Analyzing Incomplete Political Science Data: An Alternative Algorithm for Multiple Imputation

• Published in: The American political science review Vol. 95; no. 1; pp. 49 - 69
• Main Authors: King, Gary, Honaker, James, Joseph, Anne, Scheve, Kenneth
• Format: Journal Article
• Language: English
• Published: Washington Cambridge University Press 01.03.2001; American Political Science Association
• Subjects: Algorithms; Coding; Computation; Computer Assisted Research; Computer Software; Data; Data Analysis; Data editing; Data imputation; Data processing; Datasets; Estimation bias; Estimators; Evaluation; Job Performance; Mathematical Models; Mathematics; Methodology; Methodology (Data Analysis); Methods; Missing data; Modeling; Multiple imputation; Multiple imputation (Statistics); Political parties; Political science; Political science research; Political scientists; Research methods; Scientists; Software; Standard error; Statistical analysis; Statistical discrepancies; Statistical methods; Statistics; Surveys; Variables; Voting; United States
• ISSN: 0003-0554; 1537-5943
• DOI: 10.1017/s0003055401000235

We propose a remedy for the discrepancy between the way political scientists analyze data with missing values and the recommendations of the statistics community. Methodologists and statisticians agree that “multiple imputation” is a superior approach to the problem of missing data scattered through one’s explanatory and dependent variables than the methods currently used in applied data analysis. The discrepancy occurs because the computational algorithms used to apply the best multiple imputation models have been slow, difficult to implement, impossible to run with existing commercial statistical packages, and have demanded considerable expertise. We adapt an algorithm and use it to implement a general-purpose, multiple imputation model for missing data. This algorithm is considerably faster and easier to use than the leading method recommended in the statistics literature. We also quantify the risks of current missing data practices, illustrate how to use the new procedure, and evaluate this alternative through simulated data as well as actual empirical examples. Finally, we offer easy-to-use software that implements all methods discussed.