Invariance of Measures of Prevention Program Effectiveness: A Replication

Recent literature has suggested that measures of risk and protective factorsfor delinquency and substance use are not equally reliable or valid acrossgender and ethnic groups and has recommended differentiated programming andculturally specific evaluation methods. Three data sets containing up tofive ethnic groups were used to determine the degree to which risk andprotective factors are equally reliable and valid predictors of drug use anddelinquency across gender and ethnic groups. Congeneric measurement modelsand structural equation models were evaluated to determine if the factorstructures for these measures and their covariances with measures of druguse and delinquency were equivalent across gender and ethnic groups. Half ofthe risk and protective factors included in this analysis were found to beequally reliable across gender and ethnic groups. When controlling forreliability differences, all of the risk and protective factors were foundto predict both drug use and delinquncy for all gender and ethnic groups. Interms of the magnitude of these associations, no substantive differenceswere found in the validity of risk and protective factors for drug use anddelinquency. Differences in the validity of risk and protective factors weremore prevalent for delinquency than for drug use. However, all differenceswere substantively trivial. We conclude that measures of prevention programeffectiveness are invariant across gender and ethnic groups.     

Problems in the measurement of dynamic risk factors in sexual offenders

ABSTRACT

Having valid and reliable tools to measure ‘dynamic’ risk factors and minimising assessor bias is said to be critical in helping to make decisions regarding how individuals convicted of committing crime should be managed and treated, and how resources within the Justice System should be prioritised. This article outlines the problems and issues associated with the measurement of such putative dynamic risk factors (see Ward, 2015 Ward, T. (2015). Dynamic risk factors: Scientific kinds or predictive constructs. Psychology, Crime and Law, 22(01-02), 2–16. [Google Scholar] [Dynamic risk factors: Scientific kinds or predictive constructs. Psychology, Crime and Law.], 22(1–2), 2–16], this volume for a critique of current conceptions of these), and suggests ways of ensuring that measurement is conducted in the most appropriate way for the population being dealt with. Specifically, we would note that the evidence suggests that psychometric analysis, and structured professional judgment can be used to measure dynamic psychological problems using some tools within some samples. However, their use can be hindered by a number of conceptual factors including how the tests are designed and used, their reliability and validity, the context in which they are used, and the samples used. Finally, we would note that interviews are also an important part in the process of dynamic assessment.     

Striking Out: Race and Support for Police Use of Force

This research examines whites’ and blacks’ support for police use of force using a survey‐based experiment that varies the race of the offender across four different scenarios. Bivariate results show that the race of the offender influences blacks’ approval for the use of force by police, but does not affect whites’ approval. Multivariate analyses examine whether the factors influencing support for police use of force vary depending on the race of the offender. Results indicate that the predictors for approval of police use of force differ by the race of respondent, the race of offender, and the appropriateness of the use of force. The implications of the results for police–community relations are discussed.     

Manipulative relational behaviour and delinquency: sex differences and links with emotional intelligence

Trait emotional intelligence (EI) encompasses high levels of emotional understanding and is generally associated with positive outcomes. However, research has suggested that high EI might predispose some young females to delinquency. The present study investigated whether this relationship can be accounted for by a tendency towards emotionally manipulative behaviours, facilitated by high EI. Two hundred and fifty-two young adults (125 female) completed measures of EI, Machiavellianism, the managing emotions of others scale (MEOS) and self-reported delinquent behaviour. High EI females presented higher levels of delinquency, Machiavellian Tactics and morality, the supposedly prosocial MEOS behaviours enhancing and diverting and the non-prosocial behaviours worsening, inauthenticity and concealing. High-EI males reported fewer delinquent offences, high levels of MEOS prosocial enhancing behaviour and low levels of non-prosocial behaviours. We suggest that high-EI may enable manipulative relational behaviours in some females which in turn support delinquency aimed at fulfilment of social or material goals.     

