Hosmer-Lemeshow Test

[#1]

I need to conduct this test on some data predicting criminal offending status (yes or no) from age. Can anyone advise how to do this efficiently in JMP?

[#2]

Hello, I was wondering if you had figured out how to do it. I need to do the same on a logistic regression model...

Any help would be greatly appreciated!

[#3]

Greetings, profjmb and PdA!

Your best bet is probably using JMP's ability to run R.  You can, of course, calculate the Hosmer-Lemeshow goodness of fit test yourselves with formulae or a JSL script using the saved probabilities and expected values.

But, before you sink too much effort into that quest, you should be aware that even Hosmer, Lemeshow, and others wrote an article in 1997 (attached) about their 1980 test, noting many of the problems with it.  Others have found more problems (attached).  The problems are fairly disturbing, and I believe the classical Hosmer-Lemeshow test is now deprecated by serious logistic regressioners.

Good luck!