Abstract
Fifty middle-grade pupils conducted experiments on the effects of gravity, either as members of small groups who were taught skills of self-directed critical thinking (Experimental 1; E-1) or as members of a class taught by their regular teacher (Experimental 2; E-2). Sixteen other children served as controls (C), studying neither the science nor the thinking skills. Experimental 1 and Experimental 2 pupils scored significantly higher on a test of relevant science content than did the control children, but their achievement was not different from each other. When E-1, E-2, and C children were individually given new science problems to solve, E-1 pupils outperformed E-2 pupils and C pupils as predicted. Somewhat unexpectedly, E-2 children performed better on the new problems than did the uninstructed C pupils.