Embedding equivalence t-test results in Bland Altman Plots visualising rater reliability

ABSTRACT In determining rater (instrument) reliability many techniques are available. A main discrimination between methods is made by the level of measurement (binary, categorical, continuous). Most techniques, such as the Mean Absolute Difference (MAD) or the Intraclass Correlation Analysis (ICC), apply to continuous data. Two other analysis methods are equivalence t-tests (quantitatively determining significance of equivalence) and Bland Altman Plots (qualitatively judging the scatter in the plot). Here the integration of both methods is proposed and discussed. T-test results, especially the (in)significance of equivalence are visualised in the Bland Altman Plots and can be interpreted clearly, next to the qualitative judgement of the scatter in the Bland Altman Plot. This extends the value and usefulness of the Bland Altman Plotting technique.