Friday, December 4, 2020

A Theory Of Social Media Meltdowns

Hello everyone. I've been taking a long break from blogging during the Rampjaar that has been 2020 for a bunch of different reasons, but seeing as we (probably) saved The Republic by electing Joe Biden and, in God's good time, we will get some sort of vaccine to end COVID, I have decided to "get back on the horse", as they say.

(I should also say I was a bit inspired by a lot of people doing things like launching their own online projects, or leaving standard journalism to brave the wilds of Substack, or pointing out that the old internet of Blog World was in many ways a lot better that the often times insane brave new world of social media.)

So with that here's my first attempt at posting in a while.

 

One thing you hear a lot about if you spend too much time on the internet is various very online and often social media driven meltdowns*, where an organization or institution, usually closely connected to the internet, makes a questionable (or not questionable!) decision, and then all hell breaks loose inside it in a very public way you can watch in real time.

Remember that time the New York Times decided to publish Tom Cotton's terrible Op-Ed? Classic meltdown. Or see this 2019 piece about a number of the social media driven meltdowns in the world of young adult literature. Likewise columnists in top tier media outlets leaving is its own subset of meltdowns, see the saga of Bari Weiss, or the meltdown about Kevin Williamson getting the boot. Or if you'd like a "ripped from the headlines" example, see the recent meltdown at Random House over if they should publish a self-help book from best selling author, and controversial weirdo, Jordan Peterson.  

These events are hardly that important in the grand scheme of things, but for those of us who spend way to much time online (which thanks to the virus is a lot of us these days) they can be very prominent when they break.

When they happen, they tend to create a enormous amount of online content where people fight about them and "what they mean". These takes run the gambit from "this was fair" to "this was unfair" to "this shows a dangerous new ideology" to "some one is getting their just desserts" to all sorts of thing about "wokeness" and "free speech" and God only knows what else.

But I was always struck by there being a weird sort of dynamic in these meltdowns that I couldn't quite put my finger on. Then I remembered a great old post and Twitter thread by Bloomberg's Noah Smith. He compared these sorts of dynamics to the Japanese concept of gekokujo.

Gekokujo is a Japanese concept that roughly means "the lower rules the higher" or "the low overcomes the high." Once upon a time, it would refer to lords of the lower ranks in the nobility overthrowing or controlling more senior lords in pre-modern Japan. But as Noah points out one of the classic examples of this is the so called "February 26 Incident" in 1936, where a influential clique of lower level officers in the Army attempted a coup to seize control of the military and government. They failed, and many of their leader were executed, but ironically, the Japanese military and government largely adopted their policies of sidelining civilian leaders, extreme militarism, and imperial expansion.

It didn't end well, to say the least.

This example may be extreme, but that's why I thought about it in terms of various online meltdowns. A lot of this stuff is a sort of gekokujo attempt by lower level people in very online and social media adjacent organizations to "over come the high" of that organization by leveraging the internet and especially social media.

Note that as Noah points there is nothing inherently wrong with this: "As an example of how gekokujo can do good, it's wonderful that the econ profession is finally trying to rid itself of sexism. That would never have happened without the support of social media. That was a heroic episode of gekokujo!" 

You could also argue that a lot of what happened around the whole #metoo movement in Hollywood was a sort of gekokujo push to get rid of people like Harvey Weinstein and others, and as the hashtag shows social media support was a big part of this.

Noah argues that the big danger of of gekokujo is a sort of backlash it can lead too, but I don't think that's what's going on with these sorts of very online social media meltdowns. Instead I'd point out some other specific problems that keep popping up:

1. Gekokujo campaigns have a tendency to get out of control: The Bari Weiss example is telling. No her career is not some important national issue, and I don't care for her much either. But if you're a progressive media person and start out trying to criticize the writing of one of the few non-progressive writers in your newsroom and end up arguably creating a classic "hostile work environment" for your lesbian coworker, maybe you should step back and ask yourself, "what are we doing here?" Or see the strange case of David Shor, the smart progressive political data guru who who faced a classic use of gekokujo online tactics to try kick him out of the small world of progressive data campaign people for tweeting an academic paper people didn't like. (Or see some of the truly bizarre stuff from The Online Young Adult Fiction Wars.)

