Female Juvenile Risk

• Author: Traqina Q. Emeka,Jon R. Sorensen
• Published date: 01 October 2009
• Date: 01 October 2009
• DOI: http://doi.org/10.1177/1541204009334083
• Published By: Sage Publications, Inc.
• Subject Matter: Articles

Youth Violence and

Juvenile Justice

Volume 7 Number 4

October 2009 313-330

© 2009 SAGE Publications

10.1177/1541204009334083

http://yvj.sagepub.com

hosted at

http://online.sagepub.com

313

Female Juvenile Risk

Is There a Need for Gendered

Assessment Instruments?

Traqina Q. Emeka

University of Houston–Downtown, TX

Jon R. Sorensen

Prairie View A&M University, TX

There are few risk assessment instruments that address the unique risk factors for female

juvenile offenders. Typically, female risk factors are embedded within male risk factors. Thus,

the true differences between male and female juveniles may not be explored. In this study, a

generic risk assessment instrument was constructed from a pooled sample and used to identify

the level of risk posed by females in comparison with males. Though the constructed instru-

ment worked quite efficiently for males, the instrument provided only a marginal improve-

ment over chance prediction of recidivism for youthful female probationers. Lower rates of

predictive validity among the female subpopulation supports call for the development of gen-

dered risk assessment instruments.

Keywords: delinquency; female offending; gender; juvenile recidivism; risk assessment

Since the 1980s, there has been a 29% increase in female juvenile offending (Snyder &

Sickmund, 2006). Females are offending not only at higher rates, but also at earlier

ages. Given these trends, it is necessary to identify potential risk factors to ensure proper

programming and treatment for this juvenile population. Risk assessment instruments are

generally employed to address potential risk and needs of juveniles with the goal of pre-

venting future offending (Miller & Lin, 2007). Available risk assessment instruments,

however, tend to be generic, constructed, and validated and mainly utilize male samples.

Because females make up only a small percentage of juvenile offenders in a given sample,

risk factors that may most influence female offending are not identified and incorporated

in commonly available risk assessment instruments. A growing realization in recent years

that females risk factors differ from those of males have led authors to suggest that the

application of generic risk assessment instruments are likely to lead to errors in classifica-

tion when applied to females (Funk, 1999; Schwalbe, Fraser, Day, & Cooley, 2006).

Inadequate classification could lead to females either receiving more restrictive supervision

than necessary or failure to identify those most in need of services. Some suggest that

gender-focused risk assessment tools are necessary to address resource allocation and to

adequately meet female juvenile needs (Holtfreter & Cupp, 2007).

The purpose of this study is to determine whether a generic risk assessment instrument

constructed from a pooled sample adequately identifies the level of risk posed by females

vis-à-vis males. Because prevention efforts are thought to be most effective if they are

employed early before the onset of serious criminality, the sample used for this exercise

314 Youth Violence and Juvenile Justice

consists of youthful at-risk boys and girls adjudicated to probation between the ages of 10

and 14. After the construction of the general risk assessment instrument, its accuracy in

predicting female delinquency over a 2-year period is compared to that for boys. Lower

rates of predictive validity among the female subpopulation supports the call for gender-

specific risk assessment instruments.

Literature Review

Risk Assessment

According to Bonta (1996), there have been three generations of risk assessments. The

first generation of risk assessment involved employing only the use of clinical assessments.

Traditionally, clinicians attempted to gauge the probability of future juvenile offending on

the basis of their prior experience with offenders sharing similar characteristics. The prob-

lem with such an assessment was that in all but the most extreme cases, clinical judgments

tended to rely on varying factors across cases and decision makers, causing them to be

unreliable. As such, agents in the juvenile justice system often made crucial, but subjective

decisions to the benefit or detriment of juveniles.

Second- and third-generation risk assessments rely on actuarial risk assessment.

Actuarial assessments provide a more structured, less subjective approach to juvenile jus-

tice decision making. Actuarial assessments are typically best for estimating future crimi-

nality as they involve statistical models and risk factor tools. Actuarial assessments involve

predicting possible behavior based on how others have acted in similar situations and/or an

individual’s similarity to other groups. With an actuarial tool, offenders’ characteristics are

inventoried and risk is determined by the extent to which he or she possesses various risk

factors associated with recidivism (Zinger, 2004).

The second generation of risk assessment used static variables in the statistical model to

predict future offending. Ernest Burgess’ (1928) recidivism study of adults on parole is

considered the first second-generation actuarial risk prediction research in the field of

criminal justice. Burgess created a 21-point risk classification instrument to predict parole

failure. He identified static predictor variables and developed a simple additive index used

to sort parolees by level of risk. The resulting classification scheme was based on group

recidivism rates with the criterion being parole success or failure. The performance of Burgess’

actuarial assessment scale was compared to clinical assessments conducted by three prison

psychiatrists. In his study of 3,000 offenders, Burgess’ scale was more successful at pre-

dicting parole success or failure than the clinical assessments. The general format of his

scale continues to be used in determining risk scores in prediction research.

Though prediction research enjoyed a brief period of popularity in the early part of the

20th century (see Glueck & Glueck, 1950), criticism of the case study approach and sociol-

ogy’s rise to preeminence, which eschewed individual-level predictors of crime and delin-

quency, resulted in a period of dormancy for prediction research in the study of crime during

the 1930s to the 1960s. The reappearance of prediction research occurred with the efforts of

Wolfgang, Figlio, and Sellin (1972) and their work with career criminals and the chronic

offender. Since that time, the career criminal and developmental paradigms have encouraged