J Abnorm Child Psychol (2012) 40:569–581

DOI 10.1007/s10802-011-9586-3

Predictors of Boys’ ADHD Symptoms from Early

to Middle Childhood: The Role of Father–Child
and Mother–Child Interactions
Louise J. Keown

Published online: 26 October 2011

# Springer Science+Business Media, LLC 2011

Abstract This prospective 3 year longitudinal study investi- parenting factors on the development of ADHD behaviors
gated preschool paternal and maternal parenting predictors of (Johnston and Mash 2001). Current developmental psycho-
Attention-Deficit/Hyperactivity Disorder(ADHD) in a com- pathology models that seek to predict how ADHD
munity sample of 93 school-age boys. Participants were characteristics develop over time propose a number of
recruited on the basis of inattention-hyperactivity at age 4 and possible pathways with genes and environment interacting
fathers and mothers were observed interacting with their sons. in a multitude of ways, leading to variations in age of onset,
Teachers, fathers, and mothers reported children's ADHD symptom expression, severity, and developmental course of
symptoms and impairment. Results from dimensional anal- hyperactive behaviors (i.e., overactivity, inattentiveness,
ysis showed that less observed paternal sensitivity and and impulsivity). There is also agreement that these
maternal positive regard predicted higher levels of inatten- developmental pathways involve transactional processes
tiveness in middle childhood, and that intrusive paternal whereby the child and parent reciprocally shape each
behavior was predictive of hyperactive-impulsive behavior at other’s behavior over the course of time (Johnston and Mash
school. In categorical analysis, less maternal warmth and 2001; Sonuga-Barke et al. 2005). Yet current knowledge
sensitivity were predictive of later ADHD. These predictions about potentially contributing parenting factors is based
held after statistical adjustment for the effects of preschool mainly on cross-sectional studies, conducted during the
ADHD behaviors and conduct problems. At follow-up, school-age years (7–12), which provide little insight into
parents of boys with ADHD reported more negative child- how associations between parenting and ADHD develop
parent relationship perceptions than comparison parents. over time (Deault 2010; Johnston and Mash 2001). A key
Findings highlight the importance of examining responsive developmental period, for which are there few longitudinal
parenting behaviors of both fathers and mothers in relation to studies of parenting and ADHD, is early to middle childhood
multi-informant ratings of ADHD symptoms. (age 4–7 years), when the transition to school places new
demands on children’s self regulatory skills.
Keywords ADHD . Parent–child interactions . Fathers . An important reason for further research spanning this
Mothers age range is the growing recognition that ADHD symptoms
often emerge in the preschool years and persist into middle
childhood and beyond (Lahey et al. 2005; Pierce et al.
While difficulties in the family relationships of children 1999; von Stauffenberg and Campbell 2007), placing
with ADHD are well documented (for reviews see Deault children at risk for a range of negative outcomes including
2010; Johnston and Mash 2001), there is a paucity of social dysfunction, academic underachievement, delinquen-
longitudinal research that examines the influence of cy, and substance abuse (Barkley 2006). From an early
intervention perspective, a key task is to predict which
L. J. Keown (*) preschool children who show atypical patterns of inatten-
Faculty of Education, University of Auckland,
tiveness and hyperactivity will go on to develop ADHD
Private Bag 92601, Symonds St.,
Auckland 1150, New Zealand (Sonuga-Barke et al. 2005). Several maternal parenting risk
e-mail: [email protected] factors for the continuation of hyperactive symptoms from
570 J Abnorm Child Psychol (2012) 40:569–581

early to middle childhood have been identified, including Furthermore, in order to gain a better understanding of
intrusive and overstimulating behaviors (Carlson et al. 1995), the potential role of paternal and maternal responsiveness in
negative affect, and directive control (Campbell and Ewing the maintenance of ADHD symptoms several key measure-
1990; Campbell et al. 1996; Peris and Baker 2000). While ment issues need to be addressed. First, greater clarification
these studies suggest some elements of responsive maternal is needed about the extent to which associations between
parenting that contribute to the maintenance of ADHD ADHD and parenting behaviors may be explained by
behaviors, the role of responsive paternal parenting in early comorbid conduct problems (Johnston and Mash 2001).
childhood, as a risk factor for the continuation of hyperactive Recent cross-sectional evidence (Ellis and Nigg 2009)
behavior problems, remains to be examined. suggests that dimensions of ADHD (hyperactivity, inatten-
Some existing longitudinal research supports a link tiveness) have partially distinct parenting correlates, even
between an unresponsive style of paternal parenting and with comorbid conduct problems controlled, and that these
poor child self regulatory skills. Among a community associations may differ for mothers and fathers. In their
sample of families followed from early to middle child- study, Ellis and Nigg (2009) found that for fathers, but not
hood, Denham et al. (2000) found that fathers’ lack of mothers, low self-reported parental involvement and incon-
observed proactive parenting (consisting of supportive sistent discipline were related uniquely to child inattention,
presence and provision of a structured environment) at but not child hyperactivity, when conduct problems were
age 4 predicted externalizing problems at age 9. Converse- controlled. However, these types of relationships need
ly, observed paternal sensitivity (including emotional further evaluation with longitudinal designs that examine
support, lack of hostility, and respect for the child’s links between other parenting variables and ADHD
autonomy), in early childhood contributed uniquely to dimensions.
lower levels of children’s externalizing scores in middle Second, further insight into possible reasons for different
childhood (NICHD Early Child Care Research Network associations between ADHD and parenting for mothers and
2004). These findings for fathers add to the body of fathers may be provided by including a focus on informant
knowledge about the role of sensitive, supportive parenting specific ADHD dimensional outcomes. Research to date
in the development of children’s self regulation skills and that has examined the relations of ADHD symptoms to both
support the possibility that “parental difficulties in syn- paternal and maternal parenting has used two different
chronizing their actions to the child’s needs may be the approaches to measuring ADHD. Several studies (Gadeyne
mechanism that accounts for the development of disinhi- et al. 2004; Lifford et al. 2008) have used the attention
bited, poorly regulated behavior in some children” (Johnston problems subscale of the Child Behavior Checklist (CBCL;
and Mash 2001, p. 185) who develop ADHD. However, Achenbach 1991), which combines items from the behav-
these two studies concentrated on externalizing problems ioral domains of inattention and hyperactivity−impulsivity.
rather than symptoms and behaviors that relate to ADHD. However, when CBCL reports of ADHD symptoms are
Accordingly, fathers’ observed responsiveness was investi- obtained from both parents, different constructs may be
gated in the present study focusing on its relationship with tapped by mother ratings compared to father ratings
child ADHD symptoms. (Lifford et al. 2008). A second approach has combined
Emerging evidence also highlights the importance of information from parent and teacher reports to define
considering the influence of both maternal and paternal symptom dimensions (Ellis and Nigg 2009). While this
parenting on ADHD, given findings from two middle approach is consistent with the Multimodal Treatment
childhood community samples of different associations for Study of Children with ADHD design (Hinshaw et al.
mothers and fathers. For example, Lifford et al. (2008) found 1997), it may obscure unique patterns of parenting
that child reported feelings of rejection in the father−child variables that might be associated with different informant
relationships predicted later father rated ADHD symptoms, ratings of each ADHD dimension. There is some evidence
with no such relationship between mother−child relation- that parenting correlates of ADHD subtypes vary by mother
ships and mother rated ADHD symptoms. Whereas, in the and teacher defined groups (Gadow et al. 2004), but it is
study by Gadeyne et al. (2004), mothers’ reports of not known whether parenting correlates also distinguish
restrictive control predicted later teacher reports of father defined groups. However, recent findings of considerable
attention problems, however, there was no evidence of disagreement between maternal and paternal ratings of ADHD
fathers’ effects on children. These disparate findings specific symptoms (Langberg et al. 2010), suggest that it may
may be partly due to differences in the type of be a possibility.
parenting constructs measured in each study. While these The purpose of the present study was to examine
results are informative for the present study, the nature of the whether paternal and maternal parenting in early childhood
parenting dimensions assessed differed from the observed predicted boys’ ADHD symptoms in middle childhood over
parent−child responsiveness behaviors measured in this paper. and above the stability of these behaviors and the influence
J Abnorm Child Psychol (2012) 40:569–581 571

