Recent research has uncovered significant concerns about the validity of some types of college student self-reports. This study examines the extent to which student reports about a critical type of college experience--good teaching practices--may be biased as a function of students' intellectual orientations and cognitive reasoning abilities. Perceptions of instruction and instructional practices are especially important in higher education, given their increasing use for institutional quality assurance, as well as faculty rehiring and promotion processes. Using a large, multi-institutional, longitudinal dataset of first-year students, this study shows that several cognitive indicators predict perceptions of six different sets of good teaching practices and that these relationships do not seem to be explained by actual differences in students' experiences. Additional analyses indicate that halo effects, in which global evaluations of instructor quality and institutional satisfaction affect students' perceptions of their engagement with good practices, may partially explain these findings. The results provide important implications for practice and research related to college student survey data, including ways that these biases can be reduced or eliminated to more accurately capture students' engagement in good practices and the factors that may contribute to students' perceptions of their environment.