Opinion polls, economic "facts," and the sociology of knowledge

The polling results below, from The Pew Research Center back in September, tell the kind of story that should be disheartening for anyone committed to a naive-rationalist approach to understanding politics and human action.



In the real world, there is a strong tendency for all of us to hear or notice the "facts" we want to hear or notice and to blank out the rest. Furthermore, contrary to the illusions of naive empiricism, "facts" never speak for themselves, but have to be made sense of within frameworks of interpretation. And people pay more attention to the interpretations of (actual or alleged) "facts" that emanate from sources they trust and would like to believe.

It's important to note that the mechanisms involved here are not entirely reducible to features of individual psychology, or even to the dynamics of social psychology. Actually, the factors involved are sociologically complex. What we would like to hear is strongly influenced by (more or less intense) feelings of social identity (including, among other things, partisan identification), group membership, and group solidarity. One implication is that people's opinions (or rather "attitudes') on particular issues are often pretty easy to change or manipulate, since considerations of identity and group membership are often more important to individuals than their concern with, or grasp of, those substantive issues. Also, members of different subcultures are immersed, to a variable but often surprisingly great extent, in different ideological universes, within which they get different streams of messages that constantly reinforce certain world-views and images of reality while shutting out or discrediting alternative world-views. With respect to assessing polling results, there is an additional complication. The responses that people give to pollsters' questions (which are often quite ambiguous in substantive terms) may express what they actually think about a particular issue, or they may simply reflect what respondents think is the politically "correct" answer for someone with their political commitments (or some combination of the two).

And so on. A piece by Ben Smith & Zeke Miller at Buzzfeed offered a fairly perceptive analysis of that graph and its implications::

In August of 2011, Americans of all parties said the news was mostly bad, with only minor differences showing between members of different political parties.

A year later, a survey taken in early September found a "record partisan gap." A full 60% of Republicans said they were hearing “mostly bad” news. Only 15% of Democrats reported the same. And independent voters split on the question, with 36% saying they were hearing mostly bad news.

It’s not just a matter of what voters are hearing. Gallup’s tracking of Americans’ reported confidence in the economy has also seen a dramatic divergence: Democrats’ confidence reached a new high in a survey released September 25; Republicans’ reached a record low.

An online survey by the firm CivicScience, among those that has sought to develop reliable online metrics, found that among the Romney supporters it surveyed over the last three months, 96% labeled the economy “weak.” The corresponding figure for Obama supporters is 55%, according to the company's data director, Ross McGowan.

The pattern is a familiar one to pollsters and political scientists alike, though some say it has intensified with the changing and increasingly partisan media.

“Cues and signaling from the political leaders definitely influence how people experience their own lives,” said Democratic pollster Harrison Hickman, who said he’d seen the trend “aggravated in recent years.”

He attributed the vast gaps in public polls to a mixture of that psychological factor and of poll respondents’ awareness of a “politically correct” answer. [....]

The landmark 1960 The American Voter, a study of the elections of 1948 through 1956, found something similar of voter attitudes toward the Korean War, speculating that when a voters’ views conflict with his party allegiance, “allegiance presumably will work to undo the contrary opinions.”

“The influence of identification on attitudes toward the perceived elements of politics has been far more important than the influence of these attitudes on party identification itself,” the authors wrote. That is: Party identification appeared, at times, to trump voters' experience of reality.

A debate over “partisan perceptual bias” has raged in the political science literature ever since, Princeton’s Larry Bartels noted that it was particularly pronounced during Ronald Reagan’s presidency, with surveys showing that “Democrats were strikingly impervious to the good economic news.” Lee Drutman noted in Slate in 2010 that something similar seemed to apply to Republicans' reporting of their own economic conditions between 2008 and 2010.

In this light, the raft of polling showing that the economy is at the top of voters minds are a distraction. “The economy” simply means different things to different people.

None of this means, by the way, that we should give up on the effort to come to grips with reality, in politics or in other areas of life, and just surrender to cynicism, irrationalism, or uncritical relativism. These are powerful tendencies, which need to be recognized rather than ignored, but there are also ways to counteract them. And, of course, not all political and ideological subcultures, in the US and elsewhere, are equally impervious to factual evidence or equally trapped in hermetically sealed, self-reproducing, and outright delusional world-views.

Yours for reality-based discourse (with a sober recognition that this is not a simple or easy enterprise),
Jeff Weintraub