Investigating the factors influencing educational attainment across ethnic and gender groups: Structural analysis of NELS:88-2000 database

The purpose of this study was twofold: (1) to determine the relative influence of twelve variables (socioeconomic status, prior achievement, self-concept, locus of control, educational aspirations, parental expectations, parental involvement, peer influence, college plans, average grades, test scores, and coursework completion) simultaneously in predicting educational attainment (EA); (2) to locate similarities and differences in the effects of those twelve variables on EA across ethnic and gender groups.;Two central questions are addressed. What factors are major contributors to EA? What significant differences exist among ethnic and gender groups? A modified version of the Wisconsin model of EA was applied to samples in the NELS:88-2000 database. Data were selected from the base year, and the second and fourth follow-up reports. Structural equation modeling (SEM) was utilized to perform single-group analysis on a total sample (N = 12,144), and multiple-group analysis on randomly selected samples of males (N = 419) and females (N = 419) in each ethnic group. Any models containing an ethnic factor were tested separately, with a sub-sample containing only two groups to assure accurate comparisons of each ethnic group. The study also performed latent mean analysis and invariance tests to identify differences among any ethnic and gender groups.;There are several important findings. Socioeconomic status (SES) factors and prior achievement had the largest total effects on EA for all ethnic and gender groups. This was mainly through other mediating productivity factors, although direct effects of SES and prior achievement were very small or even negative. Coursework completion, average grades, and college plans had large direct effects on EA for almost all ethnic and gender groups. Educational aspirations were moderately associated with EA for Whites, Blacks, and Hispanics. All other factors had either no relationship with EA or small effect sizes on EA. Large differences were evidenced between Asian males and Black males in structural paths from parental expectations and educational aspirations to EA; between Black males and females from college plans to EA; between Hispanic males and females from parental involvement to EA; and between Asian males and females from peer influence and test scores to EA.