Behavioral and developmental pediatric research often seeks to form causal inferences from associations between variables obtained in nonrandomized studies. To do this it is necessary to distinguish the effects of the independent variable of interest from other factors with which it is correlated. We review statistical adjustment procedures for assessing the effects of the independent variable after controlling for other variables, called cofactors. We present guidelines for cofactor adjustment for different patterns of causal relationships occurring in the developmental pediatric literature. We also review procedures for selecting a subset of cofactors from a larger array of candidate variables. Finally, we examine common methodological complications related to cofactor adjustment, including the presence of measurement error in the independent variable and/or the cofactors.