This answer may be a bit more philosophical than you are looking for.
Embedded in the foundations of statistics is an element of philosophical pragmatism. This element is perhaps embodied most formally in the concept of the loss function, the idea that the purpose of a statistical estimate is to minimize ones losses from being wrong. As W. Edwards Deming noted, this introduces a subjective element into even frequentist statistics. If people have different purposes, they will have different loss functions. It is also reflected in aphorisms such as George Box's "All models are false, but some are useful" and W. Edwards Deming's much stronger version, "There is no such thing as a fact."
Under philosophical pragmatism, we learn only when our purpose is frustrated. Only when we get slammed into a wall do we have an opportunity to realize we are wrong. So long as we are achieving our goal, we think we know. In this sense, knowledge can be painful, while thinking we know is a much more comfortable and preferable state. Setting aside whether or not this is an appropriate norm for how knowledge should work, it seems a reasonable empirical description of how it in fact does work among people.
Statisticians have tried to avoid this subjective element by fixing loss functions (and p-values) by convention, ensuring everybody uses the same one. But this merely results in models that aren't fit for use and estmates that result in unnecessarily high losses because they aren't appropriate to the situation. Saying it is so doesn't make it so. Fixing conventions does not make statistics objective.
What happens when people's purposes run against each other? This is a difficulty at the heart of philosphical pragmatism. What happens when ones purpose is best achieved by deceiving others? Strict philosophical pragmatism would consider the estimate that minimizes ones own losses the best one, even if it creates losses for everyone else. The calculus is completely amoral. If ones goal is power for its own sake, then deceipt, or inventing myths that result in people aiding ones rise to power, can be a very pragmatic way to achieve it.
W. Edwards Deming's way out of this was first, to focus on purpose. Losses are with respect to a purpose, so if one doesn't know ones purpose one doesn't know ones losses and can't estimate anything. And second, to focus on cooperation. Only if people have a common purpose will they be able to have common loss functions and common estimates, and hence any common knowledge. In a society where people are at cross purposes, common knowledge simply is not possible.
We are seeing, all about us, the consequences of what happens when people operate at radically cross purposes, and when one person seeks to gain by others' losses. There can be no agreed knowledge in such a society. Only in a society where people have common purposes and cooperate on them can there be shared knowledge. Or any ability to learn as a society.
As W. Edwards Deming put it, "Will we learn? We may not. Learning is not compulsory. Survival is not compulsory. Without learning, there is no survival."
------------------------------
Jonathan Siegel
Bayer U.S. Pharmaceuticals
------------------------------
Original Message:
Sent: 01-26-2021 15:11
From: Raoul Burchette
Subject: What are "facts"
I noticed a discussion about the election and probability theory. I do agree that the modeling the election by a simple probabilistic model was probably not reasonable and perhaps we don't really have a good handle on how to model this correctly. But there were other claims in the discussion that I find curious. In particular the claims about "facts". I am under the impression that some people seem to have other ideas about what constitutes "facts" and truth, As regards this election, I have no idea about what is true and what is false, I certainly could not offer acceptable testimony in court. I am not a first hand witness, I don't know anything about the election other than I went and voted. The rest I am dependent on others to do. I cannot swear that others did their job correctly, or mostly correctly, neither can I swear that they did anything wrong. I cannot say what the facts are. To believe I am hearing truth requires a long chain of trust in each of the necessary links. These days, some of these links seem a bit tenuous. Conventional news lost its credibility in my mind when I heard them report on political topics, but when I went and read the original documents, or watched the available footage of the reported events, I could justify no such summaries or conclusions as they offered. This happened so often and consistently that I settled it my mind that there was a systematic issue. On the other hand, i suspect if we really knew the truth, we would all be surprised at what we are variously alleging as facts or truth.
I think I have been out of touch for a while, but in which court cases were the various affidavits of election fraud debunked? I heard about the affidavits but I did not hear about the court cases.
------------------------------
Raoul Burchette
Biostatistician
------------------------------