Sex and ethnic differences in past aggressive behaviors

To investigate gender and ethnic differences in experiences of violence, 415 Black and White males and females were asked about some of their aggressive behaviors in the past. In their most aggressive encounters, males were more likely than females to have received and instigated physical violence and females to have experienced violence in a sexual context; same-sex aggression was more common than cross-sex violence. More males than females had urged or screamed at others to be more aggressive, with males more likely to incite other males and females to urge other females to be aggressive. Males were more likely than females to have last been angry with a male, and a number of sex differences were found in the behaviors exhibited when last angry. Although positive consequences of aggression did not differ significantly by sex, females were more likely to have experienced negative interpersonal effects of behaving aggressively and males to have suffered physical harm or legal troubles. Relatively few differences between Black and White subjects were found, but Black males were more likely than Whites to keep their anger to themselves and to get the target in trouble; White males were more likely to yell at the target and tell the target of their anger. White females were more likely than Blacks to get the target in trouble. In general, the results are consistent with sex role stereotypes and suggest that the experiences of aggression and responses to anger may be substantially different for males and females.     

Race and Demographic Factors in Treatment Attendance for Domestically Abusive Men

This study examined the influence of race and other demographic characteristics on session attendance and dropout from a cognitive-behavioral treatment program for male domestic abusers (N = 101). Multiple regression analyses revealed that lower session attendance was significantly predicted by unemployed status and African American (versus Caucasian) race. Other demographic factors could not account for the race differences in session attendance. Race was the strongest predictor of dropout and session attendance in all analyses. Racial composition of treatment groups was not significantly related to number of sessions attended for Caucasian or African American clients. Although the combined (interactive) effect of race and referral source on number of sessions was not statistically significant, the pattern of means suggested that self-referred African Americans were at the greatest risk for dropout. Results suggest that race differences in dropout, and the mechanisms that account for these differences, deserve more clinical and research attention in treatment for domestic abusers.     

Predictors of sex offender treatment dropout: psychopathy,sex offender risk,and responsivity implications

Abstract

We examined sex offender treatment dropout predictors, in particular, the relationship of psychopathy and sex offender risk to treatment dropout in a sample of 154 federally incarcerated sex offenders treated in a high intensity sex offender treatment program. Demographic, criminal history, mental health and treatment-related data as well as data on risk assessment measures including the Static 99, Violence Risk Scale – Sexual Offender version (VRS-SO), and Psychopathy Checklist – Revised (PCL-R) were collected. Logistic regression and discriminant function analyses were used to identify predictors that made significant and unique contributions to dropout among all the variables under study. The Emotional facet of Factor 1 of the PCL-R and never being married were found to be the most salient predictors of treatment dropout and correctly identified about 70% of the cases. The implications of the findings for managing treatment dropout and for the treatment of psychopathic offenders are discussed.     

Developmental patterns of sex differences in delinquency among African American adolescents: A test of the sex-invariance hypothesis

This paper addresses a developmental issue concerning longitudinal patterns of sex differences in delinquency. Hirschi and Gottfredson argue that the age-delinquency relation is invariant across sex and that sex differences in delinquency are invariant over time as well. A combination of these two propositions generates a hypothesis, called here the sex-invariance hypothesis, that sex differences in delinquency are invariant over developmental stages of adolescents. To test the sex-invariance hypothesis, nine waves of panel data collected from a representative urban sample of African American adolescents are analyzed. The overall findings show that sex differences in delinquency tend to vary as the subjects grow older, rather than remain constant as the invariance thesis posits. Specifically, sex differences in delinquency peak at the age of 15 and thereafter declines with age. We also find that parental supervision significantly explains sex differences in delinquency for younger adolescence, but not for older adolescence.     

Gender differences in gang participation,delinquency, and substance use

The purpose of this paper is to respond to gaps in our knowledge about patterns of female gang participation and its causes and consequences. Data from the Rochester Youth Development Study, a panel study that overrepresents adolescents at high risk for delinquency, are used to compare gang participation and delinquent involvement of female and male adolescents. We then examine the role of theoretical variables associated with both female and male gang membership. The results lead us to conclude that, for females as well as males, involvement in gangs is associated with substantially increased levels of delinquency and substance use. There is also some similarity in the factors associated with gang membership for both sexes, although lack of school success emerges as a factor of particular salience for female adolescents. The results suggest that theory and intervention need to address the phenomenon of female gang membership as an important component of urban youth problems.Authors' names are listed in alphabetical order.     

