Don’t have much time? Watch a two-minute summary of our key findings.

While women in the US today earn over 50% of bachelor’s degrees, reports estimate that they only account for 25% or fewer of students earning degrees in science, technology, engineering and math (or STEM, for short). There has been no shortage of media coverage recently speculating about the causes of this imbalance, and discussing why…

The data for this project was obtained from the Integrated Public Use Microdata Series (IPUMS), which is maintained by the University of Minnesota. IPUMS consists of sample microdata from US censuses dating back to 1850. The available data for each census represents approximately 1% of the total population. The key advantage of working with this data…

Our exploratory analysis confirmed our understanding that fewer female than male students are majoring in these subjects and hinted at a few factors that may affect a female student’s decision to study STEM. In particular, factors such as ethnic background, immigration status and the presence of household members with a STEM background stood out as…

But what does this all mean? For us, the most important single takeaway from this analysis is the importance of role models! Over and over again, we kept encountering a strong association between having a household member with a STEM degree or in a STEM occupation and the likelihood of female students majoring in STEM. This…

Want to go into more detail? Check out the following additional resources. iPython process book Github repository (including processed datasets) Original (full) 2010 U.S. Census dataset