Inverse prediction from Fit curve, Fit model and Degradation

[#1]

Hi,

I have used the inverse prediction from three platforms (Fit Curve, Fit Model and Degradation).

The predicted value is the same, but the confidence interval is not. Fit Model and Degradation give the same interval, but this interval is wider than the one reported by the Fit Curve platform.

I have attched an example with mackup degradation data in which I make linear a linear fit. I want to predict the time to a cartain value.

Why is there a difference? Which one should consider?

[#2 · Accepted solution]

The fit least squares platform uses t values and the fit curve platform uses wald chisquare.