters. Discussion centering on career concerns was also significantly and positively related to retention (Terenzini. & Pascarella, 1980). Furthermore, Pascarella & Terenzini (1979) reported that, in general, faculty contact had a stronger positive influence on persistence for men than women. These findings seemed only to underscore the idea that college experiences have different effects on various individuals. Urban Commuter Institutions. When Tinto's model ‘was used as a 46 framework to explore the process of persistence/withdrawal on ‘urban commuter campuses, researchers observed different patterns of influence. Tinto (1975) predicted that the effect students' pre-college characteristics might have on persistence/withdrawal decisions would be mediated through their experiences in the institutional environment. However, when data was analyzed from urban commuter institutions, researchers (Pascarella, Duby, Miller & Rancher, 1981) found that not only (thi background characteristics significantly differentiate between those students who persisted and those who voluntarily withdrew, but also that high school achievement, academic aptitude, and gender were found to have significant and direct effects on persistence (Pascarella & Chapman, 1983; Pascarella, Duby & Iverson, 1983). Thus, these results suggested that the cluster of characteristics which urban commuter students bring to college are :1 factor of equal, if not greater, importance in their subsequent persistence/withdrawan decisions than are their actual college experiences. When the Tinto model was tested in urban commuter environments, goal and institutional commitment variables also seemed to play a different role than was expected. When they conducted a path analysis of the data collected from freshman students at the University of Illinois at Chicago, Pascarella, Duby & Iverson (1983) found that entering commitment to the institution had a direct influence on academic integration rather than on social integration as the model predicted. Furthermore, academic integration, rather than social integration, had a direct effect on subsequent institutional commitment. This finding *was confirmed by Pascarella & Chapman (1983) when they analyzed data collected from a multi-institutional sample involving both two and four-year urban 47 commuter schools. Therefore, it seemed that to students who attended urban commuter schools, institutional commitment was defined to a significant degree by successful and personally satisfying interactions with the academic rather than with the social systems of the institution. When the academic and social integration variable sets were tested on urban commuter campuses, the results were both consistent and inconsistent with expectations based (H: the modelt Pascarella, Duby & Iverson (1983) found that academic integration, consistent 'with the model, had direct and positive effects on persistence. Similarly, Pascarella & Chapman (1983) reported that academic integration indirectly influenced persistence through its direct effects on institutional commitment. On the other hand, Pascarella & Chapman (1983) found social integration had neither direct nor indirect influence on persistence, and Pascarella, Duby & Iverson (1983) reported that social integration had a negative influence on persistence-~two findings seemingly in direct conflict with the model. These findings seemed to indicate that in urban commuter environments, opportunities for social involvement were either so few or were perceived by students as so unimportant, that the concept of social integration had little ‘meaning in terms of ‘bonds to the institution. Similar to what was found on residential campuses, it was reported that on at least one urban commuter campus (the University of Illinois at Chicago), informal interaction with faculty was significantly and positively associated with personal development (Pascarella, Duby, Terenzini & Iverson, 1983). Therefore, although Tinto's (1975) model seems to be a ‘useful 48 framework for understanding the process of student persistence/withdrawal at both residential and urban commuter institutions, the patterns of influence may differ, depending on the type of environment being studied. This fact led Pascarella, Duby & Iverson (1983) to reconceptualize Tinto's (1975) model in order to afford it more explanatory power in an urban commuter institutional setting. Figure 2 illustrates their suggested reconceptualization. As Figure 2 illustrates, in the reconceptualization, Tinto's (1975) five major elements or variable sets were retained but the patterns of influence were different. For example, the revised model assumed that in an urban commuter environment, the characteristics a student brings to college would not influence his/her interactions with the college's academic and social systems, but also that these characteristics would have important direct, unmediated effects on persistence. Social and academic integration ‘were retained as major elements in the model. However, academic integration was hypothesized as having both a direct positive influence on persistence and an indirect effect through its influence on goal commitment and institutional commitment. Somewhat differently from what Tinto (1975) originally hypothesized, initial institutional commitment in the reconceptualized. model had. a direct effect on academic integration which, in turn, was shown to have a direct positive influence on persistence. Based on the research completed at urban commuter institutions, Pascarella, Duby & Iverson (1983) suggested that the effect of social integration would either be non-significant or negative. Summary. Because Tinto's (1975) model provides a potentially useful framework for understanding the complex process of student 49 .88. £3.52 a 3:0 .u__o..aona$ .30: 985... .0 cozuuizaooeooo. 2.... .N 2:9“. scan :5 co nouns 2033 can?! to 302 a 6 20.5532. 1 .268 A a / «320 8 a .85.... Emacs: hzmzesaoo .3508 20.522. A an. ..«zopaEmz. 220.4352. 882 - co..a__.=< “ \ / \ 4.5.1.: 83...? - 00 0.80083 Sushi. 35565.. I Euuwwfiaoo .28 8am , a a 5m .3. mEuatzsou «Emaeaaoo «05.54355 azaomoxo.5ou:uo>.o!h . p p «55059.0( .3 0230018 $513099..— 3 _ ace-5.5500 3:03:52... 83.8 .o 320 .o can: "Eu 3 «as» ease-$32.13“. s inc "an s .38» .832 a .2 .m .32 u a as h 5 0529.26 253 .v Cue in: 3.2.8 .a 35955 430552. n— . «co—5.8500 .ooa mp cos-=08... 250002 035m .2935“. .N 5.3033 ...—.2... .p 0220¢0¥0455— .6 .825 ..81 .8 Sea; 33:152.: 2.8.5.... 5.3 8:333... .53.. 32.8 .o .82.. .838... as: 98:: .o 8:9... 5... .4 2.6.... 2 3:52 323 :2: a. 5:952... I I . 1 ....8m / ecu... in : .333an 335:... Bo>...:ou:ao>.oi._. .3 «cos—2522 .3 0230013 mango-mm... 55 m p = 2.0.5.5500 3:25:35 .\ 858.2... / or = 2.25.5800 .000 / 3 _ 22.22.50 38:38:. + y 1 x83 .8 09.33 950: > or 5:29.... 2.5.33 up . EoE=EEo0 .30 \ 82.8 .o 3.6 .n can: "Eu: 3.5 asst-dieseau .4. A: u o "m u s . .25 3...: .u .3 .m E: u a as u : 23221.8 253 .e an: “an a 18:8 .9 $595.: #59202. 235m .3235“. d cosaoaou .3552. .— OZDO¢0¥0455— 56 paths, in the form of unidirectional arrows, are drawn from the variables taken as causes (independent) to variables taken as effects (dependent). The sign indicates whether the relationship between the variables was hypothesized to be positive or negative. These figures are a summary of the pr0positions initially used to construct the model. Each proposition takes the form: successively higher (lower) amounts of X (the independent variable) will likely produce successively higher (lower) amounts of Y (the dependent variable). The models, as depicted in Figures 3 and 4, are recursive in nature, i.e., the causal flow is presumed to be unidirectional; reciprocal causation between the variables is ruled out (Blalock, 1969; Pedhazur, 1982). In Tinto's (1975) model, the causal flow is from left to right or from the pre-enrollment characteristics to persistence. For example, in Figure 3 the model indicates that the pre-enrollment characteristics of students must ‘be taken into account in order to understand their initial commitments to the goal of graduation and to the institution itself. These commitments, in turn, influence the degree to which a student is integrated into the academic and social dimensions of the institutional environment. iMore specifically, goal commitment is expected to directly affect academic integration, and institutional commitment is exected to directly affect social. integration. Given expectations from Tinto's (1975) model, higher levels of integration should lead to increased commitment both to the goal of graduation and the institution. Both subjective (e.g., degree of satisfaction) and objective (e.g., grade point average) measures of academic and social integration are expected to directly influence subsequent commitments to the goal of graduation and the institution, respectively. Finally, it is 57 these commitments which are hypothesized as being the direct determinants of persistence. The model, as depicted in Figure 4, is similar to the one illustrated in Figure 3. Once again, pre—enrollment. characteristics are expected to influence a student's initial commitment both to the goal of graduation and to the institution. However, the pre-enroll- ment characteristics are also expected to directly affect howr many hours per week a student works-—an intervening variable which did not appear in Figure 3. A. student's initial commitments are also perceived to be direct determinants of how many hours a student works per week. Moreover, just as they did in Figure 3, goal commitment is expected to directly affect academic integration and institutional commitment is expected to affect social integration. But in Figure 4, it is hypothesized that a student's level of integration into the academic and social systems of an institution is affected not only by their initial commitments, but also by the number of hours they work per week. The remainder of the model is the same as it is in Figure 3. In causal models, a distinction is made between. exogenous and endogenous variables. "An exogenous variable is a variable whose causes are not explicitly represented in the model. Consequently, the determination of an exogenous variable is not under considera- tion in the model. An endogenous variable is a variable whose variation is explained by exogenous and other endogenous variables in the causal model" (Pedhazur, 1982, p. 581). In Tinto's (1975) model, the variables measuring a student's pre-enrollment characteristics (family background, individual 58 attributes, pre-college schooling) are exogenous. The ‘variable sets of goal commitment I, institutional commitment I, hours worked. per week, academic integration, social integration, goal commitment II, institutional commitment II, and persistence are endogenous. Each endogenous variable in a causal model is represented by a structural equation consisting of the variables upon which it is assumed to be dependent and a term representing the residuals (Nie, et a1., 1975). The structural equations for Tinto's (1975) model are as follows: Structural Equations for Model 1 (Figure 3) x13 = a + B13,1x1 + B13,2x2 +° ' ' B13,12X12 + “13 X14 = a + B14,1x1 + B14,2x2 +° ' ° B14,12x12 + ”14 x15 = a + B15,13x13 + ”15 x16 = a + B16,14x14 + u16 x17 3 a + B17,15x15 + u17 x18 = a + B18,16x16 + “18 x19 = a + B19,17x17 + B19,18x18 + “19 Structural Equations for Model 2 (Figure 4) : . . + x13 “ + B13,1X1 + B13,2X2 +' B13,12x12 p13 59 x14 = a + B14,1X1 + B14,2x2 +‘ ° ° B14,12x12 + “14 = + + . . . + + X15 “ B15,1X1 B15,2x2 + B15,12x12 B15,13x13 B x 15,14 14 + IJ15 X +B X 16 + B16,13 13 16,15 15 + “16 X17 = a + B17,14x14 + B17,15x15 + p17 x18 = a + B18,16X16 + IJ18 X19 = a + B19,17x17 + “19 x20 = a + B20,18x18 + B20,19x19 + “20 Causal models may be just identified, overidentified, or underidentified. As Pedhazur (1982) points out, when a model consists of more equations than are necessary for the purpose of parameter estimation, the model is said to be overidentified. Such is the case with Tinto's (1975) model. Although overidentified models can still be tested using ordinary least squares techniques to estimate the path coefficents among exogenous and endogenous variables, Pedhazur (1982) warns investigators of not equating the failure to reject the null hypothesis with its acceptance. One of the most commonly used overidentifying restrictions is the postulation that certain path coefficients are equal to zero (Nie, et 60 a1., 1975). In Tinto's (1975) model such hypotheses are made. For example, in Model 1 the direct path between pre-enrollment characteristics and persistence is hypothesized to be zero as is the direct path from goal commitment I to social integration. In Model 2, the direct path from hours worked per week to persistence is hypothesized to be zero, i.e., the only way hours worked per week is expected to affect persistence is through academic and social integration and through goal and institutional commitment. Data Collection This study was longitudinal with data collected through two surveys and from university records. The first survey was administered during the summer of 1983 at an on—campus new student orientation program. The follow-up survey was mailed in May 1984 after three academic quarters. The first survey (New Student Orientation Survey) collected data on the pre-enrollment characteristics, student commitment to the institution, student intention to complete college, and student intention to participate in a variety of activities that Tinto (1975) suggested as dimensions of social and academic integration. (A copy of this survey can be found in Appendix A.) One thousand one hundred and four students who completed the first survey self-reported that they were transfer students (as contrasted.‘with freshmen). ‘However, university 'records later indicated that 102 of these students were not officially classified as transfer students (one who has completed 18 or more quarter hours at another institution) and 45 CH? these students never actually matriculated, thereby reducing the initial number of respondents to 956 (51% of the total population of first-time transfer students who enrolled at UIC in the fall of 1983). 61 A follow—up survey (UIC Emperience Survey) which sought extensive information on the reality of their first year at the subject institution was mailed three quarters later to all 956 of the first—time transfer students previously identified. Accompanying the survey was a letter from the Vice Chancellor for Student Affairs and a stamped return envelope. The letter from the Vice Chancellor explained the purpose of the study, assured the students that their participation in the study was completely optional, and indicated that the students' responses would be kept confidential. The follow-up instrument was largely adapted from a survey used by Pascarella & Terenzini in 1980. (A copy of the follow-up survey and the accompanying letter can be found in Appendix B.) In order to increase the return rate, two weeks after the initial mailing, a reminder post card was sent to the entire sample asking them that if they had not done so already to complete the survey and return it. In June, a second follow-up survey was mailed to those who had not, as of that date, responded. (Eleven of the original respondents could not be reached due to incorrect addresses.) As a result of these efforts, 623 useable surveys were returned. This constituted a return rate of 66% which was determined to be acceptable (Lin, 1976). Other data used in this study was obtained from university records. The information collected from these records included: citizenship, gender, class when admitted, ethnic background, two-year/four—year transfer institution, transfer hours, transfer grade point average, amount and type of financial aid, first year cumulative grade point average, and new status (continuing or unenrolled) fall 1984. 