Abstract

This study examined the effects of combining two cognitive-motivational models of task engagement and achievement, the Control, Agency, and Means-Ends model (Skinner, 1998) and the implicit theories model (Dweck, 1999). Data were collected from 291 fifth graders attending three different schools that were similar with respect to racial composition, SES, and standardized achievement scores. In the Fall of 1994, children were administered the Control, Agency, and Means-Ends Interview (Skinner, Chapman, & Baltes, 1988) and the Implicit Theories for Ability Questionnaire (Dweck, 1999) both of which were modified to apply specifically to math. Teacher reports of student task-engagement were collected in the Spring of 1995. Standardized math achievement tests were administered by school personnel at the end of each school year. These were obtained by Tay from computerized school records for the end of the 1994 - 1995 school year as well as for the previous year. Analyses examined the interrelations between the implicit theories and CAM model constructs, the degree to which prior findings in the general academic domain could be extended to the math domain, and, most importantly, the degree to which combining the two models would provide better understanding and prediction of task engagement and achievement than either model alone. Findings were mixed. Most importantly, the results suggested that the combination of CAM and implicit theories predicted task engagement and achievement significantly better than either model alone, although the added contribution of using both models together was often small in magnitude. CAM and implicit theory constructs were often significant, with low to moderate magnitudes. This suggested that these models shared common variance, but were not redundant of each other. The use of aggregate CAM variables appeared to be helpful in predicting task engagement and achievement within the math domain, but the particular individual CAM variables that made unique contributions to the prediction of task engagement and achievement were not always consistent with previous findings. As a general trend, when predicting task engagement and achievement from the combination of CAM constructs, implicit theories, and their interactions, there was likely to be a significant unique main effect for the relevant CAM construct, but not for implicit theories, when predicting task engagement. In contrast, there was likely to be a significant unique main effect for implicit theories when predicting achievement scores. Interactions between implicit theories and CAM constructs were sometimes present when predicting task engagement, but not when predicting achievement scores. Overall, findings support the utility of combining different cognitive-motivational models when attempting to understand children's achievement-related behavior