|
Activity Number:
|
157
|
|
Type:
|
Roundtables
|
|
Date/Time:
|
Monday, July 30, 2007 : 12:30 PM to 1:50 PM
|
|
Sponsor:
|
Business and Economics Statistics Section
|
| Abstract - #308421 |
|
Title:
|
Statistics Teaching: Bayesian, Frequentist, United
|
|
Author(s):
|
Emanuel Parzen*+
|
|
Companies:
|
Texas A&M University
|
|
Address:
|
Statistics Dept, College Station, TX, 77843-3141,
|
|
Keywords:
|
statistical education ; quantiles ; posterior quantiles ; confidence quantiles ; united statistics ; data modeling
|
|
Abstract:
|
All statisticians have problems understanding distinction between confidence intervals and credible intervals (which I call statistical inference without priors and with priors, rather than frequentist and Bayesian inference). In introductory statistics education, question is how to teach Bayesian prior methods while teaching with equal emphasis frequentist nonprior methods. We propose that the practice of statistical inference requires juggling several distributions simultaneously and a village of quantile functions of distributions. One important consequence of parallel reasoning between posterior quantiles and confidence quantiles (endpoints of confidence intervals): These provide for hypothesis testing both Bayesian and frequentist solutions; we recommend using and comparing both!
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2007 program |
