A large body of literature has demonstrated a positive relationship between education and age at first birth. However, recent twin studies have shown that this relationship may be spurious because of social (and genetic) family background factors that cannot be controlled for in most research designs (Tropf & Mandemakers, 2017 Demography). However, within twin/siblings designs have high internal validity but often lack generalizability and are usually limited to only one context. We contribute to the literature by investigating and comparing the extent to which education is causally related to later age at first birth in large samples of siblings and twins in multiple different contexts. We compare within and between family (sibling/twin) estimates in -at least- five different countries and in different birth cohorts. Specifically, in the United Kingdom using the TwinsUK dataset (N=2,752 female twins) and the BHPS/Understanding Society (siblings), in the Netherlands using the NKPS dataset (N=20,000 siblings), in Germany using the Pairfam dataset (siblings), GSOEP (siblings) and the new Twinlife dataset (N=2,000 twins), in the US using the PSID datasets (siblings), and the Australian QIMR dataset (N=10,000 siblings/twins). We present novel estimates using within–sibling/twin pair models. Initial findings in the UK show that one year of additional schooling is associated with about one-half year later age at first birth in ordinary least squares (OLS) models, which reduces to 1.5-month later age at first birth for the within–identical twin model controlling for all shared family background factors (genetic and family environmental) (10% lower hazard of first birth reduces to 4% lower hazard using Cox (stratified) models). This paper will hopefully contribute to a better understanding of the paradox that within cohorts educational differences between siblings and twins do not appear to matter but that over time educational expansion is related to fertility timing between cohorts.