Public More Wary of NSA Surveillance Than Pundits Claim


Based on a bevy of polls conducted in the wake of revelations that the NSA surveiled millions of ordinary Americans' private communications, many have prematurely concluded public support or opposition to the government surveillance program (for instance herehere, and here). These polls are insufficient gauges for what Americans actually think for several reasons. First, slight wording differences result in majority support or opposition of the program as described in each particular survey question, as I've written about here. Second, the full extent of these government programs is not yet fully known; fully 76 percent of Americans think that we'll find out the programs are "even bigger and more widespread than we know even now." Third, most Americans are not even fully aware of the revealed information and its implications—according to a Time poll only 24 percent of Americans say they've been closely following the reports of the large-scale government surveillance program called PRISM.

The public's view of the information leak and revelations about these programs is complicated, as Americans strike a delicate balance between security and privacy. For instance, a Time poll found that 53 percent of Americans think the "government should prosecute government officials and others who leak classified materials that might damage security efforts," but 54 percent thought that Edward Snowden, the person who leaked the information about the secret program, "did a good thing in informing the American public." This is likely because only 30 percent, according to a CBS/New York Times poll, think these leaks will weaken U.S. security.

Examining the different poll wordings can still offer value, demonstrating how people's opinions change when they learn different details of the program. For instance, the public distinguishes between tracking ordinary Americans not suspected of any wrongdoing and collecting records of those suspected of terrorist activity. Pew/Washington Post found 56 percent thought it was acceptable for the NSA to get "secret court orders to track telephone call records of millions of Americans in an effort to investigate terrorism." However, a CBS/NYTimes poll distinguished between tracking phone records of ordinary Americans and those suspected of terrorist activity. In contrast to Pew, CBS/NYtimes found 58 percent disapprove of "federal government agencies collecting phone records of ordinary Americans" but 75 percent approve of tracking "phone records of Americans that the government suspects of terrorist activity." Americans continue to reveal their preference for targeted surveillance when 73 percent told a Rasmussen poll that the "government should be required to show a judge the reason for needing to monitor calls of any specific Americans" and 64 percent said "it is better to collect phone records only of people suspected of having terrorist connections."

Survey data also suggests Americans distinguish between government tracking phone records and government monitoring the content of online activities. Although polls have found public support for tracking phone records to investigate terrorism, most Americans draw the line at government monitoring the content of Internet activity, such as emails and chats. For instance, Pew found 52 percent think the government should not be able to "monitor everyone's email and other online activities." Likewise, when Gallup describing the government program as collecting phone records and Internet communications, 53 percent disapproved.

Surveys that assume away potential misuses and abuses of the data not surprisingly find greater support for government surveillance programs. For instance, A CNN/ORC poll, found 66 percent thought the Obama administration was "right" in gathering and analyzing data on Internet activities "involving people in other countries," while assuring respondents that the "government reportedly does not target Internet usage by US citizens and if such data is collected it is kept under strict controls." The validity of this later assertion, however, is actually at the crux of the debate for those critical of the surveillance program. In fact, according to the same CNN poll, nearly two-thirds believe the US government has collected and stored data about their personal phone and Internet activities. Moreover, Rasmussen found that 57 percent thought it was likely that government agencies would use the data collected to "harass political opponents." The fact that the public's reported support for the program jumps when survey-wording guarantees the collected data will not be abused suggests that part of the reason the public is wary of the program is the very potential for abuse. The public does not desire privacy for just privacy's sake, rather the public fears loss of privacy because of the potential for misuse or abuse. Questions that assume away this possibility are entirely unenlightening.

In sum, these data suggest the public is wary of untargeted government surveillance of ordinary Americans, especially without a warrant. They are more tolerant of government tracking phone records; however, many draw the line at government monitoring the content of ordinary Americans' Internet activity.

A version of this post also appeared on Cato at Liberty

NEXT: NYPD Faces Lawsuit Over Massive Surveillance of Muslims by Demographics Unit

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of Reason.com or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. “Second, the full extent of these government programs is not yet fully known;”

    This- Data mining and meta-data can be defined as phrase searches, text conversion, or any other chunk of content in an actual e-mail or phone call.

