Use this tool when a score or predicted probability is meant to separate two outcomes, such as diseased versus healthy or pass versus fail, and you want to know how well it does so across every possible cut-off. It suits medicine, psychology, education and machine learning.
About the roc curve
A receiver operating characteristic (ROC) curve plots sensitivity against 1 minus specificity for every possible cut-off of a continuous score. The area under the curve (AUC) summarises discrimination: 0.5 is no better than chance and 1.0 is perfect separation.
When to use it
Use it to judge a diagnostic marker, a risk score or the predicted probabilities from a logistic regression, and to choose a cut-off that balances missed cases against false alarms.
What this tool reports
The ROC curve with the chance diagonal and a highlighted cut-off point.
Area under the curve (AUC) and the sensitivity and specificity at the chosen cut-off.
Tables you can download as CSV and a ready-to-use APA7 results paragraph.
An editable figure: colours, text, sizes and line widths.
Assumptions to check
The outcome has two clear classes and every case is classified by a reliable reference standard.
Cases are independent (one row per person or item).
The sample contains enough cases in both classes for stable estimates.
How to report it in APA7
State the marker, the outcome and the sample size, then give the AUC with its interval, and the cut-off with its sensitivity and specificity. The tool writes this paragraph for you and you can edit it before downloading.
Frequently asked questions
Is the ROC Curve tool free?
Yes. You can run the analysis and read the results without an account. Create a free account or log in only when you want to download the results.
What does the AUC tell me?
It is the probability that a randomly chosen positive case scores higher than a randomly chosen negative case. Values near 0.5 mean the score does not discriminate.
How do I pick the best cut-off?
Choose the point that balances sensitivity and specificity for your purpose. A cut-off that is best for screening may differ from one that is best for confirmation.
Is my data uploaded to a server?
The analysis runs in your browser. Your data stays on your device while you work with the tool.