Yes, we have another problem
We have another problem. This has been showing up increasingly with many people pointing at it and sometimes touching it indirectly. As with all aspects of rapid paradigmatic shifts problems evolve rapidly and are often counterintuitive. They also need to be explained because they are loaded with historical and intellectual baggage.
Statistics are not ethics or morals. A valid statistical array may contain a wealth of information but it may not be true. There are many levels of truth in our world and distinguishing these levels has been a problem with truth since ancient times. Almost everyone with a basic education is familiar with the origins of logic and it’s many ways to fail. All Minoans are liars. I am a Minoan. This is not what I am concerned with although it leads back to it as everything ultimately does.
I’m seeing a growing use of intentional confusion in using statistics to suggest truth in the justification of a questionable assertion or to deny a truth that is not statistical. Statistics are relative quantifications. They say nothing of how things should be.
In a fully technological world we naturally base our knowledge on science. Science is mathematics and is based very heavily on statistical analysis of the testable characteristics of our universe. Those are the physical truths that are the foundation of our existence and, therefore of our social structures, abilities, and potential. This gives us fundamental physical truth.
This has given us, of course, our world of information processing and now Machine Learning or AI in the mass media. ML gathers data with which to build models. This is a statistical process that lives or dies on a number of critical factors the most basic of which is the validity of statistical conclusions used to build models that are then acted on by controlling systems.
Everyone is now spooked, finally, by big data systems modeling people in elections. Yes, Facebook data was gathered and delivered via Cambridge Analytica to Russian social media teams to influence the election in key swing states.This is well knows and was known in 2016. The surprising success of this process is a major cause of accelerating political collapse in the US. It was, however, as much an accident and the result of many contributing factors with normal marketing tools as it was some new and magical Artificial Intelligence system.
To keep this simple we have very new tools working in ways that most people didn’t know existed before a few months ago and probably have only a very tentative knowledge of even now. My point is not the current status of that technology but the attitudes that are now being manipulated in order to claim truth when only a statistical opinion is being presented.
Truth is not simple and is not exclusively statistical. While for decades our challenge has been to become confident with scientific fact by being able to understand the scientific process. That allows us to either evaluate the scientific facts being presented if we have the knowledge and the tools to do that but, more importantly, it gives us the confidence in how those facts were determined and the continuing process to refine and verify that set of facts as other facts evolve and are better explained.
A major problem in human society is the existence of a segment of the population that has difficulty dealing with scientific fact asm truth preferring truths that are based in people and not in process. Our basic assumption has always been that education, specifically free public education would eliminate that problem. That was a naive and we have come to see that segment as tool to be used to achieve authoritarian control by denying science and shared moral and ethical standards. Unfortunately this segment of our population may be large enough to doom us. This is the polarization that is destroying the American population and the populations in regions ruled by archaic religions and ideological authoritarians. This segment of the population in educated societies that are not as susceptible to pure religious mythology are still easily manipulated by partial or invalid interpretations of statistical evidence.
Many people understand well that statistics can be used to prove almost anything. We continuously have problems with scientific research documentation. Initial assumptions from which a research project is built are ideas based in fact but still ideas. As someone invests time and effort in understanding the data needed to answer the question posed by the initial assumption they work to successfully prove or disporve something. Criteria selected and the definition of measurements for determining proof are what require full verification. But successful verification is still not necessarily truth.
This is a well known conundrum also as knowledge is not wisdom. Wisdom is a much bigger truth than a relative statistical fact. That is why this stuff is not easy.
The political groups that have chosen to use that segment of the population most easily manipulated by things other than scientific fact or logical ethical standards to attempt to build a power base are, now, incredibly dangerous. They were always a problem and directly the cause of immense suffering and hardship as their goals are usually heavily oriented to self interest for a particular group. They offer personal power and status for their support using ignorance and fear justified by lies usually in direct defiance of the overall good. And they have no problem in manipulating data, lying, to maintain personal bonds rather to adhere to the criteria of both truth and wisdom.
Issues that are of major concern are the use of large scale statistical data that do not show clear results. Large data analysis only now is becoming analyzable because before the use of very large intelligent systems too much data was simply noise. Now we can find patterns but these are still dependent on human assumptions or programmed assumptions (algorithms) inherently corrupted to some greater or lesser degree. Knowing that degree is critical and we are not that good yet but are getting better.
Do restrictions and legal controls on gun ownership reduce violent deaths in a society? Yes, but there are many factors that may make that more true or less true for any specific area or region. Simply the inconsistency of the data collected and active restrictions on gathering relevant data can make that falsifiable. We have seen that in the US for years. From the position of cost versus benefit to human society what is the benefit of allowing open ownership of sophisticated weapons? There are none that can be identified statistically. It can be shown that some areas that may allow open ownership have no clear statistical problems with violent death. Is that relevant? I don’t see how. Does the relatively small numbers in comparison to the total society make this irrelevant? That is a clear example of non statistical truth at a higher level. Murder is morally wrong and the number murdered only make the crime more heinous. That only 17 students were murdered makes no difference to the larger moral truth.
As our students are now showing us, one death destroys the argument when there is nothing else to be gained. The logic of close regulation, licensing, right to removal, and insurance requirements have no effect on any actual legitimate use even as recreational devices.
I’ve used one example but there are a growing range. Public healthcare, free education, the right to food and shelter and, soon, the need for Universal Basic Income must be understand with statistical facts under a higher and stronger truth.
Attempting to confuse with statistics is directly related to the attempt to question Truth with claims of counter positions. We are moving to larger truths that need to be clearly identified so that they can be shared and used as ultimate standards for the good of all.