Apologies for the lengthy pause between the last entry in the Boot Camp and this one. It has been an incredibly busy summer, and truth be told, I've been a bit uncertain about what to say about the final component of the dossier: writing samples. How much can be said about them, after all, that most of us don't already know? (Your writing sample should be an awesome piece of philosophical research, didn't you know?;)
Or perhaps I missing something. Perhaps there are tips for writing sample of which I am unaware. If so, I encourage people who know them--placement directors and search-committee members in particular--to chime in! In any case, I will simply offer up a few suggestions and then open things up for discussion.
Let us begin with the issue of topic. It goes without saying, I think, that your writing sample should be "interesting", whatever exactly that means. Since you only have one shot with each committee, all things being equal you should presumably submit what you take to be your most interesting piece of work. It's also a very good idea for your sample to give readers some idea of the long-term potential of your research program--that is, its likelihood of producing multiple publications (since, at most places, a significant publishing record is necessary for tenure). So, if you have a good but very self-contained piece and an equally-good but more ambitious piece with tantalizing implications for future research, go for the latter. But there's an issue here. It's also the case that, all things being equal, you should send you most well-argued paper--and sometimes (as I know all too well) what one takes to be one's most interesting work is not always the same as one's best-argued work. What to do, then? Frankly, I don't think there's any science to it. Hopefully, your best piece of work is also your most interesting work. If not, you have to gamble. However, one "tie-breaker" might be this: the kind of department you're applying to. Research-departments at R1 schools may be happy to receive a brilliantly-argued, incredibly technical piece. But teaching schools? Remember, many of your would-be colleagues at this schools may have little background in your area of expertise, and, since their school's focus is teaching, they are interested in your ability to communicate and come across as interesting and accessible to students and other faculty members. So, I would say, if you have to choose between (A) a perfectly rigorous but technical piece, and (B) a less-rigorous, less-technical but "more interesting" piece, it makes prima facie sense to make the choice on the basis of the kind of school/department you're applying to.
Another complication arises when you're applying for a job in one of your AOS but your best pieces of polished work are in an AOS not listed in the job ad. This happened to be last year. My AOS are ethics and social-political philosophy. Yet, while I had a few published pieces in ethics and some rough (but in my view promising) work in preparation, my best polished works were in political philosophy. Hence, I faced a problem: ethics was a genuine AOS of mine, but I did not think that any of my polished work in ethics was my best work. What should one do in this case? My experience is that even if you think that your best work is in another area (in my case, political philosophy), your writing sample has to be for the AOS advertised. I tried submitting a political writing sample to some ethics jobs, and did not receive a single interview for those ones. Even if it's not your best work, if you apply for a job, your sample needs to be in the advertised AOS, not a "close by" AOS.
Now let us turn to length. Although many job ads have stated length limits on writing samples, some do not. For those that have length-limits, it's critical to respect them. If you need to abridge/cut-down your sample to meet the length limit, do it--noting somewhere on the first page that the piece is abridged (or, at least, this is what I was advised to do). When it comes to jobs with no stated length limits, I've heard that your sample should be 20-30 pages, and by no means significantly longer than that.
Another important issue (or so I have heard) is structure. Because search committees are strapped for time, it is important to submit a piece that gets to the point quickly: a piece that has a snappy introduction, not too much background exposition, etc. For what it is worth, I've found that this is also critical in publishing! You want your writing sample to "start off with a bang", captivating your reader from the very first sentence.
One further issue is whether you should submit published or unpublished work. Although I only have anecdotes to go on here, I have heard from many people that if you have published stuff, you should send published work--that it makes for much better "optics" than unpublished work (making you look more like an accomplished professional and less like a student).
Okay, then, I guess these are my only "tips." Any critiques, additions, or amendations? Fire away!