Gallup, Barna: A Tale of Two Polling Operations and Their Curious Midterm Numbers

Two polls were recently released. One indicates the Democrats are in deep trouble come November. The other not so much. Polls taken today are about as meaningful as preseason football games, but it's still a puzzlement.

Let's do a tale of the tape:

In this corner, weighing in with 1,540 registered voters queried Aug. 23-29 with a claimed 95 percent confidence level (plus or minus four percentage points): The venerable Gallup Poll.

In that corner, weighing in with 1,000 adults queried Aug. 16-22, with a claimed 95 percent confidence level (plus or minus 3.5 percent): The Barna Group.

Up close and personal:

Gallup is doubtless among the best known names in polling, with more than seven decades of experience. It was founded by George Gallup Jr. in 1935, who "resolved that he would undertake no polling that was paid for or sponsored in any way by special interest groups such as the Republican and Democratic parties," according to a company history.

Barna is exceptionally well known within a smaller circle, founded by George and Nancy Barna in 1984. The Barnas make no bones about their support of a particular definition of evangelical Christianity. Their agency seeks "to be a catalyst in moral, social, and spiritual transformation."

But the Barna methodology seems to be as bulletproof (or not) as Gallup. It offers the same kinds of claims about random, representative sampling. Same level of transparency about the questions they ask. When Barna goes all faith-talk and does a religious breakdown of its responders, it's totally specific about how it defines, say, "born-again" and "evangelical," which is not always true in such discussions. And Barna seems as happy to report interesting results that run counter to its theological mandate as it is to report findings its founders surely applaud. I take their data as seriously as I do that of any other pollster.

So riddle me this, as the old Batman villain used to say.

Gallup reports that:

"Republicans lead by 51% to 41% among registered voters in Gallup weekly tracking of 2010 congressional voting preferences. The 10-percentage-point lead is the GOP's largest so far this year and is its largest in Gallup's history of tracking the midterm generic ballot for Congress."

Barna, on the other hand, says that:

"If the election were held today, only a slim majority of voters would return the Democrats to a majority position in both the House and Senate. . . . Among registered voters, 45% wanted Democrats to rule the House while 40% desired a Republican takeover. Regarding the future of the Senate, the numbers were almost identical: 45% preferred a Democratic majority versus 41% opting for a Republican majority."

Let's grant that the two polls slice different bits out of the body politic for their biopsy. And they ask different questions.

Barna says it asked "which party the respondent wanted to dominate the House and the Senate." For the House, 15 percent did not choose either major party. For the Senate it was also 15 percent.

Gallup presented what it calls (but does not define) a "generic ballot" for Congress, presumably with an unnamed Republican facing an unnamed Democrat, and found that 9 percent of respondents did not choose either of the major parties.

Given the margins of error, both polls are pretty close on the Democrats. But Gallup sees a much larger bump for the GOP.

Digging below the overall results for Congress, we can find a few interesting details.

For Gallup: Many more GOP voters say they are enthusiastic (50 percent) than do their Democrat (25 percent) or independent (28 percent) counterparts.

For Barna: When asked if they were likely to vote for President Obama for reelection, respondents split evenly -- "45% said they would vote for him and 45% said they would not. Another 5% said it depended who was running against him and the remaining 5% did not know."

Can't get more even that that. And given the events of the summer, and the length of time until the next presidential election, I can't imagine there's a Democratic strategist in America who wouldn't be happy to take those numbers as gospel truth.

Barna being Barna, the poll sliced 'n' diced along religious lines:

"For instance, evangelicals are overwhelmingly conservative in their political views and fervently desire a Republican return to power in both chambers by about a 5-to-1 margin. Non-evangelical born again Christians, who are generally moderate with slight conservative tendencies, generally favor the Republicans taking over both houses. However, they are offset by notional Christians, a group that is moderate to slightly liberal, who prefer a Democratic majority in each chamber. Further compensation comes from the Democratic preference held by both the Skeptics (3-to-1 in favor of the Democrats) and Americans of non-Christian faiths (a 5-to-2 Democratic preference). Both the Skeptics and the adherents of non-Christian faiths are typically moderate or liberal in their perspectives."

And here's a not-particularly-meaningful but fun head-scratcher about support for Obama:

"If the election were held today, his support among evangelicals would actually have risen slightly, from 11% in 2008 to 16% now. Surprisingly, that makes evangelicals the only religious subgroup from which he has experienced increased support. (Of the various faith segments that Barna tracks, however, evangelicals remain the group that awards the president the lowest level of support, despite the minor uptick.)"

My bottom line: It's still the preseason. Campaigns really get into gear after Labor Day. Let's see what the numbers look like -- for any poll -- in a month or so.