Table 1.

Summary of the data characteristics and requirements of the compared approaches

Method	Missing genotypea		Covariates in model	Continuous exposure and covariates	External prevalence
	child	mother
Prospective likelihood	X		X	Xb	Xc
Stratified retrospective likelihood	X	X
Retrospective likelihood, unspecified distribution of GM, X	X		X	X	X
Retrospective likelihood, double-additive model for Pr(GM |X)	X	X	X	X	X
Logistic regression			X	X

^aTreatment of missing genotypes to extract information from other variables.

^bIn practice, leads to convergence problems due to large number of distinct values.

^cNumbers of cases and non-cases in population are required but can be approximated from external prevalence.