Which test to use? Two continuous variables vs one categorical variable.

#1
Hello All,

I am PhD student in the throes of data analysis and writing up. I’d like to ask for your expert opinions on which statistical test/s would be most appropriate for a particular problem, please.

I’m trying to test the relationship for two continuous variables (carbon [d13C] and nitrogen [d15N] stable isotope values for n=174 individuals) and one categorical variable (presence/absence of a certain pathological condition in those individuals [condition present for n= 12, and condition absent for n= 162]. More broadly, the question I’m trying to address with this data: does diet (represented by the d13C and d15N values) predict the likelihood of having a certain pathological condition?

Thanks so much for any advice.

PS – I’ve been reading around and it seems like maybe a multinomial logistic regression might be an appropriate choice… and I have tried one… but I’m just really not sure or confident about this, and would very much appreciate hearing the opinion of some folk who are much better with stats than me! Many thanks.
 

hlsmith

Less is more. Stay pure. Stay poor.
#2
What is your data structure, meaning did you collect C and N variables prior to the outcome event occurring? Or were they all collected at the same time point? Also, what software are you using?