High stakes standardized exams are expensive and not related to program learning (instead, most researchers found they could be predicted by prior ability). The researcher used customized pedagogy – instructor-led collaborative study groups – to help students improve their scores on standardized exams. A quasi-experiment was designed to test the predictive ability of collaborative study groups on a well know standardized exam (N=162). A high degree of experimental control was imposed and the common predictors of high stakes exams (as cited in the literature) were replicated, thus lending validity and credibility to the study's findings. Descriptive statistics, correlation analysis and ANOVA with regression were applied to test the hypothesis. The ANOVA regression model was able to capture 65% of variance on MFTB standardized exam score, where the four significant predictors were (in order of importance): SAT, pedagogy, GPA and gender. The model proved that standardized exam scores could be increased more than three grade points for students in the collaborative study group.