Vlad
Posts:
18
Registered:
7/23/12


Regression when have an "independent continuous variable" ?
Posted:
Jul 23, 2012 1:37 PM


Is it possible to use ordered logit (PLUM) in SPSS to describe the relation between a continuous IV (independent variable) and an ordinal DV (dependent variable) ?
Some authors suggest converting continuous to ordinal data, because statistical procedures are not commonly used to compare an ordinal dependent variable with a continuous independent variable (Ref.: Studying a Study and Testing a Test How to Read the Medical Evidence, page 346 )
Other studies use ordinal data instead, as here : http://humrep.oxfordjournals.org/content/15/7/1509.full.pdf .
I have an ordinal outcome variable with 15 possible values (GCS score), and a continuous predictor variable. When I run PLUM in SPSS v.20, there is a warning of complete separation of data, but pseudoR squared coefficients (Nagelkerke) are high  0.982.
[[ Warnings There are 248 (87,0%) cells (i.e., dependent variable levels by combinations of predictor variable values) with zero frequencies. The loglikelihood value is practically zero. There may be a complete separation in the data. The maximum likelihood estimates do not exist. The PLUM procedure continues despite the above warning(s). Subsequent results shown are based on the last iteration. Validity of the model fit is uncertain. ]]
I don't know how to correlate this data, and interpret PLUM's output.
Please help!

