Research in Higher Education, Vol. 36, No. 3, 1995
FACTORS AFFECTING HISPANIC STUDENT TRANSFER BEHAVIOR Barbara A. Kraemer .
.
.
o
o
o
.
. 9
. 1
.
.
.
.
7. 6 .
.
.
.
.
~ .
.
1 .
7 .
6 .
1 .
7 .
6 .
. .
.
. .
. .
. .
.
.
.
.
.
.
.
o o ~ 1 4 9.
~ 1 7 6 .
.
.
.
.
. .
. .
. .
. .
. .
. .
. .
. .
.
. o
.
. .
. .
. .
~ .
.
.
1 .
7 .
6 .
1 .
7 .
6 .
. .
.
.
.
.
.
.
.
. .
. .
. .
. .
.
.
~
.
.
.
.
. .
.
.
.
o
o
9 1 7. 6 .
o .
~ .
1 .
7 .
6
1
7
6
.
Few studies examine transfer of nontraditional-age Hispanic community college students to four-year institutions. The present study examined attitudes, cognitive and noncognitive experiences, and transfer-related behaviors of students (N = 277) who graduated from one two-year institution, in order to determine factors associated with successful transfer and make appropriate policy recommendations. An undergraduate survey based on Tinto's (1975, 1987) conceptual framework and Cabrera, Nora, and Castafieda's (1993) integrated model of student retention provided measures of family concerns, encouragement, social and academic integration, English communication skills, and educational and institutional commitments. Structural equation modeling via LISREL gave parameter estimates of causal links among variables in the hypothesized model. Variables significant for Hispanic transfer students were mathematics ability, academic achievement, and an intent to transfer. ooo . . . . . . . . . . . ooo~176 . . . . . . . . . . . . . . . . . . . . . . . . . ~ . . . . . . . . . . . . o~176149 . . . . . . . . . . ~176176 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 . . . . . . . .
Within the field of higher education, few studies can be found related to community college issues and even fewer studies on the retention, graduation, and transfer of Hispanic students attending two-year institutions. Yet Durfin (1986) pointed out that 37 percent of all students in higher education and 56 percent of all Hispanic students are enrolled in community colleges. Community colleges have greatly increased access to higher education in the United States because of open admissions policies, proximity of campuses to students, and financial assistance (Cohen and Brawer, 1989). Nationwide one-fourth of all community college students are minorities, and this proportion is higher in cities with high minority populations (Cohen, 1988). These students have the lowest retention rates and the highest transfer losses. One of the most pressing issues facing higher education today is to determine how to increase the number of minority students who earn both undergraduate and graduate degrees. Nora and Rend6n (1990) indicated that although 80 percent of Hispanic community Barbara A. Kraemer, Associate Dean of Faculty, St. Augustine College, 1333. West Argyle Street, Chicago, Illinois 60640. An earlier version of this paper was presented at the Association for Institutional Research Forum, Chicago, Illinois, May 1993. 303 0361-0365/95/0600-0303507.50/0
9
1995
Human
Scictw.es
[hr.ss,
Inc.
304
KRAEMER
college students express the intent to transfer from a two-year to a four-year college, national transfer rates for Hispanics (and most minorities) remain exceedingly low--between 5 and 20 percent. The present study examined the attitudes, cognitive and noncognitive experiences, and transfer-related behaviors of persisters to identify factors associated with successful transfer and college completion.
THEORETICAL FRAMEWORK AND LITERATURE REVIEW The study is theoretically based on Tinto's (1975, 1987) original concept of student departure and Cabrera, Nora, and Castafieda's (1993) comprehensive model of student persistence. Tinto (1975, 1987) presents a schema of dropout behavior and analyzes the role of the institution in influencing the retention of students. Factors related to family background, individual attributes, and precollege schooling are believed to influence the initial goal commitments of the student and his/her identification with a particular institution. These commitments in turn influence the student's interactions with the academic and social components of the institution, which lead to academic integration and social integration, respectively. Both integrative processes further influence the final goal and institutional commitments of the student, and ultimately impact the decision to persist or withdraw from college. Cabrera, Nora, and Castafieda (1993) merged Tinto's (1975, 1987) and Bean's (1985) theoretical frameworks, developing an integrated model of student retention which posits that the external environment has an effect on a student's academic and social integration, the student's commiUnent to the institution and to an educational goal, the intention to continue in college, and persistence behavior.
