Hi, I am trying to analyze the factors that can affect vitamin b6 levels and I performed myltiregression linear analysis with STEPWISE variable selection by SPSS.
I got 2 factors that was independently related to vitaminB6 levels but other 5 was excluded.
The question is how can I get 95% CI values for the excluded factors? I only obtained those for positive factors.
Best regards
Kensuke
Why you want to CI for excluded variables.? You can use them separately or IVs or proxies of them
Hi,
To get the confidence intervals for all regression parameters, you just need to change the method of entering variables into model from "stepwise" to "enter".
After that, you can choose CI's from the "statistics" button.
Regards
Why are you using stepwise regression? Not sure what you mean by an uncertainty value for something not in a model, but in general the uncertainty estimates (CIs, p-values, etc) from stepwise approaches need to take into account the entire search process and this is difficult (for alt, see S6.3 of https://hastie.su.domains/StatLearnSparsity_files/SLS.pdf).
- Badges
- Science method