Secondary analysis of the data available from the national sample in the Health and Lifestyle Survey allowed two samples of mothers from Social Classes III Manual and IV+V to be selected and the relationships investigated between two measures of health and a range of potential explanatory variables, including voluntary health behaviour Analysis of variance was used to establish which sets of factors were significantly and independently associated with each of the measures of health within the two samples. Health behaviour was identified as making an independent, albeit small, contribution in three out of the four analyses.