Several recent meta-analyses have revealed high correlations between negative attitudes and/or peer associations, and criminal behavior. Andrews and Bonta conducted a meta-analysis in 1994 and found that the highest correlations with risk were displayed through anti-social attitudes and associates when compared to the six major correlates of risk: 1) lower-class origins; 2) personal distress/psychopathology; 3) educational/vocational achievement; 4) parental/family factors; 5) temperament/misconduct and personality; and 6) anti-social attitudes/associates. Similarly, another meta-analysis, conducted by David J. Simourd in 1993, found even stronger correlations for antisocial attitudes and associates when compared to lower-class origins; personal distress/psychopathology; family structure/parent problems; minor personality variables; quality of parental relationship; personal educational/vocational achievement; and temperament/misconduct/self-control. Paul Gendreau, Tracy Little and Claire Goggin found similar results in 1996 when examining anti-social attitudes as a factor by itself, isolating the effects independently of anti-social associates.
The importance of attitudes when predicting anti-social behaviors remains the same when comparing male and female offenders.
When the factors correlated with risk (and subsequently recidivism) are disaggregated between males and females, the individual correlations differ slightly, but the ordering of importance remains the same. In other words, when predicting risk, and subsequently criminal behavior, anti-social attitudes are highly predictive, thereby revealing a dynamic risk factor that can be targeted through effective correctional intervention.
Risk/Needs Assessments Some of the most valid risk/needs assessments for offenders are those that incorporate dynamic risk factors through the measurement of criminogenic needs.
In addition, the strength of the anti-social attitude factor that was identified in the aforementioned meta-analyses is extended by research that has shown predictive validity when anti-social attitudes and orientations are measured by a dynamic risk/needs assessment.
When Simourd and researcher Wagdy Loza assessed anti-social attitudes using standardized and objective risk/needs assessment instruments, consistent positive correlations were displayed between these attitudes and criminal behavior. In addition, anti-social attitudes have been positively (and linearly) associated with the severity of the offense as well. Thus, the stronger the presence of anti-social attitudes, the more severe the offending may be.
Anti-social attitudes are highly correlated with other anti-social behaviors as well, such as taking drugs. Finally, Ian W. Shields and Georga C. Whitehall have shown that anti-social attitudes are highly predictive of institutional misconduct, particularly violent behavior.
One dynamic risk/needs assessment instrument that takes anti-social attitudes into account is the Level of Service Inventory-Revised (LSI-R), designed by Andrews and Bonta. LSI-R measures risk and need for service across 10 different domains. Many of the 54 items that are measured by LSI-R are considered dynamic, and thereby are subject to change through appropriate correctional intervention.
Of the 54 items that are taken into account by LSI-R, four are devoted specifically to anti-social attitudes. However, it should be noted that, because attitudes are pervasive in many other domains of everyday functioning, in reality, attitudes are being measured through several other domains within the instrument. Many studies have revealed that scores on LSI-R as a whole significantly correlate with recidivism (as well as many other correctional outcomes). In addition, correlations between the attitude and orientation components and outcome have been significant as well. These findings have been consistent for both LSI-R (the adult version of the instrument) as well as the Youthful Offender-Level of Service Inventory (YO-LSI), the juvenile version of the instrument.
In 1991, Shields and Simourd found YO-LSI helped distinguish between predatory and non-predatory offenders in an institutional setting. They first conducted several tests to determine if the individual components of YO-LSI were reliable.
The individual components, including psychological variables, were consistently higher for the predatory offenders than the non-predatory offenders. They fo