Accuracy of same-subject estimates: Are two judgements better than one
ERIC Clearinghouse on Assessment and Evaluation
EDRS No. ED451237. Paper presented at the European Meeting of the Psychometric Society (11th July 1999, Luneburg, Germany).
The accuracy of the mean of two estimates was compared with the accuracy of a single independent estimate from the same subject. A subject was asked to estimate the size of one attribute of a constant stimulus, e.g., the total of a set of numbers. The same subject was also asked to give an estimate for an upper and lower bound on the size of the same attribute of the same stimulus. The experiment was designed to ensure that the single and double estimates were independent and given by the subjects under the same experimental conditions. The experimental design compared the accuracy of estimates of two stimulus attributes using a 3 by 4 randomized block design. Participants were 187 college students competing for $10 prizes with the time severely limited. Only 50 subjects managed to complete the tasks. The frequency with which one estimate was either extremely high or extremely low suggests that the levels of task complexity were too high for the stress level of the timed competition. The results have potential ramifications for methods of collecting judgmental data, but future research should use a task of more appropriate complexity.....