Bayesian analysis for beginners

[robertb]

With regards to history, the benefit of a Bayesian approach is that it forces one to "show their work".

[shunyadragon]

Those objective hard data are evidence, yes. You were saying, in reference to Bayesian analysis, that it was "dependent on the assumptions of the information and data used." I was asking for an example of a method of analyzing evidence that does not depend on any such assumptions. Can you explain how an archeologist who has discovered some artifact might construct an assumption-free inference about what the artifact tells us about the past?

Yes, the use of Bayesian analysis is dependent on assumptions, but what I was seriously questioning was the assumptions based on subjective information, such as those used by apologists concerning the Resurrection. An example of the assumptions used for archeological finds are based on previous research results on the distribution and nature of the objects found. Patterns of trade and points of origin can be determined based on the distribution of the objects found at different sites.

[Doug Shaver]

but what I was seriously questioning was the assumptions based on subjective information,

If they're really based on information, any information, then they're not assumptions. They are inferences. If the information is unreliable, you may certainly criticize the inference on that basis. Or, if there is no apparent problem with the information, then you might find some fault in the logic by which the inference is deduced from it.

such as those used by apologists concerning the Resurrection.

I have yet to see an argument for the resurrection that is not either invalid or circular. There is no way for an apologist to fix either problem by putting the argument into a Bayesian format.

Bayes is just a form into which a valid probabilistic argument can be put. A valid argument that reaches a false conclusion has to have at least one false premise, and there is no exception just because the argument is Bayesian. If it is circular, then at least one premise either restates the conclusion or presupposes it.

[shunyadragon]

If they're really based on information, any information, then they're not assumptions. They are inferences. If the information is unreliable, you may certainly criticize the inference on that basis. Or, if there is no apparent problem with the information, then you might find some fault in the logic by which the inference is deduced from it.


I have yet to see an argument for the resurrection that is not either invalid or circular. There is no way for an apologist to fix either problem by putting the argument into a Bayesian format.

Bayes is just a form into which a valid probabilistic argument can be put. A valid argument that reaches a false conclusion has to have at least one false premise, and there is no exception just because the argument is Bayesian. If it is circular, then at least one premise either restates the conclusion or presupposes it.

OK