To me the key highlight of the article is the finding that editors generally start fairly radical and neutralize over time. Only really passionate people are willing to put the effort into Wikipedia articles which correlates well with radical opinions. But over time working as Wikipedia editors tends to de-radicalize people's work.

Contrast that with the rest of the internet, which mostly rewards radicalization and nudges people towards it.

That's some of it, but certainly Wikipedia's editorial discussions differ from most forums in that its objective remains neutral, with worldwide access.

If the number of editors were limited, it could easily develop bias (see your own Facebook page for examples).

If the subject matters were limited, it could develop bias (WikiSolarEnergy wouldn't tend to attract anti-solar-energy types).