Gadfly is pleased to report the results of a recent study, to be published in Psychological Science this fall, comparing the effectiveness of "direct instruction" (where teachers actually teach, rather than observe or facilitate) and "discovery learning" (where children are given certain materials and are expected to "discover" scientific principles on their own) in science classrooms. It turns out that students exposed to direct instruction were far likelier to be "able to design at least three out of four experiments without confounds" and were better able to critically evaluate "deeply flawed experiments." To nobody's surprise, fans of discovery learning are struggling to discredit the research, arguing that the type of discovery learning studied is rarely used in the classroom anymore. But, at least one critic, Psychologist Richard Shavelson of Stanford University, admits that the study "uses a strong research design" and says he'd like to see it replicated with the "more typical" guided discovery that many teachers use.
"Instruction versus exploration in science learning," by Rachel Adelson, Monitor on Psychology, June 6, 2004