Abstract. OBJECTIVE: To compare the incidence of small for gestation age (SGA) neonates in twin gestations conceived by ovulation induction or in vitro fertilization with that of twins conceived spontaneously. STUDY DESIGN: A retrospective cohort study was conducted. Subjects were delivered by a single obstetric practice between 2005 and 2011 at the Mount Sinai Medical Center. Maternal and neonatal data were recorded. Our primary outcome was the incidence of SGA, defined as birth weight <10th percentile, from the three modes of conception. Chi square, ANOVA, Fisher’s exact test, the Kruskal-Wallis test, the Mantel-Haenszel test, and logistic regression were used in the analysis. RESULTS: In unadjusted analysis, using the records of 756 infants from 378 mothers of twin pregnancies, twins conceived by ovulation induction had an increased incidence of SGA (28.1%) compared to those conceived spontaneously (17.1%) and by in vitro fertilization (16.5%, p = 0.006). In a logistic regression model accounting for correlated responses between twins and adjusting for gestational age, gender, chorionicity, and maternal age, the odds ratio (95% confidence interval [CI]) of SGA for the ovulation induction group compared to the spontaneous conception group was 2.64 (1.38-5.05, p = 0.003). The odds ratio (95% CI) of SGA for the in vitro fertilization group compared to the spontaneous conception group was 1.51 (0.88-2.61, p = 0.135). CONCLUSIONS: When adjusted for gestational age, gender, chorionicity, and maternal age, twin neonates conceived by ovulation induction, but not those conceived by in vitro fertilization, had increased odds of SGA compared to those conceived spontaneously.