The Validity of Teacher Rating Scales for the Assessment of ADHD Symptoms in the Classroom: A Systematic Review and Meta-Analysis

Abstract

Objective: To assess attention-deficit/hyperactivity disorder (ADHD) symptoms in the classroom, most often teacher rating scales are used. However, clinical interviews and observations are recommended as gold standard assessment. This systematic review and meta-analysis evaluates the validity of teacher rating scales. Method: Twenty-two studies ( N = 3,947 children) assessing ADHD symptoms using teacher rating scale and either semi-structured clinical interview or structured classroom observation were meta-analyzed. Results: Results showed convergent validity for rating scale scores, with the strongest correlations ( r = .55–.64) for validation against interviews, and for hyperactive–impulsive behavior. Divergent validity was confirmed for teacher ratings validated against interviews, whereas validated against observations this was confirmed for inattention only. Conclusion: Teacher rating scales appear a valid and time-efficient measure to assess classroom ADHD; although validated against semi-structured clinical interviews, there were only a few studies available. Low correlations between ratings and structured observations of inattention suggest that observations could add information above rating scales.

Psychometric properties of an observation instrument to assess ADHD symptoms in the classroom using a continuous sampling approach

“…There was some evidence for divergent validity for the Verbal Hyperactivity scale. The validity findings for the Verbal Hyperactivity scale of the GUCCI-NL when compared to teacher rating scales are in line with other classroom observation systems included in a recent meta-analysis (Staff et al, 2020a). For Oppositional Behavior, convergent validity was established when validated against a teacher rating scale but not against a clinical interview.…”

“…These findings suggest that observed Attention Problems and Motor Hyperactivity measure different aspects of behavior in comparison to rating scales or interviews. However, weak divergent validity of the observational scales has been found for many observational instruments (see Staff et al, 2020a) and is partly explained by the high correlation between symptom domains of ADHD (Willcutt et al, 2012), that cannot be judged fully in isolation (see for a detailed discussion: Staff et al, 2020a). In addition, we compared observed motor and verbal hyperactivity but not impulsivity to combined hyperactivity/impulsivity scales of a teacher rating scale and clinical interview.…”

“…ADHD symptoms in the classroom have typically been assessed using teacher ratings on DSM-based standardized behavior rating scales/questionnaires, as well as (although used in a lesser extent) the use of semistructured clinical interviews. Both measures provide information on a child's behavior over extended time spans and in multiple situations, and the psychometric properties are reported to be adequate (American Academy of Pediatrics, 2011;Staff et al, 2020a;Tannock, Hum, Masellis, Humphries, & Schachar, 2002).…”

“…Psychometric properties of these measures are generally acceptable regarding inter-observer reliability and predictive validity, although the latter is usually only examined for the total scales and not for subscales and related symptom domains. Agreement between observations of ADHD symptoms and teacher reports (i.e., construct validity) is small to moderate (Minder et al, 2018;Staff et al, 2020a). These instrument seem to be able to differentiate between children with ADHD and typically developing peers in the classroom (Kofler, Rapport, & Matt Alderson, 2008;Platzman et al, 1992).…”

Academic Achievement in Spanish Secondary School Students: The Inter-Related Role of Executive Functions, Physical Activity and Gender

“…Third, data were collected using self-reporting methods or via information from third party informants. Although they are well-proven assessment methods in different domains of social, emotional and behavioral functioning [92,93], some limitations are found. To overcome the limitations of this type of measures, it would be interesting to complement the study using distinct data collection methods.…”