post-hoc analyses when using non-linear regression

[#1]

I have data from a dose-response experiment (pharmacology) to which I fitted a modified Michael-Menten model using non-linear regression (--> specialized modeling --> nonlinear (model = A*(1-exp(-k*dose))^c + intercept)). In this experiment the accent was put on testing many different levels; the number of observations for an individual level is limited (insufficient for a direct analysis of variance). 

How can I compare if two levels of my input parameter (two doses of the test product) yield significantly different responses from each other? Using 'custom estimate' I can predict levels, but when comparing two dose levels with this I have no P values nor error terms (e.g., A*(1-exp(-k*2400))^c - A*(1-exp(-k*1200))^c). Thanks in advance. 

[#2]

Hi @tvkmez11 ,

I think what would really help would be gaining access to the data table to look at. My first thought would be to save the differences between the doses using the profiler - I have an example where I'm look at the response change with a process where Temperature is altered, I can activate the profiler option in the Non-Linear, set my temperature (or dose) values and 'Remember Settings', this will save the Potency value with the two temperature settings.

You can right click > Make combined data table and then look at stacking the Rel Potency (or your Y) values together, then performing some sort of group comparison like a T-test using Fit Y by X.

Thanks,

Ben