# Standard Agreement Calculation

The basic measure of reliability among evaluators is a percentage of agreement between evaluators. When calculating the percentage agreement, you need to determine the percentage of the difference between two numbers. This value can be useful if you want to see the difference between two numbers as a percentage. Scientists can use the percentage of agreement between two numbers to display the percentage of the relationship between the different results. To calculate the percentage difference, you need to take the difference in the values, divide them by the average of the two values, and then multiply this number by 100. For example, multiply 0.5 by 100 to get a percentage of 50%. Note that Cohen`s kappa only measures the concordance between two evaluators. For a similar degree of compliance (Fleiss` kappa) used if there are more than two evaluators, see Fleiss (1971). The Fleiss Kappa is however a multi-miss generalization of Scott`s Pi statistics, not Cohens Kappa. Kappa is also used to compare machine learning performance, but the directed version, known as Informationdness or Youdens J, is considered more suitable for supervised learning.  Multiply the quotient value by 100 to get the percentage of concordance for the equation.

You can also move the decimal to the right two places, which gives the same value as multiplying to 100. Cohen`s Kappa measures the concordance between two evaluators who divide each of the N elements into mutually excluded C categories. The definition of κ {textstyle kappa } is as follows: k=number of codes and w i j {displaystyle w_{ij} }, x i j {displaystyle x_{ij} } and m i j {displaystyle m_{ij} are elements in the weighting matrices, observed and expected. If the diagonal cells contain weights of 0 and all the off-diagonal cells contain weights of 1, this formula produces the same kappa value as the calculation shown above. A case that is sometimes considered a problem with Cohen`s kappa occurs if one compares the Kappa calculated for two pairs of evaluators, with both evaluators in each pair with the same percentage of concordance, but one pair gives a similar number of evaluations in each class, while the other pair gives a very different number of grades in each class.  (In the following cases, note B has 70 votes in favour and 30 against in the first case, but these figures are reversed in the second case.) For example, in the following two cases, there is an equality of correspondence between A and B (60 out of 100 in both cases) with respect to the correspondence in each class, so we would expect the relative values of Kappa cohens to reflect this. In calculating Cohen`s cappa for each: to calculate pe (the probability of a random match), we find that, as you can probably see, calculating percentage agreements can quickly become complicated for more than a handful of evaluators. For example, if you had 6 judges, you would have 16 pair combinations to calculate for each participant (use our combination calculator to find out how many couples you would get for multiple judges…