• We are indebted to Cindy Gibson for conducting Study 2 and to Brian Mullen for his contributions to designing the research We thank Craig Anderson, Bill Chaplin, Lew Goldberg, Bill Graziano, Bob Greene, Bob Hogan, Bob Kraut, Charlie Lord, Del Paulhus, Fred Rhodewalt, Sandy Russ, Steve Scher, Doug Schultz, Bill Swann, and Steve West for their criticisms and suggestions in response to an earlier version of this article Rations of this material were presented at the convention of the American Psychological Association in Washington, DC, in August 1986.

Address correspondence to Roy F Baumeister, Department of Psychology, Case Western Reserve University, Cleveland, CM 44106


ABSTRACT A metatrait is the trait of having versus not having a trait It refers to whether a given trait dimension or construct can be used to describe a particular personality Using attitudes as an analog to traits, we argue that the study of personality may benefit from considering metatraits Implications for the nature of traits and for the formation and disappearance of traits within a personality are discussed We review strategies for measuring metatraits and conclude that at present it seems best to use interitem variance in responses to specific, familiar trait scales Metatraits will most commonly moderate trait effects, especially such that hypotheses about traits are appropriately tested only with trailed individuals Untraited individuals may be susceptible to fluctuating states and situational manipulations Sometimes metatraits may predict behavior directly, independently of trait level Two studies are reported to illustrate metatrait moderation of trait effects and direct prediction of behavior by metatrait