Persistence Studies The Tinto (1975, 1987) model provides the theoretical foundation for numerous studies of retention of traditional college students, especially resident students at four-year institutions (Pascarella and Terenzini, 1979, 1980; Pascarella, 1980; Pascarella et al., 1981; Pascarella and Chapman, 1983). More recently, the model has been the foundation for studies of retention and transfer of nontraditional students attending two-year institutions (Nora, 1987; Nora, Attinasi, and Matonak, 1990; Nora, and Rend6n, 1990; Horvath, 1991). In general, these studies have validated the use of Tinto's (1975, 1987) attrition model to study the persistence process among different student populations. In studies of traditional students, Cabrera, Nora, and Castafieda (1992, 1993) demonstrated the value of including environmental variables as measures of the external commitments in Tinto's (1975, 1987) model. Although Tinto (1987) indicated that external commitments would have an influence on persistence, he
HISPANIC STUDENT TRANSFER BEHAVIOR
305
did not specify any of these in his model. Both Cabrera, Nora, and Castafieda (1992, 1993) studies showed that the student's satisfaction with financial support received to attend college, and support from parents and friends, influenced the persistence process. Nora's (1987) study of two-year college students found that encouragement by significant others influenced the student's integration into the social environment and his/her educational goal commitment. A number of other studies (Aitken, 1982; Bean and Metzner, 1985; Chac6n, Cohen, and Strover, 1986; Nora, 1987; Nora and Wedam, 1991; Cabrera et al., 1992; Cabrera, Nora, and Castafieda, 1992, 1993) have shown the influence of various pull factors on persistence behavior--hours of work, finances, family and personal problems, and outside encouragement. Cabrera, Nora, and Castafieda (1993) also found that the constructs of academic and social integration, goal and institutional commitments, and academic integration and academic achievement (measured by GPA) were related but not causally. Furthermore, Cabrera al. (1993) included intent to persist in college in their integrated model of student retention. The Cabrera et al. study (1993) and other retention studies (Bean, 1985; Bean and Metzner, 1985) that build on models of attitude-behavior interactions (Bentler and Speckart, 1979, 1981) have shown that the intent to persist in college influences persistence behavior. These modifications of the Tinto (1987) model represent a development of thought that can guide current educational research in explaining the complex process of student transfer-related behavior. Transfer Studies
Few studies (Nora, Attinasi, and Matonak, 1990; Nora and Rend6n, 1990; Horvath, 1991) were identified in the literature that have examined theoretical models of transfer behavior (Nora et al., 1990) or that have tested specific quantitative transfer models (Nora and Rend6n, 1990; Horvath, 1991). However, the variables in these studies and the identification of constructs represented in these conceptual frameworks were theoretically based on Tinto's (1975, 1987) model of student departure. The Nora and Rend6n (1990) study of Hispanic community college students examined Hispanic students' "predisposition to transfer," which was measured by the number of institutions to which students planned to apply for transfer; whether students discussed transfer opportunities and sought transfer information; and students' perceptions about the importance of transferring. The model supported the fact that a high degree of congruence between the student and the environment leads to a predisposition to transfer. All five factors in the model--parents' education, encouragement, initial goal and institutional commitments, social integration, and academic integration--were significant and together explained 65 percent of the variance.
306
KRAEMER
Factors in Present Study Because of the special characteristics of the population in this study (Hispanic, adult, and primarily immigrant), family responsibilities and problems will be examined for their influence on transfer. The majority of the student population is female, and most are responsible for raising children. Completion of the two-year degree is at some cost, and at times the pull of the family is stronger.than the desire to transfer to a four-year institution (Chac6n, Cohen, and Strover, 1986). An added measure of academic integration for this population is student satisfaction with the ability to communicate in English. Sanchez and Cardoza (1984), using the High School and Beyond data set, found that English-language proficiency exerted a positive influence on both educational aspirations and educational achievement of Hispanic students. It may also influence transfer behavior. Support and encouragement by persons important in the life of the student have been found to be instrumental in the persistence of Hispanic two-year college students. A study of nontraditional freshmen (commuters, older, stratified sample including more minority students) demonstrated the importance of encouragement by significant others for the student's commitment to the institution, academic and social integration, and persistence (Nora and Wedam, 1991). Support and encouragement within the academic environment by counselors and fellow students may promote not only persistence but also transfer to four-year institutions of adult learners such as those in the study population who have been away from school for an average of ten years before beginning college studies. In summary, the present study utilized the Cabrera, Nora, and Castafieda (1993) integrated model as a basis for looking at transfer-related behavior of Hispanic two-year college students (Nora, Attinasi, and Matonak, 1990; Nora and Rend6n, 1990; Horvath, 1991). Three precollege factors will be examined: the student's mathematics ability upon entering college (Tinto, 1975, 1987), family concerns (Aitken, 1982; Bean and Metzner, 1985; Chac6n, Cohen, and Strover, 1986), and encouragement to continue in college (Metzner and Bean, 1987; Nora, 1987; Nora & Rend6n, 1990; Nora and Wedam, 1991; Cabrera, Nora, and Castafieda, 1992, 1993). Academic and social integration (Pascarella, 1980; Terenzini and Pascarella, 1980; Pascarella and Chapman, 1983), satisfaction with English communication skills (Sanchez and Cardoza, 1984), and academic performance (Cabrera, Nora, and Castafieda, 1992; Cabrera et al., 1992) will be included to identify the influence of the student's experiences in college on his or her intention to transfer to a four-year institution and on actual transfer behavior. Similarly, the study will look at the influence of the student's educational goal and commitment to the two-year institution attended (Pascarella and Terenzini, 1980; Pascarella and Chapman, 1983; Nora, 1987; Nora and Cabrera, 1991) in relation to a transfer decision. Figure 1 identifies these relationships in a hypothesized model of student transfer.
