Replication and External Validation of a Bi-Factor Parameterization of Attention Deficit/Hyperactivity Symptomatology

Abstract

This study evaluated the fit and criterion validity of a recently proposed bi-factor structure for attention deficit/hyperactivity disorder (ADHD) symptoms. Participants were 1,093 children, drawn from an ongoing prospective longitudinal study, whose ADHD symptoms were rated by parents and teachers when children were in 1st grade. The criterion validity of the bi-factor model was established using a range of school-based outcomes that included treatment utilization, teacher perceptions of the need for treatment, academic functioning, and peer and teacher relationship quality. Results indicated that a bi-factor model parameterization provided an equally good fit to parent, teacher, and combined reports of ADHD symptoms as did traditional 1-, 2-, and 3-factor models. However, in contrast to traditional models, the bi-factor parameterization acknowledged both the unity and diversity of ADHD symptoms. The general ADHD latent factor explained the vast majority of the observed variation in every symptom. Whereas the general ADHD latent factor was significantly associated with all 15 outcomes, the specific Inattentive factor explained unique variation in 9 (primarily the academically oriented) outcomes and the specific Hyperactive-Impulsive factor explained unique variation in 2 outcomes. The general ADHD factor was more strongly correlated with each of the observed ADHD symptom scores (total, inattentive, hyperactive-impulsive) than was either specific factor. Results are discussed with respect to how changes in the conceptualization of the factor structure correspond to recent changes to the diagnostic criteria for ADHD, as well as whether/how individual differences in inattention and hyperactivity-impulsivity might be used to differentiate children who are diagnosed with ADHD.

Acknowledgments

Support for this research was provided by the Eunice Kennedy Shriver National Institute of Child Health and Human Development grants 1R03HD071646 and P01 HD39667, with co-funding from the National Institute on Drug Abuse. The Family Life Project Key Investigators in the second phase of the study included Lynne Vernon-Feagans, Martha Cox, Clancy Blair, Peg Burchinal, Linda Burton, Keith Crnic, Patricia Garrett-Peters, Mark Greenberg, Roger Mills-Koonce, and Michael Willoughby.

Notes

Note: P = parent; ADHD = attention deficit/hyperactivity disorder; T = teacher; P/T = combined parent and teacher using “or” rule; WJ = Woodcock Johnson; C = child (achievement testing); LEB = learning, emotional or behavioral problems; IEP = Individual Education Plan.

Note: CFI = comparative fit index; RMSEA = root mean square error of approximation; CI = confidence interval.

^a N = 1,087.

^b N = 904.

^c N = 1,093.

Note: N = 1,093. IN = Inattentive; HI = Hyperactive-Impulsive.

*Nonsignificant (p > .05) factor loadings.

Note: N = 1,093. IN = inattentive; HI = hyperactive-impulsive; OR = odds ratio; CI = confidence interval; IEP = Individual Educational Plan; On Medicated = teacher report of whether the child is taking any medication for attention deficit/hyperactivity disorder (ADHD) or related difficulties; Benefit Meds = teacher impression of whether the child would benefit from such medications; LEB = learning difficulties, emotional problems or behavior problems; R-Services = teacher intends to refer the child for services; STRS = Student, Teacher Relationship Scale; WJ = Woodcock Johnson.

*p < .05. **p < .01. ***p < .001.