Being Bayesian About Beliefs

A couple of weeks ago, I posted an OP, What Happens when Trust is Lost, about my initial reaction to the reason that the president chose to kill an Iranian general when he did. I concluded he did it for personal political reasons, not due to imminent threat to US or other lives or interests. The facts and logic behind that OP were these:

1. The president’s has a public track record of making over 15,000 false and misleading statements to the public;
2. That fact-based record constitutes an empirical basis to not trust anything the president says about anything related to his politics;
3. The president claimed he killed the Iranian general due to an imminent grave threat the general posed against Americans or their interests;
4. Despite the president’s claim of self-defense to justify the killing when it happened, it is reasonable to believe that he intentionally timed the killing to divert public attention from damaging information that had just come from information from federal courts related to his impeachment; and
5. Therefore the president lied about the reason for the killing occurring when it did.

Being Bayesian

For political beliefs, being Bayesian means that one changes one’s opinion or prediction and/or their level of confidence in it as new information comes in that would logically support or contradict the opinion or prediction. Social science research strongly suggests that people best grounded in political reality and logic are Bayesian about their political opinions (beliefs) and predictions. When new contradictory information comes to their attention, Bayesians tend to either (1) reduce their confidence in their opinion or belief that something is true or will happen, or (2) completely reverse their opinion. Similarly, awareness of when supporting evidence or logic arises, confidence in the opinion typically increases. 

People are often not Bayesian when cherished political beliefs are challenged by contrary or undermining information or reasoning. Such beliefs are simply impervious to change in the face of contrary evidence and logic. That is not uncommon in politics. It is a common basis for political beliefs and reasoning being irrational to some extent.

New evidence about killing the general

On hearing of the killing of the Iranian general, my initial logic and conclusions (beliefs) were that (i) the killing is was timed to deflect public attention from bad news about impeachment of the president, and (ii) that the general presented no imminent threat at that time. But since the story was new, it was possible that evidence could come out showing that there really was an actual threat. Thus, my level of personal confidence was pretty low, about 55% certain, that the president was lying about why the general was killed when he was killed.

In the days since the killing, evidence has come out that there was no imminent threat from the general. Comments by the president, his defense secretary and secretary of state and others in his administration were vague, contradictory, inconsistent and not accompanied any solid evidence of any real threat. People in congress, including some republicans, were unconvinced that any real threat existed at the time. Administration efforts to buttress the imminent threat argument were absurd.

Trump administration’s best people getting it’s story straight

Being Bayesian in view of the evidence to date, I have had no choice but to revise my confidence that the president lied about why he killed the general from about a 55% level to about a 95% level.