Psychosocial factors related to retention and early departure of two-year community college students.

Anthony R. Napoli, Ph.D.

Office of Institutional Research

Suffolk Community College


Paul M. Wortman, Ph.D.

Psychology Department

State University of New York at Stony Brook


The present study is based on the theoretical model of college retention developed by Tinto (1975, 1987, 1993), and subsequent validation efforts of others (Bers & Smith, Munro, 1981; Pascarella & Chapman, 1983a, b; Pascarella & Terenzini, 1983, 1991). The first goal of the study was to assess the validity of the model on a two-year community college sample. The second goal was to extend and further refine the model by examining the mediational influences of a comprehensive set of psychosocial measures (i.e., life events occurring during the first semester of college, social support, self-esteem, social competence, personal conscientiousness, psychological well-being, and satisfaction with the academic, administrative, and social systems of college) on the constructs within Tinto's (1987, 1993) model. Results confirm the generalizability of the model to two-year community college populations. In addition, the structural equations model revealed that the psychosocial measures have both direct and indirect effects on college persistence.


Tinto (1975, 1987, 1993) proposed a multivariate model of student retention in post secondary institutions to explain student departure from college prior to graduation. His model includes a comprehensive set of demographic, cognitive, psychosocial, and institutional factors. The model proposes that both student characteristics and interactions with the social and academic environments of the institution are the principal determinants of educational goals and institutional commitment. These characteristics and interactions account for the decision to persist or withdraw from college. Working backwards, Tinto (1993) proposed that persistence or departure behavior is directly influenced by two dimensions of commitment, namely institutional and goal commitment. Institutional commitment represents the degree to which an individual is motivated to graduate from a specific college or university. Recent efforts by Nora and Cabrera (1993) has demonstrated the importance of institutional commitment on both intent to persist and actual persistence behavior for students attending a commuter college. Goal commitment, or educational goal commitment, represents the degree to which the individual is committed, or motivated, to earn a college degree in general. In turn, institutional and goal commitments are directly influenced by external commitments or demands and the level of academic and social integration. Drawing from the theoretical work of Durkheim (1951) and Van Gennep (1960), Tinto (1993) proposed integration as a process in which the individual establishes membership (or fails to establish membership) in the college community. He distinguishes between social integration and academic, or intellectual, integration. The former represents the social ties that result from the day-to-day interactions. The latter results from sharing information, perspectives, and values common to other members of the community. Integration has also been defined as the extent to which an individual identifies with, or shares and incorporates the normative attitudes and values of his or her instructors and classmates, and becomes a member of the college community (Pascarella & Terenzini, 1991). Satisfying and rewarding interactions with the formal and informal academic and social systems of the institution lead to greater integration and persistence. Unpleasant or limited interactions inhibit integration and decrease the likelihood of persistence. Previous research has shown that academic and social integration are influenced by a variety of factors (Munro, 1981; Pascarella et al., 1983a, 1983b, 1983, 1985, 1986). These factors include student background characteristics (e.g., age, socioeconomic status, and personality needs, pre-college educational experiences, previous academic achievement, and initial experiences in college.

Tinto's model of student retention (1975, 1987, 1993) provides a heuristic and theoretical framework for understanding student behavior. Since its inception, the model has been the subject of many validation efforts. Munro (1981) observed significant direct and indirect effects for academic integration, goal commitment, and high school average on persistence for first-time, full-time 4-year college entrants. Social integration, however, had no significant direct or indirect effect on persistence (Munro, 1981), although the model of student retention postulates an approximate parity between social and academic integration. Pascarella and Chapman (1983a, 1983b), also employing a path analytic approach, observed that social and academic integration, while not directly influencing persistence, had significant indirect effects on persistence. This influence on persistence operated through institutional commitment and goal commitment. In addition, they found different patterns of results for different types of institutional settings (Pascarella & Chapman, 1983a, 1983b). Briefly, cluster-analysis revealed the presence of three distinct institutional groupings: (a) 4-year primary residential institutions, (b) 4-year primary commuter institutions, and 2-year primary commuter institutions. In residential institutions, social integration had both direct and indirect effects on persistence. Conversely, academic integration had neither direct, or indirect effects, on persistence. In both the 4-year and 2-year commuter institutions, however, academic integration indirectly influenced persistence. Among the commuter institutions social integration had neither a direct nor an indirect influence on persistence.

For commuter colleges, the findings for academic integration and social integration have been mixed. In a multi-institutional study of Chicano students attending two-year community college, Nora (1987) reported neither academic integration or social integration affected retention rates significantly. In another investigate involving two-year community college students, Nora, Attinasi, and Matonak (1990) found a significant direct effect for academic integration, but not for social integration, on retention. In a study of commuter students attending an urban university, direct and indirect effects of academic integration on persistence were observed (Pascarella, Duby, & Iverson, 1983). Contradicting Tinto's formulation, social integration was found to have a negative impact on persistence. That is, an increase in social integration represented an increased risk for withdrawal. Based on these and other findings Pascarella and Terenzini (1991) concluded that social integration may be a liability for persistence among commuter students. The investigators argue that commuter institutions are less likely to provide opportunities for social interaction. As a result, person-environment fit problems may occur for individuals with high affiliation needs (Pascarella & Terenzini, 1983a, 1991 ). Consistent with these results, academic integration, but not social integration, has been found to influence persistence, among disadvantaged students at a two-year commuter institution (Fox, 1986). Similarly, direct and indirect effects have been observed for academic integration, but not for social integration, among students attending a two-year community college (Mulligan & Hennessy, 1990).

Though several investigators have not found support for Tinto's prediction of social integration affecting persistence in commuter samples, others have. Both social and academic integration were found to have direct and indirect effects on long-term (Pascarella et al., 1986) and shorter term (Bers & Smith, 1991; Napoli & Wortman, 1996) persistence and graduation, among two-year community college students. In fact, the two latter studies (Bers & Smith, 1991; Napoli & Wortman, 1996) found that social integration made a larger contribution in discriminating persisters from non-persisters than academic integration.

Since narrative literature reviews and subjective appraisals of individual studies provide little guidance in synthesizing meaningful conclusions Napoli and Wortman (1996) conducted a meta-analysis on a set of studies which examined the relationship between academic and social integration on persistence among two-year community college students. Academic integration had a large and positive impact on persistence/withdrawal behavior among community college students. Also, social integration played a significant role in persistence/withdrawal decisions, and that the effect size for the impact of social integration on persistence is significantly greater for studies involving term-to-term persistence than those assessing the academic year-to-year outcome. The investigators concluded that, among first-time full-time community college students, the influence of one term's level of social integration is moderated, or attenuated, by the length of the persistence interval (Napoli & Wortman, 1996). This finding is not surprising considering the high rates of early (first semester) withdrawal and stop-out behavior (temporary withdrawal) among community college students (Bers & Smith, 1991; Tinto, 1993). Unlike four-year residential institutions, where stop-out behavior is less common and social ties and networks remain relatively stable over several academic semesters or years, social networks within community colleges are less likely to persist over time. Indeed, it is arguable that term-to-term, rather than year-to-year, persistence is a more meaningful measure when studying community college students. This conclusion is consistent with findings from several recent studies involving community college students (Webb, 1988; Bers & Smith, 1991; Romano, 1995; Napoli & Wortman, 1996) that, by studying term-to-term persistence, substantially increased the predictive power of their models.

Several investigators have also acknowledged an interplay between institutional size and academic and social integration (Pascarella & Terenzini, 1991; Pascarella & Wolfle, 1985; Tinto, 1993; Chickering & Reisser, 1993). Wilson, Gaff, Dienst, Wood, and Bavry (1975) presented evidence suggesting that institutional size is negatively associated with the amount of student-faculty contact. In an eight-institution sample, they found a significant negative relationship between the number of "out of class" student-faculty interactions and institutional size using self-report data obtained from both students and the faculty. It appears then, that observed model differences in social integration that emerge between 4-year residential and 2-year commuter colleges, may also be due to size differences, and not solely institution type. Within large 2-year community colleges social integration does occur and has a beneficial impact on persistence (Napoli & Wortman, 1996). Therefore, a comprehensive study of the factors related to persistence and attrition among two-year community college students should examine both academic and social integration. The study should also include factors, such as size, which influence social integration and may account for some of the heterogeneity found among other studies.

Psychological factors as mediators of integration, commitment, and persistence The psychological characteristics of the student have a major impact on both academic and social integration (Tinto, 1993). However, traditional psychological models have provided little utility in directly predicting academic success or departure from personality traits (Tinto 1993). Furthermore, attempts to correlate personality inventories with direct measures of academic success or persistence have produced inconsistent profile types (Tinto, 1993; Pascarella & Terenzini 1991). Psychological theories of departure invariably see student departure as reflecting a shortcoming or weakness in the individual, ignoring the impact of the institution on student behavior (Tinto, 1993). Such theories argue that attrition among college students could be substantially reduced by either improvement of student skills, by the selection of individuals with "appropriate" personality traits, or both. This argument, however, is not empirically supported.