    Typical examples of meta-data are things like phone numbers and length of call, but those are just typical. We have zero knowledge of what data the government is actually using in their mining process.

  2. Related: NSA director says ‘over 50’ terror plots foiled due to surveillance, won’t talk about them because the terrorists can’t know how they foiled the plots.

    “If we give all those out, we give all the secrets of how we’re tracking down the terrorists as a community,” Alexander said. “And we can’t do that.”

    You know, just trust us on this one.

    1. NSA data mining has prevented over 10,000 tiger attacks since 9/11/01.

      1. I have this rock for sale

      2. But how many extraterrestrial attacks has it prevented?

        1. All of them – Duh!

    2. Bullshit. Did any of these plots result in an arrest? If so, then how exactly did the NSA avoid telling the FBI and the Court how they go the initial information that lead to the investigation? Is the NSA director admitting the FBI and DOJ are lying to courts?

      And if none of these foiled plots lead to an arrest, how exactly where they “foiled”? Is the NSA allowing terrorists to roam free in the name of keeping secrets?

      The bottom line is, if there were 50 foiled plots, they could release the details of at least a few of them without doing any damage to future operations. And moreover, the release of the details of a no kidding terrorist plot would cut the knees out from under the program’s critics and go a long ways towards convincing the public the program is necessary. No way in hell would they refuse to release such details if such details actually existed.

      1. We need Penn & Teller to start a TV show on the operation of the Federal Government. The could reuse the title they already have.

      2. the release of the details of a no kidding terrorist plot would cut the knees out from under the program’s critics

        No it wouldnt. At least not this critic.

        Ends dont justify means. Ever.

        1. You miss the point. For most it would. If they came up with a bunch of plots they had foiled, the criticism of this program would die. You would be a voice in the wilderness.

  3. Which brings us to Snowden’s release strategy. Right on the eve of the G-8 summit, he released the UK/NSA-assisted hacking into the diplomat’s phones, the setting up of fake internet cafes so they could keylog the credentials, etc.

    I heard one security analyst saying Snowden had enough stuff to respond to events as they occurred. Beyond that, I wonder how he is structuring the releases so that any responses to them would be refuted in the next wave.

    Also, considering he’s said that he has some stuff that’s too damaging to report, is it possible he downloaded the whole fucking NSA spook database and is now cherry picking for maximum procedural impact?

    1. If he’s done that, I suspect the PRC will force him to turn over that data as a condition of him staying in Hong Kong.

      1. I suspect the PRC already did a long debriefing. If he wasn’t working for the Chinese before, he most certainly is now even if his payment for doing so is not being stuck in a Chinese gulag.

        1. Unless he doesn’t have the data with him and worked out some other deal. Just embarrassing the U.S. does give some value to them.

          I doubt seriously he did all of this as a PRC spy–doesn’t make much sense. But now that he’s there, I can’t imagine they aren’t imposing some conditions.

          1. I doubt he did this as a PRC spy. I am just saying once you are in China, you are pretty much going to do what the PRC asks you to or else.

    2. is it possible he downloaded the whole fucking NSA spook database and is now cherry picking for maximum procedural impact?

      It shouldn’t be. But I wouldn’t be surprised if it was and he did. Don’t forget, Bradley Manning had access to huge amounts of information that he had no reason to know and should have never been allowed access to. I am skeptical of claims that the NSA is somehow any better and would never allow a well placed contractor to have access to everything. These people are just not very competent. We have had too many leaks and too many undetected moles over the years to think that they are.

      1. That’s not quite the same:

        The NSA has its own version of Linux and was spurred the development of some of the security features in OpenVMS (RIP sniff! 🙁 ). The main focus of their efforts was in the field of access control.

        I would be very surprised if the NSA was fucking it up.

        Booze Hamilton, OTOH….

