Stat 567:Statistical Reliability
Twenty Second Class (December 1, 1997)

Weibull Anova

Class Objectives:

• Show how ignoring sources of variation can be misleading in reliability data analysis
• Demonstrate how to handle categorical variable in Weibull regression

Homework Assignments:

• Read the class notes

Class Outline and Main Points:

• Gerstle and Kunz (1983) Example
  • Failure times of Kevlar 49 fibers loaded at four different stress levels
  • They found significant variation from spool to spool
• Analysis without considering spool effect
  • MLEs
  • Residual plot versus spool
• Analysis considering spool effect
  • Likelihood ratio test for spool effect

Study Questions

• Why is it important to consider important sources of variation in reliability data analysis?
• Why the spool effects obtained with JMP are different from the ones obtained in the book?

Do you have something to tell me?