Free intraclass correlation coefficient (ICC) calculator: paste your data matrix (rows = subjects, columns = raters or measurements), answer two plain-language questions and get all six McGraw and Wong (1996) forms — ICC(1,1), ICC(1,k), ICC(2,1), ICC(2,k), ICC(3,1) and ICC(3,k) — with 95% CIs, the intermediate ANOVA table, the Koo and Li (2016) interpretation and an APA 7 or Vancouver sentence ready to paste into your manuscript.
Worked example: 2 physiotherapists measure knee flexion in 15 patients with a goniometer; absolute-agreement ICC(2,1) = .896, 95% CI [.723, .964], F(14, 14) = 19.51, p < .001: good reliability according to Koo and Li.
Which ICC model should I choose? If the same raters score every subject and are a sample of possible raters, ICC(2, absolute agreement); if they are the only raters of interest, ICC(3, consistency); if each subject has different raters, ICC(1). What value is acceptable? Below .50 poor reliability, .50 to .75 moderate, .75 to .90 good and above .90 excellent (Koo and Li, 2016), always interpreting the confidence interval.
Compute the ICC with all six McGraw & Wong forms, 95% CIs, the ANOVA table and Koo & Li interpretation. APA 7 sentence ready for your paper. Free and online.