May 15, 2006
Gallup on NSA Phone Records
Another day and yet another poll on the NSA phone records controversy. This one, conducted on Friday and Saturday by USAToday and Gallup (n=809, margin of sampling error +/- 4%), takes a far more in-depth look at the issue that helps clarify why the two polls we saw over the weekend produced such widely varying results. The poll is described and data released in USAToday (story, results) and also in a summary by Jeff Jones just posted on Gallup.com that is free today, available to subscribers only after that.
The complete results - and especially the full Gallup report -- are definitely worth reviewing in full. Here are some noteworthy highlights. The "money question," as it were, is part of two question sequence:
As you may know, as part of its efforts to investigate terrorism, a federal government agency obtained records from three of the largest U.S. telephone companies in order to create a database of billions of telephone numbers dialed by Americans. How closely have you been following the news about this?
28% very closely
39% somewhat closely
20% not too closely
12% not closely at all
<1% no answer
Based on what you have heard or read about this program to collect phone records, would you say you approve or disapprove of this government program?
6% no opinion
The net disapproval of the question is certainly closer to the result obtained by the Newsweek poll conducted on Thursday and Friday nights last week than the results obtained on Thursday night by the ABC News/Washington Post poll. See my prior posts (here and here) for more details on those surveys.
The classic question about how much attention Americans were paying to this issue over the weekend is helpful for two reasons. First, the attentiveness to this issue is, according to Gallup, "an average level of attention when compared to other prominent news events Gallup has measured over the last 15 years."
Second, it allows the pollster to check whether attitudes are different among those paying close attention, those whose opinions are more likely to be "fully formed" (as an MP commenter put it over the weekend): "Notably," according to their summary, "the Gallup Poll finds even stronger opposition among those who are following the story very closely, with 41% approving and 58% disapproving."
Probably the most interesting questions are two follow-ups Gallup asked to try to get at one of the most frequent criticisms of the ABC/Washington Post questions, that it forced a choice between between protecting civil liberties and investigating terrorism:
The poll sought to understand the reasons behind support of and opposition to the program. For example, does support stem from the belief that preventing terrorism trumps possible violations of civil liberties, or do supporters not see the program as really violating civil liberties?
Here is are the findings and interpretation from the Gallup summary:
Among those who say they approve of the program, most (69%) cite the importance of fighting terrorism, even if civil liberties are violated. Twenty-seven percent of approvers say they don't think the program "seriously violates Americans' civil liberties." The strong support for fighting terrorism echoes the ABC News/Washington Post result
BASED ON 349 ADULTS WHO APPROVE OF GOVERNMENT PROGRAM Is that mainly because -- [ROTATED: you do not think the program seriously violates Americans' civil liberties, (or is it mainly because) you think investigating terrorism is the more important goal, even if it violates some Americans' civil liberties]?
27% Does not seriously violate civil liberties
69% Terrorism more important, even if it violates civil liberties
4% No opinion
Those who disapprove of the program may do so because they see it as wrong on its face, or they may see the use of it as justifiable but not under the current circumstances. Most poll respondents who disapprove of the program reject the notion that such government action can ever be justified, saying there are no circumstances in which it would be right for the government to collect private phone records.
BASED ON 411 ADULTS WHO DISAPPROVE OF GOVERNMENT PROGRAM Do you think there would ever be circumstances in which it would be right for the government to create a database of telephone records, or would it not be right for the government to do this under any circumstances?
34% Yes, would be
60% No, would not be right under any circumstances
6% No opinion
Another way of looking at the same results is to calculate all of the percentages out of the full sample of 809 results (something done in both the Gallup summary and the USAToday results). Thirty one percent (31%) of all adults oppose the creation of a phone records database "under any circumstances," but 17% can imagine "circumstances in which it would be right." Another 6% intially oppose the program but are unsure on the follow-up. The 23% that are opposed but not under all circumstances help explain the very different result on the ABC/Washington Post survey. Odds are good that most answered "acceptable" when asked by that survey whether they consider the NSA phon records collection to be "an acceptable or unacceptable way for the federal government to listen to terrorism" (a point made by several astute MP readers in comments yesterday).
Here is what Gallup's Jeff Jones concludes about the different results produced by the three most recent surveys:
It is not entirely clear why the ABC News/Washington Post one-night poll showed a larger percentage of positive responses, although the question stated that the collected phone data were analyzed "in an effort to identify possible terrorism suspects."
Because these questions are designed to measure public opinion in an area that involves two competing principles, fighting terrorism and protecting citizens' privacy, it may not be surprising that specific phrases or nuances of wording in survey questions could cause different response patterns.
The Gallup report is worth reading in full -- do it today before it falls behind Gallup's subscription wall.
That the Newsweek and USAT polls use data from a Friday and a Saturday is also a likely factor influencing the result. Mr. Blumenthal has written on weekend polling, but this apparently escaped his notice in this case.
Posted by: Karl | May 15, 2006 5:36:56 PM
The comments to this entry are closed.