Negative Social Preference in Relation to Internalizing Symptoms Among Children with ADHD Predominantly Inattentive Type: Girls Fare Worse Than Boys

Abstract

Despite distinct peer difficulties, less is known about the peer functioning of children with attention-deficit/hyperactivity disorder (ADHD) predominantly inattentive type (ADHD-I) in comparison to the peer functioning of children with ADHD combined type. Our purpose was to examine whether child sex moderated the relations between negative social preference and internalizing/externalizing problems in children with ADHD-I. Participants included 188 children diagnosed with ADHD-I (110 boys; ages 7–11; 54% Caucasian). Teacher ratings of the proportion of classmates who “like/accept” and “dislike/reject” the participating child were used to calculate negative social preference scores. Children, parents, and teachers provided ratings of anxious and depressive symptoms, and parents and teachers provided ratings of externalizing problems. Boys and girls did not differ on teachers’ negative social preference scores. As hypothesized, however, the relation between negative social preference and internalizing symptoms was moderated by sex such that negative social preference was consistently and more strongly associated with internalizing symptoms among girls than in boys. In terms of externalizing problems, negative social preference was associated with teacher (but not parent) ratings, yet no moderation by child sex was found. Negative social preference is associated with teacher-report of externalizing problems for both boys and girls with ADHD-I, whereas negative social preference is consistently associated with girls’ internalizing symptoms across child, parent, and teacher ratings. Implications for future research and interventions are discussed.

Acknowledgments

This research was supported by a grant from the National Institute of Mental Health (NIMH) to Linda J. Pfiffner and Stephen P. Hinshaw (R01MH077671). The contents of this article do not necessarily represent the views of the National Institutes of Health and do not imply endorsement by the Federal Government.

Notes

Note. For sex, girls = 0, boys = 1. Externalizing is a composite of conduct problems/aggression composite. CP = conduct problems; PR = parent-report; SR = self-report; TR = teacher-report.

*p < .05. **p < .01. ***p < .001.

Note. For sex, girls = 0, boys = 1. For race, non–Caucasian = 0, Caucasian = 1. DV = Dependent Variable; ANX = anxiety; DEP = depression.

*p < .05. **p < .01. ***p < .001.

Note. For sex, girls = 0, boys = 1. DV = Dependent Variable; EXT = externalizing problems (conduct problems/aggression composite).

*p < .05. **p < .01. ***p < .001.

¹Because our primary aim was to examine child sex as a potential moderator of the relations between negative social preference and internalizing/externalizing symptoms, we did not initially consider covariate variables in the regression models as tests of statistical interactions are typically underpowered in nonexperimental research (McClelland & Judd, 1993). However, we did conduct supplemental analyses in order to determine if our results changed when potential covariates were included. Demographic/family characteristic considered as potentially important covariates were: child age, child race/ethnicity (dummy-coded 0 = non-Caucasian and 1 = Caucasian), child IQ, total family income, and highest parent education level. Child age, family income, and parent education were not significantly correlated with any of the child-, parent-, or teacher-reported psychopathology variables (all ps > .05). Race/ethnicity and IQ were correlated only with child-reported depressive symptoms (r = − .24, p = .001 and r = − .15, p = .04, respectively), and so the moderation model for this outcome was re-rerun while including race/ethnicity and IQ as covariates. In the final model, race/ethnicity remained significantly associated with child depressive symptoms (β = − .20, p = .006), and the Negative Social Preference × Sex interaction term was marginally significant (β = − .22, p = .053). Consistent with results from the model without covariates, there was a significant relation between negative social preference and self-reported depressive symptoms for girls (β = 0.24, p = .03), but not for boys (β = − 0.04, p = .65). Therefore, the consideration of various child and family characteristics as potential covariates did not alter our primary findings.