We agree that selection bias may arise in cohorts that are recruited via the Internet, as in other settings, but lack of sampling representativeness is not the underlying problem (Rothman et al., 2013). Indeed, a recent comparison of SnartGravid and registry data provided empirical evidence against selection bias (Hatch et al., 2015).
Comment on "Perils and potentials of self-selected entry to epidemiological studies and surveys"
Rothman, K., Hatch, EE., Poole, C., Wise, LA., Gallacher, JEJ., Lash, TL., ... Sorensen, HT. (2016). Comment on "Perils and potentials of self-selected entry to epidemiological studies and surveys". Journal of the Royal Statistical Society. Series A (Statistics in Society), 179(2), 364-364. https://doi.org/10.1111/rssa.12136