Right censored data analysis

[#1]

I have tried going through various posts regarding how to deal with data censoring situations but i havent been able to figure out the best way to deal with my situation. So here it is...

We conducted a DoE with three variables: a type of surfactant (Surf), amount of water (Water) and amount of salt (Salt) and the response is the 'minimum amount of surfactant which is effective in the process (Surf_min). Ultimately we are looking to develop a model which calculates Surf_min for a given Surf, Water and Salt.

For some of the conditions in the DoE, we couldnt go to high enough concentrations where the surfactant is effective. This number could be different under different conditions. To me this is classic case of right censoring. Now, the question is how do i analyze this data set?

I have looked at the 2017 discovery summit talks and the related posting here. It seems that i could use the 'fit model' platform with 'generalized regression' option with JMP pro. I only have the usual JMP (13.1).

Will appreciate any inputs here. I would also appreciate if someone can point me to a good reference to have on this subject so that i can learn a bit more.

Thanks

[#2 · Accepted solution]

If you have JMP Pro, then you can use Generalized Regression through Analyze > Fit Model. See the chapter about this analysis in Help > Books > Fitting Linear Models.

If you have JMP, then you can use Parametric Survival, also through Analyze > Fit Model. See the chapter about this analysis in Help > Books > Reliability and Survival.

[#3]

Thanks so much!

Was able to make much progress on the analysis.

[#4]

This Discovery presentation is a good resource for understanding how to use Generalized Regression in JMP Pro for censored data. Skip to 19:27 for a demonstration. In this case the response is left-censored because some observations are below limit of detection.