Real Spies, Fake Spies, NSA, and More: What My 2012 and 2013 National Polls Reveal

In August 2012, thanks to YouGov, I launched my first national survey to probe more deeply about what Americans know about intelligence agencies, what they think about controversial intelligence programs, and where those attitudes come from. In light of the Edward Snowden revelations, last month I asked YouGov to run another poll that asked some of the same questions, along with new ones about NSA so that I could start tracking trends over time. The poll ran Oct. 5-7, 2013, and included 1,000 people (with a margin of error of +/- 4.3 percent). I found three interesting results that admittedly invite more questions than they answer. First, confidence in the accuracy of U.S. intelligence overall has declined by 8 percent from August 2012 to October 2013. Here’s the exact question: "How confident are you that the U.S. intelligence community is giving the Obama administration accurate information about possible threats to the U.S. from places such as Iran and North Korea?"

Why the decline? One hypothesis is that Americans have more negative views of intelligence agencies in the wake of Snowden’s revelations, and confidence in intelligence analysis has suffered as a result. Or to put it differently, maybe the favorability ratings toward agencies affect confidence in those agencies’ work product. But I found that CIA favorability remained almost exactly the same from 2012 to 2013 (53 percent held a favorable view of CIA in 2012, 50 percent in 2013. That’s within the margin of error). Favorability also held steady for the FBI and Department of Homeland Security in the last year. (Unfortunately, I didn’t ask about public favorability toward NSA in 2012, so I can’t do the same year-to-year comparison). It appears that Americans like the CIA, FBI, and DHS about as much as they did a year ago. Americans just don’t trust the accuracy of intelligence provided by these or other three-letter intelligence agencies as much now. Given these numbers, the most likely explanation is that NSA’s woes have been dragging down public confidence in the accuracy of the intelligence enterprise writ large. This is interesting because accuracy has never been part of the NSA debate. News headlines have not called into question the veracity of the information the agency collects. In fact, media coverage has suggested the opposite: that NSA is an all-seeing eye that acquires vast amounts of real data about our private lives---collecting now and asking questions later. The poll’s confidence numbers suggest that NSA is accurate, all right. But NSA is not instilling greater public confidence that the intelligence system of the U.S. government overall is serving the President well. Second is the the influence of spy-themed entertainment, or “spytainment.” I have been researching for some time whether spy-themed entertainment (particularly on television and the big screen) has become adult education for intelligence. Last year, I found a strong correlation between frequent spytainment viewing and approval of aggressive counterterrorism tactics, including rendition, assassination, and harsh interrogation methods. In fact, support for torture was higher in 2012 than it was during the Bush administration. The influence of spytainment, I posited, helps explain why. Here are some of the statistically significant findings from my 2012 national poll:

• 38 percent of frequent spy TV watchers believed that waterboarding terrorists was the right thing to do, compared to 28 percent of infrequent watchers.
• 60 percent of frequent spy TV watchers thought transferring a terrorist to a country known for using torture was right versus 45 percent of infrequent watchers.
• 34 percent of frequent spy moviegoers said that they thought it was right to chain terrorist detainees naked in uncomfortable positions in cold rooms for hours. Only 27 percent of non-movie goers thought the naked-chaining-stress-position approach was right.

My October, 2013 poll reveals that spytainment viewing habits also extend to the NSA. I found that the more people watched spy-themed television shows and movies, the more they liked the NSA, the more they approved of NSA’s phone and Internet collection programs, and the more they believed the NSA was telling them the truth. In many instances, opinion differences between spytainment viewers and the rest of the country were large. The statistically significant correlations include:

• NSA Favorability: 58 percent of people who watched spy movies six times or more in the past year had favorable views of NSA, but only 34 percent of infrequent spy moviegoers reported favorable views of the agency.
• Approval of NSA Surveillance Programs: 44 percent of people who watched spy TV shows frequently or occasionally said they approved of NSA’s Internet data and telephone records collection programs. But only 29 percent of those who rarely or never watched spy TV shows approved of these surveillance programs
• NSA Telling the truth: 23 percent of frequent/occasional spy TV watchers believed the NSA was telling them the truth about not listening in on telephone call content in the agency’s “metadata” collection program, compared to just 15 percent of infrequent spy TV show watchers who said they thought NSA was telling the truth.

