Effect Coding: p-value of reference category

#1
Hi,

I fit a regression model with effect coded variables. I'm well aware of how to compute the value for the reference category but I have absolutely no idea how I can possibly compute the p-value or standard deviation for it.

I really hope anyone can help me in this matter....

Thank you very much,
Chris
 

noetsi

Fortran must die
#2
Why don't you simply use the values generated by the software (is this a class problem where you have to show some type of work)? In SAS p values and (I think) standard errors are generated automatically for reference coded variables.
 
#3
I'm using R as a software. I generate the value for the reference category by taking the negative sum of the other categories. The only solution I have so far is just doing the whole model again with a different reference category and take the p-value from that output since nothing changes in the effects.
Still I was wondering if this is really the right way to do it or if I can also conclude the p-value for the reference category from the other effects in the output...