On Testing Dependence between Time to Failure and Cause of Failure when Causes of Failure Are Missing
2007

Testing Dependence Between Time to Failure and Cause of Failure

publication Evidence: moderate

Author Information

Author(s): Dewan Isha, Kulathinal Sangita

Primary Institution: Indian Statistical Institute, New Delhi, India

Hypothesis

The hypothesis of independence between the failure time and the cause of failure is studied using conditional probabilities.

Conclusion

The proposed tests perform satisfactorily and recommend using failure time data even when corresponding causes are missing.

Supporting Evidence

  • The tests based on U-statistics perform well even when 20% of the causes of failure are not known.
  • The one-sided test based on Kendall's τ is more efficient than the test based on UPQDm.
  • The empirical powers for all three tests are higher in the case of the parametric family of distributions compared to the random sign censoring model.

Takeaway

This study looks at how to tell if the time it takes for something to fail is related to the reason it fails, even when we don't know the reason for some failures.

Methodology

Tests based on U-statistics were proposed to test for independence of failure time and cause of failure in a competing risks model.

Limitations

The study assumes that causes of failure are missing completely at random.

Statistical Information

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1371/journal.pone.0001255

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication