Wednesday, July 1, 2020

And Then They Came For the Statisticians

ORLANDO, Fla. (AP) -- The world's largest statisticians group added to a chorus of criticism this week against the recent hiring of two political appointees at the U.S. Census Bureau.  The American Statistical Association says the appointments earlier this week of Nathan Cogley and Adam Korzeniewski to top posts even though they have little experience at the agency "are in direct conflict with the bureau's mission to ensure proper, accurate, and timely delivery of statistical information to the public." Cogley is a political science professor who wrote a series of opinion pieces against the impeachment of Donald Trump.  Korzeniewski is a former campaign consultant to the pro-Trump YouTube personality known as "Joey Salads."

Mike Schneider
Associated Press
June 26, 2020 


...the suppression of its findings and the murder of those who organized it was nothing less than the obliteration of the capacity for self-analysis.  An authoritarian society, however, that is unable to form an idea of itself, whatever social engineering its leadership may have in mind, is doomed to the blind exercise of state violence.

Karl Schlögel 

     For one 24 hour period, from midnight January 6, 1937 to the following midnight, life in the entire breadth of the Soviet Union, one-sixth of the world, came to a stop.  On that day, after weeks of extensive training, one million enumerators (census takers) with a list of fourteen questions spread out over the entire country.  Their mission was to collect the information that, when analyzed and pieced together, would form a detailed picture of what Soviet society looked like twenty years after the Bolshevik Revolution.  The enumerators knocked on the doors of apartment blocks in Moscow, skied across the Arctic tundra to remote villages, and rode with passengers on the Trans-Siberian Express.  As Karl Schlögel describes in his book Moscow 1937:   

     Whether in the cities or on a river steamer, in a yurt in Kazakhstan or in a hotel in Leningrad... from different districts of the capital, from the capitals of the republics, from Kiev, from Ashbagat, from the Taiga and the Pacific ports, from the newly built suburbs.  The armada of enumerators ...not only penetrated the furthest corners of the Soviet Union; it systematically explored the social landscape...(and) discovered a Soviet Union in miniature.
     The populace was ready.  The census protocol had been highly publicized in workplaces and shops and over state media outlets.  The questions submitted to the Party leadership by the Census Board were edited by Party leader and supreme ruler of the USSR Joseph Stalin personally (he had removed eight and added one) and had been released in advance so citizens knew exactly what to expect when an enumerator showed up on their doorstep.  It was a remarkable achievement, one that required careful planning, tremendous resources and a nationwide mobilization.  What makes it even more astonishing is that, according to Schlögel, historians and demographers with access to the formerly suppressed results--made available in the 1990s after the Soviet Union's collapse--have calculated the margin of error at a low 0.5 to 0.6 percent.  
Information poster

It was Stalin who had ordered the census, and he had a lot riding on its outcome.  It had been eleven years since the last one, a time of rapid industrialization and societal upheaval.  A program of forced agricultural collectivization and two Five-Year Plans had run their course.  Stalin believed that the demographic figures would prove that the Soviet Union had built a vibrant, happy and healthy society.  High growth rates were predicted, outstripping those of rival capitalist countries.  Officially, a population figure of 170-172 million was expected (a number extrapolated from the 1926 census); Stalin was hoping for something closer to 180 million. 

Census data being processed
     They knew they were in trouble almost immediately.  Preliminary results indicated that average growth, while surpassing Germany, England, and France, was exceeded by both the United States and Japan.  Stalin's added question was about religious affiliation; he expected the state's anti-religion policies to reflect a high number of non-believers.  Yet close to 60% of the adult population identified themselves as believers.  Most alarming, the final count would be somewhere around 162 million, 8 million people shy of the official pronouncement.  I.A. Kraval, the Census Bureau Chief, ordered a recount, but only several thousand unenumerated persons were found.  In late January, the provisional results were presented to the Party leadership, and a further report in March confirmed the original figures.  Kraval and his team of demographers and statisticians, fighting for what they knew were their lives,  did their best to excuse the results, but the difference between the final count and the publicly stated number of 170 million was too great to explain away.
     Where were the missing 8 million?  And who was to blame?  The biggest discrepancies between the expected numbers and the actual count were in the regions hardest hit by the devastating 1932-1933 famine caused by Stalin's policy of forced collectivization.   Historians are still debating the actual death toll--estimates range up to 7 million--and also to what extent anyone at the time had an accurate fatality count.  (Ironically Kraval was sending false reports to Stalin, downplaying the numbers.)  In any case, the famine had been officially denied, and anyone who talked about it risked imprisonment and death.  The census data, which reflected the tragic consequences of Stalin's policy, had to be suppressed.  So did those involved with the data collection.  
     The arrests began in March.  No one was safe; from the members of the central Census Bureau to the chiefs of the regional census centers, right down to administrators at the local level.  Statisticians assigned to replace the imprisoned were soon imprisoned themselves. The statisticians and demographers were accused of sabotage and "wrecking", and labeled "Trotskyite-Bukharinite spies" and "enemies of the people".  Many of those imprisoned were eventually executed.  Kraval himself was condemned to death in August and shot.  A new census was planned for 1939.  Schlögel links this to the Great Terror:  

