Fisher’s criticism of Bayesianism in Statistical Methods for Research Workers is rather pathetic—one of his justifications went along the lines of “since other intelligent people dismiss Bayesianism, there must be some reason to dismiss it.” I would say that simple irrationality is insufficient to explain why clearly intelligent and experienced people would actively choose to ignore Bayesianism for such flimsy reasons. Instead, to explain the popularity of frequentism over Bayesianism, it is necessary to understand that scientists are motivated by more than the desire to obtain the correct answer; instead, science is a social activity in which researchers are motivated by the desire to obtain status and influence. In this regard, the fact that frequentism is more mathematically impressive and the fact that frequentism can give a better appearance of objectivity are important factors for explaining the popularity of frequentist methods.
That said, frequentist methods may indeed be better suited as the standard for mainstream science than Bayesian methods. The reason why frequentist methods seem more objective is because of their inflexibility. This same inflexibility makes it more difficult for researchers to engage in unsound practices like making multiple analyses of their data, and then only publishing the results which look the best. There is still a large enough variety of frequentist methods to make this sort of manipulation possible, but overall, frequentist methods still allow for less ‘wiggle room’ than Bayesian methods.
In contrast, if Bayesian methods were standard, all kinds of complicated hierarchical models might become the standard in academic literature. As few peer reviewers have both the background and patience to carefully scrutinize these models, the opportunity becomes much greater for the crafty researcher to “cheat” by adjusting their priors to best support their results. This may already be a problem in scientific fields where Bayesian methods are accepted.
Fisher’s criticism of Bayesianism in Statistical Methods for Research Workers is rather pathetic—one of his justifications went along the lines of “since other intelligent people dismiss Bayesianism, there must be some reason to dismiss it.” I would say that simple irrationality is insufficient to explain why clearly intelligent and experienced people would actively choose to ignore Bayesianism for such flimsy reasons. Instead, to explain the popularity of frequentism over Bayesianism, it is necessary to understand that scientists are motivated by more than the desire to obtain the correct answer; instead, science is a social activity in which researchers are motivated by the desire to obtain status and influence. In this regard, the fact that frequentism is more mathematically impressive and the fact that frequentism can give a better appearance of objectivity are important factors for explaining the popularity of frequentist methods.
That said, frequentist methods may indeed be better suited as the standard for mainstream science than Bayesian methods. The reason why frequentist methods seem more objective is because of their inflexibility. This same inflexibility makes it more difficult for researchers to engage in unsound practices like making multiple analyses of their data, and then only publishing the results which look the best. There is still a large enough variety of frequentist methods to make this sort of manipulation possible, but overall, frequentist methods still allow for less ‘wiggle room’ than Bayesian methods.
In contrast, if Bayesian methods were standard, all kinds of complicated hierarchical models might become the standard in academic literature. As few peer reviewers have both the background and patience to carefully scrutinize these models, the opportunity becomes much greater for the crafty researcher to “cheat” by adjusting their priors to best support their results. This may already be a problem in scientific fields where Bayesian methods are accepted.