Journal articles (peer reviewed ones) are 'rated' more highly in terms of things like REF and promotion criteria. But there can be benefits to book chapters, which you point out. Ideal would be to do the chapter now and then a different article towards end of PhD (as having a journal article will also help your job prospects). Of course you also need to be careful not too publish too much from your PhD if you're hoping to turn it into a book...

Also, check the regs of your institution, as some of them won't allow you to include something substantially similar to something you've published within your thesis. (Where I work will; where I studied won't)

Hmmm, in your field, I'd say that the greater REF value is a peer-reviewed journal. It's not always the case - in my field essays in books can be very prestigious and REF-safe, but if you're only 2nd year PhD, better be cautious ...

And more people will see/download/read an essay in a journal, usually.

In my experience, book chapters can take ages to come out (or at least the lag is highly variable), so if you're banking on something being out by the time you're applying for jobs, an article might be the way to go.

Yes books of collected essays can be held up for years by a single rogue contributor (I have had this happen to both as a contributor and as an editor); whereas with a journal if you do some homework to check that it has a decent turnaround time you should have a safer bet (do check first though - at least in my field some journals take forever). If you think you have a decent chance of being accepted by a good journal in your field, I'd go that route for a first publication.

I put this question to my soc colleagues earlier in the week (was interested, as it's not my discipline) - two readers, both said journal above books, because of a) the speed and b) the likelihood for better peer review and thus greater credibility on future job apps.

My discipline mainly privileges monographs, and we haven't had a shortlist of ECR applicants for a couple of years now where every single one has not had at least one book eligible for the REF (the last one or the next one, depending on when they were hired!): in sociology, though, my colleagues were saying that for an ECR, one journal article published would put someone in a really strong position for being hired (because, as we all know, despite the JD, no institution actually gives a monkey's whether you can teach: it's all about the REF) but they regularly have entire short lists of ECRs with no publications as yet. Two departments in the same Faculty, at a rising post-92 - but shows how subject specific things are!