Go to navigation (press enter key)Menu
Project Proposal

Why Women Don’t Code: Media Exposure, Education and Occupational Choices in the U.S. Labor Market

Sarah Pearlman (Economics)

Project Description

The gender wage gap, in which women earn less than men, continues to be a feature of U.S. labor markets. While the persistence of this gap remains both a concern and a puzzle, it is clear that one important contributing factor is occupation. Put simply, women are more likely than men to move into lower paid occupations. One occupation where this is most evident is computer programming, which is highly paid and yet dominated by men. Interestingly, however, this was not always the case. In 1983 the percentage of computer science majors who were women was almost twice that of today (approximately 37% in 1983 versus approximately 20% in 2015). Furthermore, before the mid-1980s the percentage of female majors in computer science rose consistently, tracking trends in other highly paid fields, like law and medicine. After the mid-1980s, however, the trend reversed and the percentage of women started to fall. As a result, the percentage of women majors in computer science today is less than half that of medicine, law and physical sciences. 

The question is what happened in the mid-1980s to push women out of computer science. Answering this question may help us understand occupational choice and the role it plays in determining the gender wage gap. This project aims to explore one theory for the change, which is that a rise in media portrayals of computer science as a profession for men discouraged women from entering the field. In the early to mid-1980s, as computers became more widely available, television shows, movies and commercials increasingly showed people using computers, but they were mostly boys and men. Meanwhile, exposure to these images increased with the expansion of cable television. The overall theory is that girls with access to cable were more exposed to images of male computer scientists and were less likely to choose this profession as a result. In general, this project follows other research in economics that finds that media can influence marriage, child bearing, and gender dynamics within the household (Guldi and Herbst 2017; Kearney and Levine 2015; Jensen and Oster 2009; La Ferrara, Chong and Duryea 2012). This project will contribute to this body of work by looking at the effect media has on educational and occupational choices.

Anticipated Project Activities

The project has a data and writing component. The data component will involve assisting in linking data on the expansion of cable networks in the U.S. to longitudinal data on individuals. If media exposure indeed played a role in reducing women’s presence in computer science, we should find that women who grew up in areas with cable will be less likely to enter computer science at the university level and beyond. The work on these data will involve multiple pieces. First, it will involve becoming familiar with a large and complex longitudinal data set and learning how to link this with other data sets. Second, it will involve doing preliminary analysis by constructing summary statistics and by using regressions. In helping with the empirical analysis the student will become familiar with the challenges associated with identifying the impact of certain policies or changes on specific outcomes, and ways in which different econometric models deal with these challenges. Third, it will involve learning how to effectively present the results of data analysis in both a paper and a presentation. 

The writing component of the project will involve writing up the summary statistics and preliminary results. This will provide the student with a valuable opportunity to practice describing data and the results of data analysis, both of which are a challenge.

Preferred Student Qualifications and Skills

The student needs to have taken Econometrics (at Vassar or elsewhere). The ability to critically analyze economic data is needed. In particular, it is important that the student understand the challenges of identifying causal impacts when non-experimental data are used, and some of the ways in which economics research tries to overcome these challenges. Furthering knowledge of econometric techniques is a main goal of the Ford project. 

The student also will need to have some familiarity with STATA, which they will have from Econometrics.

Anticipated Follow-up Teaching/Professional Activity for Student

Through the project the student will learn how to work with real data and how to use those data to answer questions about individuals’ economic decision making. This is beneficial as data analysis is a key part of many jobs, both within and outside of academia. By gaining a better understanding of how data sets are constructed and how they can be used, the student will better see the link between economic theory, data and policy-making.

The student will use STATA, a widely used statistical package. Knowledge of STATA is very beneficial if the student is interested in pursuing graduate study or a career within Economics and Public Policy. Many excellent jobs for undergraduates within both fields require a good understanding of econometrics and statistical programs. 

The student will also be encouraged to present the work inside the department, during a brownbag seminar. The student also will be encouraged to think about incorporating suggestions to better the paper. 

Finally, by exposing the student to research within Economics the experience will help the student decide if a more research oriented career within Economics appeals to them.

Project Location

Vassar College, Poughkeepsie, NY

Project Duration

Eight weeks

Project Start Date

May 30, 2017

Project End Date

July 20, 2017