Psychosocial approaches (discussed below) have produced a greater understanding of the person-environment dynamics that might lead to persistence or departure. Psychosocial factors, rather then directly impacting performance outcomes such as GPA or persistence, mediate the antecedents to these outcomes. For example, self-esteem, although not directly related to persistence, had a direct impact on three key constructs within Tinto's model, namely academic integration, social integration, and institutional commitment (Munro, 1981). Also, need for affiliation had a direct impact on social integration, and achievement need, a measure of the degree of effort and quality of effort an individual expends to surmount obstacles, was directly related to academic integration, social integration, and goal commitment (Pascarella & Chapman, 1983a, 1983b). The degree to which a student works hard on academic and non-academic tasks, or one's quality of work effort, has been associated with academic achievement (Pace, 1980, 1984). Pace investigated the influence of "quality of work effort" and observed a significant correlation between the quality of effort and actual academic achievement. Achievement need, or the degree to which an individual will work for a desired goal has also been associated with academic performance (Stern, 1970) and persistence (Pascarella & Chapman, 1983a).

Psychological adjustment to college. Baker and Siryk (1984) set out to assess psychological adjustment to college. They recognized the importance of psychological adjustment to college, as well as the importance of academic and social integration into college systems. To measure psychological adjustment to college they developed a set of self-report measures collectively referred to as the Student Adaptation to College Questionnaire (SACQ). The SACQ measures students' academic, social, and personal-emotional adjustment to college, as well as their level of institutional attachment. In assessing the predictive validity of the SACQ, Baker and Siryk (1989) reported consistent and significant correlations between the SACQ's academic and social adjustment scales and persistence. The SACQ has been standardized (Baker & Siryk, 1989) and the instrument is used by college counseling centers as a screening instrument to identify students who are experiencing difficulties adjusting to college.

Psychosocial factors may affect adjustment to college. In their review of the literature, Baker and Siryk (1989) reported that several measures of psychological maladjustment (e.g., anxiety, depression, loneliness, social avoidance, and psychological distress) interfere with adjustment and attachment to college. Conversely, measures of positive psychological health (e.g., self-esteem, psychological independence, psychological well-being, positive affect, positive personal self-concept, positive social self-concept, positive family self-concept, and positive physical self-concept) have been positively related to adjustment and attachment to college.

Life Stress. Since the earlier work by Sarason, Johnson, and Siegel (1978), it has been recognized that life demands/life stress is negatively related to GPA. Carter (1982) reported that family responsibilities were among the 5 most prevalent of 60 reasons for the attrition of older and part-time students. Berkove (1976) observed, for older female students who were married and had at least one child living at home, a significantly higher attrition rate in comparison to students with no family responsibility. Brainard (1973), Martin (1974), and Hunter and Sheldon (1980) reported family pressure and obligations were listed as major reasons for withdrawal among community college students. In a study of non-college external experiences, Metzner (1984) found that a global measure of [outside] stress was significantly related to attrition for students attending a urban commuter university.

It is unlikely that life stress has a direct impact on GPA. That is, because one may have high life demands does not mean that he or she has low academic aptitude. Rather, it is more likely that life demands influence a third factor or set of factors (e.g., a student's ability to attend class or allocate sufficient time to study) that directly impacts upon academic performance. Indeed, life demands and events inventories such as the Life Experience Survey (Sarason, et al., 1978), the Social Readjustment Rating Scale (Holmes & Rahe, 1967), and the Hassles Scale (Lazarus et al., 1980) have negatively related to adjustment and commitment measures (Baker & Siryk, 1989). These findings support the view that external life demands interfere with the ability to integrate and commit to college.

Social Support. Environmental factors have also been shown to enhance integration and persistence. Recently, Tinto (1993) incorporated social-support theory into his model by describing the positive effects of social support on adjustment to college. Tinto's (1993) acknowledgment is based on the work of House (1981), McCarthy, Pretty, and Catano (1990), Pearson (1990), and Jacobi (1991), who report that supportive relationship prevent and reduce the harmful effects of stress, and enhance individuals' ability to cope effectively with stress in specific social settings. Bean and Metzner (1985) cite five studies (Bucklin & Bucklin, 1970; Pantages & Creedon, 1978; Sexton, 1965, Spady, 1970, and Tinto, 1975) which reported positive relationship between parental support, in the form of encouragement, and persistence in college. MacMillian (1969) and Weigel (1969) found that two-year community college persisters experienced significantly greater encouragement from parents than dropouts.

Examining the influence of peer relationships on persistence decisions for a national sample of traditional age (18 to 24 years) full-time students, Anderson (1981) observed that peer discouragement to leave college was positively related to persistence. Similarly, Metzner (1984) reported that encouragement by friends to continue attending college was negatively related to intent to leave a public commuter college, for traditional age full-time students, traditional age part-time students, and older part-time students. Nora's (1987) study of two-year college students found encouragement from others significantly influenced the student's social integration and goal commitment.

According to House (1981), social support serves four important functions which involve (a) emotional support, (b) appraisal support, informational support, and (d) instrumental support. House (1981) proposed that emotional support (receiving expressions of affection, interest, and concern) increases self-esteem and perception of self-efficacy. Appraisal support involves help in understanding problems, and provides feedback about their significance. Informational support consists of advice on how to manage problems, discussing possible solutions and discussing alternative coping strategies. Instrumental support involves providing material aid and services. These resources can help the recipient cope with problems more effectively.

Just as obtaining social support from an individual can enhance persistence, so to can obtaining support from social networks. Hays and Oxley (1986) examined social support by assessing the development of social networks among college freshmen. They observed a positive impact for network density and frequency of network interactions on adapting to college. They also found that psychological disturbance, assessed with the Hopkins Symptom Checklist (HSCL; Derogatis et al., 1974), had a negative influence on establishing new acquaintances and on adapting to college. Based on these findings they concluded that social competence may mediate the extent of one's social networks (social integration). That characteristics of a person mediate the supportive behaviors of others has been well documented in the psychology coping literature (Wortman & Dunkel-Schetter, 1979; Dunkel-Schetter, Folkman, & Lazarus, 1987; Johnson, 1991; Mallinckrodt & Leong, 1992). In addition, there is evidence that the establishment of supportive personal relationships with faculty, peers, and other significant persons helps students cope better with the demands of college (Fleming, 1985; Ostrow, Paul, Dark, & Berhman, 1986). These supportive relationships, in turn, have a positive impact upon student academic success. In fact, having supportive campus and family relationships was positively related to persistence for both black and white students (Mallinckrodt, 1988).

Critique and Purpose. Even though factors such as social support may help students persist in college, the benefits may differ for two-year versus four-year college entrants. Pascarella and Terenzini (1991) conducted an extensive review of the factors associated with educational attainment among college students. They concluded that two-year community college entrants are less likely to persist than four-year college entrants. This conclusion remained intact even after accounting for a variety of relevant personal, aspirational, academic, SES, and family background characteristics. Prior to this conclusion, Clark (1960) noted that:

So great are differences [between community colleges and senior colleges], that despite their many positive functions [community colleges] can also serve to "cool out" high-aspirational but low-achieving and/or low-socioeconomic students, discouraging continued enrollment and reducing the likelihood that the two-year student would enjoy the educational benefits of the four-year college graduate. (Clark, 1960, p 569)More recently it has been suggested that community colleges do not really serve the interests of college-degree seeking students (Karable, 1972, 1986; Brint & Karable, 1989; Astin, 1977).

Indeed, there are high rates of departure among community college students. Paradoxically, enrollment in community colleges over the past three decades has increased far beyond those of private and public four-year colleges. Despite the evidence, however, there is a pervasive belief, among high school graduates, their parents, and policy makers, that attendance at a two-year community college followed by transfer to a four-year college is a viable approach to earning a four-year degree (Pascarella & Terenzini, 1991; Conklin, 1993). Such an approach represents a liability, particularly for minority and economically disadvantaged students (Astin, 1977).

There is a need to examine, identify, and better understand the factors which may lead to attrition among community college students. The present paper has, thus far, summarized the existing literature regarding its causes within the theoretical framework of Tinto's model (1975, 1987, 1993). Examining and refining Tinto's model (1975, 1987, 1993) within a community college population will help in understanding the role various factors may have on attrition.

The present study examined the impact of a comprehensive set of academic and psychosocial factors on the persistence/withdrawal decisions of two-year community college entrants. Employing Tinto's model the present study included an examination of traditional risk factors, as well as a comprehensive set of psychosocial measures to assess educational and institutional commitment, academic and social integration, social support, life events, external demands/involvements, and psychological well-being and functioning. Based on this expanded model of student departure (see Figure 1), this study examined cognitive and psychosocial measures to assess Tinto's (1993) person-environment fit hypothesis.

Major research hypotheses. (see Figure 1).

I. Initial goal and institutional commitment. Educational goal and institutional commitment were predicted to be positively associated with socioeconomic status, prior academic achievement, and quality of work effort.

II. Social and academic integration. Social and academic integration were predicted to be positively linked to initial institutional and goal commitment, social support, agreeableness, self-esteem, campus size, satisfaction with college, and psychological adjustment. Academic integration was also predicted to be positively associated with quality of work effort and first semester academic achievement. External commitments and involvements, and negative life events occurring in and out of school, were expected to negatively impact upon both academic and social integration.

III. End-of-term educational goal and institutional commitment. Initial goal and institutional commitment were predicted to be positively linked to goal and institutional commitment at the second administration. In addition, end-of-term goal and institutional commitment were predicted to be positively related to social integration, academic integration, social support, self-esteem, personal-psychological adjustment to college, and satisfaction with college. Negative and positive life events, negative events in school, and external commitments and involvements, were hypothesized to have an adverse impact on end-of-term goal and institutional commitment.

