Cognitive, Affective, Personality, and Demographic Predictors of Foreign-Language Achievement

The ability of cognitive, affective, personality, and demographic variables to predict second-language acqui sition among college students was investigated. An all possible subsets regression analysis was used to compare the propor tion of variance in foreign-language achievement explained by each variable. The analysis revealed that variables from each of the 4 domains were important predictors of foreign-lan guage achievement. Overall, academic achievement, as mea sured by GPA average, was the best predictor, explaining 11.5% of the variance in foreign-language achievement. For eign-language anxiety, the next best predictor, explained 10.5% of the variance. The educational implications of these findings for understanding foreign-language achievement are discussed, along with suggestions for future research. Foreign-language instructors are accustomed to observ ing a wide range of performance in their courses. Whereas some students excel in learning a foreign lan guage, many students underachieve, or do not achieve their desired level of proficiency. In an attempt to understand that phenomenon, researchers have investigated a multitude of factors that may affect language learning. Although Ehrman and Oxford (1995) noted that the majority of those studies focused on cognitive variables (e.g., language aptitude, cognitive ability, study habits), affective (e.g., anxiety, self perceptions), personality (e.g., locus of control, individual ism), and demographic (e.g., age, number of previous for eign languages studied) variables also seem to be related to foreign-language achievement (Ehrman & Oxford, 1995; Gardner, Tremblay, & Masgoret, 1997). Only a few studies have examined the role of cognitive, affective, personality, and demographic variables concur rently. Gardner et al. (1997) stated that "there is a lack of research examining the relationships among all those vari ables simultaneously" (p. 344). Two studies that investigat ed the relationships of several classes of variables to for eign-language achievement are Gardner et al. (1997) and Ehrman and Oxford (1995). Specifically, Gardner et al. (1997) found substantial relationships between foreign-Ian guage achievement and noncognitive variables. Those authors noted that when foreign-language achievement in French classes is assessed using relatively objective mea sures that are administered at the same time as other mea sures, foreign-language anxiety, self-confidence, and per ceived foreign-language ability are most highly related to achievement. Unfortunately, Gardner et al. tested 34 vari ables simultaneously with only 102 participants. Because the subject-to-variable ratio was only slightly more than 3:1, less than the 5:1 minimum recommended for multivari ate analyses (Thompson, 1990), the subsequent path coeffi cients likely were unstable. Ehrman and Oxford (1995), whose students comprised a large sample of adults enrolled in intensive training lan guage courses at the Foreign Service Institute, reported that cognitive variables exhibited the strongest correlations with foreign-language achievement, followed by affective fac tors, then personality variables. However, because those authors reported only zero-order correlation coefficients, the impact of each variable was considered in isolation. As such, it is not clear how each of the variables relates to for eign-language achievement in the presence of other factors. Moreover, no comparisons were made with respect to the proportion of variance explained in foreign-language achievement. That consideration is potentially important because, as Pedhazur (1982) noted, "Inspection of zero order correlations is not sufficient to reveal the potential usefulness of variables when they are used simultaneously to predict or explain a dependent variable" (p. 104). Thus, the purpose of the present study was to examine further the role of various cognitive, affective, personality, and demographic variables in foreign-language achieve ment. That is, we investigated which variable domains best predict second-language acquisition among college stu dents by comparing the proportion of variance in foreign Address correspondence to Anthony J. Onwuegbuzie, Depart ment of Educational Leadership, College of Education, Valdosta State University, Valdosta, GA 31698. (E-mail: TONWUEGB