2. Gekokujo tends to lead to group think: Many of the very elite media organizations I've outlined above that have engaged in these sorts of meltdowns seem to been stuck dumbfounded by recent political events. The New York Times is instructive: according to in-house conservative writer Ross Douthat when it came time to write their various "the case for..." pieces about who the next Democratic presidential nominee should be, none of the liberal columnists he works with were willing to do it for Joe Biden and so it fell to him. Why were these highly paid writers totally out of touch with what actual Democratic primary voters think? Likewise see the City of New York's swing of 7.6% towards Trump, more than any other state in the union, but I don't a recall a single pre-election story about this possibility. Gekokujo campaigns might cause your to lose focus on things like what voters in your own community actually think. And of course they can, one theme of these kinds of campaigns is that they take a lot of work to get going.

3. Who is the actual little guy? One of the themes in all these meltdowns is that the "lows" in gekokujo campaign aren't really that in terms of society, even if the fights are often always fought in those terms. Lower level staffer in The Atlantic or New York Times are hardly the truly disposed in our society. Is overthrowing the leaders of flagship publications really about creating a more just society? A better political press? Or could it be about something a little more self interested?

The big point here isn't that these individual meltdowns are some grave national crisis. It's that they show a peculiar sort of dynamic in a certain type of organization, some of which are quite influential, that seems increasingly common. And since this dynamic has the potential for real consequences, good and bad, it should be seen as something to actually treat seriously.


*I decided to call these events "meltdowns" in honor of the epic Gawker meltdown that was a trailblazer of this phenomena and also because I wanted to stress the similarities of these types of individual events rather than the unique circumstances that surround each one. If you are outraged and want to organize a social media campaign to cast me into the outer darkness please don't, but feel free to sound off in the comments!

Friday, January 24, 2020

Handicapping The Democratic Field V

With 10 days left before the Iowa Caucuses now seems like a great time to revisit where things in the Democratic field stand.

Following other posts I've written like this I'm going to continue to rate candidates in terms of their likeliness via broadly defined tiers. However I'm going to add something new, taking a (very, very) rough guess at the percentages of a given outcome and see where they line up with 538's cool new prediction model on who will win the nomination. So with an eye towards how idiotic this will probably look in three months, let's jump right in! (Also if you'd like to see how my thinking from about this has changed just click on the 2020 tag at the bottom of this post).

TIER 1:
Joe Biden: Simply put Biden has been and remains the front runner in the 2020 Democratic nomination cycle. Despite negative coverage by much of the press, and the never ending "gaffes" we hear so much about, he remains on top by most measures. In 538's national poll tracker he's about 6.5 points above Bernie, and over 10 points above Elizabeth Warren. Likewise he has a large lead in endorsements compared to his nearest competitor, 226 points to Warren's 81 by 538's methodology.  More over Biden's support among state legislators, another metric of party support, is the best with him having the most endorsements overall, and probably more importantly having a sizable contingent of endorsements from all four early states, unlike his competitors. And finally he has significant support from unions (see this awesome tracker somebody made), with only Bernie Sanders also having formal support among this key aspect of the Democratic coalition. To be sure, this hardly guarantees success, there is no national primary after all, and endorsements might be less valuable now that they appear to have been in past cycles. But never the less Biden is the front runner.
-538's odds of Biden winning more than half of the delegates as of 1/24: 41%
-Longwalk's guesstimate: I'd go a bit higher, as I guess his older, multiracial coalition is more durable than most suspect and there's a reasonable chance he could see a flood of party support after Iowa. Let's say 50%.

TIER 2A:
Elizabeth Warren: Warren seems to have basically been treading water since I last wrote about the race back at the beginning of December. She's in 3rd place in national polling at around 15% while generally tying Pete for fourth in Iowa and New Hampshire. Likewise she's gathered some more endorsements over January and is in second place after Biden by 538's count, but her pace of new support isn't very impressive. She picked up 13 new endorsement points in January compared to Biden's 54, Bloomberg's 22, Pete's 10, Bernie's 7, Bennet's 3, and Steyer's 1. That's not very impressive. Never the less she is well positioned as a big tent unity candidate if Biden and/or Bernie falter in early states.
-538's odds of Warren winning more than half of the delegates as of 1/24: 13%
-Longwalk's guesstimate: That seems low to me, I'll go with 25%.

