The Impact of the number of Alternatives on Estimating Reliability of the Advanced Progressive Matrices Test Using Cronbach Alpha and Stratified Alpha Coefficient

الملخص

The present study aimed at identifying the differences in Cronbach Alpha and Stratified Alpha values in estimating the reliability of the Advanced Progressive Matrices (APM) test scores in light of the difference in the number of alternatives.

For data collection we used Advanced Progressive Matrices (APM) test, where five different models are prepared only in the number of alternatives, after deleting false alternatives at random and thus three has five models of the test sample:( eight alternative, the seven alternatives, the six alternatives, the five alternatives, the four alternatives). All of these models of test have been applied on a group of 500 students from the faculty of law in the University of Damascus and sample distribution of 100 students for each model of the test models. And to answer the questions of the study, the statistical test (M) proposed by Haxstein and Wolen was used in its fourth case to determine the differences between the Cronbach Alpha and Stratified Alpha equations. The following results were obtained:

1. There were differences between Stratified Alpha and Cronbach Alpha in favor of class alpha when the number of response alternatives varies and also in favor of the number of alternatives.
2. There are statistically significant differences in the values of both the coefficient of Stratified Alpha and Cronbach Alpha, between the model of eight alternatives and models of six, five and four alternatives, in favor of the model of eight alternatives.

The results of the study reached the higher number of response alternatives, the higher values of Stratified Alpha and Cronbach Alpha, and this is consistent with the previous studies that confirmed the impact of the values of stability coefficients varying the number of alternatives to the response to the tests