degrees of freedom for random factors in anovan
2 次查看(过去 30 天)
显示 更早的评论
"anovan" seems to give the same number of degrees of freedom to a factor whether it is declared as random or fixed.
I thought that random factors "cost" less degrees of freedom. At least this is what happens in regression with mixed effects.
Any clues? Is this normal behavior?
0 个评论
采纳的回答
Shashank Prasanna
2014-4-3
Laurie, anovan uses method of moments unlike in mixed-effects models which may use ML or REML. If you are interested in the latter, you may want to use:
here the degrees of freedom will be equal to n – p, where n is the number of observations and p is ONLY the number of fixed effects, not random effects.
更多回答(0 个)
另请参阅
类别
在 Help Center 和 File Exchange 中查找有关 Logical 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!