Data validation and other tests

[TomHardwicke]

A reviewer suggests we include some advice on data validation. I can add something on this, but my current practice is pretty unsophisticated - I just sprinkle my code with a few stopifnot() statements whenever it occurs to me.

• Does anyone have a more sophisticated workflow? I've tried a few packages like validate, but they can get quite complex quite quickly and are potentially beyond our scope, or better addressed in an appendix.
• Should we treat data validation and more general testing ('defensive coding') as separate topics? We could have an appendix on testing that includes both and refer to that from both the data management and reproducible writing chapters?

[mcfrank]

@TomHardwicke what are your thoughts on this now? Should we add a note to the project management chapter? I don't think we need a full appendix.

[TomHardwicke]

Yes, something on data validation could fit quite well into the project management chapter. I very briefly mention defensive programming in the writing chapter but without any detail on how to do it practically - is it worth adding a few lines on how to do use code tests in the R Markdown appendix?

[mcfrank]

I like the idea of referencing it in the appendix as well. If you didn't mind adding a bit of prose to each that would be great! Thank you.

On Thu, Jul 21, 2022 at 1:42 AM Tom Hardwicke @.***> wrote:

Yes, something on data validation could fit quite well into the project management chapter. I very briefly mention defensive programming in the writing chapter but without any detail on how to do it practically - is it worth adding a few lines on how to do use code tests in the R Markdown appendix?

— Reply to this email directly, view it on GitHub https://github.com/langcog/experimentology/issues/14#issuecomment-1191211863, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAI25FZTKE63W2QSD7QB7CDVVEEQVANCNFSM5DY5474A . You are receiving this because you commented.Message ID: @.***>