"For independents, identifying as strongly liberal or strongly conservative was less of a factor in political engagement than strongly valuing the independent identity."

...................................................................................................................................................................
Independents Are the Hipsters of American Politics
What does identifying as an independent tell us about how somebody will vote? As it turns out, not much. Research suggests this is a (surprisingly large) group that’s focused more on image than policy.
By Eric Horowitz, January 16, 2014

The slowing of the political news cycle always presents an opportunity for polling firms to generate buzz by unleashing a startling result. Last week’s winner was Gallup, which garnered attention for a poll showing that a record 42 percent of Americans identify as independents.

The finding incited a response that fell somewhere between courteous grumbling and mild pushback. Over at The Washington Post’s Monkey Cage blog, John Sides was quick to point out that about three quarters of people who identify as independents behave just like Democrats or Republicans. That is, they support a particular party and hold more favorable views of its politicians. Even Politico echoed Sides’ sentiment that identifying as an independent doesn’t reveal much about you.

Yanna Krupnikov and Samara Klar, political scientists at Northwestern University and the University of Arizona, respectively, then followed up at The Monkey Cage with a recap of their latest research. They’ve found that because partisanship has a negative connotation, many people choose to identify as independents in order to create a more favorable impression.

Perhaps the most striking finding about the self-administered independent label comes from another recent study by Klar. Using data from a pair of 2008 and 2009 surveys of over 3,400 respondents, Klar examined how two key factors, ideological strength and the importance of political identity, were associated with political engagement. In general, a more extreme ideology—which respondents reported as moderate (weakest), slightly liberal or conservative, liberal or conservative, or extremely liberal or conservative (strongest)—and a highly valued political identity are associated with more political engagement.

Klar focused on whether ideology and the importance of identity had different relationships with political engagement among four political identities: 1) strong partisans, defined as those who strongly identified as Democrat or Republican; 2) weak partisans, defined as those who weakly identified as Democrat or Republican; 3) leaning independents, defined as those who lean toward one of the major parties; and 4) pure independents, defined as those who reported no lean or party identification. Klar hypothesized that ideological strength would be the best predictor of political engagement for partisans, but that the importance of identity would be the best predictor for independents.

In the end, that’s exactly what she found. For independents, identifying as strongly liberal or strongly conservative was less of a factor in political engagement than strongly valuing the independent identity. The reverse was true for Democrats and Republicans. For people who identified with one of the major political parties, ideological strength was most closely associated with political engagement.

It also made little difference whether political identity was reported as being strong or weak, Klar found. All that mattered was whether a person identified with a political party or as an independent. If they identified as a partisan their engagement tended to be most strongly associated with ideological strength, and if they identified as an independent their engagement tended to be most strongly associated with the importance of their independent identity.

Though inferring the exact causal relationship between identity and engagement is difficult, the idea that political identity exerts a greater influence on the actions of independents adds to the portrait of a group that’s guided by image more than policy. It’s even possible that independents engage in certain political activities for the specific purpose of maintaining and affirming their independent identities. Whereas Democrats or Republicans would be more motivated to take action that stamps their ideology on the world, independents would be more motivated to take action that allows them to increasingly be seen as independents.

Most political prognostication tends to be based on the idea that actions are driven by the desire to generate a particular ideological outcome, not the desire to be seen in a certain way. That potentially throws a wrench into conventional analysis of independents. Attempting to predict their actions is like trying to predict the preferences of a hipster (with “hipster” being defined here in the most superficial and stereotypical way). Just as a hipster’s evaluation depends not on conventional objectivity, but on a subjective judgment about whether their evaluation affirms their identity as a hipster, the actions of independents may depend less on conventional political thinking and more on subjective judgments about whether an action affirms their identity as an independent.

Given the murkiness beneath the “independent” label, it’s worth asking whether certain types of polls or projections would be better off avoiding the term altogether. At the moment, pollsters essentially use the three-point Likert Scale. What would happen if we got rid of the midpoint? Could forcing a weak independent to woefully call themselves a Republican tell you more about what they’ll do than allowing them to call themselves an independent? The future will surely shed more light on what makes independents tick, but for now you should probably take what you hear about them with a moderately sized grain of salt.
...................................................................................................................................................................