An SEO Skeptic gold star to Aaron Wall for his post today regarding PageRank scultping, Expert SEO Testing: Usually Worthless. Since the recent "announcements" that Google is now basically disregarding efforts at controlling internal linking using the nofollow attribute – "PR sculpting" – there's been a whole lot of hand-wringing in the SEO community about Google getting their messaging straight. These usually public protestations don't actually address the point Wall makes so well in his post: was there genuine value in using nofollow for PR sculpting anyway?
I can't do justice to the many good points Wall brings up (you should read the post in its entirety), but the takeaway is that the "testing" that "proved" PR sculpting was the end-all and be-all of cutting edge SEO tactics was, in itself, highly questionable (to say the least).
In many ways, dedicated SEO work is going to bring its own rewards, regardless of whether or not the focus is on one particular tactical element (here the manipulation of links using nofollow). In the absence of a truly controlled experiment (very difficult in SEO for a multitude of factors) there's every possibility that manipulation of an <a> attribute also involved manipulation of link architecture more broadly, skewing such tests left, right and center.
As Wall suggests in his post, the best approach to employing SEO tactics is a holistic one, if only because "it gets hard to isolate testing variables as sites get more established…." Experimentation is an absolute necessity if one hopes to have a sense of whether or not a particular search engine optimization tactic has validity, but you can only go so far in the real world of SEO – which must by necessity try to optimize a website on several fronts simultaneously – by trying to validate a single SEO tactic in isolation.