Matriculation and Psychometric Examinations in Mathematics and English, and the Relationship between Them

Shlomo Yitzhaki¹, Taina Pudalov², Aviel Krentzler²
(1. Department of Economics, Hebrew University, Jerusalem, Israel; 2. Central Bureau of Statistics, Israel)

Abstract: The contribution of this study lies in the research method, which enables to see whether two exams are examining the same abilities. The methodology can be described as reincarnation of variants of “Abstract Measurement Theory”, and “Item Characteristics Theory”. We examine the correlation between the scores of examinees on matriculation and psychometric exams. If the correlation is monotonic throughout the set of the scores, we can conclude that that the exams test the same abilities. However, if the sign of the local correlations change throughout the set of scores, we can conclude that the tests examine different fields of knowledge, or that the characteristics needed for success in one exam are different from those needed for success in the other exam.

Using this criterion, it can be concluded that the matriculation and psychometric examinations test indistinguishable abilities, but there is a large random error. The contributions of both exams to the total score for each student support the hypothesis that the existence of two examinations does not have a substantial impact. This conclusion applies all the more to the English exams, where the correlation between the matriculation and psychometric scores is higher than in the mathematics exams.

Key words: matriculation examination, psychometric examination, monotonicity