Measures of change in token-economy programs

Thirty-nine patients, mostly schizophrenics, on a token-economy ward were available to attend up to 30 occupational therapy (OT) sessions divided into 5 reinforcement conditions: free, patients paid, free, patients were paid, free. Four groups of patients were identified according to the condition and session in which they first responded. Group 1 patients began to attend without reinforcement by the third free session. Group 2 patients who began to attend by the second session in the second free condition were less sensitive to rewards than Group 1 members but possibly modeled their behavior after Group 1. The Group 3 members began to attend when given token reinforcers. Group 4 members never emitted the required response. Inadequacies of the grouped-data statistics used by previous researchers in evaluating token economy programs are discussed and it is suggested that in evaluation procedures patients should be separated into groups and the responses from each group be presented in cumulative curves.