Bayesian theory has been around for many years, but the predictive power of Bayesian Model Selection is showing a tremendous amount of promise that might actually offer the social sciences the Holy Grail of reproducibility… thus lifting them to the standard of the hard sciences.
Prepare yourself for awesome.
The SS Central America, a steamer carrying a cache of gold, sank off the southeast coast of the United States in 1857. Part mystery, part adventure story, “In Deep Water,” directed by Steven Leckart and presented by ESPN Films and FiveThirtyEight, recounts the tale of a group who used Bayesian theory to find the ship — and the gold.
Yes, you should watch it all.
‘Tis a great idea… so when the results of BMS demonstrate this sort of predictive success, that demonstrates incredibly utility that one could easily see put into other fields — specifically the social sciences such as economics, psychology, anthropology, etc.
Of course, the real test for Bayesian induction is practical results. A tremendous shot in the arm? If it could successfully predict the most data-driven statistical subset of all time… namely the 2016 U.S. presidential elections.
…but no worries. The lab rats are hard at work tinkering on how to get it right next time, and even if it completely missed the election results because its predicates were off? The predictive nature of this sort of abductive reasoning is way cool.