Abstract
Children of 8 to 10 yr. received 20 individual remedial reading lessons each. Four groups (ns = 4) were given token reinforcement with either material or nonmaterial back-up reinforcers and with either a response-cost contingency for errors or not. A fifth group received equal lessons but no tokens, and a sixth had no extra lessons (ns = 4). The remedial program had a limited effect on reading achievement; token reinforcement, however, did not enhance the effect of extra lessons. Material back-ups were associated with less favourable ratings of the lessons than were nonmaterial back-ups. Deducting tokens for incorrect responses led to fewer errors when the token system was in operation but more when it finished compared with having no response-cost contingency. Results are taken to demonstrate the value of control groups, and the complex operation of “reinforcement” in human cognitive activity is discussed.
