|
Activity Number:
|
559
- Spectral Analysis, Process Monitoring, and Sampling
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 11, 2022 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
|
Abstract #322501
|
|
|
Title:
|
Research Synthesis with Missing Uncertainties
|
|
Author(s):
|
Andrew Leo Rukhin*
|
|
Companies:
|
National Institute of Standards and technology
|
|
Keywords:
|
meta-analysis;
objective prior;
one-shot device
|
|
Abstract:
|
There are many collaborative studies where the data are discrepant while uncertainty estimates reported by each study are not reliable. Then the classical random effects model which explains this phenomenon by additional noise with a constant heterogeneity variance is inadequate especially when the smallest uncertainties correspond to the cases which are deviant from the bulk of data. It is suggested to view the reported uncertainty only as a lower bound on true variance and to use maximum likelihood or Bayes procedures under this restriction. The situation when reported uncertainties are missing occurs when using one-shot devices or in citizen science. The Bayes rule against non-informative prior is developed and its properties reviewed.
|
Authors who are presenting talks have a * after their name.
