Access

You are not currently logged in.

Access JSTOR through your library or other institution:

login

Log in through your institution.

If You Use a Screen Reader

This content is available through Read Online (Free) program, which relies on page scans. Since scans are not currently available to screen readers, please contact JSTOR User Support for access. We'll provide a PDF copy for your screen reader.

Posterior Predictive p-Values

Xiao-Li Meng
The Annals of Statistics
Vol. 22, No. 3 (Sep., 1994), pp. 1142-1160
Stable URL: http://www.jstor.org/stable/2242219
Page Count: 19
  • Read Online (Free)
  • Download ($19.00)
  • Subscribe ($19.50)
  • Cite this Item
Since scans are not currently available to screen readers, please contact JSTOR User Support for access. We'll provide a PDF copy for your screen reader.
Posterior Predictive p-Values
Preview not available

Abstract

Extending work of Rubin, this paper explores a Bayesian counterpart of the classical p-value, namely, a tail-area probability of a "test statistic" under a null hypothesis. The Bayesian formulation, using posterior predictive replications of the data, allows a "test statistic" to depend on both data and unknown (nuisance) parameters and thus permits a direct measure of the discrepancy between sample and population quantities. The tail-area probability for a "test statistic" is then found under the joint posterior distribution of replicate data and the (nuisance) parameters, both conditional on the null hypothesis. This posterior predictive p-value can also be viewed as the posterior mean of a classical p-value, averaging over the posterior distribution of (nuisance) parameters under the null hypothesis, and thus it provides one general method for dealing with nuisance parameters. Two classical examples, including the Behrens-Fisher problem, are used to illustrate the posterior predictive p-value and some of its interesting properties, which also reveal a new (Bayesian) interpretation for some classical p-values. An application to multiple-imputation inference is also presented. A frequency evaluation shows that, in general, if the replication is defined by new (nuisance) parameters and new data, then the Type I frequentist error of an α-level posterior predictive test is often close to but less than α and will never exceed 2α.

Page Thumbnails

  • Thumbnail: Page 
1142
    1142
  • Thumbnail: Page 
1143
    1143
  • Thumbnail: Page 
1144
    1144
  • Thumbnail: Page 
1145
    1145
  • Thumbnail: Page 
1146
    1146
  • Thumbnail: Page 
1147
    1147
  • Thumbnail: Page 
1148
    1148
  • Thumbnail: Page 
1149
    1149
  • Thumbnail: Page 
1150
    1150
  • Thumbnail: Page 
1151
    1151
  • Thumbnail: Page 
1152
    1152
  • Thumbnail: Page 
1153
    1153
  • Thumbnail: Page 
1154
    1154
  • Thumbnail: Page 
1155
    1155
  • Thumbnail: Page 
1156
    1156
  • Thumbnail: Page 
1157
    1157
  • Thumbnail: Page 
1158
    1158
  • Thumbnail: Page 
1159
    1159
  • Thumbnail: Page 
1160
    1160