Predicting Gender Gaps in Internet Use Using Facebook and Google Advertising Audience Estimates

Reham Al Tamime, University of Southampton
Ridhi Kashyap , University of Oxford
Ingmar Weber, Qatar Computing Research Institute
Masoomali Fatehkia, Northeastern University

Closing gender gaps in internet use and digital skills are important development goals. A key barrier for measuring progress towards these goals is the limited availability of gender-disaggregated data on digital gender gaps. We examine how anonymous, aggregate data from the online advertising platforms of Google and Facebook can be leveraged to predict global digital gender gaps. We generate gender gap indicators using both AdWords and Facebook data, and find that these online indicators are highly correlated with official statistics on gender gaps in internet use and low-level digital skills from the International Telecommunications Union (ITU) when available. We test different models using only online indicators, only offline development indicators, as well as those combining online and offline indicators to predict ITU digital gender gap measures. We find that the best performing predictive models are those that combine Facebook and Google online indicators with a country’s offline development indicators.

See paper

 Presented in Session 128. Using Social Media in Population Research