APS
2024 APS Annual Convention · 2024
Penalizing Optimal t Does Not Fix Its Type I Error Problem: Results from a Simulation Study on Missingness
- Rick Sperling
St. Mary's University - Fatima Gallardo Ibarra
St. Mary's University - Mia Stahl
St. Mary's University - Humberto Buenrostro
St. Mary's University - Garrette Bradley
St. Mary's University
Abstract
Previous research has shown that Optimal t has greater statistical power than paired samples t, but also inflated Type I error. Our study demonstrates that imposing a penalty on Optimal t does not decrease the risk of Type I error and leads to losses in statistical power across conditions.
Other
