# RE: st: log binomial regression - categorical independent variables

 From "Kieran McCaul" To Subject RE: st: log binomial regression - categorical independent variables Date Fri, 21 Aug 2009 08:41:31 +0800

```...

The -glm- command you have used is:
xi:glm IOTFnew highested_2 i.highested_2*IOTFnew, family(binomial)

If this is indeed the syntax you have used, then there are a couple of
obvious problems.

1. The dependent variable is IOTFnew, but this also appears as an
independent variable.
2. The independent variable, highested_2, is entered both as a
continuous variable and as a categorical variable, i.highested_2.

So I would suggest you try:

xi:glm IOTFnew i.highested_2, family(binomial) link(log) eform

______________________________________________
Kieran McCaul MPH PhD
WA Centre for Health & Ageing (M573)
University of Western Australia
Level 6, Ainslie House
48 Murray St
Perth 6000
Phone: (08) 9224-2701
Fax: (08) 9224 8009
email: Kieran.McCaul@uwa.edu.au
http://myprofile.cos.com/mccaul
http://www.researcherid.com/rid/B-8751-2008
______________________________________________
If you live to be one hundred, you've got it made.
Very few people die past that age - George Burns

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Amy Jennings
Sent: Thursday, 20 August 2009 9:43 PM
To: statalist@hsphsun2.harvard.edu
Subject: RE: st: log binomial regression - categorical independent
variables

Hi,
Many thanks for your reply it was greatly appreciated- I have tried your
suggestion but it does not seem to be working.

This if the syntax I have used:
xi:glm IOTFnew highested_2 i.highested_2*IOTFnew, family(binomial)

IOTFnew is the dependent variable (child weight status - 2 categories)
and highested_2 the independent (4 categories)
and below is the response I am getting

i.highested_2 _Ihighested_1-4 (naturally coded; _Ihighested_1 omitted)
i.high~2*IOTF~w _IhigXIOTFn_# (coded as above)
note: _Ihighested_4 dropped because of collinearity
Iteration 0: log likelihood = -130.92025 (not concave)
Iteration 1: log likelihood = -130.92025 (not concave)
Iteration 2: log likelihood = -130.92025 (not concave)
Iteration 3: log likelihood = -130.92025 (not concave)
Iteration 4: log likelihood = -130.92025 (not concave)
Iteration 5: log likelihood = -130.92025 (not concave)
Iteration 6: log likelihood = -130.92025 (not concave)
Iteration 7: log likelihood = -130.92025 (not concave)
Iteration 8: log likelihood = -130.92025 (not concave)
Iteration 9: log likelihood = -130.92025 (not concave)

Not sure where I am going wrong? I have tried a few variations but don't
seem to be getting very far,

Many thanks again,

kind regards,

Amy

----------------------------------------
> Date: Thu, 20 Aug 2009 15:00:33 +0200
> Subject: Re: st: log binomial regression - categorical independent
variables
> To: statalist@hsphsun2.harvard.edu
>
> Amy Jennings wrote:
>> Hi
>>
>> I am currently using stata to estimate prevalence ratios (or risk
ratios as I think stata refers to them) using log binomial regression
with the glm function. I am using the syntax
>>
>> glm dependent independent, family(binomial) link(log) eform
>>
>> this has worked fine but I would now like to add a categorical
independent variable with more than two categories, and obtain
coefficients for each of the categories, e.g. I am trying to calculate
prevalence ratios for childhood overweight compared with normal weight
for increasing category of parental education, is it possible to
estimate prevalence ratios for children whose parents are educated to
school-level or university-level verses those with no education?
>>
>> _________________________________________________________________
>> Windows Live Messenger: Happy 10-Year Anniversary-get free winks and
emoticons.
>> http://clk.atdmt.com/UKM/go/157562755/direct/01/
>> *
>> * 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/
>>
> Hi,
>
> Maybe what you want is:
>
> xi:glm dependent independent... i.schooling*overweight,
>
> Schooling is assumed to have various categories and stata will keep
the first one as reference. You can however control which one is used as
a reference group.
>
> Hope this helps,
>
> Evans
>
>
>
>
> *
> * 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/
_________________________________________________________________

Upgrade to Internet Explorer 8 Optimised for MSN.

http://extras.uk.msn.com/internet-explorer-8/?ocid=T010MSN07A0716U
*
*   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/
```