On the reliability of unitizing textual continua: Further developments

This paper builds on an agreement coefficient proposed by Krippendorff (Content analysis: an introduction to its methodology, 2013) for measuring the reliability of unitizing and coding continuous phenomena, for example, of texts, videos, or sound recordings. It serves three purposes: It modifies Krippendorff’s definition which turned out to behave not as expected when applied to more than two observers, coders, or annotators. It extends this reliability measure to a family of four coefficients able to assess the reliabilities of diverse properties of unitized continua. It adds a way to obtain the confidence intervals of these coefficients as well as the probability of failing to reach targeted reliability levels. And it describes and provides access to free software that calculates all values of this family of reliability coefficients.