Fig. 4

Statistical power against sample size: plots showing the relationship between the statistical power of the ADAS-CogIRT and the ANCOVA methodologies and sample size for hypothetical treatment levels of (a) d = 0, (b) d = 0.2, (c) d = 0.5, and (d) d = 0.8. The trial duration was fixed at 24 months. ADAS-CogIRT Alzheimer’s disease assessment scale-Cognitive subscale scoring methodology based on item response theory, ANCOVA, analysis of covariance