HISPANIC STUDENTTRANSFER BEHAVIOR
307
DATA SOURCES AND METHODOLOGY
The student population (N = 277) consists of all students who graduated between 1990 and 1992 from a private bilingual junior college in Illinois. The college was established to increase representation of Hispanic adults in postsecondary education and to be a bridge institution to mainstream American life. As a bridge institution, the college provides a program that enables students who wish to continue their education to transfer to four-year institutions after graduation. The sample for this study is comprised of all students who completed a survey administered prior to graduation as part of the clearance process. Other information about each student was obtained from the registrar's office and the transfer center. All students were required to complete the survey instrument. The questionnaire provided measures of latent constructs that are representative of the study's theoretical framework. Variables in the Model
Variables that are believed to have an impact on intent to transfer and transfer behavior for Hispanic students graduating from a two-year commuter institution are described below. When a scale is used to operationalize a variable, it is a Likert scale with a five-point range.
Demographics Background characteristics that reflect the unique character of the student body have been included as descriptive variables: age, gender, place of birth, marital status, children, family income, parents' educational levels, number of years out of school before coming to college, and type of high school diploma.
Precollege Factors Three exogenous variables in the model are Mathematics Ability, Family Concerns, and Encouragement. The score on the mathematics placement test, taken at admission, is the measure of Mathematics Ability. Because some students obtained theft high school diplomas in other countries or through the General Education Diploma (GED) program, it was not possible to use high school grades, class rank, or grade-point average as indica~tors of the students' potential academic ability at entrance. Nor could the English placement test score be utilized because the method of scoring the instrument was changed after the fh'st group entered, and students' results could not be compared. Thus, the mathematics placement score proved to be the only common measure of potential academic ability at entrance. Family Concerns, the responsibilities and problems that may make it difficult to study, comprise one type of external commitment, which may or may not
308
KRAEMER
have a negative effect on transfer. Student responses to the question, "To what extent do the following factors make it difficult for you to study?" for three items--caring for family members, housework, and family problems--are the multiple indicators of this construct. Encouragement to continue in college is another exogenous factor hypothesized to have a strong influence on a number of factors in the model. Indicators of this construct are student responses to the question, "To what extent did the following persons encourage and support you to continue your studies?" for two groups--fellow student(s) and counselor(s).
Intervening Variables The student's Academic Integration is measured by the frequency of student participation in classroom discussion. Because students frequently begin their college studies with a very low level of academic English, a large number must complete English-as-a-Second-Language courses before taking college English. Therefore, the student's satisfaction with his or her ability to use English in an academic setting is a separate measure of academic integration. A scale formed from a composite of five items measures student satisfaction with the following English communication skills: reading college texts, listening to a conversation, listening to a lecture, speaking to a group, and speaking in class. Another item, satisfaction with command of English when speaking in a conversation, provides an additional indicator of the construct. Cumulative grade-point average at graduation measures Academic Achievement. It is believed that the socialization process for Hispanic students at the study institution is facilitated by interactions with persons from the same background in a compatible cultural environment. Thus, Social Integration for this sample is defined as the extent to which students feel at home at the college because of Hispanic faculty and staff, other Hispanic students, and Hispanic cultural activities during break times. The final commitments of the student at graduation are divided into the student's educational goal commitment and the student's commitment to the particular institution. The importance of pursuing a college education is a single indicator of the student's educational goal. Institutional Commitment is measured by two indicators: the importance of the particular college in the student's life, and the student's willingness to choose the same college if the student were beginning college now.
Outcome Measures Both outcome measures, the intent to transfer and transfer behavior, are dichotomous. The intent to transfer to a four-year institution after graduation is determined from the student's response to the question asked before graduation,
HISPANIC STUDENTTRANSFER BEHAVIOR
309
"Do you intend to transfer to a four-year college or university immediately after graduation?" Evidence of transfer behavior is obtained later from the transfer center, which contacts the colleges and universities to which transcripts have been sent in order to determine which graduates have been admitted to another institution. Specification of the Model
The hypothesized causal model (Figure 1) consists of eight structural equations. The temporal sequence of the factors in the structural model follows the Cabrera, Nora, and Castafieda (1993) integrated model of student retention. The three exogenous variables--Mathematics Ability, Family Concerns, and Encouragement--are hypothesized to have direct effects on the intervening variables that measure Academic Integration and Academic Achievement. Since Mathematics Ability measures the student's potential academic ability, it is also hypothesized to directly affect the other construct related to the student's academic integration, namely, the student's satisfaction with his or her command of English. Because the development of English communication skills is critical for these students to succeed in college, Encouragement by counselors and peers to continue in college is also hypothesized to directly affect this variable. Family Concerns and Encouragement are believed to directly affect
FIG. 1. Hypothesized model of student transfer.
