The Academic and Social Integration of Hispanic Students into College

Barbara A. Kraemer

Figures  Tables

Introduction

Academic and social integration have been examined in numerous studies of college persistence for resident students at four-year institutions (Pascarella, 1980; Pascarella, Smart, & Ethington, 1986; Pascarella & Terenzini, 1979, 1980) and for nontraditional students attending two-year institutions (Nora, 1987; Nora, Attinasi, & Matonak, 1990; Nora & Rendón, 1990). Nora (1993) presents commonly accepted definitions of the two factors in a review of the literature on educational aspirations of minority students at two-year colleges:

Academic integration: The development of a strong affiliation with the college academic environment both in the classroom and outside of class. Includes interactions with faculty, academic staff, and peers but of an academic nature (e.g., peer tutoring, study groups). (p. 235)

Social integration: The development of a strong affiliation with the college social environment both in the classroom and outside of class. Includes interactions with faculty, academic staff, and peers but of a social nature (e.g., [End Page 163] peer group interactions, informal contact with faculty, involvement in organizations). (p. 237)

In general, the studies have validated the importance of both factors for examining the persistence process among different student populations.

Pascarella and Terenzini (1979) identified lack of integration into the college environment, due to insufficient contact with members of the institution, as perhaps the most important predictor of student withdrawal. In a study of college completion by students who began at community colleges, Pascarella, Smart, and Ethington (1986) demonstrated that academic and social integration had the most consistently positive effects on long-term persistence of all the variables in the study. Nora and Rendón's (1990) study of traditional-age Hispanic and White community college students' predisposition to transfer supported the fact that a high degree of congruence between the student and the academic and social environment of the college leads to a strong institutional commitment which positively influences the student's predisposition to transfer.

Recently, the validity of the Tinto (1987) model, which has been used as a conceptual framework for persistence studies for almost twenty years, has been called into question because it was developed for a traditional college population. For example, Tierney (1992) criticizes it and asserts instead that ethnic and racial minority students (particularly Native Americans) withdraw from college because they are required to adapt to an environment defined by the dominant culture. Furthermore, they must adapt as individuals rather than as a group, which violates their cultural orientation. Providing support for Tinto's model, however, are a few recent studies with nontraditional populations demonstrating that these students' integration into the environment of the institution, both academically and socially, was critical for their persistence and their success in college (Nora, 1987, 1993; Nora & Rendón, 1990). Population differences may mean differences in the strength of influence of one factor or the other for some students, or one of the factors may have a negative rather than a positive relationship to other constructs in the model (Nora, 1993). Pascarella and Terenzini (1979) found that "the extent of influence of various measures of social and academic integration was not independent of the particular background characteristics which students bring to college" (pp. 208-209). In any case, the constructs themselves are valid (Nora, 1993), and their validity for nontraditional college students can be confirmed through additional studies that examine valid measures for these constructs from previous studies and identify additional measures appropriate for ethnic and racial minority students. Culturally sensitive operational definitions could contribute to stronger measurements for the constructs to demonstrate their importance for college persistence of minority students. [End Page 164]

Purpose of the Study

This study raises the question of whether conventional operational definitions of academic and social integration employed in persistence studies were appropriate for Hispanics at two-year colleges. The purpose of the study is to empirically explain indicators of the student's academic and social integration into college for an older Hispanic two-year college population. Such indicators could then be used to measure the effect of the two integration factors on the persistence of Hispanic students in similar circumstances.

Literature Review

Studies of Traditional Students

Most studies of student attrition that have validated the constructs of academic and social integration have based their operational definitions of these two constructs on indicators that Pascarella and Terenzini (1979, 1980) identified in a study of traditional university students in New York State. They used principal components analysis with a varimax rotation to verify the dimensionality of the factors, which they measured by survey items that loaded on four scales: (a) academic and intellectual development, (b) faculty concern for teaching and student development, (c) informal relations with faculty, and (d) peer group relations. Cumulative grade point average (GPA) for the freshman year and frequency of involvement in extracurricular activities were added measures of academic and social integration, respectively.