Of course, the big question is causality: Is it that spytainment attracts a skewed viewership of NSA supporters? Or are these fictional shows creating real-world effects on public opinion? To begin to get at this question, we examined party identification and gender. It turns out, males who frequently watch spy television were still more likely to view NSA favorably compared to males who do not. The same was true for frequent watchers who are Democrats. Third, Americans’ knowledge of intelligence generally and of NSA specifically is pretty darn bad. A few key takeaways:

• 43 percent of respondents could identify the current Director of National Intelligence, James Clapper. But 74 percent could identify Miley Cyrus as the person who twerked at the MTV Video Music awards. (The real picture is probably worse, since we asked respondents to pick out Clapper’s name from a list of possibilities, not come up with his name out of thin air).
• 39 percent of respondents still erroneously believe (after consistently hearing otherwise from intelligence officials) that the NSA’s bulk telephone “metadata” program includes call content.
• 32 percent of respondents believe NSA conducts operations to capture or kill foreign terrorists and another 39 percent were not sure. In other words, 71 percent of Americans either didn’t know the answer to this question or got it wrong.
• 35 percent believe NSA interrogates detainees and another 42 percent were not sure. Yup—that’s more than three quarters of the population that could not answer this question correctly.

Given that nearly half of all Americans say they believe in ghosts, these numbers may not be terribly surprising. But here’s something that is: the more knowledge that Americans have about NSA and its activities, the less they support the agency. This is counterintuitive. I originally thought that more knowledge would lead to higher levels of NSA support. That’s what NSA officials think, too. NSA Director Gen. Keith Alexander, in a speech in September, argued that surveillance programs have been sensationalized by the media: "And so what's hyped up in a lot of the reporting is that we're listening to your phone calls. We're reading your emails. That's just not true." The implication is that Americans would be more supportive if reporting had been more accurate. But my poll numbers suggest otherwise. For example, knowing what metadata actually includes had no statistically significant relationship with NSA favorability. For many of our indicators of knowledge, an accurate understanding corresponded to a large decrease in NSA favorability. Consider:

• Among those who could correctly identify the DNI, 53 percent had an unfavorable view of NSA, compared to 33 percent for those who could not identify the DNI.
• Among those who said (correctly) that the NSA does not capture or kill terrorists, 64 percent had an unfavorable view of NSA, compared to 35 percent for those answered this question incorrectly.
• Among those who correctly responded that NSA does not interrogate detainees, 60 percent had unfavorable views of NSA, compared to 42 percent for those who answered this question incorrectly.

What’s going on here? I don’t think that greater knowledge is causing greater opposition to NSA. Rather, I think these correlations more likely reflect the existence of a group I will call knowledgeable skeptics—people who have serious questions and reservations about NSA activity for a host of reasons. These knowledgeable skeptics fall roughly into two camps. First, there are undoubtedly staunch civil liberties advocates who are naturally opposed to much of NSA’s activities, who pay close attention to the news, and whose attention to recent surveillance stories has increased their knowledge. For this contingent, increased knowledge is more a result than a cause of opposition to NSA surveillance. How big is this group? Gallup has found that a growing number of Americans---71 percent as of 2011, compared to about 50 percent following the September 11^th attack---believe the U.S. government should not take steps to prevent additional terror attacks if those steps would violate civil liberties. But a Gallup poll in June also found that only 4 percent of Americans listed civil liberties as their top concern for the nation’s future. This would suggest that the preternaturally anti-NSA crowd is probably a pretty small bunch. That brings us to the second group of knowledgeable skeptics—those who increasingly question the tradeoff between privacy and security and need to be convinced that the tradeoff is worth it. My poll showed, not surprisingly, that Americans with higher threat perceptions were much more likely to have favorable opinions of NSA. It also showed that Americans will give their government more leeway if they can be convinced counterterrorism tools are effective. For example, 64 percent of respondents would allow the U.S. government to assassinate terrorists if it was necessary to combat terrorism. Thirty-one percent were even willing to let their government assassinate foreign leaders for the same purpose (even though such assassinations are illegal). “Necessary to combat terrorism” is the key phrase here. If we had asked whether Americans would let their government assassinate foreign leaders and had not inserted this clause, we almost certainly would have gotten a lower level of support. But that is essentially what intelligence officials are doing with respect to surveillance tools. NSA is telling the American public what it does, but not why. What is currently missing in the NSA debate is a fulsome discussion that links those tools to the greater security they are supposed to provide. I think this probably explains both our counterintuitive results and the declining confidence that the intelligence community is giving accurate information to the President. NSA’s response to the Snowden leaks has focused inordinately on explaining the legality and oversight regime governing surveillance programs and on debunking false impressions about domestic spying. NSA has shown its programs are legal. It has not shown that they are valuable. Other than quick references to Khalid al-Mihdhar, the 9/11 hijacker who was under surveillance by NSA but whom NSA did not place inside the United States, and Najibullah Zazi, the failed 2009 New York City subway bomb plotter, the agency has not given a compelling or consistent account to the knowledgeable skeptic of how its programs are effective, efficient, and prudent in scope.