The use of human observers in psychopharmacological research.

During 1974-1978, over 40% of the nonhuman drug studies that appeared in Psychopharmacology, Pharmacology Biochemistry and Behavior, and Journal of Pharmacology and Experimental Therapeutics involved human observers; far fewer studies published in Journal of the Experimental Analysis of Behavior did so. In all of these journals, measures of interobserver agreement seldom were provided. The great majority of studies also failed to utilize one or more "blind" observers, unaware of experimental conditions. These findings are of interest in light of reports that observational data are affected by a wide range of factors and often provide an inaccurate index of behavior. The believability of observational data seemingly is enhanced by careful descriptions of recording procedures coupled with the use of two or more blind observers whose concordance in rating behavior has been determined. These procedures characteristically are followed in some behavioral sciences, such as applied behavior analysis, but not to the same degree in psychopharmacology.