IV. Persistence. Persistence/withdrawal behavior was predicted to be positively linked to Time-2 goal and institutional commitment. That is, individuals with greater goal commitment and institutional commitment would be more likely to persist. In addition, persistence was predicted to be positively and directly associated with academic and social integration.


Subjects. The final sample for the study consisted of 1,011 first-time freshmen students enrolled in randomly selected Freshmen Seminar classes. Participants were drawn from a three-campus, college-wide population of approximately 3,300 first-time, full-time, day students. A retained student is one who re-enrolled in classes by the eighth week(1) of the subsequent Spring semester. A departure outcome was recorded when a student failed to enroll in classes by the eighth week of the semester.

Demographic characteristics of the sample are reported in Table 1. Comparisons of the demographics of the study participants and the entire Freshmen class, were conducted to assess how well the sample represented the entire cohort. Chi-square analyses indicated that the sample and Freshmen class were statistically similar in gender, ethnicity, and marital status (see Table 1). A comparison of age indicated that the sample included a higher proportion of students in the 18 to 19- year age category (2(4) = 110.8, p < .0001).

Of the 1,011 students who participated in the study 72.4-percent re-enrolled in the Spring semester. The term-to-term persistence was 71.6-percent for the entire freshmen class. The persistence rates for the sample and the entire freshmen class were, indeed, equivalent (Z = .497, p > .30).

Campus Size. Suffolk Community College consists of three campuses with a combined enrollment of 22,175 students (SCC, 1994). The central county campus , is the largest facility and has a total enrollment of 12,864 students. The western region campus has an enrollment of 6,650 students. The eastern campus is the smallest campus and has an enrollment of 2,661 students. The central campus falls in the upper tenth percentile of the enrollment distribution for New York State public colleges, and represents a large college campus (based on the Fall 1994 campus enrollment distribution of the sixty-six State University of New York colleges; SUNY, 1995). The western campus falls at approximately the sixtieth-percentile and represents a medium size campus. The eastern campus, falls in the lower third of the enrollment distribution, and is a small college campus.

Design. The design of the study was longitudinal with data collected over the 1994-1995 academic year at three time intervals. Time-1 data were collected during the first and second weeks of the Fall 1994 semester. The data consisted of demographic and previous academic performance/achievement data obtained from student data files maintained by the SCC Office of Institutional Research. Time-1 self-report measures were obtained from a questionnaire administered to students enrolled in SCC Freshman Seminar classes(2). Time-2 data consisted of a second set of self-report measures administered to the Freshmen Seminar students at the end of the Fall semester.

Time-1 and 2 questionnaires were administered during class. Instructors informed the students that the college was conducting a survey of freshman students and invited the students to volunteer to participate in the study. Of the 1,800 students enrolled in Freshmen Seminar classes, 1,565 completed the Time-1 forms. Time-2 questionnaires were completed by 1,326 students. Early attrition and class absences resulted in a final sample of 1,011 students with complete data. Time-3 data were compiled during the beginning of the Spring 1995 semester and consisted of first semester GPA and Spring 1995 re-enrollment status.


Pre-college factors. Consistent with Tinto's model (1975, 1987, 1993), Time-1 self-report scales included an assessment of a variety of pre-college factors, including socioeconomic status, initial goal and institutional commitment, external commitments, agreeableness, and conscientiousness. Time-2 scales were designed to assess semesterly school-based and external experiences and the attitudes and perceptions shaped by these experiences, including external involvement, positive and negative life events, social support, self-esteem, psychological well-being and emotional adjustment to college, overall satisfaction with college, social and academic integration, and goal and institutional commitment. Time-3 data consisted of persistence/withdrawal outcomes and first semester grade point averages.

Academic achievement (Time-1). Test scores from the College Entrance Examination Board's (CEEB) Computerized Placement tests (CPT; 1990) were used to assess academic achievement. The test is a nationally standardized measure of reading comprehension, writing, and math (computational and elementary algebra) skills. Reliability estimates (test-retest) for each test were obtained on a sample of 1,800 college students (CEEB/ETS, 1990), and ranged between .90 and .92. Content, construct, and predictive validity has been obtained (CEEB/ETS, 1990; CEEB/ETS, 1986; Weber, 1991; Nold & Kuechenmeister, 1991; Pomplun, 1991; Napoli & Wortman, 1995). High school average was also used as an index of academic achievement.

At SCC, performance on the CPT test is used to assess proficiencies in the three basic skill areas, and to identify students in need of developmental (remedial) math, writing, or reading comprehension classes. Level-1 placements represent skills at, or below, the eighth-grade level. Level-2 placements represent basic skills above the eighth-grade, but below the first-year college level. Completion of a level-1 class is followed by taking the corresponding level-2 remedial class. After level-2 developmental courses are completed, a student may enroll in college-level courses. In the present study, basic skills were expressed as the number of developmental courses a student was required to take in each developmental area (where 0 = no developmental courses, 1 = level-2 courses, and 2 = both level-1 and 2 courses).

External Commitments (Time-1) External commitments at the Time-1 (EC-1) were operationalized using a three-item scale. The items were (a) marital status (married v. single), (b) number of dependent children, and anticipated weekly hours worked for the upcoming semester.

External Involvements (Time-2). External involvement at Time-2 (EI-2) was measured with a two-item scale. The items were amount of time spent socializing with (a) relatives and (b) friends who do not attend Suffolk Community College.

Institutional and Goal Commitment (Time-1 and Time-2). Time-1 institutional commitment (IC) and Time-1 goal commitment (GC) were assessed with a five-item scale taken from the Student Involvement Questionnaire (SIQ; Pascarella & Terenzini, 1980). A three-item SIQ-IC scale assesses students' initial plans to continue their education at SCC. A two-item SIC-GC scale assesses the students' initial motivation to earn a college degree, in general.

Institutional commitment and goal commitment were again assessed at Time-2 with the SIQ-IC and SIC-GC scales. Time-2 institutional commitment was also assessed with the Institutional Attachment subscale of the Student Adaptation to College Questionnaire (SACQ-IA; Baker & Siryk, 1989). This 15-item measure, is standardized for first-semester college freshmen, was designed to assess the quality of the relationship, or bond, between the student and the institution, and the strength of the student's motivation to attend that institution. Coefficient alpha has ranged from .85 to .95 (Baker & Siryk, 1989). Construct validity, convergent validity, and predictive validity for the SACQ-IA scale have been established (Baker & Siryk, 1989).

Agreeableness and Conscientiousness (Time-1). Measures of social competence (agreeableness) and quality of work effort (conscientiousness) were obtained from the NEO Personality Inventory (Costa, McCrae, & Dye, 1991). The agreeableness scale assesses interpersonal orientation ranging from compassion to antagonism in thoughts, feelings, and actions. The conscientiousness scale assesses the degree of organization, persistence, and motivation in goal-directed behavior. Internal consistency of the scales ranges from .86 to .91, and the reliability ranges from .76 to .86 (Costa, McCrae, & Dye, 1991).

Academic Integration (Time-2). Academic integration (AI) was assessed at Time-2 with two scales. One, the nine-item SIQ-AI, assessed academic-related interactions the student had (Pascarella & Terenzini, 1980). The reliability ranges from .77 to .85 (Fox, 1986). The other AI scale used was the Academic Adjustment scale of the Student Adaptation to College Questionnaire (SACQ-AA). The SACQ-AA consists of 24 items that assess educational demand characteristics. Example items include (a) keeping up to date with academic work, (b) allocating sufficient time to study, attending class on a regular basis, and enjoying the (d) academic demands of college. The reliability ranges from .81 to .90 (Baker & Siryk, 1989). Construct validity, convergent validity, and predictive validity for the scale have been established (Baker & Siryk, 1989).

Social Integration (Time-2). Social integration (SI) was assessed at Time-2 with two scales. One measure, the Student Involvement Questionnaire Social Integration scale (SIQ-SI; Pascarella & Terenzini, 1980) is a 17-item activity index reflecting the number of non-academic interactions between the student and peers and instructors. The reliability ranges from .82 to .85 (Fox, 1986). The other social integration measure, the Social Adjustment subscale of the SACQ (SACQ-SA), is a 20-item standardized measure assessing the interpersonal-societal demands inherent in adjusting to college. Example items include (a) establishing friendships at school, (b) involvement in college-based social activities, informal contact with other students and professors, and (d) overall satisfaction with the social life at the college. The reliability has ranged from .83 to .90 (Baker & Siryk, 1989). Construct validity, convergent validity, and predictive validity have been established (Baker & Siryk, 1989).

Psychological Adjustment (Time-2). Psychological adjustment was assessed with the Personal-Emotional Adjustment subscale of the SACQ (SACQ-PA). The fifteen-item SACQ-PA scale assesses general psychological distress, specifically anxiety (state and trait) and depression. Lower scores indicate fewer psychological coping resources, conflictual dependency with parents, and poorer mental health or psychological well-being. Internal consistency ranges from .76 to .86 (Baker & Siryk, 1989). Construct validity, criterion-related validity, and known-groups validity have been established (Baker & Siryk, 1989).

