ICC — Intraclass Correlation Calculator
Assesses the agreement between different raters (or repeated measurements by one rater) — the standard method for observer-agreement analysis in quantitative imaging research. Uses a two-way random-effects, absolute-agreement model and reports ICC(2,1) (single measurement) and ICC(2,k) (mean of k) with 95% CIs.
① Enter data
Each row = one subject (e.g. one patient / one measured target); each column = one rater. You can paste directly from Excel (tab, comma or space separated).
How to use & methodology
What is the ICC and when is it used?
The intraclass correlation coefficient (ICC) evaluates the agreement and reproducibility of measurements of the same object by different raters (or repeated by one rater). In imaging research it is commonly reported for inter-rater agreement of an index (e.g. lesion ADC value, diameter) between two physicians.
ICC(2,1) or ICC(2,k)?
ICC(2,1) reflects the reliability of a single measurement — the reproducibility of one physician measuring once, the value most often reported. ICC(2,k) reflects the reliability of the mean of k raters, reported when the final result is an average of several. Both use a two-way random-effects, absolute-agreement model.
What ICC value is good?
By the Koo & Li (2016) standard: below 0.5 poor, 0.5–0.75 moderate, 0.75–0.9 good, above 0.9 excellent. Quantitative imaging studies generally expect ICC above 0.75. Always report the 95% confidence interval, not just the point estimate.
How do I prepare the data?
Each row is one subject (e.g. one lesion or patient); each column is one rater. You can copy a range from Excel and paste it (tab, comma or space separated). At least 2 raters and several subjects are needed.