Robinson (1963) demonstrated that both specific and vague feedback about performance on a time-perception cask led to improved accuracy in time estimation. Feedback was given on all trials. The purpose of the present study was to examine the effectiveness of specific partial feedback on a time-production task for 42 female undergraduates who were enrolled in a class in introductory psychology. A General Electric timer was used to measure the subjects' production of time intervals. The timer was connected to a switch which the subject activated and deactivated to mark the boundaries of the assigned time interval. The subjects were randomly divided into three groups of 14. One group received feedback after each trial, the second group received feedback on 50% of the trials and the last group served as the control condition, receiving no feedback on any trial. After each feedback trial, the subject was told the interval they produced to the nearest 0.1 sec. Each subject was tested on 30 trials in five blocks of six trials. Within each block the subjects were asked to produce the following time intervals: 3, 7, 11, 15, 19, and 23 sec. The order of presentation of these intervals were randomly determined for each block As suggested by Hornstein and Rotter ( 1969), ratios of produced time to actual time were computed for all the data points. For the complete feedback group the mean ratios across blocks were 0.866, 1.011, 1.059, 1.076, and 1.056. For the partial feedback group the mean ratios were .835, 1.021, 1.094, 1.061, and 1.051 and for the control group the mean ratios were 0.715, 0.821, 0.931, 0.962, and 1.074. The pattern of results for the partial and complete feedback conditions is identical; both groups produced accurate time intervals by trial Block 2 and continued this performance for the remaining trials. Similar accuracy was not achieved by the control group until the third trial block. Newman-Keuls procedures demonstrated that the estimations of the control group in Blocks 3 through 5 differed from those in Block 1 and the estimations in Block 5 were different from those in Block 2. For both feedback groups, the estimations in Blocks 2 through 5 were different from those in Block 1. All differences were significant (p = .Ol). Unequivocally the results demonstrate the feedback on 50% of the trials in a time-production task enhances performance as effectively as 100% feedback. The improvement in the no-feedback, control group may be explained by the cumulative timeproduction error (Boulter & Appley, 1967). Briefly, on a time-production task subjects tend to increase their estimations of intervals over trials. Additional blocks of experimental trials may verify this actual difference bemeen the control and experimental groups by demonstrating that the estimations of the control group increase in magnitude and become less accurate while the experimental group continues with accurate estimates. REFERENCES
Read full abstract