Positive and Negative Life Events (Time-2). Positive and Negative life events, occurring within and outside of college during the Fall semester, were measured with the student version of the Life Experience Survey (LES; Sarason, Johnson, & Siegel, 1978). The LES is a 60-item self-report measure which assesses a wide range of life experiences and events. Subjects indicate which events, among a list, occurred over the semester, and rate the impact (positive or negative) of each event on a 7-point. The present study assessed positive life events, negative life events, positive events which occurred in school, and negative events which occurred in school. Test-retest reliability over an eight-week period ranged from .60 to .82 (Sarason, Johnson, & Siegel, 1978).

Self-Esteem (Time-2). Self-esteem was measured with the Self-Esteem scale (SE; Rosenberg, 1965). The 10-item scale measures self-acceptance, self-liking, and self-approving. Internal consistency reliability and test-retest reliability were estimated at .92 and .85, respectively (Rosenberg, 1965).

Social Support (Time-2). Support and encouragement to attend college was assessed with a 10-item self-report scale (Mallinckrodt, 1988). Seven items measured perceived support from the campus community and three items assessed support from the family. The predictive validity of the scale has been established (Mallinckrodt, 1988).

Satisfaction with College (Time-2). The Student Opinion Survey (SOS; American College Testing Program, 1994) is a 69-item, standardized questionnaire which assesses satisfaction with the college's administrative and student services, the academic and instructional environments, the physical environment, and the social environment of the institution. Test-retest reliability exceeds .90 (American College Testing Program, 1994). Content, construct, and predictive validity has also been established (Mittelholtz & Noble, 1993).

Persistence and Academic Outcomes (Time-3). Persistence/withdrawal behavior was based on Fall 1994 to Spring 1995 enrollment. Students who re-enrolled in the Spring semester were classified as persisters. Those students who failed to re-enroll for the Spring semester were classified as non-persisters. First semester (Fall) GPAs were obtained from the College's master student data files.

Overview of statistical analyses.

Structural equations modeling (SEM; Loehlin, 1992) and discriminant function analyses (Tabachnick & Fidell, 1996) were used to assess the adequacy of the expanded model (see Figure 1) in explaining persistence and early withdrawal from college. The SEM was conducted using a two-stage approach as described by Nora and Cabrera (1996). First, in a measurement stage, confirmatory factor analyses (CFA) were conducted to verify item assignments to scale, and to assess the fit of the measurement equations in defining the latent variables (Bentler, 1995). Scores for all measured variables and factors were expressed as standard scores. Second, in the causal modeling stage, the fit of the latent variables in the hypothesized expanded structural model (see Figure 1) was tested independently from the measurement models.

Since no hypothesized factor structure exists for the ACT Student Opinion Survey (SOS), an exploratory factor analysis was conducted using the SAS program (SAS, 1988). Based on an examination of the scree plot for the resulting eigenvalues it was determined that one factor sufficiently represented the common variance among the items. Therefore, in the present study, overall satisfaction with college was expressed with a single latent variable created from the 69 SOS items.

The goal of the discriminant function analysis was to assess how well the measures distinguished persisters and non-persisters. This was accomplished within the discriminant function procedure by applying classification coefficients to the scale scores, to produce predicted group membership (SPSS, 1988). Predicted group membership was then compared to actual membership to assess the proportion of students correctly classified by the measures.

Excluded measures.

Initial attempts to fit the expanded model (see Figure 1) using SEM produced two instances of collinearity and required the elimination of these measures and several paths. Specifically, a psychological well-being factor constructed from the Hopkins Symptom Checklist (Derogatis et al., 1974), which was administered at Time-2, was found to be highly correlated with the SACQ-PA factor, and was subsequently eliminated from the model. Therefore, in the present study psychological well-being was represented with the SACQ-PA scale (Baker & Siryk, 1989). Because the SACQ-PA specifically assesses psychological adjustment to college it was chosen over the HCSL.

In addition, a basic academic-skills factor derived from the CPT reading comprehension, writing, and mathematics tests, was highly correlated with the developmental course factor. This basic skills factor was dropped from the model. This collinearity was, however, expected since CPT tests scores are used to place students into developmental courses.


Results for the Confirmatory Factor Analyses.

For each psychosocial measure a confirmatory factor analysis was performed. The standardized path coefficients, or factor loadings, linking the items to their respective latent factors were assessed and only those items with significant (p < .05) factor loading were retained. The overall adequacy of each measurement equation was further assessed with the Chi-square "goodness-of-fit test" and the Comparative Fit Index. Results for all measurement equations show that Chi-square values for the latent variables were all non-significant, and the CFI statistics were all above .90. Collectively, the significant path coefficients, non-significant Chi-square tests, and the large CFIs confirm the factor structure of each latent variable(3).

Structural Equations Model.

The adequacy of the structural model (see Figure 1) was tested separately from the measurement models. The Chi-square value for the model was not significant (2(174) = 193.5, p > .10) and the Comparative Fit Index was high (CFI = .997). Thus, there was an excellent fit between the data (i.e., the correlations between measures; see Table 4), and the hypothesized causal model of student persistence/withdrawal behavior. All standardized path coefficients for the expanded model (to be discussed in detail below) were significant at the p < .05 level (see Table 2). Furthermore, the model accounted for 57 percent of the variance in student persistence/withdrawal behavior. Having satisfied the statistical criteria confirming the validity of the expanded model, the determinants of each of the major factors within the model were examined.

Initial Goal and Institutional Commitment. Initial educational goal commitment (GC-1) was positively and significantly(4) associated with socioeconomic status ( = .105), gender ( = .075), conscientiousness ( = .210), self-esteem ( = .075), and campus size ( = .081), and negatively and significantly associated with ethnicity ( = -.102), and developmental course requirements ( = -.175). Concerning the findings for gender, ethnicity, and developmental requirements, females reported significantly greater initial goal commitment than males, and non-minority (Caucasian) students showed significantly greater initial goal commitment than minority students (i.e., African American and Hispanic) students. Basic skills deficiencies had a significant negative impact on initial goal commitment. Collectively, the measures explained 12 percent of the variance in initial goal commitment.

Initial institutional commitment was significantly and positively associated with age ( = .060) and conscientiousness ( = .098). None of the other hypothesized paths were significant. In all, only 2 percent of the variance in initial institutional commitment was explained by the model. This finding is consistent with others who reported a small amount of explained variance in initial institutional commitment (Pascarella, Duby, & Iverson, 1983; Fox, 1986).

Social Support. Social support was found to be significantly influenced by a variety of background and psychosocial factors. As presented in Table 2, academic success in high school ( = .088), self-esteem ( = .377), agreeableness ( = .137), and negative life events ( = .187) were positively associated with social support. These findings indicate that students with greater high school academic success, social skills, a more positive self-image, and having more negative life events were more likely to obtain social support from others within the college environment. Age ( = -.060) and negative events in college ( = -.413) had a significant and negative impact on social support. That is, older students experience less support than younger students. Negative events occurring within college had a strong and adverse impact on social support, whereas negative life events were associated with greater social support. Overall, the structural equation model accounted for approximately 40 percent (R2 = .396) of the variance in social support.

Personal-Emotional adjustment. The SACQ-PA, used to assess general psychological well-being, was scaled so that high-scores reflect better adjustment (i.e., lower levels of anxiety and depression). Age ( = .058), agreeableness ( = .047), conscientiousness ( = .126), and self-esteem ( = .268) were all positively associated with SACQ-PA (see Table 2). The older, more agreeable, more conscientious, and higher self-esteem students had better adjustment. Gender (i.e., females; = -.076), negative life events ( = -.244), and negative events in school ( = -.072) were significantly associated with poorer psychological adjustment. Together, these measures explained approximately 27 percent of the variance in SACQ-PA.

Satisfaction with college. Overall satisfaction with college (SOS) was significantly associated with several factors contained in the model (see Table 2). Social support ( = .389), high school average ( = .058), and conscientiousness ( = .059) had positive impacts on satisfaction. Negative events in school ( = -.229) and negative life events ( = -.091) were associated with lower satisfaction with college. Together these measures explained 27 percent of the variance in overall satisfaction with college.

Social Integration. As postulated, social integration (SI), one of the key constructs in the model, was significantly and positively influenced by social support ( = .322), psychological well-being ( = .231), satisfaction with college ( = .139), socioeconomic status ( = .065), gender ( = .082), self-esteem ( = .095), campus size ( = .061), external involvement ( = .075), and negative life events ( = .122; see Table 2). Age was significantly and negatively associated with social integration ( = -.123). Also, as expected, negative events occurring in school produced significantly lower social integration ( = -.280).

Demographically, younger females with higher socioeconomic status, who were younger, and who were female were more socially integrated. As predicted, campus size was positively related to social integration with larger campuses having more integrated students. Contrary to expectations, negative life events and external involvements at Time-2 were positively and significantly related to integration. That is, those having more negative life events and external involvements were more socially integrated. Overall, 53 percent of the variance in SI was explained by the model.

Academic Integration. Academic integration (AI) was directly influenced by a variety of background, cognitive, and psychosocial factors which collectively accounted for 59 percent of its variance (see Table 2). The background factors, age ( = .087) and gender ( = .063), were significantly associated with AI. Thus, older students experienced greater AI than younger students and females evidenced greater academic integration than males.

