It was hypothesized that the use of fading in response items of stimulus-response pairs constructed to simulate learning of a foreign language vocabulary would result in fewer errors than would the use of regular flash cards used in typical trial-and-error fashion. Subjects were college women volunteers randomly assigned to one of two experimental or control groups. One half of the subjects studied word pairs using a typical flash card technique; the other half used a modified flash card technique, designed to fade the response member. A 2x2x10 mixed analysis of variance was performed on the data. On the test trials, trial-and-error subjects showed better performance, i.e., fewer errors, than did fading subjects. The learning curves crossed so that control subjects appeared to be superior during test trials. The authors believe this was an artifact of the methodology used in the study and the scoring method. The study is being replicated with some procedural changes and a finer scoring method. The experimental subjects will be given feedback with the complete stimulus-response item on each learning trial. Partial correctness will be scored instead of an all-or-none score. (DWH)