310
KRAEMER
Social Integration. Encouragement is hypothesized to have a direct effect on final educational Goal and Institutional Commitments, Transfer Intent, and Transfer Behavior, whereas, for Mathematics Ability and Family Concerns, the effects on commitments and transfer are believed to be indirect. Academic and Social Integration, Academic Achievement, and English communication skills are all hypothesized to have direct effects on educational Goal and Institutional Commitments, Transfer Intent, and Transfer Behavior. It is hypothesized that final commitments, both educational and institutional, will have a direct effect on intent to transfer and on transfer. The model further hypothesizes noncausal relationships between Academic Integration and Academic Achievement, Academic and Social Integration, and educational Goal and Institutional Commitments.
DATA ANALYSIS LISREL 7 (Joreskog and Sorbom, 1989) was used to examine the parameter estimates for both the structural and measurement models. Factor ioadings and unique variances for the measurement model are given in Table 1. T-values for TABLE 1. Factor Loadings and Unique Variances for Constructs in Measurement Model
Construct Family Concerns Encouragement Academic Integration English Social Integration
Goal Commitment Institutional Commitment
Factor Loading
Unique Variance
caring for family members housework family problems from fellow students from counselors participation in class discussions
.785 .822 .567 .712 .771 .995
.384 .325 .678 .493 .405 .010
scale: read, listen, speak (ENGSATIS)' speak in conversation Hispanic faculty and staff other Hispanic students Hispanic cultural activities during break times important for me to pursue a college education if starting college again, would choose same college this college important in my life
.962 .847 .795 .719
.074 .282 .367 .484
.744
.447
.995
.010
.739 .722
.454 .479
Measure
9Scale items include reading college texts, listening to conversation, listening to lecture, speaking to group, speaking in class.
HISPANIC STUDENT TRANSFER BEHAVIOR
311
all estimated parameters were significant, denoting that they are good measures. Measures of goodness of fit indicated that the standards set for evaluating the quality of both the measurement and the structural models were met (Pedhazur, 1982; Anderson and Gerbing, 1988; Marsh, Balla, and McDonald, 1988; Joreskog and Sorbom, 1989; Byrne, 1989). The hypothesized model is supported by a goodness-of-fit index of 0.964, an adjusted goodness-of-fit index of 0.943, and a root mean square residual of 0.056. The chi square ratio of 193.90 (df = 106) and the chi square divided by degrees of freedom ratio of 1.83 are low and meet the requirements of the tests of significance. Parameter estimates of causal links among the latent constructs in the hypothesized model are presented in Figure 2. An unweighted least squares (ULS) solution, which assumes a nonnormal underlying distribution, was produced from the correlation matrix. ULS was more appropriate than maximum likelihood estimates because of the nonnormal distribution of one of the outcome measures, transfer behavior. No differences in results occur in utilizing ULS rather than ordinary least squares or another solution.
Descriptive Variables The sample included 277 Hispanic graduates of one two-year college. They were predominantly female (78%), and their mean age at graduation was 33.63
""-,",,
,' ! \
~
-./
%
FIG. 2. Structural model of student transfer.
7 /
....
nol signif~..ant
312
KRAEMER
years. Only 14 percent were born in the United States; the other 86 percent came from Puerto Rico (30%), Mexico (28%), Central America (13%), South America (11%), or other Caribbean countries (3%). On an average, they had spent 15 years in the United States. The graduates were similar in socioeconomic background. Forty-two percent were married, and another 24 percent were either divorced, widowed, or separated. Eighty-one percent were responsible for raising children. But annual household incomes were low: 92 percent had incomes of $25,000 or less, and 67 percent of the incomes were at or below $15,000 per year. Over a third (35%) of the graduates worked 40 hours or more each week at outside jobs; another 17 percent worked 10 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 percent of the mothers and 9 percent of the fathers had completed more than twelve years of education. Close to 90 percent of these graduates were first-generation college students. Only 20 percent of them graduated from high school in the United States, another 50 percent received the General Education Diploma (GED), and almost a third of the sample (31%) 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. Twenty percent of the graduates continued their education at another institution, which is a high percentage when one considers the multiple objectives of a community college (Breneman and Nelson, 1981; Dougherty, 1991) and the low transfer rates of community college students in general (Cohen and Brawer, 1989).
Structural Equations in the Model Figure 2 presents the structural coefficients for the estimated model of student transfer. The gamma and beta weights from the standardized solution are the structural coefficients, that is, the measures of the paths. The hypothesized paths that were found to be significant are represented by solid lines, the nonsignificant paths by dotted lines. The total coefficient of determination for the structural equations was 0.621, indicating that the model of student transfer accounted for 62 percent of the variance. The LISREL estimates (standardized gamma and beta weights) and the squared multiple correlations for the individual structural equatiolas are given in Table 2. The results for each of the eight structural equations in the model are discussed below. Academic Integration. It was hypothesized that the three exogenous variable-Mathematics Ability, Family Concerns, and Encouragement--would have an effect on Academic Integration. The R 2 for the first structural equation was 0.202, that is, the equation explained 20 percent of the variance in the dependent variable. Of the three hypothesized influences, two of them were
. . . . . . . 16.0%
.078 . . . .
.390* .042
. . .
