This study was designed to ascertain whether reading achievement and attitude toward reading of students in grades two, four, and six can be improved through implementation of a token system within regular reading programs. Two schools were assigned to the experimental and two to the control treatment, with an equal number of classrooms from the three grades in each treatment group. All subjects (Anglo and Mexican American students of lower-middle and lower socioeconomic levels) were pretested in speed and accuracy, vocabulary, comprehension, and attitude toward reading. A reward system of tokens redeemable for reinforcing events was implemented in the reading programs of the experimental groups. The control group received conventional reading instruction without the token system. Analysis of posttests on reading achievement and attitude toward reading showed that the token system was effective for all indices of reading achievement but not for reading attitude. (Author/JM)