Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

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

Re: st: Confidence intervals from -margins-

From   Roger Newson <>
To   "" <>
Subject   Re: st: Confidence intervals from -margins-
Date   Thu, 29 Apr 2010 21:35:44 +0100

A possible fix, in your case, might be to use -nlcom- after -margins-, to get confidence limits for the logit, or possibly the log, of the predicted probability. I regularly use this method to get sensible confidence intervals for population attributable fractions (PAFs).

I hope this helps.

Best wishes


Roger B Newson BSc MSc DPhil
Lecturer in Medical Statistics
Respiratory Epidemiology and Public Health Group
National Heart and Lung Institute
Imperial College London
Royal Brompton Campus
Room 33, Emmanuel Kaye Building
1B Manresa Road
London SW3 6LR
Tel: +44 (0)20 7352 8121 ext 3381
Fax: +44 (0)20 7351 8322
Web page:
Departmental Web page:

Opinions expressed are those of the author, not of the institution.

On 29/04/2010 21:20, Eric Johnson wrote:
Hello all,

I'm transitioning some old code to Stata 11, and I noticed that -margins- is producing negative predicted probabilities in its confidence intervals after a logistic regression. Its values match up exactly with -adjust- on the log odds scale (using the xb option), but not on the predicted mean scale.

It looks like -margins- is assuming the predicted probabilities follow a normal distribution and gets its confidence intervals there, while -adjust- works on the log scale, then exponentiates the confidence bounds.

Am I missing something here? I'd prefer to use margins, but adjust is providing the correct answer as far as I can tell.


*   For searches and help try:
*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index