Cross-Cultural Studies of Self-Esteem, Self-Differentiation and Stress in School Students

Article excerpt


For too long psychology has been dominated by North American concepts and models, which often ethnocentrically assume that the world, psychologically speaking, is like America, and that US psychological measures and concepts can be exported and applied in other cultures. Counter to these ideas, the discipline of cross-cultural psychology has emerged, distinguishing emic and etic concepts.These concepts apply respectively to psychological concepts, measures and processes which are more or less specific to particular cultures (emic), and to culturally universal concepts which have psychological validity for human beings in a variety of cultural contexts (etic).

The concept of self-esteem (an affective appraisal of self and self-concept (a combination of affective appraisal and cognitive descriptions of salient features of oneself) has developed in both dynamic psychology (with European origins and American applications) and in the social psychology of interaction and self-perception arising from the American work of William James and G. H. Mead. In younger children selfesteem is fragmented, but by adolescence a more integrated or global appraisal and understanding of self-characteristics emerge, and can be an important predictor of future success or failure in academic and other areas (Young and Bagley, 1982)

The most comprehensive demonstration of this comes from the work of Kaplan (1980), who obtained data on the Rosenberg Self-esteem Scale (Rosenberg, 1966) for several thousand children entering junior high schools in Texas. Self-esteem at age 12 or 13 was a strong and consistent predictor of a variety of social pathologies, including suicidal ideas and actions, drug use, school failure and drop-out, and unwanted pregnancies (Kaplan, 1980). The applicability of this model to the situation of disaffected youth in Hong Kong (e.g. those joining Triad societies) has been demonstrated by Leung and Lau (1989).

Cross-cultural studies of various self-esteem measures (measuring negative and positive appraisals of self characteristics, and performance in various social roles) indicate that such measures are normally distributed. These studies (e.g. in Hong Kong and the Philippines) indicate that measures of self-appraisal have social and behavioural correlates which indicate construct validity (Church,1987; Bagley et al., 1996; Chan and Lee, 1993; Chan and Verma, 1995).

One problem in using Western constructs and measures imported into Asian countries is that cultural factors may have a strong influence on how individuals evaluate themselves and present themselves to others. Researchers comparing the responses of Japanese, Chinese and Filipino subjects have observed that those populations tend to devalue themselves, in a form of psychological deference, when presented with self-appraisal questions by older individuals of higher status (Markus and Kitayama, 1991).

We have argued that although many Western psychological instruments can be used in Eastern cultures, they cannot be used for comparing one cultural group with another, e.g. it would not be meaningful to compare the mean scores of British adolescents with those of Indian adolescents on a self-esteem measure (Bagley and Verma, 1982). We have found, for example, that Canadian high-school students, on first or superficial examination, have significantly 'higher' scores on the Coopersmith self-esteem measure than British students (Bagley, 1989); Chinese students have 'lower' mean scores than either British or Canadian scores (Bagley and Mallick, 1995); whilst, comparatively speaking, Japanese students have extremely 'low' scores on the self-esteem measure used (Bagley et al., 1983).

While it may not be meaningful to argue that students in one culture have 'better' self-esteem than students in other cultures, it may well be a valid exercise to compare the factor structure of instruments between cultures and correlations which indicate construct or behavioural validity (Bagley and Verma, 1982), since mean values for a scale do not usually influence correlations and factor structure, provided that there are some basic, or etic, similarities between the cultures. …


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.