The Faces Pain Scale was simple to use, readily understood by the children, and showed a realistic distribution of scores with respect to the type of pain being measured. With the exception of verbal reactions (which were not meaningfully related to self-report), observer ratings based on detailed coding of the child's behavior correlated only poorly to moderately well with self-report scores (r = .39 to r = .58). Similarly, although confident in their judgments, the nurses' ratings showed only moderate agreement with those of the children. In estimating the child's pain, all observers appeared to appropriately weigh changes in the child's facial behavior, which showed evidence of being the most sensitive behavioral index to the intensity of short sharp (needle) pain in 4 to 6 years olds.