The clinical and scientific efficacy of behavioral analysis is dependent upon interveners' accurate and reliable detection and measurement of target behaviors. This study compared the accuracy and reliability of observers' detection and recording of a designated target behavior when different forms of a target behavior description were used. Using an intra-subject design, undergraduate college students were asked to count the number of target behaviors depicted on a videotape under each of two conditions. Conditions differed only to the extent that each contained a different description of the target behavior. Results showed that participants' detection and recording of the target behavior was more accurate and reliable when the target behavior description used a verb (in the present tense, active voice) depicting an action with an observable and discrete beginning and end and omitted modifiers requiring observers to make subjective or relative judgments. Analysis of the data using methods developed by Signal Detection Theory demonstrated the potential utility of this approach for studying observer detection of target behaviors.
Keywords: Signal Detection Theory; accuracy; description; gold standard; reliability; target behavior.