We move from considering individuals and situations and expand outwards to collective groups of individuals interacting within a collective situation: culture. Immediately questions are raised about the structure and measurement of personality in different cultures. Do we find the same five
factor traits in the USA or UK that we do in, say, China? We saw in that one of the criteria that broad traits should meet is cultural universality (Costa and McCrae) – in effect, cross-situational universality of trait structure.
-->
There are a priori arguments why we might expect structural models of traits to replicate across cultures. If traits do have a biological basis, then they should be a property of homo sapiens rather than of any particular culture, although the way the biological substrate is expressed in behaviour may be culture bound.
Irrespective of biology, it is likely that different cultures face somewhat similar adaptive challenges. All people must cope with threats to well-being, form social relationships with others, obtain a livelihood, and so forth. Goldberg (1990) has loosely related the Big Five to Power, Love,Work, Affect and Intellect (i.e., E, A, C, N and O). It is likely that these five areas of life may be identified in all or most cultures, even if there are important cross-cultural differences. More generally still, Pinker (1994) has suggested thatWestern culture is rediscovering the concept of human nature. The twentieth century was characterised by what Pinker called the Standard Social Science Model (SSSM), which states that human behaviour is wholly or largely determined by culture-bound social learning (popularly, but wrongly, described as ‘conditioning’). Pinker claims that anthropologists have overstated the malleability of behaviour, and have frequently ignored similarities between cultures. If our species does have a common ‘human nature’, which may be biologically influenced, wemight expect that individual difference dimensions should show some similarities across cultures.
Nevertheless, there are potential obstacles to establishing trait universality. Cultural specificity may be strong enough to substantially alter the relative impor- tance of traits. For example, given that cultures differ in the value placed upon achievement motivation (McClelland, 1961), it might be that conscientiousness is less salient in some non-Western societies than it is in our own, or that C is not expressed through achievement striving. We can measure a distinctive ‘Protestant work ethic’ (Furnham, 1990) trait related to Western cultural values. Conversely, other traits might be more important in societies other than our own. For example, Bond (1979, 2000) discusses a ‘filial piety’ or ‘Chinese tradition’ trait found in Chinese cultures, which places high value upon respect for parents and upholding Chinese ways. In addition, a sixth factor - Interpersonal Relatedness - was obtained in factor analyses of the NEO- PI-R and the Chinese Personality Assessment Inventory (Cheung et al., 2001). Hence, we cannot, a priori, be confident that the Western ‘Big Five’ is assessing universal traits as opposed to traits that reflect the preoccupations of our culture. In addition, there are methodological difficulties in translating Western questionnaires into the languages of other cultures, because item content and differences in compliance of responding are culture bound. One of the versions of the 16PF includes items asking, variously, about interest in improvements in production and marketing, in Indian murders, in photography and in becoming a research chemist!
In the next days, we review empirical studies of the cross-cultural generality of two major descriptive frameworks for personality, the Eysenckian three and the Big Five. We focus on questions of dimensional structure. This is a distinct issue from that of cross-cultural differences in mean scores on trait dimensions, though comparison of means is sensible only if commonality of dimensional structure is established (see Lynn andMartin, 1995, for a survey of data). Triandis (1997) describes the structural approach as ‘etics’. This contrasts with ‘emics’, in which traits specific to individual cultures are identified. There has also been considerable research on generality of factor structure across different groups within the same culture, such as comparisons across sex, age and different ethnic groups. Summarises studies of sex differences. Normally, factor structure is highly replicable across different demographic groups (e.g., Costa,McCrae and Dye, 1991). In the sections that follow, we concentrate on etics first, and emics second - both are important in building our understanding of personality cross-culturally.