Profile Likelihood Estimation of the Correlation Coefficient in the Presence of Left, Right or Interval Censoring and Missing Data

Authors

• Yanming Li, University of Michigan-Ann Arbor
• Brenda W. Gillespie, University of Michigan-Ann Arbor
• Kerby Shedden, University of Michigan-Ann Arbor
• John A. Gillespie, University of Michigan-Dearborn

Date of this Version

12-2018

Document Type

Article

Citation

The R Journal (December 2018) 10(2); Editor: John Verzani

Comments

Copyright 2018, The R Foundation. Open access material. License: CC BY 4.0 International

Abstract

We discuss implementation of a profile likelihood method for estimating a Pearson correlation coefficient from bivariate data with censoring and/or missing values. The method is implemented in an R package clikcorr which calculates maximum likelihood estimates of the correlation coefficient when the data are modeled with either a Gaussian or a Student t-distribution, in the presence of left, right, or interval censored and/or missing data. The R package includes functions for conducting inference and also provides graphical functions for visualizing the censored data scatter plot and profile log likelihood function. The performance of clikcorr in a variety of circumstances is evaluated through extensive simulation studies. We illustrate the package using two dioxin exposure datasets.