The aim of the study was to examine the association between stress and pregnancy outcome after adjustment for possible confounding and mediating variables. A prospective cohort study of 5511 pregnancies was conducted in 2001-03 in the Netherlands. A standardised questionnaire collecting demographics and mental health data was administered at 14 and 30 weeks of pregnancy. Medical data on the pregnancy and delivery were obtained from obstetricians and midwives. The results showed that a high level of perceived stress at 14 weeks of pregnancy increased the risk for delivery of an infant that was small-for-gestational-age (OR = 1.26 [95% CI 1.01, 1.56]), but the association was reduced after adjustment for the possible confounding effects of demographic variables (OR = 1.16 [95% CI 0.92, 1.47]). The results do not support a direct relationship between perceived stress and adverse pregnancy outcome. Demographic variables may explain the association between psychosocial stress and pregnancy outcome to a significant degree.