of early conduct problems. In particular, the study exam- approaches. For the dimensional approach, in order to
ined three theoretically important aspects of responsive discern potential differences in the pattern of relationships
parenting: observed sensitivity (parent is “tuned in” to, and by informant, parenting predictors of ADHD symptom
appropriately responsive to the child’s cues, interests and dimensions were examined separately for teacher, mother,
mood); warmth (demonstrations of affirmation and affec- and father ratings. For the categorical analysis, maternal
tion towards the child); and intrusiveness (over control of and paternal predictors of ADHD group status, formed on
the child’s play instead of allowing for child preferences). the basis of symptoms and impairment, were examined.
Given evidence that sensitivity and intrusive parenting may Given the study focus on parental responsiveness, a second
contribute independently to developmental outcomes aim of the study was to compare parental perceptions of the
(Keown et al. 2001; Pungello et al. 2009), and that effects child−parent relationship of children who did and did not meet
may vary by parent gender (Cabrera et al. 2007), the unique criteria for ADHD in middle childhood. There is some
association between each of these parenting variables on evidence from clinical samples that parents of children with
ADHD outcomes was investigated for both fathers and ADHD have more negative perceptions of the parent−child
mothers. Warmth was also examined separately given cross- relationship quality than comparison parents (Gerdes et al.
sectional findings linking preschool inattention−hyperactivity 2007), and that these perceptions are similar for mothers and
with observed parental warmth (Goldstein et al. 2007; Keown fathers of children with ADHD (Gerdes et al. 2003). One of
2011). A new contribution of this study was the inclusion of these studies also found less perceived warmth in the
observed paternal parenting as a predictor of later ADHD parent−child relationship in children with ADHD on the
behaviors. To date, the few longitudinal studies relating to basis of both paternal and maternal ratings (Gerdes et
parenting and ADHD across this age range include maternal al. 2007), while in Gerdes et al. (2003) this finding
observations only (Campbell and Ewing 1990; Campbell et applied to mothers only. However, few, if any studies have
al. 1996; Carlson et al. 1995). The hypothesis tested in this investigated paternal and maternal perceptions of child−parent
study was that boys’ ADHD symptoms in middle childhood relationships in community samples of school-age children
would be associated with less observed parental warmth and with ADHD symptoms, who were followed up from
sensitivity and higher rates of observed parental intrusiveness preschool. As clinic-referred cases of hyperactivity have been
in early childhood, after statistical adjustment for the effects differentiated from community-identified cases of hyperactivity
of preschool ADHD symptoms and conduct problems. by less optimal parenting practices (Woodward et al. 1997), a
Given the state of the literature, specific predictions were community sample was chosen to identify the parenting
not made about differences in these aspects of responsive correlates of ADHD symptoms given the potential bias
parenting for mothers versus fathers. However, both theory associated with clinic referral.
and past research suggest that, particularly in early
childhood, fathers and mothers may have distinct and
complementary parenting roles that contribute uniquely Method
to the prediction of some child outcomes (Grossmann et
al. 2002; NICHD ECCRN 2004). For example, drawing Participants
on Bowlby’s theory that fathering involved a greater
emphasis on mentoring and play than on nurturing Participants for this study were recruited as part of a follow-
interactions, Grossmann et al. (2002), argue that fathers’ up study of boys’ behavioral development that began when
relationships with their children may be particularly the children were 4 years-old. One hundred and ten 4 year-
important for supporting the development of child com- old boys and their fathers and mothers were initially
petence outside the family, such as adjustment to the recruited for the study through preschools using a two-
demands of the classroom. During the preschool years step procedure. Children in the study came from 47
sensitive parenting during father−child play interactions preschools located in a range of socioeconomic areas (low
are also likely to be particularly important in the SES: 22.8%; mid SES: 38.1%; high SES: 39.1%) within the
development of children’s attentional and self-organizing Auckland (New Zealand) urban area. Details of recruitment
skills (Sonuga-Barke et al. 2006). Furthermore, research and selection are provided in Keown (2011). Briefly, letters
suggests that fathers are more likely to be involved with containing a consent form and the Strengths and Difficul-
sons (Pleck 1997) who are at higher risk than daughters of ties Questionnaire (SDQ; Goodman 1997) were sent home
developing ADHD. to fathers and mothers of 4 year-old boys to complete.
Based on the stance that ADHD is best viewed using Parents were asked permission for their son’s teacher to
both a dimensional and a categorical approach (Hinshaw et complete an SDQ on their son’s behavior at preschool.
al. 1997), relationships between responsive parenting and Completed SDQ questionnaires were returned by 373
ADHD symptoms were examined using both of these parents, with 290 parents agreeing to further contact about
572 J Abnorm Child Psychol (2012) 40:569–581

