Skip Navigation

Close and don't show again.

Your browser is out of date.

You may not get the full experience here on National Journal.

Please upgrade your browser to any of the following supported browsers:

How 'Independent' Are Independents? How 'Independent' Are Independents?

This ad will end in seconds
Close X

Want access to this content? Learn More »

Forget Your Password?

Don't have an account? Register »

Reveal Navigation



How 'Independent' Are Independents?

Tea Party Polling Finds Far Fewer Maverick Voters Than Advertised

Updated at 10:34 a.m. on Feb. 22.

Buried in a report on a recent survey of the Tea Party movement was the apparently innocuous observation that, "according to the survey, most Tea Party activists describe themselves as independents."

Read more and comment on this column at

Pollsters from CNN and the Opinion Research Corporation had identified 11 percent of Americans who said they have given money, attended rallies or taken other "active steps" to support the Tea Party movement. Of those who had, more identified themselves as independent (52 percent) than Republican (44 percent) or Democratic (4 percent).

But the article then adds a caution from CNN pollster Keating Holland that some will find puzzling: The apparent independence of Tea Party supporters "might be slightly misleading," he said, "because 87 percent say they would vote for the GOP candidate in their congressional district if there were no third-party candidate endorsed by the Tea Party."

I highlight these results because they raise some fundamental questions about how pollsters identify political independents and analyze their attitudes, a point I knew would resonate with my friend, George Washington University political scientist John Sides.


Sides has been on something of a crusade of late to remind journalists, pundits and anyone else who follows politics of something that political scientists have long understood: Very few Americans are truly "independent." In fact, most of those typically identified as independents in political surveys are, to quote the seminal text on the subject, "largely closet Democrats and Republicans."

Sides wrote a must-read blog post in December that goes into this issue in detail, but I'll try to summarize: Most pollsters ask about partisan attachments using the question asked by American National Election Studies for nearly 60 years: "Generally speaking, do you usually think of yourself as a Republican, a Democrat, an independent or what?" (Sometimes they use minor variations.)

The independent label on most polls applies to the respondents who choose that category. Most national surveys over the past year show slightly more than a third of Americans initially identifying as independents (although some polls will show a higher number depending on how they handle respondents who offer the name of a third party or who say they are unsure).


However, the ANES surveys have always included a follow-up question asked of the independents: "Do you think of yourself as closer to the Republican Party or the Democratic Party?" As Sides point out, "the vast majority" of those who initially identify as independent will "lean" toward a political party, so the number of remaining "pure independents" is typically small -- only about 10 percent of the population.

Far more important, independents who lean to one party typically "act like partisans," as Sides puts it. They vote for their side's candidates as often as those who initially identify with a party but describe their attachment (on yet another follow-up question) as weak.

The chart below (originally created by Sides but updated with the assistance of Jon Cohen and Jennifer Agiesta of the Washington Post) shows that independent leaners look much like partisans in terms of their approval rating of President Obama. Only the "pure independents" (just 7 percent of the population) are divided in their views of Obama.

So if the tiny swath of pure independents is a better representation of truly "independent" Americans, why do so many pollsters typically report results for the larger, potentially misleading group? Some omit the follow-up question to save precious interview time. More often, they report on the larger group because the "pure independent" rarely yields sufficient interviews for reliable analysis. A national sample of 1,000 interviews, for example, will yield less than 100 interviews among pure independents (and Cohen informs me that the Post will not report poll numbers for subgroups of less than 100).

That said, results tabulated for the larger independent category will typically resemble those among pure independents, mostly because the larger group includes a mix of Republican and Democratic leaners that roughly balances out.

But not always, and that brings me back to the CNN poll. Remember the 52 percent of Tea Party activists who initially identify as independent? It turns out that virtually all of them lean Republican. According to CNN, 88 percent of the activists identify or lean Republican, 6 percent identify or lean Democratic and only 5 percent fall into the pure independent category.

Remember that CNN pollster Holland reported that 87 percent of the Tea Party activists would vote Republican if there were no Tea Party-endorsed third-party candidate running? That makes perfect sense for a group that is 88 percent Republican.

Holland said that the 52 percent independent number is "slightly misleading." Sides said via e-mail that he considers it "highly misleading," since most independent leaners vote loyally for their party, even among the Tea Party activists identified in their survey. That "the Tea Party activists are mostly composed of Republicans, not independents," he writes, "combined with the relatively small number of activists, may hamper the electoral prospects of any Tea Party candidates."

This also tells us that Republican candidates will suffer if Tea Party activists start to mount third party candidacies.

CLARIFICATION: The American National Election Study has not changed its partisan attachment question.

comments powered by Disqus