Here's The 20% Statistician, Daniel Lakens, on why blogs have better science than Science.
Lakens observes that blogs (usually) have open data, sources, and materials; open peer review; no eminence filter; easy error correction; and open access.
I would add that blogs are designed to fit human cognitive capacities. To reach a broad audience, they are written to be broadly comprehensible -- and as it turns out, that's a good thing for science (and philosophy), since it reduces the tendency to hide behind jargon, technical obscurities, and dubious shared subdisciplinary assumptions. The length of a typical substantive blog post (500-1500 words) is also, I think, a good size for human cognition: long enough to have some meat and detail, but short enough that the reader can keep the entire argument in view. These features make blog posts much easier to critique, enabling better evaluation by specialists and non-specialists alike.
Someone will soon point out, for public benefit, the one-sidedness of Lakens' and my arguments here.
[HT Wesley Buckwalter]