Demographers and others often study the interaction between two types of individual-level behaviour, such as migration and childbearing. Unfortunately, one can get estimation bias if one compares childbearing before and after migration from data confined to migrants, say, as is sometimes done to...