In this article, we report the findings of research that was designed
to identify factors associated with learning mathematics in groups wit
h computers.(1) The study was influenced by different theoretical pers
pectives on social interaction and learning mathematics, and we descri
be how we attempted to integrate these approaches into the research de
sign. To cope with complex data based on eight groups of six students
(aged 9-12 years), we developed a methodology that involved moving bet
ween quantitative and qualitative approaches in an iterative spiral. I
n this article, we focus on the patterns of learning associated with t
wo group tasks incorporating the use of Logo. Quantitative analysis of
learning measures indicated positive learning gains as a result of th
e groupwork, with no differences across gender or ability; qualitative
and quantitative analysis of process factors pointed to explanations
for the differing profiles of success across groups. Although balanced
coconstruction at the computer, coupled with the coordination of othe
rs' perspectives, was most advantageous for learning conceptually base
d mathematics, this was not the case with ''technology-driven'' mathem
atics, where fragmentation and concentrated work at the computer prove
d to be more efficient. These findings suggest that detailed specifica
tion of the learning goal is crucial when evaluating groupwork within
educational settings.