The GEM Cross-Sectional dataset covers a total of 155 nations and contains one observation for each variable and country. It thus includes the latest available observation for each country, using 2015 as the target year. If no data from 2015 are available, data for 2014 are included and so on. This means that the CS-GEM contains variables with observations that mostly were collected between 2010 and 2015.
The GEM Time-Series dataset also covers a total of 155 nations. Data from 1995 to 2015 are included and the unit of analysis is country-year (e.g. Sweden1995, Sweden1996 and so on up until Sweden2015).
You can find a description of all data sources and variables in the GEM Codebook. There are detailed descriptions of all variables and both datasets and codebook are organized in three themes qualities, causes and consequences, consisting of different types of variables.
The datasets are free to use for non-commercial purposes, but citations and references to both the GEM dataset and the original data sources are mandatory. All users are also advised to consult the original studies, codebooks, and the documentations where more extensive descriptions of the original data collection and variables can be found.
When using the GEM datasets, make sure to cite both the original source and our publication:
Färdigh, Mathias A., Monika Djerf-Pierre, Maria Edström, with Carolyn M. Byerly, Sarah Macharia, Claudia Padovani, and Karen Ross. 2020. “The GEM Dataset, version 1.” Comparing Gender and Media Equality (GEM) Project, Department of Journalism, Media and Communication (JMG), University of Gothenburg.