GEORGE TEMPLETON
COMMENTARY
By George TempletonGazette Columnist
Data Wizards
In the Wizard of Oz, Dorothy was not to look behind the curtain, but like Alice in Wonderland,
she became “curiouser and curiouser” as she opened the telescope and
looked backwards, down at her seemingly receding feet. That is what
happens when you look through the wrong end of the telescope, when your
world view insists that up is down.
Benghazi
I
watched the eight hour Benghazi fact finding hearing recently. There,
Republican Chairman Trey Gowdy could be found hiding behind the curtain
of an alleged “all the truth and nothing but the truth, a final
definitive accounting” which he stated was “my perspective”. We were
invited to entertain conspiracies that were “within the realm of
possibility”, but the fact is that we live in an unavoidably uncertain
world. The improbable can occur, and by definition it is not foreseen.
If
the proceeding had anything to do with improving the security of our
overseas embassies it would have invited specialists who could address
the situation at every location. Funding for improvements would have
been discussed and planned for. Trey informed us that the hearing was
not a prosecution, but you would think otherwise if you counted the
number of lawyer congressmen present and the blaming tone of the
proceeding. Trey’s angry outburst exclaiming “Just wait until the next
round”, I’ll get you then, was not about inquiring. It came from his
wounded competitive pride. It was not about fairness, promoting
harmony, or statesmanship, putting the good of the American people above
self-interest. It was about leading questions, “Isn’t it true that …”,
and creating distrust.
Fox
News had already run more than 1000 segments on Benghazi and there had
been eight previous investigations. More time and money was wasted by
tax payer funded political theater. It cost us 4.7 million dollars.
Big Data
Trey
Gowdy’s final assessment suggested that his “more” could be measured,
counted, and added up to create a statistical labyrinth of truth.
Statistics is about the collective, but only the individuals are real.
They are described by the distribution of the data. They are free to do
as they choose.
The
digital world and big data is our latest rage, but the world is more
than numbers. Numbers confuse fact with subjective reality. Things
that are qualitative and complex are mapped into the quantitative and
easy, but numerical data cannot begin to capture the many facets of real
situations. Will surveying millions of people, “just like you”, lead
to understanding and helping you? They say that data can change your
world, but will it do it in ways that you want? We must remember that
data collection requires categories, forms, and rules, and that
officials with a mind-set and world-view will do their design.
The
World Health Organization (WHO) has classified meat as carcinogenic,
like cigarettes. It combined eight-hundred studies to argue that
“probably?” red and processed meat (how much, how often, prepared how?)
increased the chances of colon cancer by eighteen percent (above what?)
in some people, somewhere. The study was a sample and as a consequence
that eighteen percent has a tolerance, a confidence interval that was
not reported. At least the WHO did not plagiarize research like the
intelligently designing creationists do when they reinterpret findings.
What about combining data, sensitivity to initial conditions,
evolutionary behaviors, the branching tree of reality, hidden and
interacting variables, and uncontrolled parameters?
Business wants to be your friend and will give you “points”. It will help you without your even asking.
Claims
have been made that the fountain of youth and cures for chronic
diseases that take a lifetime to develop have been found by simply
surveying available world-wide data at a moment in time. Studies
justify this, claiming that they are “almost statistically
significant”. Findings are “linked together”, implying “relationships”
while disregarding mathematics and doubt about “more of the same”.
It
has been claimed that we should believe contradictory studies. The
easiest solution to this quandary is to disregard all of them until a
consensus of opinion between experts develops. A better solution
requires learning a few things about data, probability, statistics,
science, and experimental design. The public school system has been
deficient in this regard, but leadership industries have found it
necessary, developed complete curriculum, and successfully taught such
things to selected high school graduates.
We
must realize that the man behind the curtain is he who controls the
computer that he uses to micromanage. His flaw is that he does not have
the courage to admit that he does not understand. That makes it
impossible to learn.
In
March of last year I wrote about Nelson Goodman, categories, and
self-referential ambiguity. Ever increasing complexity and a growing
numbers of categories can become a “now you see it, now you don’t” shell
game used to deceive and rob the poor and obscure. Science believes
that quantification is an important thing for understanding. But Nelson
Goodman creeps into that. Things cannot be measured without a precise
definition of what is and what is not an instance of them.