62 Sample Due to financial constraints, strict random sampling techniques were not used in this study. Instead, transfer students yd“) attended. a voluntary new student orientation program during the summer of 1983 prior to matriculation were asked to complete the survey. As was mentioned before, 51% of the total population of first-time transfer’ students attended the orientation ‘prograna and, therefore, completed the first survey. The characteristics of the transfer students who attended the orientation program. and the total population of first-time transfer students entering the University of Illinois at Chicago in the fall of 1983 are show in Table 2. T-tests (used to determine the similarity of the sample and the population) indicated that the transfer students who attended the orientation program were representative of the first-time transfer student population with respect to gender, ethnic background (white, nonwhite), first year cumulative grade point average, and transfer institution (two-year versus four—year). However, the group attending the orientation program had a slightly higher mean transfer grade point average and a larger percentage were still enrolled one year later. The 623 transfer students who responded to the second survey were more representative of the transfer students who attended the orientation program than they were of the entire population of first-time transfers students. This was not an unexpected phenomenon because this was the group from which the sample was drawn. T-tests indicated that the respondents were representative of the students who attended the orientation program with respect to gender, ethnic background (white, 63 ANav mm Anwv ow ANQHV «on possum ANwmv «mm ANomv Hum ANNMV Nam no“::h ANwm. .mN ANamv mmm Anamv mm. «Hoaonaom ANm.. am ANQ.. mm. ANw.V .Nm amanmuum mmmao o.- m.- N.mm Acmmzv mw< ANqNV NmH ANqNV «mm ANNNV moo ouHLB coz ANomv Hue ANomv Nmn Ammmv oomH ouHLB vasouwxomm oacsum Aquv mum Aquv Noe AquV cow mamamm ANomv wqm ANmmv mam ANomv aqofi mam: Hovcmo Mme omm mmwfi muamwaum mo Honasz maaamm coaumuaofiuo wafivcouu< coaumaaaom muaumauouomumzo muaowsum nowmcwue HmuOH avSum mHLH :H mucmwzum uwwmamue mo mHmEmm mnu wnm .aoaumqufiuo.wanCmuu< mufiovsum Hommcmue may .mucovSum Homwcmue oaaelumufim mo cofiumasmom Houoe onu mo mofiumfiumuumumfiu oHEovmo< a ownmmuwoama mnu mo aomwummaou < N 5.9.9 64 ANQNV as. AN.m. mam qu. ea. Ausoaouov umfiaosamc: ANQ.V was ANme. soc ANam. amo. chssaauaoo. emfiaoucm umumq your mac msumum q..m mo.m ...m «mu m>fiumasaa=o um.» umHHm Aqu. o.~ ANmm. m.m flNos. we. .mowuusom ANocv m.q ANao. wmo ANOOV so.. swowsose coauaufiuwcH ummmamuH o.mm m.mm w.qw musom smmmsmue .m.m ow.m .m.m Acmozv swam N can H mauve: - mumm manmwum> uamuommwn 0% can ommmuoaH mm m manmh 78 for both Models 1 and 2. As the table indicates, Model 1 explained 21.52 (adjusted R2 = .185) of the variance in the dependent variable persistence; Model 2 explained 22.1% (adjusted R2 = .189) of the variance in persistence. Although modest, these percentages are consistent with the R2 reported by other investigators who tested Tinto's (1975) model in both residential (Pascarella & Terenzini, 1980) and urban commuter institutions (Pascarella & Chapman, 1983; Pascarella, Duby & Iverson, 1983). In each model, three of the variable sets were associated with statistically significant increases in R2: background characteristics (4.982); academic and social integration (11.52 - Model 1, 12.02 - Model 2); and subsequent goal and institutional commitments (4.8% - Model 1, 4.7% - Model 2). The finding that the integration variables were associated with the largest increase in R2 in both models tended to support Tinto's (1975) proposition that persistence is predicated, to a great extent, on a student's integration into the academic and social systems of institu- tion. However, this finding is inconsistent with what Pascarella, Duby & Iverson (1983) found when they tested Tinto's (1975) model among freshman students at a commuter institution. They found that the pre-enrollment characteristics were associated with the largest increase in R2 which led them to conclude that the "cluster of characteristics which commuter stu— dents bring to college might be a factor of equal, if not greater, importance, in their subsequent persistence/withdrawal. decisions than 5The adjusted R2 statistic is a more conservative estimate of the percentage of variance explained because it has been adjusted to the number of independent variables in the equation and the number of cases (Nie, et a1., 1975). 79 their actual experiences of college once enrolled" (Pascarella, Duby & Iverson, 1983, p. 93). Direct Effects The use of path analysis in this study required the regression of each endogenous variable on each of the exogenous variables and all other causally antecendent endogenous variables in the model. Tables 4 and 5 display the standardized and unstandardized regression coefficients, as well as their standard errors, for each structural equation in Models 1 and 2, respectively. Although path analysis can be performed with either the standardized coefficients or’ the unstandardized coefficients, ‘the standardized coefficients were used in this study because the independent variables were measured in different units and the main interest was in assessing the overall effect of one variable over another variable in the same sample (Nie, et a1., 1975). Those standardized regression coefficients (beta weights) which were significant at the .05 level are indicated by an asterick following the coefficient. For example, both Tables 4 and 5 indicate that when goal commitment I (variable #13) was regressed on the twelve exogenous variables preceding it in the model, only one standardized regression coefficient was significant at the .05 level--parents' education. Another way to represent the data presented in Tables 4 and 5 is to draw path diagrams using only the standardized regression coefficients which were significant at the .05 level. Figures 5 and 6 portray the significant standardized regression coefficients for Models 1 and 2, respectively. The arrows from outside the model pointing to each endogenous variable indicate the residual path coefficients (Jq:E§) for that variable. 13() Table 4 Standardized and Unstandardized Regression Coefficients For All Structural Equations (Modelgl) Variables i) ‘9 ‘5 1‘ i7 10 19 ‘a 'lrtfl20° an”. .015 .100“ e019 e035 “.030 .6010 Education .1531.066) .036(.092) .172(.063) .056(.076) .06~(.o69) -.0311.039) -.001(.012) 2. Financial :026 .065 -.019 -.026 .016 .020 .036 Status .026(.0bb) .066(.062) -.01h(.062) -.032(.os1) .019(.o&6) .015(.026) .oo1(.0o1) 3. Sender (1-Hale. -.026 -.061 -.050 .001 -.000 .001 .077' O-Feneie) -.111(.163) -.37h(.25~) -.212(.175) .035(.211) -.oo1(.191) .020(.106) .0661.033) 4. Ithnic Background -.079 -.056 .139-9 .016 .000 -.066 .022 (I-Hhite. O-Ionuhite) -.389(.210) -.36&(.291) .6S6(.201) .10$(.212) .025(.221) -.236(.12h) .022(.039) 5. Age -.000 .1509. .126-- .000 -.066 .011 ..095 -.000(.022) .097(.030) .056(.021) .0001-025) -.03b(.023) .000(.013) -.001(.000) 6. Heritai Status -.017 .012 -.020 -.000 -.103' .1069. .016 (1 Single, O-Herried) -.10~(.279) .1o6(.967) -.123(.266) -.000(.320) -.729(.290) .b61(.16h) .020(.051) 7. r011 Tine Status -.0oo -.066 .1460. .096. -.000 .015 -.017 (1-r1, 0-r1) -.001(.203) -.~7~(.261) .706(.191) .Sho(.233) ..000(.21k) .051(.120) -.017(.037) 3e. freshen no“ -.107 .047 -.113 -.09‘1 -.o63 -.039 -.271(.361) -.905(.529) .277(.36~) ~.76S(.636) -.633(.399) -.3~3(.225) o.046(.070) 80. Sophomore -.130 -.116 -.000 -.017 -.068 -.110 -.001 ~.567(.337) -.730(.h67) -.000(.322) -.067(.366) -.343(.351) -.36~(.196) -.071(.062) 0c. Junior -.0‘3 -.053 -.093 -.055 -.0‘5 -.069 .Ofi9 -.191(.330) -.332(.h$6) -.ho7(.316) -.276(.379) -.226(.3bh) -.212(.19£) .016(.06o) 9.. Business 4.026 .076 -.103 -.116 .001 .062 -.017 Administration -.133(.631) .593(.67s) -.556(.601) -.730(.72§) .0hk(.657) .31h(.370) -.019(.11A) 96. Engineering -.036 -.026 o.162 -.101 -.060 .091 .001 o.226(.666) -.2ss(.921) -1.090(.635) -.796(.76s) -.622(.695) .h)5(.392) .001(.120) 9c. Architecture -.016 .041 -.091 -.079 .016 .056 .056 5 Art -.119(.663) .434(.919) -.660(.632) -.66h(.76o) .131(.690) .265(.369) .067(.120) 96. Liberal Arts .013 .016 c.116 -.116 .069 .132 -.122 .056(.610) .010(.6h6) - 5091.561) -.569(.700) .3~2(.635) .402(.356) -.105(.110) 10. Achieve-ant .022 -.035 .1119. -.019 .0611 -.072¢ -.001 .0291.057) c.067(.079) .1h7(.054) -.076(.06$) .126(.06o) -.066(.033) -.000(.01o) 11. Transfer institution -.037 c.1940. .042 .020 -.000 -.069 -.001 11-rour Yr. 0-Tuo Yr.) -.166(.169) o1.256(.262) .166(.163) .106(.221) -.050(.200) -.220(.112) -.001(.035) 12. High School .046 -.001 .016 .037 .00 .055 -.016 Activities .036(.032) -.001(.085) .01~(.031) .031(.037) .000(.031) .025(.019) -.001(.001) 13. Goal Col-it-ent 1 -.026 .001 .952»11 .065 .010 -.026(.0~0) .001(.019) .606(.0bh) .032(.0zs) .000(.001) 16. institutions! .1031 .1669. .059 .2620- -.06s Commitment 1 .0711.029) .1911.035) .0671.032) .126(.016) ..oo1(.0o1) 15. Academic Integration .156" .399" .260" .161(.066) .276(.027) .052(.001) 16. Social Integration .032 .126“ -.046 .0321.039) .o76(.022) -.001(.0o1) 17. coal Cal-lteent 11 -.016 ’aw°(e”') 10. "1621211250001 .z'nee toe-item: 11 .077(.013) 19. Persistence 0’ .026 .096 .113 .051 .215 .991 .215 Iota. '0 (.05 9.01:.01 Top amber is the standardised regression coetfactnt, batten amber is the metric sufficient, ache: in parentheses is the standard error. Table 5 Standardized and Unstandardized Regression Equations (Model 2) 81 Coefficients For All Structural Variables 13 16 15 16 17 16 19 20 1. Parents' .09“ .015 .067 .116“ .037 .030 -.032 ..925 tdutatIon .153(.066) .036(,092) .631(.361) .165(.063) .066(.076) .070(.070) -.036(.039) -.0o1(.012) 2. Financial .026 .065 -.63399 -. 7 -.075 -.000 .031 .077 s:.:., .0261.066) .0661.062) -2.701(.229) -.071(.667I -.093(.0$6) -.ooo(.051) .023(.029) .016(.0o1) 3. Gender (I-Ha1a, -.026 -.061 .057 -.063 .013 .000 .000 .072 o-r.n.1.) -.111(.163) -.376(.255) 1.66(.565) -.1021.17s) .067(.2101 .001(.192) .016(.106) .0621.033) 6. Ethnic sactgroand -.079 -.056 .013 .1 99 .020 .000 -.066 .019 (I-Uhits, O-Nonuhlte) -.369(.210) -.366(.291) .361(1.063) .666(.200) .116(.261) .031( 221) -.236(.125) .019(.0361 5. Age -.000 .15099 -.050 .12299 -.000 -.067 .012 -.051 -.000(.022) .097(.030) -.133(.113) .05$(.021) -.000(.ozs) -.035(.023i .000(.013) -.001(.000) 6. nar1t61 status -.017 .012 .060 -.015 .000 -.1o19 .10599 .016 (I-Singla, o-narra.d) -.1o6(.279) .106(.367) 1.660(1.636) -.093(.26S) .0321.319) -.716(.291) .6561 166) .017(.os1) 7. r011.11.. sr.r.. o. o -.066 -.16199 .12699 .076 -.001 .013 -.000 (1-r1, 0-r1) o.0o1(.203) -.676(.261) -6.590(1.066) .609(.196) .636(.236) -.033(.217) .066(.122) -.000(.037) 6.. rrstnaan -. 6 -.107 .019 .050 -.110 -.093 -.063 -.061 -.17l(.3011 9.9061.519) .670(1.956) .291(.362) 9.7501.637) -.626(.399) -.366(.2251 -.066(.069) 80. Sophomore 9.130 -.116 .0“ .000 -.012 -.066 -.119 9.085 ._ 71.337) -.130(.667) 1.166(1.737) .021(.320) -.0601.366) -.336(.351) -.366(.1961 -.076I.061) 80. Junior 0. 3 -.053 .013 -.092 -.053 -.M5 -.069 .050 -.19|(.330) -.332(.657) .327(l.697) -.600(.313) -.268(.377) -.225(.363) -.212(.196) .063(.O60) 9e. Business 9.025 .076 .010 -. 02 -.115 .001 .002 9.016 Administration -.133(.631) .593(.I7$) .320(3.264) -.SSI(.5961 -.723(.7211 .062(.‘S7) .3161.370) -.017(.116) 90. Engineering -.036 -.026 -.131 -.179 -.116 -.066 .095 .021 -.226(.667) -.256(.925) -S.233(3.626)-1.205(.633) -.913(.763) -.669I.695) .656(.393) .026(.121) 91:. Architecture -.016 .041 -.054 -.090 -.005 .013 .058 .052 6 Art -.119(.663) .AJ~(.919) -2.290(3.600) -.706(.620) -.715(.757) .112(.690) .296(.389) .076(.120) 9d. Liberal Arts .016 .016 -.016 -.121 -.116 .067 .133 -.119 .056(.610) .010(.666i -.659().137) -.S16I.576i -.5601.697i .335I.635) .605(.356) -.103(.1101 1o. Achieve-Int .022 -.o35 -.023 .10699 -.052 .0699 -.0729 -.001 .029(.os7) -.o67(.079) -.162(.291) .163(.053) -.061(.065) .1271.060) -.066(.036) -.ooo(.010) 11. Transfer Institution -.037 -.196“‘ -.069 .033 .012 -.012 9.068 .001 (1-rv, 0-TV) 6.166(.169) -1.256(.262) —1.636(.990) .166(.163) .06s(.221) -.063(.201) -.216(.113) .001(.035) 12. High School .066 -.000 -.001 .017 .036 .000 .065 -.063 Activities .036(.032) -.000(.06s) -.039(.166) .013(.031) .033(.037I .000(.036I .025(.0191 -.001(.001) 13. Goal ton-itaont 1 -.020 -.026 .000 .35199 .065 .011 -.117(.217) -.026(.060) .000( 066) .605(.066) .092(.025) .0oo(.001) 16. Institutional -.029 .0999 .16399 .059 .26399 -.066 _ commit-ant 1 -.116(.157) .0691.029) .1311.035) .067(.032) .126(.016) -.oo1(.001) 15. Moms Uorked Per Heat -.12“. «113' -.0‘0 .026 .091. -.021(.001) -.022(.0o1) -.001(.001) .000(.001) .0001.000) 16. Aces-lit Integration .153“ .3 n .260“ .1761 066) .260( 027) .0s6(.001) 17. social Integration .030 .12799 -.038 .030(.060) .077I.023) -.001(.001) 16. coal Colnlt-ent 11 -.001 -.000(.001I 19. Institutional .267“ Omit-ant II .076(.013) 20. Persistence A’ .026 ..036 .260 .126 .061 .217 .332 .221 lot—e. Top amber is the undefined regression coefficient. bottu was is the let!“ coefficient. enter is parentheses is the standard error. ‘01t.05 “s t .01 82 m..~.fl«¢ so. ... 362.5283: 52. 53.6.6568 8.3:? 66.03286; acamu 33. 0mm. . S. v .... 8. v ... up 3:53 .028 ...... a. 5.1.. .r... c 35...... 6.2.2» i>§mt¢o>1¢3h . p —. ...—2:03.132 .3 2.5028 mom-3001mm: hp 31.8... 68.0... 9...... E... a 2.88 25.5.3553“... .31.. 61: 2.5a 3.3.... 8e... .32.... in: 2:29.80 2.23... anginagd $50.55 .2355... .. .55558 A .. .30 . 0.6K . 0:; .8935". d con-0:8 ...-.2...— .— Ozacaxg >43..— N3. 08. ..u .88.. .5... Sean... 5:. ... 3365.8 5.82.... 86.28:... ...-0.....5 ... 65...... S.V a: an"... to. . 8... 8... 8... 3. v ... 00:33 .0038 :2: .m— ..C n o a... u s 8.55.... ...-5..» i>£30utco>gh . 2. «62:25.51 .3 83 02:09.8 33.3.2... 2.. .. . ... 3 .m. .. 225.558 . .5558 .8328... 5.255.... . 82.8 d . .... ...... c... u .. ...... n a 3.8 5.5.3.5.... ... a 11 .... .... ...... .. 8.5%.... 8 .38 5...... .6 8.. ... m. «1.08 .32.... .31 c s ..e 9.3.9.68 0.53 .e .. $.11... .16.... tunincgu w. . ...» a $50.55 €35.02. 2,... ..v. ml op .6 aw:— / .. Ease—5.8 1 . . :55. e8 ,, n‘. _80 esmfl .019 g 1‘“ . \xv . 9. _ z > 9. . .w. 26 ee”\. ~.\ 2 o. .38 38...... a 8:052... 4 :9. 1|. 3.... coaaoaou 3:03.. .p 2288.. 335.83 5.3... mg. 8m. 33. 