Maybe IQ is the wrong word when it comes to subject matter excellence. Maybe it's a combination of knowledge and IQ. But I certainly do not enjoy reading the Toronto Sun.
As for assessing the intelligence of an online community, it shouldn't be too hard. It's been proven that for native English speakers vocabulary directly correlates to IQ, even when controlling for linguistic IQ. So you could put together a managed LDA generative model that trains on various documents either classified as academic research all the way down to 5th grade writing and see what the topic drift is over time. Should correlate relatively well.
Have you found evidence contrary to this? Or are you taking issue with the parent's use of the term "higher IQ" rather than "higher perceived intelligence" or something similar? I'm asking out of genuine interest because the statement seems plausible to me (a layman) based on anecdotal evidence.
What is your evidence for this factual assertion? (I have done a huge amount of research on IQ testing, and I have never seen a trace of a finding like this in any of the research literature.)
But I've grown and the community probably became a bit more mainstream and less intelligent.
How would we assess the "intelligence" of an online community?
Why not rather just help everyone acculturate?