Monday, January 12, 2015

Truths too good to be true

From One of Science’s Most Famous Quotes Is False by Michael Specter.
I recently began a Daily Comment about the growing threat of viral epidemics with one of modern medicine’s more famous quotes. “In 1967, William H. Stewart, the Surgeon General, travelled to the White House to deliver one of the most encouraging messages ever spoken by an American public-health official,’’ I wrote. Then I continued with the quote, “ ‘It’s time to close the books on infectious diseases, declare the war against pestilence won, and shift national resources to such chronic problems as cancer and heart disease,’ Stewart said.”

[snip]

I liked Stewart’s quote so much that I had used it before—once, several years ago, in a piece for this magazine about Nathan Wolfe, a scientist who travels the world hunting for deadly viruses. Stewart’s quote also appeared in a piece that I wrote more than twenty years ago, when I worked at the New York Times. It was part of a series on the resurgence of tuberculosis in the United States. I was hardly alone in pointing to the foolishness of the Surgeon General. The quote has been used dozens of times by journalists, scientists, and public-policy officials. That’s easy to understand. One rarely sees a better example of that kind of inept thinking. There was just one problem.

Stewart never said it.
Specter then documents his discovery that this iconic quotation was conjured from thin air. The yeoman's work tracing the supposed quotation was done by Brad Spellberg and Bonnie Taylor-Blake in their paper On the exoneration of Dr. William H. Stewart: debunking an urban legend.
Specter concludes with the observation
Once a fact, an assertion, a quote, or a meme has been launched into cyberspace, it will orbit there forever. Unless somebody shoots it down.

I regret the repeated error, of course. More than that, though, I hope that the next time a journalist reaches for this handy example of myopia among our public-health leaders, he or she latches onto a foolish statement that somebody actually made. Sadly, there are plenty to go around.
I encounter this situation all the time. I come across a quotation of a third party in a book or article that I want to capture in Thingfinder. I start looking for a source. Many times I have to abandon the supposed quotation entirely because I cannot trace a source for it. Other times, the quote has morphed with repetition, becoming pithier over time. Sometimes the quote is a translation from another language which raises both issues of attribution as well as accuracy of translation.

But Specter is touching on a larger and, to me very interesting set of subjects which I have alluded to as cognitive pollution and Gramscian memes.

Specter's example here is one of cognitive pollution - a fact (the quotation) that is widely disseminated and referenced and is universally accepted as true and yet is flat out wrong. William H. Stewart never said it.

Closely related are the Gramscian memes. Assertions based on ideology (belief systems) rather than logic, rationality or empirical evidence and yet so seductively packaged and smoothly presented as to not elicit any skepticism at all.

Most Gramscian memes are advanced by advocates with or without a particular agenda. For example Post-Ferguson and New York, there has been the whole #Blacklivesmatter movement which isn't really a movement at all. It is a rallying cry for many fringe interests to come together in protest. It is dispersed, there is no central leadership, there are no definitive demands for action specific action, merely an inchoate call to action.

But if you are not an ideologue, what do these protests mean? At a simplistic level there is the anger associated with any citizen's death at the hands of agents of the government. A black man killed by police.

But definitions, a context, and measurement all matter. Death is always in some fashion a tragedy to be avoided, but what is really being argued here? The skeptic can turn around the headlines from Ferguson and New York from "Black Man Dies at Hands of Police" to "Criminal Dies While Resisting Arrest" with no loss of accuracy. Why do some gravitate to the first headline while others to the second and why does one elicit more emotional conviction than the second when both are factually true?

A big part of it is obviously confirmation bias shaped by ideology and culture. But if you are a rational empiricist, you want something deeper. You want the Truth.

In that case #Blacklivesmatter is a nice rhetorical cry but what is the underlying truth?

It is irrelevant, in that case, whether there is a disparate impact on mortality by race if that disparate impact arises from reasons other than bias. What the empirical rationalist would look for is not the simple existence of disparate impact but the reasons why there might be a disparate impact.

What you want to know is whether the death rates by race match the rates of violent encounters with the police. If one group is 50% of the population but they commit 75% of the violent crimes and 85% of the violent confrontations with police, then you would expect that the mortality rate for that group from violent encounters with police would be something between 75-85% and not 50%.

In other words, are they killed in proportion to their violent acts, regardless of their representation in the population at large?

That is what this set of analyses suggests in the case of the Ferguson protests, Race and Police Killings: Additional Thoughts by Robert VerBruggen.

Between 100-200 officers are killed in the line of duty each year, with fairly significant variations year-to-year. Likewise, between 400-1,200 citizens are killed each year in confrontations with police (also varying by year but also by report as there is not a standard tracking of such data.) You want to bring both those numbers to Zero.

Structured programs and investments in police training have led to falls in police shootings in many major metropolitan areas. There is much that has been learned that can be scaled nation-wide and that should continue. However, the other side of the coin is larger and needs more effort. How do you stop citizens reacting to police with violence in the first place. That clearly varies by race and that is where much more can be done but it is basically at the frontiers of our knowledge. We know how to improve the care with which officers behave but we have a much harder time determining how to improve the behavior of the public, particularly the criminal portion of the public.

What does this have to do with William Stewart? The fact that for coming on six months we have been working with a proposition, #blacklivesmatter, which is a great rallying device (just as the Stewart quote was a great example of scientific hubris) but it does not align with known reality. If you want to be effective in bringing citizen/police shootings and deaths down to zero, it has a lot less to do with bias and race than it has to do with behaviors, both police and citizen. As long as we are answering the siren song of race as the root cause, we won't actually solve the problem.

No comments:

Post a Comment