Abstract

The present study was conducted to explore the reliability and the validity of the Raven Progressive Matrices Test as an intelligence test in Koreans. The booklet form of the Standard Progressive Matrices (SPM) was administered to 405 college students (202 males, 203 females). Among them, 100 volunteers were also tested with the KWIS. The mean of the SPM scores was 52.5(s.d.=4.0). The sex difference was not found. The estimate of the test-retest reliability was .62 and the internal consistency (Cronbach's alpha) .73. The construct validity of the SPM was investigated by factor-analyzing the 5 set scores. A single factor that explained 39.2% of the variance was extracted from the 5 sets, indicating the high internal consistency of the SPM. The fact that there were significant correlations between the SPM and KWIS's Full, Verbal, and Performance Scale IQs indicates that the SPM is a valid measure of intelligence. That the Vocabulary subtest having the highest correlation with Verbal Scale IQ was barely correlated with the SPM, however, suggests that both the SPM and the vocabulary scale should be considered together for assessing general intelligence(g). According to the item analysis based on the item response theory, it was found that the SPM consists of too many easy items and the average "guessing" was in the slightly high range, while the average "item discrimination" was in the moderate to high range. In surn, the present results indicate that the SPM is a reliable and valid measure of intelligence, although these results should be validated by the future studies that will be conducted to other aged normal and patient groups.