Race as a Factor in the Judicial Decision of Bail Amount in a Midwestern Jurisdiction

Abstract

Data collected on all felony defendants during 1996 from a district court in a Midwestern county were examined for racial differences in the setting of bail. While we hypothesized that African Americans would receive higher bail amounts than whites, we found no independent effect of race on bail amounts either before or after applying controls for legal variables-seriousness of offense and prior arrests-and extralegal variables-sex, age, residency, and type of attorney. Thus, we were compelled to reject our hypothesis. Seriousness of offense and the interaction between offense seriousness and race were statistically significant in the regressions.     

Drug use and pretrial misconduct in New York city

Is recent drug use significantly associated with pretrial misconduct? Does consideration of recent drug use enhance risk classification among a sample of persons who have time free pending the disposition of their cases? Using data on arrestees in Manhattan, this paper examines these issues and some related questions. To measure recent drug use, urine samples were collected from persons shortly after their arrest and tested for four drugs: heroin, cocaine, PCP, and methadone. Two measures of pretrial misconduct are considered: whether a defendant fails to appear for a scheduled court date (FTA) and whether a defendant is rearrested prior to case disposition. Censored probit models are used to estimate the statistical association between drug test results and pretrial misconduct. Results show that drug test results are significantly associated with pretrial misconduct over and above the information typically available to judges at the time release decisions are made. Some implications of these findings for pretrial decision making are discussed.     

Racial and ethnic differences in the risk factors associated with bully victimization

Bullying victimization has been recognized as a social issue facing a large proportion of America’s children and adolescents. Although important contributions to the knowledge base have been made regarding risk factors for bullying, little is known about whether a relationship between the potential victim’s race/ethnicity and their likelihood of bully victimization exists. Further, whether the factors that place persons at risk for bullying victimization are invariant across groups is unknown. The present study attempts to fill these voids in the literature by using a national sample and incorporating a more comprehensive list of predictors compared to what has been used in previous studies. Results indicate that risk factors for bullying are largely invariant across race and ethnicity.     

Continuity in crime: Sex and age differences

This study deals primarily with the question of continuity in serious criminality. The study follows a cohort consisting of 15,117 individuals from the age of 15 years through the age of 30 years. The age-based transition probabilities are studied, both for males and for females, showing that the transitions for females conform to a first-order Markov chain, so that the original matrices can be seen as estimates from a single parent matrix. This was not the case for males. The predictive power of prior criminality for later transition probabilities is also studied. Finally, the difference between stating that a given percentage will relapse in crime and the ability to make real prediction is emphasized.     

Improving the quality of drug court clinical screening: a call for performance measurement policy reform

Despite the widespread use of the drug court model, standardized performance measures for drug courts are not uniformly utilized, and rarely include process measures. To ensure that drug courts are being implemented in the most effective manner, the use of performance measurement tools should be considered for wide scale adoption. Drug court effectiveness is moderated by participant characteristics, and is most effective for individuals with the highest substance use needs. Therefore, having quality clinical screening processes is crucial to ensuring that drug courts are serving the population for which they are effective. This paper examines clinical screening in drug courts, to answer the following 1) what is the current state of screening, 2) what works, and 3) why measurement matters. It also proposes a clinical screening performance measure to improve fidelity and ensure appropriate participant enrollment. The creation of a performance measure would create opportunities to improve drug court outcomes, and leverage pay-for-performance models.     

Sex differences in the experience of child sexual victimization

This study analyzed 390 (303 girls and 87 boys) cases of child sexual abuse to understand sex differences in child sexual victimization. Analyses focused on three domains that have received considerable empirical attention: victims' relationships with the offender, type and extent of abuse, and disclosure. Findings largely replicated previous efforts that reported sex differences in terms of victims' relationships with offenders. Although results also replicated previous findings about sex differences in disclosure patterns, these sex differences were minor in comparison with similarities in the way most instances of abuse were eventually identified. In terms of type and extent of abuse, results were the opposite of previous findings: girls, not boys, were the victims of more physical injury, violent threats, and use of force; and girls, not boys, were more likely to have physically resisted abuse. Divergences in findings were interpreted in terms of memory biases and social forces operating differently on the sexes.     

