[R] Using MCMC sampling to estimate p values with a mixed model

[ps0u5145]

Hi,

Thank you for the replies.

Yes I would say it does resemble a randomized block design in that each
person completed 4 out of a possible 16 versions/conditions of the vignette.
There was approximately 20 responses in each condition.

The commands I have been using are;

modela <-lmer(DV ~ IV*IV*IV*IV + Co-variate + Co-variate + (1|Subject))
anova(modela).

I have been working out the denominator degrees of freedom myself at the
minute, just for reporting. Do you mean I should use the df and look up in F
critical value table rather than use pMCMC bbolker?

Robert I see what you mean, some of my hypotheses are two tailed, so for
these F < 1 is a signficant finding, just in the other direction. For my one
tailed hypotheses then a value of less than 1 would not be signficant? 

Thanks again it is really helpful as most of the lecturers in my university
are familiar with SPSS and use R very rarely. 

--
View this message in context: http://r.789695.n4.nabble.com/Using-MCMC-sampling-to-estimate-p-values-with-a-mixed-model-tp3606654p3611453.html
Sent from the R help mailing list archive at Nabble.com.