PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
January 1, 1979Biometrika1,031 citations

Linear regression with censored data

View Full Paper
JBJonathan D. BuckleyIJIan James

Key Points

  • To develop a consistent linear regression estimation method for models with censored dependent variables and unspecified residual distributions.
  • Modified the classical normal equations directly rather than altering the sum of squared residuals.
  • Derived large-sample asymptotic properties heuristically and evaluated estimator performance using numerical simulations.
  • Reanalyzed clinical heart transplant survival data to benchmark against previous estimation methods.
  • Direct modification of the normal equations resolved the asymptotic inconsistency issues present in Miller's residual-based approach.
  • Simulation studies substantiated the theoretical large-sample properties and demonstrated stable slope parameter estimation under censoring.

Abstract

We give a method of estimating parameters in the linear regression model which allowB the dependent variable to be censored and the residual distribution to be unspecified. The method differs from that of Miller (1976) in that the normal equations rather than the sum of squares of residuals are modified and this appears to overcome the inconsistency problems in Miller's approach. Large sample properties of the estimator of slope are derived heuristically and substantiated by simulations. Some of the heart transplant data reported and analysed by Miller are reanalysed using the present method.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Buckley et al. (1979) studied this question.

synapsesocial.com/papers/6a193932f2eb401dc788c9a2https://doi.org/10.1093/biomet/66.3.429
Ask AI
Helpful
Bookmark
Share
View Full Paper