A STUDY OF THE EFFECTS OF THE NUMBER OF CHOICES ON THE RELIABILITY OF THE MULTIPLE-CHOICE ACADEMIC ACHIEVEMENT TESTS

The present study seeks to investigate the effect of the number of the academic achievement tests using a four-choice achievement test on Islamic theology. The tests were constructed by omitting the distracters, so two-choice or three-choice test formats were reproduced from the original. The test was administered to a population of 600 last-year high school students in Andimeshk, a southwest city in Iran. They were selected on a stratified random basis with equal ratio. The obtained results showed the reliability of four-choice, three-choice and two-choice tests items to be 0.77, 0.76 and 0.65 respectively. All of these three coefficients were significant with the probability of p ? 0.001. The results of the present research suggested that there was no significant difference between the reliability coefficients of the four-choice, and three-choice tests with the probability of less than 0.05 (p ?0.05). It also showed that there is a significant difference between the reliability coefficient of the four-choice and two-choice test, and between the reliability coefficients of the three-choice and two-choice test items with the probability of less than 0.01 (p ? 0.01).The pedagogical implication of the study could be the point that instructors who are in some way concerned about the reliability of their tests should as far as possible avoid using two choice items e.g. true false items.