Our first stop was Gallup, the polling organization, which has recorded the national appetite for capital punishment for more than seven decades. Gallup's most recent survey, conducted October 2011, asked respondents whether they were "in favor of the death penalty for a person convicted of murder." The response was 61 percent in favor, 35 percent opposed and 4 percent without an opinion.

Half true, in that it is true as far as it goes but leaves a lot out in a misleading way.  Citing only the latest number, which is possibly a blip in response to the Troy Davis circus, is a tad misleading.  So is ignoring the fact that Gallup itself has noted that their standard question doesn't necessarily capture the true extent of support for the death penalty for the worst cases.  Another question that Gallup sometimes asks captures public opinion on the real question better.  "In your opinion is the death penalty imposed too often, about the right amount, or not often enough?"  People opposed to the death penalty in all cases surely think it is imposed too often, and that number has never exceeded 23%.

Generally, the data that Gallup has on hand shows a growing discomfort nationally with the death penalty. In the mid-1960s, opposition eclipsed those in favor, but it soon began to fall until the mid-'90s when opposition hit an all-time low of 16 percent. Since then, the tide has turned again, and opposition has been on a slow climb. Of course, it's still clear the majority of those polled favor capital punishment in some cases.

Half true.  The paragraph gives the impression of a slow but steady rise in opposition.  In fact, there was a sharp "reversion to the mean" jump in the late 90s after the all-time low, followed by a period of stability. In a 2010 report, Frank Newport of Gallup noted "a trend that has shown little change over the last seven years."

One other poll from late September 2011, this one done by CNN, showed that 50 percent of respondents favored life in prison as a penalty for murder, while 48 percent favored the death penalty.

Half true.  Failure to mention that this question is worded in a way that implies the respondent must choose one punishment for all murders, which is not at all the question to be decided, is a seriously misleading omission.  Also seriously misleading, in light of the way this number is used later in the article, is the failure to mention that results on this question have been in the split-down-the-middle range for years, so this 50-48 result is not much of a change.

We called Marquis to see whether he had any evidence to back up his claim. He had a lot to say.

First, he clarified that he wasn't talking nationally, he was referring to Oregon specifically when he made that 25 percent comment. As for his evidence, he said that back in 2002, there was a group that was considering mounting a campaign to abolish the death penalty in Oregon. But, he said, they backed off after polls showed just 20 to 25 percent opposed.

We pointed out that Gallup showed similar numbers 10 years ago, but since then the number opposed has changed.

That last statement is mostly false.  Overall, the Gallup Poll has found little change over the last decade, as noted above.

Marquis was unconvinced. He said the poll's question made the issue appear too black and white....

We talked to Stuart Elway, an independent pollster up in Seattle who works regionally. First we asked him what he thought about the Gallup poll and whether Marquis' criticism seemed fair. He said it's possible that the question, if asked differently, could elicit a slightly changed result -- but only a couple points in either direction.

"Liar, liar pants on fire!"  This is PolitiFact's lowest rating, and they have earned it themselves in spades with this one.  As we have described several times on this blog, including this post, the numbers vary wildly in death penalty surveys depending on how the question is asked.  In the same poll, the results were 67-28, 48-43, or 83-16 depending on how the question was asked.  That's way more than "a few points," and this truth is not at all hard to find.  Elway's statement and PolitiFact's uncritical repetition of it is either an intentional lie or a statement made with reckless disregard of the truth, which is pretty much the same thing.  (See New York Times v. Sullivan.)

He did, however, agree with Marquis that it wasn't necessarily fair to tie the national polls to Oregon. Attitudes differ regionally and could certain affect the results by 5 to 10 points.

Still, Elway said, the fact that Oregon was trending with the nation in 2002, probably indicates that the trend Gallup has found over the past few decades is somewhat present in Oregon as well.

True as far as it goes, but since they mischaracterized the trend Gallup found, this statement misleads the reader as to the bottom line conclusion.

All that said, the facts are these: There is no recent Oregon-specific polling regarding the death penalty and national polls tend to vary pretty wildly, depending on how the question is asked.

Um, right, but the "vary pretty wildly" flatly contradicts what they said just a few paragraphs earlier.

Marquis is basing his statement on numbers from at least a decade ago. The Gallup and CNN polls seem to indicated that the national mood is changing.

Mostly false.  Except for one recent bump, which may be a temporary effect from intense but soon forgotten media coverage of one case, the Gallup Poll shows remarkable stability.  The CNN poll does not indicate a change.  The question worded that way always shows radically different numbers from the standard question, as they seemed to recognize with the "vary pretty wildly" comment but now seem to have forgotten.  (Pretty short attention span.)

Marquis' statement might have the smallest grain of truth, but it ignores some important details.

"Smallest grain of truth" is false.  His statement has a lot of truth.  In the best-worded question asked by any poll, opposition was only 16%.  That was Connecticut and not Oregon, to be sure, but both states have tended to trend with the country as a whole as shown by answers on differently worded questions.  In the best-worded question Gallup asks nationally, opposition runs around 21%.  So 25% may actually be an overestimate.  "Ignores some important details" is an apt description of PolitiFact's own analysis.

Overall, PolitiFact's fact-checking rates a "mostly false."  The points on which it is true are the relatively minor ones, while the points on which it is false or at least misleading are the more important ones.