Verbatim · Inter-coder reliability

Fleiss' kappa

Generalisation to a FIXED number n of raters per unit (raters may differ across units); nominal.

kappa = (Pbar - Pe) / (1 - Pe); P_i = (sum_j n_ij^2 - n)/(n(n-1)); Pe = sum_j p_j^2

Measurement level
nominal
Coders
n raters per unit, fixed; n>=2 required (n-1 in P_i denominator).
Seminal reference
Fleiss (1971), Psychological Bulletin 76(5):378-382

Benchmark bands

Benchmarks are reporting guidance, not strict cut-offs — the engine presents them with their source.

AgreementReadingSource
0.80 – 1.0almost perfectLandis & Koch (1977)
0.60 – 0.80substantialLandis & Koch (1977)
0.40 – 0.60moderateLandis & Koch (1977)
0.20 – 0.40fairLandis & Koch (1977)
0.0 – 0.20slightLandis & Koch (1977)
-1.0 – 0.0poor or below chanceLandis & Koch (1977)

Verbatim computes Fleiss' kappa with a deterministic, golden-verified math kernel — never a language model — with a bootstrap confidence interval and full provenance (coder + unit + kernel + fixture + run-id). The number is measured; the interpretation stays yours.

Measure this in the workshop →