The Effect of Varying the Number of Alternatives and the Location of the Correct Alternative on the Reliability of the Achievement Test and Its Internal Consistency between Classical Measurement Theory and Paizi's Statistical Model

Abstract

The research aims to determine the degree to which the variation in the number of alternatives to the objective question and the location of the correct alternative affects the stability and internal consistency of a test in the measurement and evaluation course between the Bayesian statistical model and the classical measurement theory among King Khalid University College of Education students. An objective accomplishment exam with four possibilities was created by (100) students who studied the measuring and assessment process. The questions were designed to accommodate options two and three. The JASP statistical software was utilized to apply Bayesian statistics, aiding in the execution of both conventional and Bayesian statistical tests. The researcher reached the following results: the greater the number of alternatives, the greater the stability of the test, and the differences between the value of the stability coefficient in Bayesian statistics and its value in traditional statistics were simple differences, and there was an internal consistency between the questions and the total score of the test in the case of the number of alternatives being four, and a large agreement was found. between the results of the Bayesian statistic and the traditional statistic, while on the contrary, in the case of the two alternatives, a weak internal consistency was found in the case of the traditional statistic and medium In the case of Bayesian statistics, where the correlation indicatives and their inclination towards the alternative hypothesis were more in the case of Bayesian statistics than in the case of the alternative statistics, the stability coefficient is higher in the Bayesian statistics than in the traditional statistics, and that the location of the correct alternative in the Bayesian statistics has a slight effect on the value of the stability coefficient (0.693 vs. 0.700), while it has a more clear effect in the traditional statistics (0.728 vs. 0.756), and there is a similarity in the level of significance between the traditional theory and Bayesian statistics, and all correlation coefficients were in favor of the alternative hypothesis, which indicates the presence of a significant positive correlation between the question score , and the total score of the test, with the exception of one question, was in favor of the null hypothesis, which indicates that there is no significant positive correlation between the score of the question and the total score of the test, and there is a similarity in the level of significance between the traditional theory and Bayesian statistics, and all correlation coefficients were in favor of the alternative hypothesis, which indicates There is no significant positive correlation between the score of the question and the total score of the test, except for three questions among the ten, which were in favor of the null hypothesis, which indicates that there is no significant positive correlation between the score of the question and the total score for testing.