Based on strong research literatures, we conjectured that social processing of feedback by cooperating in a small group setting--with social incentives to ask questions, give explanations and discuss disagreements--would increase learning. We compared group and individual feedback, using two technologies: (1) Technology-mediated, Peer-Assisted Learning (TechPALS), which uses wireless handheld technology to structure feedback in small groups as they solve fractions problems and (2) a popular desktop product, which provides feedback to individual students as they solve fractions problems individually. Three elementary schools participated in a randomized controlled experiment conducted in the 2007-2008 school year. Students in the TechPALS condition learned more than did the control group students, with effect sizes ranging from d = 0.14 to d = 0.44. Analysis of observational data confirmed that students in the TechPALS condition participated socially in questioning, explaining, and discussing disagreements, whereas students in the individual condition did not. We conclude that an integration of technology, cooperative activity designs and broader educational practices can lead to impact on students' mathematics learning.