F-distribution vs. T.distribution in Factorial Design
- July 31, 2011 at 8:11 pm #53855
QiuParticipant@Minghua Include @Minghua in your post and this person will
be notified via email.
When looking at the ANOVA (based upon F-distribution) output and Effect Analysis table (based upon T-distribution) in a typical Fatorial Design data output, I have noticed exactly the same P-values between these 2 methods for any given factor.
I understand logically that a big P-value in the effect analysis will imply a big P-value in the ANOVA output since the effect is nealy zero (insignifiant) therefore the population variance for that given factor should be equal to or less than that of the error in the F model.
What is the statistical relationship between these 2 analysis models, or is there is a function for conversion between the T model and F model? They agree with each other perfectly well.
Thanks a lot.
You must be logged in to reply to this topic.