Odds Ratio

[waqas imran]

May any body tell me the Command to calculate unadjusted and adjusted ODDS Ratio in SPSS. I have earch alot in the software but i haven't found it yet. Help me please.

Thank You

[Bruce Weaver]

waqas imran wrote

May any body tell me the Command to calculate unadjusted and adjusted ODDS Ratio in SPSS. I have earch alot in the software but i haven't found it yet. Help me please.

Thank You

Look up LOGISTIC REGRESSION.  Exp(B) = the odds ratio.  for unadjusted (or crude) odds ratios, run a model with only one explanatory variable; for adjusted odds ratios, add the variables you wish to adjust for to the model.

[Bruce Weaver]

Bruce Weaver wrote

waqas imran wrote

May any body tell me the Command to calculate unadjusted and adjusted ODDS Ratio in SPSS. I have earch alot in the software but i haven't found it yet. Help me please.

Thank You

Look up LOGISTIC REGRESSION.  Exp(B) = the odds ratio.  for unadjusted (or crude) odds ratios, run a model with only one explanatory variable; for adjusted odds ratios, add the variables you wish to adjust for to the model.

I should have added that you must have exactly the same cases going into the two models.  One way to ensure that is to run a single LOGISTIC REGRESSION command with multiple /ENTER sub-commands.  E.g.,

LOGISTIC REGRESSION VARIABLES Y
  /METHOD=ENTER X
  /METHOD=ENTER C
  /PRINT=CI(95)
.

Here, Y is the binary dependent variable (with values of 0 or 1); X is the explanatory variable of primary interest; and C is some other variable you wish to control for.  For this syntax, both X and C are continuous variables.  

For any categorical explanatory variables, you'd have to add /CONTRAST sub-commands.  E.g., the following syntax treats both X and C as categorical (with the first level of each as the reference category).

LOGISTIC REGRESSION VARIABLES Y
  /METHOD=ENTER X
  /METHOD=ENTER C
  /CONTRAST (X)=Indicator(1)
  /CONTRAST (C)=Indicator(1)
  /PRINT=CI(95)
.

[Ryan]

Click on the link below for relevant information regarding logistic regression:

http://www.listserv.uga.edu/cgi-bin/wa?A2=ind0705C&L=sas-l&P=R12617

Ryan

On Tue, Dec 21, 2010 at 4:04 PM, Bruce Weaver <[hidden email]> wrote:

> Bruce Weaver wrote:
>>
>>
>> waqas imran wrote:
>>>
>>> May any body tell me the Command to calculate unadjusted and adjusted
>>> ODDS Ratio in SPSS. I have earch alot in the software but i haven't found
>>> it yet. Help me please.
>>>
>>> Thank You
>>>
>>
>> Look up LOGISTIC REGRESSION.  Exp(B) = the odds ratio.  for unadjusted (or
>> crude) odds ratios, run a model with only one explanatory variable; for
>> adjusted odds ratios, add the variables you wish to adjust for to the
>> model.
>>
>>
>
> I should have added that you must have exactly the same cases going into the
> two models.  One way to ensure that is to run a single LOGISTIC REGRESSION
> command with multiple /ENTER sub-commands.  E.g.,
>
> LOGISTIC REGRESSION VARIABLES Y
>  /METHOD=ENTER X
>  /METHOD=ENTER C
>  /PRINT=CI(95)
> .
>
> Here, Y is the binary dependent variable (with values of 0 or 1); X is the
> explanatory variable of primary interest; and C is some other variable you
> wish to control for.  For this syntax, both X and C are continuous
> variables.
>
> For any categorical explanatory variables, you'd have to add /CONTRAST
> sub-commands.  E.g., the following syntax treats both X and C as categorical
> (with the first level of each as the reference category).
>
> LOGISTIC REGRESSION VARIABLES Y
>  /METHOD=ENTER X
>  /METHOD=ENTER C
>  /CONTRAST (X)=Indicator(1)
>  /CONTRAST (C)=Indicator(1)
>  /PRINT=CI(95)
> .
>
>
>
> -----
> --
> Bruce Weaver
> [hidden email]
> http://sites.google.com/a/lakeheadu.ca/bweaver/
>
> "When all else fails, RTFM."
>
> NOTE: My Hotmail account is not monitored regularly.
> To send me an e-mail, please use the address shown above.
>
> --
> View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Odds-Ratio-tp3313037p3314100.html
> Sent from the SPSSX Discussion mailing list archive at Nabble.com.
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[katarina]

Hello!!

 I have one question! I need to calculate adjusted odds ratio, but I have four variables I want to adjust for to the model. i wondering am i  doing things properly: I only put four categorical variables as independent  variable , and run the binary logistic ratio. I sthat ok?
Thanks in advance!

[waqas imran]

Yes Ms Katarina your are on right path. Adjustment means u have to see the net effect of a number of independent catagories simultaiously on dependent binary or multinomial outcome variable. Whether its catagorical or any other form in case of adjusted ODS ratio, you have run the equation alond with set of all independent variables at the same time in a model. For unadjusted you have to run the equation separately for each independent variable or subset of variables from a complete set of alll independent variables.

For both situations your interpretaions will be different.

You can ask me everything that confuses you regarding logistic regression.

Waqas Imran,
Research Associate,
NIPS

[katarina]

Thank you! I ll be free to contact you again if i have some problems with logistic regression!

 Best regards!

 Katarina

[waqas imran]

Its my pleasure to help others

[romibri]

Hi,

All the advice you are giving is really helpful! I am currently working with logistic regression and was wondering if you could help me. Nobody in my office uses SPSS except for me, so I have nobody to ask.

When we talk about bivariate analysis, are we refering to running the logistic regression analysis with the outcome and each independent variable separately? (without controlling for any variable). If so, is this what we would call "unadjusted odd ratios"? I know people run bivariate analysis using chi square tests, so I get really confused.

Also, when you run a multivariate logistic regression model with the outcome and all the independent variables at once, what the output shows in "variables in the equation" is the odds ratio of the outcome for each independent variable and controlling for the rest? For example, if you have gender, age, and race as covariates (indep variables) and use of drugs as the outcome/dependent variable, would you be able to get odd ratios for the odds of using drugs if you're female controlling for age and race, the odds of using drugs if you are white controlling for gender and age, etc.??

Thank you so much in advance for your help!! I really need help with this.