Political polls far from ‘scientific’
Gordon L. Weil
“Lies, damned lies and statistics.”
That old saying about the levels of lying both reveals and
warns about the excessive use and misuse of numbers.
People love numbers. Rankings
and ratings are the basis of decisions that may determine matters ranging from which
country is a world power to what toaster browns bread best to where a student should
go to college.
These decisions matter, but they may be based on flimsy
data. Because we place much confidence
in numbers, they become the tools or weapons of those seeking to impress or
convince us. Every day, we are flooded
with facts and factoids by supposed experts who keep pumping out statistics.
The reason why numbers seem to be the gateway to the truth is
that science relies heavily on numerical data.
A scientific law can be verified all the time by data and observation. People who disagree on almost everything can
accept scientific proof of a law of nature.
Sir Isaac Newton discovered the law of gravity, and its existence
can be endlessly proved by observation and measurement. It was not merely his opinion, wrapped in
numbers.
Science distinguishes laws from theory. Theory can be tested against alternative
explanations. Many numbers today would give
us the same impression of certainty that we get from a scientific law. In fact, they are so far from the real truth
that they wouldn’t even qualify as theory.
Take the Summer Olympics, soon to occur in Paris, which will
produce numbers that are widely accepted as the truth. Athletes will be proclaimed as winners, the
best in the world. Countries will engage
in a race for the most medals, which somehow will indicate their superior
standing among the nations of the world.
Russia seeks to pile up medals in the belief that winning
more medals is an indicator of its superiority.
So, it cheats by doping its
athletes. The late, unlamented East
Germany used extensive doping to boost its international reputation.
Some events, like gymnastics and artistic swimming, are
judged not scored, but awarded numerical rankings. Judges rate break dancing, a newly added
athletic event. Baseball, which produces
a score, has long been excluded.
Athletic judges may make personal judgments, yet they
contribute to the same medal count as timed races. In fact, most ratings and rankings are based
at least in part on the opinions and values of the raters and judges and not on
standards required by good science.
Among the most popular numbers are the college rankings that
help young students decide where to pursue their education. The numbers are unreliable. The components may change, undermining
long-term comparisons. Information provided by colleges to raters may
be falsified, fudged or incomplete. Yet
the rankings are revered.
When they hand out the Academy Award Oscars, do they truly
go to the best picture or best performer?
The current ranked choice voting could produce a distorted result
reflecting popularity and prejudice more than quality. Some big mistakes have been made in the past.
The best picture could be everybody’s second choice.
When it comes to ratings, from military power to movies, always
beware of the word “best.”
Probably the subject where we most willingly follow questionable
numbers too closely is public opinion polling about elections. Numbers are published with so-called “margins
of errors” that give the data the appearance of science. Yet we know little or nothing about who
participated or whether questions were fair. What was the bias of the pollster?
And a survey is only a snapshot; the world can change. Six months ago, few responses would have been
influenced by events in the Middle East, but the Israel-Gaza situation is now a
major issue. How and how much will it
influence voter preferences about six months from now? And many other key issues may look different by
then.
The message that emerges from our excessive reliance on
numbers of dubious validity is that they should not be taken at face value.
None of the daily flood of data that drowns us is truly scientific, no matter
any claim. Still, it may be as good as
we can get it.
Perhaps the most useful information we derive from rankings
and ratings is whether they change over time and, if so, by how much. Biden’s
or Trump’s ratings moving by a point or two over a few days are probably
meaningless. If a trend continues long
enough, a voter may begin to understand who is gaining or losing.
If we are not skeptical of statistics, we risk accepting
politics as if it were nothing more than an athletic competition. Poll numbers, like other statistics, are not
perfect, and they are often overanalyzed.
The biggest risk is that another old saying may apply:
“Figures don’t lie, but liars figure.”
