RESUMEN
The kappa statistic is a frequently used measure of interobserver agreement when two or more observers are asked to rate the same items or subjects on some criterion. The advantage of the kappa statistic over simple agreement is that it corrects for agreement by chance. In dental research, because multiple examiners may be involved in assessing subject variables such as attachment loss, bleeding on probing, or periodontal pocket depth, some statistical measure of agreement is needed. The program described here was developed for estimating agreement among periodontal examiners, but also could be used in clinical teaching applications. Assumptions and limitations of the kappa statistic are discussed. Examples of outputs illustrate applications of the program.