Thank you for this great video. I have a question: When should one use the T stat or r2 among the statistically significant items to determine the next item to add to the model instead of the lowest p among those statistically significant?
Hi Isaac, I am glad you liked the video. For T-stats and P-value, they are like two-side of the same coin. P-value and T-stats just have different threshold, but the variables selected under the P-value should also be selected by using T-stats. We prefer the P-value, because the P-value threshold (0.05, or 0.01)is always the same regardless what kind of test you are performing in any situation (F-test, T-test etc.). Regarding R2, we have different selection method, called stepwise selection, we will use R2 to form a F-test. However, it will be computation intensive, if you only have excel. If you have other software, such as SAS, MATLAB, STATA, they have pre-built functions allowing you to construct F-test to do the model selection. If you only have excel, Forward-selection using P-value is preferred. Please let me know if this is helpful. Good luck!
![Eminem - Somebody Save Me (feat. Jelly Roll) [Official Music Video]](http://i.ytimg.com/vi/Vwa0HenQMi4/mqdefault.jpg)
This helped me immensely. Thanks!
thank you very much......explained systematically........really superb...
You are welcome!
Thanks a lot doctor
I am glad it's helpful! You are welcome!
nicely explained
Thank you for teaching, really helpful
You are very welcome! I am glad it is helpful.
GREAT VIDEO, it helped me a lot for my college homework, thank you so much
You are welcome!
Thank you teacher
You are welcome
Thank You Doctor
You are welcome!
Thank you for this great video. I have a question: When should one use the T stat or r2 among the statistically significant items to determine the next item to add to the model instead of the lowest p among those statistically significant?
Hi Isaac, I am glad you liked the video. For T-stats and P-value, they are like two-side of the same coin. P-value and T-stats just have different threshold, but the variables selected under the P-value should also be selected by using T-stats. We prefer the P-value, because the P-value threshold (0.05, or 0.01)is always the same regardless what kind of test you are performing in any situation (F-test, T-test etc.).
Regarding R2, we have different selection method, called stepwise selection, we will use R2 to form a F-test. However, it will be computation intensive, if you only have excel. If you have other software, such as SAS, MATLAB, STATA, they have pre-built functions allowing you to construct F-test to do the model selection. If you only have excel, Forward-selection using P-value is preferred.
Please let me know if this is helpful. Good luck!
@DrJiang Jingze your response was very clear and helpful!! Thank you for your help, doctor. Your videos are very educational and a blessing.