To calculate pe (the probability of a random agreement), we note that: where is the match observed among the advisors (identical to accuracy), and pe is the hypothetical probability of the agreement of luck, using the data observed to calculate the probabilities of each observer randomly each category. If the advisors are in complete agreement, it`s the option ” 1″ “textstyle” “kappa – 1.” If there is no agreement between advisors who are not expected at random (as indicated by pe), the “textstyle” option is given by the name “. The statistics may be negative,[6] which implies that there is no effective agreement between the two advisers or that the agreement is worse than by chance. Nevertheless, important guidelines have appeared in the literature. Perhaps the first Landis and Koch[13] stated that the values < 0 were unseable and 0-0.20 as light, 0.21-0.40 as just, 0.41-0.60 as moderate, 0.61-0.80 as a substantial agreement and 0.81-1 almost perfect. However, these guidelines are not universally accepted; Landis and Koch did not provide evidence, but relied on personal opinion. It was found that these guidelines could be more harmful than useful. [14] Fleiss`[15]:218 Equally arbitrary guidelines characterize Kappas beyond 0.75 as excellent, 0.40 to 0.75 as just to good and less than 0.40 bad. There are several formulas that can be used to calculate compliance limits. The simple formula that was given in the previous paragraph and which works well for sample sizes over 60,[14] is then No. 1 if the debtors fully agree.

If there is no agreement between the councillors (other than what you might expect), it is ≤ 0. If statistical significance is not a useful guide, what is Kappa`s order of magnitude that reflects an appropriate match? The guidelines would be helpful, but other factors than the agreement may influence their magnitude, making it problematic to interpret a certain order of magnitude. As Sim and Wright have noted, two important factors are prevalence (codes are likely or vary in probabilities) and bias (marginal probabilities are similar or different for both observers).