This study investigated the performance of graduate students enrolled in introductory statistics courses. The course in Fall 2005 was delivered in a traditional face-to-face manner and the same course in Fall 2006 was blended by using an online commercial tutoring system (ALEKS) and making attendance of several face-to-face classes optional. There was no significant difference in the t-test comparing performance in the courses, which used the students' combined score on two mid-terms and the final exam to indicate performance. The ANCOVA analyzing influences on performance in the blended class yielded no significant influence for gender, ethnicity, age, or class type (traditional vs. blended), but a significant influence from students' incoming GRE-Quantitative score. Seven Likert questions on students' perception of blended learning were not correlated with student performance. Three focus groups--comprised of low-, medium-, and high-performing students--revealed three themes and several subthemes and differences based on students' performance level. (Contains 4 tables.)