A Study of Item Response and Anchor Bias in Economics Objective Tests among Senior Secondary School Students in Osun State

Abstract

The study investigated the extent to which test option keys exhibit middle bias, the influence of primacy, recency, and middle attractions on item difficulty and discrimination in multiple-choice items with a view to improving the quality of objective test. The population for the study consisted of Senior Secondary School three (SSSIII) students in Osun State. The sample consisted of 620 SSSIII Economics Students randomly selected from 20 randomly selected schools in five purposively selected local government areas. The research instruments used for the study were a 50-item Economics Multiple-Choice (MC) test adopted from WAEC questions with 4-option length and its adapted equivalent MC test with 5-options length. The results revealed that correct answers were placed in the middle positions in 49.2% of the times in 4-option MC test items as against 64.4% in the 5-option MC test. Also, primacy and recency effects were significant on item difficulty in 4-option MC test (?2=164.16; p<.05) and the 5-option MC test (?2=39.69; p<.05). Further, the results showed that middle attraction had significant effect on item difficulty in 5-option MC (?2 =519.50; p<.05) but had a not significant effect in the 4-option MC test (?2=3.66; p>.05). In addition, middle attraction had significant effect on item discrimination in both test formats {(?2=41.93;p<.05) (?2=135.63; p<.05)}.The study concluded that test developers tended to exhibit middle bias in placing correct keys in 5-option MC format than in 4-option MC and that key balancing was more pronounced in 4-option MC test.

DOI: 10.5901/ajis.2013.v2n3p173