84 The path diagram summarized in Figure 5 contains only those variables or variable sets originally included in Tinto's (1975) theoretical model of college student persistence. Only three variables (academic integration, institutional commitment II and gender) had significant direct effects on persistence, controlling for all other variables in the model. Of these three variables, both academic integration (beta = .26) and institutional commitment (beta = .27) were equally important in predicting persistence among first-time transfer students at the subject institution. Sex was somewhat less important (beta = .08), but still statistically significant.6 All of these relationships ‘were in. a 'positive direction. which. suggested. that the higher the level of academic integration and commitment to the institution, the more likely the student was to be enrolled at the subject institution one year later. Since gender was a dummy coded variable (l-male, O—not male), the data suggested that after one year at the subject institution, male students were ‘more likely’ than. female students to re-enroll. Tinto (1975) hypothesized that only goal commitment II and institutional commitment II ‘would.'have significant direct effects on 6In order to determine if the data should be partitioned on the basis of gender, an additional structural equation. was created ‘which regressed persistence on all of the variables preceding it in the model and a set of cross-product terms representing the interaction of gender and all other variables. Since none of the interaction terms were found significant at the .05 level, it was judged that separate equations based on gender were not needed. More information about interaction effects can be found later in this chapter. 85 persistence, and that the effects of the background characteristics and the integration variables would be indirect, transmitted through the commitment variables. The data did not totally support this hypothesis. Although institutional commitment II had a significant effect on persistence, goal commitment II did not have a significant direct effect on persistence. Moreover, instead (M5 indirectly affecting persistence through the endogenous variables preceding persistence in the model, both academic integration and gender had significant direct effects on persistence. The sixteen variables preceding goal commitment II in the model accounted for 21.52 (adjusted R2 = .187) of the variance in goal commitment II. The four best predictors of goal commitment II were: goal commitment I (beta = .35), academic integration (beta = .16), marital status (beta = -.10) and achievement (beta. = .08). .All of these relationships were in a positive direction with the exception of martial status. This finding suggested that successively higher (lower) amounts of academic integration, goal commitment I and achievement produced successively higher (lower) amounts of goal commitment II. Since marital status was a dummy coded variable (l-single, O-not single), the data suggested that after one year at the subject institution, single students were not as committed to the goal of graduation from college as were married students. As was 'mentioned earlier, institutional commitment II (measured after one year at the subject institution) was an important intervening variable in this study. Thirty—three percent (adjusted R2 = .308) of the variance in institutional commitment II was accounted for by the sixteen variables preceding it in the model. Academic integration (beta 8 .39) 86 was by far the most important variable in predicting institutional commitment II, followed by institutional commitment I (beta -= .26), social integration (beta = .13), marital status (beta = .11) and achievement (beta 8 -.07). Although the positive direct path from social integration to institutional commitment II seemed to confirm what Tinto (1975) suggested in his theoretical model, it is important to note that the direct effect between academic integration and institutional commit- ment II was nearly three times as strong. This finding suggested that one year after matriculation, commitment to the institution was defined largely ‘by successful, and ‘personally satisfying interaction.*with the academic dimension of the institution and to a lesser degree by interaction with the social systems. The significant inverse relationship between achievement and institutional commitment II seemed to indicate that after one year at the subject institution, higher ability students were less committed to the institution than were students with lower transfer grade point averages and high school grades. 0n the other hand, the positive sign of the beta weight between marital status and institutional commitment indicated that single students were more committed than married students to the institution one year after matriculation. Based on the data collected for this study, the second best predictor of persistence among first-time transfer students was the variable set measuring academic integration. However, only' a small proportion of the variance (11%; adjusted R2 = .085) in academic integration was accounted for by the fourteen variables preceding it in the model. Six variables had a significant and positive influence on academic integration: full-time status (beta = .15), ethnic background 87 (beta = .13), age (beta == .13), achievement (beta == .11), parents' education (beta = .11), and institutional commitment I (beta = .10). The positive sign of the beta weights for these six variables indicated the following: 1. Transfer students who attended school on a full-time basis were more integrated into the academic system of the university than were students who attended school on a part-time basis; 2. White students were more integrated into the academic system than were nonwhite students; 3. The older the transfer student the better integrated he/she was in the academic system of the university; 4. Successively higher levels of prior academic achievement resulted in successively higher levels of academic integration; 5. The more formal education the student's parents had, the higher the level of academic integration; 6. The more committed the transfer student was to the institution prior to matriculation, the higher the level of academic integration. Contrary to what Tinto (1975) suggested, initial goal commitment did not have a significant direct effect on academic integration. The same fourteen variables which preceded academic integration in the model also preceded social integration. However, these same variables accounted for an even smaller proportion of the variance (52; adjusted R2 = .02) in social integration. As hypothesized, institutional commitment I had the greatest effect on social integration (beta = .17). 88 The more committed the transfer student was to the institution prior to matriculation, the higher the level of subsequent social integration. However, not surprisingly, the pre-enrollment variable measuring full-time status also had a significant positive direct effect on social integration (beta = .10). Students attending school on a full-time basis were integrated more fully into the social system of the institution than were students who attended school on a part-time basis. Finally, the twelve exogenous variables preceding goal commitment I and institutional commitment I accounted for 2.8% (adjusted R2 = .0009) and 9.8% (adjusted R2 = .073) of the variance in those variables, respectively. The only variable which had a significant effect on goal commitment I was parents' education (beta = .10). The more formal education a transfer student's parents had, the more committed the student was to the goal of graduation from college. Two variables significantly influenced institutional commitment 1: age 'was positively related (beta. = .15), while transferring from. a four-year institution was negatively related (beta = -.19) to institutional commitment. In other words, the older the students the more committed they were to the institution prior to matriculation, and students transferring from four-year institutions were less committed to the institution than. were students transferring from two—year institutions. Figure 6 displays the significant standardized regression coefficients found in Model 2 in path diagram form. Some interesting variations were found when the significant path coefficients in this model were compared with those in Model 1 (see Figure 5). Although academic integration and institutional commitment were 89 still the best predictors in persistence (beta weights of .27 each), the third variable which significantly influenced persistence was the number of hours worked per week. Somewhat surprisingly, the positive sign of the beta weight indicated that the more hours a student worked per week, the more likely he/she was to be enrolled one year later at the subject institution. Furthermore, it was interesting to note that the direct effect of gender on persistence found in Model 1 disappeared when the variable measuring how many hours a student worked per week was added to the model. As predicted, the variable measuring the number of hours a student worked per week had a significant negative effect on both academic (beta = -.12) and social (beta = -.11) integration. However, this finding seemed somewhat incongruous with the fact that the variable measuring hours worked per week had a significant positive direct effect on persistence. There was a significant negative direct path ‘between financial status and hours worked per week (beta = -.43). That is, the more financial help a student received either from parents/spouse or financial aid, the fewer hours he/she worked per week. Although a student's financial status did not seem to affect persistence either directly or indirectly in Model 1, there was an indirect influence in Model 2 through the intervening variables hours worked per week, academic integration, social integration, and institutional commitment II. The remainder of the significant paths found in Model 2 were the same as those which were found in Model 1. Interaction Effects As was discussed in Chapter 4, in order to determine whether the 90 effect of the independent variables was additive, several sets of interaction terms were formed and included as new predictor variables in the regression equations. Only one significant interaction was found at the .05 level in both models: the interaction between goal commitment II and institutional commitment II. The unstandardized regression coefficient for the interaction between goal commitment II and institutional commitment II was -.0111 in both Models 1 and 2. When the negative sign of the unstandardized regression weight for the interaction terms was compared with the positive signs of the unstandardized regression weights for goal commitment II (.0582 - Model 1, .0586 - Model 2) and institutional commitment II (.0794 - Model 1, .0784 - Model 2), a compensatory relationship seemed to exist between the two commitment variables. In other words, the nature of this interaction suggested that institutional commitment had its most positive effect on persistence when the level of goal commitment was relatively low, and goal commitment had its most positive effect on persistence when the level of institutional commitment was relatively low. This finding lended support to 'Tinto's (1975) hypothesis that high levels of goal commitment would tend to compensate for low levels of institutional commitment, and vice versa. Similar results were reported by Pascarella & Terenzini (1979) when they examined the inter-relationships between and within the variable sets in Tinto's (1975) model. The lack of significance between the remainder of the interaction terms tested in this study indicated that the relative influence of the variables in the models did not differ for different kinds of students. For example, although men were more likely than women to be enrolled at 91 the subject institution one year later, the relative influence of the variables in explaining their behavior were the same for both groups. Summary In summary, although the findings seemed to support a number of Tinto's (1975) theoretical expectations, the data was not totally consistent with the model. First, as expected, in Model 1 the influence of the pre-enrollment characteristics on. persistence ‘was primarily transmitted through the transfer student's first year experience at the subject institution. The only exception to this was gender. Controlling for all other variables in the model, men were significantly more likely to persist than were women. Interestingly, however, when the variable measuring the number of hours worked per week was added to the equation, the significant direct effect of gender on persistence disappeared. The reason for this was not clear. The role of the exogenous variables in Model 2 were more in keeping with what Tinto (1975) hypothesized: there were no significant direct effects between the pre-enrollment characteristics and persistence. That is, the effects of the pre-enrollment characteristics on persistence in Model 2 were transmitted through the intervening variables such as hours worked per week, academic/social integration, and goal/institutional commitment. Secondly, contrary to what was expected, neither measure of goal commitment (goal commitment I (n: goal commitment II) seemed to play an important role in the model, whereas both measures of institutional commitment (institutional commitment I tun! institutional commitment II) were significant predictors. In both Models 1 and 2, institutional commitment II had a significant direct effect on persistence; goal 92 commitment II did not. (Goal commitment II did have a positive zero-order correlation, however, as shown in Appendix D.) Furthermore, entering goal commitment did not have a significant direct influence on academic integration as predicted. Instead, entering commitment to the institution had 21 significant direct influence on academic integration and social integration, rather than just social integration as expected. The significant interaction found between goal commitment II and institutional commitment II, lended support to Tinto's (1975) hypothesis that high levels of goal commitment would tend to compensate for low levels of institutional commitment, and vice versa. The finding that the integration variables were associated with the largest increase R2 in both models tended to support Tinto's (1975) proposition that persistence is predicated, to a great extent, on a student's integration into the academic and social systems of the institution. However, the roles played by the ‘variables ‘measuring academic and social integration were somewhat different than what was expected. Tinto (1975) hypothesized that the influence of academic and social integration on persistence would be indirect, transmitted through subsequent measures of goal and institutional commitment, respectively. Consistent with expectations based on the model, in this study social integration had a significant positive effect on institutional commitment II and institutional commitment II, in turn, had a significant positive influence on 'persistence. ‘The data collected in. other studies and reported in the literature (Pascarella & Chapman, 1983; Pascarella, Duby & Iverson, 1983) did run: support this hypothesis. Pascarella, Duby & Iverson (1983) reported that among freshman commuter students, social 93 integration has a significant negative effect on persistence. Although there was a negative beta weight between social integration and persistence in this study, it was not significant at the .05 level. Contrary to what was expected, academic integration had both a significant direct effect on persistence and an important indirect effect on persistence which was transmitted through institutional commitment II (instead of goal commitment II as expected). Thus it appeared that in both Models 1 and 2, academic integration played a major role in determining whether or not a transfer student re—enrolled at the subject institution one year after