The main aim of this study was to examine twin specific risk factors that influence educational achievement in primary school. We included prenatal factors that are not unique to twins, except for zygosity, but show a higher prevalence in twins than in singletons. In addition, educational achievement was compared between twins and their nontwin siblings in a within-family design. Data were obtained from parents and teachers of approximately 10,000 twins and their nontwin siblings registered with the Netherlands Twin Register. Teachers rated the proficiency of the children on arithmetic, language, reading, and physical education, and reported a national educational achievement test score (CITO). Structural equation modeling showed that gestational age, birth weight, and sex were significant predictors of educational achievement, even after correction for socioeconomic status. Mode of delivery and zygosity did not have an effect, while parental age only influenced arithmetic. Mode of conception, incubator time, and birth complications negatively affected achievement in physical education. The comparison of educational achievement of twins and singletons showed significantly lower ratings on arithmetic, reading, and language in twins, compared to their older siblings, but not compared to their younger siblings. Low gestational age and low birth weight were the most important risk factors for lower educational achievement of twins in primary school. It seems that the differences observed between twins and their nontwin siblings in educational achievement can largely be explained by birth order within the family. Copyright © Cambridge University Press 2012.