Related Experiment Videos
Improving interpretability: gamma as an alternative to R(2) as a measure of effect size
D D Reidpath1, M R Diamond, G Hartel
1School of Health Sciences, Deakin University, 221 Burwood Hwy, Burwood VIC 3125, Australia. reidpath@deakin.edu.au
Statistics in Medicine
|May 18, 2000
Summary
A new statistic, gamma, better captures long-term performance deficits from disease exposure than traditional R-squared. Gamma estimates the probability of an unexposed individual outperforming an exposed one, reflecting cumulative effects.
Area of Science:
- Biostatistics
- Epidemiology
- Health Outcomes Research
Background:
- Traditional effect size measures like R-squared (variance explained) may not fully capture cumulative deficits from long-term disease exposure.
- R-squared focuses on group differences at a single point, potentially underestimating chronic health impacts.
Purpose of the Study:
- To introduce and evaluate an alternative effect size statistic, gamma, for assessing long-term performance deficits.
- To demonstrate how gamma can more effectively convey the cumulative impact of disease exposure compared to R-squared.
Main Methods:
- Proposed a novel statistic, gamma, defined as the probability of an unexposed individual outperforming an exposed individual.
- Compared the utility of gamma against traditional R-squared in scenarios with cumulative performance deficits.
Main Results:
- Gamma provides a more intuitive measure of cumulative deficits resulting from long-term exposure.
- Unlike R-squared, gamma can reflect the accumulating effect of a deficit over time.
Conclusions:
- The proposed gamma statistic offers advantages for quantifying the long-term impact of disease on performance.
- Gamma enhances the understanding of chronic health effects by capturing cumulative deficits.