Audience response systems (ARS) are effective tools for improving learning outcomes and student engagement in large undergraduate classes. However, if students do not accept ARS and do not find them to be useful, ARS may be less effective. Predicting and improving student perceptions of ARS may help to ensure positive outcomes. The present study expands on previous work by simultaneously evaluating several student and practice variables that may predict student perceptions of ARS. Perceived learning outcomes, engagement, interaction in class, enjoyment of ARS, attendance, and preparation for class were examined using multivariate regression analysis in two undergraduate samples. Student variables, including gender, grade, and year of program predicted student perceptions of ARS. Practice variables, including the consistency of ARS use, question type, question difficulty, and ease of use also predicted student perceptions. The nature and implications of these findings are discussed. (Contains 3 tables, 1 figure and 2 notes.)