matriculation. Similar findings regarding the importance of academic integration in urban commuter environments are reported in the literature (Pascarella & Chapman, 1983; Pascarella, Duby & Iverson, 1983). Three aspects of :1 student's financial status were considered in this study: the percentage of help received from parents/spouse, the amount of financial aid awarded, and the number of hours worked per week. In Model 1, the exogenous variable measuring a student's financial status did not have a significant direct effect on any of the endogenous variables including the initial commitment variables. In Model 2, as hypothesized, hours worked per week had a significant direct negative effect on both of the integration variables, and, somewhat surprisingly, a significant direct positive effect on persistence. In other words, the more hours student worked per week, the less integrated he/she was in the academic and social dimensions of the institution. However, the more hours a student worked, the more likely he/she was to be enrolled at the subject instution one year later. Finally, although gender was the only exogenous variable which had 94 a significant direct effect on persistence, other pre-enrollment characteristics had important indirect effects on persistence through their direct effect on the endogenous variables preceding persistence in the model. The two endogenous variables of particular importance in this study were academic integration and institutional commitment II. The exogenous variables which were significant determinants of academic integration were: 'parents' education, ethnic background, age, full-time status, and achievement. ‘These findings suggested that the older the students, the better educated their parents' were, and the higher the level of academic achievement prior to matriculation, the more integrated the students were in the academic dimensions of the subject institution. Furthermore, white students and students attending school on a full-time basis were more likely than nonwhite students and students attending school on a part-time basis to do well academically. The exogenous variables which directly affected institutional commitment II were: marital status, and achievement. Single students were more committed than married students to the institution one year after matriculation, and successively higher levels of prior academic achievement were associated with successively lower amounts of institutional commitment one year after matriculation. Two other exogenous variables which indirectly affected institutional commitment II through their direct effect on institutional commitment I were: age and transferring from a four-year institution. Older students were initially more committed than the younger students to the institution, and students transferring from four-year institutions were initially less committed to the institution than students transferring from two-year institutions. The conclusions that can be drawn from these findings are presented in Chapter 6. CHAPTER 6 CONCLUSIONS, IMPLICATIONS AND SUGGESTIONS FOR FURTHER RESEARCH Overview of the Study The purpose of this study was to test the explanatory power of Tinto's (1975) theoretical model of college student attrition in describing the persistence/withdrawal behaviors of first-time transfer commuter students at an urban university. In addition to the variables originally included in the model, the investigator also tested the influence of _finances. on student persistence/withdrawal within the framework of Tinto's (1975) model. A complete test of the model was conducted at the University of Illinois at Chicago (UIC), an urban commuter institution with an undergraduate pepulation of approximately 20,000 and a graduate population of nearly 5,000. The sample was drawn from 1855 transfer students who enrolled at UIC for the first time in the fall of 1983. Data was collected in a two—step longitudinal process. The first survey was administered in the summer of 1983 prior to matriculation; the followeup instrument was mailed in May, 1984, after three academic quarters. Other information was ascertained through university records. In order to adequately test whether or not the data was consistent with the causal model, two statistical techniques were employed: multiple regression and path analysis. The findings of this study suggested that Tinto's (1975) model is a useful framework for understanding the process of persistence/withdrawal 95 96 among transfer commuter students in urban university environments. While the amount of variance explained in the dependent variable, persistence, was modest (22%), it is comparable with. what is reported in other studies. Although the patterns of influence found to exist between_ the variables in this study seemed to support a number of Tinto's (1975) theoretical expectations, the data was not totally consistent with the model. First, the finding that the integration variables accounted for the largest percentage of variance in the dependent variable, _persistence, tended 1x) support Tinto's (1975) proposition that persistence is predicated, to a great extent, on a student's integration into the academic and social dimensions of the institution. Furthermore, academic integration had both a significant direct effect on persistence and an important indirect effect on persistence which was transmitted through institutional commitment II (rather than goal commitment II as expected). This finding suggested that among transfer students at the subject institution, persistence was defined largely by successful and personally satisfying interaction with the academic dimension of the institution and to a lesser degree by interaction with the social system. As expected, the influence of a transfer student's background characteristics on persistence was primarily indirect, with the exception of gender. Controlling for all other variables in the model, men were more likely than women to be enrolled at the subject institution one year later. Moreover, a transfer student's financial status had both a direct and an indirect effect on persistence. Three aspects of a student's financial stuatus were considered in this study: the percentage of help 97 received from parents/spouse, the 999939.9fi financialgaid.awarded, and ...—a H"- .-- .fi _- __j the number 0f,399I§-Y93k99 per weak- The variable measuring how much help a student received from either parents/spouse and/or financial aid indirectly affected persistence through its indirect effect on hours worked per week. In short, the more financial helpawstudent.regeived, ”mm "5", av- I‘Abm‘t the fewer hours he/she worked per week. The number of hours a student ..... ‘W‘i-‘W ...—u- oa...’ ... worked, on the other hand, had both a direct and an indirect effect on persistence. The more hours a_student worked per week, the less well ’.,_-Iw... university. However, despite this negative influence, the more hours a student worked per week, the _ greater ‘ the... likelihooimhelshe. ,wouldd be ”'M” 'Ms "a J.— enrolled at the subject institution one year later. \ d “Hg-“.AMH fim‘findwfi‘dfl’wfl These findings have several. implications for ‘progran1 and 'policy M‘s-mp.” action which might be helpful to university decision-makers. For example, efforts to reduce attrition among transfer commuter students are more likely to succeed if they are focused on what happens to students after their arrival on campus rather than on the characteristics the students possess prior to matriculation. Since academic integration was a ”1"“A1‘9'JJ‘ effigy, mugs-«muggy: firs-0'51"” particularly salient variable in explaining the persistence/withdrawal ...-um behaviors of the transfer students in this study, particular attention '31». .....- Pm .. should be given to developing policies or programs aimed at increasing _ =,-..-,..t .-' student involvement in that dimensionrrofcampusvlifet Presentation and Discussion of Major Findings In keeping with Tinto's (1975) model, six hypotheses were developed and tested in this study. Each hypothesis is discussed briefly below. 1. The integration variables, as a set, will account for the largest amount of variance in dependent variable, 98 persistence. In order to determine the unique influence of different variable sets in Tinto's (1975) model of persistence, a setwise regression procedure was employed in this study. Sets of independent variables were entered in the regression equation in an §_priori manner according to the causal ordering of the variables in the model. In each model, three of the 'variable sets ‘were associated.‘with statistically significant increases in R2 (see Table 3, p. 77): background characteristics (4.92); academic and New“ W'I'o “.1" UN.- ... 1 social integration (11.52 - Model 1, 12.02 - Model 2); and ‘wn M-n—a-a subsequent goal and institutional commitments (4.8% - Model 1, 4.7% - Model 2). The finding that the integration variables were associated with the largest increase in R2 tended to support Tinto's (1975) proposition that persistence is " fir." NM ”......‘y um predicated, to a great extent, on a student's integration WAui‘wa “‘fia ‘JR'nJ,uH,.-.n‘a ‘ ‘9: into the academic and social systems of the institution. __ t . ,-.‘F.'.;'4 as ..-. ‘.,fl’M-I.7n n-..p,~-"..4—_r«-'¢‘u ’ .v The influence of the pre-enrollment characteristics (including a measure of a student's financial status) on persistence will be mediated by the student's commitments and first year experiences at the subiect institution. In order to test this hypothesis, the use of path analysis was employed. Each endogenous variable in both Models 1 and 2 was regressed on all of the exogenous variables and all other causally' antecendent. endogenous 99 variables in the model. It was expected that none of the exogenous variables would have significant direct effects on persistence. In Model 1, one of the exogenous variables had a significant effect (n1 persistence (see Figure 5, {L 82). That variable was gender. Controlling for all other variables in the model, men were significantly more likely than women to be enrolled at the subject institution one year after matriculation. The effects of the rest of the exogenous variables in Model 1 on persistence were indirect, transmitted through the commitment and integration variables as expected. The roles of the exogenous variables in Model 2 were more in keeping with the hypothesis (see Figure 6, p.33). There were no significant direct effects between any of the pre-enrollment characteristics and persistence. As hypothesized, the effects of the ‘pre-enrollment. charac- terisistics on persistence in. Model. 2 ‘were transmitted through the intervening variables that included, hours worked per week, academic/social integration, and goal/institutional commitment. Initial commitments to the goal of graduation and the subject institution will directly affect academic and social integration, respectively. As Figures 5 and 6 illustrate (see pp.82 and 83), initial commitment to the goal of graduation from college (goal commitment I) did not have a significant direct 100 effect on academic integration in either Model 1 or 2. Instead, entering commitment to the institution (institutional commitment I) had a significant direct influence (n1 social integration 22d academic integration, rather than just social integration as hypothesized. The influence of academic and social integration on persistence will be indirect, transmitted through subse- quent measures of goal and institutional commitment, respectively. Consistent with expectations based on the model, in this study social integration had a significant positive effect on institutional commitment II, and institutional commitment II, in turn, had a significant positive influence on persistence. The data collected in other studies and reported in the literature (Pascarella & Chapman, 1983; Pascarella, Duby & Iverson, 1983) did not support this hypothesis. In fact, Pascarella, Duby & Iverson (1983) reported that when they tested Tinto's (1975) model among freshman commuter students, social integration had a significant negative direct effect on persistence. Although there was a negative beta weight between social integration and persistence in this study, it was not significant at the .05 level. Contrary to what was expected, academic integration had both a significant direct effect on persistence and an important indirect effect on persistence which was transmitted through institutional commitment II (instead of 101 goal commitment II as expected). The direct, positive influence of Tinto's (1975) central concept of academic integration on persistence is quite consistent with the results of previous studies conducted at both residential (Munro, 1981; Terenzini & Pascarella, 1978) and commuter institutions (Pascarella & Chapman, 1983; Pascarella, Duby & Iverson, 1983). Goal commitment and institutional commitment (measured one year after matriculation) will be the only two variables that will have significant direct effects on_persistence. As such, they will be the best predictors of presistence. In this study, the two best predictors of persistence were academic integration and institutional commitment measured one year after matriculation. Both of these variables had significant direct effects on persistence as portrayed in Figures 5 and 6 (pp.82and 83). Although goal commitment II did not have a significant direct effect on persistence as expected, commitment to the goal of graduation did have a positive zero-order correlation with persistence as shown in Appendix D. The number of hours a student worked sper week will indirectly influence persistence through several inter- vening variables. More specifically, the more hours a student works per week, the less integrated he/she will be in the academic and social dimensions of the university. Furthermore, the less well integrated the student is, the more likely it will be that the 102 student will not be enrolled at the subject institu- tion one year after matriculation. As anticipated, the variable measuring the number of hours a student worked per week had a significant negative effect on both academic and social integration. However, this seemed somewhat incongruous with the fact that the variable measuring hours worked per week had a significant positive direct effect on persistence. In other words, the more hours a student worked per week, the less integrated he/she was in the academic and social dimensions of the institution. However, the more hours a student worked, the more likely he/she was to be enrolled at the subject institution one year later. Conclusions Several general conclusions seemed warranted on the basis of the findings derived from this study. 