...having destroyed the analytical matrix which disclosed the contours of the nation and its people, and having sacrificed the very instrument that would enable them to interpret these things, the leadership... (was) overcome by a blind flight into terror, an intensification of violence whose excesses would surpass the very disasters that the census had just diagnosed.  As a result, for the catastrophe that followed there were no longer any instruments that might have diagnosed what was to come.     

      Not surprisingly, the new census resulted in a population count of 170.6 million people.  Just what Stalin had ordered.  The question about religious affiliation was eliminated.  There would not be another census until 1959.      

     Suppression of truth and truth-tellers isn't exclusive to the former Soviet Union.  It's not hard to come up with examples of our own.  The data linking smoking to cancer and the cover-up of abuses by clergy in the Catholic Church are two that come to mind.  We know stories of individual whistleblowers coming to bad ends (see Karen Silkwood) and we also know of consumer advocates forcing changes to promote public safety (see Ralph Nader).  Thankfully, we have avenues to access information, mechanisms to uncover truths, protections for those who would speak truth to power and independent media outlets to publicize those truths that would be unthinkable in the Soviet Union under Stalin.  Our system of government is based on checks and balances, and while it may not always work the way we want it to, it does provide a measure of protection against the worst autocratic tendencies of an executive, protections that, again, would have been unthinkable to Stalin.
     Consider the case of our census, currently in progress.  When the Trump administration tried to add a question related to citizenship, it was met with resistance and criticism, not only from progressives, but from statisticians and demographers, who argued it would threaten its integrity, estimating that inclusion of the question could lead to an undercount of 9 million people.  The pretext for adding the question was seen as strictly political; it was a way to depress the count in heavily Democratic areas, which would have consequences for proportional representation in Congress, how many electoral votes a state receives, and how and where some $1.5 trillion in federal money is allocated. A lawsuit challenging the question made its way to the Supreme Court.  In a fractured and complicated 5-4 decision, the court ruled that the justification for adding the question was invalid.  Unable to delay the implementation of the constitutionally mandated decennial census, the administration threw in the towel.  
     However the recent appointments of Cogley, a frequent radio commentator and former head of the department of government, legal studies and philosophy at Tarleton State University and Korzeniewski, who once worked as a consultant for the failed Staten Island congressional run of Joey "Joey Salads" Saladino, a YouTuber famous for racist pranks, and whose primary qualification seems to have been that he once worked in a census bureau field office, to top positions at the U.S. Census Bureau should be cause for alarm.  It certainly is to Kenneth Prewitt, a former Bureau director.  "These are two people ill equipped to actually manage the census," he said in a Politico report:
They're very well equipped to advance political interests, especially those of the Republican Party.  That's their background and their career goals.  It's unprecedented for two political appointees to be added to the bureau in the middle of a census count in the recent history of the Census Bureau.

     What influence will they exert on the counting and reporting of the census figures?  Will they be guided by best statistical and demographic practices?  Do they even know (or care) what those practices are?  Were they put there to do the bidding of the administration that appointed them?  Will the final census be an accurate reflection of America in the year 2020?  And if it isn't, will we even know?

    All this is to say that our community has a very important role to play.  Math helps build the models, math looks at the data, math runs the statistical analysis.  It's our job to call bullshit when we see it...

...and it's our job to give our students not only the tools they need to be discriminating and critical consumers of information, but the confidence to use their voices to speak out when they see data being used to deliberately manipulate and confuse; whether it's related to: 
something as innocuous, like how many people attend an inauguration, 
something important, like a census count, 
something serious, like the death toll from the coronavirus, or 
something that threatens the very existence of life on the planet, like climate change.  
So whether you're in kindergarten teaching kids how to count with one-to one correspondence or in high school teaching AP Stats, or anywhere else above, below, in between or sideways, please keep your shoulder to the wheel.    


No comments:

Post a Comment