When
the immeasurable is quantified and computerized, it becomes “garbage
in, garbage out”. For example, if you take a pill, you must have the
problem it was taken for. The illness is the pill. That’s the easy way
to see it. An assisted living facility will charge you for it even if
the pill controls things; there has been no instance of it, no action
taken, and no future plan. It is self-referential and circular because
if you stop taking the pill, by definition that makes you well. The
pill is no longer needed, but what was responsible for the pill in the
first place could reoccur. Worse yet, every pill you take can be double
counted, once for their administration and again for their implied
crisis. Conveniently, this can be managed by executives who are
thousands of miles away from families and caregivers.
Attaching
a number to things does not just measure them. It changes them. When
corporate rates numerical points (0.01 out of 2000) they imply authority
and precision that does not exist. That number simplifies, overlooks,
and forever more becomes part of your assessment. So, because a number
and an expense will be attached to your condition, you might abandon a
medication that works.
You
have been taking that pill for 15 years, so long ago that you don’t
remember what started it in the first place. If it’s working, the
doctors continue it, and so it becomes an ambiguity, a self-referential
justification for continuing the prescription. The assisted living
center will charge you more and report it to government surveys
indicting that you have a lifestyle interrupting problem even though it
does not exist.
The
truth is rarely simple and complete. It is not served by top-down
corporate or government power and control. Unfortunately, this is the
sort of thing that goes into statistics where subjectivity and
incompleteness is ignored.
Some
things cannot be put in words because doing so is only an
approximation. There is such a thing as seeing without vision and gut
feeling. They are important. But seeing is not just with the eyes.
Biologically, we all have a blind spot, and the mind fills in for that
flaw. There is such a thing as human reality and the metaphor that
means different things to people in different situations. But the man
behind the curtain will not have it that way. He will not embrace the
unavoidable uncertainty that we have to learn to live with.
Understanding Data
According
to Fox News, the facts do not support man-caused global warming. They
are only computer simulations that should be ignored, but shapes
matter. Bends in a response curve reveal time constants and energy
movement. The sharpness of those changes reveals whether the measured
behaviors are simple or the result of many coupled things. It is
forensic evidence, a signature that goes beyond the magnitude of
temperature rise and the difference in slope between Centigrade and
Fahrenheit degrees.
The
bell curve is another shape that has impacted nearly everyone’s life.
We were graded by it. It is the signature of probability, expected
value, Las Vegas profits, and a random variable. What we have to
consider is that all statistics is an abstraction. The only thing that
is real is the data. Correlation is not causation. Scientists should
be respected because data, measurement, and statistics are central to
their work. They live with the problems of measurement and data every
day.
The
computer has made it possible to acquire and generate massive amounts
of data. Those of us who worked within this situation recognize that a
large portion of their job was to destroy data. It was necessary to
survive. What we wanted was meaning. In science and math there are
only a few simple but powerful ideas that apply to myriad situations.
What we look for is the courage and insight of those providing mountains
of data to interpret it. But that is only the first step.
A
simplification is to fit a mathematical curve, line, or “fit” to the
data. Then, all we have to do is consider the curve, how good the fit
is, and whether it could have happened by chance or if there could be
something that is partly responsible for it. Although these measures
exist they are not communicated to the public because “the layman would
not understand” or maybe because the author has a hidden objective.
The
problem is that all the data points may not fall perfectly on the
line. If we make the equation, the “fitting function”, more complex
with lots of variables, we can make our curve go perfectly through every
data point, but things become extremely complicated, confounding our
original intent that was to simplify so we could see.
The
simple straight line is the thing that high school graduates need to
understand most of all. Fitting functions are an example of art work
that can also use curved lines and equations such as those that are
quadratic or cubic, but they are more likely to become wild and
unstable.
In
mathematics there can be any number of dimensions. Each variable used
in a big data model adds another dimension. We can see up to three with
difficulty. More than that improves the fit, but destroys
simplification and our ability to see and to undertake the follow-on
work that is necessary to understand what is going on.
The Clinton Inquisition
So,
what does all this have to do with the Benghazi hearings? They were
for the man behind the curtain who was growing “curiouser and curiouser”
thinking that he might be able to discover, to fabricate a connection
that would hurt his opponent.
There
is a moral imperative to this, but it is not the touchy, feely,
personal morality of the Seven Deadly Sins. It is a mean spirited
morality, about expediency, duty, and pragmatism that emphasizes winning
above policies that care about people. This is the human reality that
we must weigh and decide.
No comments:
Post a Comment