1. The findingsfi of this study suggested that Tinto's (1975) model is a useful framework for understanding the process of persistence/withdrawal among transfer commuter students in urban university environments. While the amount of variance explained in the dependent variable, persistence, was modest (222), it is comparable with what is reported in other studies (see Chapter 3). However, we are still a considerable way from "explaining" the persistence/withdrawal decisions of transfer students. The weak explanatory power of the model could be a function of inadequate operational definitions of the model's variables. In addition, some important determinants of 103 persistence may not be specified in the model, and undoubtedly, some of the determinants of persistence are due to influences outside of the institution or personal propensities which are difficult to measure. Recently, Bean (1981) has attempted to identify some of these additional determinants of persistence, both inside and outside of the institution, which are not included in Tinto's (1975) model. Some of these determinants are: family responsibilities, opportunity to transfer or get a job, perceived practical value of education, major and occupational certainity, family approval of major and institution, and absenteeism. Astin (1974) suggested several other variables which may also be important determinants of involvement and persistence: student's perceived locus of control and attributional inclinations, financial aid policies, and rules governing residency. Still other variables which might explain the persistence/withdrawal behaviors of transfer commuter students are: level of involvement in activities outside of college, timing of classes, and parking regulations. What happened during the transfer student's first year at the institution was more important in explainingg sub— sequent persistence/withdrawal behavior than were the characteristics the studentgpossessed prior to matriculation. The patterns of influence found to exist between the variables in Tinto's (1975) model when it was tested among transfer students at an) urban commuter institution were more similar to what was found when the model was tested among 104 freshman students at residential institutions than what was found when the model was tested among freshman students at urban commuter institutions. For example, when Pascarella, Duby & Iverson (1983) tested Tinto's (1975) model. among freshman. students at 21 commuter institution, they found that the pre-enrollment characteristics, as a set, were associated with the largest increase in 1R2. This finding, coupled. with the fact that several of the pre-enrollment characteristics (e.g., gender, academic aptitude, secondary school achievement) had significant direct effects on persistence, led these researchers to conclude that "the cluster of characteristics which commuter students bring to college may be a factor of equal, if run: greater, importance in their subsequent persistence/withdrawal decisions than their actual experiences of college once enrolled" (Pascarella, Duby & Iverson, 1983, p. 93). However, the findings in this study did not support this conclusion. As hypothesized, among the transfer commuter students in this study, the integration variables were associated with the largest increase in R2, and the effects of the pre-enrollment characteristics on persistence were mainly indirect, transmitted through the college experience variables. In other words, what happened during the transfer student's first year at the subject institution seemed more important than the student's pre-enrollment characteristics. Interestingly, this finding was similar to the results of 105 tests of the model among freshman students at residential institutions (see Chapter 3). That is, what happened during the freshman student's first year was more important than their pre-enrollment characteristics in subsequent persistence/with— drawal decisions. Secondly, when they tested Tinto's (1975) model among freshman commuter students, Pascarella & Chapman (1983) found that social integration had neither a direct nor an indirect effect on persistence. 0n the other hand, Pascarella, Duby & Iverson (1983) reported that social integration had no effect on institutional commitment and £1 negative effect on persistence. These findings led these researchers to conclude that among commuter students in. urban. university settings, opportunities for social involvement were either so few or were perceived by students as so unimportant, that the concept of social integration had little meaning in terms of bonds to the institution. Furthermore, they suggested that the socially integrated student may leave the commuter institutional environment in order to find increased opportunities for social involvement consistent with their personalities. In contrast, the data. analyzed in this study' did. not replicate these findings. As ‘hypothesized, in this study social integration had a positive direct influence on institutional commitment, and institutional commitment, in turn, had a significant positive effect on persistence. (Although there was a negative ‘beta. weight between social integration and persistence in this study, it was not 106 significant at the .05 level.) Once again, this finding was similar to the results of tests of Tinto's (1975) model among freshman students attending residential campuses (Munro, 1981; Terenzini & Pascarella, 1977, 1978). One possible explanation for the variation in these findings could be that in terms of their background characteristics and the way in which they interacted with the university environment, the transfer students in this study were more similar to freshman residential students than freshman commuter students. In fact, some of the students in this study may have actually been freshmen at residential institutions one or two years before. This suggests that commuter students may, indeed, be more of a heterogenous rather than a homogeneous group, and that different groups of commuter students may differ widely both in terms of their pre-enrollment characteristics and the way in which they interact with and become involved in the academic and social dimensions of the urban university. Another explanation for the fact that the findings in this study did not replicate the findings in other studies which tested Tinto's (1975) model among commuter students may lie in the limitations of most of the studies: small sample size, single year sample, and single institution. Furthermore, the dependent variable in this study included both voluntary and involuntary withdrawal; in other tests of the 'model, only voluntary withdrawals were included. Therefore, the conclusions drawn from studies must remain suggestive until the 107 model is tested among other groups of transfer students via multiple institutional studies or replicated at other single institutional sites in similar studies. The results of this study suggested that what happens in a transfer student's academic life might be more influential than his/her social experiences in subsequent persistence/withdrawal decisions. Similar findings regarding the importance of academic integration III subsequent persistence/withdrawal decisions among ‘both residential and commuter students are discussed in Chapter Three (see pp. 44 and 47 ). Therefore, it can be concluded that efforts to reduce attri- tion are more likely to succeed if they are focused on helping transfer students (or particular groups of transfer students) become integrated into the academic dimension of the univer- 22:1; Programs aimed at increasing a student's academic integration might involve such things as pre-enrollment placement exams, academic advising, learning assistance programs, and orientation (Beal & Noel, 1980). Astin (1984) places some of the responsibility for student involvement in the academic dimension on the faculty members. He suggests that in the classroom, faculty members should focus less on content and teaching techniques and more on what students are actually doing--a pedagogical approach which directs attention toward the motivation and behavior of the student (Astin, 1984). Gender was the only pre-enrollment characteristic which had a 108 significant direct effect on persistence. Men were more likely than women to still be enrolled at the subject institution one year after matriculation. However, the fact that none of the interaction terms tested in this study were significant implied that although women were more likely to drop out, the relative influence of the variables in the models was the same for both groups. Therefore, it was difficult to draw any conclusions from this finding. It may be that in order to get a better understanding of the persistence/withdrawal ‘behaviors of female transfer students, the variables in Tinto's (1975) model would need to be redefined, new variables added, or even an entirely new model conceptualized. There is growing support in the literature for the idea that an environment that encourages the growth and development for men may be quite different than an environment that encourages the growth and development for women (Forrest, Hotelling & Kuk, 1984; Gilligan, 1982). Lack of finances is frequently the reason students cite for dropping out of college. As the literature suggests, this is particularly true of transfer students and commuter students. However, based on the results of this study it can be conclud- ed that a student's financial status did notgplay a major role in explaining the spersistance withdrawal behavior among the transfer commuter students at the subject instutition. One indicator of a student's financial status is the number of hours a student works per week; while attending college. When he studied the impact of work on persistence, 109 Astin (1975, 1977) found that having a part-time job usually increased a student's chances of persisting. On the other hand, for students who worked full-time, the positive effects of employment were not only lost, but actually reversed. Interestingly, in this study, the more hours a student worked per week, the less well integrated he/she was in the academic and social systems of the institution, but at the same time, the more hours the students worked per week, the more likely the student was to be enrolled at the subject institution one year later. One plausible explanation for this finding is the fact that the urban universities are located in large metropolitan areas and are usually easily accessible, therefore making them particularly attractive to students who are working part-time or full-time and for whom time is a critical element. Although they may not be as well integrated into the academic and social systems of the university, the fact that they are still enrolled one year later' 'may' be indicative of their determination rather than their level of involvement. A second explanation for why a student who works more hours per week was more likely to be enrolled one year after matriculation in this study is that the more hours a student works, the more money he/she has and, therefore, the more likely he/she will have the money to pay the tuition in the future. Astin (1975) offers still another explanation for this finding. If the student establishes a particular pattern of 110 employment early in his/her tenure at a particular university, the impact of working a considerable number of hours may not be as negative as it ‘would ‘be if the student suddenly 'began working more hours sometime after matriculation. While the nature of these findings are somewhat difficult to interpret, it does seem clear that transfer students can, apparently, tolerate a substantial invest- ment of time in work without suffering, a completely debilitating effect on their academic progress. Implications The results of this study have several implications which might be useful to university decision-makers who are interested in increasing the retention rate of the transfer commuter students at the subject institution. Inasmuch as the transfer students in this study are similar to transfer students at other urban universities, the results of this study may be helpful to decision-makers at other urban universities. 1. More attention should be paid to identifying, implementing, and evaluating institutional interventions which encourage transfer student involvement in the academic and social dimensions of the university. Although we are 21 long way from fully understanding the persistence/withdrawal behaviors of transfer commuter students in urban university environments, there is sufficient evidence to suggest that what happens during a transfer student's first year may be more important in subsequent persistence/withdrawal decisions than the particular background characteristics, commitments, 111 aspirations (n: aptitudes that student brings to college. Therefore, intervention strategies aimed at increasing a transfer commuter student's level of integration or involvement in both the academic and social dimensions (but primarily the academic) of the university should prove most beneficial in increasing retention rates. As Beal & Noel (1980) point out, those designing such intervention strategies must keep in mind that particular target groups are best aided by particular action programs. In other words, different intervention strategies may need to be pflanned for students with different pre-enrollment characteristics. In this study, five pre-enrollment charactistics tuui significant direct effects (n1 academic integration: full-time/part-time status, ethnic background, age, achievement, and parents' education. These findings suggest five target groups for whom intervention strategies might be focused: part-time students, nonwhite students, students particularly young in age, students who have lower transfer grade point averages and/or high school grades, and students whose parents are not highly educated. Below are a number of suggestions for increasing student involvement among transfer students: a. Develop as ‘many opportunities as ‘possible for students to work on-campus. b. Encourage students to participate in extra— curricular activities. Since, in terms, of 112 retention, what happens in a transfer student's academic life might be more influential than his/ her social experiences, providing extra- curricular activities through a student's college and/or with faculty members might be most beneficial. Find ways to encourage greater student involve- ment with faculty. Some ways in which this might be done include: mentor programs, an academic advisement program involving intensive work with faculty, or undergraduate research assistantships. Provide academic support/learning services. Some of these services might include mini courses, reading labs, personal counseling, or supplemental class instruction through a student learning center. Expand orientation activities to include such things as: a mentor program, a special orienta- tion session designed for transfer students, special workshops for nontraditional students. "Create a greater sense of community and more opportunities for student involvement through the use of 'cluster' arrangements--small and relatively autonomous colleges created within the larger institution" (Astin, 1977, p. 257). Train students to work as peer advisors in a 113 variety of areas ‘within. the ‘university, e.g., financial aid, academic advising, tutoring. Although most of these suggestions are fairly specific, Astin (1984) takes a more global approach to the need for developing strategies that 'will increase student involvement. He suggests that, . . all institutional policies and practices—— those relating to academic and nonacademic matters-- can be evaluated in terms of the degree to which they increase (n: reduce student involvement. Similarly, all college personnel-—counselors, student personnel workers as well as faculty and administrators-~can assess their own activities in terms of their suc- cesses in encouraging students to become involved in the college experience (Astin, 1984, p. 307). Intervention strategies should be aimed at faculty as well as students. In this study, academic integration played a major role in the subsequent persistence/withdrawal decisions of transfer students. Three different dimensions of academic integration were measured: cumulative grade point average, perceived level of intellectual development, and perception of faculty concern for quality teaching and student de- velopment. Therefore, efforts to reduce attrition among transfer students should. not. only focus on. helping the students become integrated into the academic dimensions of the institution (see p.111), but also on encouraging the faculty to examine their approaches of pedagogy and how those approaches either encourage or’ discourage student integration or involvement in the learning process. Astin (1984) suggests that faculty members should 114 focus less on content and teaching techniques and more on what students are actually doing. 