NA X in ezMixed

[pulpo]

Hi,

Running the example from the help file:

er = ezMixed(
    data = ANT
    , dv = .(error)
    , random = .(subnum)
    , fixed = .(cue,flank,group)
    , family = 'binomial'
)
print(er$summary)

gives me this output:

    NA X - cue 
    NA X - flank 
    NA X - group 
    NA X - cue:flank 
    NA X - cue:group 
    NA X - flank:group 
    NA X - cue:flank:group 
Time taken for ezMixed() to complete: 3 seconds
d> print(er$summary)
           effect errors warnings bits
1             cue   TRUE    FALSE   NA
2           flank   TRUE    FALSE   NA
3           group   TRUE    FALSE   NA
4       cue:flank   TRUE    FALSE   NA
5       cue:group   TRUE    FALSE   NA
6     flank:group   TRUE    FALSE   NA
7 cue:flank:group   TRUE    FALSE   NA

I tried this after finding the same issue with my data. I'm in R 3.6.2 GUI 1.70 MacOS El Capitan build (7735) running ez 4.4-0. 

Can you help, please?

Thanks,

[Mike Lawrence]

Huh, thanks for notifying me. I'll look into it. Probably something changed in lmer or some other package that ez::ezMixed uses, and I didn't notice because I've moved to using Stan for modelling/inference in my daily work. I've actually been considering updating ez to completely remove ezMixed because:

1. I don't think it's very popular (maybe I'm wrong?)

2. The model comparison approach it takes violates Barr et al (2013)'s proscription to "keep it maximal"

3. I no longer espouse it's philosophy of statistics (evidence quantification; I prefer Bayesian parameter estimation these days)

--
You received this message because you are subscribed to the Google Groups "ez4r" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ez4r+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ez4r/a8fec6db-75d5-4335-80b5-e3db1aeb6dab%40googlegroups.com.

[Mike Lawrence]

Ah, seems the problem is that something's changed in how package dependencies are detected/loaded (likely exacerbated by my hack-y use of `base::eval()`). Until ez is updated, a quick fix is to simply run `library(lme4)`. For example, this now works:

    library(ez)

    library(lme4)

    data(ANT)

    er = ezMixed(
        data = ANT

        , dv = .(error)
        , random = .(subnum)
        , fixed = .(cue,flank,group)
        , family = 'binomial'
    )