The limits to personal self-categorization on key demographics.

Any research survey is likely to apply some set of rules in order to classify respondents according to age, gender, social class, ethnicity, disability, and so on. But the rules governing how people are categorized may be less consistent than we suppose. In respect of gender and ethnicity, there is an unspoken assumption that the researcher should accept unquestioningly the category with which the data subject self-identifies. But in respect of social class, an equally important aspect of identity, respondents are content to have categories applied externally, for example, on the basis of their occupation. The article argues that rather than seeking to achieve a supposedly “correct” classification for each data subject, it might be better to apply at least two categorizations, one reflecting a person’s own self-identification, the other on how they are socially perceived. These different forms of identification are relevant to different types of research application. [ABSTRACT FROM AUTHOR]