'Moreover, rather than rewarding faculty members solely on the basis of their scholarly research, it might be wise to also reward faculty for their teaching skills, their’ success :hn encouraging students to become involved in the college experience, and/or their out-of-class contact with students. Beal & NOel (1980) identify several other activities which focus on encouraging faculty awareness of retention and faculty development: a. A monthly faculty forum discussing teaching excellence and improvement of instruction. b. A seminar in college teaching for faculty available for graduate credit. c. Improvement of instruction grants. d. A two-day, pre-semester workshop for new faculty members on effective advising, teaching, and teaching evaluation. Effort should be made to maintain or create a student's sense of loyalty to the institution. In this study, institutional commitment or attitude toward the institution was an important predictor of persistence among the transfer commuter students in the sample. Several variables which preceded institutional commitment in the model were influential in determining institutional commitment and, therefore, should 1”! con- sidered when determining strategies to maintain or create a 115 student's sense of loyalty to the institution: academic integration, institutional commitment I (measured prior to matriculation), social integration, marital status, and prior academic achievement. In other words, students who had a positive attitude toward the subject institution prior to matriculation were more loyal to the institution one year later. Students who were involved in the academic and social dimensions of the subject institution were more loyal to the institution one year after' matriculation. Single students were more committed to the institution than were married students. And, finally, the higher the level of prior academic achievement the lower the level of commitment to the institution one year after matriculation. These findings suggest that when creating strategies that will encourage the development of loyalty to the institution, consideration should be given both to students' impressions of the university prior to matricu- lation as well as their experiences in the environment after matriculation. For example, marketing strategies may need to be developed that realistically portray the strengths (M? the institution. Of particular importance would be those features of the university environment which would be appealing to students transferring from lother four-year institutions (a major determinant of institutional commitment I) and students with higher levels of prior academic achievement. Once enrolled, opportunities for transfer students to 116 become involved in the academic and social dimensions of the institution need to be made available. Some of these are discussed more fully on page 111. Sgggestions for Further Research Six suggestions are made for future research on transfer student persistence/withdrawal behaviors: 1. The causal model tested in this study failed to account for 88% of the variance in the persistence/withdrawal behaviors of transfer students. The 'main task. of future investigators will be to identify the missing determinants or the interaction effects which will reduce the unex- plained variance. Some of the determinants which might be tested are discussed on p. 103. 2. Although academic integration and institutional commitment were demonstrated to be important predictors of persistence, they were not well explained themselves. Better predictors of these variables should be found. 3. One of the advantages of path analysis is that it is possible to determine both the direct and indirect effect (total causal effect) of one 'variable (n1 another. As Pedhazur (1982) points out, when investigators 'want to study the differential effects of several variables on an endogenous variable, it is their total effects which should be compared. 'However, until recently, the calculation of indirect effects from even a moderately complicated path diagram. was not only tedious and cumbersome, but the potential for errors of logic and computation was great. 117 Furthermore, even if the indirect effects were calculated, it was not always possible to determine whether the indirect effect was statistically significant. Recently, however, two researchers (Wolfle & Ething— ton, 1984) have developed a computer algorithm for the calculation of standard errors for indirect effects. With this information, the computation of significance tests for indirect effects will be relatively easy. Therefore, identifying the significant indirect effects of the varia- bles in the model on persistence is another important direction for further research. Institutional research studies need to be conducted that determine what about the university environment either encourages or discourages female transfer students to stay or withdraw from particular institutional settings. As several authors suggest, it may well be that an entirely new model of attrition will need to be developed before we fully understand the differences between men and women and how those differences affect one's interactions with the institutional environment. Still another fruitful area for research is in assessing different forms of involvement or integration and how that involvement has an impact upon both student development and retention. For example, as Astin (1984) suggests, "a time diary could be valuable in determining the relative importance of various objects and activities to the student" (p. 306). It would also be useful to assess the 118 quantity and quality of effort students devote to various activities within the academic and social dimensions of the institution. Finally, it would also be useful to explore the influence that various individuals have on a student's level of involvement in the learning process, including peer groups, faculty members, counselors, and student affairs professionals. 6. The relationship between student employment and persistence is still another question that should be addressed in future research. For example, how important is the number of hours worked? Does it matter if the work takes place on campus or off campus? Does the type of work in which a student is engaged make a difference in subsequent persistence/withdrawal decisions? A Final Note The field of student attrition has grown tremendously over the past two decades. Although it would be foolhardy and counterproductive to believe that institutions of higher education can eliminate attrition, institutions can act to reduce, within reason, drOpout among particular groups of students. As Tinto (1982) points out, The difficult question, of course, is the net cost and benefit of such efforts. In pondering this question, one should note that those institutions that act to improve the total quality of their educational activities are more likely not only to retain more of their abler students but also to attract a greater share of students during the next two decades . . . Not infrequently, successful retention programs become opportunities for institutional self-re- newal, an outcome which, in the long-run, may be more beneficial to the institution's well—being than the simple reduction of drop out rates (PP. 698-699). APPENDIX A NEW STUDENT ORIENTATION SURVEY llD-ll/ l12/ I”! [14-15/ /16/ [17/ /18/ /19/ /20/ ,121-26/ /27/ l28/ 1.0. 12. 13. 1‘. 15. 119 APPENDIX A IEH STUDENT ORIENTATION SUIVEV hue 2. Social Sec. 0 [14/ :2::::.z°:;,:'::.r:'.::.:‘::::::°:- m: 2': .::::';:::::::::,':::;:t'::,::.i.::°.:::::z.:t :11, what was your averaoe nade in high school? (Please check out) (1)_ All. (2)_ lo (3)_ III. (4)_ I- (S)_ CICO (S)_ c. (7)__ D or below Ihat is the highest academic degree you expect to obtain? (1)_ less than a Bachelor's Degree (5)__ II.D.. D.D.S.. D.V.h.. etc. (2)_ Bachelor's Degree (S)_ u..l. or J.D. (loo) (3)__ Haster's Degree (I.A., its.) (7)__ Other (¢)_ Ph.D. or Ed.D. Now any of your close friends fro high school eill also be attending UIC this fall? (If none. out 0) in applying to colleges. was UIC your: (1)___ 1st choice; (2)_ 2nd choice; (3)____ 3rd choice; (4)___ 4th or lower choice? if UIC uas _ngg your first choice. as it because of the absence of on-cenous Musing? (1l__ Yes (2)_ ho ho- ieoortant is it to you to graduate frn college? (Please check Ill!) (1)_ Extra-sly inortant (Z)___ Very inortant (3)_ Some-hat lloortant (i)__ has at an Ieoortsnt ho- iroortant is it to you to receive a degree fro- Hit? (1)_ Extra-e1, lmrtant (2)__ Very lmut (3)_ Smat Inortant (a)_ not at an [Want How confident are you that you node the right decision in choosing to attend Ult? (Please check DRE) (1) Estreeely Confident (1) Very Confident (4) Somewhat Confident (4)_ no: at an Confident what '1?!” geLt estieate of your parents' icuoined) total income during the last year? }__ Do you consider yourself financially independent of your parents? (1)___ (2 Ihen you register at DID this fall, will you he: cg 8 u 8 (1)__ a ne- fresinanf (2)_ a new transfer? Ihat is the higgest level of fornal education obtained by your parents? (Check out category for ggch parent) fatnr I29! I”! Dru-tar School or less (14 years) Some high School (9-11 years) high School Graduate (12 years) Sal! College toll Graduate (Bachelor's Degree) Sale aduate Study Received Graduate Degree llllllll" ‘2' “ARAAAA l) 2) 3) ‘) 5) 6) 7) (o'er) I31] [32/ [33/ /3‘/ l35/ [36/ I37] 138/ 139/ [Sal I‘ll I62! ll3/ /“/ I45! /‘6/ l‘7/ [‘8/ [49/ [50/ l51/ [52/ [53/ 16. 17. 120 Hhat is your best guess as to the chances that you will: (Circle ONE number for each ital) Very Ver Good Sale Litt e lo Chance Chance Chance gance Make at least a '8' average 3 2 3 4 Have to work during college 1 2 3 4 Fall one or lore courses 1 2 3 4 le satisfied with UIC l 2 3 a Droo out of college permanently (exclude transferring) 1 2 3 4 Transfer to another college before graduating 1 2 3 a Droo out of college tenoorarily (exclude transferring) l 2 3 4 Find it easy to lake friends with other students at UlC l 2 3 4 Find 01C to he intellectually stiaulating 1 2 3 4 Find sufficient oooortunities for extracurricular involve» lent at UIC l 2 3 Q Find anole oooortunity for social activities at UIC 1 2 3 I Find high quality acadeaic programs at 01C 1 2 3 8 indicate the innortance to you personally of each of the following: (Circle ONE nuaoer for each item) Very Somewhat hot Essential lnoortggt Important lnoortant Living on campus 1 2 3 ‘ Interacting frequently with faculty outside of class 1 2 3 ‘ Having cabs. student friends on calous 1 2 3 O Studying with other students 1 2 3 4 having soleone froe your high school or DfeleuS college as a friend on caaous 1 2 3 4 Deco-ing accolnlished in one of the perforeing arts (for example. acting, dancing. etc.) 1 2 3 4 Deco-ing an authority in ay field 1 2 3 4 Influencing the political structure 1 2 3 4 influencing social values 1 2 3 ‘ Having adainistrative responsibility for the sort of others 1 2 3 4 Raising a faeily 1 2 3 4 I54] /55/ /56/ [57/ 158/ I59] 150/ /61/ l62/ l53/ [6‘] /65/ /67/ ISO] [69/ [70/ I71/ I72! I73! I"! 175/ [76/ 20. 121 Very Somewhat Essential lroortggt mortant Mortant leing well-off financially l 2 Helping others who are in difficulty 1 2 Making a theoretical contribution to science 1 2 writing original works (peels. novels. short stories. etc.) 1 2 Creating artistic work (painting sculpture. decorating) 1 2 2e13, successful in a business 17 CI! 1 2 Deco-in, involved in prograls to clean up the environaent 1 2 Developing a aeanin ul philosophy of life gf l 2 Participating in a calamity action progran 1 2 Keepi up-toodate with pol tical affairs 1 2 i Gaining a broad, liberal arts education and appreciation of ideas 1 2 Gaining knowledge and skills directly app lcable to a career 1 2 Learning nore about ayself, ey values. and an life's goals 1 2 Learning how to get along with different kinds of people and enhance ly ‘ interpersonal skills 1 2 how do you rate yourself on each of the following traits compared with the avergge person of your age? Give your,nost accurate estimate.of yourself. (Circle DliE nunoer for each iten) Above Acadeaic ability 1 2 Motivation to achieve 1 2 Leadership ability 1 2 lathe-atical ability 1 2 Hechanical ability 1 2 Originality 1 2 Popularity 1 2 Popularity with the opposite sea 1 2 Self-confidence (social) 1 2 3 3 lot I 4 Ielow Averagg UUUUUUUUU (over) 122 Above Ielow 5228 M 92.22% [77/ Self-confidence (intellectual) 1 2 3 I73] lhderstanding of others 1 2 3 [79/ writing ability 1 2 3 [80/ Artistic ability 1 2 3 III Public speaking ability 3 2 3 I2! Athletic ability 1 2 3 /3/ Physical attractiveness 1 2 3 I8! Detergination 1 2 3 lS-S/ 21. which of the following is your st likely academic najor at UIC? (Please check only 0i ) ( 1)_ Social Sciences (for asanole; Psychology, Econoaics. Sociology. Anthropology) ( 2) Natural Sciences (for eamle; Chemistry. Physics. Biology) ( 3) Pathanatics or Colouter Science ( a) Manities (English, Philosophy. Music) i S)___ Education ( 6)__ business ministration i 7)__ Engineering ( a)_ Art and Architecture ( S)_ health/Physical Education (10)_ Undecided THINK '01! FOR YOUR PARTICIPATION APPENDIX B UIC EXPERIENCE SURVEY 123 APPENDIX B The University of Illinois at Chicago Transfer Student Experience Survey 1. Name 1.9 2. Social Security Number PART I Instructions: Please complete the following questions to the best of your knowledge. Respond to every item. IO 3. What is the highest academic degree you expect to obtain? (Check one) a. less than a Bachelor's Degree e. _. M.D.. D.D.S., D.V.M., etc. b. __ Bachelor’s Degree f. __ LLB. or JD. (law) c. Master's Degree (ILA, MS.) g. __ Other d. _ Ph.D. or Ed.D. n 4. How important is it for you to graduate from college? (Check one) . a. Extremely Important b. Very Important c. Somewhat Important d. Not At All Important 12 5. How important is it for you to receive a degree from The University of Illinois at Chicago? a. Extremely Important b. Very Important c. Somewhat Important d. Not At All Important ‘ 13 6. How confident are you that you made the right decision in choosing to attend The University of Illinois at Chicago? a. Extremely Confident b. Very Confident c. Somewhat Confident d. Not At All Confident 14-16 7. During the last year, approximately what percentage of your total school expenses (tuition. books and supplies. transportation) did your parents or spouse contribute? % 17-18 8. During the last year. approximately how many hours per week did you work either on or off campus? hours per week 19 9. How difiicult has it been for you to secure financing to attend The University of Illinois at Chicago? a. Extremely Difi'tcult b. Very Difficult c. Somewhat Difficult (I. Not At All Difl'tcult 124 10. How difficult do you think it would be for you to leave this university and do each of the following: Extremely Very Somewhat Not At All Difficult Difiicult Difficult Difficult 20 a. Leave this university and get a full-time job 21 b. Transfer to another college. university. or junior college 22 c. Wertoscollegecruniversityasgoodasthiscne 23 11. Are you presently: a. married b. separated c. single (never married) d. divorced e. widowed 2e—25 12. Students have a variety of contacts with faculty members. Please estimate the number of times since coming to The University of Illinois at Chicago you have met with a faculty member outside of the classroom for any reason either informally or formally. (number of times) PART II I 13. Following is a list of statements characterizing various aspects of academic and social life at The University of Illinois at Chicago, and with which you may or may not agree. Using the scale to the right of each statement. please indicate the extent of your agreement or disagreement with each statement as it applies to your UIC experience by checking the appropriate response. Please check ONLY ONE response for each statement. M Not ' Strongly Agree Agree Sure Disagree Disagree 26 Fewofmycoursesthisyearhavebsenin- tellectually " ‘ " _ 27 lamsatisfiedwitbmyacademicexpsrienceat me as Iammcrelihelytoattendacultunlevenufor example.aconcert. lecturemrartshowlnowthan I was before coming to UIC 89 Iamsstisfiedwiththesatentofmyintellectual development since enrolling at UIC so Inadditiontorequiredreadingsasignmsnts.l 81 typicallyresdmanycftherecommendedboobsin ll! lyinterestinidsssandintellecuialmattarabu increassdsincscomingteUIC Ibavencidesatallwbatlwanttomaicrin..... MyacademicexpsriencestUlChasbadastrong positiveinfiuenceenmyiasellecnialgmuhaad interesting..