Confidence Intervals in Fit Curve Platform

[#1]

I feel like this question has been asked before, but I can't seem to find the discussion. I am fitting a 2P Exponential decay model to my data in the Fit Curve platform. The prediction profiler for this fit has confidence intervals but there is no option to save the confidence intervals, or their formulas.

I can think of 2 options to get confidence intervals:

1. Save the predicted values and the standard error of the predicted values from the exponential 2P model. Use these to calculate 95% confidence intervals.

2. Save the parametric prediction formula and use this in the Nonlinear platform, where you have the option to save confidence limits (and their formulas).

These 2 options yield the same values for my 95% confidence interval. See the last 4 columns of my attached data table for this comparison. However, this set of confidence intervals differs from the ones plotted in the Fit Curve platform (see pictures below).

Is there any way to get the confidence intervals used in the fit curve platform? And also, does anyone know why they would be different? My guess is in Fit Curve platform either some different estimation method is being used, or perhaps a different estimate of the degrees of freedom is being used.

Thanks,

Robert

[#2 · Accepted solution]

Hi @rcast15 : It looks like perhaps Fit Curve uses a t critical value for the interval, while the other uses z (normal) critical value. This can be seen by taking the ratio of the two ranges and observing it is the same as t(0.975,df=15)/Z(0.975) = 2.13145/1.95996.

[#3]

@MRB3855 Ah ya that's gotta be it. Good find thanks!