My goal is to fit a cox regression model in SPSS. As I am still new to regression methods, I would appreciate a little of your help.
First, I checked all variables in a univariate regression and select those, which have a p-value less than 0.05. The next step is that these significant variables (p-value < 0.05) are considered in a model with backward conditional.
My question now is, whether I can find out, why one variable is removed in the backward selection or which variable rather explains the influence? I am surprised, that this variable drops out, although it is very significant in the univariate regression.
Jignasa this is about confounding. That dropout variable's influence is controlled when you performed a multivariable regression. So it became insignificant and removed by the procedure.
Stepwise methods have problems. So be careful when using them. lf a variable should conceptually be included in your model then put it into your model regardless of its removal by the procedure.
When selecting candidate covariates it is common to raise the p level to 0.1 or even more.
BTW when performing backward stepwise selection there is no point of preselection of covariates by univariate analyses.
Dear Jignasa,
Check this link out:
http://www.medicine.ox.ac.uk/bandolier/painres/download/whatis/cox_model.pdf
Jignasa this is about confounding. That dropout variable's influence is controlled when you performed a multivariable regression. So it became insignificant and removed by the procedure.
Stepwise methods have problems. So be careful when using them. lf a variable should conceptually be included in your model then put it into your model regardless of its removal by the procedure.
When selecting candidate covariates it is common to raise the p level to 0.1 or even more.
BTW when performing backward stepwise selection there is no point of preselection of covariates by univariate analyses.
I suggest that you look at the Cox regression material from the website below. stepwise has problems that lasso doesn't. I would do the variable selection this way. best wishes
http://www4.stat.ncsu.edu/~boos/var.select/lasso.adaptive.html
Jignasa if I am correct you want to ask why should a variable dropout in multivariate if it is significant in univariate analysis.
When you do multivariate analysis it gives you a values adjusted for other factors (i,e adjusting for confounding), Hence when you use step down method you loose out on some factors which fail to achieve predecided statistical significance.
Hence if you are sure that some factor has valuable influence on the outcome and want to adjust for the same try using enter method in cox regression.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS