Paper
Washington, Ebonya L. (2008). American Economic Review, 98(1): 311–332
Read the original paper → Opens at the publisher; use the DU library / JSTOR login if it asks for access.
In one paragraph
Sociologists had shown that parenting daughters increases feminist sympathies. Washington asks whether children, like neighbours or peers, change their parents’ behaviour in a measurable way: do legislators with more daughters vote differently? Conditional on the total number of children — which parents choose — the sex mix is as good as random, so the number of daughters is a clean source of variation. Each additional daughter raises a member of Congress’s propensity to vote liberally on women’s issues, with the strongest effect on reproductive-rights votes.
Setting and question
Roll-call votes in the US House and Senate over several congresses, scored on women’s issues by interest-group ratings. The question is whether a personal characteristic that is plausibly exogenous shifts legislative behaviour once party and constituency are held fixed.
Data and identification
Legislators’ numbers of sons and daughters matched to voting scores. Identification rests on the sex of each child being random conditional on family size — the paper is a model of stating and defending that assumption, and of showing results are not driven by selection into having children.
Main results
A positive, statistically significant daughter effect on liberal voting on women’s issues; largest for reproductive-rights votes; robust to controls for party, region and ideology.
Why it is on the reading list
A canonical “random conditional on something” natural experiment. Read together with Green, Hyman-Metzger, Sood & Zee (2023), which re-examines the finding over a longer period — the pair is a case study in replication and in how much weight a single natural experiment can bear.
No interactive explainer for this paper yet. If you build one, see Contribute to get it added.