Previous studies on cognitively activating instruction have mainly shown positive effects on students’ performance. However, the impact of such learning opportunities is rarely explored among educationally disadvantaged learners. Hence, this article focuses on the use and effect of cognitively activating instruction with students in non-academic track schools. We examined (1) whether there is a correlation between the observed cognitive activation and the self-reported cognitive activity of the students, and (2) whether the degree of cognitive activation and self-reported cognitive activity could explain the learning gains of the students in the videotaped lessons. For this purpose, cognitive activation was assessed in 60 videotaped mathematics lessons using high-inference rating items; the corresponding eight grade students (N = 608) were asked about their cognitive activity in the videotaped lessons and their performance was assessed in a pre-post design. Multi-level analyses predicted relevant factors for cognitive activity and performance of the students. Against our expectation, the results revealed that cognitive activation and cognitive activity are not related. Nonetheless, the findings confirm the assumption that students’ performance is predicted, among other things, by the extent of cognitive activation moderated by pre-test scores. Implications for further research are discussed.