I'm a newbie when it comes to SPSS and Statistics, and in need of some help.
I would like to know how to calculate a multi level factor odds ratio in SPSS, for an outcome (0 vs 1 : no disease vs disease) according to X protein serum levels (normal, borderline, high). And also if there's any way to compare these odds ratio to another Y and Z protein.
Thanks in advance, Carlos
I think the multilevel-analysis tag here is not indicating what you think it is or what you want. "Multilevel analysis" is typically used to refer to analyses involving data measured at different levels in hierarchies, such as repeated observations on the same subjects or basic units, and/or where these basic units are nested within larger entities, such as clinics or hospitals, or classrooms or schools, etc.
The first question seems like a simpler problem, that of obtaining odds ratios for a binary outcome and a three-level categorical predictor or factor (one with multiple levels). The way the primary question is worded seems to request a single odds ratio for this, but that doesn't exist unless you're willing to assume a linear relationship across the levels of that predictor. If you run a logistic regression with the predictor specified as a categorical factor, you can get two odds ratios comparing (for example) borderline vs. normal and high vs. normal. Odds ratio estimates in SPSS Statistics logistic procedures typically are in a column labeled Exp(B) in the parameter estimates output.
The additional question about comparing to other proteins depends on whether you're doing this all on the same cases (you have both or all types of proteins measured for each case) or not (each case has only one type of protein measurement). Both situations can be analyzed in SPSS Statistics, but using somewhat different approaches.
