Omission training compared with yoked controls and extinction in multiple-schedule discrimination learning

After rats were trained on a multiple schedule with variable interval 30-sec reinforcement in both stimulus components, omission, yoked, and extinction procedures were applied in S? while the variable interval continued in S+. In S?, omission training reduced response rate faster and to a lower terminal level than either response-independent yoked reinforcement or extinction, which were approximately equivalent. In S+, the extinction group exhibited elevated response rates representing behavioral contrast, while the omission and yoked groups showed reduced response rates. These results attest to the effectiveness of omission training as a response elimination method. They also contradict theories which posit reduction of response rate in S? to be necessary and sufficient for behavioral contrast.