The effects of play-like, verbal-feedback, and nonverbal-feedback testing conditions on three creative ability measures and the reliabilities of those measures were studied using 175 lower-class preschool children. The creativity measures were found to be largely unaffected by variations in testing conditions; but at the same time, different forms of the measures hypothesized to be parallel proved to have only small relationships with each other. However, split-half reliabilities of the individual measures were fairly high. (Author)