Academic Achievement
9T-values for these coefficients indicated that they are significant,
20.2%
.359* . . . . . . .
Encouragement A c a d e m i c Integration Academic Achievement English Social Integration Goal C o m m i t m e n t Institutional C o m m i t m e n t Transfer Intent
R2
.257* .084
M a t h e m a t i c s Ability Family C o n c e r n s
E s t im ate s
Academic Integration
. . . . . . 19.6%
.
.410"
.168 --
English
.
. 27.0%
. .
.498*
.146"
Social Integration
. . . 12.1%
.
,155 .022 .205* .037 ,146
.
Goal Commitment
48.2%
.338 - .211 * .006 .122 .450*
Institutional Commitment
13.0%
.307 .052 .162* - ,051 - .099 .101 .184
Transfer Intent
TABLE 2. LISREL Estimates and Squared Multiple Correlations (R 2) for Structural Equations
20.2%
- .093 - .027 .137* .114 -.049 .052 .051 .387*
Transfer Behavior
314
KRAEMER
found to be significant, Mathematics Ability (gamma = 0.257) and Encouragement by counselors and fellow students (gamma = 0.359). Family Concerns, a pull factor, did not have a significant effect on Academic Integration, either positive or negative. The dependent variable, Academic Integration, was thus explained by the two factors of Mathematics Ability and Encouragement, the respective T values being 2.116 and 1.834. Academic Achievement. The same three exogenous variables were hypothesized to have an effect on the next dependent variable, Academic Achievement. The R 2 for the second structural equation was 0.160 (16%). Of the three hypothesized variables, only Mathematics Ability had a significant effect on Academic Achievement (gamma =0.390); the T value was 4.252. English Communication Skills. Two exogenous variables, Mathematics Ability and Encouragement, were hypothesized to have an effect on the other concept related to academic integration, namely, satisfaction with English communication skills. The R 2 for the structural equation was 0.196 (20%). Encouragement had a significant effect on English skills (gamma =0.410) and was the only variable that explained English skills. The T value was 2.228. Social Integration. Two exogenous variables, Family Concerns (gamma = 0.146) and Encouragement (gamma = 0.498), were believed to and did have a significant effect on Social Integration; the stronger influence came from Encouragement. The R 2 for the equation was 0.270 (27%). The T values for Family Concerns and Encouragement were 1.671 and 2.559, respectively. Goal Commitment. The student's educational Goal Commitment was hypothesized to be affected by a number of variables: Encouragement, Academic Integration, Academic Achievement, English Communication Skills, and Social Integration. The R 2 for the equation was 0. 121 (12%). The results indicated that only Academic Achievement (beta =0.205) had a significant direct effect (T = 2.626) on Goal Commitment and explained this dependent variable. Institutional Commitment. The next dependent variable was the student's commitment to the particular institution. The same five independent variables hypothesized to affect Goal Commitment were believed to influence Institutional Commitment. The R 2 for the structural equation was 0.482 (48%). In this case, only two variables were significant, Academic Integration (beta = - 0 . 2 1 1 ) and Social Integration (beta =0.450), the greatest portion of the variance being explained by Social Integration. The T values for these two variables were - 2 . 1 9 7 and 4.075, respectively. The influence of Academic Integration on Institutional Commitment was negative. Intent to Transfer. The first outcome measure was the student's intention to transfer to another academic institution. It was believed that Transfer Intent would be influenced by the same variables as for the preceding two equations-Encouragement, Academic Integration, Academic Achievement, English Communication Skills, and Social Integration--plus Goal and Institu-
HISPANIC STUDENT TRANSFER BEHAVIOR
315
tional Commitments. The R 2 for this equation was 0.130 (13%). Only one hypothesized variable, Academic Achievement, proved to be significant (beta =0.162) in explaining the dependent variable. The T value was 1.805. Transfer. The second outcome measure, Transfer Behavior, was believed to be influenced by the same seven variables as in the previous equation, plus Transfer Intent. The R 2 for the final structural equation was 0.202 (20%). Two variables, Academic Achievement (beta =0.137) and Transfer Intent (beta =0.387), were found to be significant. The dependent variable was explained primarily by intent to transfer. The T values for Academic Achievement and Transfer Intent were 1.930 and 5.529, respectively. Noncausal Relationships
Based on previous studies, three noncausal relationships were hypothesized in the structural model: relationships between Academic Integration and Academic Achievement, Academic and Social Integration, and Goal and Institutional Commitments. None was significant for this population. The LISREL estimates were: Academic Integration and Academic Achievement, 0.111 ; Academic and Social Integration, 0.025; and Goal and Institutional Commitments, 0.063. Direct, Indirect, and Total Effects An examination of the direct, indirect, and total effects of each construct on the outcome measures, Transfer Intent and Transfer Behavior, in the final structural model showed that three factors had a significant impact on Transfer Behavior: Transfer Intent, Academic Achievement, and Mathematics Ability. The total effects were 0.387, 0.220, and 0.116, respectively. Both Academic Achievement and Mathematics Ability also had a significant influence on the other outcome variable, Transfer Intent. The total effects were: Academic Achievement, 0.184, and Mathematics Ability, 0.095. Finally, even though there were significant internal paths for some of the other variables in the model, these variables did not have an impact on the outcomes. The model substantiated that Hispanic students with higher levels of mathematics ability at admission were more likely to achieve academically, express an intent to transfer to a four-year institution, and actually succeed in transferring after graduation. DISCUSSION
