The Development of a Test for Measuring Primary Mental Abilities Based on Thurstone’s Primary Mental Abilities for Prathom Sueksa Five Students in Schools Under the Jurisdiction of Bangkok Metropolitan Administration

Pariya Wonglek, Taviga Tungprapa, Kamontip Srihaset

Abstract

In this thesis, the researcher develops and constructs (1) a test to measure mental abilities; determines (2) the quality of the test; and determines
(3) standardized normality for the test. Research and development was divided into three phases taken seriatim: Stage One: Developed and constructed the test; Stage Two: Determined the quality of the test; and Stage Three: Determined standardized normality for the test.
Using the multistage random sampling method, the researcher selected a sample population from a research population of 4,653 students.
The data collected were analyzed using the index of item-objective congruence (IOC), the difficulty index, the discriminatory power, Cronbach’s
coefficient alpha method for testing the reliability of the test, confirmatory factor analysis (CFA), and normalized T-score. Findings are as follows:
1. The construction and development of a test for measuring primary mental abilities resulted in five sub-tests: No. One: a test for measuring
language ability; No. Two: a test for measuring number ability; No. Three: a test for measuring reasoning ability; No. Four: a test for measuring spatial ability; and No. Five: a test for measuring perceptual ability. Each test had thirty items with four choices. 2. In analyzing the quality of the test, the following was found: (1) Insofar as concerns content validity, applying the index of item-objective congruence (IOC), it was found that all items had an IOC of 1.00 and were thus congruent with test objectives. (2) The difficulty index and discriminatory power satisfied standards, since the difficulty index fell between 0.23 and 0.80 and the discriminatory power between 0.20 to 0.74. (3) The reliability levels for all five tests were 0.83, 0.78, 0.83, 0.78, and 0.86, respectively. (4) In applying second order CFA, it was found that the test as a whole showed construct validity.
3. In developing standardized normality in the form of normalized T- score for each test, the results were as follows: No. One on language
showed a normalized T-score ranging from T32.24 to T72.87; No. Two on number from T31.42 to T73.52; No. Three on reasoning from T32.82 to T74.80; No. Four on spatial ability from T31.35 to T74.41; and No. Five on perception from T32.24 to T72.61.