The Structure and Measurement of Self-Construals: A Cross-Cultural Study of the Self-Construal Scale
Miramontes, Lilia G.
MetadataShow full item record
The cultural dimension of individualism-collectivism (I-C), and how best to measure it, continues to be controversial in cross-cultural psychology. An important aspect of I-C at the individual level is self-construals. Markus and Kitayama (1991) distinguished independent and interdependent self-construals and Singelis (1994) adopted this two-dimensional representation of self-construals in developing the Self-Construal Scale (SCS). Although the SCS has been translated into many languages and widely administered across cultures, few researchers have systematically tested its psychometric properties or the generalizability of its structure across diverse cultures. The purpose of the present study was to examine the psychometric properties, dimensionality, and cross-cultural structural equivalence of SCS. Archival data from six studies that used the 30-item version of the SCS was analyzed. Participants included college students from two individualistic countries, the United States and Australia, and four collectivistic countries, Mexico, Philippines, Malaysia, and Japan. Hypotheses were: 1) A hierarchical model, with multiple independent and interdependent facets, will provide the best structural representation of self-construals as measured by the SCS and will replicate fairly well across cultures, and 2) Gender differences in selected aspects of self-construals will be identified. In particular, in all cultures, it was expected that females would endorse relational interdependence items or scales more than males, and that males would endorse assertiveness items or scales more than females. Structural analyses were conducted using Comprehensive Exploratory Factor Analysis (Hardin et al., 2004) and confirmatory factor analysis. Hypothesis 1 was partially supported. The structure of self-construals, as measured by the SCS, was better represented by multiple specific factors than by broad independent and interdependent dimensions (i.e., a two-dimensional model). Also, the pattern of correlations among the specific factors was not consistent with distinct higher-order independent and interdependent dimensions. The psychometric properties (i.e., reliability, item-total correlations) of the SCS were marginal in most countries and the structure of the SCS did not replicate well across most cultures. Hypothesis 2 was rejected because gender differences for relational interdependence and assertiveness were not found. The results of the study have implications for both theory (e.g., the structure of self-construals) and measurement in cross-cultural research.