Findings of both explanatory and confirmatory factor analyses indicate that the psychometric properties of the items and the scale used to provide measures
316
KRAEMER
of the latent constructs in the structural model were highly reliable and valid. In particular, three factors were found to be significantly related to transfer for the graduates in the sample: the student's mathematics ability, academic performance, and the student's intention to transfer to a four-year institution after graduation. The direct effect of the student's mathematics ability (as measured by the mathematics placement test) on academic integration and academic achievement is similar to findings in persistence studies. Pascarella and Chapman (1983) found that high school GPA, an indicator of academic ability for traditional two-year and four-year college students, had a direct effect on academic integration, but its effect on persistence was mainly indirect. Although Dunin (1986) presented a contradictory finding (namely that high school grades and college admission test scores were not good predictors of college grades for Hispanic students at selective institutions), the results of the present study indicated that for this Hispanic student population graduating from a two-year institution, admission test scores were a good predictor of college grades. Previous studies by Pascarella and Terenzini (1979, 1980) that incorporated academic ability (SAT scores) and high school performance (high school rank) in their investigation of traditional university students found that these and other precollege factors had no predictive validity for persistence, reinforcing Tinto's (1987) supposition that what happens in college is more important than what happens before college. However, Nora's (1987) study substantiated the importance of high school grades for persistence of Chicano two-year college students, even though the effect was not direct. Moreover, Nora and Rend6n (1990), in a study of transfer-related behavior of Hispanic and white two-year college students, found that academic ability directly influenced goal and institutional commitments and indirectly influenced persistence and the predisposition to transfer. Olivas (1979), citing evidence from aggregate National Longitudinal Study data on the high school class of 1972, reported that students with high ability and aptitude persist, transfer upward, and graduate. The present study confirmed the indirect influence of mathematics ability on intent to transfer and transfer behavior for one group of Hispanic two-year college graduates. Studies of persistence (Cabrera, Nora, and Castafieda, 1992, 1993; Metzner and Bean, 1987) and predisposition to transfer (Nora and Rend6n, 1990) of both traditional and nontraditional, two-year and four-year college students have shown that GPA, as a measure of academic achievement, has had a strong influence on persistence and transfer-related behavior. Nora (1987) concluded that achievement, persistence, and transfer are interrelated. The results of the present study provided further evidence of the importance of academic achievement on both the intent to transfer and transfer behavior for Hispanic two-year college students. A large number of minority community college students initially intend to
HISPANIC STUDENT TRANSFER BEHAVIOR
317
continue their education at a four-year institution. Richardson and Bender (1987) cited a figure of 40 to 50 percent. In this study, 56.7 percent of the sample expressed their intention to transfer at the point of graduation. Although not as many actually transfer as intend to transfer, the results of the current study showed that intent to transfer had the strongest influence on actual transfer behavior for the study population. Implications for Practitioners
From the results of the study, it can be concluded that institutions serving Hispanics with similar characteristics should identify the ability levels of students at the time of enrollment; faculty and departments can develop students' academic potential and implement means to improve students' academic achievement; and, academic advisers and transfer center personnel need to identify early which students intend to transfer to a four-year institution and then actively foster transfer behavior. The recommendations that follow are directed toward strengthening college programs and environments in order to enhance transfer of Hispanic students and to promote their completion of the bachelor's degree.
Mathematics Ability In this study, mathematics ability, a measure of students' academic potential at entrance, was associated with students' academic achievement and indirectly affected both the intent to transfer and actual transfer behavior. It is this author's opinion that, since both administrators and faculty need to be aware of the academic potential and level of preparation of minority students, the institution should utilize several common instruments to assess the quantitative and verbal abilities of entering students. With this information, the faculty would be able to develop programs to strengthen students' academic abilities as a step toward improving academic performance and increasing the transfer rate of community college students. Faculty development programs could focus on (a) recognizing student intellectual potential, (b) developing academic abilities-analysis, synthesis, problem-solving--in class, (c) promoting good study skills, and (d) challenging students to work up to their potential. The curriculum of the freshman seminar, a required semester-long course in many community colleges, could include specific classes designed to augment reasoning abilities. Small classes and tutorial sessions would give the faculty the opportunity to work productively with students individually or in small groups and monitor student progress. In addition, developmental workshops could .be provided for students as needed for the further development of their reasoning skills.
