Abstract
In a typical school‐organised experiment comparing programmed and conventional instruction there were significant differences between the attainments of the two groups. Differences were present at pre‐test and post‐test which suggested that the programmed instruction group had overtaken the conventional group. Further investigation of the data revealed that the two groups had learned different things and that their test scores could not be justifiably compared since the post‐test/ pre‐test regressions were significantly different.