the next stage of the study. Invitations to participate in a The occupational groupings of the families were distrib-
home visit were sent to 204 parents, consisting of parents of uted across all occupational categories (Elley and Irving
all potential hyperactive group boys (n=79), 81 potential 2003), with most fathers employed in either specialist
comparison group children, and 37 other boys who were professional/high level management (28%), professional/
borderline for inclusion in the hyperactive group. One managerial (28%), or technical/administrative (29%) occu-
hundred and sixteen parents consented to participate in this pations. The majority of mothers either worked part-time
next stage of the study. Three of these parents subsequently (48%), or were not in paid employment outside the home
withdrew and three proved to be unsuitable for inclusion in (27%). On average, mothers had slightly higher levels of
the study for reasons such as insufficient English compre- tertiary education (mean years = 3.14, SD = 2.09) than
hension to complete questionnaires and observations. fathers (mean years=2.64, SD=2.16). Maternal age ranged
Comparisons between parents who did and did not consent from 29 to 50 years (M=39.79, SD=4.85) and from 31 to
to participate in the next stage of the study or who 55 years for fathers (M=41.39, SD=4.83).
withdrew/were unsuitable, did not differ significantly in
terms of father, mother, and teacher rated SDQ scores and Procedure
the SES rating of the preschool attended by the children.
The second step involved an interview with each child’s At Time 1, home visits were made to each child’s mother
primary caregiver using the Parental Account of Children’s and father on separate occasions, to collect questionnaire
Symptoms (PACS; Taylor et al. 1991), to reconfirm the and interview data on parenting and child behavior and
group status of each child. On the basis of the questionnaire videotape parent−child observations. During one of the
and interview ratings, boys were assigned to hyperactive home visits (usually to the mother), the child was
(n=41) and comparison (n=38) groups. Thirty one boys who administered the British Picture Vocabulary Scale (BPVS
did not meet criteria for inclusion in either group were II; Dunn et al. 1997), a standardized measure of children’s
retained in the study to provide a full range of ADHD receptive language skills. Details of these measures can be
symptoms for analysis from a dimensional perspective. Only found in Keown (2011). Only the interview measures of
boys were included because the prevalence of hyperactive child conduct problems, the BPVS and the parent−child
behavior problems is higher in boys than in girls (Taylor et observations are discussed in this paper.
al. 1991) and the constraints of the study made it impractical The follow-up (Time 2) consisted of two home visits
to screen the numbers of children needed to obtain a large during which mothers and fathers were interviewed
enough community sample of boys and girls to analyse data separately and completed questionnaire measures on child
by both parent and child gender. behavior and functioning. Seven families who had moved
Approximately two−and−a−half years after the initial overseas or to other parts of the country, provided data by
assessment, families were asked to participate in a follow- questionnaire only. In addition, teachers were asked to
up study (Time 2). Of the original 110 children and their complete questionnaires about participant boys’ behavior
parents, 11 families refused to continue in the study and and social relationships at school. Teacher questionnaires
five families could not be located. Comparisons were made were returned by teachers of 83 (89%) boys. It should be
between these 16 families and the 94 families who took part noted that these were different teachers to those who
in Time 2, on the basis of family background, parent, and participated in the preschool assessments of child behavior.
child characteristics. Families who did not continue had a At both time points trained interviewers who were unaware
significantly lower total annual income at Time 1 (NZ$40- of the children’s initial group status administered all child
50,000 pa compared to NZ$50-70,000 pa for continuing and family measures.
families); however, there were no other differences between
participants and non-participants. Of the 94 families in the Measures
study at Time 2, two fathers declined to participate. The
Time 2 participants also included three parents (two fathers, Adults’ Reports of Children’s ADHD Symptoms, Conduct
one mother) who took part in the study at Time 2, but not at Problems, and Impairment
Time 1. Thus, the number of parents who contributed data
to this study was 89 fathers and 93 mothers. Time 1 Parent and teacher ratings on the Inattention−
Children retained in the follow-up study had an average Hyperactivity subscale of the SDQ (Goodman 1997) were
age of 7 years (SD=9.43 months). The majority of the boys used to initially identify boys for the study, and in this
were of New Zealand European descent (79.8%), with paper, to control for initial symptoms levels when predict-
smaller numbers from Maori (7.4%), Samoan (3.2%), ing later ADHD outcomes. The SDQ has demonstrated
Indian (3.2%), and non-NZ Caucasian and other ethnicities sound psychometric properties in large community samples of
(6.4%), and were primarily from intact families (n=90). young children, with the Inattention−Hyperactivity subscale
J Abnorm Child Psychol (2012) 40:569–581 573

correlating strongly with clinical assessments of ADHD over time may be the rule rather than the exception” (Lahey
(Hawes and Dadds 2004). In the present study (N=94) et al. 2005, p. 901), and given the extent of cross-situational
Inattention−Hyperactivity scores were significantly cor- impairment. The modified ADHD criteria enabled the
related for parents (mothers and fathers) and teachers (r=0.39, sample size of the ADHD group, and hence power, to be
p<0.001), and mothers and fathers (r=0.78, p<0.001). increased. Using these criteria, 24 boys (five with five
Independent interviewer ratings of inattention, hyperac- symptoms, 19 with six symptoms), were assigned to the
tivity, and conduct problems were obtained using the PACS ADHD group. Twenty nine boys who had three or fewer
(Taylor et al. 1991). The PACS ratings of conduct problems symptoms were assigned to the control group. Consistent
were utilized in this paper to control for the influence of with other ADHD follow-up studies from early to middle
early conduct problems in predicting later ADHD symp- childhood (Lahey et al. 2005), symptoms (a rating of
toms. The PACS is a standardized semi-structured interview “often” or “very often” on the ADHD Rating Scale-IV)
measure of child behavior across a range of situations were deemed to be present if reported by either parents or
(home, shopping, outings). Parents are asked detailed teachers.
questions about hyperactive and inattentive behaviors, and Child impairment was assessed using the SDQ impact
conduct problems such as aggression, temper tantrums, supplement (Goodman 1999), the SDQ Peer Problems
lying, disobedience and destructiveness over the previous subscale, and PACS interviewer (Taylor et al. 1991) ratings
6 months. Using operationalized criteria, the interviewer of child functioning at home and school. Parents and
rates parental responses in terms of the frequency and teachers were asked to rate each participant child on the
severity of each child’s hyperactive and conduct problem SDQ impact supplement (Goodman 1999), which assesses
symptoms. The PACS has good construct validity and high whether the respondent thinks the child has a problem with
reliability with interrater correlations in the current study concentration, behavior, or being able to get on with
ranging from 0.91 to 0.99. PACS hyperactivity scores were people. If so, parents were asked to indicate which areas
significantly correlated with SDQ hyperactivity scores for their child had problems with and further questions, rated
mothers (r=0.71, p<0.001) and fathers (r=0.65, p<0.001). on a 4-point Likert scale, were asked about the chronicity,
distress to the child, social impairment, and burden to
Time 2 Mothers, fathers, and teachers completed the others, associated with the problem(s). The social impair-
ADHD Rating Scale-IV (DuPaul et al. 1998) to measure ment questions cover classroom learning, peer relations and
the frequency of the target child’s ADHD symptomology. friendships, home life, and leisure. These areas of func-
The ADHD Rating Scale-IV consists of 18 DSM specific tioning are known to be impaired in children with ADHD.
items that are rated on a 4-point Likert scale ranging from Items on distress and social impairment can be summed to
never or rarely (0) to very often (3). This scale was chosen generate an impact score, or scores can be classified as
as it taps age-appropriate behaviors and tasks across both normal, borderline, or abnormal using SDQ “caseness”
home and school contexts of relevance to 7 year-old guidelines (Goodman 1999). SDQ impact scores have been
children (e.g. finishing school work, remaining seated in shown to discriminate between clinic and community
the classroom, completing tasks, such as homework, that subjects better than total symptoms scores (Goodman
require sustained mental effort). The ADHD Rating Scale- 1999).
IV has been used extensively with school-age children and The SDQ Peer Problems subscale was completed by
has demonstrated good reliability and validity (DuPaul et each child’s mother, father, and teacher. The subscale
al. 1998). In this study the ADHD rating scale total score consists of five items (solitary, bullied, unpopular, lacks a
demonstrated very good internal validity (teacher α=0.95; good friend, gets on better with adults) that are rated on a 3-
mother α=0.93; father α=0.90). Relationships between point Likert scale to indicate how much each behavior
parent and teacher ratings on each subscale were in the applies to the child. SDQ scores can be used as continuous
moderate range, as follows: Inattention 0.42 (father/teacher), variables or scores can be classified as normal, borderline,
0.44 (mother/teacher), 0.59 (father/mother); Hyperactivity/ or abnormal to identify likely cases with mental health
Impulsivity 0.50 (father/teacher and mother/teacher), 0.59 disorders, using normative SDQ guidelines (Goodman
(father/mother) (all p’s<0.001), similar to the parent/teacher 2001). In this paper “caseness” scores were used to identify
agreement reported by DuPaul et al. (1998). children with impaired functioning on the basis of SDQ
For the purpose of categorical analysis, boys were said impact and peer problems scores.
to meet modified criteria for ADHD if they had at least five As part of the follow-up interview with each child’s
symptoms in one of the ADHD symptom domains and primary caregiver (83 mothers and four fathers of those
showed impairment across two settings. A cut-off of five parents interviewed), semi-structured questions based on
symptoms was based on evidence “that fluctuations above the PACS interview (Taylor et al. 1991) were asked about
and below the boundaries of nominal diagnostic categories concerns with the child’s behavior, school work, and
574 J Abnorm Child Psychol (2012) 40:569–581

