The purpose of this study was to investigate the causal relationship between gender, prior math performance, perceived sources of mathematics self-efficacy, goal setting, mathematics self-efficacy, attributional styles, and math performance of high school student in Ahvaz, Iran. The sample for the study consisted of 390 sophomore students (192 boys and 198 girls) who were selected randomly from the high school population in Ahvaz, and had passed math 2 (prior math performance), and were studying math 3 (math performance. Participants were administered the Perceived Sources of Mathematics Self-Efficacy Information Scale (PSMES), the Mathematics Self-Efficacy Scale (MSES), the Attributional Style Questionnaire (ASQ), and the Goal-Setting Scale (GS). Also, the prior and current math grades were obtained from high schools' archives. In addition to descriptive statistics (mean, standard deviation and Pearson correlation coefficient), the following statistical methods were used to analyze the data: factor analysis, and structural analysis. The primary research question was whether the main model (the relationship between gender, prior math performance, perceived sources of mathematics self-efficacy, goal-setting, mathematics self-efficacy, Attributional styles, and math performance) fits the data. In other words, whether the above variables influence math performance. The results indicated that the main model fits but one path (from goal-setting to math performance) was not significant. Also, the path coefficient from mathematics self-efficacy to math performance was negative. The non-significant path, then, was eliminated from the model, and a modified model was formed. The results of testing this modified model showed that it was a better model over the previous one. In short, from nine direct hypotheses only one was not confirmed, and one of them was confirmed in the opposite direction.