Response cost and impulsive word recognition errors in reading-disabled children.

Subjects were 30 fourth grade children with average intellectual ability but reading achievement at least 1.5 years below grade level. Each child was given two word-recognition lists, the first one as a pretest and the second list under one of three different experimental conditions: control, positive reinforcement (1 nickel for each word read correctly), and response cost (1 of 40 nickels taken back for each word read incorrectly). Relative to the control condition, positive reinforcement led to a significant increase in response latency but no change in errors, while response cost led to both a significant increase in latency and a significant decrease in reading errors. The entire group was found to be impulsive on the Matching Familiar Figures test. The successful reduction in impulsive reading errors was interpreted as support for Kagan's hypothesis that the impulsive child evidences low concern about errors on such academic tasks.