I want to test a nonlinear hypothesis. More specifically, in a logistic regression, I would like to calculate the standard error of the ratio of two parameters. I know how to approximate this manually from the parameter covariance matrix using the Delta method. But can this be done directly in the procedure (or indeed has anyone written macros to do this sort of non-linear testing of parameters).
Why not just include the ratio as a third derived predictor in the model specification? That way, you could test it using the available statements. If it presents no strong evidence for predicitng your response, why bother testing it at all? Or am I missing something?
I am interested in the ratio of parameter estimates, not the effect of a ratio variable as predictor. That is,
y = G(a + bX + cY)
I am interested in the standard error of b/c.
I found this interesting code from Russell Millar. He uses proc NLMIXED to calculate the delta method for arbitrary (2 variable) functions. Once I get my head around how nlmixed works, I might try generalising this to work from the COVB matrix output.
https://www.stat.auckland.ac.nz/~millar/730/ComputerCode/DeltaMethodMacro.sas
Don't miss out on SAS Innovate - Register now for the FREE Livestream!
Can't make it to Vegas? No problem! Watch our general sessions LIVE or on-demand starting April 17th. Hear from SAS execs, best-selling author Adam Grant, Hot Ones host Sean Evans, top tech journalist Kara Swisher, AI expert Cassie Kozyrkov, and the mind-blowing dance crew iLuminate! Plus, get access to over 20 breakout sessions.
ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.
Find more tutorials on the SAS Users YouTube channel.