Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: Paradoxical RRR with logit and auto.dta


From   Michael Ingre <Michael.Ingre@ipm.ki.se>
To   statalist@hsphsun2.harvard.edu
Subject   st: Paradoxical RRR with logit and auto.dta
Date   Mon, 1 Nov 2004 18:41:59 +0100

Dear list

Paradoxes do exist, but this one might be attributed to my ignorance. Below is an example using the auto.dta that to me at least, is difficult to understand. Two -logit- models are fitted and then, Relative Risk Ratios (RRR) are calculated using -nlcom- that contrasts the probability at different levels of the independent variable. In the first model, the greater the contrast the greater the significance. However, in the second model, the opposite apply.

If anyone could explain this apparent paradox for me, or point me to my mistakes, I would be very happy.

Thanks,

Michael

Paradoxical code below (output omitted):

sysuse auto

logit foreign mpg

// Contrast probability of mpg=1 with mpg=10
nlcom (exp(1 * _b[mpg] + _cons) / (1+ exp(1 * _b[mpg] + _cons))) / ///
(exp(10 * _b[mpg] + _cons) / (1+ exp(10 * _b[mpg] + _cons))) //

// Contrast probability of mpg=1 with mpg=100
nlcom (exp(1 * _b[mpg] + _cons) / (1+ exp(1 * _b[mpg] + _cons))) / ///
(exp(100 * _b[mpg] + _cons) / (1+ exp(100 * _b[mpg] + _cons))) //

// Paradoxical model
logit foreign length

// Contrast probability of length=1 with length=10
nlcom (exp(1 * _b[length] + _cons) / (1+ exp(1 * _b[length] + _cons))) / ///
(exp(10 * _b[length] + _cons) / (1+ exp(10 * _b[length] + _cons))) //

// Contrast probability of length=1 with length=100
nlcom (exp(1 * _b[length] + _cons) / (1+ exp(1 * _b[length] + _cons))) / ///
(exp(100 * _b[length] + _cons) / (1+ exp(100 * _b[length] + _cons))) //






*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* 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