Logistic Regression

[#1]

I am trying to do a multivariate logistic regression and I wanted to know how do we chose a baseline (referent) to compare the odds to.

For example, I am comparing different medications, and I want to choose the referent and then have the odds come out comparing it

[#2]

Hi,
If you use the Fit Model > Nominal Logistic platform and select the Odds Ratio from the output window (red triangle at the top left), JMP will generate the odds ratios for all possible combinations of pairwise values of your nominal variable including their inverse. Hence, you do not need to nominate a specific treatment / drug / state to obtain the Odds Ratio for a specific combination.
Does that make sense or did I misinterpret your question?
Best,
TS

[#3]

Thank you for your message, but when I use SPSS, I am able to use a referent. 

Let me give you an example

If we are looking at the dependent variable that has 4 options for example, how can I select the referent from these 4.

For example, let us consider these 4 options as Drug A, B, C, D.  Indep variable let us say adverse event (yes/no). Can we look at odds ratio for B, C, D in reference to A

Like Drug B has 10 times odds of having an adverse reaction compared to drug A. 

Let me know what you think

[#4]

Hi,
I'm afraid that I may not understand you problem because when you run a logistic model as you described and ask JMP to calculate the Odds Ratios, JMP will list the following:
A vs B
A vs C
A vs D
B vs C
B vs D
C vs D
and the inverse comparisons. Hence, in your case, you would need to capture the A vs B, A vs C, and A vs D Odds Ratios to meet your goal.
Sorry if I'm not providing the help you need. Maybe, if you could share some screen capture of where you are experiencing a problem.
Best,
TS

[#5]

You can specify the target level of interest in the Fit Model launch dialog after selection logistic regression fitting personality. There is a drop-down menu of response levels in the upper right corner above the Run button.

The default target level is the first level. Order is determined alphanumerically or by one of the built-in lists (e.g., months of the year, days of the week, et cetera). You can impose your own order by adding the Value Order column property to the response data column before beginning your analysis.