In the original Swansea study respondents’ occupations were classified according to the classification system used in the decennial census, Registrar General's Social Class. This was designed to produce six ‘social classes’ which are similar to the SEGs/occupational groups in Table 2.3. Thus, classes I & II correspond to the uppermost of the three classes in that table, classes IIIM, IV and V (skilled manual and manual supervisory, semi-skilled and unskilled) correspond to ‘Manual’, and IIIN (defined as Non manual but not in I or II) corresponds to ‘other non-manual’. For purposes of comparison we had to follow this procedure when coding occupation in the 2002 survey. But we then ran up against the problem that, in the original study, married women were not classified. This was because it was assumed that the occupation of the head of household determined the social honour and life chances of its members and therefore the classification of the household. Obviously, under contemporary cultural and labour market conditions it was inappropriate to use the Registrar General's male-based system of classification, but to abandon it would have made comparison between the 1960 and the present surveys problematic. In the 2002 survey women's occupations were therefore coded twice; one coding (called ‘male-based’) assigned a woman's male partner's occupation to her even where she already had an occupation of her own (as would have been done in 1960); the other coding simply coded the woman's occupation without reference to her male partner's occupation.

We then had further problems in comparing our survey data both with the 2001 Census and with the original study. We could not compare the Registrar General's Social Class distribution of the 2002 survey with the 2001 Census because this variable had been abandoned by ONS. We could not straightforwardly compare it with the 1961 Census because that classified males only. We could not compare it with the full 1960 survey male-based distribution because Rosser and Harris (1965) does not contain full data. What it does contain, however, is a collapsed version of their RGSC 1960 data to form three occupational categories by combining classes I & II and IIIM-V, leaving IIIN on its own. Rosser and Harris called these three categories Managerial, Artisanal and Clerical respectively (1965, 102).