Confidence intervals for individual slopes

[StefanC]

In an unequal slopes ANCOVA model the individual slopes can be estimated with the Estimates > Custom test or Estimates > Compare slopes option in the output for Fit model. Standard errors and p-values are also given but I miss confidence intervals for the slopes. This should also be possible if additional random effects are added to the model (a mixed model).

[kayleewarren]

@StefanC Thank you for your suggestion! We have captured your request and will take it under consideration.

[SamGardner]

@StefanC this is an interesting idea.  Can you provide some additional details on what type of problem you would use this for and why it is important to you?  A data example of where you would like to do this would also be helpful.  

[StefanC]

@SamGardner Thank you for considering my suggestion. It would be useful for analysing stability data where a common slope across batches cannot be assumed. In this case the individual slope estimates including confidence intervals would be of great interest to estimate the worst case degradation. I am aware of the degradation platform but to my knowledge it does not give confidence intervals? Also, sometimes you want to include a random effect to the model, like an assay effect for example.  

[SamGardner]

Using the sample data file Stability.JMP, if you fit the ANCOVA model, you get this output

If you right click on the Parameter Estimates Table, you can turn on the display of the confidence intervals for the estimates, which includes the individual slopes

You can also do this if you include random effects in both Fit Least Squares and in JMP Pro's Fit Mixed platform.  

Degradation > Stability Test does not show confidence intervals for the parameter estimates, but you can now use the pooled MSE (in v17) for the stability predictions and for the ICH shelf-life calculations.

[StefanC]

Thanks, but that is not quite what I am looking for. In order to get the individual slope estimates, you need to add the parameter estimate of Time to the parameter estimate of Batch*Time which can be done with the Estimates > Custom Test option

It would be nice to be able to include confidence intervals of the estimate to the output. In this simple case I should be able to calculate the confidence intervals myself from the Std error and degrees of freedom but it is still inconvenient. My version of JMP is 16.0.