search

runehaubo / lmerTestR

Repository for the R-package lmerTest
50 stars 9 forks source link

Satterthwaite method fails for binomial model #29

Closed Benambridge closed 5 years ago

Benambridge commented 5 years ago

Hi

I'm trying to fit this model using lmerTest, but it fails (i.e., summary just outputs the normal lme4 p values from the z distribution, presumably meaning it has failed to calculated the Satterthwaite df):

Full1=glmer(correct ~ MLU + BPVS_raw + Chi_Dir_Loggroupcondition + (1+Chi_Dir_Log|Participant) + (1|item), data=Data, family=binomial, glmerControl(optimizer = "bobyqa"))

Is there a problem using the Satterthwaite method for binomial models, or is it something else?

Thanks Ben

runehaubo commented 5 years ago

Satterthwaite aids the F-test in the multivariate normal distribution that arises in the linear mixed models. Since similar F-tests are not defined for non-normal outcomes the method can only be applied for LMMs.

Cheers Rune

ons. 23. okt. 2019 kl. 10.04 skrev Benambridge notifications@github.com:

Hi

I'm trying to fit this model using lmerTest, but it fails (i.e., summary just outputs the normal lme4 p values from the z distribution, presumably meaning it has failed to calculated the Satterthwaite df):

Full1=glmer(correct ~ MLU + BPVS_raw + Chi_Dir_Loggroupcondition + (1+Chi_Dir_Log|Participant) + (1|item), data=Data, family=binomial, glmerControl(optimizer = "bobyqa"))

Is there a problem using the Satterthwaite method for binomial models, or is it something else?

Thanks Ben

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/runehaubo/lmerTestR/issues/29?email_source=notifications&email_token=ABHZZLIJ2WRDO3TO2QFQDYLQQAARHA5CNFSM4JD5O4CKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HTXQX5A, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABHZZLJFTUR6LRYFNDNAQL3QQAARHANCNFSM4JD5O4CA .