Hi!

Apologies if this is a simple question but would appreciate some help!

I'm looking at reasons women delay seeking help for cancer symptoms. I have 13 different reasons & some are significantly correlated with delay in three of my cancer symptoms.

I have also collected demographic & lifestyle data (including age, marital status, frequency of exercise, smoker/non-smoker, frequency & amount of alcohol consumed & whether the participants have experienced the symptom I am asking about).

I'm trying to run a logistical analysis as my data are non-parametric & also categorial rather than continuous.

My question is about running the analysis in SPSS v.14

I have gone to binary logistic (as my data is binary) and have added delay for the first symptom in the dependent variable. I have then added those demographic & lifestyle variables that ARE significantly correlated with delay for that cancer symptom.

My question is: Do I just have to run the regression firstly with the demographic variables that are correlated & then run a separate regression with the reasons for delay (which are coded in a binary fashion)?? Or can I put them all in together!?

Is there anything else I need to be aware of?!

Thanks for the help & sorry if it's a bit waffly! :D

missy

You should run it together as otherwise you will not hve contorlled for either group in the other group. Meaning you have a reason to avoid seeking cancer treatment e.g. time and then you also have the demographic low socioeconomic status which probably are highly correlated and thereby should be in the same model to control for them. But be aware you do not have an unlimited number of predictors. Only a 10th of the outcomes (delay in seeking care), so if you have a 100 women delaying care then you can only use 10 predictors.
sorry forgot to mention the binary variables you have to tell SPSS that they are categories and NOT continuous, otherwise you will not get the correct answer. Most of your data seems to be in categories e.g. married yes/no and Socioeconomic status so all of those need to be defined as categories and the other continous ones such as age you just put in the model.
Hope that helps