I haven't personally done any research in this area - but a friend of mine has, and he assures me that Google is still picking off bad backlink sources (like spammy directories) as if they were skeet shooting.
@SevenCubed -- I wouldn't let concerns about scrapers prevent me from undertaking a project that I really wanted to do.
As for directories, several years ago I started a small directory for a fairly-obscure niche that I have a special interest in. It is a stricly "hand-made" direcory in basic html, no ads, nothing for sale. The only way anyone can submit a site is to send me an email. I'm very selective about what I include, and I review each site carefully and write my own descriptions. I enjoy doing it because it gives me a chance to recognize people who have put a lot of time and effort into creating good sites and then reward them by sending them traffic. Google has never de-indexed this directory, and in fact it is at the top of their SERPs for it's main keyword (Of course there isn't much competition since this is an off-the-beaten-path niche).
Occasionally I come across a similar directory that someone else has created for a small niche, and it seems to me that these are the type of directories that Google should give give some credence to in their algorithm since the listed sites have already been screened and evaluated by an independent judge.
Very well said aristotle, the whole post. The emerging niche I am studying is of personal interest too and that's what adds value to it -- first hand knowledge of the topic. That helps me weed out the fakers from the real stuff. If I know I can build a directory that is going to potentially become popular I do not want to give a microphone to websites that are promoting fluff. And like you said, I will analyze them under a microscope, both from a content perspective and also from a behind the scenes perspective.
However the scraper thing still does weigh me down. I know there is a lot I can do to lessen it as well as methods of containing it when it does get hit but I want to continue to explore preventing it for a while first before launching.