In a national sample of over 3,000 36-year-olds, neurosis was more prevalent in men of lower status as assessed by housing tenure but not occupational class, education or personal income. Symptoms were more frequent in the unemployed, especially those not seeking work. In women, rates did not differ by husband's occupational class but were higher for those in rented accommodation, with unskilled manual jobs, without paid employment, with poor qualifications and with unemployed husbands, the strength of associations being dependent upon family structure. Many of these associations were reduced or eliminated by adjustment for reported financial hardship. Accounts of mental health problems in unemployed men and in married women (especially with children) may have underestimated the importance of financial circumstances.