Thanks for posting your syntax.  It looks generally okay to me, assuming that each variable listed in a CONTRAST sub-command is a categorical variable. (One might ask why you are treating things like age and BMI as categorical, but that's a different issue.)

In the original post, you asked about two things:

Regarding the first question, I wonder if you have empty categories for some of your categorical variables.  You can check that by generating a frequency table for each one of them. 

Re the second question, there should be a column headed Exp(B) in your table of coefficients.  That column displays adjusted ORs.  To get the corresponding 95% CIs, add this sub-command:  /PRINT = CI(95).

HTH.

Please post your command syntax, and indicate which variable you think ought to be treated as categorical.  My guess is that you didn't include the sub-command needed to indicate that it is a categorical variable. 

Dear Kulandaivelan

One way round this problem might be to turn your multinomial variable into several dummy binary variables and include them all in your model rather than using an automated feature (we always recommend doing this anyway). The only type of adjustment I am aware of to parameter and odds ratio estimates is bias corrected confidence intervals using bootstrapping.

Thanks Bruce for your help, i will contact you with my data soon. Thank you Peter for your inputs.

@ Kulandaivelan:  We don't need the data (at this point).  Just post your command syntax.  Thanks. 

@ Bruce: Is this command syntax? or else how will be get them from default SPSS commands?

Using this command i got individual sub-category OR also, but i don't know how to get adjusted OR...please help me...

In this example...i got significant results for age, sex, WC, education, income, hand-grip strength. now i want adjusted OR for hand-grip strength as age, sex, wc might influence hand-grip strength....thanks in advance for your valuable time sparing...

I believe the problem is that in SPSS you have to specify the categorical variables to be categorical (on the top of the right hand side there is a key named CATEGORICAL). Of course, if a variable has k categories, then one category will be the reference category and you will get k-1 regression coefficients (the reference category will have regression coefficient 0) and k-1 odds ratios (the OR for the reference category will be 1).

Thanks Francois.... i just did OR calculation and found values from my syntax for knee pain command is always lesser than simple calculation of OR.... whether are they adjusted, if yes for what? @ Bruce

@ Kulandaivelan:  If you are using the menu system to perform your analysis, just exit via the PASTE button when you are finished rather than clicking on OK.  This will paste the command syntax to a syntax editor.  You can then copy it, and paste it into a post here.  This will allow others to see whether you are in fact treating categorical variables as categorical.  (You will of course have to indicate in your post which variables are categorical and which are quantitative.) 

See section 1 of the page linked below for some more discussion of the Paste button. 

http://www.theanalysisfactor.com/3-spss-buttons-easier/

@ Bruce: I followed your instruction and got command similar to yesterday one. Please do see and give your opinion. Thanks in advance...

@ Bruce: Am i frustrating you? please help me to understand how to do adjusted OR from default menu system.

Thanks for posting your syntax.  It looks generally okay to me, assuming that each variable listed in a CONTRAST sub-command is a categorical variable. (One might ask why you are treating things like age and BMI as categorical, but that's a different issue.)

In the original post, you asked about two things:

Regarding the first question, I wonder if you have empty categories for some of your categorical variables.  You can check that by generating a frequency table for each one of them. 

Re the second question, there should be a column headed Exp(B) in your table of coefficients.  That column displays adjusted ORs.  To get the corresponding 95% CIs, add this sub-command:  /PRINT = CI(95).

HTH.

Thanks Bruce

Hi everyone,

i have issue on how to compute AOR in SPSS. I am doing a study on prevalence of hypertension and it risk factors. i got the OR to determine the risk factors of hypertension such as age, marital status, occupation, smokeless tobacco use and BMI. But the issue lies on how to calculate the AOR please I need guidance on how to go about this.

Thank you

I must sadly and repeatedly say that there is no Normal Distribution or Gaussian Distribution in this Universe or Nature at all.

Then, Non-parametric Statistics should be utilized in the Scientific Articles (please see file; Dr. Terentyeva Urine BIN).

I have recently found that the Active Transport of Fucoidan obeys the Non-parametric Distributions and the Median values indicates the true results; i.e., unexpectedly the Average or the Mean value usually gives false results (please see file; JCB Fucoidan/Fucoidin Transport).

By the way, Albert Einstein and Karl Schwarzschild have falsely estimated that this Universe may be Symmetric and may obey the Normal Distribution. However, this Universe obeys to the Non-parametric Distribution. Therefore, notorious black hole is not present in this Universe (my unpublished estimation).

Following