The impact of incomplete data on quantile regression for longitudinal data

Authors

  • Anneleen Verhasselt Data Science Institute, I-Biostat, Universiteit Hasselt, Diepenbeek, Belgium
  • Alvaro J Florez Data Science Institute, I-Biostat, Universiteit Hasselt, Diepenbeek, Belgium. School of Statistics, Universidad del Valle, Cali, Colombia
  • Ingrid Van Keilegom Research Centre for Operations Research and Statistics, KU Leuven, Leuven, Belgium
  • Geert Molenberghs Data Science Institute, I-Biostat Universiteit Hasselt, Diepenbeek, Belgium. I-Biostat, KU Leuven, Leuven, Belgium

Keywords:

Dropout; Inverse probability weighting; Missing data; Multiple imputation; Quantile regression

Abstract

We investigate the performance of methods for estimating the conditional quantile of a response based on longitudinal data, when outcomes are incomplete and when the correlation between repeated responses is ignored. In a simulation study, we compare the performance of the quantile regression estimator based on the complete cases, the available cases, quantile-based multiple imputation, and quantile-based inverse probability weighting. In the data setting considered, quantile-based multiple imputation is the most promising method with the best bias-efficiency trade-off. A potential drawback, however, is its computation time.

Journal of Statistical Research 2021, Vol. 55, No. 1, pp. 43-58

Abstract
27
PDF
18

Downloads

Published

2021-12-09

How to Cite

Verhasselt, A., Florez, A. J., Keilegom, I. V., & Molenberghs, G. . (2021). The impact of incomplete data on quantile regression for longitudinal data. Journal of Statistical Research, 55(1), 43–58. Retrieved from https://banglajol.info/index.php/JStR/article/view/56566

Issue

Section

Articles