5 women sat at a table at the ICWSM conference in May. Four had baby boys at home, the 5th was expecting. Two of us (Karrie and I) thought this was a funny coincidence, and meeting again in July at the Google faculty summit, we thought: let’s do a survey. The survey would be about women with PhDs having kids, but we would gather data on the child’s sex as well.
Several months and several hundred survey responses later we know the following:
- If danah boyd tweets about your survey, you’ll not only get ~2,000 page views, and ~150 completed surveys, but varied, global exposure, compared to e.g. posting to a campus-wide list for women engineers (which cascaded to a similar number of responses). And yes, Eszter Hargittai is right, there are various potential sources of bias: snowball sampling could have skewed participants toward those who are like Karrie and me (or more likely, like danah), women who had more/less difficulty juggling career and family may be more/less motivated to respond, etc.
- You can hastily compose a survey, not have a recruitment plan, and still learn a lot from it informally. I’ve compiled the freetext responses, and produced some summary statistics (with again the caveat of sampling bias): when did PhDs have children? how much time off did they take and what childcare options did they select? How did having a child impact their work hours?
- Finally, at the 0.01 significance level, there is no effect of PhD-attainment, career choice, or field on the child’s sex. This is a relief, because statistically significant results may have landed us on Andrew Gelman’s blog as an example of how not to do statistical analyses. It’s also a relief because there’s no plausible reason for a connection. So this weekend, at a baby shower (a female faculty member is expecting a boy), I thought, “there’s no pattern. Think statistically! There’s no pattern.”
I am very grateful to those who have taken the survey and hope that these few pages will be a possible resource (in addition to much great published research) to those weighing their choices or wondering what others’ experiences have been like. Personally, it has opened my eyes to the range of experiences, struggles, triumphs, and resourceful solutions out there, and has helped me to calibrate my own experience.
Dooo deee doo….. Thanks for sharing what you found!!!!
When I was in graduate school, a frequent topic of conversation among those of us who were married was when was the best time to have (or bear) children. The consensus seemed to be that: (a) there’s never a good time to have children, but (b) having them while still in graduate school may be the lesser of inconveniences.
I changed fields (AI -> HCI / CSCW) after finishing my Ph.D., so I rarely see any of my friends from graduate school, and don’t know how their estimates have panned out. You’ve probably already thought of this, but it might be interesting to track responses to your survey over time, and perhaps invite responses from those who are at earlier stages of their careers.