Data Imputation in Multi-Level Quantile Regression

• Activity Number: 181
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Social Statistics Section
• Author(s): Luke Fostvedt*+ and Mack Shelley
• Companies: Iowa State University and Iowa State University
• Keywords: quantile regression ; multi-level ; imputation

Abstract

In any large scale study, the presence of missing data is a near certainty. In most cases, there is a non-random pattern to the missingness that needs to be taken into account. Complete case analyses make untenable assumptions and throws away important information. Multiple imputation methods have been proposed for quantile regression (Wei et al. 2012; Geraci, 2013). By specifying the conditional quantile function(s), chained equations can be used to impute on a variable-by-variable basis with the quantile regression model specified for the sampling process. The imputation approach from Geraci (2013) will be applied to a multi-level quantile regression scenario and the performance of the approach will be assessed.