The diagnoses in agreement are located on the main diagonal of the table in Figure 1. Thus the percentage of agreement is 34/50 = 68%. But this figure includes agreement that is due to chance. E.g. Psychoses represents 16/50 = 32% of Judge 1’s diagnoses and 15/50 = 30% of Judge 2’s diagnoses. Thus 32% ∙ 30% = 9.6% of the agreement about this diagnosis is due to chance, i.e. 9.6% ∙ 50 = 4.8 of the cases. In a similar way, we see that 11.04 of the Borderline agreements and 2.42 of the Neither agreements are due to chance, which means that a total of 18.26 of the diagnoses are due to chance. Subtracting out the agreement due to chance, we get that there is agreement 49.6% of the time, where