Journal of Personality and Social Psychology 2000, Vol. 78, No. 4, 670-689 Copyright 2000 by the American Psychological Association, Inc. 0022-3514/0<y$5.O0 DOI: 10.1037//0022-3514.78.4.67O Displaced Aggression Is Alive and Well: A Meta-Analytic Review Amy Marcus-Newhall Scripps College William C. Pedersen, Mike Carlson, and Norman Miller University of Southern California Content analysis of 122 social psychology textbooks confirmed that displaced aggression received a surge of attention immediately following J. Dollard, L. W. Doob, N. E. Miller, O. H. Mowrer, and R. R. Sears (1939), but subsequent interest sharply declined. Contemporary texts give it little attention. By contrast, meta-analysis of the experimental literature confirms that it is a robust effect (mean effect size = +0.54). Additionally, moderator analyses showed that: (a) The more negative the setting in which the participant and target interacted, the greater the magnitude of displaced aggression; (b) in accord with N. E. Miller's (1948) stimulus generalization principle, the more similar the provocateur and target, the more displaced aggression; and (c) consistent with the contrast effect (L. Berkowitz & D. A. Knurek, 1969), the intensity of initial provocation is inversely related to the magnitude of displaced aggression. In a commonly used anecdote to illustrate displaced aggression, a man is berated by his boss but does not retaliate because he fears losing his job. Hours later, when he arrives home to the greeting barks of his dog he responds by kicking it. Conceptually, displaced aggression can be defined as a level of aggression toward a target that, in terms of the tit-for-tat rule (Axelrod, 1984), incommensu- rately exceeds that which is ordinarily seen as justified by the level of provocation emitted by that target. In exceeding the aggression warranted by the target's behavior, it reflects the failure to respond aggressively toward the source of a temporally antecedent provo- cation, or in this case the berating boss. The notion that frustration leads to aggression is commonly known as the frustration-aggression hypothesis (Dollard, Doob, Miller, Mowrer, & Sears, 1939). Several conditions influence the Amy Marcus-Newhall, Department of Psychology, Scripps College; William C. Pedersen and Norman Miller, Department of Psychology, University of Southern California; Mike Carlson, Department of Occupa- tional Science and Occupational Therapy, University of Southern Califor- nia. Portions of this research were presented at me 1993 Western Psycho- logical Association Conference, Phoenix, Arizona, and the 1995 Western Psychological Association Conference, Los Angeles, California. This re- search was facilitated by National Science Foundation Grants BSN- 8719439 and SBR-9319752. We thank Robert Bonakdar, Heather Logelin, Brooks Park, and Karen Tomlinson for their assistance in providing the judgments of the moderator variables; Addy Chulef and Sharon Gross for providing judgments of the moderator variables on an earlier version of the manuscript; Deborah Little for assistance in data entry; and Brad Bushman for preparation of Figure 1 and for many helpful comments on the manuscript prior to its final revision. Correspondence concerning this article should be addressed to Amy Marcus-Newhall, Department of Psychology, Scripps College, 1030 Co- lumbia Avenue, Claremont, California 91711, or to William C. Pedersen or Norman Miller, Department of Psychology, University of Southern Cali- fornia, Los Angeles, California 90089-1061. Electronic mail may be sent to amarcusn@scrippscol.edu, billp@rcf.usc.edu, or nmiller@rcf.usc.edu. intensity and/or frequency of aggression: (a) greater levels of frustration, (b) stronger expectations of reaching a goal, and/or (c) increased interference with goal attainment (Berkowitz, 1989). Although aggression frequently is directed toward the agent per- ceived to have provoked it, sometimes other features of the situ- ation elicit restraint. Miller (1941) proposed several constraining factors: (a) the provoking agent is unavailable (e.g., the provoca- teur has left the immediate environment), (b) the source of frus- tration is intangible (e.g., bad weather or a foul odor as in Konecni & Doob, 1972; Rotton, Barry, Frey, & Soler, 1978), and (c) retaliation or punishment is feared from the provoking agent (e.g., the provocateur is one's boss or has other sources of power). When any of these constraining factors are present, direct aggression is often controlled (e.g., Bandura, 1973; Baron, 1971; Taylor, Schmutte, & Leonard, 1977). Instead, it is alleged to be redirected toward or displaced onto less powerful or more available targets, as described in our opening vignette. Baron and Bell (1975) provide an empirical example that is based on the second of these restraining factors in that the source of a frustrating initial provocation was intangible. Thus, in the first stage of their study, the ambient temperature of a room was manipulated (i.e., hot and humid vs. normal) during a filler task. In the second stage, anger arousal was manipulated by a confederate who either insulted (or did not insult) the participant. In the final stage, the same confederate served as the learner in a modified teacher/learner paradigm in a new room and thus was available as a target of displaced or triggered displaced aggression (depending on whether insult was absent or present in the second stage of the experiment). The dependent variable was the duration and inten- sity of shock across 20 trials. In sum, there was a manipulation of the presence of an initial provocation (i.e., hot and humid vs. normal temperature) and a subsequent opportunity to aggress against a target who had or had not provided an act (insult or no insult) that by itself could function as a triggering provocation unrelated to the initial provocation. Their results showed that the initial provocation increased aggressive responding irrespective of 670