Our research builds upon two existing types of data: cross-national, longitudinal survey studies as well as macro data of political regimes. Here we provide information how to access and harmonizes these datasets.
Step-by-step guide to data harmonization of survey data
As part of this project, we produced a new dataset, which harmonizes numerous existing public opinion surveys from across the world to create a unique global public opinion dataset.
- World Value Survey (WVS), 1981-2014
- International Social Survey Programme (ISSP), 1985-2013
- Latinobarometer (LB), 1995-2013
- Americas Barometer (AB), 2004-2014
- Asian Barometer (ANB), 2001-2014
- Asia Barometer (AsiaB), 2003-2008
- Afrobarometer (AFB), 1999-2015
- European Values Study (EVS), 1981-2010
- European Social Survey (ESS), 2002-2014
- Eurobarometer (EB, Mannheim Trend file), 1970-2002
- Central & Eastern European Barometer (CEEB), 1990-97
- Comparative Study of Electoral Systems (CSES), 1996-2015
These studies consists of over 1,100 individual country-year datasets. Putting all these together, covers 160 countries and over 3 Million respondents. Here a list of countries and years that allow for longitudinal analysis (min. 10 years of coverage per country).
Due to data ownership issues, we are not able to offer a ready-downloadable harmonized dataset of these different studies. Instead, we provide a step-by-step guide to allow other researchers to build their own working files. In order to create the harmonized dataset four steps need to be followed:
- Download all datasets. Click here to access data and documentation.
- Decide which countries, years and variables are needed. We have complied three documents that should be used for this. For more information on each, please see step-by-step guide:
- Waves by Country: provides detail on coverage for each country. Use this file to see the coverage over time for each country. We list which study included each country-year.
- Variables by Dataset Wave: This spreadsheet gives the names of variables contained in each dataset. Each tab refers to a group of variables, and these tabs correspond to the tabs in the “Question Wording” spreadsheet. The focus of coverage is on political attitudes and behavior. The coverage is not complete, but quite extensive. See more in step-by-step guide.
- Question Wording: question wording for each variable in each dataset. Use this dataset to assess comparability in question wording between the datasets for the same variable.
- Prepare datasets that belong to each study first by recoding the variables into common variable names and categories and if necessary merge single files for each study to create longitudinal files. We provide a template STATA do-file for step 3 called “template_within_data_prep.do”. You can also download the do-file to create country labels in the zip-folder below.
- Merge all separate, prepared studies into one big dataset. We provide a template STATA do-file for step 4 called “template_merge.do”.
We have prepared a codebook of all the data and variables that we harmonized for our project following this step-by-step data harmonization.
You can also download a zip-folder here with all relevant files. All STATA do-files used in this project are available in this zip-folder.
Macro political regime data
Several major projects have coded political regime characteristics, covering most countries in the world over time (often annual data). Here we provide links to the data that we used: