Bookmark and Share

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


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

Re: st: Logistic regression interpretation - continued questions with ONY binary predictors


From   Michael Morrison <[email protected]>
To   [email protected]
Subject   Re: st: Logistic regression interpretation - continued questions with ONY binary predictors
Date   Wed, 22 Sep 2010 09:38:56 -0500

Hi Bill,

I'm not sure I understand the issue. With two binary predictors you have 4 basic combinations (1-1, 1-0, 0-1 and 0-0), producing 4 different logits and 4 different odds; exp(logit). With the 0-0 combination the logit is equal to the constant.

Mike




On 9/21/2010 6:24 PM, Dr. Bill Westman wrote:
Mike.  Thanks.  It seems that this book and others move to models that
are seem to be unrelated to a binary outcome and two binary predictors
(specifically).  With a second binary predictor, many of the
techniques suggested by Long&  Freese seem irrelevant.  There isn't
really a "mean" risk - it is either High or NOT.   The fact that I do
not have any continuous predictors makes this case easier -- I think,
but still perplexing to me.  If any listers have insight, please let
me know



On Tue, Sep 21, 2010 at 3:04 PM, Michael Morrison<[email protected]>  wrote:
I recommend you read Long&  Freese (2006) /Regression Models for Categorical
Dependent Variables Using Stata. It provides interpretation of odds,
_controlling for covariates_ and ways to test your model's goodness of fit.
Also of interest is their Spost commands.

Mike

/

On 9/21/2010 4:36 PM, Dr. Bill Westman wrote:
Dear Listers: I have more of a stats question than a stata question,
but here goes:

A simple logistic regression was run - outcome - Survived/Non Survived
(1/0), with Group(Treatmeant/Control -- 1/0)&    Risk(High/Low -- 1/0)

The results

------------------------------------------------------------------------------
        Survived | Odds Ratio   Std. Err.      z    P>|z|     [95%
Conf. Interval]

-------------+----------------------------------------------------------------
     Risk |   4.790295   2.589136     2.90   0.004     1.660699    13.81763
    group |   2.111181   .9107981     1.73   0.083     .9063649    4.917538

------------------------------------------------------------------------------

Assuming the .10 level of significance was acceptable (for discussion
purposes)

Can I assume that odds of treatment survival were 2.1  times higher in
the treatment Group for LOW risk patients?
And for High Risk patients in the treatment group is the adds 4. 8 +
2.1 (or 6 times higher)?

Other than test for interactions, can I run a graphs of these two groups?

Can I estimate the "average" effect of Treatment across both high and
low risk patients?  If so, how?

What other tests might I wish to do in Stata?

Thanks

Bill
*
*   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/





*
*   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/

*
*   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/





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