Related Experiment Videos
Meta-analysis of screening and diagnostic tests
1Center for Health Policy Research and Education, Duke University.
Psychological Bulletin
|January 1, 1995
Summary
This study introduces a universal measure for comparing the accuracy of screening and diagnostic tests across psychology, medicine, and education. The standardized distance between population means offers a unified approach for both continuous and dichotomous outcomes.
Area of Science:
- Psychometrics
- Medical Diagnostics
- Educational Assessment
Background:
- Screening and diagnostic tests are widely used in psychology, medicine, and education.
- Multiple tests often compete, necessitating accurate methods for comparing their performance.
- Existing measures may not be universally applicable to different types of outcome data.
Purpose of the Study:
- To introduce a generalizable measure for evaluating the accuracy of screening and diagnostic tests.
- To provide a unified metric applicable to both continuous and dichotomous outcome measures.
- To facilitate informed decisions when selecting among competing tests.
Main Methods:
- The proposed measure is the standardized distance between the means of two populations.
- For continuous outcomes, this corresponds to the standard effect size.
- For dichotomous outcomes, it is proportional to the sum of the log odds of sensitivity and specificity.
Main Results:
- The measure is easily computed for both continuous and dichotomous data.
- It offers a consistent way to quantify test accuracy.
- Properties of the measure are discussed with illustrative examples.
Conclusions:
- A single, versatile measure can effectively compare the accuracy of diverse screening and diagnostic tests.
- This metric aids in selecting the most accurate tests in psychology, medicine, and education.
- The standardized distance provides a robust foundation for test performance evaluation.