Several studies of traditional freshman populations in the Southwest and the Midwest (Cabrera, Castañeda, Nora, & Hengstler, 1992; Cabrera, Nora, & Castañeda, 1992, 1993) selected items from the Pascarella and Terenzini (1979, 1980) research and added other items to provide multiple indicators of academic and social integration. The Cabrera, Castañeda, Nora, and Hengstler (1992) study used the 4-item scale of academic and intellectual development that Terenzini, Lorang, and Pascarella (1981) had determined to be good measures of academic integration; two items measuring the noncognitive component of academic integration ("I have performed academically as well as I anticipated I would," and "I am satisfied with my academic experiences"); and GPA as a measure of academic performance. The Cabrera, Nora, and Castañeda studies (1992, 1993) used the two noncognitive items above, plus "I am satisfied with my course curriculum," as indicators of academic integration; GPA was a measure of academic performance. The three Cabrera et al. (1992, 1992, 1993) studies contained the same two indicators of social integration: developing close personal relationships, [End Page 165] and ease in meeting and making friends. Both academic and social integration were significant predictors in the persistence of traditional students at four-year institutions (Cabrera, Nora, & Castañeda, 1992, 1993; Cabrera, Castañeda, Nora, & Hengstler, 1992).

Pascarella (1985) conducted a longitudinal study using Cooperative Institutional Research Program data on a traditional student population from 352 four-year colleges. A sum of two items measured academic integration: average undergraduate grades and membership in a scholastic honor society. A sum of six items about students' involvement with peers and faculty measured social integration: knowing professor or administrator personally, being president of one or more student organizations, serving on a university or departmental committee, taking a major part in a play, winning a varsity letter, and editing the school paper, yearbook, or literary magazine. This study also supported the validity of incorporating academic and social integration in Tinto's (1987) persistence model.

The Pascarella, Smart, and Ethington study (1986) of a national sample of four-year residential students who transferred from community colleges used the same measures of academic and social integration as Pascarella (1985), omitting one item for social integration (serving on a university or departmental committee). For both men and women students, academic and social integration had positive direct effects on their completion of a bachelor's degree.

In 1983, Pascarella and Chapman compared freshmen from different types of institutions: four-year and two-year, residential and commuter. They used a 9-item scale for academic integration: first semester GPA, expected second semester GPA, hours spent studying per week, number of unassigned books read for pleasure, number of cultural events attended, informal contact with faculty on academic topics, peer conversations on academic topics, participation in an honors program or accelerated course, and participation in a career development program. They measured social integration by an eight-item scale: average number of weekends on campus per month, participation in organized extracurricular activities, number of best friends on campus, "Is there a person on campus you date regularly?", "Do you spend time with college friends on vacation?", peer conversations on social or personal topics, and informal contact with faculty on social or personal topics. The results of this study confirmed the importance of both academic and social integration for persistence, but social integration was a stronger factor in the persistence of resident students at four-year institutions, whereas academic integration was more important for commuter students at both two-year and four-year institutions.

Nora, Attinasi, and Matonak's (1990) persistence study of community college freshmen focused on students who were academically underprepared. [End Page 166] Items based on other studies (Pascarella & Terenzini, 1980, 1983; Nora, 1987) operationally defined academic and social integration. Eleven items measured academic perceptions, academic involvement and study behavior; seven items were indicators of social perceptions and social involvement. Both factors had significant effects on retention, but the effect of social integration was negative, consistent with previous research on commuter students (Pascarella, 1980).

Studies Including Minority Students

Nora's (1987) study of Chicano community college students from three institutions in the Southwest identified two indicators of academic integration: (a) student perceptions of academic experiences, based on interactions with faculty, counselors, and administrators, and (b) student perceptions of career preparation received at the institution. Students' informal contacts with faculty, counselors, and other students measured social integration. The results of the study were only minimally supportive of the Tinto (1987) model in demonstrating the effect of academic integration on persistence and could not demonstrate any effect of social integration on persistence. These findings led Nora (1987) to conclude that additional studies were needed to identify the factors and the underlying structural patterns affecting Hispanic student retention.

