How to Deal with Missing Data and Galton’s Problem in Cross-Cultural Survey Research: A Primer for R

Multiple imputation (MI) has become the preferred method for dealing with missing data in surveyresearch. MI involves three steps: creating m multiply imputed complete datasets; estimatingmodels on each of the m datasets using any standard statistical procedure; combining the resultingmultiple estimates of each statistic of interest. This paper provides R programs for MI, and offerssome advice for employing MI with data drawn from the Standard Cross-Cultural Sample (SCCS).A second set of R programs combines estimates from the m imputed data sets, and also deals withthe problem of network autocorrelation effects, i.e., Galton’s Problem or the non-independenceof cases, using two-stage instrumental variables (IV) regression. The objective of the paper isto provide programs, advice and explanations that will help researchers employing cross-culturalsurvey data, especially the SCCS, to deal with the twin problems of missing data and networkautocorrelation effects, using the open source statistical package R. The paper is intended tocomplement a recent suite of publications by Dow and Eff where both theoretical and empiricalissues underlying these two problems are discussed in detail.