Saturday, January 17, 2015

Two interesting developments in on-campus politics reinforce points I've made previously about bias in the modern university:

1.  A political science professor at Marquette University has been suspended for a blog posting criticizing a fellow instructor for allegedly restricting in-class debate on gay marriage.  An obvious parallel can be drawn with the Salaita affair, in which the University of Illinois administration vetoed a job offer to a Native American studies professor based on the virulently anti-Zionist content of his Twitter feed.  Yet a web search for pages containing the names of both professors turns up remarkably few hits, most of them explaining why the cases are actually completely different.  (One distinction offered is that criticizing a fellow professor is far more egregious a breach of civility than, say, preaching hate for all the citizens of an entire country.  I'll leave it to the reader to infer the implications of that argument.)

What this juxtaposition demonstrates is that although academics remain adamant about imposing their political preferences on academia to the greatest extent possible, they're equally adamant in refusing to admit that that is in fact what they intend to do.  Arguments about "civility", consistently mustered against only certain points of view in any political debate, are transparent pretexts for the imposition of political limits on that debate. 

And those limits would be eminently defensible--if only those imposing them were willing to fess up and concede their intentions.  Marquette University, for instance, is officially a private Jesuit institution, and could at one time have been understandably expected to impose a Catholic-friendly atmosphere on its students, even at the expense of stifling "debate" about, say, the empirical falsehood of this or that tenet of Catholic dogma.  If it is instead now a bastion of liberal dogma, then why shouldn't it proudly so declare itself, and impose its moral principles accordingly?

The question answers itself, of course:  if it did so, then many (though certainly not all, and maybe not even most) students would refuse to fork over its hefty tuition, being more interested in a rigorous education undistorted by those particular doctrinal restrictions.  So instead it lies, and pretends that it is a non-partisan champion of free and open intellectual inquiry, taking no position on where it may lead.  In this respect, Marquette is no different from virtually every other university in America--as William F. Buckley Jr. pointed out more than fifty years ago.

Unfortunately, this pretense not only criminally defrauds the students who pay enormous sums to receive what they imagine to be a non-partisan education; it's also responsible for both originating and exacerbating the problem it's designed to cover up.  A university with a clearly stated mission has at least the foundation of a defense against being co-opted by a faction with a conflicting agenda, but a university embracing empty neutrality is defenseless:  between a leadership hamstrung by its obligation to at least appear to make all its decisions impartially, and a group of partisans ready to advance their cause by any means at their disposal, there's simply no doubt which side will win every political or bureaucratic battle.  Indeed, that's no doubt how the erstwhile Jesuits of Marquette University came to be completely dominated by partisan leftists in the first place. 

2.  A recent research paper has taken the highly unusual step of arguing that (per its title), "Political Diversity Will Improve Social Psychological Science".  It claims that the uniformity of political opinion in the field of social psychology introduces biases into its research, and that the solution is to increase the diversity of political opinion among researchers in the field.

The obvious question to ask is why they stop at the political biases of social psychology researchers, omitting, say, the research assistants who collate survey results, the subjects on whom the research studies are performed, the employees of the printer companies whose products print the surveys used in the studies, and so on.  Mightn't their political biases affect research results as well?

Of course, if your research methodology allows the political bias of your RAs, subjects or printer manufacturing workers to bias your research results, then there's a serious problem with your research methodology, and the solution is to fix that problem, not attempt to root out imperfect balances in political opinion wherever they might conceivably interact with your research.  But couldn't the same thing be said of the researchers themselves?  If the quality of their research depends on their collective political balance, then how can they ever even conceivably do good research, given that they will inevitably be collectively biased in some direction or other (say, in the direction of increasing government grants for social psychology research, perhaps)?  And if the research is subject to political bias, then what other kinds of bias might also be seeping into research results?  Racial bias?  Gender bias?  Religious or cultural bias?

The problem of bias--political and otherwise--in experimental results is hardly a new one, and it's actually rather shocking that social psychologists are only now beginning to discuss grappling with it.  And the fact that introducing political diversity into the field is considered a plausible and reasonable approach to the problem is a demonstration of just how pitifully naïve and confused the social sciences are in dealing with it.  Generating unbiased experimental results--or even getting a reasonable handle on the possible biases in one's experimental results--is extraordinarily hard, and that's one reason why scientists (supposedly) undergo such extensive and rigorous training, and why their work is (supposedly) subjected to such intensive peer scrutiny before being published.

In practice, of course, those standards have long disappeared, and much published research--even in the hard sciences, as my co-blogger is fond of pointing out--is actually transparently shoddy.  So when a social psychologist advocates increasing political diversity in the field as a way of reducing experimental bias, he should be understood to be saying, not "here's a previously-undetected source of subtle bias in our research, and here's a clever way to reduce it", but rather, "we all know that our work is shot through with bias of all kinds, which we frankly can't be bothered to try to mitigate significantly, but this particular type of bias is likely to be both obvious and annoying to the non-scientists who pay our salaries, so perhaps we should at least make some pretense of trying to address it."

2 comments:

LTEC said...

1) Regarding l'affaire Marquette:
A very long time ago an (otherwise) wise man once told me that the extreme-left/feminist control of college campuses was seeing its "last gasp". (Perhaps that would have happened if it weren't for the recent epidemic of rape.)

2) Social science research is crappy, even compared to the low level of other experimental science. But there is one reason to be more concerned about the political bias of social science research than about its other crapitude. The reason is that by and large, the other biases are mainly towards getting interesting results. And the (otherwise) unbiased "interesting" results are usually so stupid and unimportant and obviously irreproducible that only another social scientist could possibly care about them. For example, consider that "famous" result (famous in the world of social science) that people become more racist in the presence of litter. Like most people, I had never heard about it until it came out that the work was not merely the usual crap, but completely fraudulent even by social science standards. (See http://www.nytimes.com/2013/04/28/magazine/diederik-stapels-audacious-academic-fraud.html?pagewanted=all&_r=0)

Politically biased results, on the other hand, have a great deal more of an effect on the world. Whenever some study shows that conservatives are stupid or that republicans are ignorant or that "climate deniers" are conspiracy theorists, the rest of the world takes notice and such results become instant folk knowledge.

By contrast, in most other sciences, even politically unbiased crap can have real importance. When someone declares a new treatment for cancer that winds up being bogus, this is important.

Dan said...

1) Yes, he was (and still is) a wise man, but being relatively new to the world of academia, he naively overestimated its institutional strength and health. He therefore assumed that it would easily shrug off the attempted takeover by political partisans. Of course, now he knows better, and realizes that it's academia as a whole that's seeing its last gasp.

2) I don't think political bias is fundamentally different from "interestingness" bias--in fact, they're really just variants of the same phenomenon: the generation of bogus results that happen to be appealing to a certain target audience, and are therefore likely to be believed by that audience. Many bogus social science results are variations on "your romantic/financial/social problems aren't your fault", or "people like you are generally superior to others", or "you and your friends are absolutely right about the world", and are therefore believed and embraced by the audiences to which they're addressed. One particular variant is, "people who disagree with you politically are scum", and the fact that that particular set of results tends to appeal only to half the population, rather than all of it, isn't all that significant, since it doesn't convince anybody of anything they didn't already ardently believe. Indeed, I would guess that far more people are hurt by bad romantic advice they got from a bogus social science study than by bad political advice they got from a bogus social science study.