Main Ideas
- Null hypothesis significance testing (NHST) is the prevalent statistical method in biology, but it has significant limitations.
- NHST fails to provide two critical aspects:
- 1. the magnitude of the effect of interest, and
- 2. the precision of that estimate.
- Understanding biological importance relies on effect magnitude rather than statistical significance.
- It is recommended that all biological research include effect size statistics and confidence intervals (CIs) to facilitate better data analysis.
- Using effect sizes alongside CIs offers a more comprehensive understanding of data relationships compared to p-values.
- Presenting effect sizes can help researchers contextualize their results and contribute to future meta-analyses.
- The article discusses standardized effect size statistics (d statistics and r statistics) that can be applied across various study designs and are essential for meta-analysis.
- The authors also acknowledge the importance of unstandardized effect sizes and offer solutions to common challenges in calculating effect sizes and CIs, including issues with covariates, estimation bias, non-normal data, and non-independence of data.
- The paper serves as both a practical guide for beginners and a call to improve statistical practices within biological research.
Reference
Nakagawa, S., & Cuthill, I. C. (2007). Effect size, confidence interval and statistical significance: a practical guide for biologists. Biological Reviews, 82(4), 591–605. doi:10.1111/j.1469-185x.2007.00027.x