Using classification and regression trees (CART) and random forests to analyze attrition: Results from two simulations
Article
Hayes, T, Usami, S, Jacobucci, R et al. (2015). Using classification and regression trees (CART) and random forests to analyze attrition: Results from two simulations
. PSYCHOLOGY AND AGING, 30(4), 911-929. 10.1037/pag0000046
Hayes, T, Usami, S, Jacobucci, R et al. (2015). Using classification and regression trees (CART) and random forests to analyze attrition: Results from two simulations
. PSYCHOLOGY AND AGING, 30(4), 911-929. 10.1037/pag0000046
In this article, we describe a recent development in the analysis of attrition: using classification andregression trees (CART) and random forest methods to generate inverse sampling weights. These flexiblemachine learning techniques have the potential to capture complex nonlinear, interactive selectionmodels, yet to our knowledge, their performance in the missing data analysis context has never beenevaluated. To assess the potential benefits of these methods, we compare their performance withcommonly employed multiple imputation and complete case techniques in 2 simulations. These initialresults suggest that weights computed from pruned CART analyses performed well in terms of both biasand efficiency when compared with other methods. We discuss the implications of these findings forapplied researchers.
