By combining the behavior ratings from two or more evaluators it is possible to obtain 'synthetic' ratings, which show higher interrater reliabilities and higher validities than individual ones. The extent of this superiority is demonstrated by reliability and validity coefficients obtained in an experiment in which tour nurses rated 25 patients, using the Warnsveld Ward A djustment Scale, a not yet fully developed scale of the Lucero and Meyer type. These ratings were compared with evaluations made by means of two existing stales.