Related Experiment Video
Updated: Jun 15, 2026

A Naturalistic Setup for Presenting Real People and Live Actions in Experimental Psychology and Cognitive Neuroscience Studies
Published on: August 4, 2023
Is an ROC-type response truly always better than a binary response in observer performance studies?
David Gur1, Andriy I Bandos, Howard E Rockette
1University of Pittsburgh, Department of Radiology, Radiology Imaging Research, Pittsburgh, PA 15213, USA. gurd@upmc.edu
Comparing performance metrics for mammography interpretation, this study found that binary (yes/no) ratings offered larger differences in performance comparisons than receiver-operating characteristic (ROC)-type ratings, especially when considering the observer
Area of Science:
- Radiology and Medical Imaging
- Observer Performance Studies
- Statistical Analysis in Medical Research
Background:
- Evaluating diagnostic performance is crucial for medical imaging, particularly for breast cancer screening technologies like full-field digital mammography (FFDM) and digital breast tomosynthesis (DBT).
- Observer performance studies rely on rating data to compare imaging modalities, with different data types (binary vs. ROC-type) potentially influencing results.
Purpose of the Study:
- To compare performance assessment methods using binary and ROC-type rating data in observer studies.
- To evaluate differences in comparing full-field digital mammography (FFDM) versus FFDM plus digital breast tomosynthesis (DBT).
- To understand how rating scale choice impacts the detection of performance differences between imaging techniques.
Main Methods:
- Eight radiologists interpreted 77 digital mammographic examinations using both binary (yes/no) and ROC-type (0-100) rating scales.
- Performance was quantified using the area under the ROC curve (AUC) for ROC-type data and Youden's index for binary data.
- Differences in reader-averaged AUCs between FFDM and FFDM+DBT were compared, considering reader variability.
Main Results:
- Absolute differences in performance metrics were larger on average for binary ratings (0.12) compared to ROC-type ratings (0.07).
- Standardized differences also indicated larger effects for binary ratings (2.32 vs. 1.63).
- The observed discrepancies were influenced by the specific operating points chosen by individual readers.
Conclusions:
- The choice of an observer's operating point is critical when designing performance studies.
- While ROC-type ratings offer detailed information, binary ratings may better reflect actual observer behavior and provide statistical advantages in certain applications.
- Binary response analysis can offer statistical advantages in specific clinical scenarios for observer performance studies.
Related Concept Videos
Receiver Operating Characteristic Plot
Dose Response Curve: Conventional Versus Nonmonotonic
Actor-Observer Effect
Region of Convergence of Laplace Tarnsform
Consider a decaying exponential signal that begins at a specific time. When deriving its Laplace transform, the time-domain variable is replaced with a complex variable. This substitution...
Observational Studies
There are three types of observational studies – Prospective, retrospective, and cross-sectional.
Prospective Study
Prospective studies, also known as longitudinal or cohort studies, are carried out by collecting future data from groups sharing similar characteristics. One example of...
Dose-Response Relationship: Overview

