Psychometric Properties of ADHD Symptoms in Toddlers

Abstract

The purpose of this study is to examine the psychometric properties of the 18 Diagnostic and Statistical Manual of Mental Disorders (DSM) symptoms of attention deficit/hyperactivity disorder (ADHD) in 2-year-old children. ADHD is typically diagnosed in elementary school, but research suggests that many children with ADHD first show symptoms during the toddler years. An important first step in identifying toddlers who are at high risk for developing ADHD is to better understand the properties of DSM symptoms of ADHD in young children. Parents of 2-year-old children (N = 434; 240 boys, 194 girls; M = 29.07 months, SD = 3.42) across the United States were recruited online through Amazon’s Mechanical Turk to complete surveys about their children’s ADHD symptoms and temperament. Confirmatory factor analysis showed that 2- and 3-factor models fit well. All but one verbal hyperactive/impulsive symptom loaded highly on their respective factors. Relations between ADHD symptoms and temperament traits provided some support for convergent and divergent validity of the symptoms. Finally, item response theory analyses showed that items showed moderate to high levels of discrimination between toddlers with high and low levels of ADHD symptoms. Psychometric properties in 2-year-old children were generally comparable to those in older children, with similar factor structure, good reliability and validity, and good though somewhat lower discrimination, particularly for verbal symptoms. The study provides support for conducting prospective studies to determine whether these symptoms have utility for identifying at-risk toddlers.

ACKNOWLEDGMENTS

We acknowledge Craig Wells, Ph.D., and Alice Coyne, M.S., University of Massachusetts Amherst, who each provided statistical consultation.

Notes

¹ To be consistent with the larger literature on ADHD, we use the term “symptom” to refer to any of the behavioral features that characterize ADHD (e.g., “fidgets”). Note that it is not clear whether these behaviors at this age reflect early symptoms of the disorder, precursors to the disorder, or part of normal development.

² There were 854 invited to the survey, but some were not U.S. residents (N = 17), did not consent (N = 44), stopped partway through and did not submit the survey for payment (N = 38), or were duplicate participants (N = 47; an early error we rectified by changing the MTurk settings).

³ This model was constructed by setting the correlation between the two factors to one so that the two-factor and one-factor models were nested, allowing for direct model comparison between these factor models.

⁴ For this set of analyses only, if the convergent and divergent subscales were inversely related, the convergent subscale was reverse scored so that the direction of the relation to H/I or ATT was the same.

⁵ Several items on the temperament scale that are highly similar to ADHD symptoms were removed (Items 12, 75, 76, 78, 95, 44, 45) and analyses were repeated. Results were highly similar, with all three temperament domains showing significant relations with ADHD symptoms in the expected direction.

⁶ Several items on the temperament scale that are highly similar to ADHD symptoms were removed (Items 12, 75, 76, 78, 95, 44, 45) and analyses were repeated. Results were similar. Most temperament subscales showed the same relation with ADHD symptoms, with one exception: Higher Perceptual Sensitivity was now significantly related to ATT (b = −.08, SE = .03, p = .006).