Of the cognitive and psychosocial factors, goal commitment at Time-1 ( = .107), social support ( = .136), psychological well-being ( = .342), satisfaction with college ( = .148), first-semester grade point average ( = .197), academic success in high school ( = .058), and conscientiousness ( = .076) had a direct, positive, and significant impact on academic integration. Similar to the findings for social integration, negative school events produced significantly lower academic integration ( = -.127), whereas, negative life events were associated with greater academic integration ( = .061).

From these results, the more academically integrated students had higher academic achievement and initial goal commitment. Psychologically, academic integration was higher among students who were more conscientious, who had greater self-esteem, and who were more psychologically adjusted. Experientially, academic integration was greater for individuals with greater social support, who were more satisfied with college, and who had fewer negative events in school. Contrary to initial expectations, academic integration was positively associated with negative (non-school) life events. Thus, individuals with greater negative life events experienced greater academic integration.

First-semester Grade Point Average (GPA). First-semester GPA was significantly and positively associated with social support ( = .093), psychological well-being ( = .084), age ( = .172), gender ( = .121), and high school average ( = .147). Negative school events had a significant, and large adverse impact on GPA ( = -.475), as might be expected. Negative life events, however, were associated with higher GPAs ( = .123). This unexpected finding is consistent with the influence negative life events had on social and academic integration. Combined, the measures explained 34 percent of the variability in first-semester GPA.

Based on these results, older individuals with greater social support, psychological adjustment, academic success in high school, and with fewer negative school events, obtained higher first-semester GPAs. In addition, females earned significantly greater GPAs than males.

End-of-term Goal Commitment (GC-2). Consistent with Tinto's (1993) model, goal commitment at Time-2 was significantly and directly linked to both academic integration ( = .506) and social integration ( = .122), with the former having the greatest impact of all the variables on GC-2 (see Table 2). Significant positive and direct influences were also observed for initial goal commitment ( = .101), conscientiousness ( = .087), social support ( = .071), and self esteem ( = .048). Significant negative associations were found for ethnicity ( = -.059) and external involvements at Time-2 ( = -.065).

Like the findings for initial goal commitment, minority students showed lower goal commitment. Unexpectedly, external involvement was positively related to goal commitment. Overall, 52 percent of the variance in GC-2 was explained by the model.

End-of-term Institutional Commitment (IC-2). Institutional commitment, just prior to the beginning of the Spring semester, was significantly and directly related to initial institutional commitment ( = .163; see Table 2), social support ( = .098), psychological well-being ( = .158), satisfaction with college ( = .084), social integration ( = .197), academic integration ( = .381), age ( = .066), and negative life events ( = .104). Negative school events had an adverse impact on IC-2 ( = -.184). Replicating the result for initial institutional commitment, older students had greater institutional commitment at Time-2. Overall, the model accounted for 71 percent of the variance in IC-2.

Persistence. Term-to-term persistence was significantly and directly linked to institutional commitment at Time-1 ( = .057), social integration ( = .197), academic integration ( = .064), first-semester GPA ( = .268), goal commitment at Time-2 ( = .068), and negative life events ( = .123; see Table 2). External commitments at Time-1 ( = -.100) and Time-2 ( = -.070) and negative school events ( = -.450) each had a significant and detrimental impact on persistence.

Discriminant Function Analysis.

To recapitulate, the goal of the discriminant function analysis was to assess how well the set of background and psychosocial measures distinguish between the persisters and non-persisters. The analysis included those measures with at least one significant path in the expanded model (see Table 2). The single canonical function consisted of twenty-four retained measures which significantly discriminated between persisters and non-persisters (2(24) = 835.9, p < .0001).

To assess how well the measures discriminated persisters from non-persisters, classification function coefficients were computed and applied to the discriminator variable scores to produce predicted group membership (SPSSX, 1992). The background and psychosocial measures correctly classified 76.6% of those students who withdrew and 93.9% of the term-to-term persisters (see Table 3). Overall, 89.1% of the cases were correctly classified.Discussion

The goal of the present study, was to assess the validity of Tinto's (1975, 1987, 1993) student departure model in explaining persistence/withdrawal behavior of community college students employing a prospective design. Using structural equations modeling, the study assessed the role and relative importance of each of Tinto's constructs (i.e., academic and social integration, institutional and goal commitment), as well as additional psychological, psychosocial, and institutional factors, which were hypothesized to mediate the relationships among the constructs. The organization of subsequent discussion reflects the temporal pattern of causality represented by Tinto's model (1993). In this manner, the impact of each construct on the subsequent "down stream" construct, and the associations between the constructs, will be presented in their hypothesized causal order (see Figure 1).

Early goal commitment and institution commitment. Evidence presented in the present study, and from previous work (Pascarella et al., 1983), clearly links demographic characteristics and prior academic achievement with early goal commitment. In the present study, non-minority students with higher socioeconomic status and better academic preparedness showed higher initial goal commitment. As was hypothesized, both conscientiousness and self-esteem had direct and positive influences on commitment. In fact, conscientiousness, or the degree of organization, persistence, and motivation in goal-directed behavior (Costa, McCrea, & Dye, 1991), emerged as the strongest predictor of initial goal commitment. This finding replicates past research that found a direct link between achievement need and goal commitment (Pascarella & Chapman, 1983a, b).

The finding for self-esteem showed that, even after controlling for all other factors, it has a significant and positive impact on initial goal commitment. Thus, the initial goal of earning a college degree is not only related to demographic factors, but also to one's overall willingness to commit to goals in general and to having a positive self-image.

Older students and females had a greater initial institutional commitment. Post hoc analyses showed that older students have more domestic demands (i.e., dependent children and greater financial demands) so they may be unable to attend an out-of-area institution or more expensive local institution. Thus, producing greater institutional commitment to the less expensive, local community college.

Social and academic integration. First, SI and AI exert both direct and indirect effects on persistence through goal commitment and institutional commitment (see Figure 1). An examination of the determinants of these constructs was conducted to elucidate possible causes. It was predicted that social and academic integration would be positively and directly linked to initial institutional and goal commitment, social support, agreeableness, self-esteem, campus size, overall satisfaction with college, and psychological well-being. Further, it was anticipated that external commitments and involvements, and negative life events occurring in and out of school, would have direct negative effects on social and academic integration. Results from the SEM show these factors significantly, but differentially, influenced SI and AI.

For social integration, social support was found to be particularly important. Indeed, it had the largest impact on SI (see Table 2). Based on previous work, agreeableness was also expected to have a direct and positive effect on social integration (Pascarella & Chapman, 1983a, b). Nevertheless, this was not the case. Although the simple correlation between agreeableness and SI was significant (r = .146, p < .0001), after controlling for other variables in the model, the path coefficient linking agreeableness and SI was not significant (p > .10). This result suggests a third variable may be responsible for the correlation between agreeableness and SI. Previous investigators have proposed that "social competence" may influence a student's ability to use network members as a source of support (Hays & Oxley, 1986). Under this assumption it would be expected that agreeableness would have a significant and direct impact upon social support. In a post hoc analysis this hypothesis was, indeed, confirmed. Results show that agreeableness has a significant, positive, and direct effect on social support, and an indirect effect on social integration through social support.

From their study of social network development, Hays and Oxley (1986) found psychological well-being to be positively related to social network density and negatively related to network conflict, among college students. Based on these findings, it was expected that social integration would be higher among students with greater psychological well-being. As observed in this study, psychological well-being has a significant and positive impact on social integration. Additionally, self-esteem was directly related to social integration. So, students who are relatively free from anxiety and depression and who have a greater positive self-image are more adept in forming social relationships in college.

In the second hypothesis it was also predicted that negative events occurring in school would have an adverse influence on social integration. This hypothesis was supported. Negative school events, such as conflicts or problems within the social, academic, and administrative systems of the college inhibited social integration, and ultimately persistence.

Based on Tinto's (1993) model, it was predicted that negative life events occurring outside of school would interfere with social integration and other antecedents of persistence (i.e., academic integration, grade point average, goal commitment, and institutional commitment). However, this hypothesis was not supported. Rather, positive relationships consistently emerged between these constructs and negative life events. That is, each increased as levels of negative life events increased. Although these findings were unexpected, the consistency of the effect across the measures suggests that they are not spurious. Explaining these findings requires a reconceptualization of the direction of causality originally proposed. Recall, the a priori hypothesis was that negative life events outside of college would adversely affect social involvements and academic achievement in college. However, it is also possible and more consistent with the data, that social and academic involvement in college adversely affects external, non-school based experiences and relationships. For example, students may experience greater life stress and strain as a consequence of trying to meet the demands of college.

Campus size. Tinto (1993) proposed that larger institutions would have a greater variety of social and intellectual communities, thereby providing greater opportunities for social integration. Results from the SEM support this proposition. Specifically, campus size, which varied considerably in the present study, was significantly and directly linked to social integration. Further, post hoc comparisons showed that, when controlling for all other factors, students from the larger campus have significantly higher SI than the smaller campuses. Moreover, campus size had a significant indirect impact on persistence through social integration.

Academic integration (AI). The analyses of the factors which influence academic integration provide additional support for Tinto's model (1975, 1987, 1993). In addition, these results replicate earlier findings that age, SES, and previous academic achievement influence academic integration (Pascarella & Chapman, 1983a, b). In the present study conscientiousness also had a positive impact on AI.

