Simpson's paradox in survival models

Clelia Di Serio, Yosef Rinott, Marco Scarsini

Research output: Contribution to journalArticlepeer-review


In the context of survival analysis it is possible that increasing the value of a covariate X has a beneficial effect on a failure time, but this effect is reversed when conditioning on any possible value of another covariate Y. When studying causal effects and influence of covariates on a failure time, this state of affairs appears paradoxical and raises questions about the real effect of X. Situations of this kind may be seen as a version of Simpson's paradox. In this paper, we study this phenomenon in terms of the linear transformation model. The introduction of a time variable makes the paradox more interesting and intricate: it may hold conditionally on a certain survival time, i.e. on an event of the type [T>t] for some but not all t, and it may hold only for some range of survival times.

Original languageEnglish
Pages (from-to)463-480
Number of pages18
JournalScandinavian Journal of Statistics
Issue number3
Publication statusPublished - Sep 2009


  • Cox model
  • Detrimental covariate
  • Linear transformation model
  • Omitting covariates
  • Positive dependence
  • Proportional hazard
  • Proportional odds model
  • Protective covariate
  • Total positivity

ASJC Scopus subject areas

  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Fingerprint Dive into the research topics of 'Simpson's paradox in survival models'. Together they form a unique fingerprint.

Cite this