A predisposition to transfer study (Nora & Rendón, 1990) of 422 Hispanic and 147 White community college students along the U.S.-Mexican border included four multiple indicators of academic integration: academic perceptions (same items as in Nora 1987 study), transfer perceptions, behavior counseling, and academic behavior. The latter consisted of nine specific items including using the library to study, taking class notes, asking the instructor for help with writing skills, etc. The six indicators of social integration were: faculty contact outside of class, involvement in extracurricular activities, informal conversations with faculty, reading the college paper, looking at bulletin boards for announcements, and participating in freshman orientation. True to the specifications of the Tinto (1987) model, both academic and social integration were significant factors that affected students' predisposition to transfer but indirectly, through institutional and educational goal commitments. The study did not estimate whether the influence of each of these factors was stronger on Hispanic or White students. A question that remains unanswered is whether academic and social integration were less important than institutional and educational goal commitments for this predominantly Hispanic two-year college population, or whether the measures used in the study did not capture the meaning of the concepts for this group. [End Page 167]

Academic Performance

While earlier studies included a student's grade point average (GPA) as an additional indicator of academic integration (Pascarella & Terenzini, 1979, 1980; Pascarella & Chapman, 1983; Pascarella et al., 1983, 1986), researchers (Cabrera, Nora, & Castañeda, 1992; Cabrera, Castañeda, Nora, & Hengstler, 1992) later recognized that GPA, though related to academic integration, is a separate measure of academic performance. Tinto (1987) defined the concept of academic integration as the individual's evaluation of the academic system, rather than the student's academic achievement.

Operational Definitions of Constructs in This Study

Academic Integration

Previous studies have defined academic integration in terms of the student's cognitive development and student perceptions of satisfaction with elements in the classroom environment and with certain academic behaviors, such as interactions of an academic nature with faculty, staff, and peers. I incorporated several items from previous studies into the survey as possible indicators of the student's affiliation with the academic environment of the college: frequency of the student's participation in classroom discussions (Q1), and meeting with instructors outside of class to discuss academic topics (Q5) (Pascarella & Terenzini, 1979; Pascarella & Chapman, 1983; Nora, 1987; Nora & Rendón, 1990). I incorporated use of facilities on campus, specifically the library (Q2) and a computer lab (Q4), as potential academic indicators, based on the hypothesis that commuting students who made use of these facilities outside of class would be more involved in their cognitive development, whether they needed to use the resources available to complete assignments or simply wanted to study away from the distractions of their home environment. Such behavior arguably demonstrates a commitment to pursuing academic tasks. I also hypothesized that if students seek assistance from others on campus, either tutoring (Q3) or academic advising (Q6), they would be more involved in intellectual development and academic-related activities than if they did not. Therefore, the definition of academic integration included the use of academic facilities and services.

I measured the six items with Likert scales ranging from 5 ("very often") to 1 ("never"). (Academic performance, a separate construct from academic integration, was measured by the student's cumulative grade point average at graduation.)

Social Integration

Social integration is usually defined as informal interactions between students and faculty, between students and counselors, and among peers. [End Page 168] Some researchers have also included the student's social perceptions and participation in extracurricular activities. This study was based on the assumption that the construct of social integration in persistence studies is valid, but it also raises the question of whether operational definitions of social integration that have been used in the past are appropriate for Hispanic students, specifically adult Hispanic students--the majority of whom are immigrants--attending a two-year bilingual commuter institution. This study attempted to identify culturally sensitive measures of the concept suitable for this specific population, as recommended by earlier studies (Pascarella & Terenzini, 1979, 1980; Nora, 1987). Furthermore, indicators from other studies seemed more appropriate for traditional college students who are free from family or work responsibilities and who can spend lengthy periods of time on campus or who reside on campus. The population in the study consisted of commuter students, the majority of whom were responsible for raising children and many of whom were working off campus. They had little time for extracurricular activities. Because of the nature of the student body, the study institution has few student organizations, only one sports team (soccer), and a school newspaper that is published only once a semester. As a result, I did not use traditional indicators of social integration.

