Human Aggression While Under the Influence of Alcohol and Other Drugs: An Integrative Research Review

Abstract

For hundreds of years, it has been assumed that people behave more aggressively while under the influ of alcohol and other drugs. Correlational evidence provides some support for conventional wis dom. Many studies have found a strong relation between drug intoxi cation and violent crimes (e.g., homicide, assault, rape). These stud ies generally find that approximately 50% of the assailants were intoxi cated at the time the violent crimes were committed.1 This is an alarm ing statistic. Are responsible for half of the violent crimes com mitted in the United States? Do cause violent crime? More generally, is there a causal relation between and aggression? If there is such a causal relation, what is its magnitude? Unfortunately, it is difficult to draw causal inferences about the ef fects of on aggression from correlational data. Some of the com I plications surrounding correlational evidence are these: The aggressor may misreport drug ingestion as an excuse or to avoid punishment; drug consumption may accompany par ticipation in group events that could lead to violence; some drug addicts involve themselves in crimes to sup port their habits; and drug-related bungling of crimes may increase the probability of capture.2 The experi method, however, avoids these and other pitfalls by allowing the researcher to control the occur rence of events. Consequently, it is much easier to draw causal infer ences from experimental data than from correlational data. It becomes useful, then, to determine what the experimental literature tells us about the relation between and ag gression. The data base for the present review is limited to experi studies of drug-related hu man Jules Henri Poincar? said, Sci ence is built up with fact, as a house is with stone. But a collection of facts is no more a science than a heap of stones is a house.3 The task for the reviewer is to organize and integrate the facts or results obtained from a collection of studies. There are two general approaches to ac complishing this task: the narrative (or qualitative) approach and the meta-analytic (or quantitative) ap proach. In the traditional narrative review, the reviewer uses mental algebra to integrate a collection of studies and describes the results ver bally. In the meta-analytic review, the reviewer uses statistical proce dures to integrate a collection of studies and describes the results us ing numerical effect sizes. Tradi tional narrative reviews are more likely than meta-analytic reviews to depend on the subjective judg ments, preferences, and biases of the reviewer. Meta-analytic proce dures are used in the present review. Meta-analysis is not without its own shortcomings. One of the most common criticisms of meta-analysis is that it combines studies that use different operationalizations of the independent and dependent vari ables. This problem, known as the and oranges problem, can be recast in terms of threats to the construct validity of putative causes and effects. Is the hypothetical con struct apples, or is it fruit? If the construct is apples, it would be inappropriate to combine apples and oranges; yet if the construct is fruit, it would be appropriate to combine apples and oranges. It is therefore the responsibility of the meta-analyst to provide clear opera tional definitions of the independent and dependent variables. In the fol lowing sections, operational defini tions are given for the constructs drugs and aggression.