Thursday, August 4, 2016

What do you mean I live in an epistemological bubble?

This would seem an example of the subtle media bias and spin that so many believe is characteristic of the mainstream media and academia. The tweet is from Conrad Hackett, a demographer who works for the Pew Research Center. I do not know Hackett nor do I know his ideological affiliations (if any).

The general issue, which we already know, is that there is a tendency for people to live in racial, financial, class, educational, religious, and ideological bubbles. Most the people with whom you electively spend time have some high degree of affiliative similarity with yourself. Nothing particularly surprising about that but there is plenty of political hay to be made out of it. Charles Murray has a fun Bubble Test that focuses on the tendency of people to live in class bubbles.

Pew has just completed some research and Hackett tweets the results:



Fair enough. Interesting and confirmative of what we know. In this case, the supporting evidence is that a third of Trump supporters say they have no close friends who support Clinton. By drawing attention to Trump's supporters, Hackett seems to be imputing that this is the most notable finding and that Trump supporters are distinguished by the degree of their epistemological closure. It's as if to say, "Can you imagine? 31% don't know anyone who supports Clinton?" Included in the tweet is the table with the supporting data.

But looking at the data, it appears that Hackett has diverted our attention from the more striking result. Among Clinton supporters, 47% do not have any friends who support Trump. In other words, Clinton supporters are 50% more epistemologically closed than Trump supporters.

So why wasn't the closed mindedness of Clinton supporters the headline of the tweet instead of the lesser closedness of Trump supporters? Meta-epistemological closure on Hackett's part? Confirmation bias? Motivated reasoning? Sheer ideological hackery? It's hard to tell. Hackett clearly did not misunderstand the data or overlook the difference.

From the originating research Hackett notes that:
Nearly half of Clinton supporters (47%), and 31% of Trump supporters, say they have no close friends who support the opposing candidate.
I suspect that Hackett is a thorough professional. I am inclined to doubt that this was politically motivated.

But I also suspect that this is an instance of the the type of epistemological closure so characteristic of academia, media and the hot-house Acela corridor where vested interests are so tightly and incestuously interconnected. I am guessing that in Hackett's circles, it is to be expected that Trump supporters are small-minded, prejudiced, epistemologically closed individuals and therefore it was very easy to see that the data proved that to be the case. And, apparently, very easy to overlook the worse numbers for Clinton supporters.

All speculation of course, but the routine instances of this blindness of the chattering classes to their own preferences and biases speaks volumes to everyone else.

UPDATE: Apparently much ado about nothing. Hackett had a subsequent tweet shortly after the first one which drew attention to the HRC numbers. Following so many research sites, even a minute's difference between tweets can mean that paired tweets are not seen in pairs.



No comments:

Post a Comment