Background: Previous research on small-group learning has found that level of constructive activity (solving or explaining how to solve problems using ideas stated or implied in the explanation provided by a partner) was a better predictor of post-test achievement than either a student's prior achievement or the quality of help received (Webb, Troper, & Fall, 1995). Aims: The purpose of this study was to extend this research by examining the influence of additional factors, in particular, achievement goals and comprehension monitoring, on low achieving students' constructive activity after receiving help from a high achieving peer. Sample: Thirty-two low achieving upper elementary students from an urban school district in the mid-west of the United States were paired with high achieving partners. Methods: Videotape data from a previously reported study on peer collaboration were transcribed and reanalyzed. In that study, dyads were randomly assigned instructions designed to induce either a learning or performance goal and were videotaped as they worked together to solve a set of mathematical word problems. The following day, students were individually post-tested on problems similar to the ones worked on in pairs. Results: Consistent with previous research, low achieving students' level of constructive activity predicted post-test performance. In addition, constructive activity was found to mediate the relationship between achievement goals and learning. However, achievement goals were not related to low achievers constructive use of help. Instead, achievement goals were related to low achievers' relative accuracy in comprehension monitoring, which in turn was related to level of constructive activity. Conclusions: The meaning of these results for understanding the processes by which low achievers learn from peer help and implications for classroom practice are discussed.