Related Experiment Videos
The power of analysis: statistical perspectives. Part 1
A E Pulver1, J J Bartko, J A McGrath
1University of Maryland School of Medicine, Maryland Psychiatric Research Center, Baltimore 21228.
Psychiatry Research
|March 1, 1988
Summary
Statistical power is crucial for interpreting "negative" results. Many studies with nonsignificant findings have inadequate statistical power, making their conclusions unreliable.
Area of Science:
- Biostatistics
- Research Methodology
Background:
- Interpreting nonsignificant results requires understanding statistical power.
- Low statistical power can lead to false negative findings.
Purpose of the Study:
- To assess the statistical power of published studies reporting nonsignificant t-test results.
- To highlight the implications of inadequate power in research.
Main Methods:
- Analysis of 154 published studies with nonsignificant t-test outcomes.
- Calculation of statistical power using a standardized difference of unity for effect size.
Main Results:
- Over 50% of the examined studies demonstrated inadequate statistical power.
- This suggests a significant portion of "negative" results may stem from insufficient sample sizes.
Conclusions:
- Failure to consider statistical power can lead to misinterpretation of research findings.
- Many published nonsignificant results may be inconclusive due to low power, not the absence of an effect.