Abstract: | Previous studies suggest that group work would yield conceptual gains in physics if tasks specifically promoted dialogue about the factors which determine observed outcomes. This might be done by constraining task items to present critical tests of the effects of individual factors; and, additionally, by requiring participants to generate or select explicit rules about those effects. In order to test this hypothesis, groups of primary school children were asked to work on one of four versions of a task concerning object flotation: Random (i.e. a control in which there were no task constraints), Critical Tests, Rule Generation, and Rule Selection. Learning was assessed via change between a pre-test and two post-tests. Although the effect was delayed, children from the Random and Rule Generation conditions showed an increased awareness of relevant factors which was strongly related to prior group activity. This gain was absent in the Critical Tests and Rule Selection conditions. The results indicate that group dialogue can be productively shaped by task design, but also that its effects may take time to appear, and are negated by overmuch external direction of attention. |