What’s missing? The effect of missing data and imputation techniques on predictive performance in forecasting civil war violence

Abstract:

Missing data is a ubiquitous problem in science in general, and social science in particular. It is well-known that not dealing properly with the missing data problem may cause inferences to be biased. Less attention has been given to the implication of missing data and imputation techniques on the predictive performance of dierent models. This paper explores these issues through generating missing data in a data set used for forecasting the number of fatalities from political violence in all countries during 1989-2021. The results show that single model-based and non modelbased imputation techniques perform best with regards to the predictive performance of the model on the outcome of interest, while multiple imputation techniques fare generally fare worse. These results highlight that the guidelines for how to handlemissing data in inferential studies are not generally transferable toforecasting studies. Instead, this paper suggests that the imputation technique should be specifically tailored to the research objective and to the variables which are to be imputed.

Authors:

David Randahl