Comparing Some Methods of Split-Halve Reliability

Abstract

The current research aims to Comparing Some Methods of Split-half Reliability. To achieve the research objective, the researcher adopted the intelligence test (Alfred W. Munzert, 1994), which is published by the American institution (Prentice Hall) and consists of 60 items. The researcher then prepared the test according to the procedures followed in the development of psychological tests and presented it to 17 experts to verify the test's apparent validity. The experts provided their feedback on the test items, leading to all necessary modifications and the removal of 4 items. Subsequently, the researcher administered the test to a pilot sample of 60 male and female students in the intermediate stage, paving the way for its application to the statistical analysis sample of 400 students, selected through stratified random sampling from the fourth and fifth grades in intermediate schools in Kirkuk Governorate during the academic year (2023-2024). The students' responses were then corrected, and the results were organized into a data table using Excel, after that, I made a comparison between the classifications (iterative Split, random split, equal difficulty, equal discrimination, equal reliability). Then, we used methods for split-half reliability (Spearman-Brown, Horst, Guttman, Flanagan-Rulon, Angoff-Feldt) and created a matrix for each classification method and each reliability method. This way, each reliability method was calculated with each classification method, and the results indicated that there were no statistically significant differences between the methods.