Did something go wrong with exclusions? about mb2-analysis HOT 10 CLOSED

no, this is deliberate because of issue #46 . This issue has been noted in past MB studies and we need to ensure that all subid is unique for further analyses

from mb2-analysis.

ok.

from mb2-analysis.

It seems that the subid modification comes from this line:
mutate(subid = paste0(subid, sep = " _ ", lab, trial_num))

I think we should at remove the trial_num (did you mean to put trial_id?) from the subid column, because otherwise group_by(subid) starts treating each trial, rather than each subject, differently. We could add experiment_num to subid though.

from mb2-analysis.

Got it, thanks for your explanation, may be we shouldn't make it too unique to avoid problems like that.....I honestly think that subid + lab is already unique enough to do the trick...perhaps just combining these two names. thoughts? @GalRaz @mahowak

from mb2-analysis.

subid needs to uniquely identify each subject. 

subid + lab should usually do it, but because labs assign their own subid, they might call a particular subject mb2-1 in both Pilot 1a and Pilot 1b. 

What @GalRaz said sounds good: subid + lab + experiment_num should be unique

I am also a little thrown by having the sep arg right in the middle of the paste like that, how about:

mutate(subid = paste(subid, lab, experiment_num, sep="_"))

from mb2-analysis.

subid+lab+exp_num sounds good.

I can remove the sep arg, cause I don't think putting it at the end helps anything. I like putting it in the middle for me to see the original subid, but I am okay if you want to remove the sep

Do you want me to change the codes, or @GalRaz?

from mb2-analysis.

from mb2-analysis.

sounds good to change it to paste()

from mb2-analysis.

just put in a pull request; closing

from mb2-analysis.

Just to close the loop here, paste0 doesn't actually take a sep arg. It just concats all the args with no sep. But it lets you pass garbage names as arg names.

> paste0("this", sillyarg="is", fakearg="a", dsfj9iw="test")
[1] "thisisatest"

from mb2-analysis.