Work Item
ASTM WK65967
New Guide for Binary Logistic Regression
1. Scope
This guide describes the analysis of data that is developed from test methods that produce go/no-go, Bernoulli trial outcomes, such as ignition or no-ignition or reaction/no-reaction data produced under a specific set of test conditions. Bernoulli trial outcome refers to test data with only two outcomes (i.e., flipping a coin heads or tails, etc.). This standard does not purport to address the actual test methodologies, which are assumed to be performed according to other recognized consensus standards. The values stated in SI units are to be regarded as the standard. The values given in parentheses are for information only.
Keywords
bernoulli trial, bernoulli likelihood, binomial distribution, logistic regression, logit, negative binomial distribution, repeatability, reproducibility
Rationale
There currently lacks strong central guidance in the form of a consensus standard for the handling of go/no-go data. Other committees have proposed the use of binary logistic regression as part of a test standard (ASTM G74 and E2862, for example), and this guide seeks to harmonize with the existing standards. Users would include anyone analyzing go/no-go data produced from a test method seeking a standardized approach for the data handling.
The title and scope are in draft form and are under development within this ASTM Committee.
Work Item Status
Date Initiated: 11-20-2018
Technical Contact: Nicholas Linley
Item: 000
Ballot:
Status: