Discriminant effects of performance-oriented and structure-oriented mathematics tasks on achievement growth

Abstract

Studies concerning the superiority of low-level or high-level teacher questions for mathematics achievement growth show conflicting results. We suggest that this unsatisfying state of affairs can be overcome (1) by using a modified classification system for low- and high-level questions or, more generally, for tasks (“performance-oriented” vs “structure-oriented”) and (2) by assessing discriminant effects of these two types of tasks on different aspects of achievement rather than by assessing general superiority of one type. We tested these ideas in a field study (33 third-grade classrooms and their teachers) in which we employed low-inference classroom observations to record the teachers' use of different kinds of tasks and then looked at mathematics achievement growth as a function of task type. The data confirmed our expectations. The theoretical and practical implications of our results and the restrictions of the present study are discussed.