“…To our knowledge, this is the first time that an explicit association between SE and kappa has been reported, although given that the former is a measure of response variability and the latter of agreement, the negative correlation between them would be expected. Indeed, SE-based measures of consensus and dissention for Likert scale data have been suggested before [70,71], although they do not appear to have been used except by their originators. For both datasets (NOW and THEN), a kappa value of 0.4 corresponds to SE of 1.02; thus, an intermediate or better value of kappa, using Fleiss's definitions, corresponds to a SE of 1.02 or less.…”