Philosophy and the practice of Bayesian statistics in the social sciences
This chapter presents our own perspective on the philosophy of Bayesian statistics, based on our idiosyncratic readings of the philosophical literature and, more importantly, our experiences doing applied statistics in the social sciences and elsewhere. Think of this as two statistical practitioners...
Main Author: | |
---|---|
Other Authors: | |
Format: | Text |
Language: | English |
Published: |
2010
|
Subjects: | |
Online Access: | http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.180.3161 http://www.stat.columbia.edu/%7Egelman/research/published/philosophy_chapter.pdf |
Summary: | This chapter presents our own perspective on the philosophy of Bayesian statistics, based on our idiosyncratic readings of the philosophical literature and, more importantly, our experiences doing applied statistics in the social sciences and elsewhere. Think of this as two statistical practitioners’ perspective on philosophical and foundational concerns. We are motivated to write this chapter out of dissatisfaction with what we perceive as the standard view of the philosophical foundations of Bayesian statistics. Here’s what we take as the standard view:- Bayesian inference--“inverse probability”--is inductive, learning about the general from the particular. The expression p(theta|y) says it all. This is in contrast to classical statistics which is based on hypothesis testing, that is, falsification.- The paradigmatic setting of Bayesian inference is the computation of the posterior probability of hypotheses. To give a quick sense of our position, we agree essentially entirely with Greenland (1998), who attributes to Karl Popper the following attitude toward inductive inference: “we never use any argument based on observed repetition of instances that does not also involve a hypothesis that predicts both those repetitions and the unobserved instances of interest. ” To put it another way, statistical models are a tool that allow us to make inductive inference in a deductive framework. |
---|