This paper analyses the quantity-quality trade-off and other household-level influences on educational outcomes of children. Nineteenth and early-twentieth century census micro-data is used, providing material from Western and Eastern Europe, the USA and Canada. This is the time-frame of the demographic transition and the onset of modern economic growth; when the quantity-quality trade-off should be most important. Besides the number of siblings, other household features can be of influence as well, such as the gender composition, the inclusion of members outside the nuclear family and the position of women in the household. We find mixed evidence for the quantity-quality trade-off, as well as positive effects of the presence of (upwards) extended family members and parental, especially maternal, literacy.