Related Concept Videos
Sample Size Calculation
The sample size for the given experiment or sampling effort is fundamental to any study design. Sample size decides the number of...
Prediction Intervals
However, the point estimate is most likely not the exact value of the population parameter, but close to it. After calculating point estimates, we construct interval estimates, called confidence intervals or prediction intervals. This prediction interval comprises a range of values unlike the point estimate and is a better predictor of the observed sample value, y.
Estimating Population Mean with Unknown Standard Deviation
William S. Gosset (1876–1937) of the...
Margin of Error
Receiver Operating Characteristic Plot
Survival Tree
Building a Survival Tree
Constructing a...
You might also read
Related Articles
Articles linked to this work by shared authors, journal, and citation graph.
What influences whether researchers adhere to healthcare reporting guidelines successfully? A thematic synthesis.
A scoping review of guidance on sharing trial results with patients and the public finds inconsistent recommendations and limited methodological development.
Stakeholder perspectives on potential open-label bias in patient-reported outcome results from cancer clinical trials: evidence from the IMPROvE project.
Related Experiment Video
Updated: Sep 17, 2025

Automatic Image Processing to Determine the Community Size Structure of Riverine Macroinvertebrates
Published on: January 13, 2023
Extended sample size calculations for evaluation of prediction models using a threshold for classification.
Rebecca Whittle1,2, Joie Ensor3,4, Lucinda Archer3,4
1Department of Applied Health Sciences, School of Health Sciences, College of Medicine and Health, University of Birmingham, Birmingham, UK. r.l.whittle@bham.ac.uk.
Calculating adequate sample size is crucial for accurate individualised risk prediction models. This study provides new formulas and code to determine the minimum sample size for precisely estimating threshold-based performance measures, enhancing model evaluation.
Area of Science:
- Biostatistics
- Medical Informatics
- Epidemiology
Background:
- Accurate sample size calculation is essential for reliable individualised risk prediction models.
- Existing guidance focuses on calibration, discrimination, and net benefit, but often overlooks threshold-based measures.
- Threshold-based performance metrics are frequently reported in clinical practice.
Purpose of the Study:
- To extend existing sample size guidance for prediction models.
- To provide methods for precisely estimating threshold-based performance measures.
- To offer tools for calculating minimum sample size in external validation studies.
Main Methods:
- Developed closed-form solutions for sample size estimation of accuracy, specificity, sensitivity, positive predictive value (PPV), and negative predictive value (NPV).
- Utilized an iterative method for estimating sample size for the F1-score.
- Incorporated user-defined target standard error, expected performance values, and outcome prevalence.
- Considered extensions for time-to-event outcomes.
Main Results:
- New formulae and computational tools (Python, R, Stata) are provided for sample size calculation.
- In examples, the required sample size for threshold-based measures was often less than for calibration slope estimation.
- The methods enable precise estimation of various performance metrics in external validation.
Conclusions:
- The developed methods and tools facilitate accurate sample size determination for threshold-based performance measures.
- These criteria should complement existing guidance for a comprehensive evaluation of prediction models.
- Researchers can now more reliably assess the minimum sample size needed for robust model validation.