- Gsttinggecduahsissetilqcrtnttoms ........ Ihsvepsri‘remsdaadsmicallyaswellaslantici- Minoan Hy interpersonal relatioubips with other duo dentsatU'ICbavebadapositivsinfiusacecnmy WMMonmyidsas ..................... fincemmingtoUICJhsvedsvelopsdclcsspsr- serial relationships with nth. Mia............... Icstsl’thesthsraudsntalhowatUICare ssriousstidsMs summit-”Wain havebsmpumllynimulatin: 41 42 47 49 51 57 125 The student friendships I have develped at UIC have been intellectually stimulating .................... My interpersonal relationships with other stu- dents at UIC have had a positive influence on my personal growth, values, and attitudes ................. It has been difficult for me to meet and make friends with other I am dissatisfied with my dating relationships at UIC (answer only if Few of UIC students I know would be willing to listen to me and help me if I had a personal problem Most students at UIC have values and attitudes which are different fi'om my own ......................... I am satisfied with the opportunities to par- ticipate in organized extracurricular activities at Illfl I am satisfied with the opportunities at UIC to meet and interact informally with faculty mem- horn Strongly AW Few ofthe UIC faculty members I have bad con- tact with are willing to spend time outside of class to discuss issues of interest and importance to students Since coming to UIC I have developed a close. personal relationship with at least one faculty member My non-classroom interactions with UIC faculty members have had a positive influence on my intellectualgrowthendinterestintdees My non-classroom interactions with UIC faculty have had a positive influence on my personal growth, values. and attitudes My non-classroom interactions with UIC faculty membershavehadepoeitiveinfiuenceonmy cureergeelsnnd FewdtbeUlescultymemberaIbavehsdcon- tact with are genuinely outstanding or superior teacher- FewofthelanacultymembersIhavebadcon- tactwitharegenuinelyinterestedinteaching... float of the UIC faculty members I have had contact with are genuinely interested in etudnts IoetdtbeUICfacultymmberalhavebad contact with are interested in helping students growinmorethanjustacademicareas ............... Conflicting family responsibilities have made it difficult for me to regularly attend classes. ......... Hy family approves of my attending UIC .......... Ilyfamilyapprovesofmyehoeenmqior ............. Itisunlihelythatlwillmerrybeforegreduetion (answer imly if applicable) ‘l‘ransportationtoandfromechoolbasnotbeena poblemfesme Not Sure Disagree Strongly Disagree PART In 14. Indicate the importance to you personally of each of the following: (Circle ONE number for each item) 8828 70 71 73 74 75 76 73 II 126 Very Somewhat Not Essential Important Important Important Becoming accomplished in one of the performing arts (for example. . acting. dancing. etc.) I 2 3 4 Becoming an authority in my field I 2 3 4 Influencing the political structure I 2 3 4 . Influencing social values I 2 3 4 Having administrative responsibility for the work of others I 2 3 4 Raising a family I 2 3 4 -Being well-off financially l 2 3 4 Helping others who are in difficulty I 2 3 4 Making a theoretical contribution to science I 2 3 4 Writing original worhs (poems. novels. short stories. etc.) I 2 3 4 Creating artistic work (painting. sculpture. decorating) I 2 3 4 ..Beingsucceesfulinabusineesofmyown I 2 3 4 Becoming involved in programs to clean up the environment I 2 3 4 Developing a meaningful philosophy of life I 2 3 4 Participating in a community action program I 2 3 4 Keeping up-to-dete with political afi‘airs I 2 3 4 Gaining a broad. liberal arts education and appreciation of ideas I 2 3 4 Gaining knowledge and skills directly applicable to a career I 2 3 4 Learning more about myself. my values. and my life's goals I 2 3 4 Ianrningbowtogetalongwithdiflaentkindsofpeopleand enhance my interpersonal skills I 2 3 4 Living on campus I 2 3 4 Interacting frequently with faculty outside of class I 2 3 4 Having close student friends mi ampus I 2 3 4 THANK YOU FOR YOUR PARTICIPATION! 127 OF IUJNGS AT CHICAGO Office of the vice Chmcellor tor Student Affairs 2705 UfllVOYSII’y Hell (MIC 500) Bax 4348. Chicago. Illinois 60680 (312) 996-7654 lane Address City State ZIpcode Dear Hr./Hs. Lastnaae, In the Fall of 1983 you were one of more than 1500 new transfer students who started classes at The UnIversIty of Illinois at Chicago. Prior to the beginning of classes you attended an orientation program for new students. Among other things that day, you completed a survey which asked you a lot of questions about your expectations for your first year. Now we want to compare your expectations wIth what your fIrst year was actually like. He will use your perceptions to help sake changes In both the academic offerings and the supporting services of the University for transfer students. Please take a few sinutes and complete the enclosed survey; It should only take you about fIfteen aInutes. when you are finished, return the survey In the self-addressed stamped envelope which Is enclosed. He would like to have the completed surveys back before finals week. Since tIae and costs prevent us froa contacting all of the new transfer students, we have confined our survey to a sample. That aeans your candid response Is vItal to the validity of the information we gather. Our abIlIty to generalize the results of the survey to all new transfer students depends entirely on your willingness to help us. Response to the survey Is coapletely voluntary. Your decision to participate or not participate wIll In no way becoae part of your permanent record at UIC. tour Identity and responses will be held In strict confidence and used only for the purposes discussed above. We sincerely appreciate your help In our efforts to understand student life at UIC. Thank you. Sincerely, Thomas I. Beckhaa Vice Chancellor for Student Affairs APPENDIX C RESULTS OF FACTORIALLY DERIVED SCALES MEASURING ACADEMIC/SOCIAL INTEGRATION AND GOAL/INSTITUTIONAL COMMITMENT 128 Hm.l mm.l Hn.l mm. 05. mm. mm. em. 9:5. @533 onum sac ma Scum uaouommfiv movauwuum vow mosHm> o>mn mufimum>fia= mHnu um muaovsum umoz SmHnoua Hmdowuon m on: H Ha we aHmn can we ou amumHH ou waHHHH3 on vHoos aodx H mucowsum ecu mo 3mm mudowsum uwsuo nuHB mvamfium oxwa can uoofi ou we now uHaonmHv down was uH mmmvfi :H umoumunfi can nusouw HmauooHHmudH he do ouaoaHHaH o>HuHmoa m on: o>mn muaovouw Honuo Sofia waHanOHuMHmu Hmaomumaumudw >2 mmaHm> wan .movSu IHuum .ausouw Hmcomuma he do moaoanaH o>HuHmom a pm: o>m£ muaovsum Honuo Sufi? manmQOHumHou HmdomuonumucH %: wafihmmfiumm mHHm Idomuma coon m>m£ hufiwuo>fido mqu on wmQOHm>mv m>ms H maHnmwcoHum ucmvnum mnw mucovoum Hosuo nuHB mmfinm:0HumHmu Hmaomuoa omOHo chOHm>mw o>m5 H wufimuo>fid= mHSu cu mcHEoo moafim macauomuoucH msouu poem ”H onom BouH\mHmom Aome .HdHNsonH w «HHoumommmv uaoEuHano HmaOHusuHum:H\Hmou can cOHuwuwmudH HmHuom\oHEmwmo< waau=mmoz monom vo>Hpmo hHHmHuouomm mo muHomom o magma 129 mm.l NH.I Hm.l Nw. no. Nu. mu. mm. ow. mm. mLQH< wanmOH onom mudowsum ou moamuuoaaa can ummuoucH mo mosmma mmsumfic ou mmmHo mo mva luno maHu woman cu waHHHH3 mum :uHa uodu 1:00 was o>m£ H mumnaoa huHsomm onu mo 30m muosumou Hafiuoanm no devoMuwuoo hHoafiocmw mum saws uomuaoo on: m>mn H muonama muHsomm onu mo 30m mucmvsum ca wmumoumuca xHoawoaow mum nufia nomuaoo on: m>m£ H muonama >uH=umm msu mo 3mm wcH£umoH can unmadem>oo uamwnum How mcumoaoo HuHsumm "HHH mHmum wuonama kanumm nufi3 hHHmauomaH uowuoucw can woos ou moHuHasuHomao onu Laws wmammaumm an H menace huHaumm mac ummmH um sufis afinm IGOHumHmH Hmaomuoa .omOHo w onOHo>mv o>mn H huHmuo>Hcs mfinu ou waHEou ouch maOHumuHmmm can mHmow umoumo he so oocmSHHaH m>HuHmoa N no: o>m£ zquomH spas macauumumuaw EooummmHocoa %z mmowa CH umopouafi can :uBouw HanuooHHoucH %a do ooamnHde m>Hunom m on: w>mc muHSUmm nuH3 meowuumuouafi Booummeucoa >2 mowsuHuum van mmsHm> .suaouw Hmcowuoa >8 no mocoaHm IaH o>HuHmom m on: o>m£ hquUdu nufis maoauomuoucfi Eooummeodoc m2 zuHsomm nuHa maowuomumucH ”HH onom amuH\onom A.e..aoue s mange 130 an. mna~< mHmum HH. me. He. me. mm. mm.l cc. we. we. cm. mm. wchmOH oonHoo 8ouw oumnvmuw ou m8 you ucmuuoaaH mH uH uamauHEEOU Hmoo wow HmaOHusuHumcH u> mHmom vHoos H wmquHoHucm H mm HHM? mm >HHmUH8mwmom vmahomuma m>m£ H >uHmuo>Hos mHsu ou mafiaoo daemon mos H cmnu 308 Aaozm who no .oHSuooH .uuoucoo m .mHa8mxo Howv uam>o HmusuHou m vacuum ou >meHH ouo8 8m H >unuo>Hca mHnu Cu wsHEOU moaHm vmmmmuodH mm: mumuuma HmnuooHHmuaH can mmovH 8H uwmumucH >2 waHumH88Hum >HHm=uowHHou8H Goon o>m£ umw> anu momunoo >8 mo 3mm >uHmuo>Has mHsu um mocmfiuma IXm UHEmvmum >8 Laws vameumm 8m H momvH 8H umw lumuoH can Luzouw HmouooHHmuaH >8 so moamsHmcH o>HuHmoa m on: mm: moaoHuoaxo oHamwmom >2 >uHmuo>Hao wwnu 8H waHHHouao moaHm u8m880H0>mw HmsuooHHouaH >8 mo ucmuxo mnu nuH3 vonmHumm an H u8m8a0Ho>ma HmauumHHmudH wan oHEowmo< ”>H onom wcHnommu 8H vmummuoucw >HmaHsaow mum suHs uomuaoo cm: m>m£ H muon8m8 >uHoomw umoz mmmum 0H80wmum push amnu 0808 8H 30pm masocSum waHmHm: 8H cmuwmumudH mum nuHs uomudoo on: w>mn H >quumH mzu mo umoz amuH\mHmom A.e.uaouv o mfinme 131 qq.| o8 ou ucmuuoa8H uoc mH movmum voow wdHuuou mq.l 8H Homm8 ou and: H umnz HHm um mowH 08 o>mn H mm.| >uHmHm>H8= mHnu 808m mumovmuw ou m8 ou uamuuoaafi uoc mH uH No. HHmm uxm: >uHmuo>Hcs mess on Houmfiwmu Haas H owed mfimxaa ma uH mo. >uHmuo>Ha= mHnu vacuum ou waHmoonu 8H GOHmHomv uanu may ovm8 H umnu udmemdou 8m H m£QH< wswmeH 8ouH\meom onom A.@.ufioov o QHQNB APPENDIX D MEANS, STANDARD DEVIATIONS AND ZERO-ORDER CORRELATIONS BETWEEN INDEPENDENT AND DEPENDENT VARIABLES IN MODELS 1 AND 2 132 a~.. cos. coo-....... .ou 3.... main 2 .83.! 7833...... .m— 3.& 28. ...ilzela_!J.-. mo:.~ ooo. oo..o.oo.c. ...ooa .H. ....u ooo. oo..o.oo.¢. u_-ooo< .o. «3.2 23.: ..ee) 5: 3.33 :33. .m- noohn ooo. . .ooe._aeou _ooo..a....o_ .o. “ntu 98. .zllzela.!8 A. coo." no... ...._»..u¢ .ooou. ...: ... Hag. Han. co..a._..o. ...-a... ... mom.. ooo. .oono...¢u< .o. 3:. 3m. :3 .201: in H3. mac. .3 a 3:33.53 .3 oz. 2.. octezzzu.oa ooo.. «an. a... oo_.o....¢_-o< ..oe_.ao .oa om_. ooo.. mo.. osn. eo_osq .9. so: .3: 8a.. a... 8.. 9.9.2.8 .8 -o.- nnn.- onn.. ooo._ non. om_. eoug.o.. ..- aso. .oo. moo.. .oo. ooo.. oH:. ~ch. .3...» oe_»-..oa\oe.»-_.aa .s m.o. ..o.. “no. moo. m... ooo.. “on. ~oo. .3...» .....43 .o m.o.- a... on... _o... on..- no..- ooo. ~o~.o a~m.- coo .u no.. ~ao. o:o.- ago.. on.. moo. coo. ooo.. one. omH. oeoocoaooo u.og.u .o «~o. «mo. ono.. ~oo.. ono.. moo. .mo. Noo. ooo.. soc. om“. ..oooo .H ono.. «.o. moo. moo. «Ho. moo. .m.. «o... .oo.. ooo.. coo.~ ooo. s....“ ...uooo_a .u :~o. ooo.- ~oo. ..o. ono. «no. om_. moo. coo. .no. ooo.. Han.. noo.n =o_.ooao. ...ooeoa .. on no so no H o m c H n _ on : o.n~_.oo Saiozs N can H mHovoz 8H moHanum> unencunwd vcm ucovconuvcH :oo3umm mcoHumHouuoo uovuonouom vac mcoHuoH>on unaccoum undue: Q xHszmm< h oHnaH 133 80.— HHH. So. one. non. oeo . on.. "no. H... oeo.. nnn. one. .... .nn. ooo.. nee. «mo.. mne.. Hoe.. noo.- ooe.. Hoe. .on. .... a... who. one.. ooo.. - o.o. «2. 3n. 3e. m3. 2.... 2“. 8o.— o~o.- one. o.o. one. o.o. nuo.- ano.-. one. oeo.. nno.. .no.. .oo. nee. o.o. «no.- o.o.. o.o.. one. ooo . nae. .no.- “no. see.. n... ono.- o~o.- Hue. one. o.o.- ooo.. no..- «no. see. nee. nne. «so. one.. «no. noo.- n... o.~.- eoe.. one. ..e.. nee. nee. one. one.. nee. o.o.. one. one. one. nan.- ooe.. nae. sne.- oo..- «no.. «...- no... o.o.. o.o.. “He.. mm... mm.. o.o.. o.... ooo.. nee. moo. Heo.- noo.- one. .no. m.o. ane.- one.. on..- n... n.n.. on..- ....- .... o.o. o.o. noo.- Hne.- nno.- one. «no. nae. no..- an.. on~.. o.o.- nan. ..... ano.- o.o.- one. ..o. .no. «no.. o.o.. o.o. n.o. no..- no.. o.o. no..- «no.. mne.. nno.. one.. one. o.o. «ne.. o.o. Hme.. no.. «so.. H... n.o. «a... one. 30. H3. .3. a... an..- «we... 3.... nme. one. 3.. 3... .3. o.o. ..o. o.o. nno.- ooo.- ane.- one. nne.- o~e.- see. one. «no.- one. o.o. nno.- H.o.- nae. o.o. Hoe. o.o. noo.- .n.. nee. one.. o.o.- .oe. o.o.- nae.. one. nno. ..e.- nee. .ne. on.. one. .ne.- o.o.- nno.- .co. one. n.... Hoe. noo.. «no. mne.. o.o.- noo.- ....- moo. o.o.- ono.- Hoo.- sno.- nno.- an..- o.o. non. one. one. o.o. -o.- o.o.- no..- o.o. one. o.o. nee. nae. «no.- o.o. one. fine. moo.. .me. one. no.. «no. o.o.- one. neo.- on.. one. ..o.- one. o.o.- on a. n. a. e. n. o. n. n. ._ e. on on an «o.o....» ..o..oooo . o.ooe LIST OF REFERENCES LIST OF REFERENCES Alwin, D.F., & Hauser, R.M. The decomposition of effects in path analysis. American Sociological Review, 1975, fig, 37-47. Anderson, E.F. A comparison of transfer and native student progress at the University of Illinois at Chicago-University Center. Unpublished report, University Office of School and College Relations Research Memorandum 83-1, University of Illinois at Chicago, 1983. Arthur, S. Designing ways to serve the commuting student. Liberal Education, 1977, 63, 316—321. Astin, A.W. Preventing students from dropping out. San Francisco: Jossey-Bass, Inc., 1975. Astin, A.W. Four critical years. San Francisco: Jossey-Bass, Inc., 1977. Astin, A.W. Minorities in American higher education. San Francis- co: Jossey-Bass, Inc., 1982. Astin, A.W. Student involvement: A developmental theory for higher education. Journal of College Student Personnel, 1984, 22, 297-308. Avakian, A.N., Mac Kinney, A.C., & Allen, G.R. Race and gender differences in student retention at an urban university. College and University, 1982, 21, 160-165. Baird, L. The effects of college resident groups on students' self-concepts, goals, and achievements. Personnel and Guidance Journal, 1969, 31, 1015-1021. Baumgart, N.L., & Johnstone, J.N. Attrition at an Australian university. Journal of Higher Education, 1977, fig. 553-570. Beal, P.E., & Noel, ‘L. What works in student retention. Iowa City, IA: American College Testing Program and National Center for Higher Education Management Systems, 1980. Bean, J.P. Path analysis: The development of a suitable methodo- logy for the study of student attrition. Paper presented at the Annual Meeting of the American Educational Research Association, San Francisco, April 1979. 