        1. But if this guy was an IT guy visa vie an analyst, who knows what he had access to. You can’t firewall everyone. Someone has to maintain the system. And system administrators are generally the worst about security. It would not be surprising at all, in fact it would be typical, if the various system admins working for Booze Hamilton shared passwords and gave each other access to the whole system.

        2. They probably do the same thing most people do when deploying SELinux.

          Turn off SELinux.

          (it has it’s place, but it sure can be a PITA)

    3. I hope he does have a shit ton of data on the program. How amazing would it be if he released the files that the NSA was keeping on specific people. For instance, he could wait long enough to see who is calling him a nutjob or who all is saying that this is overblown and they aren’t collecting data on normal americans… Then release the data on them that the NSA has.

      It’s one thing to bury your head in the sand when it’s something that is potentially, maybe happening to someone else. It’s an entirely other thing when it’s definitely happening to you. Besides, I’d love to see DiFi’s file.

      1. My guess is the NSA has surprisingly few files on specific people. Collecting this information and making use of it are two different things. Collecting is easy, making use is really hard. So, I bet there is a lot more collecting than using going on.

        1. I assume they have some sort of search mechanism that segregates communications that involve certain key words or involve suspects. Then they can narrow from there.

          I’d be surprised if they don’t have individualized data, and lots of it. And, given the latest IRS scandal, I fear that such information just might be misused.

          1. I assume they have some sort of search mechanism that segregates communications that involve certain key words or involve suspects.

            Key words are useless. There is no key word that isn’t used billions of times in innocent context. And as far as specific suspects, sure. But how many of those were there? My guess is few if any. If there were, they would be leaking the details of all of these plots they helped foiled instead of saying “trust us we stopped all of these plots.”

            1. Google does its thing entirely based on key words.

              1. Google, Bing, Yahoo, and on and on.

                It is amazing how few keywords are required to find exactly what you are looking for.

                1. But Kinnath, you know what you are looking for. That is the point. The NSA doesn’t know what it is looking for. Sure, if t hey know the person sending it or specific information like that. But if they have that, they can get a warrant.

                  1. The NSA doesn’t know what it is looking for.

                    My only hope is that they are idiots and don’t know what they are looking for.

                    The starting point is to look at phone numbers (or email addresses). You build networks of phones that call phones that call phones that call phones and so on. That’s why the metadata is so fucking important, and the government assertions that they don’t listen to phone calls are meaningless bullshit, because the network analysis is crucial to tracking people down.

                    Once they have networks of who connects to whom, then listening for names (suspects) or places (targets) it becomes really easy to narrow down which conversations require in depth analysis. It doesn’t matter if aliases or code names are used. Patterns are patterns, and they show up to lexical analysis.

              2. Sure. But Google knows what it is looking for. It knows that it is looking for a particular kind of site. And it also is told to put some sites above others based on what the site has paid google.

                The NSA in contrast is looking through billions of separate communications. It is much harder to figure out which one of those hits are more likely to be relevant.

                1. I don’t think it’s technologically unfeasible at all. For example, spam filters are remarkably good at recognizing when a given email is spam without preventing Aunt Nancy’s cherry pie recipe from getting through.

                  1. For example, spam filters are remarkably good at recognizing when a given email is spam

                    Sure they are. And the data set that is all of the emails you receive is how many? A few hundred maybe? In contrast the data set that is all of the emails sent within or to the United States is 144 billion per day.


                    I don’t care how great your search engine is. You are not finding the magic email in that data set without knowing where to look.

                2. By the way, what makes Prism so fucking dangerous is that it really is hard work to find a terrorist that is hiding in plain sight in the US.

                  Finding anyone and everyone that ever had a phone call with Edward Snowden is pretty fucking easy in comparison.

                  1. Finding anyone and everyone that ever had a phone call with Edward Snowden is pretty fucking easy in comparison.

                    Exactly Kinnath. It is useless for stopping terrorism. But it is money if your goal is to get blackmail material on someone. It’s only real value lies in its misuse. Using the system properly is a complete waste of time and money.

                    1. I would say it is nearly useless for stopping terrorism. If conventional intelligence work gives you a suspect, then you could do to that suspect what you could do to Edward Snowden.

