There is now a colossal research literature on the development of racial prejudice among young children (for overviews see Milner 1983; Aboud 1988). Studies first pioneered in the 1920s and 1930s and repeated in a variety of formats since then have consistently shown that children have the capacity to recognise racial differences and to develop negative attitudes and prejudices towards certain groups from the age of three onwards. While this body of work has undoubtedly played an important role in highlighting the reality of racism in the lives of young children, it has attracted a significant amount of criticism over recent years most notably from a number of social psychologists (Billig 1985, 1987; Reicher 1986; Potter and Weatherall 1987; Condor 1988) and, more recently, sociologists (Troyna and Hatcher 1992; Connolly 1996, 2001; Van Ausdale and Feagin 1996, 2001). A central focus of the criticism has been the dominance of structured, experimental designs within the research to date and their tendency to encourage rather crude and simplistic understandings of the nature and influence of 'race' in young children's lives.
Moreover, in failing to study young children's behaviour within its 'natural settings', it has been argued that the research has tended to reify the concept of 'race'. Thus rather than studying the particular processes that encourage children to develop and reproduce racial categories, those categories have often simply been taken for granted (Reicher 1986; Condor 1988). As a consequence, as Billig (1985, 1987) has argued, such work has also tended to reproduce the view that racial prejudice is natural and inevitable. Without an understanding of social context, much of the research has tended to explain the development of young children's racial prejudices as an inevitable result of normal perceptual processes of categorisation. What is needed instead, Billig contends, is for this way of thinking to be turned on its head. Thus rather than examining how people's 'natural' tendency to categorise leads on to racism, researchers should be focusing on how racism, and the particular structures and relations associated with it, lead people to then categorise others in certain ways. This is precisely what some researchers have recently begun to do (see Holmes 1995; Van Ausdale and Feagin 1996, 2001;