Abstract
Dugard and Todman (this issue), in this journal, present convincing arguments for the greater utilisation by educational researchers of the analysis of co‐variance when analysing data from experiments employing pre‐test‐post‐test control group designs. We applaud this initiative, but caution against too readily employing co‐variance procedures where it may be inappropriate to do so and repeat some of the cautions concerning ‘pitfalls for the unwary’ previously voiced (Lambourne & Wheldall, 1979).