Tuesday, April 12, 2016

Measurement Without A Cause

One of the arts in social science modeling is to distinguish endogenous variables, the values of which are determined by the model, from the exogenous parameters that get set outside the model.  Typically we write the variables on the left hand side of the equation with the parameters on the right hand side.  Causality runs from right to left, at least when the model is well specified.  So, for example, with a model that was popular at around the time I left graduate school, sunspots cause the business cycle (perhaps) but the business cycle does not cause sunspots (definitely true).  As I said, this distinction between endogenous and exogenous is something of an art and depends on the nature of the study at hand.  In a model of consumer expenditure, income is often treated as exogenous, which is what we do when we teach intermediate microeconomics.  On the other hand, the model can be readily extended to make income endogenous, via decisions about labor supply, how much to save, and how to hold one's financial portfolio.

Increasingly we seem to have social science analysis performed by pundits who write Op-Ed columns and in the case of the New York Times this happens more when the columnist is a known Conservative than otherwise.  My conjecture on why this happens follows.  Most of the Times readership is Liberal.  It is a challenge to write for an audience of doubters.  One way to address that challenge is to wrapper the argument in a layer of social science analysis, presumably objective and therefore not itself subject to reasonable critique.  The Liberal columnists don't need to provide such a wrapper to get the readership to accept the arguments, so quite often they don't.

But there are are some occupational hazards with this approach.  One stems from a desire to moralize in these pieces, to correct the readers in their misguided views and set them on the straight and narrow path.  Why else would a tried and true Conservative agree to write such a column on a regular basis?   However, it is a mistake for this motivation to find its way into the columns.  As a reader, I don't want to be moralized to.  I'm okay on reading opinion that runs contrary to my own, but please, spare me the moralizing. Several years ago I wrote a post, Taking a Sabbatical from David Brooks, with that as the reason.  I went cold turkey on his columns for quite a long time.  Now I will look at them and decide on a case by case basis whether to read through a piece or not.

A second hazard is to argue one side only and not bring up counterarguments.  Liberal columnists might do likewise, but then they have a different writing task in persuading their readers about their arguments.  Without bringing up counterarguments, the reader can't tell whether the Conservative columnist is aware of them or not and even if they are aware whether they've thought them through.

A third hazard is cherry picking - both on the published research used to support the argument and on the model the author comes up with to make the case.  And here let me return to the exogenous/endogenous issue.  That really needs to be reconsidered when making a persuasive argument to people of a different political persuasion.  Treating what might sensibly be taken as endogenous as if it were exogenous will raise the hackles of readers like me with a reasonably strong social science background.  It looks too much like the author is trying to pull a fast one.  If there is a hidden agenda and that ultimately comes out, the author is doomed.  At that point the audience is permanently lost.  So a better approach is to lay one's cards on the table and then make the best hand from that in clear view of the reader.  If the hand is weak, saying otherwise is not helpful.  Credibility is found by telling it like it is.

With this as background, let's consider the piece by Arthur C. Brooks from the Week in Review called Bipartisanship Isn't For Wimps, After All.  Brooks begins this piece talking about polarization, that it is worse now than it was 20 years ago, and this is happening both at the individual level and the political party level.  For Brooks polarization has inexorably intensified in that time period and he is quite comfortable treating polarization as his exogenous parameter, itself not requiring any explanation.  One consequence of this approach is to argue symmetrically about both the hard right and the hard left, not entertaining at all that it is quite possible for polarization to increase with one endpoint remaining entirely fixed as along as the other endpoint moves further in its own direction.

There actually seems to be a cottage industry of books on this score.  I was previously aware of Mann and Ornstein's It Is Even Worse Than It Looks, having seen Ornstein on the NewsHour discussing some of its findings.  (Maybe that was on Charlie Rose, I don't really remember.)  It now seems that every time I Google a book title and look it up at Amazon.com, that title or something similar shows up in my Facebook feed.  (I wonder how that happens - smirk, smirk.)  In this case I got a promo for a book called The Party Is Over by Mike Lofgren, which is notable to me mainly because Lofgren was a Republican insider, yet his conclusions seem largely the same as those of Mann and Ornstein.  It is one thing for E.J. Dionne to make these sort of arguments.  It is quite different to hear it from other authors who are not of the Liberal persuasion.  The upshot is that the right has moved a lot more to the right.  One might ask why, but Brooks doesn't do that.

I am willing to accept that there are multiple causes for this rightward shift beyond the Reagan Revolution.  One clear cause is the Koch Brothers, whom I first became aware of in this piece from summer 2010, Covert Operations.   While the focus of that piece is how the Koch brothers laid the foundation for the Tea Party, it makes clear that they have been funding substantial think tank operations that favor their anti-government Libertarian views and have been doing so for the preceding 20 years or more.   Fox News is probably a separate distinct cause.  Rush Limbaugh is still another.  Somebody else who pays attention to right wing media can probably supply quite a few more members to this list, if they care to do so.

Instead, let me ask a different sort of question.  Just because the media offers inflammatory stuff, that doesn't mean I will change my point of view.  Indeed, and in spite of what Brooks argues, while I'm Moderate to Liberal I don't think I've drifted leftward much at all.  So if that is not happening, what is actually going on, because even if polarization is endogenous and somewhat one-sided, surely it is happening.  That much of what Brooks reports is real.