getting on with others at home and school (Heptinstall Scale (CPRS; Pianta 1992). The CPRS was designed to
1993). On the basis of parental descriptions, interviewer assess the parent’s report of the study child’s attachment to
ratings of child behavior problem severity were made on a the parent and items tap the parent’s feelings and beliefs
4-point scale from no problem (0) to definite and marked about his/her relationship with the study child, and about
problem (3). the child’s behavior toward the parent. Items are rated on a
Cross-situational impairment was defined as (a) parent 5-point scale from 1 (definitely does not apply) to 5
(mother or father) reported problems at home or with peers (definitely applies). Example questions include “My child
according to SDQ impact or SDQ peer problems scores that easily becomes angry with me” (conflict with child scale);
were in the borderline or abnormal range, or interview “I share an affectionate, warm relationship with my child”
ratings of definite problems getting on with others away (closeness with child scale). Internal consistency for the
from school; (b) parent (mother or father) SDQ impact current sample was α=0.83 (fathers) and 0.85 (mothers) for
report of problems with classroom learning in the border- parent−child conflict, and α= 0.76 (fathers) and 0.75
line or abnormal range, or interview ratings of definite (mothers) for parent−child closeness.
problems getting on with others at school, or teacher
reported problems at school according to SDQ impact or
SDQ peer problems scores in borderline/abnormal range. Data Analytic Plan
Mothers, fathers, and teachers also completed the SDQ
Conduct Problems subscale. In this paper teacher SDQ The results of the study are presented in three stages. First,
ratings were used to control for conduct problems in the bivariate correlations between predictors and criteria are
analyses for parent perception of the child−parent relationship presented. Hierarchical multiple regression analyses were
at Time 2. then used to investigate the relationship between observed
parenting during the preschool years to later child ADHD
symptoms (as measured by continuous scores on the
Parent−Child Interactions and Relationships ADHD Rating Scale-IV), controlling for initial levels of
these behaviors and early conduct problems. As Time 1
Time 1 Father−son and mother−son dyads were observed in child BPVS scores were negatively correlated with teacher
separate home visits, during 10 min of semi-structured free ratings of inattentiveness (r=−0.21, p=0.027), receptive
play. Visits to each parent were counter-balanced to reduce language skill was included as a covariate in models
order effects. These play sessions, which were videotaped predicting teacher rated inattentiveness. Last, taking a
for later coding, involved giving each parent−child dyad categorical approach (ADHD and comparison groups),
three bags of toys to play with. Three parenting scales were these relationships were also examined using a series of
examined: sensitivity (parent is “tuned in” to, and appro- logistic regression analyses.
priately responsive to the child’s cues, interests and mood), Following preliminary analysis of the data, to meet
intrusiveness (physical or verbal overcontrol of the child’s linear regression requirements, Time 1 mother and teacher
play), and positive regard (demonstrations of affirmation, SDQ scores and ADHD outcome variables were trans-
warmth, and affection toward the child). Each of these formed using square root transformations to normalize the
behaviors was coded using 7-point scales (1 = very low, 7 = distribution of each variable. To reduce the impact of
very high) by extensively trained raters blind to all other univariate outliers on the Time 1 observed parenting
information about the dyads. Coders noted relevant behav- variables, as recommended by Tabachnick and Fidell
iors observed during interactions and made global ratings (2007), changes were made to scores on the outlying cases
after several viewings of the entire play session. The to make them less deviant.
observational protocol and coding scheme were adapted
from the Early Head Start Research and Evaluation Project
“Three Bag” assessment (Brady-Smith et al. 2000). Inter- Results
observer reliabilities, based on gamma coefficients for
ordinal data, were 0.80, 0.93, and 0.94 respectively for Bivariate Correlations Between Parenting Behaviors
parental intrusiveness, positive regard, and sensitivity. and Child ADHD Symptoms
Descriptive statistics for the observational measures may
be found in a prior publication (Keown 2011). Table 1 shows correlations between parenting predictors
and child outcomes. Paternal sensitivity was negatively
Time 2 Fathers and mothers each completed the Caregiver related to maternal and paternal ratings of inattentiveness
Conflict with Child (7 items) and Caregiver Closeness with and teacher ratings of hyperactivity−impulsivity, while
Child (8 items) scales from the Child−Parent Relationship paternal positive regard was negatively related to mother
J Abnorm Child Psychol (2012) 40:569–581 575

Table 1 Zero-order correlations between Time 1 parenting predictors and Time 2 child ADHD symptoms

Measure Father ratingsa Mother ratingsb Teacher ratingsc

IA HI IA HI IA HI

Observational Ratings
Father codes
Sensitivity -0.23* -0.13 -0.28** -0.14† -0.14 -0.22*
Positive regard -0.17 -0.12 -0.19* -0.07 -0.22* -0.13
Intrusiveness 0.10 0.04 0.21* 0.09 0.14 0.23*
Mother codes
Sensitivity -0.19* -0.04 -0.11 0.03 -0.15† 0.04
†
Positive regard -0.20* -0.15 -0.23* -0.09 -0.14 -0.03
Intrusiveness 0.18* 0.05 0.11 -0.02 0.22* 0.01
†
p<0.10. * p<0.05. ** p<0.01
a
Father observations with parent ADHD ratings, n=89
b
Mother observations with mother and father ADHD ratings, n=91–92
c
Teacher ratings with mother and father observations, n=81–82
IA Inattention; HI Hyperactivity−Impulsivity

and teacher ratings of inattentiveness. Conversely, model predicting teacher rated hyperactivity/inattentiveness
paternal intrusiveness was positively related to maternal (because of the absence of a significant bivariate association
ratings of inattentiveness and teacher ratings of hyper- between early parenting and mother and father ratings of later
activity−impulsivity. Thus, fathers rated as more sensi- hyperactivity/impulsiveness, these relationships were not
tive and less intrusive during early childhood had boys explored further in multivariate analyses).
who were later rated as demonstrating fewer problems On the first step of each regression equation, Time 1
with hyperactivity/impulsivity at school and inattentiveness at hyperactivity scores as reported by the mother, father, or the
home. Whereas, according to mother and teacher report, child’s previous teacher, and Time 1 PACS conduct
paternal warmth was related to higher rates of attentiveness problems scores were entered. The rationale behind
across both settings. entering these variables first was that it allowed examina-
Mothers’ observed sensitivity was associated with tion of the stability of hyperactive/inattentive behaviors and
fathers’ reports of fewer inattentive symptoms at Time 2. controlled for the possible confounding effects of conduct
Likewise, maternal positive regard was associated with problems. It also provided a conservative test of the unique
lower rates of inattention on the basis of both father and role of observed parenting from early childhood, once the
mother ratings. Maternal intrusiveness in early childhood continuity in behavioral symptoms was taken in account.
was positively related to both teacher and father ratings of Given the pattern of correlations, for the models
inattention. Some of these significant findings parallel those predicting to mother and father rated inattentiveness,
found for fathers, where early warmth, sensitivity, and parenting variables were added in two steps to examine
intrusiveness were associated with later ratings of inatten- the relative contribution of maternal and paternal variables
tiveness by the other parent. However, unlike the results for on mothers’ versus fathers’ ratings of child inattentive
paternal parenting, observed maternal parenting was unre- behavior. To predict mothers’ ratings of inattentiveness,
lated to teacher ratings of child hyperactivity/impulsivity, maternal variables were entered first and paternal variables
while maternal intrusiveness, rather than warmth, was second. The order was reversed for fathers’ ratings. To
related to more child inattentiveness at school. reduce the number of variables on the third step composite
variables representing maternal or paternal responsiveness
Multiple Regressions Predicting ADHD Symptoms (the sum of sensitivity, positive regard, and reversed
intrusiveness) were created. Thus, in the prediction of
Hierarchical multiple regression was used to predict child- maternal ratings of child inattention, mothers’ positive
ren’s inattentive and hyperactive/impulsive symptoms at regard was added at step 2 and paternal responsiveness at
Time 2, as reported by their fathers, mothers, and teachers. step 3. While, in the prediction of fathers’ ratings of child
Altogether four sets of analyses were conducted: three inattention, paternal sensitivity was included at step 2, and
models predicting child inattentive behaviors, and one maternal responsiveness at step 3. For the models predict-
576 J Abnorm Child Psychol (2012) 40:569–581

