Warnings with ezANOVA function.

[MylifeisinHishands Amen]

Hi Mike,

I am trying to the use the function ezANOVA for within-subject design and got the warnings in the picture below. I guess they are warnings that I should ignore but just want to confirm this from you. 

. 

Concerning warning 1, my data frame contains all the levels that I want for the variable "Products" although the data frame is a subset of a larger data frame. Both data frame have the same number of levels only that the subset has less observations but I don't think they are connected in anyway because the name of the new data frame is different from the name of the larger data frame. Am I right on this or this is a warning I should not ignore?

Concerning warning 2: I don't want the subset data frame to be considered in the context of the larger dataset so I guess no need to use within_full.

Best Regards,

Wuraola.

[Mike Lawrence]

Hm, I don't think warning 1 should appear if the situation is as you describe. Can you run:
```r
length(unique( my_big_data$Products ))
length(unique( my_subset_data$Products))
unique( my_subset_data$Products)[ ! (unique(my_subset_data$Products) %in% my_big_data$Products)]

```

And for the 2nd warning, that occurs when it finds multiple observations per Product and combination of conditions supplied to `within`. If the list provided to within is truly the full set of within-subject variables (i.e. you're not doing a follow-up analysis on a subset of the within-subject variables), then yes you can ignore that, but note that the existence of multiple measurements of this kind means that your data are actually of a "hierarchical" type and you are leaving information on the table by using anova versus a proper hierarchical modelling tool.