Evidential Calibration of Confidence Intervals

06/24/2022
by   Samuel Pawel, et al.
0

We present a novel and easy to use method for calibrating error-rate based confidence intervals to evidence-based support intervals. Support intervals are obtained from inverting Bayes factors based on the point estimate and standard error of a parameter estimate. A k support interval can be interpreted as "the interval contains parameter values under which the observed data are at least k times more likely than under a specified alternative hypothesis". Support intervals depend on the specification of a prior distribution for the parameter under the alternative, and we present several types that allow data analysts to encode different forms of external knowledge. We also show how prior specification can to some extent be avoided by considering a class of prior distributions and then computing so-called minimum support intervals which, for a given class of priors, have a one-to-one mapping with confidence intervals. We also illustrate how the sample size of a future study can be determined based on the concept of support. Finally, we show how the universal bound for the type-I error rate of Bayes factors leads to a bound for the coverage of support intervals, holding even under sequential analyses with optional stopping. An application to data from a clinical trial illustrates how support intervals can lead to inferences that are both intuitive and informative.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
02/07/2022

Valid confidence intervals for μ, σ when there is only one observation available

Portnoy (2019) considered the problem of constructing an optimal confide...
research
05/10/2019

Confidence intervals with maximal average power

In this paper, we propose a frequentist testing procedure that maintains...
research
03/04/2022

Coverage Properties of Empirical Bayes Intervals

This note is an invited discussion of the article "Confidence Intervals ...
research
11/27/2022

An Empirical Bayes Approach for Constructing the Confidence Intervals of Clonality and Entropy

This paper is motivated by the need to quantify human immune responses t...
research
07/21/2021

Frequentist inference for cluster randomised trials with multiple primary outcomes

The use of a single primary outcome is generally either recommended or r...
research
09/20/2019

Novel algorithm for confidence sub-contour box estimation: an alternative to traditional confidence intervals

The factor estimation process is a really challenging task for non-linea...
research
09/20/2019

A novel algorithm for confidence sub-contour box estimation: an alternative to traditional confidence intervals

The factor estimation process is a really challenging task for non-linea...

Please sign up or login with your details

Forgot password? Click here to reset