Adjusted odds ratio

[emmyb...@yahoo.com]

I need to compute odds ratios determining the association between one
variable (i.e. uterus shape) and an outcome (i.e. infertility), while
adjusting for a confounding variable (smoking status). Any help would
be appreciated. I am most familiar with using the dialog boxes, not
typing commands. I don't typically use SPSS (I use STATA or SAS) so I
am not very familiar with it. Thanks in advance!

[Bruce Weaver]

Have you tried it yet? The GUI is pretty straight forward. Click
on the "Categorical" button in the main dialog to set up how you
want to deal with categorical predictor variables (e.g., indicator
variables with first or last category as the referent). In
another sub-dialog (possibly under Options?), you have to check a
box to get 95% CIs for Exp(B) in the output.

--
Bruce Weaver
bwe...@lakeheadu.ca
http://sites.google.com/a/lakeheadu.ca/bweaver/
"When all else fails, RTFM."

[Bruce Weaver]

On Sep 29, 6:56 pm, Bruce Weaver <bwea...@lakeheadu.ca> wrote:

> emmybet...@yahoo.com wrote:
> > I need to compute odds ratios determining the association between one
> > variable (i.e. uterus shape) and an outcome (i.e. infertility), while
> > adjusting for a confounding variable (smoking status).  Any help would
> > be appreciated.  I am most familiar with using the dialog boxes, not
> > typing commands.  I don't typically use SPSS (I use STATA or SAS) so I
> > am not very familiar with it.  Thanks in advance!
>
> Have you tried it yet?  The GUI is pretty straight forward.  Click
> on the "Categorical" button in the main dialog to set up how you
> want to deal with categorical predictor variables (e.g., indicator
> variables with first or last category as the referent).  In
> another sub-dialog (possibly under Options?), you have to check a
> box to get 95% CIs for Exp(B) in the output.
>

The dialogs I was referring to there are for logistic regression. But
it turns out the OP was trying to obtain the adjusted odds ratio via
CROSSTABS. If there is only one additional categorical covariate,
that can be done by asking for the Mantel-Hanszel chi-square. There
is an example here:

http://www.angelfire.com/wv/bwhomedir/spss/pooling_odds_ratios.txt

[alina]

I am trying to do the same things but I want two get the adjusted odd ratio for more cofoun ders like age, area of residence and schoolig level. I use SPSS but I don't understand how to get adjusted odd ratio from logistic regresion without sintax. Is it possible to do so?

[Bruce Weaver]

Analyze - Regression - Binary Logistic

Then click Help, followed by Show Me.

The adjusted OR for your variable of interest is in the Exp(B) column in
the table of regression coefficients.

--
Bruce Weaver
bwe...@lakeheadu.ca
http://sites.google.com/a/lakeheadu.ca/bweaver/Home

[micah...@gmail.com]

On Monday, September 29, 2008 6:14:39 PM UTC-4, (unknown) wrote:
> I need to compute odds ratios determining the association between one
> variable (i.e. uterus shape) and an outcome (i.e. infertility), while
> adjusting for a confounding variable (smoking status). Any help would
> be appreciated. I am most familiar with using the dialog boxes, not

> typing commands. I don&#39;t typically use SPSS (I use STATA or SAS) so I

> am not very familiar with it. Thanks in advance!

We are also trying to adjust for age and sex, but are unsure of what to use for age in the analysis... our age data is 19, 20, 21, etc, or we can find a mean age, but we are unsure if there is another way to change age so as to find the OR (for example, a binomial outcome?) Thanks!

[Bruce Weaver]

It appears you have people's actual ages, so why would you not simply
enter age as a continuous explanatory variable? Carving it into
categories us almost always a bad idea, because you throw away
information. Google <Streiner breaking up is hard to do> to find a nice
article on that.