Sunday, June 16, 2013

NSA and Father's Day

Happy Father's Day to all those in the relevant group!

I saw my father yesterday, and we, like much of the country, spent some time talking about the NSA leaks and Snowden.   My father asked how I felt about it, and I answered in a way only a daughter who's been debating her father for decades could answer:  You already know how I feel about it Dad, we debated this years ago when Bush was President.  He was testing me. Nothing makes my Dad happier than knowing he raised kids who keep their opinions consistent regardless of who's in power.

At that point my Dad mentioned that he had seen a survey that showed that Democrats and Republicans have switched places when it comes to supporting programs like this.  Under Bush, Republicans supported NSA surveillance programs, now the don't.  Vice versa for the Democrats.  

I didn't have a chance until today to look up the survey my Dad was talking about, and I found a good breakdown at here.

There are actually 3 different polls cited:  one from 2002, one from 2006, and one from just recently.  The numbers do, in fact, flip (and 2006 is more dramatic than either of the other two years).  Eugene Volokh however, does an interesting take on the numbers, and points out a different spin:
If the 38% of Republicans who said no still say no today, and the 45% who say yes new said yes in 2002, that amounts to 83% (out of the average of 93.5% responding) whose answers were the same. Likewise, if the 41% of Democrats who said yes still say yes today, and the 43% who say no now said no in 2002, that amounts to 84% (out of the average of 94% responding) whose answers were the same. (I oversimplify here by assuming that the same people were surveyed today as before, despite the changing composition of the public overtime; but if you relax that assumption, then the consistency rate might be even higher.)
Those numbers actually sound pretty reasonable to me.  One also has to wonder how many of those 16/17% would actually admit they legitimately changed their minds.  11 years is a long time.  Even if you took the more dramatic 2006 numbers, about 75% of each party maintained their beliefs.

Now obviously it was not very likely that the same exact people were polled, so we don't actually have evidence that any individual changed their mind.  The one thing to keep in mind when you see polls like this talking about Democrat vs Republican attitudes is that the type of person who identifies themselves with either party is changing.  Here are the breakdowns of Dem vs Rep vs Independent for the 3 years listed:

              Dem      Rep      Ind
2002      31         30         30
2006      33         28         30
2012      32         24         38

  Even if these survey had polled the exact same group of people and they all had answered identically, the numbers would have changed based on changing political affiliation (or lack thereof).  Things to ponder.


  1. What the numbers actually show is how many people think things through for themselves (what ever that means) and how many say what they are told to say.

  2. Volokh is a smart guy, and his looking at the numbers from a different angle is helpful.

    I believe there is indeed this partisan switch, that has hypocrisy, or tribalism, behind it. But I think there are other things happening as well, each of which dent the impact of the numbers.

    There is a fear component. People were more worried about terrorists in 2002 and 2006, and thus more willing to say "sure, do whatever you need to do. Get 'em."

    Surveillance is ancient, but this increased electronic sophistication is new. People may not have yet thought about it much in 2002 and 2006, or not realised what the downstream implications were. We focus on present dangers.

    Similarly, people may not be thinking this through much at all. It may have only gradually dawned on them "Oh yeah, what am I going to think when it's my party/the other party holding the microphones?" That could be partisanship, but it could also be duh, not thinking hard for years and just starting now. There may continue to be oscillation as governments change, but it might be dampening already.

    And related to all the above: the surveillance is getting wildly better, and there may be threshold changes as people think more about the implications.

    Subtract those from Volokh's 16 or 25% and it may not be so deeply partisan as advertised. And I write that as one prepared to believe that virtually everything in American politics is driven more by fashion than thought.

  3. I forgot to mention: Gillespie's article tacitly assumes that both parties are equally untrustworthy using secret information. Do we know that to be true? By history, none are wholly trustworthy with power, but some are much worse than others. People's emotional experience of the trustworthiness of their party and the danger of the other guys might be based on mere advertising - or it might be based on something more solid.

    If one side really is worse, it mucks up the experimental data.