Thursday, May 17, 2018


This what you read to get yourself centered and started.

There two big components to the Cambridge Analytica scandal, neither of which is really very well understood.  One is the personality test that Facebook promoted and that is the source of one part, the idea that a personality test is a magic route to reality ad control. 

The other part is the idea of scraping data, extracting from a mass of scores some kind of insight to be used for manipulation.  This common strategy is the source of my doctor’s judgement of how much metformin I should take for my diabetes.  it’s not about my particular and unique body, but rather comes from studies that add up the scores from a lot of people (probably white male college students) and then looking to see who among them had the best outcomes.  It’s currently estimated that there are four different causes of “diabetes two”, not all of which are properly treated the same way.  But the idea that there is a “typical” treatment, derived from averaging, means I’m getting generic advice.  It turns out that what works for one, doesn't work for another.

Consider the enormous data base that is recorded by grocery cashiers and sent to some mysterious entity which compiles it all “for inventory” which is how the store knows to give you coupons and other encouragement to buy things.  Also, consider a store that accepts checks which are instantly investigated online so that if you have enough money to cover the the purchases, the money is quickly removed before anyone else can claim it.  That database now has all your banking information.  Same with ATM’s.  It is not limited by privacy considerations.

It might not seem that food tells anyone anything, but suppose you’re buying Mexican food, or what our culture considers Mexican food (beans and tortillas).  Political preferences can be derived from what you eat.  (Although I’ve never understood why educated white people have a preference for Thai food.  Also, I've never understood why a taste for fish eggs is a sign of elitism (caviar).)

Food, like politics, is part of how we live.  In a mercantile culture how we dress, arrange our housing, and so on are all both merchandizing opportunities and what we call politics, our ideas about what government should do so we can buy more.  Our “material culture” — our buying habits — become stand-ins for our “personalities,” so that someone who sees that you eat a lot of frybread can assume that you’re “Native American” and attribute to you a lot of whatever they think are NA traits, and how you are likely to vote, what you value, what flash-points you might have.

Now let’s spend time on personality tests.  Of course, you already know about Myers-Briggs, a little four-letter system derived from Jung that turns out to be a great pickup gimmick, better than horoscopes.  (I’m an INTP but as a result I avoid singles bars.)  

More seriously, there is the MMPI — the Minnesota Multiphasic Personality Inventory.  The mysterious person who anonymously makes entries in Wikipedia says,  “The Minnesota Multiphasic Personality Inventory (MMPI) is a standardized psychometric test of adult personality and psychopathology. Psychologists and other mental health professionals use various versions of the MMPI to help develop treatment plans; assist with differential diagnosis; help answer legal questions (forensic psychology); screen job candidates during the personnel selection process; or as part of a therapeutic assessment procedure.”

The Unitarian Universalist Association, a "religious" denomination, used this instrument to decide whether persons should aspire to be future ministers.  The director of aspiring ministers at the time I aspired, David Pohl, told me it was faulty but the best thing they had at the time. (1978)  Also, a candidate was supposed to be interviewed by a psychologist, but two of them told me they were NOT religious and had no concept of  what made a good clergyperson.  Their ideas came from pop culture and  were carried into the denominations that are supposed to use intelligence, experience and meaning to figure out what works.

We resort to polls.  All the UU’s were asked to decide whether they are theist, atheist, pantheistic, and so on.  Of course, it’s clear that the person who is asking the questions finds “theism” to be central, which leaves out all the members who do not, finding public service or mystic experience or simply friendship as the center of their self-identification.

Politically, we also use polls to determine how people consider themselves to be affiliated, though since so many fail to vote, it doesn’t work universally to interview people as they come out of the voting location.  And our political parties are much split now, far more than halved by party or in thirds, if you include the independents.  I would like to see a personality test that somehow sorted people according to what makes them vote.  I think we assume they are poor and dark.

Some people say it is not the content of the test, but rather the ability of computers to sort categories “granularly” — into small subtypes according to whatever evidence is in the base.  We’re in favor of granularity.  If you put one kind of toothpaste on the store shelf, you will sell a certain number of tubes.  If you present five kinds of toothpaste (separated by reputation for mintyness, brightness, tartar removal, what pretty girls buy, and cheapest) you will not sell five times as many tubes, but will sell more than if there is only one kind.  We like choosing.

Some political strategists settle for simply suppressing all votes from categories of people who are not likely to vote for their candidate.  The test that will count will be the one that identifies people who hate your candidate and easily lose faith in voting anyway.  This last election favored populism and fearfulness.  That probably means poor, dark people.  But you can’t ask, “are you poor and dark?”  So you ask, “what do you drive?”  If they don’t own a car or drive an old cheap model, they might be poor and dark.

The whole idea of a personality test/culture test/personality test/material-culture test is that we are what we own, what we do when we choose.  It’s not magic — in fact, it’s simply stereotyping, very effective except when what is sorted is people who don’t think or don’t vote, no matter how granular they are.

This is the test used by Cambridge Analytica.  
You will see that a high score is supposed to mean you are happy and will live a long time, based on five qualities.  Openness, Conscientiousness, Extroversion, Agreeableness, and Neuroticism.  These are the qualities that make Michigan State University researchers get along well.  It's their culture.  I'm guessing they're prosperous and white.  There's much more to say.

(I took the online Facebook personality test -- it said I was an INTP.  We were not in a bar.)

No comments: