The Numbers Game

It's been said that five-fourths of all Americans don't understand fractions.  Well, that is true 53% of the time, if you know that 68% of all statistics are made up on the spot.  Enough with the dad jokes!!!  But I digress...But there is truth in the fact that most Americans don't understand numbers all that well (our nations math scores are often lower than most developed countries) Specifically, statistical data that is often used in the media to try quantify some aspect of society, the economy, or political phenomena.  Added to this issue is that there is a plethora of statistics out there and often are contradictory, or at the very least conflicting.  "Lies, damned lies, and statistics" as Mark Twain remarked (he was not the originator of the quote), holds some value these days.  Numbers, data, statistics are often used to deliberately present a picture that the author wants you to see.  Often hidden from those numbers are other numbers, or shall I say other sets of variables that need to be put into context so that we can see the real picture. Statistics and data are somewhat of different language that the average American doesn't speak.  When you are not fluent, nigh, minimally proficient in a language, competency, or skill, one tends to acquiesce to the experts and accept what they say.  If you are not a car person and you take your car to a mechanic, there is always a feeling that this guy is speaking in a language you don't understand, so you go along a pay the $1,000 for a new fuel-equalizer intake apparatus.  I felt like this the other day when I was reviewing my home and auto insurance.  There were terms and numbers and formulas.  I realized I was at the mercy of this agent and I hope to God, he is trustworthy. 

One of the ways data can be misleading is what I call it the simplification of complex issues, also known as univariate analysis.  It is when one attempts to analyze a set of numbers with only a single variable.  The problem with this is that life is much more messy and complex to limit an explanation to one variable.  But we often allow the media and others to present data in this manner resulting in an inaccurate understanding of the issue, and ultimately leading to really poor policy.  It has to be poor policy because the policy is attempting to address a variable that probably is not the cause of the problem in the first place.  This is what is termed a causation error: attributing cause to a variable, as opposed to a correlation.  The famous example of the cause-correlation issue is ice cream consumption and drowning.  As ice cream consumption increases, the amount of drownings increases as well.  Ah, if we eat less ice cream, not as many people will drown.  Ban the ice cream!!!  Hopefully, you get the point.  There are more variables at play than eating ice cream and drowning.  The media often uses this sort of data analysis when referring to race and gender.  For example, the gender pay gap is one of the most widely used as an example that women are still discriminated against.  Many on the left attempt to reduce almost every issue down to the single variable of race.  In fact, there have been a number of recent articles discussing the outbreak of COVID-19 along racial lines.  Having a single variable as the cause is appealing to us.  It makes it simple to quantify and understand.  It's like my old sociology professor in college used to say, we are "cognitive misers."  The single variable prevents us from having to do the hard work of true analysis.

Along with univariate analysis, is a related concept and that is disproportionality.  Disproportionality is the extent to which something is too big or too small in comparison to something else.  In other words, if the percentage of blue eyed people in America was 10%, then one would expect that nearly all categories of people (types of jobs, death rates, cancer patients, etc) would be roughly 10%.  If, say 30% of all cancer patients were blue eyed, we would say that is disproportional.  Disproportionality is the liberals playground, especially when it is combined with the univariate analysis, which it often is.  Not to say that concerns about disproportional distribution is unwarranted, but it often lacks adequate understanding of the complexities that exist within the issue.  It also has a hidden inference or underlying premise, and that is the belief that all outcomes should be equal.  The equality of outcomes is an important battleground of the left and disproportionality is a sword with which they fight.  So to the left, if the population is 50% female, then 50% of the CEOs, scientists, politicians, etc should be female.  But that is not the case.  In response the Left attempts to fashion policies that will attempt to increase the number of women in these roles.  Same can be said for race and ethnicity.  There are many prestigious universities in which their concern is the disproportionally high number of Asian students they have.  To the Left, the data showing disproportionality is proof that America is racist and sexist.  However, they conveniently use disproportional data only when it services their narrative.  You will not here the left complain that men disproportionally are convicted of crimes.  Better yet, you won't hear the left say that because almost all of those convicted of sexual assault are men, that somehow sexual assault laws are inherently or institutionally sexist.  You don't hear the Left bemoan the fact that most nurses are women.  Again, it is not to say that looking at disproportions is of no utility.  To the contrary, it is very important and should be a part of a larger analysis.  But maybe a more interesting question that needs to be answered is what amount of disproportional outcomes are tolerable in a free society? And, when disproportionate outcomes exist, how do we correct them?

Finally, the last issue often deals with economic data and is two fold.  The first is that there is often an infinite way to present economic data so that one can arrive at the conclusion they wish.  One of the most widely used economic data sets is income distribution or wealth stratification.  However, every economist or statistician can create their own groupings of people:  the top 10%, the top 5%, the middle 20%, the bottom 20%, the bottom 30%, etc.  Then one could add qualifiers such as "after taxes", "gross wages"," percentage of wealth."  All this does is add to the over-saturation of numbers and leaves many with the feeling that they don't really know what is happening.  Essentially, most people will then turn to those who influence their beliefs (Fox News, MSNBC, CNN).  The other issue is when one attempts to make observations of the data and assumes that a variable can be altered without impacting the other variables that resulted in the data.  In other words, you cannot assume that if the tax rate in 2018 was increased by 10%, that the GDP, unemployment rate, and S&P would have yielded the same 2018 results.  But that is precisely what many people do, especially on the left.  So when the New York Times puts out these charts that discuss the how the wealth is distributed in America, they do so with the assumption that all that wealth would have still been created, regardless of what ever policy they suggest to limit the disparity in distribution.  But that is not how it works.  It is much more like how Dr, Brown explains to Marty McFly  alternative realities.  Every action you take has a domino effect. 

Quantifying phenomena as a means to provide an explanation or deepen understanding is a worthy endeavor.  However, it requires people be able to understand the process of quantifying and what it means.  Data is never the problem.  It is the interpretation of the data that needs to be addressed.  Too often, we accept the data we like and dismiss the data we don't.  Next time you are presented with numbers, start asking questions, dig deeper, think harder.  As much as quantifying can help reduce issues to more granular levels, it does not limit the actual complexities that are often the reality.  It is the same reason that there are so many conflicting and competing views on health and nutrition.  The fact is, there is no way to limit the variables that go into health into small enough components to develop an absolute.  So, one day you will hear that a plant-based diet is the most healthy and the next you will hear that a ketogenic diet is the most healthy.  Both will have studies, research and numbers.  And both will have different conclusions.