This paper uses fresh data to confirm and extend our original findings (Britten and Heath, 1983) which were criticized by Goldthorpe (Sociology, November 1983). Given that the manual/non manual divide has little relevance for women's jobs, we demonstrate that office work and sales work entail very different conditions of employment for women, although the market situation of both groups is inferior to that of men in manual work. The career paths for women across three distinct labour markets (semi-professional, office and unskilled) demonstrate women's attachments to particular kinds of work in spite of interrupted employment histories. It is concluded that the intermittent nature of women's employment does not detract from our approach. In fact we show that women's qualifications have a more significant association with their careers than do their husbands' class positions. Using a new class schema, we confirm the importance of women's jobs in explaining their voting and fertility behaviour.