False discovery rates: a new deal.

Stephens M. · 2017

Read False discovery rates: a new deal. by Stephens M. (2017) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc5379932.txt.

Description

Matthew Stephens offered a new deal for false discovery rates in Biostatistics in 2017. Big studies test thousands of claims at once, from genes to brain spots. Many flagged hits are false, and the false discovery rate gauges that share. Standard tools sum up each test in one number and treat all effects as alike, which wastes what the data hold.

His method takes two numbers per test: the size of the effect and its error bar. It also bakes in a plain prior hunch, that true effects bunch at zero with one hump. That hunch makes the math a simple, sturdy fit and sharpens the calls when it holds. The output adds ranges for the true effect of each hit, not just a yes or no. Labs can then rank hits by real size, not by noise.

Catalog record

CREATORStephens M.
YEAR2017
COLLECTIONPAPERS
FORMATTXT
FILE SIZE150 KB
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC5379932

Browse the full library · About the Tower of Babel