Factors related to persistence in engineering, particularly for women, were examined in 278 upper level undergraduate women and men who, as entering students, enrolled in the college of engineering at a large midwestem university. An overall model of persistence was created by integrating constructs derived from a number of theories relevant to women's career behavior, including self-efficacy theory, expectancy-valence theory, interest congruence theory, and influences on women's career development. Hierarchical logistic regression revealed that ability, self-efficacy, support-barriers, and interest congruence each added significantly to the model predicting persistence, which correctly identified 92.6% of persisters and 623% of nonpersisters. Gender and expectancy-valence variables were not significant predictors, nor were any of the 2-way interactions with gender. Implications for future research and interventions are discussed.Women's career development has received increasing attention since the 1940s when women began entering the workforce in significant numbers. Although they currently make up over half the employees in the entire U.S. workforce (U.S. Department of Labor, 1993), women are drastically underrepresented in mathematics, the physical sciences, and engineering (Betz, 1994; U.S. Department of Labor, 1994). In 1992, for instance, only 9% of engineers, 13% of physicists and astronomers, and 11% of geologists were women (National Science Board, 1993).Given projected shortfalls of engineers, mathematicians, and scientists (American Association of University Women, 1991;Brush, 1991; National Research Council, 1991), the small numbers of women in these fields may have negative consequences for the future workforce. Women are needed not only to fill vacant positions but also to bring new questions, ideas, and perspectives to these fields and to increase the likelihood that the United States will remain competitive in technological development well into the next century (Wilson, 1992). In addition, increasing women's numbers in the sciences, mathematics, and engineering would benefit individual women, who would be that much less restricted to a range of lower status and lower paying "pink collar" occupations (Betz, 1994;Betz & Fitzgerald, 1987). In fact, engineering has been identified as among the top five most lucrative potential occupations for women (U.S. Department of Labor, 1990).

Model with ability block entered first 109 Model with ability block entered last 110 Examining the model for parsimony 110 Logistic Regression Analysis using Pre-Choice Measures RESULTS 113 Persistence Status 113 Sample Representativeness 114 Questionnaire Form 115 Preliminary Analyses 116 Scales with Experimental Items 116 Math Self-Efficacy Courses Subscale (MSE-COURSES) 116 Outcome Expectations (OE) 117 Classroom Environment Scale (CES) 118

No abstract