Examining the Measurement Invariance and Psychometrics of the Drug Abuse Screening Test for Adolescents (DAST-A) in Justice-Involved Youth

• Published date: 01 September 2024
• DOI: http://doi.org/10.1177/00938548241246437
• Author: Alexandra Mogadam,Tracey A. Skilling,Michele Peterson-Badali,Liam Hannah
• Date: 01 September 2024
• Subject Matter: Articles

CRIMINAL JUSTICE AND BEHAVIOR, 2024, Vol. 51, No. 9, September 2024, 1413 –1432.

DOI: https://doi.org/10.1177/00938548241246437

Article reuse guidelines: sagepub.com/journals-permissions

© 2024 International Association for Correctional and Forensic Psychology

1413

EXAMINING THE MEASUREMENT INVARIANCE

AND PSYCHOMETRICS OF THE DRUG ABUSE

SCREENING TEST FOR ADOLESCENTS (DAST-A)

IN JUSTICE-INVOLVED YOUTH

ALEXANDRA MOGADAM

TRACEY A. SKILLING

University of Toronto

Centre for Addiction and Mental Health

MICHELE PETERSON-BADALI

LIAM HANNAH

University of Toronto

Substance abuse is a serious mental health concern and reoffense risk factor for justice-involved youth. The Drug Abuse

Screening Test for Adolescents (DAST-A) is used to assess drug abuse in different contexts, yet its psychometric properties

have not yet been thoroughly explored in youth justice samples. We examined the measurement invariance and psychometrics

of the DAST-A in a diverse sample of 741 justice-involved youth (Nyoung men = 636). The tool showed strong reliability in the

overall sample and subgroups (ω = .88–.94), and good convergent and concurrent validity. Logistic regression results indi-

cated that, with each unit increase in DAST-A score, the odds of an substance use disorder (SUD) diagnosis increased by 23%

(overall sample). The predictive validity findings were more robust for White youth than Black youth and as a result, a dif-

ferent cut-off score was explored for Black youth. The DAST-A demonstrated measurement invariance across gender and

race. Practice implications are discussed.

Keywords: drug abuse; substance use disorder; justice-involved youth; measurement invariance; psychometrics

Substance use disorders (SUDs) are characterized by co-occurring cognitive, behav-

ioral, and physiological symptoms that result in recurring use or craving of a sub-

stance, despite the negative substance-related sequelae (American Psychiatric Association,

2022). According to a 2010 national mental health survey in the United States, the life-

time prevalence of SUDs in youth was 11.4% (6.4% for alcohol abuse/dependence and

8.9% for drug abuse/dependence; Merikangas et al., 2010). SUDs are highly overrepre-

sented in the youth criminal justice arena, with prevalence rates ranging between 22% to

AUTHORS’ NOTE: We have no known conflict of interest to disclose. Correspondence concerning this article

should be addressed to Michele Peterson-Badali, Department of Applied Psychology & Human Development,

Ontario Institute for Studies in Education, University of Toronto, 252 Bloor Street West, 12th Floor, Toronto,

Ontario, Canada M5S 1V6; e-mail: m.petersonbadali@utoronto.ca.

1246437CJBXXX10.1177/00938548241246437Criminal Justice and BehaviorMogadam et al. / Psychometrics of DAST-A in Justice-Involved Youth

research-article2024

1414 CRIMINAL JUSTICE AND BEHAVIOR

51% in justice-involved youth; rates for cannabis abuse/dependence range from 8% to

45% (Teplin et al., 2002; Wasserman et al., 2005).

These high SUD rates in justice-involved youth are of clinical concern generally and of

significant concern specifically within the criminal justice space, as substance abuse is a

well-established recidivism risk factor in adults and youth (Dowden & Brown, 2002;

Stoolmiller & Blechman, 2005). Recidivism risk assessment tools typically provide an

overall estimate of a person’s risk for reoffense and identify specific domains of “crimino-

genic need,” including substance abuse, to be targeted in rehabilitative interventions.

Clinical best practice (Hoge & Andrews, 2011; Vincent et al., 2012) recommends multi-

method, multiinformant assessment, including the use of formal tools to assess functioning

in criminogenic need domains. Within the substance use domain specifically, the Drug

Abuse Screening Test for Adolescents (DAST-A; Martino et al., 2000) has been in use for

decades. Given the high prevalence rates of SUDs in youth justice populations and the need

for effective substance abuse screening in the context of criminogenic needs assessment, we

investigated the DAST-A’s psychometrics and cut-off scores within and across subgroups of

justice-involved youth that have been underresearched (young women) and are overrepre-

sented (Black youth) in the justice system.

THE DAST AND DAST-A

The Drug Abuse Screening Test (DAST; Skinner, 1982a) is a brief screener for drug abuse

over the past year. Developed and validated in a sample of adult treatment-seeking SUD

patients (N = 223, 72% men, sample racial make-up unknown), it consists of yes/no questions

that assess for any type of drug use. The recommended cut-off score to identify drug use of

clinical concern is >5 in the 20- and 28-item versions of the test (Skinner, 1982a).

Beyond its initial validation, the DAST’s psychometric properties have been found to be

strong in other contexts, including employment (El-Bassel et al., 1997) and criminal justice

settings (Saltstone et al., 1994); in different languages, including Turkish (Evren et al.,

2014) and Mandarin (Y.-T. Chen et al., 2020); and in various populations, such as different

mental health populations (e.g., Cassidy et al., 2008; McCann et al., 2000). The tool has also

demonstrated good internal consistency (α = .74–.998; e.g., Cassidy et al., 2008; Y.-T.

Chen et al., 2020) and test–retest reliability (α = .75–.85; e.g., Y.-T. Chen et al., 2020;

El-Bassel et al., 1997).

With regards to the DAST’s factor structure, some studies have found the scale to be

unidimensional (e.g., Y.-T. Chen et al., 2020; Skinner, 1982a), and others as multidimen-

sional, although usually with one dominant factor (e.g., El-Bassel et al., 1997; Saltstone

et al., 1994). In terms of convergent and concurrent validity, findings in different popula-

tions include correlations between the DAST and other measures of drug use and addiction

with medium to large effect sizes (e.g., Cocco & Carey, 1998; Evren et al., 2014), and small

to medium associations with mental health symptoms (e.g., anxiety, depression, thought

disorder), alcohol abuse, and work performance (e.g., Cocco & Carey, 1998; El-Bassel

et al., 1997; correlation coefficient effect size descriptors derived from Cohen, 1988). The

tool has demonstrated robust predictive power, with areas under the curve (AUCs) ranging

from .77 to .94 in the 20 to 28 item versions, predicting issues with substance use and cur-

rent diagnosis/identification of drug abuse/dependence in the studied populations (e.g., Y.-

T. Chen et al., 2020; Wolford et al., 1999).