The relationships between self-esteem, self-efficacy, and test anxiety: A cross-lagged study

The aim of the present study was to illuminate the causal relationships between self-esteem and test anxiety, as well as between general self-efficacy and test anxiety using two-wave longitudinal research design with a sample of 252 Chinese college students. After controlling for gender, grade and autoregressive effects, the results revealed that (1) self-esteem at T1 did not significantly predict test anxiety at T2; (2) general self-efficacy at T1 did not significantly predict test anxiety at T2; (3) test anxiety at T1 significantly and negatively predicted self-esteem at T2; and (4) test anxiety at T1 marginally significantly and negatively predicted general self-efficacy at T2. These results suggest that test anxiety is more likely to affect self-esteem and general self-efficacy rather than vice versa, and that the causal relationship between self-esteem and test anxiety are clearer than the causal relationship between general self-efficacy and test anxiety. However, given the limitation of the longitudinal research using cross-lagged analysis for revealing causality, these results should be viewed with caution.