ing Time 2 teacher ratings of child hyperactivity/impulsive- significant when added at step 3. Proceeding to mothers,
ness and inattentiveness, Time 1 parenting variables that Table 2 shows that early observed maternal positive regard
were associated at a statistically significant level with predicted maternal ratings of child inattention in middle
teacher ratings of ADHD symptoms were entered at step childhood when mother rated preschool overactivity,
2. Because paternal sensitivity and intrusiveness were impulsivity, inattentiveness, and PACS conduct problems
correlated (r=−0.83) they were not examined simultaneously were controlled. Fathers’ observed responsiveness at Time
in the model predicting teacher ratings of child hyperactivity/ 1 made a significant contribution to the prediction of child
impulsiveness. inattentiveness as rated by mothers, when included at step 3.
Regression results are presented in Table 2. Beginning Regression results predicting Time 2 teacher ratings of
with fathers, the results show that observed paternal child ADHD symptoms are also shown in Table 2. For the
sensitivity in early childhood predicted paternal ratings of model predicting child attentional problems, when age 4
child inattention in middle childhood, when father rated teacher rated inattentiveness, hyperactivity/impulsiveness,
preschool ADHD symptoms and PACS conduct problems PACS conduct problems, and receptive language scores
were controlled. Early maternal responsiveness was not were controlled, early maternal intrusiveness did not
significantly predict Time 2 teacher ratings of child
Table 2 Predictors of paternal, maternal, and teacher ratings of child
behavior. Paternal positive regard did not contribute
ADHD behaviors at Time 2 uniquely to teacher rated child inattentiveness. While for
the model predicting to child hyperactivity/impulsiveness,
Predictor ΔR2 β p after controlling for initial child problems at age 4, early
Criterion:T2 Attentional Problems (Father rateda)
paternal intrusiveness added unique variance to Time 2
Step 1. 0.27 <0.001
teacher reports of child symptoms. Paternal sensitivity did
not add significantly to the prediction of increases in child
T1 SDQ hyperactivity (father rated) 0.49 <0.001
hyperactivity/impulsiveness.
T1 PACS conduct problems 0.07 0.459
Step 2. T1 Paternal sensitivity 0.06 -0.24 0.009
Logistic Regressions Predicting ADHD
Step 3. T1Maternal responsiveness 0.02 -0.16 0.081
Criterion: T2 Attentional Problems (Mother ratedb)
Logistic regression models were used to assess the extent to
Step 1. 0.28 <0.001
which observed parenting in early childhood predicted later
T1 SDQ hyperactivity (mother rated) 0.41 <0.001
ADHD, after adjusting for the effects of age 4 child
T1 PACS conduct problems 0.21 0.032
inattentive−hyperactive symptoms and conduct problems.
Step 2. T1 Maternal positive regard 0.06 -0.24 0.008
Given between-group differences in maternal education, the
Step 3. T1 Paternal responsiveness 0.04 -0.20 0.022
c
confounding effect of maternal education was also consid-
Criterion: T2 Attentional Problems (Teacher rated )
ered. When Time 1 observed parenting variables were
Step 1. 0.18 0.001
compared for fathers in the Time 2 ADHD and comparison
T1 SDQ hyperactivity (teacher rated) 0.36 0.002 groups, no significant differences between the two groups
T1 PACS conduct problems 0.04 0.717 were found. Therefore, logistic regression models were run
T1 BPVS receptive language score -0.14 0.204 for maternal parenting variables only.
Step 2. 0.03 0.099 Two models were fitted for each of the parenting
T1 Maternal intrusiveness -0.17 variables of interest. The first model examined the
Criterion: T2 Hyperactivity/Impulsiveness (Teacher ratedd) relationship between the risk of ADHD and each parenting
Step 1. 0.24 <0.001 variable after statistical adjustment for the effects of earlier
T1 SDQ hyperactivity (teacher rated) 0.42 <0.001 inattentive−hyperactive symptoms and conduct problems.
T1 PACS conduct problems 0.15 0.148 This model was then extended to include maternal
Step 2. 0.04 0.040 education.
T1 Paternal intrusiveness 0.20 The results of this analysis are summarized in Table 3,
a which shows the adjusted and unadjusted odds ratios for
n=87
b those parenting variables with a significant univariate
n=89
c
association with ADHD. After adjusting for the effects of
n=82
d
Time 1 inattentive−hyperactive behaviors and conduct
n=81
problems, observed maternal sensitivity and positive regard
The variation in sample size is due to the different numbers of fathers,
were significantly associated with ADHD, while observed
mothers, and teachers who contributed data to this study. In addition, a
multivariate outlier was deleted for the regressions predicting maternal maternal intrusiveness was no longer significant after
and paternal ratings of child attentional problems statistical adjustment for the effects of preschool ADHD
J Abnorm Child Psychol (2012) 40:569–581 577

Table 3 Odds Ratios (OR) before and after adjustment for the effects of T1 ADHD symptoms, conduct problems, and maternal education

Univariate With T1 ADHD symptoms With T1 ADHD symptoms and conduct

Parenting Measure OR (95% CI) and conduct problems problems, and maternal education
OR (95% CI) OR (95% CI)

Free Play Observational Ratings

Maternal sensitivity 2.10* (1.34-3.88) 2.37* (1.08-5.20) 2.28* (1.04-5.00)
*
Maternal positive regard 2.14 (1.09-4.20) 3.53* (1.21-10.31) 3.88* (1.33-11.24)
Maternal intrusiveness 1.72* (1.00-2.94) 2.27† (1.06-4.88) -

n=51. Analyses are based on mothers in the T2 ADHD and control groups with T1 observational data. There were two cases for whom the model
did not fit well (one for positive regard, the other for sensitivity and intrusiveness). These cases were removed from the final models
†
p<0.06. * p<0.05

