Factor Structure, Measurement Invariance, and Scoring Practices of the Strengths and Weaknesses of ADHD-Symptoms and Normal Behavior

We investigate the factor structure, concurrent validity, internal consistency, measurement invariance (sex and parents' geographical/cultural background), and practical scoring practices of the Strengths and Weaknesses of Attention-deficit/Hyperactivity Disorder (ADHD)-Symptoms and Normal Behavior scale (SWAN) in junior high school students. With a sample of 650 parents of Mexican junior high school students (55.84% girls; mean age of 13.15 and SD = 0.97 years) who completed the SWAN scale, a bidirectional instrument, and the Barrios and Matute Questionnaire of ADHD symptoms (BMQ-ADHD), a traditional Mexican unidirectional instrument, we obtained SWAN' psychometric properties by a series of confirmatory factor analyses. Two and three-correlated factors of the SWAN and bifactor models fitted well to the data. The bifactor model with one general dimension and two specific dimensions (BF 2S) showed most favorable psychometric properties and was invariant regarding sex and cultural background. The analyses of the BF 2S revealed that only the general factor was sufficiently reliable for scoring. Percentiles divided by sex better approximated SWAN averaged-based scores to the general factor derived from the SWAN bifactor model, measuring the trait with acceptable precision for norms development. The study provides evidence that SWAN measures a trait best represented by a reliable general domain, that the average-based SWAN score is closer to the general SWAN trait when the score is segregated by sex, and that it is invariant with respect to sex and cultural background to be reliable in a culturally diverse population.

Automated summary

In this study, we investigated the factor structure, validity, consistency, measurement invariance, and scoring practices of the SWAN scale in junior high school students. The results indicated that the bifactor model with one general dimension and two specific dimensions showed the best psychometric properties and was invariant across sex and cultural background. Additionally, dividing scores by sex better approximated the general factor derived from the SWAN bifactor model. The study provides evidence for the reliable use of the SWAN scale in a culturally diverse population.