Related Experiment Videos
Note on comparison of small proportions in large-scale experiments
Journal of the National Cancer Institute
|May 1, 1977
Summary
This study introduces a statistical method for comparing treatment and control groups with binary outcomes, like tumor presence. It provides a confidence interval for the ratio of response rates, suitable for large samples with few positive results.
Area of Science:
- Biostatistics
- Clinical Trials
- Medical Research
Background:
- Comparing treatment efficacy often involves binary outcomes (e.g., tumor presence/absence).
- Existing statistical methods may be inadequate for large sample sizes with sparse positive responses.
Purpose of the Study:
- To develop a statistical confidence interval procedure for comparing a treatment to a control group.
- To provide a method suitable for binary response data when sample sizes are moderately large but positive responses are few.
Main Methods:
- A statistical confidence interval procedure was developed.
- The method focuses on the ratio of two response rates.
Main Results:
- The procedure yields a confidence interval for the ratio of response rates.
- It is specifically designed for scenarios with moderately large sample sizes and a small number of positive responses.
Conclusions:
- The proposed statistical method offers a reliable way to compare treatment and control groups under specific conditions.
- This approach enhances the analysis of binary outcome data in research settings with limited positive events.