W5:ANOVA Tests for Three or More Groups

Biostatistical considerations are an important part of any clinical trial even before the first subject is enrolled. Sample size calculations need to consider the desired power of the study, the number of locations at which it will be conducted, an estimate of a clinically significant treatment effect for the outcome variable, and its variance. All descriptive, independent, and primary and secondary outcome variables need to be identified and defined and the data analysis plan must be clearly articulated.

To preparefor this Discussion, read the Cleland article, Comparison of the Effectiveness of Three Manual Physical Therapy Techniques in a Subgroup of Patients with Low Back Pain from this weeks Learning Resources.

Based on your knowledge of the assumptions involved in ANOVA,a response to the following questions:

• Is ANOVA the appropriate data analysis method? Why or why not?
• The data analysis plan anticipates 93% power to detect a difference in the primary outcome across the four locations but only 83% power within a single location. Is that difference important? Why or why not?