G2SPSSAdvancedAbout 30 min to complete
Agreement and reliability between raters
Why percentage agreement is not enough.
Slides · Dataset
About 30 minutes to complete
Enrol to open the slides and download the data.
What this module covers
why percentage agreement is not enough; Cohen's kappa and what it corrects for; weighted kappa for ordered categories; Fleiss' kappa for more than two raters; interpreting kappa magnitude and the kappa paradox with skewed prevalence; intraclass correlation for continuous measurements; Bland-Altman analysis for method comparison, with the plot built step by step; how to report agreement.
By the end of it
- Why percentage agreement is not enough
- Cohen's kappa and what it corrects for
- Weighted kappa for ordered categories
- Fleiss' kappa for more than two raters
- Interpreting kappa magnitude and the kappa paradox with skewed prevalence
- Intraclass correlation for continuous measurements
- Bland-Altman analysis for method comparison, with the plot built step by step
- How to report agreement
The dataset
Records review and device comparison — Sigma teaching dataset
A teaching dataset of 60 cases, generated from a fixed seed so every figure in the module reproduces exactly.
- CodebookIncluded
- Dataset (CSV)Included
- Dataset (CSV)Included
- SPSS syntaxIncluded