Predictors of Academic Achievement and Retention among College Freshmen: A Longitudinal Study.

Abstract

This research examines potential psychosocial predictors of freshman academic achievement and retention. College students were assessed on various dimensions, (i.e., demographics, prior academic record, smoking, drinking, health-related quality of life, social support, coping) during the first week of their freshman year, and at the beginning of the next academic year. A multiple linear regression equation predicting cumulative GPA using 10 predictors accounted for 56% of the variance in academic achievement while a logistic equation predicting retention rates was not statistically significant. The amount of variance accounted for in first year cumulative GPA (56%) represents a substantial improvement in prediction over using highschool GPA and SAT scores alone (25%; Wolfe & Johnson, 1995). However, similar to past research, some health and psychosocial variables (e.g., smoking, drinking, health-related quality of life, social support, and maladaptive coping strategies) were related to retention. This model may be used as a tool to proactively identify students at high risk for poor academic performance during their freshman year and to provide direction regarding proactive intervention strategies for maladaptive behaviors predictive of poor academic performance (e.g., smoking, binge-drinking, social support, coping). ********** The freshman year represents a stressful transition for college students (Lu, 1994). Despite a multitude of social, academic, and emotional stressors, most college students successfully cope with a complex new life role and achieve academic success. Other students are less able to successfully manage this transition and decide to leave higher education during or at the end of their freshman year. It is estimated that 40% of college students will leave higher education without getting a degree (Porter, 1990) with 75% percent of such students leaving within their first two years of college (Tinto, 1987). Freshman class attrition rates are typically greater than any other academic year and are commonly as high as 20-30% (Mallinckrodt & Sedlacek, 1987). The implications of leaving college without obtaining a degree are many. Each student that leaves before degree completion costs the college or university thousands of dollars in unrealized tuition, fees, and alumni contributions. The decision to leave college is also frequently economically deleterious to the college dropout, whose decision to leave often leaves him or her in a position to earn much less over a lifetime of work (National Center for Educational Statistics, 1989). Despite these considerable negative consequences for universities and students, attrition rates have not changed appreciably over the last few decades (Porter, 1990). This fact has provided an impetus to understand risk factors for college student attrition. If such risk factors can be identified, then intervention programs can be designed to increase retention rates (Clark & Halpern, 1993). There is a consistent relationship between college academic achievement and retention, with higher performing students persisting in their studies to a greater degree than their lower achieving cohorts (Kirby & Sharpe, 2001; McGrath & Braunstein, 1997; Ryland, Riordan, & Brack, 1994). Given the consistent relationship between these variables, it is prudent and efficient to identify common risk factors for these student outcomes in order to best develop targeted intervention programs. This study will examine a number of potentially predictive variables of academic achievement and retention that have been examined in prior studies and also examine some new potential risk factors, such as student health status, that have yet to be examined. The goal of this research was to create a multidimensional risk model that would optimize prediction of both academic achievement and attrition. The following is a brief literature overview and rationale for the inclusion of certain variables in this comprehensive risk model. …

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 686 citations exist upstream.