Turning 1 project into 2 publications

[seym8842]

Anyone have an opinion about using data from 1 project (in this case, human testing in a psychologically-based experiment) for 2 papers? The project used multiple outcome measures with a behavioral manipulation pre and post, but there are so many data points in the pre testing alone that I'm considering making 2 papers out of it. There are 2 very different and interesting stories.

Obviously, there would need to be an explicit indication in the paper that "Some of these data were previously reported in Author et al."

[Puget]

You don't want to mince things into least publishable units, but it's perfectly normal and appropriate to get more than one paper out of a study (often many papers from large studies with many measures) so long as (a) they are each testing distinct hypotheses, and (b) you are clear on what has already been published from the study (you seem to already know this).

It took me all of grad school and most of a postdoc to finally get it through my head that each paper needed to tell just one story-- try to pack in too much and the main message gets lost and no one can figure out let alone remember what that paper was about. I constantly have to dial back my own grad students now and tell them that their cool idea for some additional analyses is indeed a cool idea, but for the next paper, not this one, so they should finish writing this one then preregister* the next.

*If you aren't already doing so, I'd urge you to start preregistering papers before looking at the data. It may be too late as it sounds like you have already analyzed, but something to consider for next time. There is a thread about that and other good data practices here:
http://thefora.org/index.php?topic=277.msg4250#msg4250

[polly_mer]

A series of related, good papers is a wonderful thing.  Puget is correct that you don't want to slice into LPUs, but different chapters for a steadily unfolding saga is a good way to go.

[Hibush]

I will chime in with agreement that it is the story that matters. If you have two distinct stories, they will be better appreciated as two papers. If you were able to test two distinct sets of hypotheses, then you should be good to go.

OTOH, if you have four or five partial and highly interrelated stories, then you are in a pickle. (Preregistraion is good for preventing this particular pickle).

[saramago]

Absolutely. Do go ahead with your two papers. Just be transparent about any overlap.