I believed that the development of Hispanic students' strong affiliation with the social environment of the college was facilitated by interactions with persons from the same background in a compatible cultural environment. Thus, I defined social integration for this sample as the extent to which the Hispanic environment and interactions with other Hispanics at the college gave students a sense of being an integral part of the institution and feeling welcomed. Students were asked, "To what extent did each of the following help you to feel at home at the college?" The following items provided potential indicators of the construct: Hispanic faculty and staff (Q21), other Hispanic students (Q22), Hispanic cultural activities during break times (Q23), Hispanic music at dances and other events (Q24), and Hispanic food in the cafeteria (Q25). I measured these five items by Likert scales ranging from 5 ("a great deal") to 1 ("not at all"). The hypothesized factor structure for both academic and social integration is presented in Figure 1.

Data Sources

The student population consisted of all students from the 1990 to 1992 graduating classes of a private bilingual junior college in the Midwest (student population 1,400). The college has an "open door" admissions policy, and students may begin their studies in Spanish while also learning English. Students who wish to continue their education transfer to four-year institutions after graduation. [End Page 169]

The sample (N = 217) was comprised of all students who completed a survey administered prior to graduation. I obtained cumulative grade point averages from the registrar's office and evidence of transfer from the transfer center. I based the measures of academic and social integration in the questionnaire on some survey questions from previous studies (Pascarella & Terenzini, 1979, 1980; Pascarella & Chapman, 1983; Nora, 1987; Nora & Rendón, 1990), plus other items added because of characteristics of the study population.

Data Analysis

I hypothesized a base-line model of six indicators for academic integration and five indicators for social integration. Using LISREL 7 (Joreskog & [End Page 170] Sorbom, 1979), I then conducted confirmatory factor analysis (CFA), which allows the researcher to test how well the covariances among the observed variables are explained by a theoretically based factor structure (Nora & Cabrera, 1993). Separate CFAs examined the psychometric properties of the items measuring each construct--academic integration and social integration. I tested two alternatives to the baseline model for academic integration before confirming the final three-factor model that best represented the dimensionality of constructs in the broad category of academic integration. I employed two confirmatory factor analyses to substantiate the dimensionality of social integration. Using these procedures improved the fit and let me determine good independent measures of each construct for the final measurement model. (Correlation matrix, means, and standard deviations for all of the indicators are available on request.)

Because the data were ordinal, I used PRELIS (Joreskog & Sorbom, 1988) to prescreen the data and obtain a polychoric correlation matrix for the measurement model (Joreskog & Sorbom, 1979). A review of the data revealed that the multivariate normality of the data was good; therefore, I employed maximum likelihood estimates.

Measures of goodness of fit assess how well the covariances produced in the estimation model reproduce sample covariances. These indicators for each successive measurement model included chi square (X²), chi square to degrees of freedom ratio (X²/df), goodness-of-fit index (GFI), adjusted goodness-of-fit index (AGFI), and root mean square residual (RMR). Byrne (1989) proposes ranges for the various measures, similar to standards established by other authors (Pedhazur, 1982; Joreskog, 1989). Chi square divided by degrees of freedom should be lower than 2.00. The goodness-of-fit index and the adjusted goodness-of-fit index (adjusted for the degrees of freedom in the model) for the hypothesized measurement model should be close to 1.00. Pedhazur (1982) specifies that these two indicators should be at least .90. The root mean square residual, which is the error term, should be small, less than .05. None of these indicators should be relied upon alone to determine the fit of the model (Nora & Cabrera, 1993). In addition, I calculated the Normed Fit Index (NFI), a nonincremental index, to compare the goodness of fit of successive measurement models of each factor, a value of .90 or above indicating an acceptable measure of the construct (Mulaik et al., 1989).

Results

Descriptive Variables

The sample consisted of 217 Hispanic graduates at one two-year college. The students were predominantly female (78 percent), with a mean age at [End Page 171] graduation of thirty-four years. Fourteen percent were born in the United States; 86% came from Latin American countries: Puerto Rico (30%), Mexico (28%), Central America (13%), South America (11%), and other Caribbean countries (3%). On an average, the students had spent fifteen years in the United States.

The graduates were similar in socio-economic background. Forty-two percent were married and another 24% were either divorced, widowed, or separated. Eighty-one percent were responsible for raising children. Annual household incomes were low: 92% had incomes of $25,000 or less, and 67% of the incomes were at or below $15,000 per year. Over a third (35%) of the graduates worked forty hours or more each week at outside jobs; another 17% worked ten or more hours per week at on-campus jobs. Parents of over half of the graduates had received no more than six years of formal schooling, although 8% of the mothers and 9% of the fathers had completed more than twelve years of education.

