Burnt Orange Report

True Lies

By Jim Dallas

Kevin Drum has a post on a poll that shows a collapse in support for the Bush agenda; but Kevin being the circumspect fellow that he is, feels compelled to note that a significant part of the sample probably lied when responding whether they were registered voters.

That there might be a few people out there misrepresenting facts about their own voting history isn't surprising; it's more than the rule than the exception. Moreover, how people lie can shed some interesting light on what they think the "correct" (as in socially-sanctioned) answer is. For example, there's a tendency for people to say they voted when they didn't, since voting is considered good, and not voting is considered bad. And, perhaps out of a subconscious fear that the thoughtpolice will arrive at their door if they tell the pollster the "wrong" answer, people lie.

Another traditional pattern of misrepresentation (innocent or intentional) is that respondants often tend to claim they voted for the winner of an election in rates out of proportion to the actual voting results (even accounting for margins of error and sample bias). Here are some recent examples:

• In a Gallup poll finished on January 13, 1997, 54 percent of respondants reported voting for Bill Clinton, 32 for Bob Dole, and 7 for Ross Perot. The actual election results were 49-41-8. Note especially the big drop off in reported support for the loser, Bob Dole.

• In a Gallup poll conducted on December 28, 1992, 51 percent of respondants reported voting for Bill Clinton, 29 for George Bush, and 15 or Ross Perot. The actual results were 43-37-19.

• In a Gallup poll conducted during Operation Desert Shield (October 28, 1990), 48 percent of respondants reported voting for George Bush in 1988; 23 percent reported voting for Michael Dukakis. A substantial number could not remember or said they did not vote; so what's relevant here is the 48:23 ratio, which is more than a 2:1 split for Bush. The actual 1988 split was 53:46.

Similar, albeit less dramatic disparities between reported and actual votes, can be found after the Reagan-Mondale election in 1984, and the Carter-Ford election in 1976. The data in 1980 is mixed, perhaps because Reagan inherited a bad economy.

I would argue that the rate at which respondants lie is tied to the winner's job approval rating. Note different polling results in 2001, before and after the terrorist attacks on September 11 (which sent Bush's approval into the 90s). In a Gallup poll conducted on August 5, 2001, 47 percent of respondants reported voting for Bush, 48 for Gore. In a Gallup poll conducted on November 4, 2001, 50 percent reported voting for Bush, and 41 for Gore.

Why do I bring all this up? Well, in the clip of the New York Times poll which Kevin Drum critiques, the respondants claimed the following: 41 for Bush, 41 for Kerry, 6 percent "won't say". This response is backed up by Gallup's January poll (finished January 16), where 47 percent claimed to have voted for Bush, and 45 percent claimed to have voted for Kerry.

Simply put, it would appear that very few, if any, poll respondants are willing to lie for Bush right now; falsely claiming to be a Bush-backer is not perceived to improve one's self-image.

There's a pattern here. As such, I'm not too concerned about Kevin's objections and I put stock in the Times's original analysis:

Americans say President Bush does not share the priorities of most of the country on either domestic or foreign issues, are increasingly resistant to his proposal to revamp Social Security and say they are uneasy with Mr. Bush's ability to make the right decisions about the retirement program, according to the latest New York Times/CBS News poll.

The poll underscores just how little headway Mr. Bush has made in his effort to build popular support as his proposal for overhauling Social Security struggles to gain footing in Congress. At the same time, there has been an increase in respondents who say that efforts to restore order in Iraq are going well, even as an overwhelming number of Americans say Mr. Bush has no clear plan for getting out of Iraq.

Again, when people lie, they can reveal just as much as when they're telling the truth. Is that a scientific conclusion? Perhaps not, but I'd be willing to stake my fortune on it.