318
KRAEMER
Academic Achievement Academic achievement was found to be highly related to both intent to transfer and actual transfer behavior. It is the author's belief that promotion of academic achievement should be an institution-wide effort and not the responsibility of the faculty alone. A criticism of community college faculty (Dougherty, 1991; Richardson and Bender, 1987) is that they do not enforce high academic expectations and therefore courses are not up to university standards. Community college administrators have the responsibility of establishing and maintaining academic standards, so that students acquire a good academic foundation for continuing their education at a four-year institution. The following recommendations are offered as a contribution toward accomplishing that goal. Administrators could engage the faculty in discussion of academic standards and determination of strategies for ensuring compliance with standards. Administrators could also oversee a regular evaluation of the degree to which faculty and departments adhere to these standards, which would include, for example, a review of faculty grading practices. But establishment of academic standards involves more than grades. For more than a decade, colleges have been implementing programs of competence- or ability-based education (Mentkowski and Doherty, 1984; Farmer, 1988), and in 1992 assessment based on outcomes was incorporated into the areas evaluated by the North Central Association of Colleges and Schools for the further accreditation of any institution of higher education in its region (North Central Association, 1992). Faculties could engage in a process of defining student outcomes for each academic program and writing course objectives for syllabi in terms of outcomes. These outcomes would be expressed both as knowledge of course content and as abilities a student needs to develop, such as analysis or problem-solving. Once this foundation is laid, each department could work systematically with its faculty members so that teaching would be directed toward achieving these outcomes. Then program assessment based on student outcomes could be employed to improve academic programs. Another recommendation is that faculty notify academic advisers early in the semester if students are not achieving at a satisfactory level in a given class, so that the adviser can intervene. The college could introduce a system of performance notices for the early identification of nonachievers, so that faculty and advisers would cooperate in helping students improve their performance before midterm examinations. Because many students are underprepared and need special help, community colleges have furnished remedial and developmental programs to assist these students, but have not as frequently provided programs to challenge the more talented. A further recommendation is to offer an honors program to capable students in each area of the curriculum, and to develop other methods to stimulate all students to maximize their learning, especially in
HISPANIC STUDENTTRANSFER BEHAVIOR
319
courses in their major, thus ensuring that they are well-prepared for upper division work. Intent to Transfer
The intent to transfer had the strongest impact on actual transfer behavior in this study. Therefore, it is important that students' desire to continue their education at another institution be fostered early in their academic life. This could be accomplished through transfer center activities directed toward first-semester freshmen to awaken their interest in the possibility of transfer. The center could also arrange for alumni who have completed the bachelor's degree to meet with potential transfer students and initiate mentoring relationships with them to strengthen the possibility that the graduates would continue their education at a four-year institution. A career counseling program to help students assess their talents, refine their career goals, and learn about the educational requirements for their desired careers could also foster the intent to transfer. Implications for Researchers
The findings of the study suggest the validity of the basic theoretical framework for examining the transfer process of nontraditional Hispanic students at two-year institutions. This transfer model could be tested with Hispanic students graduating from other community colleges, where they are not a majority, to see if the same factors have an influence on their intent to transfer and transfer behavior. A comparative study could be done between Hispanic and non-Hispanic students, to see if the factors in the model are applicable to both populations. The model could also be tested with different student populations to confirm the impact of intent to transfer on transfer behavior. In the present study, family concerns had a positive effect on social integration but no relationship to transfer. This variable should be examined further in order to understand how family problems and responsibilities are dealt with by those who succeed in transferring. Minority students in this study who transferred may have developed time management skills, which enabled them to meet the challenges of transition to another college. Since 43 percent of the graduates in the study indicated that they worked between 20 and 65 hours a week off-campus, work could be included as another external commitment in the model. CONCLUDING REMARKS
The continued underrepresentation of Hispanics in higher education closes the door to greater contributions by Hispanics to American society as citizens,
320
KRAEMER
members of the labor force, and parents of the next generation. How to increase the number of minority students who earn college degrees is a compelling issue facing higher education today. Since so many of these students attend community colleges, the issue becomes how to increase the number of minority students who transfer to four-year institutions As indicated previously, many Hispanic community college students intend to transfer to four-year colleges, but national transfer rates for Hispanics (and most minorities) are exceedingly low (Nora and Rend6n, 1990; Rend6n & Mathews, 1989). Unfortunately, within the field of higher education, few studies can be found related to community college issues and even fewer studies on the retention, graduation, and transfer of Hispanic students attending two-year institutions. The overall significance of this study lies in its ability to focus attention on specific factors important to increasing the transfer rate of Hispanic community college students, and to lead to interventions that would further the participation of educated Hispanics in American society.
