Mike Lawrence

Check out [loggr](https://github.com/smbache/loggr) for this.

> I'm indeed thinking about changing the current workflow, so that it ... Check out my (very shoddy) work at [here](https://github.com/mike-lawrence/ezStan/blob/master/R/start_stan.R) doing exactly this (but for the niche use case...

Interesting idea. I'm not familiar enough to know whether there is an analytic solution to deriving a CI for the generalized eta square provided by ezANOVA; the post you link...

I wasn't able to reproduce this when it was originally reported. I should have posted a message saying so. A user just emailed me separately with a reproducible example in...

Hm, I can't replicate your colleague's observation that setting the contrasts option makes the SS's of $ANOVA and $aov match: ``` r library(tidyverse) library(ez) #> Registered S3 methods overwritten by...

Though that does remind me that ezANOVA internally sets `options(contrasts=c('contr.sum','contr.poly')` and I should really warn users about that (I'd even forgotten about it and was confused for a bit why...

Hm, in my simulations it's the latter that yields a high rate of false alarms, and only in the case where the covariate is correlated with the independent variable, where...

Please have a look at my simulation [here](https://gist.github.com/mike-lawrence/3122329), which produces [this plot](https://github.com/mike-lawrence/ez/files/2033255/ancova.pdf). Feel free to correct any code or interpretation error I've made.

That discussion highlights that the method I use is only a problem when the covariate is correlated with one or more predictors, in which case yes, you should be doing...

That message simply means that ez detected that you previously had more subjects in the data frame you're giving it. If you intentionally excluded some subjects, you can just ignore...