[R] Question about comparison of nested models and F-statistics

2011-12-10 Thread олег ---
Hello, I have a linear model with 10 regressors, and summary() gives me F-statistics. After deleting a term with smallest p-value the new model was obtained with F-statistics higher then it was for full model. Does this mean that parsimonius model is better? The distributions of dependent variable

Re: [R] Question about comparison of nested models and F-statistics

2011-12-10 Thread Bert Gunter
This is not a statistical help list. Post to stats.stackexchange.com or a similar list to have someone tutor you in regression and explain why what you are doing is likely to produce utter nonsense. Or, better yet, get local statistical help in person. You also clearly need to do some reading on