A large body of literature has reported associations between socioeconomic position and adverse pregnancy outcomes even in affluent egalitarian welfare states. This study explored the nature of this relationship by examining women who changed socioeconomic position between pregnancies and women who were siblings but were different in terms of socioeconomic position. Data consisted of 471,215 live born singletons born in Denmark 1997-2007 with at least one sibling or one first cousin. We examined parental educational attainment and household income in relation to preterm birth and small for gestational age using Cox regression. Household income was only weakly related to these outcomes. Paternal education was strongly associated with the outcomes only in the cohort analyses. Maternal education was inversely associated with preterm birth only in the cohort analyses, where the least educated women had the highest risk. Maternal education was inversely associated with the risk of small for gestational age in cohort analyses, attenuated between mothers who were siblings, and not present between children who were siblings. For example, the hazard ratio of preterm birth of women with a college/university degree when compared to women with only mandatory education was 0.64 (95% confidence interval: 0.60-0.67) in the cohort analysis, 0.90 (0.78-1.04) between mothers who were siblings, and 1.01 (0.82-1.24) between children who were siblings. The corresponding hazard ratios of small for gestational age were 0.54 (0.52-0.56), 0.72 (0.63-0.83), and 1.02 (0.84-1.24). This suggests that the associations were partly explained by factors shared between mothers who are siblings. In conclusion, the early life circumstances of mothers appear to be important in understanding the association between education, preterm birth and small for gestational age.