The influence of factors acting during early life on health outcomes of offspring is of considerable research and public health interest. There are, however, methodological challenges in establishing robust causal links, since exposures often act many decades before outcomes of interest, and may also be strongly related to other factors, generating considerable degrees of potential confounding. With respect to pre-natal factors, the degree of counfounding can sometimes be estimated by comparing the association between exposures experienced by the mother during pregnancy and outcomes among the offspring with the association of the same exposures experienced by the father during the pregnancy period and offspring outcomes. If the effects are due to an intra-uterine exposure, then maternal exposure during pregnancy should have a clearly greater influence than paternal exposure. If confounding by socio-economic, behavioural or genetic factors generates the association then maternal and paternal pregnancy exposures will be related in the same way with the outcome. For early life exposures it is also possible to compare outcomes in siblings who are concordant or discordant for the exposure, which will reduce the influence of family-level confounding factors. A different approach is that of Mendelian randomization, which utilises genetic variants of known effect that can proxy for modifiable exposures and are also not in general related to potential confounding factors, or influenced by disease. In other settings the use of non-genetic instrumental variables is possible. A series of examples of the application of these approaches are presented and their potentials and limitations discussed. Other epidemiological strategies are briefly reviewed. It is concluded that the naïve acceptance of findings utilising conventional epidemiological methods in this setting is misplaced.