Multiculture-Multiresponse Matrices of Correlations as a Measure of Construct Validity of a Premarital and Sexual Attitude Inventory Given to Iranian and British Subjects

Abstract

Based on the assumption that for a valid culture-bound instrument, one can expect that groups from similar culture will show greater similarities in the patterns of their responses to the instrument than groups of different culture, a method was proposed to judge the construct validity of the instrument. To apply the method on actual data, a 50-item attitude inventory toward sex and marriage was given to 392 Iranian college students (199 males, 193 females). Percentages of endorsement of different responses to the items were converted to rank-ordered numbers for Iranian subjects and for a British sample of 1541 (780 males, 761 females). Rank-order correlations of different responses across the items for different samples were shown in a table that we called a “multiculture-multiresponse” matrix. With respect to this table, relationships among correlations provided evidence for the construct validity of the instrument.