An Examination of Direct and Indirect Effects of Maltreatment Dimensions and Other Ecological Risks on Persistent Youth Offending

• Date: 01 June 2010
• DOI: 10.1177/0734016809360327
• Author: P.J. Verrecchia,Matthew D. Fetzer,John H. Lemmon,Thomas L. Austin
• Published date: 01 June 2010
• Subject Matter: Articles

An Examination of Direct

and Indirect Effects of

Maltreatment Dimensions

and Other Ecological Risks

on Persistent Youth Offending

P. J. Verrecchia,

1

Matthew D. Fetzer,

2

John H. Lemmon,

3

and

Thomas L. Austin

3

Abstract

The study examined the ecological model by testing the direct and indirect effects of four

maltreatment dimensions (supervisory neglect, age at onset, recurrence, and severity) on

persistent youth offending. A path model was constructed hypothesizing that maltreatment,

family functioning, and community risks would increase behavior and academic problems in

childhood and delinquency in adolescence. The design featured within-group analysis that included

methodological and statistical controls determining how variations in maltreatment affect delinquent

behavior among maltreated children. Supervisory neglect produced direct and indirect effects. Mal-

treatment severity produced direct effects only. Family functioning and community risks produced

indirect effects. The results support the ecological explanation of the maltreatment/delinquency link.

The key implication of the study is that accurate and early identification of maltreatment, coupled

with interventions that improve parental discipline and supervision, will reduce persistent youth

offending while also mediating the effects of other risk factors.

Keywords

maltreatment dimensions, ecological risks, juvenile delinquency

Although thelink between maltreatment anddelinquency is not clearly understood, some important

questions have been answered thus far. Studies using improved methods featuring quasiexperiments,

prospective designs, and multiple datasources (Bolton, Reich, & Gutierres,1977; English, Widom, &

Branford,2002; Lemmon, 1999; Ryan & Testa,2005; Smith & Thornberry,1995; Stouthamer-Loeber,

Wei, Homish, & Loeber, 2002; Widom 1989a, 1989b; Zingraff, Leiter, Johnson, & Myers, 1993)

1

Behavioral Sciences Department, York College of Pennsylvania, USA

2

School of Criminal Justice, University at Albany, NY, USA

3

Criminal Justice Department, Shippensburg University, PA, USA

Corresponding Author:

John H. Lemmon, Criminal Justice Department, Shippensburg University, 305 Shippen Hall, PA 17257, USA.

Email: jhlemm@ship.edu

Criminal Justice Review

35(2) 220-243

ª2010 Georgia State University

Reprints and permission:

sagepub.com/journalsPermissions.nav

DOI: 10.1177/0734016809360327

http://cjr.sagepub.com

220

indicate that the presence of maltreatment is a significant predictor of chronic and violent youth

offending even in the presence of other ecological (e.g., gender, race, family, social class, and com-

munity) risks.Although there is ample evidence that the presenceof maltreatment is a strong predictor

of youth offending, literature on the effects of maltreatment dimensions is limited and inconclusive.

Some studiesexamining maltreatment dimensions indicatethat the relationship is eitherspurious (Zin-

graff et al., 1993) or so confounded by other risk factors (Smith & Thornberry, 1995) that the actual

relationship is exceedingly difficult to discern.

Thus far, research on the maltreatment/delinquency link has established a direct link between the

presence of maltreatment and various measures of the initiation, persistence, and severity (violent)

of youth offending. This genre of research (Lemmon, 1999; Smith & Thornberry, 1995; Widom,

1989b; Zingraff et al., 1993) has been based on comparison group designs featuring maltreated ver-

sus nonmaltreated youth. Important as these findings are, methods that have used between-group

comparisons fail to take into account that maltreatment varies along different dimensions (see dis-

cussion in Smith & Thornberry, 1995).

Because it has been established that the presence of maltreatment affects youth offending in

comparison group studies, a logical course of action to continue this line of inquiry is to examine

the effects of maltreatment dimensions. This can be facilitated through within-group analyses of

maltreated children only. This is not a novel idea. A number of researchers have pointed out (Carter,

Sterk, & Hutson, 1998; Howling, Wodarski, Kurtz, Gaudin, & Herbst, 1990; Widom, 1989c) that

aggregating abuse and neglect into a single category of maltreatment for between-group compari-

sons with nonmaltreatment groups masks the actual effects on delinquency.

Within-group designs also have an advantage of reducing measurement error. The use of nonmal-

treatment comparisons tend to distort actual relationships because comparison group scores on

delinquency measures (i.e., the number of arrests) tend to cluster around 0 whereas maltreatment

group scores vary upward (Lemmon, 2006; Smith & Thornberry, 1995). Lemmon (1999) reported

that the use of comparison groups in studies of maltreatment dimensions created artificial linear rela-

tionships that inflated maltreatment effects on various measures of offending.

The purpose of the study was to examine an ecological model of delinquency (Bronfenbrenner,

1977, 1979; Catalano & Hawkins, 1996; Garbarino, 1999; Loeber & Farrington, 2001) by testing the

direct and indirect effects of four dimensions of maltreatment (supervisory neglect, age at onset,

recurrence, and severity) on a measure of persistent youth offending. The maltreatment effects were

examined in the context of individual (i.e., behavior and academic) and other environmental (e.g.,

family functioning and community) risks also correlated with youth offending. A path model was

constructed hypothesizing that the maltreatment dimensions along with family functioning and com-

munity risk would increase the subjects’ behavior and academic problems in childhood which, in

turn, would increase persistent youth offending during adolescence.

The design used a cohort of low-income, urban, male children. The selection of the cohort was

based on ecological risks (e.g., poverty, urban life, and gender) associated with youth offending,

which served as methodological controls to isolate the maltreatment effects. The cohort

(n¼632) was comprised of 330 youth who had received child welfare services (52%) and 302 non-

maltreated (48%) youth.

1

The study featured the 330 maltreated youth and the results were based on

recursive path modeling. In addition to the methodological controls, a set of individual (i.e.,

behavior and academic) and environmental (e.g., family functioning, and community) risk indices

were used in the path models to map out the temporal sequence of risks on the persistent youth

offending (i.e., the number of delinquent referrals).

The current study offers a more precise understanding of the maltreatment/delinquency relation-

ship than currently exists in the extant literature. First, within-group analyses facilitate examinations

of the effects of specific dimensions of maltreatment on youth offending. This approach provides

more exact assessments of the relationship that can be offered by the current genre of studies, which

Verrecchia et al. 221

221