[Skip to Content]
[Skip to Content Landing]
Views 518
Citations 0
Comment & Response
April 16, 2019

Lowering the P Value Threshold—Reply

Author Affiliations
  • 1Oklahoma State University Center for Health Sciences, Tulsa
JAMA. 2019;321(15):1533. doi:10.1001/jama.2019.0574

In Reply Mr Adibi and colleagues raise salient points about the prior probability of replication of late-phase RCTs and the increased sample size requirements necessary for a P value threshold of .005. Although we agree that the .005 threshold would require adjustments to funding strategies and trial design, we remain unconvinced that the .005 threshold would come with “little added benefit.” Ultimately, we remain supportive of the new .005 threshold,1 for this proposal was meant to serve as a temporizing measure while improved statistical frameworks are developed.