Two additional psychological measures, self-esteem and psychological well-being, were found to have direct and positive effects on academic integration. These findings, which are similar to those obtained for social integration, indicate that students who possess a positive self-image and who are relatively free from anxiety and depression experience greater adjustment to the academic demands of college. These results support earlier research that found self-esteem to have a direct influence on AI (Munro, 1981), and that students who are relatively free of psychological disturbance report greater adjustment to college (Hays & Oxley, 1986).

The proposition that satisfying and rewarding encounters with the academic and social systems of the institution would lead to greater integration was also tested (Pascarella & Terenzini, 1991). To test this hypothesis overall student satisfaction with college and negative events occurring within college were assessed. Indeed, both social and academic integration were positively and directly influenced by overall satisfaction with college, and adversely effected by negative college events.

Social support has many implications for academic integration. Tinto (1993) proposed beneficial effects of social support on adjustment to college. The data show that social integration is enhanced by social support. Further, the results also show social support is directly and positively associated with academic integration. Applying House's (1981) view of the adaptive benefits of social support, it is likely that the supportive encounters between students had a positive impact upon social and academic integration by (a) promoting effective emotional behaviors and self-efficacy, (b) assisting in cognitive appraisals and clarifying problems, (c) providing specific information to solve and manage problems, and (d) providing needed instrumental support and services. Further study on social support needs to be conducted, however, to fully understand precisely how it may lead to greater social and academic integration.

An analysis of the influence of social support on grade point average was conducted. Previous work found social support to moderate the performance-reducing effects of test anxiety in undergraduate students (Sarason, 1981). In the present study, social support had a positive and direct impact on GPA. Negative events occurring in school had an adverse impact on GPA. To further assess the stress buffering effects of social support, post hoc regression analysis was computed in which GPA was regressed on social support, negative school events, and their interaction. The main effects for social support and negative school events were significant, and produce standardized regression weights nearly identical to those obtained in the SEM (see Table 2). The interaction was also significant (F(1,1007)= 56.7, p < .0001) and indicates a stress buffering effect for social support. To illustrate this effect we computed the mean GPA for high-stress students (negative events scores exceeding one standard deviation above the mean) with low social support (social support below one standard deviation from the mean) and high-stress students with high social support (social support scores one standard deviation above the mean). On average, the high-stress low-social support students earn a GPA of 1.5. In comparison, the high-stress high-social support students earn a substantially higher mean GPA of 2.63. In this regard, social support from family and classmates moderates, or reduces, the adverse effects of negative school events on GPA.

End-of-term Goal Commitment (GC-2). Tinto (1975, 1987, 1993) asserted that goal commitment is one of the most important determinants of persistence. He stated that "Once the individual's ability is taken into account, it is the student's commitment to the goal of completing college that is most influential in determining college persistence" (Tinto, 1975, p. 102). In their review of the literature, Cope and Hannah (1975) similarly concluded that personal commitment to an educational goal is the single most important determinant of persistence in college. Evidence from the present study and from the others (Allen & Nora, 1995; Axelson & Torres, 1995; Bers & Smith, 1990; Munro, 1981; Pascarella & Chapman, 1983a; Terenzini, Lorang, & Pascarella, 1981) shows that goal commitment does have an important influence on persistence.

So, the direct contribution of a variety of psychosocial factors on goal commitment were examined. Results indicate that initial goal commitment, social support, academic integration, social integration, conscientiousness, and self-esteem are the principal determinants of goal commitment. These factors represent a complex set of individual attributes and person-environment interactions. Collectively, they indicate that persisters have a positive self-image, approach goals in a conscientious and organized manner, and have the ability to interact with members of the college community.

In this study academic integration exerted the strongest influence on goal commitment. In fact, academic integration explained 25 percent of the variance in goal commitment, when controlling for all other factors. Thus, a student's interaction with his or her classmates and instructors involving academic issues serves as an important function in promoting and reinforcing one's commitment to earning a college degree.

End-of-term Institutional Commitment (IC-2). Among the demographic factors studied, only age showed a significant and positive effect on institutional commitment at Time-2 (IC). This finding was consistent over repeated administrations. Again, older students may show greater institutional commitment due to their limited ability to attend other institutions.

Tinto (1975, 1987, 1993) proposed that factors shaping goal commitment also influence institutional commitment. Results from the SEM generally support this view. Specifically, as in the case of goal commitment, academic integration, social integration, and social support exerted positive influences on IC-2. The findings for overall satisfaction with college, and negative events in college, deviate from this pattern, however. Both measures failed to influence significantly goal commitment, but were significantly linked to institutional commitment. This finding is reasonable as a student may experience overall dissatisfaction and negative events within a specific institution and still maintain a high or constant level of goal commitment. In fact, these developments may be the primary cause for transferring to another institution prior to graduation. It is unlikely that pregraduation transfers were common among the non-returning students, however. Post hoc analysis revealed that 68 percent of these students obtained a GPA below the C level. Thus, making it unlikely they transferred to another institution. This indirect GPA evidence, although compelling, cannot substitute for direct evidence of transfer. Therefore, future efforts should attempt to contact non-returning students using reliably data collection strategies, such as telephone interviews, to more accurately estimate pre-graduation transfer rates.

Persistence. Tinto (1975, 1987, 1993) suggested that academic and social integration may not equally affect persistence. Although social integration influences persistence, academic integration (i.e., attending class, studying, completing assignments, and maintaining a minimum GPA) is a mandatory condition for it. From the meta-analysis recently reported by Napoli and Wortman (1996) there were moderate to large effects for both SI and AI on persistence.

With regard to the generalizability of Tinto's model (1975, 1987, 1993) to community colleges, results from the SEM show that the proposed constructs influence college persistence among two-year community college students. In addition to observing positive and direct effects for social integration and academic integration (and relatedly GPA), institutional commitment and goal commitment positively impacted on persistence/withdrawal decisions.

Negative events occurring in school had strong and consistently adverse affects on persistence and the antecedents of persistence. In fact, of all of the paths tested, negative school events had the largest direct influence on persistence. This finding supports Tinto's (1975, 1987, 1993) and Pascarella and Terenzini's (1991) assertions, that unpleasant interactions occurring within the academic, administrative, and social systems of the college inhibit integration and decrease the likelihood of persistence.

Moreover, Tinto (1993) and others (Chacon et al., 1983; Weidman, 1985; Schwartz, 1990) proposed that departure from community college is shaped by external forces. Indeed, in the present study, external demands (i.e., family, work) had a significant and negative impact on persistence. Other investigators have also reported detrimental effects of external demands on community college persistence. Specifically, Bers and Smith (1991), and Axelson and Torres (1995) found non-persisters had greater work demands (e.g., the number of hours/week in off-campus employment) than persisters. Mulligan and Hennessy (1990) reported greater external demands (outside employment, family responsibility, and financial demands) among non-persisting community college students.

Pascarella and Terenzini (1991), in their extensive review of the literature, concluded that two-year community college entrants are less likely to persist than four-year college entrants. This relationship occurs even after holding constant a variety of relevant personal, aspirational, academic, SES, and family background characteristics (Pascarella & Terenzini, 1991). It is likely that the greater departure rates among community college students are related to problems associated with meeting demands from multiple communities. In this regard, community college students are not only faced with problems of adjusting to the demands of college, but also adjusting to the demands of external communities (i.e., family, friends, and work). In attempting to cope with these added demands, community college students are more likely to experience greater strain, leading to a reduced ability to participate and persist in college. By contrast, freshmen college students who attend residential institutions are more likely to be isolated from the day-to-day demands of family, friends, and work. These students have fewer distractions and greater opportunities to focus, almost exclusively, on within college issues increasing their likelihood to persist.

In sum, results of this study contribute to understanding the factors which influence student decisions to persist or withdraw from two-year community colleges. The model developed and tested accounted for a large part of the variability in persistence/withdrawal behavior. The present study's improvement over these past studies was the addition of psychosocial measures of conscientiousness, agreeableness, psychological well-being, self-esteem, social support, student satisfaction ratings, negative life events, and negative school events. These additional factors, alone, explained a significant (Finc.(8,980) = 35.6, p < .001), and sizeable portion of the variance (R2 = .21) in persistence/withdrawal behavior. Results from the discriminant function analysis confirm the utility of the overall model. In fact, 89 percent of the actual persistence/withdrawal outcomes were correctly identified by the model.

The study has several limitations. As with all correlational studies, causality is inferred, but cannot be demonstrated. In addition, selection and mortality bias (Cook & Campbell, 1979) may have influenced the results. In this prospective study, participation was voluntary, and although the final sample was demographically similar to the entire freshmen population, it is possible that the sample differed from the population in its intent to persist or graduate, resulting in selection bias. It is equally, possible that mortality bias could have occurred as a result of very early withdrawals (withdrawal during the first few weeks of the semester).

The generalizability of the findings to other institutions is also questionable. Unlike the casual path models developed by Munro (1981) and Pascarella and Chapman (1983a, 1983b) which employed multi-institutional samples, the present study was conducted using a single suburban community college sample. Therefore, the generalizability of the present findings to urban or rural populations is unclear. An additional problem concerns the demand characteristics of the study (Orne, 1969). Specifically, participants were obtained exclusively from Freshmen Seminar classes. The purpose of this course is to present methods and techniques which students can adopt to promote their perseverance and success at the college. Therefore, participants may have adjusted their responses to the questionnaire items in an attempt to satisfy these course objectives.


Allen, D. and Nora, A. (1995). An empirical examination of the construct validity of goal commitment in the persistence process. Research in Higher Education, (36), 5 509-533.

