[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: linear probability model (LPM)

From   David Airey <[email protected]>
To   [email protected]
Subject   Re: st: linear probability model (LPM)
Date   Tue, 15 Jan 2008 09:14:21 -0600


I seem to have missed this thread, but this post suggests it was a little heated or somebody is feeling a little punchy.

Quite a few pages at the UCLA ATS Stata site describe interpretation of the logit/probit models in terms of predicted probabilities. That seems to also be the basis for a package from that group (Mitchell) on interpreting main effects and interactions in logit models (vibl), through graphics of the predicted probabilities. So even when using the logit models, thinking in terms of probabilities seems worthwhile, and doesn't require a linear probability model (with constraints to keep the predictions between 0 and 1). The Spost package and literature is also available to help sensibly and intelligently interpret logit and other categorical dependent variable outcome models.


On Jan 15, 2008, at 9:00 AM, Nick Cox wrote:

Well, Maarten is sensible, intelligent and a non-statistician...
Perhaps if you a non-statistician you need to be very sensible
and very intelligent to understand odds ratios...


Paul Seed

My experience is that even the most
sensible & intelligent & non-statistician
is likely to be defeated by odds ratios.
Time & again, I have known respected colleagues
present them as though they are risk ratios,
& they seem quite surprised when I point out the


Maarten buis

I disagree, both measures are perfectly understandable. With odds
ratios you are representing chance by at odds instead of probability.
Both are easy to understand: odds give you the expected number of
successes for every failure, while probability gives you the expected
proportion of successes. With odds ratios groups are compared by
calculating the ratio: e.g. the odds of success for men is 10% higher
than the odds of succes for women. With risk difference you are
comparing the groups by computing differences.
*   For searches and help try:
*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index