Test taking motivation and item performance

Abstract

[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] The effect of test taking motivation on item performance is tested. The study proposes that 1) the item performance difference between high and low motivation groups varies with item difficulty; and 2) aberrant response patterns can be associated with motivation. Statistical techniques related to item level analysis are employed to test the hypotheses. Particularly, a DIF index is used to quantify the performance difference between high and low motivation groups, and is correlated with item difficulty. A person fit statistic is used to measure the difference between aberrant response patterns and expected response patterns, and is correlated with motivation indices. Results partially support the hypotheses. The higher the motivation, the more likely examinees ignore the item difficulty and have optimal performance. The lower the motivation, the more likely examinees perform under the influence of their perception of item difficulty. Motivated examinees are less likely to make random errors causing aberrant response patterns, as compared with less motivated examinees. However, the magnitude of motivation effects on item performance differs by the indices chosen.