As we’ve discussed many times, I prefer blogs to twitter because in a blog you can have a focused conversation where you explain your ideas in detail, whereas twitter seems like more of a place for position-taking.
An example came up recently that demonstrates this point. Jennifer sent me a blurb for her causal inference conference and I blogged it. This was an announcement and not much more; it could’ve been on twitter without any real loss of information. A commenter then shot back:
Do you see how your policies might possibly negatively impact an outlier such as myself, when you arbitrarily reward contestants for uncovering effects you baked in? How do you know winners just haven’t figured out how you think about manipulating data to find effects? How far removed from my personal, actual, non-ergodic life are your statistical stories, and what policies that impede me unintentionally are you contributing to?
OK, this has more words than your typically twitter post, but if I saw it on twitter I’d be cool with it: it’s an expression of strong disagreement.
It’s the next step where things get interesting. When I saw the above comment, my quick reaction is, “What a crank!” And of course I have no duty to respond at all; responding to blog comments is something I can do for fun when I have the time for it: it can be helpful to explore the limits of what can be communicated. In a twitter setting I think the appropriate response would be some snappy response.
But this is a blog, not twitter, so I replied as follows:
That’s a funny way of putting things! I’d say that if you don’t buy the premise of this competition, then you don’t have to play. Kinda like if you aren’t interested in winter sports you don’t need to watch the olympics right now. I guess you might reply that our tax money is (indirectly) funding this competition, but then again our tax money funds the olympics too.
Getting to the topic at hand: No, I don’t know that the research resulting from this sort of competition will ultimately improve education policy. Or, even if does, it presumably won’t improve everyone’s education, and it could be that students who are similar to you in some ways will be among those who end up with worse outcomes. All I can say is that this sort of question—variation in treatment effects, looking at effects on individuals, not just on averages—is a central topic of modern causal inference and has been so for awhile. So, to the extent that you’re interested in evaluating policies in this way, I think this sort of competition is going in the right direction.
Regarding specifics: I think that after the competition is over, the team that constructed it will publicly release the details of what they did. So at that point in the not-so-distant future, you can take a look, and, if you see problems with it, you can publish your criticisms. That could be useful.
I’m not saying this response of mine was perfect. I’m just saying that the blog format was well suited to a thoughtful response, a deepening of the intellectual exchange and a rhetorical de-escalation, which is kind of the opposite of position-taking on twitter.
P.S. Also relevant is this post by Rob Hyndman, A brief history of time series forecasting competitions. I don’t know anything about the history of causal inference competitions, or the extent to which these were inspired by forecasting competitions. The same general question arise, of what’s being averaged over.
Then again, a blog can also end up with a comments section like that of Marginal Revolution’s in which position-taking and word diarrhea are combined into an extra-foul little soup, and the blog’s authors don’t even bother replying to anyone anymore because it’s such a wasteland of thought.
Will:
One thing I haven’t figured out about Marginal Revolution in that, in his posts, Cowen seems often to be talking to a hypothetical audience of political liberals or people on the center-left, but his commenters seem to lean hard-right. These commenters don’t seem to coincide with the audience of his posts. I guess that makes sense—commenters are a small subset of readers so no need to suppose they’d be representative of an audience of interest—but still it’s striking to me.
As a blogger, I write to some imagined audience—I have no idea who’s reading these posts—and I don’t usually feel that I’m writing to the commenters, but I do learn from comments and I often engage in the comments section (as here), I think as a way to avoid real work!
Robin Hanson found in a Twitter survey that the people most likely to reply were those with the minority response to the survey:
https://twitter.com/robinhanson/status/1574033112748945409
Wonks:
I followed the link to the Hanson post and I don’t understand it at all. He writes, “comments don’t reveal a silent majority”—but nobody would ever make such a claim. By definition, commenters are not silent. They’re the loud people! Which is fine—I really appreciate commenters. It’s just . . . of course they don’t represent a silent majority. They literally are a loud minority.
Andrew: One might assume the commenters would represent the views of the majority: that only a small percentage of readers are motivated to comment, but that those that don’t comment might share the same views as those that do. That is, although the commenters are not themselves silent, they might represent the views of the majority. I think that’s the assumption that Hanson thinks needs to be debunked.
Phil:
Perhaps it’s just that as a blogger I’ve already thought a lot about these issues . . . In any case, I don’t see why anyone would take as a default assumption that a loud minority would represent a silent majority. The whole point of the “silent majority” idea is that they have opinions which are not well represented by the loud people. That’s the reason people talk about a “silent majority” in the first place. So, once you bring up the “silent majority” concept, I don’t think it makes much sense to expect they’d be revealed by a loud minority.
I remain convinced that Cowen is (most of the time) intentionally trolling his readership.
Andrew, you say “nobody would ever make such a claim” [that commenters represent non-commenters]. You may be right but even if you are, that’s not necessarily relevant to understanding Hanson’s post. Perhaps Hanson thinks some people would make such a claim.
You know and I know and Hanson knows that comments don’t reveal the beliefs of the majority, and maybe we even think most people know that, but perhaps Hanson thinks that not _everybody_ knows that. If that’s the case then the comment makes sense, yes? I agree with you that if that’s the intent then it’s weird to refer to a ‘silent’ majority, but not everybody is as careful with language as we would like.
Exactly… I’ve noticed this too… same thing at Scott Alexander’s blog, though not as extreme…
MR’s comment section is one of the ones which prove the rule, and that’s why it so often comes up as an example of a bad comment section. Something weird happened something like a decade ago, and now it’s in a strange equilibrium. The salient aspect is not that it’s right-wing but that it’s filled with weirdos and rejects from the Island of Misfit Toys, like a bunch of recycled Usenet posters. Typically, any comment section which goes weird like that simply gets disabled or moderated by the author (especially on a blog with multiple contributors) – assuming it lasted long enough for anyone to notice instead of simply going defunct as almost all blogs do. (MR started in August 2003, which means it’s not much younger than the word ‘blog’ itself, and exceeds the mean lifespan by like 20x.) I wouldn’t generalize from it too much.
Cowen does get good comments, it’s just his real commentariat is elsewhere: Twitter, private emails, in-person conversations, miscellaneous blogs etc.
Cowen will often link to comments on his posts. So he does apparently read them.
I wouldn’t say ‘often’; 1 in 20 or 30 posts maybe? It feels pretty rare. He will post excerpts from emails more often than that, so if you take post frequency as your metric… For relatively rare stuff like comment links, there are other explanations besides ‘Cowen avidly reads every comment’ – how much of that is the staff (apparently Cowen/Tabarrok don’t do the tech stuff themselves) selecting ones, or the other commentariat linking/repeating stuff, or them only reading comments on a few key posts?
I wonder how many people don’t notice that if they leave their comments section open and don’t moderate it that will hurt their reputation. Even if you distinguish between the two, you will associate the writer or site owner with the awful things in the comments.
Marginal Revolution is like Twitter that way
An argument for Twitter > Blogs (at least sometimes):
Today via a Twitterbot @StatRetro, I read an old blog post on this website: “Cross-validation and selection of priors” from 2006, and through which I further clicked two posts “Posterior distributions vs bootstrap distributions” and “Cross-validation vs data splitting”, both were from Bush time but surprisingly related to what I am currently thinking about. Paradoxically, Twitter is better than blog in refreshing memories from pre-twitter-era materials.
Hey, I’m glad that somebody’s following StatRetro!
Happy customer testimonials: “@StatRetro is great, we should all follow it.”
Nowadays, Andrew, it’s all about the cross-posting. Perhaps a StatTikTok next?