Failure Discounting

[DManuel]

Does anyone know of a statistical method for Failure discounting (i.e. applying effeciveness values) in the Crow-Extended model?
The RGA Reference book shows a method when using the Lloyd-Lipow, Logistic and Gompertz models, but not the Crow-Extended.
Thank you,

[David]

Hi DManuel,

When using the Crow Extended model, effectiveness factor (EF) values must be supplied for the BD modes. BD modes are modes that will receive a delayed fix at the end of the test. EF values are not determined by the software, but by engineering judgment.

Additional information on the Crow Extended model and EF values can be found at http://www.weibull.com/RelGrowthWeb/Crow_Extended.htm#background.

I hope this helps.

[DManuel]

Thanks David,
I was trying to determine if there was a statistical method like that used in the Lloyd-Lipow, Logistic and Gompertz models.
We currently use the SWAG method also.

It seems that a good statistician could come up with a method to "refine" the effectiveness from data in subsequent periods (like that in the other models). I'm not one, so I thought that I would throw that out!

D