My profs have advised me to ignore the suggestions for writing sample length, and to just send in the proofs from a paper I published (about 23 single spaced pages or 12 000 words). I guess it's a bit of a double edged sword: on the one hand, it's nice to have it in the journal's type setting, and there may be implicit bias knowing that it is a published paper; on the other, some departments may throw it out right off the bat for not meeting their requirements.
What are your thoughts?