Research suggests that parents are effective at scaffolding their children'
s learning to help them become self-regulated problem solvers. Yet little i
s known about parents' effectiveness at assisting their children with probl
ems that parents find unfamiliar and, thus, do not have at hand either the
solution or strategies that a novice child could profitably implement. In t
his study, 20 dyads of parents and their preadolescent children spent 45 mi
n solving a scientific reasoning problem that entailed generating and inter
preting a series of experimental trials to understand the causal structure
of a moderately complex system. As the dyads worked, researchers tracked th
eir experimentation strategies, patterns of interaction, and changes in the
ir domain-specific beliefs about the system. In comparison to solo particip
ants observed in previous research, the dyads performed well on strategies
for generating and interpreting evidence. However, parents assumed most of
the difficult conceptual tasks, including recording data and making inferen
ces, and delegated the logistical roles-the actual manual operation of the
equipment-to the children. Parents did not cede the conceptual roles to chi
ldren as the session progressed. On the positive side, parents shared contr
ol of the problem solving and engaged in collaborative discussions with the
ir children. Parents also provided valuable assistance of many kinds, usual
ly during the evidence-generation phase of the trials. They missed key oppo
rtunities for helping children interpret evidence, and as a result, childre
n failed to achieve the gains in understanding that the parents did. The be
liefs of parents and children did not come into closer alignment over the c
ourse of the study. We conclude, especially as children reach school age, t
hat parents who wish to assist their children's learning may increasingly n
eed to understand how their children think-not just in general, but about c
entral kinds of problems and content domains.