The Outstater
Bad Statistics, Worse Policies
“There are lies, damned lies and statistics,” — Mark Twain
WE TOUCHED ON THIS the other day, correlation but not causation, the inability to legitimately deduce a cause-and-effect relationship between two variables solely on the basis of an observed association. What if the number of Indiana hate crime complaints was not the same thing as the number of hate crimes, we asked.
There’s more to it. indeed, an entire industry is built on the confusion of correlation and causation. You can see it at work whenever your city councilman discovers a statistical difference between groups and then asks for money to make things fair — in arrests, in housing, in grades, in healthcare and so forth.
Such a study was central to the winning argument in Brown vs. Board of Education. But we now know that black children don’t learn either better or worse simply because they are in a classroom with white children. Similarly flawed studies fill the feminist literature — you know, the pay disparity thing. But when researchers adjust for child-rearing and other life choices, women are paid equally with men. Ditto for claims that women are safer drivers than men, but only if you ignore miles driven.
Housing is more complicated, but the neighborhoods of certain groups, even with low incomes and regardless of skin pigment, have relatively less crime. That, of course, is a huge determinant of housing value and availability. Redlining and white flight have had more to do with public safety than with racial bias. Nobody seems to want to look into that.
The same with education. Thomas Sowell has famously noted that the children of Vietnamese families that spoke little if any English somehow excelled in our schools. Statistics didn’t help us predict these outcomes. How about inundating sea levels? Glacial melting? Polar Bear extinction? Job estimates of the latest economic-development project? And need we point out that arbitrary statistical quotas in the military, higher education and government procurement have proved counterproductive to the point of ruin?
In 2021, statistics were quoted that black infants died almost twice as often when their physician was white. “Study Finds Black Babies Are More Likely to Survive When Their Doctors Are Black,” read the Scripps headline. The Indianapolis Star piled on with an investigative series, “Joy in Jeopardy,” on infant mortality generally and black births particularly.
“Babies are dying. That’s not a political statement. That’s what’s happening, and it’s unacceptable,” the study’s author told the World Economic Forum.
Think about that: It has been widely accepted in the media for the past three years that there is a sizable number of white physicians who would purposely cause the death or at least disregard the health of babies because of skin color. Profession-wide “malicious bias,” the research said. Despicable. Something has to be done — prosecute white doctors, adjust standards at medical schools, award healthcare reparations, etc.
But wait. This week in an obscure medical journal it is reported that researchers now say that the original study was in error. It omitted birth weight, a key indicator of neonatal mortality. What the data reflected was seriously ill infants with guarded prognoses being referred to specialists, of which the great number still are white. So never mind. The black babies were not being ignored but being given extraordinary treatment.
Do you see a pattern in all of this? So did Sir John Cowperthwaite, a British civil servant skeptical of statistics and data collection.
Cowperthwaite, the financial secretary of Hong Kong, refused to compile GDP statistics, arguing that such data was useless in managing an economy and would lead to official meddling. He was asked what the key thing that poor countries could do to improve their growth. Abolish the office of national statistics, he replied.
Instead, he used the principles of Adam Smith and free-market logic. Within a generation, the citizens of Hong Kong went from eating bark off of trees to being among the richest per-capita citizens in the world.
Too much focus on quantitative metrics, Cowperthwaite had learned, distorts the true nature of economic activities and dampen genuine progress. Unquestioned statistics tempt the use of science as a cover for political agendas and allow presumptions to run wild.
Clearly, statistics are dangerous things. They cannot be trusted in the hands of dullards in the bureaucracy and the media. — tcl
Comments...