There were two author errors (one terminological and one relating to data) in this paper by Dr Eachus and others (1999;53:603–11). Firstly, the authors referred to the index relating socioeconomic position to New Zealand score of severity of hip disease as the relative index of inequality, whereas the statistic presented is actually the slope index of inequality. Secondly, a programming error led to miscalculation of the correlation coefficients and slope indices of inequality presented in table 5. A corrected table is presented here. The direction of associations is the same as for the incorrect results presented in the original paper, but the effect sizes and significance level are both substantially greater when the correct data are seen, in particular for the associations of social class and Townsend deprivation score with hip disease severity. In the light of the correct data the discussion that was included on why the social class association was weak is no longer applicable.