Introduction Illustration: Trust in institutions Step 1: Preparation and coding of technical variables Step 2: Selection of source variables for harmonization Step 3: Mapping source values to target values Step 4: Harmonization Results: Availability of trust items Comparability of sample aggregates Appendices: Code examples Appendix 1: Data preparation Appendix 2: Codebook from labelled data in R Appendix 3: Values crosswalk Appendix 4: Harmonization Introduction Ex-post (or retrospective) data harmonization refers to procedures applied to already collected data to improve the comparability and inferential equivalence of measures collected by different studies.
Working with categorical data, such as from surveys, requires a codebook. After spending some time unsuccessfully looking for a function that would create a nice, searchable codebook from labelled data in R, I decided to write my own. What I want to achieve is a simple table with variable names, labels, and frequencies of labelled values like the one below, to search for specific keywords in the value labels and to see distributions of various variables.
So You Want to Write a Fugue? Glenn Gould So you want to write a fugue? You’ve got the urge to write a fugue You’ve got the nerve to write a fugue So go ahead and write a fugue that we can sing Pay no heed to what we’ve told you Give no mind to what we’ve told you Just forget all that we’ve told you And the theory that you’ve read
Sample correlations Sample correlations by gender Sample correlations by age Sample correlations by education Contrast Conclusion One of the reasons for the harmonization of personal income in addition to household income was to check if the two correlate highly enough to use household income as a substitute for personal income in analyses where economic status is a control variable. This would be great, because household income variables are available in 1177 surveys out of 1721 analyzed in the Survey Data Recycling dataset (SDR) version 1, while personal income only in 453 surveys.
Data Number of response options Item non-response Distributions Harmonized target variables Next steps with Przemek Powałko Individual economic status is a necessary element of almost all sociological analyses, including studies of political attitudes and behavior. To supplement the already harmonized variables in the Survey Data Recycling dataset (SDR) version 1 and for the purposes of my resesarch of the effects of education on political engagement, Przemek and I harmonized two additional variables: personal income and household income1.
Cross-national survey projects conduct surveys on representative samples of adult populations. How do the distributions of respondents’ age vary across surveys carried out in the same country in different years and different projects? Like in a couple of previous posts (here, here and here) I use data from the Survey Data Recycling dataset (SDR) version 1, which includes selected harmonized variables from 22 cross-national survey projects. SDR only includes surveys that claim to have samples representative for adult populations.
Data Differences within country-years Differences by groups Gender Age Urban/rural residence Education Sampling scheme The growth in cross-national survey projects in the last decades leads to situations when two or more surveys are carried out in the same country and the same year but in different projects, and contain overlapping sets of survey questions. Assuming that the surveys are based on representative samples - a claim that major cross-national survey projects typically make - it could be expected that estimates from surveys carried out in the same country and year are reasonably close.
Educational attainment data OECD data SDR data Cleaning and merging SDR and OECD data Results The curious case of ISSP Switzerland Conclusion Appendix with Przemek Powałko General population surveys with representative samples should have a similar education structure as shown by data from administrative sources, especially if survey weights are used. In this post we compare sample aggregates from 15 cross-national survey projects (including the European Social Survey, the World Values Survey and the European Values Study, and others) from the Survey Data Recycling database with educational attainment statistics from the OECD.
Instructions References In the previous post I wrote about downloading and exploring the Survey Data Recycling (SDR), version 1 dataset, which consists of selected harmonized variables from 22 survey projects, 1966-2013. The SDR project will develop a website for browsing, subsetting, downloading, and visualizing data from the SDR project. This website is currently under construction. Meanwhile, I made a Shiny app with basic functionalities of the future on-line browsing and subsetting tool (also serves as its mock-up): https://mkolczynska.
Introduction Downloading the SDR data Exploring SDR: availability of variables by project Exploring SDR: availability of variables with different formulations Identifying surveys containing selected variables Subsetting the Master File Country coverage plot Combining data from different survey projects creates new opportunities for research, alas, at the cost of increased volume (obviously) and complexity of the data. The Survey Data Recycling project created a dataset with data from 22 international survey projects.