Should there always be two lines in the results file when you pass both “first” and “last”? If so, should there be a value somewhere in the line reporting “first” / “last” / “first-last” (in case only one selection happens) / “NA” (in case no selection happens)?
That would be my preference by default, since it’s not very convenient to delineate things in R by which line they appear on (and doubly so if the number of lines isn’t consistent). I don’t know how your code works but it seems like “first”/”last” etc. would be a good candidate to go in the area that just says “choice” now, since the particular choice on each selection is one of the choices that they made, meaning that “choice” is now ambiguous. I’m not sure how that would work with “all”, though, since you wouldn’t want to put in 14 lines for every scale if they selected on one of them 14 times…but in that case it might be enough to differentiate them by number. Since we end up with our data in wide format, the consistently two-line version is good for me, but I don’t know what other people use.
At first I thought I can’t really use the newScale strategy you suggested, since we want them to be able to replay the scene as many times as they want. But since we’re not collecting all of the selections, just two of them, would it be fine to just create a secondScale every time they press replay? Or would that cause errors with the name overwriting?
(Another side note: I don’t know if this is a one-or-two-off thing or a more generalized thing, but https://www.pcibex.net/wiki/selector-test-index/ this page’s text also doesn’t appear in “try it!”, even when I add print() to it.)