Abstract

Psychological tests such as attitudinal or personality tests are often adapted or translated for use in many languages and cultures. In this situation, the construct equivalence is a fundamental issue because having equivalence of measures is a prerequisite for obtaining valid comparisons across cultural groups. To completely accept results from quantitative comparisons across groups, evidence for construct equivalence must be established. The purpose of this study is to describe a general approach for empirically investigating the construct equivalence of personality tests. The Sixteen Personality Factor (16PF) Questionnaire was administered in English to 844 American college students and in Korean to 538 Korean college students. Two statistical methods were utilized in a complementary way: (a) Principal Component Analysis, and (b) Multi-group Confirmatory Factor analysis. The results showed that the extraversion scale of the 16PF had the same factor structure across the two groups only in that it had the same number of factors and the same items on each factor. However, the results indicated that the factor loadings and error variances were not equivalent across the two groups. Practical and theoretical implications are discussed.