Tribe watch: Jonathan Chait sees an absolute difference!

WEDNESDAY, JUNE 22, 2016

Tribal delusions and war: We've often said that our own liberal tribe has started to ape the very bad practices of players like Rush and Sean.

Jonathan Chait disagrees. In a new post, he reviews a discussion by Jonathan Rauch of our failing political culture.

To Chait, conservatives are Goofus and we liberals are still Gallant. For starters, consider this passage:

CHAIT (6/22/16): The more closely we look at the composition of the two parties, the more obvious it is that only one of them truly exhibits the tendencies [Rauch] describes. Over the last decade, writers like me, Jacob Hacker and Paul Pierson, and Thomas Mann and Norman Ornstein have written about the growing asymmetry between the two parties. The GOP, but not the Democratic Party, is fully identified with an ideological movement. The almost-all-white Republican Party is far more ethnically monolithic than the polyglot Democratic Party, and more ideologically monolithic, too—more than two-thirds of Republicans identify themselves as conservative, while fewer than half of Democrats call themselves “liberal.” (Self-identified moderates and conservatives comprise a majority of the party’s supporters, albeit a shrinking one.) Democratic voters rely on news sources that, whatever their unconscious bias, strive to follow principles of objectivity and nonpartisanship. Republican voters mostly trust Fox News and other party organs that merely amplify the party’s message.

That's straight outta Goofus and Gallant! Goofus mostly trusts Fox News and other party organs. By way of contrast, Gallant relies on news sources which sometimes display unconscious bias despite their good intentions.

Liberals, please! Our direct counterpart to Fox News is MSNBC. It's silly to think that MSNBC displays only an "unconscious bias," nothing more, despite its good intentions.

Already, Chait is displaying a severe myopia. In the passage we highlight below, the pundit goes over the edge.

Except as an example of Hard Dogmatic Tribal Vision, the highlighted claims makes zero sense. That is especially true when you click Chait's link:

CHAIT (continuing directly): The political scientists Matt Grossmann and Dave Hopkins have found that Democrats tend to conceive of their policies in concrete terms, while Republicans present theirs in ideologically abstract terms. The pragmatic deal-making Rauch venerates is simply far more compatible with the style of the modern Democrats than the Republicans...

A series of polls have all found that Democratic-leaning voters want their leaders to compromise, while Republican-leaning voters do not. Many Democrats feel frustrated with the system, but they want to make it work. Republicans do not feel this way at all.

According to Chait, "Democratic-leaning voters want their leaders to compromise, while Republican-leaning voters do not." In case that flattering tribal contrast wasn't stark enough, he proceeds to make his claim absolute:

"Republicans do not feel this way at all."

Republicans don't feel this way at all? To anyone but a hard-core believer, the claim sounds absurd on its face. It turns into a species of lunacy when you click Chait's link to review his supporting data.

Holy tribal belief! Chait links to this survey by Pew from October 2015. The relevant passage says this:

PEW (10/12/15): Partisan compromise is viewed less positively by Republican primary voters than by Democratic ones. Just 41% of possible GOP primary voters say they would be more likely to vote for a candidate who will compromise with Democrats.

Beneath that passage is a chart which tabulates the responses from Democratic-leaning voters. In the corresponding figure, 62% of Democrat-leaners say they'd be more likely to vote for a candidate who will compromise with Republicans.

It's true! In Pew's survey, Democratic voters were more likely to favor compromise. But as anyone can see, respondents from the two parties were much more alike than different.

This wasn't a case where 100% of one group said yes to some proposition while 100% of the other group said no. This wasn't anywhere close to that. But this is the only survey to which Chait links—and in this survey, responses from the two groups were much more alike than different.

(More alike than different? The large majority of Democratic voters could be matched with a Republican voter who answered the relevant question in the same way.)

Chait makes a very basic type of interpretive error here. He sees a pair of statistics in which one group favors A (on balance) while the other group favors B (on balance).

He then takes this (somewhat minor) difference in degree and describes it as a difference in kind. Quite explicitly, he turns it into an absolute difference.

"Republicans do not feel this way at all," he says, making a claim which is flatly false. As he does, he ignores an additional fact—a lot of Democrats didn't respond to the relevant question in the way he favors.

Chait took a relatively minor difference in degree and turned it into an absolute difference in kind. In recent years, we've seen an endless array of liberals making this type of error.

In fairness, it's a very old type of error. People have always thought this way as they happily marched toward the war their tribal loathing helped start.