'Race' and ethnicity are widely studied as explanatory variables in health research and their apparent independent effects are reported to explain many health variations between different groups within and between populations. Both, however, are associated with major problems of definition and confusion of socio-cultural and biological concepts which call into question the validity of many of the 'independent' effects attributed to them. This paper seeks to unravel the basis of these concepts and explore their scientific validity using examples from the child health literature.