Poster #47 - Longitudinal invariance for parents and teachers SNAP-IV ratings of ADHD: Evidence from a community-based sample

• In Event: 1-115 - Poster Session 03
  In Poster Session: PS 03 Section - Developmental Disabilities

Integrative Statement

Background: According to DSM-5, attention-deficit/hyperactivity disorder (ADHD) is a neurodevelopmental condition encompassing symptoms of inattention, hyperactivity, and impulsivity which interfere on person’s daily functioning and its development. Although the ADHD categorical classifications are developmentally unstable, especially during preschool age, the ADHD symptom dimensions are largely stable. To demonstrate that a construct can be reliably measured at different time points, it is necessary to test its stability over construct items’ and their variances. Objective: The present study aims to investigate temporal measurement invariance for the inattention and hyperactivity/impulsivity (HI) scales of the short form of the SNAP-IV in a community-based sample of Brazilian preschoolers. The SNAP-IV is a DSM-based questionnaire composed by 26 items, 18 of which pertaining to the ADHD scales. It presents two versions, for teachers and parents respondents. Methods: The sample was composed by 578 preschoolers nested in 27 schools from a 6-months clinical trial, which were randomly assigned to one of the following groups: executive function training, language training, or control group. Data from the baseline and after six months intervention evaluations were used. We tested the longitudinal invariance of SNAP-IV with 18 items considering the original ordinal items as dichotomous. We dichotomized the items to reduce the number of thresholds to be constrained following the diagnostic criteria for ADHD, where 0 and 1 scores were grouped and named as 0 and scores 2 and 3 were grouped and named as 1. In terms of measurement model, for both versions, we used SNAP-IV correlated-factor solution with two-correlated factor (attention, constituted by nine items; HI constituted by nine items). WLSMV-estimator was used considering the non-independence of the preschoolers across the 27 schools adjusting, then, to the obtained standard errors for the complex design. Three levels of invariance were tested: configural, metric, and scalar. The analysis was controlled for the groups of intervention. Results: Across the three different constrains, the model for parents and teachers had excellent fit indices, achieving scalar invariance (where factor loadings and thresholds constrained to be equal across time). Moreover, the constrained did not worse the model between metric and scalar model, as ΔCFI for parents and teachers was zero, which are below than the suggested by Maede et al. (2008) of 0.002. The results attest for the invariance of the SNAP-IV across the time in both versions for the preschooler’s sample. Conclusion: Model parameters for SNAP-IV are invariant across time points and these results provide critical information about the psychometrics properties of the instrument. The results have important implications for experimental and longitudinal studies, as it is possible to distinguish between changes in scores due the effect of interventions or time from that ones due the symptom’s different relationships with the factors at the second time point. Implications for clinical practice are discussed.

Authors

• Patricia Silva Lucio, Universidade Estadual de Londrina (State University of Londrina)
  Presenting Author

• Hugo Cogo, Freie Universität Berlin
  Non-Presenting Author

• Guilherme V. Polanczyk, Universidade de São Paulo
  Non-Presenting Author