Factor Structure and Invariance of an Adolescent Risks and Needs Assessment

Abstract

The Washington Assessment of Risks and Needs of Students is a youth self-report screening instrument developed for use by high school and juvenile court personnel faced with the legal and practical challenges of high truancy and dropout rates. It purports to measure six facets of risks and needs of youth relevant to improving school outcomes. In this study, a bifactor model measuring a general factor and six specific factors was examined for fit and invariance across different groups defined by sex and race/ethnicity, with a sample of court-petitioned high school students ( N = 937; ages 13-17 years). The results of multigroup confirmatory factor analysis revealed an essentially invariant bifactor structure across the groups. Further analysis of reliability support the use of the general factor to guide decision making for youth at risk for truancy and school failure, and scores deriving from the six specific factors as providing insight on specific areas of risk and need.

Equivalence Testing of a Youth Risk and Needs Assessment

“…A bifactor model met fit criteria, except for the CFI for the African American sample. However, we continued with the bifactor model, as it was consistent with the WARNS use and practice (Strand et al, 2019). Table 2 provides reliability estimates by group, showing much variability.…”

“…The WARNS is a 40-item self-report instrument with six specific factors (i.e., aggression-defiance, depression-anxiety, substance abuse, peer deviance, family environment, and school engagement) and a general factor supported via a bifactor model (Strand et al, 2019). Response options range from 0 = Never or hardly ever to 3 = Always or almost always .…”

“…First, a bifactor model, supported by evidence and use (Strand et al, 2019) was fit for each group with diagonally weighted least squares estimation with the lavaan package in R. Model fit was assessed via (a) fit criteria, χ 2 test, comparative fit index (CFI ≥ .90), and root mean squared error of approximation (RMSEA ≤ .08; Brown, 2015), (b) out of bound parameter values, and (c) theory for interpretation and use. Internal consistency reliability was estimated via omega (ω; McDonald, 1999), ω hierarchical (ω H ), and ω hierarchical subscale (ω HS ; Rodriguez et al, 2016) to highlight reliability differences when accounting for the general factor.…”

Improving the measurement of effectiveness in the Irish Youth Justice System

“…• The Washington Assessment of the Risks and Needs of Students (WARNS) is a self-report survey designed to collect, identify and synthesise information relevant to youth (13-to 18-year-old school students) who may be at the early stages of engagement with juvenile justice and child welfare services (Strand et al, 2017). The survey gathers data in relation to aggression and defiance, depression and anxiety, substance misuse, peer deviance, family environment, and school engagement and measures experiences related to healthy social, emotional, and educational development (Strand et al, 2017).…”

Differential Item Functioning of a Truancy Assessment

“…Given score use and race/ethnic differences in truancy and related behaviors, providing score validation evidence is essential for measures such as WARNS (American Educational Research Association, American Psychological Association, & National Council on Measurement in Education, 2014; Kane, 2013). Some validity evidence does exist for the WARNS including internal factor structure (e.g., Strand, Gotch, French, & Beaver, 2017) and latent profiles of risk and needs (e.g., Iverson, French, Strand, Gotch, & McCurley, 2016), yet there lacks evidence of item invariance across groups for which the instrument is widely used. Differential item functioning (DIF) refers to when the probability of endorsing a certain item response is unequal for different groups who possess the same ability level (Zumbo, 1999).…”