Research aimed at the empirical evaluation of infertility treatment including assisted reproductive technologies (ART) on child health and development is hampered by investigators' inability to methodologically separate possible treatment effects from underlying fecundity impairments. While the literature continues to identify ART as a risk factor for many child health outcomes, less attention has been paid to the methodologic rigor needed to answer this question. We identify aspects of fecundity and the nuances of medical practice that need to be considered and captured when designing epidemiologic investigations aimed at assessing ART and child health. These include: (i) the use of prospective study designs in which the unit of analysis (cycle versus individual versus couple) is defined; (ii) data collection on relevant time-varying covariates at, before and during treatment; and (iii) the use of statistical techniques appropriate for hierarchical data and correlated exposures. While none of these issues in and by itself is unique to ART research, attention to these issues has been lacking in much of the published research limiting our ability to evaluate health consequences for children. Longitudinal studies of children conceived with ART will benefit from attention to these issues and, hopefully, produce answers to lingering questions about safety.