Let me offer two different hypotheses that can explain the polarization.  The first I'll call politics-makes-me-nauseated, which is how I feel, increasingly often, when I watch the news.   Why get upset if you don't have to?  This is a recipe for tuning out, which it seems an increasing fraction of the electorate is doing.  If centrists tune out more than those at the extremes, you get polarization of those who are likely to continue to participate.  This particular hypothesis favors a symmetric view of the polarization.

They other hypothesis I'll call politics-as-sports-substitute, which it seems to me is the style of the overheated version of reporting and analysis that is now fairly common today but simply didn't exist when I was a kid and we only had TV via over the air networks.  The networks have figured out that tone matters, as does content.  More viewers would prefer gossipy stuff to real news; the latter is often boring and detailed, while the former appeals to the more prurient interests.  Sex and violence sells, at least for some potential viewers.  This one correlates inversely with education, and is therefore not symmetric with respect to audience.  Fox News has a much larger audience than MSNBC.

From polarization Brooks moves onto contempt.  Readers of Malcolm Gladwell's book Blink will recall that contempt is discussed in the very first chapter, where the work of the psychologist John Gottman is taken up and his ability to predict from quick observation of a couple whether their marriage is in trouble or not. The telltale sign occurs when one of them rolls their eyeballs.  It is a sure giveaway that the relationship is doomed.  Once a level of contempt has been reached, there is no coming back from the dead.  So on the one hand, I think Brooks is right here that if bipartisanship is ever to be restored that there needs to be tolerance for alternative views.  Indeed, if you take a look at my recent post, Might it be possible to restore majority rule in Congress?, which was about getting rid of the Hastert rule and restoring a bipartisan majority in the House, with collegiality restored as the mode of discourse to support that, I am certainly on the side of promoting tolerance as a search to finding where the center is.

Now we are getting closer to the real issue, which is exemplified by the Republican leadership in the Senate and their refusal to hold hearings on the Merrick Garland nomination.  I am a fairly regular reader of Jeffrey Toobin's writing in the New Yorker, and he is clearly contemptuous for how this nomination is being filibustered.  So am I.  I have read The Prince and I believe I can adequately apply Game Theory to analyze a strategic situation.   If there were some clear strategic advantage to be applied from blocking this nomination, I might grudgingly respect the decisions of Messrs.  McConnell and Grassley, even if I otherwise didn't agree with it.  As it is now, none of that is evident.  This seems to be about ego only, nothing more.  McConnell is filibustering because he can.  There is no other reason.

Does McConnell's behavior regarding the Garland nomination deserve contempt as a response?  If so, then Brooks' argument clearly needs some modification.  There may be some behaviors by Conservatives that merit contempt from Liberals, while other behaviors merit a collegial response.   Let's say for now that is true.  How then should a Liberal respond to a Conservative, who is himself not contemptuous of other Conservatives in their behavior that is sufficiently offensive to warrant that sort of response?  For example, while there is now a burgeoning 'Stop Trump' movement among Conservatives, there doesn't seem to be anything analogous regarding a 'Stop McConnell' movement.

Yet I am aware of one Conservative who has expressed his disgust at the McConnell filibuster.  See this open letter to Senators Hatch and Lee written by Jon Mott.  (Mott lives in Utah so it is appropriate that he express his views to his own Senators.)  Mott is a learning technologist, as I was before I retired.  I learned of this piece via my people network from then that remains partially intact in Facebook.  And I knew Mott a little bit back then.  He had an essay from spring 2010 in Educause Review that cites and quotes from a column I had written.  I saw him present on this piece at the Educause Learning Initiative conference around that time and had a brief face to face conversation with him as I was chatting with Gardner Campbell there.  I then had a subsequent email thread with him about an online grade book.  That ultimately went nowhere, but that's because we didn't have our act together on the Illinois end.

Mott was perfectly collegial in all of those interactions.  Indeed, he is a model of the behavior that Brooks would actually like to see.  Yet Mott was able to forcefully critique members of his own party.  Where is Brooks on that?  Nowhere, as far as I can tell.  Instead, he quotes the Dalai Lama.  Under other circumstances, that sort of argument might work.  But in the present circumstance, the obstruction of Congress either must become an object for Conservative pundits to critique or they have lost their Liberal potential audience.

How could it be otherwise?  Do they really expect the following argument to work.  Readers, you and I know that Congress is being unreasonable, but I will lose standing within my own party if I say so, so I'm asking you to be tolerant on this score so that progress might eventually be made when things do settle down, without directly taking on the current leadership now.  Don't ask don't tell was the policy in the military for quite some time, until we were ready for a more realistic approach.  That's where we are now on bipartisanship.  Please see it that way.

Brooks, in fact, doesn't even make this argument.  It is an argument that requires a lot of patience.  But at the moment the electorate seems to be taking its mantra from Marat/Sade - We want our revolution...now.  So Arthur C. Brooks, if you feel lonely as a Conservative columnist at the NY Times, this explains why.

No comments: