While many JMP methods are described or identified in platform output (e.g., Nonlinear saying "Solved by Analytic Gauss-Newton") others are not. When different JMP platforms give different results, understanding and explaining that difference is difficult. My client (US government) uses confidence limits for policy decisions, so I need to know why the methods differ. Example (JMP 16.0.0): Nonlinear vs. Fit Curve, both using Exponential 3P model, even though the corresponding parameter (and Std Err) estimates agree to the number of digits shown. Extremely useful would be JMP having a central Help page where one can easily locate the identification of the statistical method employed. Including a reference and/or summary of the statistical details would also be helpful.
Nonlin vs FitCurve using Exp3P.jrn