Race and the probability of pleading guilty

The legal ramifications of pleading guilty and findings of an interdependence between pleading guilty and sentence severity suggest that the guilty plea decision is a significent turning point in case processing. The present research examines the variables affecting the probability of pleading guilty. The first analysis involves estimating a single probit equation of main effects of variables previously found to be related to pleading guilty. A second analysis is conducted estimating the same equation separately for black defendants and white defendants. Findings from the first part of the analysis indicate that physical evidence, number of charges, and confessing to the crime during police/prosecutor interrogation increase the probability of pleading guilty, whereas the number of witnesses, use of a weapon, and offenses carrying a minimum penalty of 5 years in custody with no maximum prison term decrease the probability of pleading guilty. Findings from the second analysis indicate that the effect of marital status, prior record of felony convictions, type of counsel, number of charges, and use of a weapon on the probability of pleading guilty varies by defendant's race. The research concludes by offering several competing explanations of these findings in hope of stimulating further research on the variables affecting the route of case disposition in felony processing.     

Measurement Properties of the Communities That Care<Superscript>®</Superscript> Youth Survey Across Demographic Groups

Prevention science has produced information about risk and protective factors that predict adolescent drug use and related problem behaviors. This paper investigates the Communities That Care Youth Survey that measures multiple risk and protective factors. Using a sample of 172,628 students who participated in surveys administered in seven states in 1998, analyses were conducted to test the factor structure of these risk and protective factors and to test the equivalence of the factor models across five racial/ethnic groups (African Americans, Asians or Pacific Islanders, Caucasians, Hispanic Americans, and Native Americans), four grade levels (6th, 8th, 10th, and 12th) and both gender groups. Results support the construct validity of the surveys risk and protective factor scales and indicate that the measures are equally reliable across males and females and five racial/ethnic groups. Implications of these findings for science-based prevention planning are discussed.     

Predictors of Support of Legislation Banning Juvenile Executions in Oklahoma: An Examination by Race and Sex

The Supreme Court has recently decided to re‐examine the constitutionality of executing individuals under the age of 18 at the time of the offense. The Supreme Court’s reliance on public opinion as evidenced through opinion polls and changing laws in the 2002 Atkins decision has suggested that public opinion may play a role in the Court’s decision regarding juvenile executions. There is considerable evidence that the majority of Americans favor a ban on juvenile executions. In the current study, we use Oklahoma data collected in 2003 by the Oklahoma University Public Opinion Learning Laboratory to examine more closely the factors that predict a support of a ban on juvenile executions. Interestingly, only one fourth of Oklahomans oppose such a ban. Earlier research suggested that religious fundamentalism is linked to support of juvenile executions, but we did not find this, suggesting that public opinion may be shifting. We then analyzed the data separately by race and then by sex. Our findings suggest that there may be differences between groups in the predictors of support for a ban on juvenile executions, at least in Oklahoma, indicating the need for further research.     

Gender differences and patterns of arrest across two generations of Cuban immigrants in the United States

The present study is based on a secondary-data analysis of a subsample of Cuban immigrants residing in the United States. While there are no significant differences in arrest rates between U.S. natives of Cuban descent and their foreign-born counterparts, life transitions have a differential impact on behavioral outcomes when young men and women of Cuban ancestry are compared. Job instability increases the risk of arrest for men, while being a single parent increases the risk of arrest for women. Growing up in a two-parent family and marriage have crime protective effects for men, while religiosity decreases the probability of arrest for women. In both subsamples, having a family member arrested was the strongest predictor of one’s arrest status in early adulthood.     

In the eye of the beholder: sex and race estimation using the human orbital aperture

From the works of Broca and Krogman to modern-day Jantz and Buikstra, the orbit has been used for both quantitative and qualitative sex and race estimation. This study evaluates the practical value of these estimations. Orbital height and breadth were measured to determine the orbital index and assess differences between men and women or black people and white people in the Hamann-Todd Collection. Replicability of these measures was also examined. Finally, a geometric morphometric study was performed to assess shape differences using the entire margin. Significant differences were found in both the index and the geomorphometric study; however, further investigation revealed that the differences were of little practical use. The measurement differences were found to be smaller than intra-observer error, while the geometric morphometric analysis demonstrated that minimal percentage of variation in shape was attributable to group differences. Thus, these techniques should not be used to estimate sex or race.