Home | Business News | Browse by Publication | C | Contemporary Drug Problems

Differences in predictors of alcohol-related aggression and non-alcohol-related violent perpetration: an exploratory analysis in a sample of young adult drinkers.

Article Excerpt
Violence among young people is an important public health problem (Dahlberg 1998; Simon, Mercy & Perkins 2001) and therefore it is imperative that researchers gain a better understanding of factors that explain its occurrence. However, studies aiming to better understand aggression among youth vary considerably in terms of the types of aggression that are examined. In particular, some studies have focussed on aggression involving alcohol consumption (i.e., measures that included some assessment of whether the respondent and/or opponent were drinking) (e.g., Coll, Shott & Morris 1999; Engs & Hanson 1994; Haworth-Hoeppner, Globetti, Stem & Morasco 1989) while others have examined aggression without alcohol (i.e., measures with no assessment of whether alcohol was consumed prior to or at the time of aggression) (e.g., Duncan, Albert, Duncan & Hops 1997; Fergusson & Horwood 2000). As suggested by Pernanen (1991), there may be different causes or combinations of causes of aggression depending on the type of aggressive behavior examined. Thus, it is important to determine whether a different set of explanatory variables explains alcohol-related aggression as compared to non-alcohol-related aggression.

In the present study, drinking behaviors and variables reflecting early behavioral problems and risk-taking are assessed in terms of whether they differentially predict two forms of aggression: Arguments or fights after drinking (i.e., alcohol-related aggression) and hitting or attacking someone (i.e., non-alcohol-related violent perpetration) in comparison with respondents reporting no aggression. These two main forms of aggression differ not only in terms of the presence of alcohol but also in terms of the severity of aggression (i.e., arguments or fights versus hitting or attacking) and the role of respondents (involvement versus perpetration), making it difficult to determine which differences in the two forms of aggression explain differences in associations. Therefore, this study is largely exploratory.

This research takes advantage of a unique longitudinal data base that includes childhood measures of problem behaviors as well as measures of current alcohol consumption, risk taking, and different forms of aggression. Therefore, while the measures of aggression may not be perfect for comparison, the results may provide direction for future research in terms of developing and testing hypotheses about how predictors of aggression may differ depending on the form of aggression examined. In addition, a third group of respondents who reported both arguments or fights after drinking and hitting or attacking someone are compared with those reporting no aggression, providing additional information about whether predictors are different for those engaging in multiple forms of aggressive behavior.

Methods

A secondary analysis of Young Adult data from the National Longitudinal Survey of Youth (NLSY79) was conducted. In 1979, multistage stratified sampling was used to obtain a national probability sample of young men and women aged 14 to 21 (as of December 31, 1978) residing in the United States. African Americans and Hispanics, socially disadvantaged non-African Americans and non-Hispanics, and members of the active military force were over-sampled. However, the socially disadvantaged and military samples were excluded from 1990 onward. Women from the 1979 sample who had children were administered questions regarding the development of their children. In 1994 and every second year thereafter, youth aged 15 and up who were offspring of the original female sample were administered confidential Young Adult Self Report Booklets (YASRB) on personal topics including delinquent activities and substance use. The present analyses focus on these young adult data.

The sample for the present analyses included respondents 17 to 21 years of age who reported drinking alcohol in the prior year. A composite data file was used from the years 1994, 1996, and 1998 in order to increase the size and heterogeneity of the sample. For those who responded to more than one year, the most recent year of data collection was selected to avoid duplication of cases across years. The resulting sample included 911 cases.

Measures

Alcohol-and non-alcohol-related aggression

Alcohol-related aggression (arguments or fights after drinking) was assessed by asking respondents how often they had "gotten into an argument or fight" during or after drinking in the previous 12 months with four response options (never, once in a while, fairly often, very often). Due to the highly skewed distribution of this variable, responses were recoded as never or ever. Non-alcohol-related violent perpetration (hitting or attacking someone) was assessed with two questions that were combined to form one variable. Respondents were asked whether they had hit or seriously threatened to hit someone and whether they had attacked someone with the idea of seriously hurting or killing them in the previous 12 months. A dichotomous variable was created reflecting hitting and/or attacking someone versus no aggression. Notably, for this variable, respondents were not asked whether they had been drinking prior to the incident. Therefore, it is possible a proportion of these incidents involved alcohol. A final outcome variable was computed for the present analyses consisting of four categories: 1) No aggression 2) Argument or...