Dear everyone! I recently submit a paper.My aim is to identify independently risk factors. I include all parameters which are not significant on univariable logistic analysis into multivariable logistic analysis because i think these parameters may be related to each other. The reveiwers say that only parameters found to be significant on univariable analysis should be go into multivariable analysis . Is it rational to include all parameters which are not significant on univariable analysis into multivariable analysis ? is there any reference or paper about this topic ? or is there any example paper? Thanks a lot !
HI, You might look at what I did for a similar problem. Some of the references are pretty good. Best wishes, David
you are required to include only significant risk factors in multivariate analysis. if you feel these can be related check the multi colinearity or control their effect.
Dear professor Booth and professor Jyoti,,thank you for your kind reply !
- Badges
- Science topic
- Similar topics
- References