| Laura Gonzalez
North Carolina State University
The college choice process of Latino students: Influence of demographics, academic preparation, and self-efficacy beliefs
FINAL REPORT:
National data has shown that Latino students entering college have selected two-year institutions at a greater rate than any other group. Some reasons for this tendency have been suggested (e.g., financial or academic difficulties), but they have not explained the phenomenon satisfactorily. The current study addressed this issue with logistic regression analysis, utilizing variables from the Education Longitudinal Study (ELS:2002) of the National Center for Education Statistics. The outcome variable was level of college intended by Latino and White high school students in the spring of 2004 (when many of them were seniors), with the reference level of the outcome being planning for four-year college. Predictor variables were related to demographic factors, academic preparation factors, and academic self-efficacy beliefs. Academic self-efficacy beliefs had not been utilized in prior studies examining college choice, and were therefore the focus of the current study. Six main predictors were significant and positive in their effects on the reference outcome for all students (e.g., valuing the academic reputation of the college as a factor in choice, the highest level of high school math, ELS reading and math test scores, student expectations for future education, teacher expectations for the student's future education, and an academic self-efficacy item related to persistence in studying). There were also significant predictors that had a negative effect on the reference outcome for all students (e.g., percent of free and reduced lunches at the high school, valuing low expenses or ability to live at home as factors in college choice, not planning to take the SAT). Thus, many Latino and White students planning for two-year college were responding to similar environmental factors.
To further examine differences, predictors were interacted with race/ethnicity. Only one interaction term was significant using the traditional interpretation of the regression coefficient (p<.05) - race/ethnicity times student expectations for future education, one of the academic self-efficacy variables. An alternate evaluation of interaction terms (using a cross derivative as recommended by Ai and Norton in 2003) found some areas of significance for the interaction effects of race/ethnicity with teacher expectations and math and reading test scores, in addition to student expectations. For Latino students with otherwise high predicted probabilities of planning for four year college, all three of these interactions had a negative effect (as compared to Whites). Socioeconomic status was held constant in the analysis of interaction terms, so the Latino students were similar in economic terms to the White students. However, even with high aspirations, high expectations from teachers, and good math and reading test scores, Latinos had a 16.5% lower predicted probability of the reference outcome. Therefore, that subgroup of Latinos could have been choosing two-year colleges for other reasons, such as family ties, social networks, or other dynamics common to collectivistic cultures. Findings are discussed in terms of Bandura's social cognitive theory. Academic self-efficacy and social/familial influences are promising areas for future research. School level variables may be influencing the three negative interaction effects in ways not identified in this study, but deserving of more attention. Continuing support for students in poorly resourced schools (which includes many Latinos) is important. In addition to those areas, other policy implications include the importance of understanding diverse experiences within racial/ethnic groups. The covariates in this model illuminated the differences between Latinos with a low predicted probability of the reference outcome (but a positive interaction effect of student expectations as compared to Whites) and Latinos with a high predicted probability of the reference outcome (but a negative interaction effect of student expectations as compared to Whites). The main effects suggested that students (both Latino and White) who planned for four-year college were focused on academic goals and persistent in their efforts. This could direct program planning efforts and educational interventions.
Back to Funded Dissertation Grants Page |