Free Markets, Free People


Poll Battle

Do Americans support the stimulus bill proposed by Congress, or hate it? The only way to glean a credible answer is by looking to reliable polls. Bruce did that earlier with respect to the ATI-News/Zogby poll which found that:

Amidst all the rhetoric surrounding President Barack Obama’s first signature piece of legislation, a massive $800 billion economic “stimulus” bill, one thing is clear: a majority of Americans reject the President’s handiwork. A just-released ATI-News/Zogby International poll shows that clear majorities of Republicans and Independents are against it.

On the other hand, a USA Today/Gallup poll found that support for the bill has dramatically risen [HT: Hot Air]:

Public support for an $800 billion economic stimulus package has increased to 59% in a USA Today/Gallup poll conducted Tuesday night, up from 52% in Gallup polling a week ago, as well as in late January.

So which is it? Is support up or down? Frankly, I don’t think we can really tell. Here’s why.

Both polls reveal the number of people questioned, and break down the results by party affiliation (although the ATI-News commissioned poll did not provide any numbers for Democrats). However, neither poll details how many participants of each party were polled, and/or whether the results were weighted. In short, if the ATI-News poll included substantially more Republicans and Republican leaning people among the 7,010 voters questioned, then the results should predictably skew towards the Republican side of the issue. Similarly, if there were significant number of Democrats and Democratic leaning independents among those 1,021 national adults polled by the USA Today poll, then we should expect that poll to favor the Democratic side.

Because we aren’t informed as to the breakdown of the total by party affiliation, we really can’t say how reflective the polls are of the country as a whole. Seeing as how the polls contradict one another, it’s safe to say that neither one accomplishes that task.

It’s tempting to conclude that, since the ATI-News poll was conducted over 5 days, as opposed to one, and interviewed almost 7 times as many people as the USA Today poll, the larger sample provides a more accurate picture. Moreover, the poll showing that the public is against the stimulus bill claims a margin of error (+/- 1.2%) that is far lower than the other poll (+/- 3.5%). Yet, the confidence interval for the latter poll is 95% and none is given for the ATI-News offering. If it was only 90%, I think (but could be wrong) that makes the USA Today poll slightly more accurate. In addition, without knowing how many answers came from each party (D/R/I), it’s impossible to say just how representative the poll actually is.

By the same token, the USA Today poll appears to offer a more comprehensive look at those questioned, and the questions asked seem less likely to evoke biased answers. For example, the main USA Today poll question was this:

As you may know, Congress is considering a new economic stimulus package of at least $800 billion. Do you favor or oppose Congress passing this legislation?

Compare that question to the following:

Most Republicans oppose the currently proposed stimulus bill supported by President Obama because they say there is too much money being spent for non-stimulus items. Do you agree or disagree that too much money is being spent on items that won’t improve the economy?

The first question above is simple, straightforward, and doesn’t present any potential bias words with respect to the issue. The second, however, sets up a premise, attaches “Republican” to it, and then asks for agree or disagree. Not surprisingly, the second question elicited a much stronger response from Republicans (93% agreed) and Independents (66%) than the first (56% Independents; 28% Republicans). Perhaps then the USA Today poll, despite its small sample, is the more accurate?

Once again, we don’t know how many of each party were questioned. If it was overwhelmingly Democratic Party leaners, then the results would have to be expected. In addition, the USA Today poll questioned all adults, while the ATI-News poll only queried voters, whom one might assume are somewhat better informed. Finally, the fact that any poll of voters could find a string correlation between the words “agree” and “Republicans” suggests that the wording was not causing any undue bias (unless, of course, it was mostly Republicans interviewed, which is pretty unlikely).

In the end, I don’t know how to view these two contradictory polls in a way that sheds any light on how the populace is actually feeling about the stimulus bill. Other than the glaring fact that Democrats overwhelmingly favor its passage, while Republicans do not, there is nothing definitive to be learned. I do agree with Bruce’s assessment that Independents are the best to look for answers, however the poll numbers we have don’t seem to match up.

I guess its possible that the a majority of people are ambivalent about the stimulus bill — yeah it’ll probably be a big screw up, but we have to do something, don’t we? — which would explain some of the apparent contradiction. And maybe Obama’s sales job made the difference in the numbers (the ATI-News poll ended on the 9th, while the USA Today poll was taken on the 10th).

Whatever the reason for the contradiction, I think it’s interesting that each day we have a different poll telling us that the public loves/hates the stimulus package, yet we never see any polls testing the public’s knowledge of what’s in the bill (much less anyone in Congress). Maybe if people were better informed about the contents of the legislation we see more consistent polling. Instead of constantly reading polls asking if the Republicans are right or wrong, or if $800 billion is a good number to spend, perhaps we’d learn more about what the public really thinks if we asked them how stimulated they would be by $4.2 billion for “neighborhood stabilization activities,” or $34 million to renovate the Department of Commerce headquarters, or $88 million to help move the Public Health Service into a new building, or $55 million for Historic Preservation Fund, or $6.2 billion for the Weatherization Assistance Program, or $2.4 billion for carbon-capture demonstration projects. Now there’s a poll I’d like to see.

  • Facebook
  • Twitter
  • LinkedIn
  • Tumblr
  • Digg
  • Reddit
  • email
  • Print
  • Google Bookmarks

7 Responses to Poll Battle

  • Ah, you may need to read a quick primer on confidence intervals.  A 95% confidence interval of (n-z) just means that you are “95% confident the true value falls within this range”.  Confidence intervals are a reflection of the statistic and the error within measurement.  I’m not really familiar with how confidence intervals apply to poll questions (as the data is inherently scored or binary, not really quantifiable).  The quick equation for a 95% CI = mean +/- 1.96*error.  It looks like they’re talking about the confidence in the error rate, which involves a bunch of calculations in the background.

    Long story short: they’re just saying they’re pretty sure the error on the latter poll is +/- 3 points.  I’d wager they’re pretty sure the error on the former poll is 1.2 points, they just didn’t specifically say “95% confidence”.  As a general rule, all else being equal, the larger a survey group, the lower the error rate will be.

    • I completely missed my greater point though haha.  Your instincts are correct.  Given the (lack of) information, the larger survey is more reliable as it will have a lower error rate and ostensibly samples a greater cross-section of the general population.  The polls need to be taken with a huge grain of salt though, because as you mention, no sampling information is given nor weightings considered.

      Your points looking at the wording of the questions are very valid.  I like your take on it, although honestly I think the cross-section and weighting data carry much more influence than the wording in this case.  The MOST interesting thing, I think, is looking at the independents.  When the question is attached to the Republican qualifier, 66% of independents side with it.  It’s too early to call, but could this signal a shift?

    • Thanks, calvaria.  I had a feeling I was wrong about that, but it’s been almost 20 years since I took a statistics course.

  • Can staticians put that selves out of a job?

  • Polls are meaningless, the public are sheep.

  • Informal poll where I work, 100% union workers, 10 Dems of which 5 are hard core progressives, 4 conservatives of various stripes, 12-2 against.