Statistical models for time-to-pregnancy data are usually formulated in discrete time. Two approaches were surveyed, including methods of accounting for both known heterogeneity (covariates, possibly time-dependent) and unknown heterogeneity (frailty). Delicate censoring and truncation patterns arose for prospective and, particularly, retrospective sampling designs. The inclusion of several pregnancies per couple presents new challenges and possibilities.

Key terms cross-sectional sampling; discrete-time survival analysis; frailty; prospective sampling; retrospective sampling; survival analysis; truncation