Designing Reliable Studies with Coefficient of Variation
Author Information
Author(s): Shoukri Mohamed M, Elkum Nasser, Walter Stephen D
Primary Institution: King Faisal Specialist Hospital and Research Centre
Hypothesis
Can the within-subject coefficient of variation be effectively used to estimate sample size requirements for continuous and binary variables?
Conclusion
The study provides methods for efficiently designing reliability studies using the within-subject coefficient of variation for both continuous and binary variables.
Supporting Evidence
- The study proposes a variance-stabilizing transformation to improve reliability estimates.
- Sample size estimation is crucial for designing studies that use inferential statistics.
- The within-subject coefficient of variation is suggested as a reliable measure for reproducibility.
Takeaway
This study helps researchers figure out how many people they need to include in their studies to get reliable results, whether they're measuring things like blood pressure or yes/no answers.
Methodology
The study uses Monte Carlo simulations to validate confidence intervals and discusses optimal allocation of sample resources for reliability studies.
Limitations
The study does not account for systematic errors in measurements.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website