I have a population of 135 athletes, im seeing the frequency of a disease and its association with return to play (number of days taken for them to get back to their game)
Frequency is a yes/no question - nominal
I have the return to play data in scale and also in ranges in an ordinal format. I dont know which is the correct method, keeping them as numerical in days or making them categorical.
I was advised for a chi square test but my table is 2x5 and many less than 5 counts are seen when p value comes out as 0.000
what could be the correct test for this?
what type of variable is the return to play (categorical or numerical?)
Hello there
I am not an expert in this field of statistics, I remember using the attached program written by the experts in point processes...your interval data falls into this category...
>>>
Reference
Cox, D.R. & Lewis, P.A.W. (1966). The Statistical Analysis of Series of Events. Methuen, London.
Hi Wajeeha,
if return to play is assessed as the number of days till the participant returned, I would assumed it to be at least interval scaled, because the difference between 2 and 3 days is equal to the difference between 14 and 15 day (just an example, but this is what we call equidistance). The only problem might be the distribution of this variable (is it appr. normally-distributed?).
However, I would recommend to use either point-biserial correlation analysis (if you are interested in the association), a linear regression model with return to play (criterion) regressed on frequency (dichotomous predictor) or a generalized linear model/logistic regression model with frequency (dichotomous criterion) regressed on return to play (predictor; both last models depending on whether you want to predict one variable based on the other).
Maybe you could give some more information on your research question and the meaning of the frequency variable.
Best,
Brian
Brian Schwartz
Frequency variable is: experience of stinger (disease), kept as nominal: yes/no
so im looking at stinger's frequency and its association with return to play (days taken)
now, im not sure if my return to play data is equally distributed, im guessing its not (not good with stats, im a DPT)
how do you suggest i add the return to play variable in spss, do i keep it in scale and add days as numbers, or do i assign ranges to them in ordinal?
and then, do you still suggest the above tests?
So the question is whether participants with the experience of stinger need more (or less) days to return to play?
Well, to be honest, with an N of 135 athletes, I would not worry so much about skewed distributions. But maybe others who are more into the mathematics of statistics wouldn't agree with me.
I'd sugget to keep the return to play variable in scale and add days as numbers. In general, I am not convinced of categorizing continuous variables. And yes, I would still recommend the above mentioned tests.
Brian Schwartz correct, thats the question.
Alright, could you guide me about logistic regression on spss? (binary logistic?) and what would be the dependant and covariate in my case?
thank you SO much for your help.
- Badges
- Science topic
- Similar topics
- Correction