I have indeed met Flemming and have put forward his approach as a good example of an accountable, scientifically based methodology, and discussed his input/output analysis method (and yes, I go a bit oopsy-la thinking about it myself). Indeed, his results, in differing from anything else publicly available, were an object lesson in how influence and popularity differ.

The NxE ‘study’, however, is not a study at all. It really does seem to be totally based on *subjective* preference, not *objective*, umm, anything.

As I say, I do accept criticisms levelled at studies that just use public metrics such as Technorati. Having watched my blog grow in *popularity* but not, I think, *influence*, I understand more and more the difference.

At the simplest level, one has to define “influence” loosely as “capable of causing a measurable reaction in the wider environment” (v. clumsy, I know – but better, I think than “I like this”).

If we agree on that point, there are various mathematical tools that can be used to assess influence — from simple regression analysis (simple, that is, to brief in to a maths whizz) to network analysis, and input-output analysis.

I tend to fall over a bit when it comes to thinking about matrix maths. But I’m pleased to think that I know a few people who don’t.

Thanks for linking to my blog. You’re right, they probably selected the bloggers because they like them, and there’s another thing: even if half of them mention their story on their blogs, it is a big linkbait opportunity for NxE.