REFERENCES Aitken, N. (1982). College student performance, satisfaction and retention. Journal of Higher Education 53(1): 32-50. Anderson, J. C., and Gerbing, D. W. (1988). Structural equation modeling in practice: A review and recommended two-step approach. Psychological Bulletin 103:411423. Bean, J. P. (1985). Interaction effects based on class level in an exploratory model of college student dropout syndrome. American Educational Research Journal 22(1): 35-64. 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., and Speckart, G. (1979). Models of attitude-behavior relations. Educational Research Journal 22: 35-64. Bender, P. M., and Speckart, G. (1981). Attitudes "cause" behaviors: A structural equation analysis. Journal of Personality and Social Psychology 40(2): 226-238. Breneman, D. W., and Nelson, S. C. (1981). Financing community colleges. Washington, DC: Brookings Institution. Byrne, B. M. (1989). A Primer of LISREL: Basic Applications and Programming for Confirmatory Factor Analytic Models. New York: Springer-Verlag. Cabrera, A. F., Castafieda, M. B., Nora, A., and Hengstler, D. S. (1992). The convergence between two theories of college persistence. Journal of Higher Education 63: 143-164. Cabrera, A. F., Nora, A., and Castafieda, M. B. (1992). The role of finances in the persistence process: A structural model. Journal of Higher Education 63(2): 143164. Cabrera, A. F., Nora, A., and Castafieda, M. B. (1993). College persistence: Structural equations modeling test of an integrated model of student retention. Journal of Higher Education 64: 123-139. Chac6n, M. A., Cohen, E., and Sta'over, S. (1986). Chicanas and Chicanos: Barriers to
HISPANIC STUDENT TRANSFER BEHAVIOR
321
progress in higher education. In M. A. Olivas (ed.), Latino College Students (pp. 296-324). New York: Teachers College Press. Cohen, A. M. (1988). Degree achievement by minorities in community colleges. Review of Higher Education 11(4): 383-402. Cohen, A. M., and Brawer, F. B. (1989). The American Community College. San Francisco: Jossey-Bass. Dougherty, K. (1991). The community college at the crossroads: The need for structural reform. Harvard Educational Review 61:311-336. DurAn, R. P. (1986). Hispanics' Precollege and Undergraduate Education: Implications for Sciences and Engineering. New York: College Entrance Examination Board. Farmer, D. W. (1988). Enhancing Student Learning: Emphasizing Essential Competencies in Academic Programs. Wilkes-Barre, PA: King's College. Horvath, F. (1991). Causal Factors in the Retention of Transfer Students: The lmport'ance of Financial Aid. Unpublished doctoral dissertation, University of Southern California. Joreskog, K. G., and Sorbom, D. (1989). LISREL 7. Mooresville, IN: Scientific Software.
Marsh, H. W., Balla, J. R., and McDonald, R. P. (1988). Goodness-of-fit indexes in confirmatory factor analysis: The effect of sample size. Psychological Bulletin 103: 391-410. Mentkowski, M., and Doherty, A. (1984). Abilities that last a lifetime: Outcomes of the Alvemo experience. American Association for Higher Education Bulletin 36: 5-14. Metzner, B., and Bean, J. P. (1987). The estimation of a conceptual model of nontraditional undergraduate student attrition. Research in Higher Education 27(1): 15-38. Nora, A. (1987). Determinants of retention among Chicano college students: A structural model. Research in Higher Education 26(1): 31-58. Nora, A., Attinasi, L. C., Jr., 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. (in press). The construct validity of institutional commitment: A confirmatory factor analysis. Research in Higher Education. Nora, A., and Rend6n, L. I. (1990). Determinants of predisposition to transfer among community college students: A structural model. Research in Higher Education 31(3): 235-255. Nora, A., and Wedam, E. (1991, April). Off-Campus Experiences: The Pull Factors Affecting Freshman Year Attrition on a Computer Campus. Paper presented to the meeting of the American Educational Research Association, Chicago. North Central Association of Colleges and Schools, Commission on Institutions of Higher Education (1992). A Guide to Self-Study for Commission Evaluation. Chicago: Author. Olivas, M. A. (1979). The Dilemma of Access: Minorities in Two Year Colleges. Washington, DC: Howard University Press. Pascarella, E. T. (1980). Student-faculty informal contact and college outcomes. Review of Educational Research 50: 545-595. Pascarella, E. T., and Chapman, D. (1983). A multi-institutional path analytical validation of Tinto's model of college withdrawal. American Education Research Journal 20(1): 87-102. Pascarella, E. T., Duby, P., Miller, V., and Rasher, S. (1981). Preenrollment'variables and academic performance as predictors of freshman year persistence, early with-
322
KRAEMER
drawal, and stopout behavior in an urban, non-residential university. Research in Higher Education 15(4): 329-349. Pascarella, E. T., and Terenzini, P. T. (1979). Interaction effects in Spady's and Tinto's conceptual models of college dropouts. Sociology of Education 52: 197-210. 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. Pedhazur, E. J. (1982). Multiple Regression in Behavioral Research (2nd ed.). New York: Holt, Rinehart and Winston. Rend6n, L. T., and Mathews, T. B. (1989). Success of community college students. Education and Urban Socie~. 21 : 312-327. Richardson, R. C., and Bender, L. W. (1987). Fostering Minor#), Access and Achievement in Higher Education. San Francisco: Jossey-Bass. Sanchez, A. A., and Cardoza, D. (1984). Aspects of Educational Success Among Low Socioeconomic Status Hispanic Youth: A Causal Analysis Using the High School and Beyond Survey. Los Alamitos, CA: National Center for Bilingual Research. Terenzini, P., and Pascarella, E. T. (1980). Toward the validation of Tinto's model of college student attrition: A review of recent studies. Research in Higher Education 12: 271-282. 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. Chicago: University of Chicago Press