134 135 Bean, J.P. Dropouts and turnover: The synthesis and test of a causal model of student attrition. Research in Higher Educa- tion, 1980, 155-187. Bean, J.P. Student attrition, intentions, and confidence: Interaction effects in a path model. Part I. The twenty— three variable model. Paper presented at the annual meeting of the American Educational Research Association, Los Angeles, April 1981. (ERIC Document Reproduction Service No. ED 202 443) Bean J.P. Conceptual models of student attrition: How theory can help the institutional researcher. In E. Pascarella, (Ed.), New directions for institutional research: studyingpstudent attrition. San Francisco: Jossey—Bass, Inc., 1982a. Bean, J.P. Student attrition, intentions, and confidence: Inter— action effects in a path model. Research in Higher Education. 1982b, 11, 291—320. Bishop, J.B., & Snyder, G.S. Commuter and residents: Pressures, helps, and psychological services. Journal of College Student Personnel, 1976, 11, 232—235. Blalodk, H.M. Theory construction. New Jersey: Prentice Hall, Inc., 1969. Braddock, J.H. Desegration and black student attrition. Urban Education, 1981, 12, 403—418. Buckley, H.D. A comparison of freshman and transfer expectations. Journal of College Student Personnel, 1971, 12, 186—188. Burnett, D. Traditional-aged commuter students: A.review of the literature. NASPA Forum, 1982, 2, 6-7. Call, R.W. A comparison of resident students' quality point averages with those commuting students. Fort Lauderdale, Florida: Nova University, 1974. (ERIC Document Reproduction Service No. ED 094-809) Carnegie Council on Policy Studies in. Higher Education. Three thousand futures: The next twenty years in higher education. San Francisco: Jossey-Bass, 1980. Chickering, A.W. Commuting versus resident students. San Francisco: Jossey—Bass, 1974. Chickering, A.N., & Kuper, E. Educational outcomes for commuters and residents. Educational Record, 1977, 22 255-261. Christian, C.E. A comparison of resident and non-resident student pgrceptions of UC-Irvine. Irvine, CA: University of California-Irvine, 1973. (ERIC Document Reproduction Service No. ED 093-258) 136 Cohen, J., & Cohen, P. Applied multiple regression/correlation analysis for the behavioral sciences. New York: John Wiley & Sons, 1975. Cope, R.G. Why students stay, why they leave. In L. Noel (Ed.), Reducing the dropout rate. San Francisco: Jossey-Bass, Inc., 1978. Cournelis, J.S., & Dolan, F.A. Perceptions and needs: The full time undergraduate student at the University of San Francisco. San Francisco, CA: lkdxersity of San Francisco, 1974. (ERIC Document Reproduction Service No. ED 094 611) Cross, P. Higher education's student. Junior College Journal, 1969, 32, 38-42. Davila, E. Urban university study: Progress report #1. Unpub- lished report. The College Board, June 1982. Davis, J.L., & Caldwell, S. An intercampus comparison of commuter and residential student attitudes. Journal of College Student Demos, G.D. Problems integrating the commuter college student to the college campus. American Journal of Orthopsychiatry, 1966, 39, 336-337. Donato, D.J. Junior college transfers and a university environment. Journal of College Student Personnel, 1973, 14, 254—259. Drasgow, J. Differences between college students. Journal of Higher Education, 1958, 22, 216-218. Dressel, P.L., & Nisula, E.S. A comparison of the commuter and non—commuting student. East Lansing, MI: Michigan State University, 1966. (ERIC Document Reproduction Service No. ED 011 967) Eddy, E. The college influence on student character. 'Washington, D.C.: American Council on Education, 1959. Fenske, R.H., & Scott, G.S. A comparison of freshmen who attend college in their home community and freshmen who migrate to college. Iowa City, IA: American College Testing Program, 1972. Flanagan, D. The commuter student in higher education: A synthesis of the literature. NASPA Journal, 1976, i3, 35-41. Forrest, L., Hotelling, K., Kuk, L. The elimination of sexism in universityienvironments. Paper presented at Student Deve10p- ment Through Campus Ecology, Second Annual Symposium, Pingree Park, C0., 1984. 137 Foster, M.E., Sedlacek, W.E., & Hardwick, M.W. A comparison of dependent commuters, independent commuters, and resident students. Journal of National Association of Women Deans and Counselors, 1978, £2, 36—42. Francis, C.M. College enrollment trends: Testing conventional wisdom against the facts. Washington, D.C.: American Council on Education, 1980. George, R.L. Resident or commuter: A study of personality differences. Journal of College Student Personnel, 1971, _L2, 216-219. Gilligan, C. In a different voice. Cambridge, MA: Harvard University Press, 1982. Glass, J.C., & Hodgin, H.H. Commuting students and cocurricular activities. The Personnel and Guidance Journal, 1977, 22, 253-256. Goodale, T.E. & Sandeen, A. The transfer student: A research report. NASPA Journal, 1971, 2, 248-263. Goodall, L.E. The urban university: Is there such a thing? Journal of Higher Education, 1970, 32, 44—54. Graff, R.W., & Cooley, G.R. Adjustment of commuter and resident students. Journal of College Student Personnel, 1970, ii, 54-57. Gusfield, J., Kronus, S., & Mark, H. The urban context and higher education: A delineation of issues. Journal of Higher Education, 1970, 42, 29—43. Hills, J.R. Transfer shock: The academic performance of the junior college transfer. Journal of Experimental Education, Hood, A.B. A method of comparing student achievement levels at different. colleges. Personnel & Guidance Journal, 1967, ‘42, 799-803. Iverson, B.K. Faculty-student informal contact and educational aspirations in commuter college freshmen (Doctoral disser- tation, University of Illinois at Chicago Circle, 1982). Dissertation Abstracts International, 1982, 32, 954A-1322A. (University Microfilms No. DA 8220016) Johnson, E.F. Characteristics and ‘needs of Indiana University freshman commuter students (Doctoral dissertation, Indiana University, 1981). Dissertation Abstracts International, 1981, 32, 1A—428A. (University Microfilms No. 8114958) 138 Keller, B.V. Significant impacts on the freshman commuter student at Bowling Green State University (Doctoral dissertation, Bowling Green State University, 1980). Dissertation Abstracts International, 1980, 41, 1252A-1811A. (University Microfilms No. 8022841) Kenny, D.A. Correlation and causality. New York: John Wiley & Sons, Inc., 1979. Kerlinger, F.N., & Pedhazur, E.J. Multiple regression in behavior- al research. New York: Holt, Rinehart & Winston, Inc., 1973. Klotsche, J.M. The urban university. New York: Harper & Row Publishers, 1966. Knoell, D.M. Focus on the transfer program. Junior College Knoell, D.M., & Medskar, L.L. Fromgjunior to senior college: A national study of the transfer student. Washington, D.C.: American Council on Education, 1965. Kysar, J.E. Mental health in an urban commuter university. Archives of General Psychiatry, 1964, ll, 472-483. Lara, J.F. Differences in quality of academic effort between successful and unsuccessful community college transfer students. Paper read at American Educational Research Assoc- iation, Los Angeles, April 1981. Laudicina, RJAH Toward effective counseling for two—year transfer students. NASPA Journal, 1974, 21, 57-61. LeMoal, M.J. Residents and commuters: A study of change in attitudes, values, and personality factors (Doctoral disser- tation, Columbia University, 1980). Dissertation Abstracts International, 1980, 42, 1252A—1811A. Lenning, O.T., Beal, P.E., & Sauer, K. Retention and attrition: Evidence for action and research. Boulder, CO: National Center for Higher Education Management Systems, 1978. (ERIC Document Reproduction Service No. ED 156 098) Lindsay, C.A., Marks, E., & Hamel, L.S. Native and transfer bacca- laureate students. Journal of College Student Personnel, 1966, 7, 5-13. Lin, N. Foundations for social researCh. New York: McGraw-Hill, 1976. Liu, R., & Jung, L. The commuter student and student satisfaction. Research in Higher Education, 1980, 12, 215-226. 139 Mayhew, L.B. Surviving the eighties: Strategies and procedures for solving fiscal and enrollment problems. San Francisco: Jossey—Bass, 1980. Minkevich, G., Rickey, L.G., & Marshall, J.C. Personality differ- ences for two and four-year college commuters. College Student Journal, 1972, p, 87-90. Moore, K.M. The transfer syndrome: A pathology with suggested treatments. NASPA Journal, 1981, 28, 22-28. Munro, B.H. Dropouts from higher education: Path analysis of a national sample. American Educational Research Journal, 1981, 12, 133-141. Nickens, J. The relationship of selected variables to performance of junior college transfer students at Florida State Univer— sity. The Journal of Experimental Education, 1970, 28, 61-68. Nie, N.H., Hull, C.H., Jenkins, J.C., Steinbrenner, K., Bent, D.H. SPSS: Statistical package for the social sciences. New Ybrk: McGraw—Hill Book Company, 1975. Nolan, E.J., & Hall, 1L1“ Academic performance of the community college transfer students: A five year follow-up study. Journal of College Student Personnel, 1978, i2, 543—548. Palm, R.J. A comparison of study of high school extracurricular activities and college extracurricular plans as related to the matriculation at a commuter university or a residential university (Doctoral dissertation, University of Nebraska- Lincoln, 1980). Dissertation Abstracts International, 1981, 42, 4682A—5271A. (University Microfilms No. 8111682) Pantages, T.J., & Creedon, C.F. Studies in college attrition: 1950-1975. Review of Educational Research, 1978, 42, 49-101. Pascarella, E.T., & Chapman, D.W. A1 multi-institutional, path analytic validation of Tinto's model of college student withdrawal. American Educational Research Journal, 1983, 22, 87-102. Pascarella, E.T., Duby, P.B., & Iverson, B.K. A test and reconcep- tualization of a theoretical model of college withdrawal in a commuter institution. setting. Sociology of Education, 1983, 29, 88-100. Pascarella, E.T., Duby, P.B., Miller, V.A., & Rasher, S.P. Ewe- enrollment variables and academic performance as predictors of freshman year persistence, early withdrawal, and stopout behavior :h1 a urban, nonresidential university. Research in Higher Education, 1981, 12, 329-349. 140 Pascarella, E.T., Duby, P.B., Terenzini, P.T., & Iverson, B.K. Student faculty relationships and freshman year intellectual and personal growth in.zi nonresidential setting. Journal of College Student Personnel, 1983, 24, 395-402. Pascarella, E.T., Terenzini, P.T. Patterns of student-faculty informal interaction beyond the classroom and voluntary freshman attrition. Journal of Higher Education, 1977, 48, 540-552. Pascarella, E.T., & Terenzini, P.T. Interaction effects in Spady's and Tinto's conceptual models of college dropout. Sociology of Education, 1979, 22, 197-210. Pascarella, E.T., & Terenzini, P.T. Predicting freshman persis- tence and voluntary dropout decisions from a theoretical model. Journal of Higher Education, 1980, 21, 60—75. Pedhazur, E.J. Multiple regression in behavioral research. New York: CBS College Publishing, 1982. Pettyway, S.B. A comparison of college resident and commuter students (n1 selected psycho—social characteristics. Disser- tation Abstracts, 1968, 22, 997-1983. (University Microfilms No. 997-1983) Phlegar, A.G., Andres, L.D. & McLaughlin, G.W. Explaining the academic performance of community college students who transfer to 21 senior institution. Research in Higher Education, 1981, _1—5-3 99-108e Rudnick, A.J. The American university in the urban context: A status report and call for leadership. Washington, D.C.: National Association of State Universities and Land Grant Colleges, 1983. Rue, P., & Stewart, S. Toward a difinition of the commuter student population in ‘higher education, NASPA Forum, 1982, ‘2, 8-9. Schuchman, H. Special tasks of commuter students. Personnel & Guidance Journal, 1974, 22, 465-470. Spaights, E. Toward a definition of an urban university. Urban Education, 1980, 12, 369—375. Spady, W.G. Dr0pouts from higher education: An interdisciplinary review and synthesis. Interchange, 1970, 2, 64-85. Stark, M. Commuter and residence hall students compared. Person- nel and Guidance Journal, 1965, 44, 277-281. Summerskill, :L Dropouts from college. In N. Sanford (Ed.), The American College, New York: Wiley & Sons, Inc., 1962. 141 Terenzini, P.T., Lorang, W.G., & Pascarella, E.T. Predicting freshman persistence and voluntary dropout decisions: A replication. Research in Higher Education, 1981, 12, 109-127. Terenzini, P.T., & Pascarella, ERIK Voluntary freshman attrition and patterns of social and academic integration: A test of a conceptual model. Research in Higher Education, 1977, p, 25-43. Terenzini, P.T., & Pascarella, E.T. The relation of students' pre-college characteristics and freshman. year experience to voluntary' attrition. Research in Higher Education, 1978, .2, 337-366. Terenzini, P.T., & Pascarella, E.T. Toward a validation of Tinto's model of college student attrition: A review of recent studies. Research in Higher Education, 1980, i2, 271-282. Tinto, V. Dropout from higher education: A theoretical synthesis of recent research. Review of Educational Research, 1975, 42, Tinto, V. Limits of theory and practice in student attrition. Journal of Higher Education, 1982, 22, 687-700. Trivett, D.A. The commuting student. Washington, D.C.: ERIC Higher Education Research Currents, 1974. (ERIC Document Reproduction Service No. ED 090 887) User's Guide: SPSS. Chicago: SPSS, Inc., 1983. Ward, R.F., & Kurz, T.E. The commuting student: A study of facili- ties at Wayne State University. Detroit, MI: Wayne State University, 1969. (ERIC Document Reproduction Service No. ED 031 901) Webb, S.C. Estimated effects of four factors on academic perfor- mance before and after transfer. Journal of Experimental Ed- ucation, 1971, 22, 78-84. Welty, J.D. Resident and commuter students: Is it only the living situation? Journal of College Student Personnel, 1976, ‘il, 465-468. Wessell, T.R., Engle, K., & Smidchens, V. Reducing attrition on college campuses. NASPA Journal, 1978, 19, 26-32. Wolfle, L., & Ethington, C. A program for standard errors of indirect effects in recursive models. Paper presented at annual meeting of the American Educational Research Associa— tion, New Orleans, April 1984. 142 Wray, F.E., & Leischuck, G.S. Predicting academic success of junior college transfers. College and University, 1971, Al, 10—16. Young, W. Admissions of the transfer students. Personnel & Guidance Journal, 1964, A2, 60-62. Zaccaria, L., & Creaser, J. Factors related to persistence in an urban commuter university. Journal of College Student Personnel, 1971, i2, 286—291. Zusman, B.J. Retention of undergraduate transfer students: A study of the fall 1982 cohort. Unpublished report, Office of Academic Planning, University of Illinois at Chicago, 1983. HICHIGQN STQTE UNIV. LIBRRRIES 111111111111111111111||1|1ll|||1|11|lll1HI 31293106671351