search

mike-lawrence / ez

Development repository for the `ez` package for R
52 stars 17 forks source link

Way to output the corrected DFd and DFn if GG is in use #42

Open wuschLOR opened 9 years ago

wuschLOR commented 9 years ago

Hi there,

I was just wondering if i have to calculate the adjusted degrees of freedom by hand or is there an easy way to get the output done with an ez option ?

sadly I'm not firm enough with R to insert the formulas to the code by myself.

mike-lawrence commented 9 years ago

Hm, the bits of code that compute the GG correction were taken from elsewhere a long time ago. Taking a look, I can't see where the degrees of freedom could be extracted.

wuschLOR commented 9 years ago

hi, thanks for answering. meanwhile i did some digging and found a semi suitable answer.

Mainly GG and HF need corrected degrees of freedom but how to calculate them is pretty trivial. Its just the normal formula

dfn= (p - 1) dfd =(p - 1) (n-1)

and multiply it with the GG epsilon or HF epsilon

epsilon = # epsilon p = # number of repeated measurements n = # number of participants

f.dfn <- function (epsilon, p, n){ DFn = epsilon * (p - 1) return (DFn) } f.dfd = function (epsilon, p, n){ DFd = epsilon * (p - 1) * (n - 1) return (DFd) }

all with the hint if original epsilon > .75 use HF original epsilon < .75 use GG

Is there a way to output the original epsilon ? Another thing that came to my mind: would it be hard to built in an option where you get the output for the not significant Mauchly's and its corrections just to get more feeling for the data you are working with?

literature used Bortz 2010 S.301 (doi 10.1007/978-3-642-12770-0) Field 2012 Discovering Statistics Using R https://en.wikipedia.org/wiki/Mauchly%27s_sphericity_test#Violations_of_sphericity