symptoms and conduct problems. These findings for coexisting child conduct problems. Paternal perceptions
maternal sensitivity and positive regard held when maternal of child−father conflict continued to be significantly
education was added to the models. As indicated in Table 3, associated with ADHD group status after statistical
boys whose mothers showed less positive regard towards adjustment for the effects of conduct problems (OR 1.29; CI
their sons at age 4 were 3.8 times more likely to meet 1.1–1.52), whereas maternal perceptions of child−mother
criteria for ADHD at Time 2, and 2.2 times more likely to conflict and child−mother closeness were no longer
meet ADHD criteria than comparison boys if less maternal significant.
sensitivity was evident during preschool mother−son
interactions.
Discussion
Parental Perceptions of Child−parent Relationships at Time 2
The purpose of this study was to examine whether observed
T tests were used to compare the parental perceptions of maternal and paternal responsive parenting during the
child−parent conflict and closeness between parents of boys preschool years predicted middle childhood ADHD symp-
in the ADHD group and parents of boys in the comparison toms, while taking into account the effects of early ADHD
group. There was significantly more child−parent conflict symptoms and conduct problems. A unique aspect of the
reported by both fathers and mothers of ADHD group boys study is the inclusion of observed paternal parenting.
than by parents of comparison group boys (Mothers: Furthermore, few, if any, previous studies, have examined
ADHD group M=19.33, SD=7.01; Control group M= whether independently observed maternal and paternal
14.07, SD = 4.36, t (37) = −3.20, p = 0.003, d = 0.93), parenting predict later ADHD symptoms as perceived by
(Fathers: ADHD group M=19.43, SD=5.57; Control group mothers, fathers, and teachers, above and beyond early
M=14.39, SD=4.86, t (49)=−3.45, p=0.001, d=0.97). ADHD symptoms and conduct problems. Both regression
Significantly less closeness in the child−parent relationship (ADHD dimensions) and between-group (ADHD/compari-
was reported by mothers of boys with ADHD (M=34.87, son) yielded evidence that specific aspects of responsive
SD=4.08) than mothers of comparison boys (M=37.17, SD maternal and paternal parenting uniquely predicted ADHD
=2.30), t (35)=2.46, p=0.019, d=0.72. Fathers of boys symptoms. Finally, new information is provided about
with and without ADHD did not differ significantly in their maternal and paternal perceptions of the child−parent
ratings of father−child closeness. Given these between- relationship quality of boys with ADHD symptoms.
group findings it was also of interest to compare whether Correlations between important aspects of responsive
fathers and mothers within the same family differed in their parenting when the boys were 4–5 years old and later
child−parent relationship scores. Within-group comparisons ADHD dimensions revealed a number of significant
using paired t tests showed that within the control group associations, with some variation by rater. Paternal sensi-
fathers had significantly lower scores than mothers on tivity was linked to less inattentiveness (based on paternal
closeness with their son (p=0.015). There were no other and maternal ratings), and less hyperactivity/impulsivity
significant differences in parental perceptions of child−parent (rated by teachers) over time. Maternal sensitivity was also
relationship between parents within the same family in related to fewer problems with inattention, but only on the
either group. Logistic regression models were then basis of maternal ratings. Paternal positive regard also
used to assess the extent to which between-group predicted higher ratings of child attentiveness at home
differences in parental perceptions of the child−parent (rated by mothers) and at school. While predictions from
relationship could be explained by the effects of maternal positive regard were linked with child attentive-
578 J Abnorm Child Psychol (2012) 40:569–581

ness in the home setting only (based on both maternal and they address the question of a link between parenting
paternal ratings). Lastly, both maternal and paternal responsiveness and ADHD symptoms left open by earlier
intrusiveness were related to later ratings of child inatten- cross-sectional research (Seipp and Johnston 2005). The
tiveness by the other parent. Whereas, according to teacher results also add weight to the suggestion that maternal
report, maternal intrusiveness predicted child inattention warmth and criticism may act together with genetic factors
and paternal intrusiveness was associated with later child to alter the severity of ADHD (Sonuga-Barke et al. 2008).
hyperactivity−impulsivity. Lastly, the findings also raise the possibility that maternal
These parenting variables were measured antecedent in sensitivity has a greater impact on the functioning of
time to child outcomes, suggestive of parental influence, children with ADHD than on inattentive symptoms alone,
nonetheless the direction of effect in correlational designs given that maternal sensitivity predicted later ADHD group
cannot be specified. Accordingly, the analytical approach status (as defined on the basis of both symptoms and
taken allowed for the prediction of parenting contributions, functioning) but not the level of inattentiveness.
with initial child problems at age 4 controlled, making it a In relation to ADHD symptoms at school, fathers’
stringent test, given the known strong continuity in early intrusive behavior in early childhood was the only
emerging disruptive behaviors (Denham et al. 2000). significant parenting predictor of teacher ratings in middle
Consistent with previous research (Lahey et al. 2005; von childhood. In particular, fathers who were overcontrolling
Stauffenberg and Campbell 2007) there was considerable during play interactions at age 4, were more likely to have
continuity in ADHD symptoms over time. Despite this sons who teachers perceived to be overactive and impulsive
stability, parenting variables were found to predict parent in the classroom. These findings suggest that paternal
ratings of child inattentiveness and teacher ratings of child behaviors that interrupt or take over the child’s activity, and
hyperactivity−impulsivity, in addition to the contributions limit the child’s influence on content and pace of play, may
made by early child behavior. compromise the development of behavioral self-regulation
More specifically, the results show that lower levels of in children showing early signs of ADHD. The findings are
observed paternal sensitivity predicted higher levels of also consistent with research that highlights the function of
subsequent child inattentive behavior at home according to high quality father−child play interactions for encouraging
fathers’ perspectives. Furthermore, paternal responsiveness the development of child competence outside the family
also made an independent contribution to the prediction of (Grossmann et al. 2002). A possible reason for the different
child inattentiveness as rated by mothers. While these results findings in relation to teacher ratings of hyperactivity/
are new, they are supported by the research about the impulsivity and teacher ratings of inattention may be that
significance of paternal sensitivity for lower levels of child when controlling for conduct problems variability was
externalizing behaviors (NICHD ECCRN 2004), and they reduced more so for hyperactivity/impulsivity than inattention,
extend this prior research to ADHD inattentive symptoms. due to the stronger correlation between conduct problems and
Current findings also support the possibility that, for some hyperactivity/impulsivity than inattention.
children who develop ADHD, one mechanism that accounts The variation in the parenting variables that predicted
for their poorly regulated behavior may be paternal difficul- informant specific ADHD ratings may partly reflect rater
ties in synchronizing their actions to the child’s cues and disagreements on the extent of child inattentive and
adapting their behaviors accordingly (Johnston and Mash overactive−impulsive behavior. Consistent with other re-
2001). Sensitive parenting attuned to child needs is thought search, there was a moderate level of agreement between
to be particularly important in the development of children’s mothers and fathers on ADHD symptom specific ratings
attentional and self organizing skills (Sonuga-Barke et al. (Langberg et al. 2010), with lower levels of agreement
2006), and is likely to be more difficult to sustain during between parents and teachers than between mothers and
play interactions with young boys who are inattentive, fathers (Achenbach et al. 1987). These discrepancies may
hyperactive and impulsive (Johnston and Mash 2001). reflect variation in rater perceptions and context differences
Additionally, findings showed that lower levels of in child behavior. However, in agreement with, and
observed maternal positive regard also predicted higher extending the findings of, studies based on mother and
levels of subsequent child inattentiveness at home as rated teacher defined ADHD subtypes (Gadow et al. 2004), the
by mothers. These results for mothers were echoed in the present results indicate that parenting predictors of ADHD
categorical analysis in that fewer signs of early maternal symptom dimensions may vary by mother, father, and
warmth, as well as less sensitive mother−child interactions teacher ratings. Furthermore, while these findings require
also predicted later ADHD group status. These findings replication in other samples, they highlight the utility of
support the possibility that less responsive maternal examining these types of relationships on the basis of
parenting is a risk factor for the continuation of early separate informant ratings of child behavior in order to
inattentive−hyperactive behavior problems. Furthermore, identify factors that predict later problems.
J Abnorm Child Psychol (2012) 40:569–581 579

