MEASUREMENT OF CONSTRUCTS USING SELF-REPORT AND BEHAVIORAL LAB TASKS: IS THERE OVERLAP IN NOMOTHETIC SPAN AND CONSTRUCT REPRESENTATION FOR IMPULSIVITY?

Abstract

Although highly emphasized in psychological research, there has been little empirical evidence examining the overlap in meaning for self-report measures and construct representation for behavioral lab tasks in most psy-chological constructs. Using the personality trait of impulsivity as an exam-ple, the authors completed a meta-analysis of 27 published research studies examining the relationship between these methods. In general, although there is a statistically significant relationship between multidimensional self-report and lab task impulsivity (r = 0.097), practically, the relationship is small. Examining relationships among unidimensional impulsivity self-report and lab task conceptualizations indicated very little overlap in self-report and behavioral lab task constructs. Significant relationships were found between lack of perseverance and prepotent response inhibition (r = 0.099); between lack of planning and prepotent response inhibition (r = 0.106), delay re-sponse (r = 0.134), and distortions in elapsed time (r = 0.104); between negative urgency and prepotent response inhibition (r = 0.106); and be-tween sensation seeking and delay response (r = 0.131). This little conver-gent validity evidence for impulsivity as measured by self-report and behav-ioral lab tasks could indicate that these two measures are assessing different constructs. If these are different constructs, referring to them in the litera-ture as “impulsivity” influences one to think of them as representing a uni-tary underlying construct, when, in fact, we may be measuring disparate constructs. When disparate measures are described using the same multidi-mensional moniker, little forward progress can be made in figuring out how a trait relates to a criterion of interest. Researchers should take care to specify which particular unidimensional constructs are operationalized with not only impulsivity, but with all traits. If self-report and lab task conceptu-alizations measure disparate aspects of impulsivity, we, as a field, should not expect large conceptual overlap between these methods.