American College Testing Program. (1994). The ACT evaluation/service for educational institutions and agencies: Users Guide (6th ed.). Iowa City, Iowa, ACT. Author.

Anderson, K.L. (1981) Post-high school experiences and college attrition. Sociology of Education, (54), 1-15.

Astin, A. (1977). Four critical years: Effects of college on beliefs, attitudes, and knowledge. San Francisco: Jossey-Bass.

Axelson, R., and Torres, D. (1995). Modeling first-semester integration and departure of community college students. Paper presented at the Annual Forum of the Association for Institutional Research, Boston.

Baker, R.W., and Siryk, B. (1984). Measuring adjustment to college. Journal of Counseling Psychology, 31, 179-189.

Baker, R.W., and Siryk, B. (1989). SACQ student adaptation to college questionnaire manual. Los Angeles, CA, Western Psychological Services.

Bean, J.P., and Metzner, B.S. (1985). A conceptual model of nontraditional undergraduate student attrition. Review of Educational Research, 55(4), 485-540.

Bentler, P.M. (1995). EQS Structural Equations Program Manual. Encino, CA: Multivariate Statistics Inc.

Berkove, G. (1976) Returning women students; A study of stress and success. Paper presented at the annual convention of the Western Social Science Association, Tempe, AZ (ERIC Document Reproduction Service No. ED 156 971).

Brainard, S.R. (1973) Student Attrition: The student personnel climate. In R.L. Alfred (Ed.), Student attrition: Strategies for action (pp. 72-79). Kansas City MO. Metropolitan Junior College District. (ERIC Document Reproduction Service No. ED 085 064).

Bers, T.H., and Smith, K.E. (1991). Persistence of community college students: The influence of student intent and academic and social integration. Research in Higher Education 32(5), 539-556.

Brint, S., and Karable, J. (1989). The diverted dream: Community colleges and the promise of educational opportunity in America, 1900-1985. New York: Oxford University Press.

Bucklin, R.W. and Bucklin M.L. (1970). The psychological characteristics of college persisters and leavers: A Review. Washington, DC: U.S. Department of Health, Education, and Welfare. (ERIC Document Reproduction Service No. ED 082 751).

Campbell, D.T., and Stanley, J.C. (1966). Experimental and quasi-experimental designs for research. Chicago: Rand McNally.

Carter, B.L. (1982). Exit interview summary, fall 1981. Indianapolis, IN: Indiana University-Purdue University, Office of Student Services.

Chacon, M.A., Cohen, E.G., and Strover, S. (1983). Chicanas and Chicanos: Barriers to progress in higher education. Paper presented for the Conference on the Latino College Student. Princeton: Educational Testing Service.

Chapman, D.W., and Pascarella, E.T. (1983). Predictors of academic and social integration of college students. Research in Higher Education, 19(3), 295-322.

Chickering, A.W, and Reisser, L (1993). Education and Identity (2nd ed.). San Francisco: Jossey-Bass

Clark, B. (1960). The "cooling-out" function in higher education. American Journal of Sociology, 64, 569-76.

College Entrance Examination Board. (1990). Coordinator's guide for computerized placement tests, 3. Princeton, NJ: Author.

Conklin, K. A. (1993). Constituent perceptions of a community college: An "image" study. Journal of Applied Research in the Community College, 1(1), 49-58.

Cook, T,. and Campbell, D. (1979). Quasi-experimentation: Design and analysis issues for field settings. Chicago: Rand McNally.

Cope, R., and Hannah, W. (1975). Revolving college doors. New York, John Wiley and Sons.

Costa, P.T., McCrae, R.R., and Dye, D.A. (1991). Facet scales for agreeableness and conscientiousness: A revision of the NEO personality inventory. Personality and individual differences, 12, 887-898.

Derogatis, L.R., Lipman, R.S., Rickels, K., Uhlenhuth, E.H., and Lino, C. (1974). The Hopkins Symptom Checklist: A self-report symptom inventory. Behavioral Science, 19, 1-15.

Dunkel-Schetter, C., Folkman, S., and Lazarus, R. (1987). Correlates of social support receipt. Journal of Personality and Social Psychology, 53, 71-80.

Durkheim, E. (1951). Suicide. Translated by J. A. Spaulding and G. Simpson. Glencoe: The Free Press. Originally published as Le suicide: Etude de sociologie. Paris: Felix Alcan, 1897.

Fleming, J. (1985) Blacks in college. San Francisco, Jossey-Bass.

Fox, R. (1986). Application of a conceptual model of college withdrawal to disadvantaged students. American Educational Research Journal, 23, 415-424.

Hays, R.B., and Oxley, D. (1986). Social network development and functioning during a life transition. Journal of Personality and Social Psychology, 50, 305-313.

Holmes, T.H., and Rahe, R.H. (1967). The social readjustment rating scale. Journal of Psychosomatic Research, 11, 213-218.

House, J. S. (1981). Work Stress and social support. New York: Random House.

Hunter, R., and Sheldon, M.S. (1980). Statewide longitudinal study, part 3. Sacramento: California Community Colleges, Office of the Chancellor. (ERIC Document Reproduction Service No. ED 188 714).

Johnson, T.P. (1991). Mental health, social relations, and social selection: A longitudinal analysis. Journal of Health and Social Behavior, 32, 408-423.

Karable, J. (1972) Community colleges and social stratification. Harvard Educational Review, 42, 521-562.

Lazarus, R.S, Cohen, J.B., Folkman, S., Kanner, S., and Schaefer, C (1980). Psychological stress and adaptation: Some unresolved issues. In H. Selye (Ed.), Selye's Guide to Stress Research. New York: Van Nostrand, Reinhold.

Loehlin, J.C. (1992). Latent variable models: An introduction to factor, path, and structural analysis (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.

Martin, F.H. (`1974) Student follow-up 1: Attrition study. Harriman, TN: Roan State Community College. (ERIC Document Reproduction Service No. ED 099 087).

MacMillian, T.F. (1969). NORCAL project phase 1, final report. Los Angeles: University of California. (ERIC Document Reproduction Service No ED 031 240).

Mallinckrodt, B. (1988). Student retention, social support, and dropout intentions: A comparison of Black and white students. Journal of College Student Development, 29, 60-64.

Mallinckrodt, B. and Loeng, F.T.L. (1992). Social support in academic programs and family environments: Sex differences and role conflicts for graduate students Journal of Counseling and Development, 70, 716-723.

McClelland, K., Mueser, P., Olneck, M., Schwartz, J., Ward, S., and Williams, J. (1979). Who gets ahead in education? The determinants of economic success in America. New York: Basic Books.

Metzner, B.S. (1984) An application and evaluation of a model of student attrition using freshmen at a public urban commuter university. Dissertation Abstracts International, (44), 2387A. (University Microfilms No 83- 28,8080).

Mittelholtz, D.J., and Noble, J.P. (1993). The validity of evaluation/survey service instruments for reflecting institutional change. ACT research report series. Iowa City, Iowa: American College Testing Program.

Mulligan, S.C., and Hennessy, J.J., (1990). Persistence in a community college: Testing attrition models. Paper presented at the annual meeting of the American Educational Research Association, Boston.

Munro, B. (1981). Dropouts from higher education: Path analysis of a national sample. American Educational Research Journal, 81, 133-41.

Napoli, A., and Wortman, P.M. (1995). Validating college-level reading placement test standards. Journal of Applied Research in the Community College, 2(2), 143-151.

Napoli A., and Wortman, P. (In press). A meta-analytic examination of the relative importance of academic and social integration among community college students. Journal of Applied Research in the Community College.

Noel and Levitz, R. (Eds.) Increasing student retention. San Francisco: Jossey-Bass.

Nold, D. and Kuechenmeister, M. (1991). Aims Community College computerized placement tests pilot study: Summary Report. Computerized Placement Tests: Background Readings. New York: College Entrance

Nora, A. (1987). Determinants of retention among Chicano college students: a structural model. Research in Higher Education. 26 (1), 31-59.

Nora, A., Attinasi, L.C., and Matonak, A. (1990). Testing qualitative indicators of precollege factors in Tinto's Attrition Model; A community college student population. Review of Higher Education,(13), 3, 337-356.

Nora, A., and Cabrera, A. (1993). The construct validity of institutional commitment: A Confirmatory factor analysis. Research in Higher Education, 34 (2), 243-263.

Nora, A., and Cabrera, A. (1996). The role of perceptions of prejudice and discrimination on the adjustment of minority students to college. Journal of Higher Education, 67 (2), 119-148.

Orne, M. (1969). Demand characteristics and the concept of quasi-control. In R. Rosenthal and R.L., Rosnow (Eds.). Artifact in behavioral research, pp 143-179.

Ostrow, E.S., Paul, S., Dark, V., and Berhman, J. (1986). Adjustment of women on campus. Effects of stressful life events, social support, and personal competencies. In S.E. Hobfoll (Ed.) Stress, Social Support, and Women. Washington, D.C.: Hemisphere.

Pace, C. (1980). Measuring the quality of student effort. Current Issues in Higher Education, 2, 10-16.

Pace, C. (1984). Measuring the quality of college student experiences. Los Angles: University of California, Higher Education Research Institute.

Pantages, T.J., and Creedon, C.F. (1978). Studies of college attrition: 1950-1975. Review of Educational Research, (48), 49-101.