The results for parental perceptions of the child−parent parents how to better pick up on child cues, respond to
relationship in middle childhood show some similarities child initiations, and show positive regard for their child. To
with Gerdes et al. (2003), who found that both mothers and further advance knowledge about effective parenting
fathers of children with ADHD had more negative interventions for preschool ADHD, these components of
perceptions than comparison group parents, but that only responsive parenting could be included in randomized
mothers of children with ADHD perceived less warmth in controlled trials to test whether improvements in child
the relationship. This research also extends Gerdes et al. symptoms and functioning are linked to improvements in
(2003) findings to a community sample. However, there specific responsive parenting behaviors for each parent.
were some differences in the types of parental perceptions Furthermore, if the current findings are replicated in clinical
assessed in each study. Gerdes et al. (2003) tapped parents’ samples, they highlight the potential importance of including
views about their use of power assertive discipline and the fathers in the evaluation and treatment of family components
degree of warmth in the relationship. Whereas, the present of ADHD, which could be tailored to the symptom domain
study asked about the child’s behavior towards the parent as most affected (Ellis and Nigg 2009).
well as the parents’ feelings and beliefs about the conflict While the current study identified specific aspects of
and closeness in his/her relationship with the child. Current responsive parenting that may play a role in maintaining
results suggest the possibility that maternal perceptions of ADHD symptoms, future studies need to assess parent−child
more conflict and less child−parent closeness may have interactions and child ADHD symptoms at both time
been present since early childhood, given the finding that points in order to examine bi-directional influences
less maternal warmth and sensitivity at age 4 predicted later between responsive parenting and ADHD. It would also be
ADHD group status. However, findings also suggest that by helpful to include three or more time points and to incorporate
middle childhood, maternal perceptions of the child−mother genetic factors in order to examine gene−environment
relationship were shaped by comorbid child conduct interactions and analyse complex patterns of change across
problems. The different findings for mothers and fathers time.
in early warmth and later parent−child closeness are This study had a number of limitations. First, a larger
also consistent with the theory that the mothering role sample would allow comparisons of responsive parenting
has a greater emphasis on nurturing than the fathering predictors to be made on the basis of DSM-IV ADHD
role (Grossmann et al. 2002). subtypes as well as ADHD dimensions, and parent gender,
At Time 2, both fathers and mothers of boys with ADHD and thus improve generalizability of findings. In particular,
reported experiencing parent−child interactions that includ- larger categorical group sizes would provide more power to
ed constant struggles and child negative affect directed examine paternal predictors of ADHD. Missing teacher
towards them. As parent−child relationships play a major questionnaire and parent interview data at Time 2 poten-
role in children’s development and adjustment, these tially reduced the number of children who could be
findings suggest that interventions should be targeted categorized as meeting ADHD criteria using impairment
towards enabling both fathers and mothers of children with data, and the size of the groups available for the categorical
ADHD and comorbid conduct problems to utilize effective analysis. Also, the present study did not use a diagnostic
child behavior management strategies that will reduce interview measure to define ADHD group status. However,
parent−child conflict and increase the level of positive ADHD symptom counts were based on ratings from three
parent−child interactions. informants and child impairment across settings was based
In addition, the longitudinal findings that early respon- on a comprehensive range of information. Second, the
sive parenting predicted later ADHD symptoms, offers sample in the present study was limited to boys. A larger
further support for the clinical utility of training parents of sample that includes girls would allow child gender effects
children with ADHD in strategies for improving responsive on parenting to be tested. Third, the results may only
parenting (Seipp and Johnston 2005), and for these generalize to two-parent families. The relationships be-
interventions to be targeted at both fathers and mothers of tween parenting and ADHD may differ in other family
preschool children presenting with signs of ADHD. It may contexts (e.g., single-parent families). Furthermore, as boys
be helpful for clinicians to tailor interventions suited to in this study were predominantly from mid to high SES
individual families that focus on specific aspects of homes, these data may not be applicable to children living
responsive parenting. For example, in families where in low SES homes. Future research also needs to consider
parent–child interactions are intrusive and overcontrolling, the extent to which relations between parenting and child
clinicians could focus on increasing parent behaviors that ADHD may be explained by the parent’s own ADHD
allow the child more opportunity to respond at his/her own symptoms.
pace. While interventions that include components for In conclusion, within the limitations noted, these results
increasing parental sensitivity and warmth could teach tentatively suggest that lower levels of paternal sensitivity
580 J Abnorm Child Psychol (2012) 40:569–581

and maternal positive regard in early childhood are behavior problems from early to middle childhood: The role of
parental socialization and emotion expression. Development and
uniquely predictive of higher levels of inattentiveness in Psychopathology, 12, 23–45.
middle childhood and that intrusive paternal behaviors Dunn, L. M., Dunn, L. M., Whetton, C., & Burley, J. (1997). The
during father−son interactions are predictive of hyper- British picture vocabulary scale (2nd ed.). Berkshire, UK:
active−impulsive behaviors at school. While, preschool NFER-NELSON.
DuPaul, G. J., Power, T. J., McGoey, K. E., Ikeda, M. J., &
mother−son interactions characterized by less optimal Anastopoulos, A. D. (1998). Reliability and validity of parent
levels of warmth and sensitivity are uniquely predictive and teacher ratings of attention-deficit/hyperactivity disorder
of later ADHD symptoms. These predictions held over symptoms. Journal of Psychoeducational Assessment, 16, 55–68.
and above the stability of early ADHD behaviors and Elley, W. B., & Irving, J. C. (2003). The Elley-Irving Socio-economic
index: 2001 Census revision. New Zealand Journal of Educational
the influence of early conduct problems, and highlight Studies, 38, 3–18.
the importance of considering the unique associations Ellis, B. M. A., & Nigg, J. P. D. (2009). Parenting practices and
between ADHD symptoms and responsive parenting attention-deficit/hyperactivity disorder: New findings suggest
behaviors of both fathers and mothers. partial specificity of effects. Journal of the American Academy
of Child and Adolescent Psychiatry, 48, 146–154.
Gadeyne, E., Ghesquiere, P., & Onghena, P. (2004). Longitudinal relations
Acknowledgement This research was supported by grants from the
between parenting and child adjustment in young children. Journal
Auckland Medical Research Foundation, Oakley Mental Health
of Clinical Child and Adolescent Psychology, 33, 347–358.
Foundation, and the University of Auckland. Special thanks to my
Gadow, K. D., Drabick, D. A. G., Loney, J., Sprafkin, J., Salisbury,
research assisstants, Teresa Clark, Joy Davidson, Rachel Haselden,
H., Azizian, A., et al. (2004). Comparison of ADHD symptom
Richard Keown, Sundeia Lomberg, Dorothy MacArthur, Melanie
subtypes as source-specific syndromes. Journal of Child
Palmer, and Heather Summerfield, for help with data collection,
Psychology and Psychiatry, 45, 1135–1149.
processing, and coding.
Gerdes, A. C., Hoza, B., & Pelham, W. E. (2003). Attention-deficit/
hyperactivity disordered boys' relationships with their mothers
and fathers: Child, mother, and father perceptions. Development
References and Psychopathology, 15, 363–382.
Gerdes, A. C., Hoza, B., Arnold, L. E., Hinshaw, S. P., Wells, K. C.,
Hechtman, L., et al. (2007). Child and parent predictors of
Achenbach, T. M. (1991). Manual for the Child Behavior Checklist/4-18 perceptions of parent-child relationship quality. Journal of
and 1991 Profile. Burlington, VT: University of Vermont Attention Disorders, 11, 37–48.
Department of Psychiatry. Goldstein, L., Harvey, E., & Friedman-Weieneth, J. (2007). Examining
Achenbach, T. M., McConaughy, S. H., & Howell, C. T. (1987). subtypes of behavior problems among 3-year-old children, part III:
Child/adolescent behavioral and emotional problems: Implications Investigating differences in parenting practices and parenting stress.
of cross-informant correlations for situational specificity. Psycho- Journal of Abnormal Child Psychology, 35, 125–136.
logical Bulletin, 101, 213–232. Goodman, R. (1997). The Strengths and Difficulties Questionnaire: A
Barkley, R. A. (2006). Attention-deficit hyperactivity disorder: A research note. Journal of Child Psychology and Psychiatry, 38,
handbook for diagnosis and treatment (3rd ed.). New York, NY: 581–586.
Guilford Press. Goodman, R. (1999). The extended version of the Strengths and
Brady-Smith, C., O’Brien, C., Berlin, L., Ware, A., & Fauth, R. Difficulties Questionnaire as a guide to child psychiatric caseness
C. (2000). 36-Month child-parent interaction rating scales for and consequent burden. Journal of Child Psychology and
the three-bag assessment. Unpublished Manuscript. National Psychiatry, 40, 791–799.
Center for Children and Families, Teachers College, Columbia Goodman, R. (2001). Psychometric properties of the Strengths and
University. Difficulties Questionnaire. Journal of the American Academy of
Cabrera, N. J., Shannon, J. D., & Tamis-LeMonda, C. (2007). Fathers' Child and Adolescent Psychiatry, 40, 1337–1345.
influence on their children's cognitive and emotional development: Grossmann, K., Grossmann, K. E., Fremmer-Bombik, E., Kindler, H.,
From toddlers to Pre-K. Applied Developmental Science, 11, 208– Scheuerer-Englisch, H., & Zimmermann, P. (2002). The uniqueness
213. of the child-father attachment relationship: Fathers' sensitive and
Campbell, S. B., & Ewing, L. J. (1990). Hard-to-manage preschoolers: challenging play as a pivotal variable in a 16-year longitudinal
Adjustment at age nine and predictors of continuing symptoms. study. Social Development, 11, 307–331.
Journal of Child Psychology and Psychiatry, 31, 871–889. Hawes, D. J., & Dadds, M. R. (2004). Australian data and
Campbell, S. B., Pierce, E. W., Moore, G., Marakovitz, S., & Newby, psychometric properties of the Strengths and Difficulties Question-
K. (1996). Boys’ externalizing problems at elementary school naire. The Australian and New Zealand Journal of Psychiatry, 38,
age: Pathways from early behavior problems, maternal control, 644–651.
and family stress. Development and Psychopathology, 8, 701– Heptinstall, E. (1993). Revised manual for the PACS interview.
719. Institute of Psychiatry, University of London.
Carlson, E. A., Jacobvitz, D., & Sroufe, L. A. (1995). A develop- Hinshaw, S. P., March, J. S., Abikoff, H., Arnold, L. E., Cantwell, D.
mental investigation of inattentiveness and hyperactivity. Child P., Conners, C. K., et al. (1997). Comprehensive assessment of
Development, 66, 37–54. childhood attention-deficit hyperactivity disorder in the context
Deault, L. (2010). A systematic review of parenting in relation to the of a multisite, multimodal clinical trial. Journal of Attention
development of comorbidities and functional impairments in Disorders, 1, 217–234.
children with Attention-Deficit/Hyperactivity Disorder (ADHD). Johnston, C., & Mash, E. J. (2001). Families of children with
Child Psychiatry and Human Development, 41, 168–192. attention-deficit/hyperactivity disorder: Review and recommenda-
Denham, S. A., Workman, E., Cole, P. M., Weissbrod, C., Kenziora, tions for future research. Clinical Child and Family Psychology
K. T., & Zahn-Waxler, C. (2000). Prediction of externalizing Review, 4, 183–207.
J Abnorm Child Psychol (2012) 40:569–581 581

