Yes, but many of these are testable. Thus for example, Oscar’s hypothesis that “Things are only tropes if they happen more often in fiction that in reality, so to detect them you need an accurate map” is testable. You could take a random sample of people who edit TVtropes and test their map accuracy in completely separate areas (say things that can be often estimated with a Fermi calculation) and compare that to a general sample of people. Oscar’s hypothesis suggests that the Tropers will do better.
RobinZ’s point is difficult to test, but presumably if one examined in detail what pages have historically stuck around and which have been merged or deleted, one could get data that would test it.
I would also consider my thesis undermined were it demonstrated that the rate of rationally-insightful contributions to TV Tropes was significantly higher than for other notable Wikis (e.g. Wikiversity).
How would you measure the rate of rationally-insightful contributions? I’m also not sure which wikis would be useful to test this on. Some wikis (such as say the various Wikipedias) have prohibitions on original research. Other wikis have narrow goals that will mimimze the number of rational insights. Thus, I’d expect a very low insight rate on say Wikispecies since that is devoted to cataloging existing biological knowledge.
Good points. What I was attempting to measure was the relative measure* of rationalists on TV Tropes versus other nerd communities. The part of my thesis being tested is that no notable difference need be hypothesized to explain EY’s perception of unusual rationality in the wiki.
(Was I mistaken to believe that EY thought TV Tropes was unusually rational compared to other nerdy Internet communities, as opposed to compared to other Internet communities, full stop? I agree that TV Tropes is nerdier than most of the Internet.)
* i.e. fraction of population weighted by intensity of participation.
This was a nice exercise in generating a host of just-so stories.
Yes, but many of these are testable. Thus for example, Oscar’s hypothesis that “Things are only tropes if they happen more often in fiction that in reality, so to detect them you need an accurate map” is testable. You could take a random sample of people who edit TVtropes and test their map accuracy in completely separate areas (say things that can be often estimated with a Fermi calculation) and compare that to a general sample of people. Oscar’s hypothesis suggests that the Tropers will do better.
RobinZ’s point is difficult to test, but presumably if one examined in detail what pages have historically stuck around and which have been merged or deleted, one could get data that would test it.
I would also consider my thesis undermined were it demonstrated that the rate of rationally-insightful contributions to TV Tropes was significantly higher than for other notable Wikis (e.g. Wikiversity).
How would you measure the rate of rationally-insightful contributions? I’m also not sure which wikis would be useful to test this on. Some wikis (such as say the various Wikipedias) have prohibitions on original research. Other wikis have narrow goals that will mimimze the number of rational insights. Thus, I’d expect a very low insight rate on say Wikispecies since that is devoted to cataloging existing biological knowledge.
Good points. What I was attempting to measure was the relative measure* of rationalists on TV Tropes versus other nerd communities. The part of my thesis being tested is that no notable difference need be hypothesized to explain EY’s perception of unusual rationality in the wiki.
(Was I mistaken to believe that EY thought TV Tropes was unusually rational compared to other nerdy Internet communities, as opposed to compared to other Internet communities, full stop? I agree that TV Tropes is nerdier than most of the Internet.)
* i.e. fraction of population weighted by intensity of participation.
I noticed that—I believe it is a classic case of (warning: TV Tropes) the Rhetorical Question Blunder.
(In my defense, I tried to make mine testable.)