Close to 90% of these graduates were first-generation college students. Furthermore, they had stopped their formal schooling for an average of nine years before enrolling in college. Only 20% graduated from a U.S. high school, 50% received the General Education Diploma (GED), and 30% completed high school outside the United States. Seventy percent of the graduates indicated that they intended to earn a bachelor's degree or a degree beyond the bachelor's.

Confirmatory Factor Analysis

I conducted separate confirmatory factor analyses for both academic integration and social integration. Results for both sets of analyses met the standards set for evaluating the quality of the measurement model. (See Table 1.) The baseline measurement model for academic integration was based on the hypothesis that six variables would be independent indicators of the construct. The results showed a low goodness-of-fit index (GFI) of .851 and a low adjusted goodness-of-fit index (AGFI) of .687, with a root mean square residual (RMR) of .157. An improvement of fit resulted by eliminating two variables, one measuring use of the computer lab outside of class (Q2), and the second, meeting with one's academic counselor (Q6). I eliminated the first variable because computer facilities were rather recent additions at the college and these students, on the verge of graduation, may not have had ready opportunities to use computers as students unless they were majoring in computer information systems. I eliminated the other variable since academic counselors also do personal counseling, which may have confounded students' interpretation of the item in the survey.

Because the AGFI was still low (.736), the RMR high (.131), and the nonincremental fit index (NFI) low (.793) for the first alternative model, I hypothesized a second alternative: a three-factor model under the broad [End Page 172] category of academic integration. I identified the three constructs as formal faculty-student interaction, study behavior, and informal faculty-student interaction. This model showed an improvement of fit over the previous one-factor model, with a GFI of .958 and an AGFI of .896. However, the error term was still too high, .102, and the NFI was low, .793. The reliability of one variable, seeking tutoring (Q3), was low; and I therefore eliminated it from the model. The factor analysis for the final model with three constructs, each measured by a single indicator, showed a definite improvement of fit over the previous models. The GFI for the final model was .980, the AGFI .959, the root mean square residual .072, and the Normed Fit Index (NFI) .931. Thus, all of the indices were within acceptable ranges.

Rather than one latent construct--academic integration--with three single indicators, the final model identified formal faculty-student interaction, study behavior, and informal faculty-student interaction as three separate constructs that represented the broad theoretical concept of academic integration. Each of these constructs was operationally defined in this study by a single indicator. Although it would be preferable to identify multiple indicators for each construct, sometimes in practice only a single indicator is available. In the absence of an alternative, a single indicator is valid, although it is unlikely that it will perfectly estimate the construct (Marasculo & Levin, 1983).

I used two confirmatory analyses to determine the construct validity and the reliability of indicators for social integration. (See Table 1.) The baseline model hypothesized that five variables would be independent measures of the single construct, social integration. Initial results showed a poor fit. The item measuring whether students felt at home at the college because of Hispanic food in the cafeteria (Q25) had lower reliability than the other items. I eliminated this item and another measuring whether students felt at home because of Hispanic music at dances and other events (Q24), limiting indicators in the final model to student involvement in the social environment of the college. The model thus incorporated only items culturally sensitive to the study population. The indicators ultimately shown to be the most reliable and valid indicators of social integration were: Hispanic faculty and staff (Q21), other Hispanic students (Q22), and Hispanic cultural activities during break times (Q23). The improvement of fit was significant: GFI .989, AGFI .931, RMR .032, and NFI .984.

Thus, although additional research is needed to identify reliable and valid scales which measure each of the constructs, the findings from the confirmatory data analysis provided some indicators to measure the constructs of academic and social integration in the final measurement model. T-values for the estimated parameters were significant, denoting that they were good measures of the constructs. The factor loadings and unique variances are given in Table 2. [End Page 173] [Begin Page 175]

In summary, I identified three constructs related to the academic integration of students. Frequency of student participation in classroom discussion (Q1) was the single indicator of formal faculty-student interaction. Study behavior, the second construct, was measured by one indicator, frequency of use of the library (Q2). And informal faculty-student interaction was measured by one indicator, how often students met instructors outside of class to discuss academic topics (Q5). Multiple indicators of social integration, appropriate for the college population in the study, measured the extent to which students felt at home in the college environment due to the presence of Hispanic faculty and staff (Q21), other Hispanic students (Q22), and Hispanic cultural activities during break time (Q23). The final measurement model is presented in Figure 2.