                      But sifting through random communications will produce a thousand false hits (or maybe a million false hits) for every useful piece of information that you gather.

                      In general, the practice is counter productive.

                    2. But it is money if your goal is to get blackmail material on someone.

                      See, I agree with you on this. I’m not trying to say that the NSA has a flawless system designed to ferret out terrorists. I think it’s more that they just don’t really care to, not that they couldn’t.

                      Fearmongering is what got them the power they have and it’s the only thing that will let them keep the power. It’s much harder to fearmonger when there is no real threat. It would not surprise me in the least to find out that they were well aware of what the Tsarnaevs were up to but did nothing because it wasn’t a threat to their interests. A few lives lost very publicly goes a long way towards justifying the surveillance state. Now, if they detected a plot to blow up Air Force One, they’d act on that info. A plot to blow you or I up… Not so much.

            2. Well, there are some sophisticated algorithms used just in commercial searches that could be employed, even to root out code terms. I’d be shocked if they aren’t using such things.

              There was a NOVA on the NSA, I think. Criminal episode?

              1. Yes. There are sophisticated algorithms. But the usefulness of them is limited. The bottom line is, you can’t find something unless you know what you are looking for. And if you know what you are looking for, you have already found it.

                There is a logical limit to how well all of this stuff can work. They would like you to think that there is this magical formula that will allow them to sort through every phone call and email made in America and find the one or two dozen out of trillions that actually are valuable. And that is just not true. The extent to which they find anything is because they had some kind of outside information that told them what to look for.

                1. I agree that such things are limited and that we can’t rely on remote monitoring for everything, but it’s not totally useless, either. And it’ll get less useless the more powerful computing gets.

              2. Well, there are some sophisticated algorithms used just in commercial searches that could be employed

                The NSA are the largest employer of mathematicians in the country. And one of the largest employers of computer scientists as well. Gee, what do you suppose they’re doing?

                1. And one of the largest employers of computer scientists as well. Gee, what do you suppose they’re doing?

                  Making a lot of money building systems that don’t really solve the problem.

            3. Key words are useless.

              No they aren’t, John. A keyword is useless. Just because I sent a text to my buddy containing the word bomb doesn’t mean I have any intent to build or use a bomb. However, that, combined with other keywords, run through analytics can create a much shorter list that could be given greater scrutiny.

              Context does matter, and even ubiquitous programs like browsers implement contextual analytics. The spell check on Chrome will tell you if you used the wrong “its” or if you should use “whom” instead of “who”. Google’s search analytics are also creepily good at contextual searches.

              Remember back in the early days of search engines? You’d type some keywords and then scroll through pages of results because every page that contained those words would come up. Now when you search those same keywords, it analyzes the words and comes up with some context. How many times have you searched for something that wasn’t quite exactly the thing you were looking for, but Google figured it out for you and gave you proper results.

              If Google can do this with a search engine that has to look at the entirety of the internet, don’t you think the NSA can do this with a much narrower focus and a more limited data set?

              1. However, that, combined with other keywords, run through analytics can create a much shorter list that could be given greater scrutiny.

                Sure. And when you are talking about the entire universe of texts, phone calls and emails within the US, the “shorter list” means millions or perhaps billions instead of trillions. The amount of information is so large that even something that could effectively limit it to 1/1000 of the original data set would still leave too much data to sort through.

                Look at it this way, they have been collecting this data for years. Yet, people like the Boston bomber and Hassain went unrecognized. Hassain was sending emails directly too fucking Al Alwalacki and it was the Army IT people not the NSA who noticed. The Boston bombers practically put it Facebook that they planned to blow something up. They traveled back and forth to several areas known for Islamic radicalism and were in communication with lots of known bad people. Yet, the NSA never noticed them.

                You guys are giving this technology way too much credit. Finding a cell phone call that talks about a terror plot is a hell of a lot different than finding that cherry pie recipe from Good Eats last night.

