Bayesian thinking and the Statistical analysis of the Book of Mormon

News from a recent Exmo presentation rocked the world of exmos and apologists alike when a presentation suggested two entirely new books as sources for the Book of Mormon.  The Late War of the United States against Britain and The First Book of Napoleon

In an attempt to defend the faith, a blogger approaches the subject here

http://www.mormoninterpreter.com/a-bayesian-cease-fire-in-the-late-war-on-the-book-of-mormon/

And I can always applaud anyone who uses Bayesian thinking.  However, I think some discussion of his approach is worth thinking about.

I think this section is key:

A person’s belief in H (That the Last War was pivitol in the creation of the Book of Mormon) before encountering E (The sheer volume of matching material) can be expressed as a probability P(H), called the prior probability of H. If someone has no opinion about H, but wants to use probability theory as a mind game for thinking about potential results, it is convenient to set P(H)=0.5. The person’s disbelief in H before encountering E is denoted by P(H*)=1-P(H), which is also 0.5 when P(H)=0.5.

What he is saying, in essence, is that now we know that a certain book is similar to the Book of Mormon, let’s calculate probabilities based on someone learning about the matches.

What is key here is that he (or she) is working backwards now that we have a match, to discuss what impact that has on belief.

The people who ran the experiment had their experiment based on the idea that the Book of Mormon was a unique book.  Something that was unlike the rest of 1800’s writing entirely, in both phrases and topics.  This is something I’ve heard claimed by Parley P. Pratt and other more modern general authorities.

Their hypothesis was that NO books would come forward.  That’s it.  They tested it, and several books popped out that could clearly have been used in the books construction.

The author then states:

In my opinion, P(E|H) ( = the probability of the evidence assuming that LW is a direct source for the Book of Mormon) is at most 0.8. 

And this is the rub.  We’ve thrown gut intuition into a mathematical formula.  The original authors were asking “Is there any probability that books were used in the compilation of the book of Mormon, and the blog writer is saying “There is a chance I feel that it could be true…”

The shocking news that there are books that match VERY CLOSELY to the Book of Mormon in terms of phrase and content should be enough to cause one to re-evaluate belief, but if you use your personal bias in the calculation by taking a gut swag, you are due to tip the statistical outcome.

In other words, one should seek to reject the null hypothesis (The Last War and Napoleon could have been used as sources to the Book of Mormon).  Any test that could prove this false would be a good approach, but arguing that belief should remain because you FEEL the probability should be… that’s not a good use of Bayesian mathematics for science, that is abusing Bayesian mathematics to resist change.

As to the study, I’m unconvinced that The Late War and Napolean were the only sources.  I see more significance in context around Spaulding-Rigdon.  However, I’m willing to be dissuaded.  As soon as I can find a way to prove or disprove it, I’ll write about the method, but as for now, I cannot reject the null hypothesis; and I’m not willing to simply interject my beliefs into the calculation.

I”m comfortable saying “I don’t know, it could be the case that these books were sources for the Book of Mormon”.  The believer should be as well.

This entry was posted in Book of Mormon, Current issues. Bookmark the permalink.
Last edited by Mithryn on August 15, 2014 at 7:44 pm

2 Responses to Bayesian thinking and the Statistical analysis of the Book of Mormon

  1. J. says:

    Can I ask your history with Bayesian statistics is? Bayesian statistics are very nuanced, and often misused. I agree that the article you references does a disservice by inserting “gut feelings” into a mathematical formula. But I believe your null hypothesis and other statements are equally uninformed.

    Either way, the only conclusion that can be drawn from your analysis or the Mormon Interpreter analysis is “There is not enough evidence to convince someone that these books were sources for the Book of Mormon.” (You say “I’m unconvinced that The Late War and Napoleon were the only sources” and the Mormon Interpreter article says ” No slam dunk has occurred”)

    • Mithryn says:

      I am a novice with bayesian mathematics. I have an economic degree and work with data in my profession, but have only been working with bayes’ formula for 4 years

Leave a Reply to J. Cancel reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.