English abstract
The current study aimed to compare the effectiveness of the five person fit indices (WSR, USR, LZ, WT, and UT) using the numerical ability test data of the Gulf Scale for Multiple Mental Abilities. The study sample was 4206 students from all Gulf countries. Appropriate statistical methods were used to verify the assumptions of the 3 parameters model, and item fit of the numerical ability test to the 3 parameters model. Means, standard deviations, kurtosis and skewness were computed to examine the statistical characteristics of the frequency distribution for each index. The proportions of individuals with non-fitting responses were also computed for each index. The study also examined the validity of classification of individuals into fitting/non-fitting for each index using the classification of two judges. WSR, and Ly were the most valid indices with relative to other indices. The two judges classified the non-fitting persons into six aberrant patterns (sleeping, plodding, extremely creative examinees, cheating, guessing, alignment errors) using a new evaluation form proposed in the current study. Cheating was the most common aberrant pattern found in this data. Non-fitting responses were found more common with males than with females in all Gulf countries (Bahrain, Saudi Arabia, Kuwait, Qatar, UAE and Oman). The results were discussed in the light of the theoretical framework and previous studies, and some recommendations and suggestions related to the topic of the study were presented
