Effects of Reward Schedule Changes on Children's Speech Fluency

Speech disfluency indices were obtained for 50 young normal-speaking males who told stories about 30 situational pictures under continuous and/or partial (random) reward conditions. Four groups of ten children each received 100% rewards for the first ten responses (Condition I) and either 100%, 75%, 50%, or 25% rewards for the last 20 responses (Condition II). A fifth group of ten children received 50% rewards for the first 20 responses (Condition I) and 100% rewards for the last ten responses (Condition II). The four groups having reward schedule changes tended to have lower proportions of vocal segregates and higher proportions of revisions in Condition II than in Condition I, whereas the group receiving 100% rewards in both Conditions showed only slight changes and in the opposite directions. Significant changes in total disfluency were not obtained. The three groups changed from continuous to partial reward tended to have higher disfluency indices for those responses following non-reward than for those following reward in Condition II for all disfluency categories, except revisions.