Document is current

Article History

Accepted: 26 May 2022

First Online: 29 June 2022

Declarations

:

: The author has no conflicts of interest or competing interests.

: The research involves no human or animal subjects.

: No human subjects.

: No human subjects.

: This research involves computer simulations to validate a new method for generating absolute precision confidence intervals on unstandardized mean differences using sequential methods. No simulations require NHST methods or power analysis. All simulations use between 10,000 and 5 million iterations Sequential methods can be abused in the generation of confidence intervals in a fashion like P-hacking in the NHST world, and the major goal of the paper is to demonstrate how to avoid those pitfalls. The method is Frequentist rather than Bayesian. I report the results of exact replications of many of the simulations and provide code and instructions for investigators to replicate the simulations themselves.

Free to read: This content has been made available to all.