The Emergence of Optimal F: Applying Optimal t Logic to ANOVAs
Files
Publication Date
Summer 2024
Digital Publisher
Digital Commons at St. Mary's University
Collection
McNair Scholars Symposium
Keywords
t-test; ANOVA; analysis of variance; within-subject design; missing data; missingness
Description
Practitioners often deal with missingness when using with-subject designs. While there are useful approaches for dealing with missingness, such as imputation, most are too sophisticated for practitioners to apply. Those that are more manageable, such as deletion methods, have deleterious effects on the interpretability of results. In attempts to reduce the loss of statistical power through preservation of data, research on the use of between-subjects test in within-subjects scenarios have been examined (Avila et. al.,2021). Rather than using a paired samples t-test on pretest/posttest data, an independent sample t-test is used for its potential to increase statistical power form the preservation of data. This method was applied in a previous study in the development of Optima t, where R code is used to identify and perform the t-test with the highest statistical power based on specified parameters. (Cuevas et al., 2023). This study examined whether the logic of Optimal t can be applied to within-subjects scenarios that have three or more measurement points. We considered specified sample sizes (25, 50, 75), correlation (0,0.15, 0.3), missingness (0, 0.15, 0.3), and differences in means (0 sd, 0.25 sd, 0.5 sd). Results were encouraging as this new method, which we refer to as Optimal F, showed promising outcomes with respect to both Type I and Type II error rates. Recommendations for future research are discussed.
Disciplines
Analysis | Dynamical Systems | Logic and Foundations
Format
MOV
Medium
video
City
San Antonio
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.