                1. No algorithm can penetrate Alton Brown’s cleverness.

                2. John, are you aware that, not only is the NSA the biggest employer of mathematicians and up there in rank of computer science employees, they also have the larges bank of supercomputers in the world. We’re talking about trillions of calculations per second. Sorting and analyzing that much data is not a problem.

                  Again, going back to Google, their computers process 20 billion searches per day. The NSA’s computers wouldn’t even hiccup at that kind of load.

                  Saying they can’t find what they are looking for because they don’t know what they are looking for is idiotic. They do know what they’re looking for. Patterns. And patterns, analyzed by these massive supercomputers are easy to pick out.

                  1. John, are you aware that, not only is the NSA the biggest employer of mathematicians and up there in rank of computer science employees, they also have the larges bank of supercomputers in the world

                    Sure I am. And are you aware that just because you hire a bunch of people and put them in a room doesn’t mean you can solve the problem?

                    Again, if these programs worked, guys like Husain, the shoe bomber, the underwear bomber and the Boston bombers would have never come near to completing their plots. Yet, they never showed up on the NSA radar. The underwear bomber was so obvious about his intentions, his own father reported him to the US embassy in Amsterdam saying his son was likely to do something dangerous. Yet, the NSA and all of their computer scientists and mathematicians had no fucking idea the guy existed. Can you say boondoggle?

                    Here is the thing, I can point to a good number of cases that, had these systems worked as advertised, should have never occurred yet did. And no one on the other side can point to a single case where these systems have successfully identified a terror plot before it happened. They claim such cases exist but refuse to give any details.

                    Given those facts, can you really still believe this bullshit? Doesn’t it seem a lot more likely that none of this shit works and the entire program is nothing but a case of fraud waste and abuse?

                    1. I clarify my position above, but I agree with you that it’s worthless. Not because of any lack of computing power, but because of lack of caring.

                      worked as advertised

                      That’s the key right there. These systems weren’t put in place to catch terrorists. They were put in place to monitor EVERYTHING. As you said, most of the actual terror plots that have been foiled have been either someone ratting on them or old fashioned police work, not some magic data that the NSA coaxed out of their spying.

                      Again, not because they lack the capability, but because that’s not what they want that data for.

                    2. I think it is because of lack of capability. The problem is not just finding the emails, which I still claim is well neigh impossible, it is also what to do even if you find them. They read Hussain’s emails to Al Alwalacki. And they didn’t do anything. Why? Mostly because the emails consisted of Hussain asking Al Alwalacki whether he should kill the infidel or not. They were reading the guy’s email and it still wasn’t clear what the hell he was going to do if anything.

                      That is the other problem with the whole concept, it assumes that every terrorists or even most or some terrorists are going to in some obvious way say or write what they plan on doing well before they do it. And that is just not how these things work. Even if you know where to look or have this magic algorithm you guys claim exists, chances are very good the communications are not going to tell you shit.

  4. For once, Ekins got the correct amount of alt-text.

  5. No one in this world, so far as I know — and I have researched the records for years, and employed agents to help me — has ever lost money by underestimating the intelligence of the great masses of the plain people. Nor has anyone ever lost public office thereby.

    Mencken (needless to say)

  6. When I was growing up in the 80s, foreign governments doing this kind of shit to their people was the reason we waged the Cold War.

    Seems that we lost.

    1. Not “the” reason, but “a” reason.

      *heads the pedants off at the pass, whew*

    2. No, we just sided with fascists against communists. It’s OK if it’s your own team doing it.

  7. everyone are intered in jerseys can feel free to
    http://mallsjersey.blogspot.com/ to us or go to our shop
    http://modernjerseys.org/ to know about jerseys details.
    1,nfl nike jerseys 1=22$,5=21$
    2,nhl jerseys 1=33.79$,5=32.9$
    3,mlb jerseys 1=17.3$/pcs,5=16.5$
    4,nba jerseys 1=19$,5=18.5$
    5,ncaa jerseys 1=17.3$,5=16.5$
    6,soccer jerseys 21$/pcs
    and other items price here.

Please to post comments

Comments are closed.