TIER 2B:
Bernie Sanders: I've been bearish on Bernie's chances for a long time, mainly because he's a factional candidate running in a party that values coalition building and big-tentism. But I can't deny he has real support in the form of close polling in Iowa with Joe Biden and a small recent lead in New Hampshire while being second nationally. Likewise he recently passed Klobuchar and moved into third place in endorsement support. So despite my longstanding doubts, it is quite possible Bernie can succeed. Having said that I still think Warren and Biden are far more likely.
-538's odds of Bernie winning more than half of the delegates as of 1/24: 23%
-Longwalk's guesstimate: Due to the above factors I still think this is high, but having gotten this far and retaining real support means Bernie has a chance. I'll go with 10%, and feel free to mock me when he wins (I can then point out you don't know how probability works).

TIER 3A:
Pete Buttigieg: Since early December Pete has gone down in national polls from about 10% to 7.5% basically tying him with a rising Bloomberg. In addition, his support in Iowa and New Hampshire has declined making him roughly tied for fourth with Warren in both. Likewise, he remains weak when comes to party support, with his tiny roster of state legislators being pretty noteworthy, at least to me. But despite these difficulties it's his total lack of support among non-white voters in general and black voters in particular, that remains his biggest challenge. Could he win Iowa and somehow dramatically upend this race? I guess anything is possible, but at this point in the race he strikes me as a longshot.
-538's odds of Pete winning more than half of the delegates as of 1/24:7%
-Longwalk's guesstimate: Seems about right!

TIER 3B
Amy Klobuchar, Michael Bloomberg: I see both of these candidates as extreme long shots, but for very different reasons. Klobuchar has run a classic "Iowa or bust!" campaign hoping to use a victory there to launch a bandwagon that everyone will climb aboard, similar to what happened with John Kerry in the 2004 cycle, or Jimmy Carter way back in '76. Unfortunately while she has improved her standing in Iowa into the high single digits, she's still way behind everyone else. Moreover she seems to have little to no support in states after that. In that sense I really don't see it happening for her.
-538's odds of Klobuchar winning more than half of the delegates as of 1/24: Essentially Nil (538 gives everyone but the "Big Four" a .6% combined chance of winning).
-Longwalk's guesstimate: Hey, Trump's president right? So let's say 3%.

Michael Bloomberg has been trending up in national polls after spending God knows how much money of TV ads and hiring an army of staffers. Likewise he's been able to parlay his extensive connections with mayors and other Democratic political actors into some real endorsements, earning him 33 points by 538's counting, right behind Pete's 36. At this point he's no longer someone who can be dismissed. But at the same time the fact that he's skipping the early states, something that has never worked for other candidates who've tried it, and has a record that's deeply out of touch with the contemporary Democratic Party makes him a major long shot in my view. But then again look who's president!
-538's odds of Bloomberg winning more than half of the delegates as of 1/24: Essentially Nil.
-Longwalk's guesstimate: Hey, Trump's president right? So let's say 3%.

TIER 3C:
Contested Convention: As far back as I can remember journalists and pundits have fantasized about this outcome and thus every four years people try desperately to try to find ways bring the dream to life, and this cycle is no exception. But a mid-20th Century style convention in Milwaukee where the ultimate nominee is unknown when the convention is gaveled in remains really unlikely. Why? Well Dave Hopkins summed this up nicely last spring when professional take writers first got excited about it. Simply put such an outcome is unlikely because the winnowing based process we've already seen happening will only accelerate once voting starts, the front loaded calendar might end the process soon rather than later, the Democratic Party isn't "highly fractious", and party elites will work incredibly hard to prevent what would undoubtedly be a disaster in the modern era where there simply are no "brokers" to "broker" the convention.
-538's odds of a contested convention as of 1/24: 16%.
-Longwalk's guesstimate: Something really remote, let's say 2%.

TIER 4:
Everybody else. Sorry, you don't have to go home, but you can't stay here.