In Substance Use Patterns of Young Offenders and Violent Crime, Lennings, Copeland, and Howard (2003, p. 415) argue their purpose is to ôvalidate the possibility that a connection exists between substance use or types of substances used and engaging in violent crime.ö The research hypothesis is that ôillicit substance use is related to a history of violent crimeö (Lennings, et al. 2003, p. 416). The sample population was a group of adolescent detainees who were provided with a survey over a one-month period in Queensland, Australia. The uses of screening measures were employed, including AUDIT (Alcohol Use Disorders Identification Test), which measures frequency, dose, and psychosocial harm related to alcohol use. After choosing the primary substance of use, the LDQ (Leeds Dependency Questionnaire) was provided on the specific substance chosen.
The independent variable used by the researchers was the admission of violence. Dependent variables were age, the AUDIT score, a 5-item depression scale developed by Tambs and Moun, and a general substance dependency score calculate from the LDQ (Lennings, et al. 2003). To measure the significance of each variable, the multivariate analysis of variance (MANOVA) was used on two groups from the sample, those admitting to violent offense and those not. A score greater than 10 on the AUDIT indicates ôsignificant risk of problem alcohol useö (Lennings et al. 2003, 416). The overall MANOVA was found to be significant (RaoÆs R = 3.39(4), p<.05) and analysis of the univariate FÆs revealed that only the AUDIT score discriminated between the two groups, ôwith those convicted of a violent offence reporting a higher mean AUDIT score than those that did not (F = 9.65 (1,84) p<.01 violent offenders mean û 21.00, non-violent offenders = 15.36)ö (Lennings, et al. 2003, p. 416).
The researchers also performed a second analysis using the frequency of life-time use of substances, however...