This study of persisters at one institution used operational definitions of academic and social integration believed to be appropriate for a nontraditional Hispanic population. Even though the results of the current study are preliminary and, in relation to academic integration, limited to single indicators of underlying constructs, the results suggest that the concept of academic integration into college may not be unidimensional but rather represents a cluster of subcomponents and not a single construct. The study [End Page 175] suggested the validity of both concepts--academic and social integration--for older (over twenty-five) full-time Hispanic community college students. More research needs to be conducted to confirm the validity of these indicators and to identify additional indicators for these constructs appropriate for this population.

Implications of the Study

Implications for Researchers

This study identified three factors that contribute to the academic integration of Hispanic students: formal faculty-student interaction, informal faculty-student interaction, and study behavior. Faculty-student interaction, both formal and informal, has been shown to be a strong predictor of integration into college and a strong influence on academic achievement and persistence (Pascarella & Terenzini, 1979, 1980; Nora & Rendón, 1990). Since Hispanic adult commuter students have limited time on campus and [End Page 176] often have family and/or work responsibilities, class participation may be the primary way they interact with faculty regularly and are integrated into the academic environment.

Nonclassroom interactions with faculty have been significant in studies of Hispanic students (Nora, 1987; Nora & Rendón, 1990). The indicator of Informal Faculty-Student Interaction in this study specifies meeting instructors outside of class to discuss academic topics, thus focusing on the academic dimension of the interaction. Both formal and informal faculty-student interactions need to be included in future studies with an adult Hispanic commuter population.

The third factor in academic integration was study behavior, measured by use of the library. The 1990 Nora and Rendón study of a sample with 75% Hispanic students contained nine multiple indicators of academic behavior, one of which was use of the library. Although my study included other possible indicators, only one was determined to be valid for the study sample. Future persistence studies of minority students need to include additional potential measures of this concept.

Previous persistence studies of Hispanic students have defined social integration in terms of peer relationships (Nora, 1987) or involvement on campus outside of class (Nora & Rendón, 1990). Tierney's (1992) critique of the concept of social integration in the Tinto (1987) model is that it does not recognize that minority students need to adapt to a different cultural environment and do this as individuals, not as a group. In this study, students are in an environment in which they do not need to adapt to another culture. Therefore, the cultural dimension of social integration is taken into account in measuring the concept for this population. The presence of Hispanic personnel and other Hispanic students leads to the extensive use of Spanish on campus, and all interact in their own cultural environment rather than adapting to a milieu in which they are a minority. A future study could incorporate into the traditional definition of social integration this cultural dimension to see whether more than one factor is present.

This study was an initial investigation with a sample of persisters (graduates). Longitudinal research has been initiated at the study institution with a cohort of all incoming freshman students in an attempt to identify additional measures for academic and social integration and to document the importance of these factors in the persistence process. Other studies should be undertaken at the study institution and at other two-year colleges to investigate various measures of academic and social integration and other variables appropriate to different minority student populations. Such methodological clarification in relation to the operationalization of variables may lead to the validation of constructs important for the persistence of minority students at community colleges. [End Page 177]

Implications for Practitioners

Both student and institutional characteristics limit the generalizability of the findings to two-year institutions and to Hispanic students with very similar characteristics. This institution-specific orientation follows the recommendation of Tinto (1987) for persistence studies, namely, that one must study each institution and the experiences of individuals within each institutional setting to know why students depart. Mow and Nettles (1990), who have studied minority students, suggest that single-institution studies, in addition to providing insights to help explain the college experiences of minority students, can help the institution meet its "own unique challenges" and also "raise issues that should be addressed at other campuses" (p. 48). Given the significance of students' integration into the academic and social environment of the college for persistence and success in college, it is important that these concepts be operationalized correctly for each student population. Once the validity of the indicators is established, the institution can provide experiences to strengthen the academic and social integration of their students.

