A case-base study in a population of 214,108 commercial and clerical female workers in Denmark during the period 1983–1985 investigated whether increasing job stress, defined as increasing job demands and decreasing job control, increased the risk of an adverse outcome of pregnancy. Information on the cohort, the outcome of 24,362 pregnancies, was obtained by linkage with nationwide health registers. Six case groups were selected: 1) 2,248 spontaneous abortions, 2) 209 stillbirths or deaths within the first year of life, 3) 661 infants with congenital malformations, 3) 593 preterm deliveries, 4) 587 infants with term low birth weights, 5) 988 infants with light-for-date birth weights, and 6) a random sample of 2,252 pregnancies that constituted the reference group. Information on exposure was obtained by mailed questionnaire. When dichotomized scales on data concerning job demand and control were used, there was an increased relative risk of spontaneous abortion (odds ratio = 1.28, 95% confidence interval (Cl) 1.05–1.57) and term low birth weight (odds ratio = 1.46, 95% Cl 1.05–2.04) for women experiencing high job stress. For the other case groups, the odds ratios were as follows: congenital malformation, 1.23 (95% Cl 0.93–1.63); preterm delivery, 1.03 (95% Cl 0.77–1.39); light-for-date birth weight, 1.08 (95% Cl 0.83–1.40); and stillbirth/death within the first year of life, 1.42 (95% Cl 0.90–2.24). No substantial response bias was found. When occupational titles were used as an exposure matrix, no increased risk was found, except for term low birth weight. Recall bias is one possible explanation. Thus, the results must be interpreted with caution. Am J Epidemiol 1992;135:302–11.