Keown, L. J. (2011). Fathering and mothering of preschool boys with Pungello, E. P., Iruka, I. U., Dotterer, A. M., Mills-Koonce, R., &
hyperactivity. International Journal of Behavioral Development, Reznick, S. J. (2009). The effects of socioeconomic status, race,
35, 161–168. and parenting on language development in early childhood.
Keown, L. J., Woodward, L. J., & Field, J. (2001). Language Developmental Psychology, 45, 544–557.
development of pre-school children born to teenage mothers. Seipp, C. M., & Johnston, C. (2005). Mother-son interactions in
Infant and Child Development, 10, 129–145. families of boys with Attention-Deficit/Hyperactivity Disorder
Lahey, B. B., Pelham, W. E., Loney, J., Lee, S. S., & Willcutt, E. with and without oppositional behavior. Journal of Abnormal
(2005). Instability of the DSM-IV subtypes of ADHD from Child Psychology, 33, 87–98.
preschool through elementary school. Archives of General Sonuga-Barke, E. J. S., Auerbach, J., Campbell, S. B., Daley, D., &
Psychiatry, 62, 896–902. Thompson, M. (2005). Varieties of preschool hyperactivity: Multiple
Langberg, J., Epstein, J., Simon, J., Loren, R., Arnold, L., Hechtman, L., et pathways from risk to disorder. Developmental Science, 8, 141–150.
al. (2010). Parent agreement on ratings of children’s attention deficit/ Sonuga-Barke, E. J. S., Thompson, M., Abikoff, H., Klein, R., &
hyperactivity disorder and broadband externalizing behaviors. Brotman, L. M. (2006). Nonpharmacological interventions for
Journal of Emotional and Behavioral Disorders, 18, 41–50. preschoolers with ADHD: The case for specialized parent
Lifford, K., Harold, G., & Thapar, A. (2008). Parent–child relationships training. Infants and Young Children: An Interdisciplinary
and ADHD symptoms: A longitudinal analysis. Journal of Journal of Special Care Practices, 19, 142–153.
Abnormal Child Psychology, 36, 285–296. Sonuga-Barke, E. J. S., Lasky-Su, J., Neale, B. M., Oades, R., Chen,
NICHD Early Child Care Research Network. (2004). Fathers' and W., Franke, B., et al. (2008). Does parental expressed emotion
mothers’ parenting behavior and beliefs as predictors of children’s moderate genetic effects in ADHD? An exploration using a
social adjustment in the transition to school. Journal of Family genome wide association scan. American Journal of Medical
Psychology, 18, 628–638. Genetics. Part B, Neuropsychiatric Genetics, 147B, 1359–1368.
Peris, T. S., & Baker, B. L. (2000). Applications of the expressed Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics
emotion construct to young children with externalizing behavior: (5th ed.). New York: Pearson Education.
Stability and prediction over time. Journal of Child Psychology Taylor, E., Sandberg, S., Thorley, G., & Giles, S. (1991). The
and Psychiatry, 41, 457–462. epidemiology of childhood hyperactivity (Maudsley Monographs
Pianta, R. C. (1992). Child-parent relationship scale. Unpublished No. 33). New York, NY: Oxford University Press.
measure. Charlottesville, VA: University of Virginia. von Stauffenberg, C., & Campbell, S. B. (2007). Predicting the early
Pierce, E. W., Ewing, L. J., & Campbell, S. B. (1999). Diagnostic developmental course of symptoms of attention deficit hyperactivity
status and symptomatic behavior of hard-to-manage preschool disorder. Journal of Applied Developmental Psychology, 28, 536–
children in middle childhood and early adolescence. Journal of 552.
Clinical Child Psychology, 28, 44–57. Woodward, L., Dowdney, L., & Taylor, E. (1997). Child and family
Pleck, J. H. (1997). Paternal involvement: Levels, sources and factors influencing the clinical referral of hyperactive children: A
consequences. In M. E. Lamb (Ed.), The role of the father in research note. Journal of Child Psychology and Child Psychiatry,
child development (pp. 63–103). New York: Wiley. 38, 479–485.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.