Abstract
Three groups of university students were exposed to the same computer‐administered programmed instruction in numerical differentiation with different degrees of access to the output from the typewriter terminal. Analysis of covariance showed no significant difference on post‐test scores between students who were allowed to keep the output and those who were not, nor between those students who could look back during the session at previous output and those whose view was restricted to the most recent output. This evidence lends partial support to the conclusion that character display terminals for computer‐administered programmed instruction are not necessarily enhanced by the provision of a feature for the production of hard copy.