Barbara A. Kraemer is on the faculty of the Public Services Management Program at De Paul University in Chicago. An earlier version of this paper was delivered under the title of "The Dimensionality of Academic and Social Integration for Persistence of Hispanic Students" at the ASHE Annual Meeting on November 4, 1993, in Pittsburgh, Pennsylvania.

References

Byrne, B. M. (1989). A primer of LISREL: Basic applications and programming for confirmatory factor analytic models. New York: Springer-Verlag.

Cabrera, A. F., Castañeda, M. B., Nora, A., & Hengstler, D. (1992, March/April). "The convergence between two theories of college persistence." Journal of Higher Education, 63, 143-164.

Cabrera, A. F., Nora, A., & Castañeda, M. B. (1992). "The role of finances in the persistence process: A structural model." Research in Higher Education, 33, 571-194.

Cabrera, A. F., Nora, A., & Castañeda, M. B. (1993, March/April). "College persistence: Structural equations modeling test of an integrated model of student retention." Journal of Higher Education, 64, 123-139.

Joreskog, K. G. (1989). Advances in factor analysis and structural equation models. Cambridge, MA: Abt Books.

Joreskog, K. G., & Sorbom, D. (1979). LISREL 7. Mooresville, IN: Scientific Software.

Joreskog, K. G., & Sorbom, D. (1988). PRELIS: A program for multivariate data screening and data summarization (2nd ed.). Mooresville, IN: Scientific Software.

Marasculo, L. A., & Levin, J. R. (1983). Multivariate statistics in the social sciences: A researcher's guide. Monterey, CA: Brooks/Cole.

Mow, S. L., & Nettles, M. T. (1990). "Minority student access to, and persistence and performance in, college: A review of the trends and research literature." Higher Education: Handbook of Theory and Research, Vol. 6, 35-106.

Mulaik, S. A., James, L. R., Van Alstine, J., Bennett, N., Lind, S., & Stillwell, C. D. (1989). "Evaluation of goodness-of-fit indices for structural equation models." Psychological Bulletin, 105 (3), 430-445.

Nora, A. (1987). "Determinants of retention among Chicano college students: A structural model." Research in Higher Education 26 (1), 31-58.

Nora, A. (1993). "Two-year colleges and minority students' educational aspirations: Help or hindrance?" Higher Education: Handbook of Theory and Research, Vol. 9, 212-247.

Nora, A., Attinasi, L. C., Jr., & 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., & Cabrera, A. F. (1993, Winter). "Examining graduate education through structural modeling." In L. Baird (Ed.), Increasing Graduate Student Retention and Degree Attainment, (pp. 41-48). New Dimensions for Institutional Research, Vol. 80. San Francisco: Jossey-Bass.

Nora, A., & Rendón, L. I. (1990). "Determinants of predisposition to transfer among community college students: A structural model." Research in Higher Education, 31 (3), 235-255.

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

Pascarella, E. T. (1985). "Racial differences in factors associated with bachelor's degree completion: A nine-year follow-up." Research in Higher Education, 23, 351-373.

Pascarella, E. T., & Chapman, D. W. (1983). "A multi-institutional path analytical validation of Tinto's model of college withdrawal." American Education Research Journal, 20 (1), 87-102.

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

Pascarella, E. T., & Terenzini, P. T. (1979). "Interaction effects in Spady's and Tinto's conceptual models of college dropout." Sociology of Education, 52, 97-210.

Pascarella, E. T., & Terenzini, P. T. (1980). "Predicting freshmen persistence & voluntary dropout decisions from a theoretical model." Journal of Higher Education, 51 (1), 60-75.

Pascarella, E. T., & Terenzini, P. T. (1983). "Predicting voluntary freshman year persistence/withdrawal behavior in a residential university: A path analytic validation of Tinto's model." Journal of Educational Psychology, 75 (2), 215-226.

Pedhazur, E. J. (1982). Multiple regression in behavioral research (2nd ed.). New York: Holt, Rinehart and Winston.

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

Tierney, W. G. (1992, November/December). "An anthropological analysis of student participation in college." Journal of Higher Education, 63, 604-618.

Tinto, V. (1987). Leaving college. Chicago: University of Chicago Press.

[an error occurred while processing this directive].