Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: S-Curve/logistic curve regression


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: S-Curve/logistic curve regression
Date   Wed, 1 Apr 2009 18:00:19 +0100

Some people would use -nl- here, but assuming that you can scale values
to values / maximum, you can still use a logit link, or linear
regression on a logit transformed variable. The only obvious problem is
if the bounds are attained, because logit(0) and logit(1) are
indeterminate. 

There are some related comments in 

SJ-8-1  gr0032  . . . . . . .  Stata tip 59: Plotting on any transformed
scale
        . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  N.
J. Cox
        Q1/08   SJ 8(1):142--145                                 (no
commands)
        tip on how to graph data on a transformed scale

which explains how to plot continuous responses on a logit scale. 

Nick 
n.j.cox@durham.ac.uk 

Sven Wiese

I have a twodimensional dataset of two continious variables and want to
fit a regression to the data, but I want the outcome to be an s-Curve
(or logistic curve). I think it might be like fitting a polynomial to
the data, but as mentioned I want to fit an s-curve, which can be
expressed by the exponential function.
I am aware of the logistic regression, the difference is, that my
dependant variable is not binary.

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index