Pascarella, E.T. (1980). Student-faculty informal contact and college outcomes. Review of Educational Research 50, (4) 545-595.

Pascarella, E.T., and Chapman, D. (1983a). A multi-institutional, path analytic validation of Tinto's model of college withdrawal. American Educational Research Journal 20(1), 87-102.

Pascarella, E.T., and Chapman, D. (1983b). Validation of a theoretical model of college withdrawal: Interactive effects in a multi-institutional sample. Research in Higher Education 19(1), 25-48.

Pascarella, E.T., and Terenzini, P.T. (1980). Predicting freshmen persistence and voluntary dropout decisions from a theoretical model. Journal of Higher Education, 51, 60-75.

Pascarella, E.T., and Terenzini, P.T. (1983). Predicting voluntary freshmen year persistence/withdrawal behavior in a residential university: A path analytic validation of Tinto's model. Journal of Educational Psychology, 75, 215-216.

Pascarella. E.T., Duby, P.B., and Iverson, B. (1983). A test and re-conceptualization of a theoretical model of college withdrawal in a commuter institution setting. Sociology of Education, 56, 88-100.

Pascarella, E.T., and Wolfle, L.M. (1985). Persistence in higher education: A nine-year test of a theoretical model. Paper presented at the annual meeting of the American Educational Research Association, Chicago.

Pascarella, E.T., Smart, J.C., and Ethington, C.A. (1986). Long-term persistence of two-year college students. Research in Higher Education, 24(1) 47-71.

Pascarella, E.T., and Terenzini, P.T. (1991). How college affects students: Findings and insights from twenty years of research. San Francisco: Jossey-Bass.

Pomplun, M. (1991). Correlations between CPTs scores and course grades. Computerized Placement Tests: Background Readings. New York: College Entrance Examination Board.

Romano, R.M. (1995). First-year attrition and retention at a community college. Journal of Applied Research in the Community College, 2(2), 169-177.

Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton University Press.

Sarason, I.G., Johnson, J.H., and Siegel, J.M. (1978). Assessing the impact of life changes: Development of the life experience survey. Journal of Consulting and Clinical Psychology, 46(5), 932-946.

Sarason, I.G. (1981). Test anxiety, stress and social support. Journal of Personality. 49, 101-114.

SAS Institute Inc. (1988). SAS/STAT Users Guide, Release 6.03 Edition. Cary, NC:SAS Institute.

Schwartz, S. (1990). Application of a conceptual model of college withdrawal of technical college students. Paper presented at the annual meeting of the American Educational Research Association, Boston.

Sexton, V.S. (1965). Factors contributing to attrition in college populations: Twenty-five years of research. Journal of General Psychology, (72) 301-326.

Spady, W.G. (1970). Dropouts from higher education: An interdisciplinary review and synthesis. Interchange, 1, 64-85.

SPSS, Inc. (1988). SPSS-X Users's Guide (3rd ed.). Chicago: SPSS.

Stern, G. (1970). People in context: Measuring person-environment congruence in education and industry. New York: John Wiley and Sons.

State University of New York (1995). Official Credit Course Enrollment, Fall 1994, (Statistical Release No. 299). Central Office of Institutional Research, State University of New York.

Tabachnick, B.G., and Fidell, L.S. (1986). Using Multivariate Statistics. (3rd ed.) New York: Harper and Row.

Terenzini, P., Lorang, W., and Pascarella, E. (1981). Predicting freshmen persistence and voluntary dropout decisions: A replication. Research in Higher Education, 15, 109-127.

Tinto, V. (1975). Dropout from higher education: A theoretical synthesis of recent research. Review of Educational Research 45, 89-125.

Tinto, V. (1987). Leaving college: rethinking the causes and cures of student attrition. Chicago: University of Chicago Press.

Tinto, V. (1993). Leaving college: rethinking the causes and cures of student attrition (2nd ed.). Chicago: University of Chicago Press.

Van Gennep, A. (1960). The rites of passage. Translated by M. Vizedone and G. Caffee. Chicago: University of Chicago Press. Originally published as Les rites de passage. Paris: Nourry, 1909.

Webb, M.W. (1988). Freshman year retention at three campuses of a large urban community college district: 1983-1986. Community and Junior College Quarterly of Research and Practice, 12, 213-242

Weber, C. (1991). Statistical analysis of the class of 1994 at the USMMA. Computerized Placement Tests: Background Readings. New York: College Entrance Examination Board.

Weidman, J. (1985). Retention of non-traditional students in postsecondary education. Paper presented at the annual meeting of the American Educational Research Association, Chicago.

Weigel , M. (1969). A comparison of persisters and non-persisters in a junior college. Coon Rapids, MN: Anoka-Ramsey State Junior College. (ERIC Document Reproduction Service No. ED 044 115).

Wilson, R., Gaff, J., Dienst, R., Wood, L., and Bavry, J. (1975). College professors and their impact on students. New York: Wiley-Interscience.

Wortman, C. B., and Dunkel-Schetter, C. (1979). Interpersonal relationships and cancer: A theoretical analysis. Journal of Social Issues, 35, 120-155.

Table 1. Demographic comparison of sample and entering cohort.

  Sample (N = 1,011) Fall Cohort (N = 3,438)














Age Category, p. < .0001




35 and above









Marital Status







Table 2. Standardize path coefficients for the expanded model.

Goal Commitment - 1

.105 SES

.075 Gender

-.102 Ethnic

-.175 Developmental Placements

.210 Conscientiousness

.075 Self Esteem

.081 Campus Size

.939 Error

Institutional Commitment - 1

.060 Age

.098 Conscientiousness

.992 Error

Social Support

-.060 Age

.088 High School Average

.137 Agreeableness

.377 Self Esteem

-.413 Negative School Events

.187 Negative Life Events

.777 Error

Psychological Adjustment

.058 Age

-.076 Gender

.047 Agreeableness

.126 Conscientiousness

.268 Self Esteem

-.072 Negative School Events

-.244 Negative Life Events

.856 Error

College Satisfaction

.389 Social Support

.058 High School Average

.059 Conscientiousness

-.229 Negative School Events

-.091 Negative Life Events

.854 Error

Social Integration

.322 Social Support

.231 Psychological Adjustment

.139 College Satisfaction

.065 SES

-.123 Age

.082 Gender

.095 Self Esteem

.061 Campus Size

-.280 Negative School Events

.075 External Involvement - 2

.122 Negative Life Events

.688 Error

Academic Integration

.107 Goal Commitment - 1

.136 Social Support

.342 Psychological Adjustment

.148 College Satisfaction

.197 GPA

.087 Age

.063 Gender

.058 High School Average

.076 Conscientiousness


.093 Social Support

.084 Psychological Adjustment

.172 Age

.121 Gender

.147 High School Average

-.475 Negative School Events

.132 Negative Life Events

.813 Error

Goal Commitment - 2

.101 Goal Commitment - 1

.071 Social Support

.122 Social Integration

.506 Academic Integration

-.059 Ethnic

.087 Conscientiousness

.048 Self Esteem

-.065 External Involvement - 2

.695 Error

Institutional Commitment - 2

.163 Institutional Commitment - 1

.098 Social Support

.158 Psychological Adjustment

.084 College Satisfaction

.197 Social Integration

.381 Academic Integration

.066 Age

-.184 Negative School Events

.104 Negative Life


.057 Institutional Commitment - 1

.197 Social Integration

.064 Academic Integration

.268 GPA

.068 Goal Commitment - 2 (3)

-.100 External Commitment - 1

-.450 Negative School Events

-.070 External Involvement - 2

.123 Negative Life Events

.654 Error

Goodness of Fit Summary

All paths significant at p. < .05-level.

CFI = .997.

2(174)= 193.46, p =.123.

Table 3. Classification results for discriminant function analysis.

Predicted Group Membership

Withdrew Persisted - N (%)
Actual Group N % N %
Withdrew (N = 279)

Persisted (N = 732)









Percent of cases correctly classified








*p < .0001

Table 4. Correlation Matrix.

Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)  



Socioeconomic status

Developmental placement

H.S. average

Grade point average

Campus size



Goal commitment -1

Institutional commitment - 1

External commitment -1

SIQ-Social integration





































































































































































































  (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14)
SIQ-Academic integration

Goal commitment - 2

Institutional commitment - 2

External commitment - 2

Self esteem

Social support

SACQ-Academic adjustment

SACQ-Social adjustment

SACQ-Psychological adj.

SACQ-Institutional comm.

Negative life events

Negative school events

Satisfaction with college




















































































































































































































  (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) (26) (27)  
Goal commitment - 2

Institutional commitment - 2

External commitment -2

Self esteem

Social support

SACQ-Academic adjustment

SACQ-Social adjustment

SACQ-Psychological adj.

SACQ-Institutional comm.

Negative life events

Negative school events

Satisfaction with college

























































































































































































Figure 1. Expanded model of student departure.


1. Since the college offers "late-start" courses which begin eight weeks into the semester, recording retention or departure was delayed until the eighth week of class.

2. The Freshman Seminar class is designed to assist students in dealing with the academic and non-academic demands of college (e.g., developing note-taking and study skills, time management skills, introduction to library services, registration services, and advisement services). It is a required course for all first-time full-time entrants.

3. Confirmatory factor analysis results are available from the first author upon request

4. In the present paper, significance is defined as p < .05, or below.