Surveys of smoking in the United Kingdom point to an increasingly strong association between cigarette smoking and socio-economic status among both men and women. They have uncovered a clear class gradient in smoking prevalence, with high socio-economic status associated with low smoking prevalence. These surveys rely on a particular set of procedures to determine the socio-economic position of women. The paper enters a caveat about these procedures. It reports on re-analyses of national data which suggest that the relationship between women's smoking and socio-economic status is more complex than that revealed through conventional measures of women's socio-economic position. In these re-analyses, women in some non-manual categories emerge with smoking prevalence rates that match those found among women in manual groups. The patterns uncovered in these re-analyses point to dimensions of women's smoking behaviour which are currently obscured in the surveys which inform health promotion and tobacco control policies in the UK.