How to measure national stereotypes?

BECAUSE IT IS PARTICULARLY DIFFICULT TO EVALuate the accuracy of national stereotypes, the Report by A. Terracciano et al. (“National character does not reflect mean personality trait levels in 49 cultures,” 7 Oct. 2005, p. 96) examining the relations between ratings of national character and ratings of individuals in 49 different cultures represents quite a technical achievement. Studies of stereotypes usually suggest that stereotypic beliefs contain a kernel of truth: The perceived differences between groups do in fact exist, but they are smaller than the stereotype would suggest (1, 2). Terracciano et al. instead found that, on average, there was no relation between national stereotypes and self and other descriptions. Some methodological weaknesses of their study must be considered, however. One issue is their almost exclusive reliance on college student samples. Although there is some evidence that cross-cultural comparisons between college students may generalize to broader populations (3), there is also substantial evidence that findings with college students frequently do not so generalize (4). These findings do not invalidate college student samples as representations of broader national populations, but neither do they justify assuming college students provide an acceptable proxy for the population as a whole. A second issue is whether the authors have provided a sufficient evaluation of national character. The authors reduce national character to personality traits. This ignores other potential elements of stereotype, most particularly differences in values, beliefs, or perceptions that are not adequately included in the measures used in this study. Finally, Terracciano et al.’s measures of perceived national character were the mean ratings of the culture by members of that culture. Stereotypes are usually defined in terms of perceptions of the target group by outside observers. Moreover, their measure of actual national character was the mean ratings of oneself or a significant other. In other words, the measurement of national character was based on the ratings of a culture, whereas the measurement of actual character was based on the ratings of a person. The contexts of the two kinds of assessments were quite different and potentially not comparable. It is increasingly evident that context is an important contributor to outcomes on rating scales (5). There is even evidence that cultural differences by themselves can produce differences in the context of the measurement (6). A person familiar to the respondent will likely be evaluated in relation to other individuals familiar to the respondent, while a person asked to rate the culture will rate it in relation to other cultures. It is not surprising then to find that these ratings were on average unrelated to ratings of the country’s national character. It is possible that there really is no relation between national stereotypes and actual behaviors. One must wonder, however, what is the source of the variability in the ratings of cultures. Why, for example, do the German Swiss believe they are so conscientious? Even more curious is why Indonesians and Chileans accept that they are not. It seems likely that when asked to rate themselves on conscienLetters to the Editor