The Influence of Child Gender on the Prospective Relationshipsbetween Parenting and Child ADHD

David H. Demmer1,2 & Francis Puccio3,4 & Mark A. Stokes1,2 & Jane A. McGillivray1,2&

Merrilyn Hooley1,2

Published online: 2 March 2017# Springer Science+Business Media New York 2017

Abstract The aims of the current study were to (i) explore thepotential bidirectional, prospective relationships between par-enting and child ADHD, and (ii) explore whether these rela-tionships differed on the basis of child gender. Data wereobtained from waves 1 (children aged 4- to 5-years) to 5 (chil-dren aged 12- to 13-years) of the Longitudinal Study ofAustralian Child (LSAC) dataset (child cohort). In order toexamine dimensions of both mothers’ and fathers’ parenting,a subsample of nuclear families with mothers, fathers andchildren present at all waves was extracted (final sam-ple = 1932; sons = 981, daughters = 951). Child ADHD mea-sures included the hyperactive-impulsive subscale of thestrengths and difficulties questionnaire for symptoms, andparent-report question for diagnosis. Mothers and fatherscompleted scales on dimensions of Angry, Warm andConsistent Parenting. A cross-lagged panel model demon-strated (i) higher child ADHD symptoms at wave 1 led to aglobal increase in less-than-optimal parenting at wave 2, and(ii) child ADHD symptoms and Angry Parenting shared a

prospective, bi-directional relationship (whereby increases inone predicted increases in the other over time) during earlieryears of development. Latent growth curve models demon-strated that increases in Angry Parenting across timewere significantly predicted by increases in childADHD symptoms. A logistic regression demonstratedthat both mothers’ and fathers’ Angry Parenting at wave1 significantly predicted an ADHD diagnosis in childrenat wave 3. No predictive relationships differed betweenchild genders; thus, it appears these prospective path-ways are similar for both sons and daughters.

Keywords Childhood psychopathology . ADHD .

Externalizing disorder . Parent . Gender

Attention-deficit/hyperactivity disorder (ADHD) is aneurodevelopmental disorder involving difficulties with atten-tion and/or hyperactive and impulsive behaviours (AmericanPsychiatric Association [APA] 2013). ADHD affects approx-imately 5% of children worldwide, with prevalence estimatesincreasing or decreasing based on time period and/or cultureconsidered (Polanczyk et al. 2014). ADHD can have negativeimpacts on various domains of functioning, including familyand peer relationships (Hoza 2007; Keown and Woodward2002) and academic achievement (Barbaresi et al. 2007;Reid et al. 2004), and is also a significant predictor of bothconcurrent and later psychopathology (Biederman et al. 2006;Burke et al. 2005; Meinzer et al. 2016). Therefore, under-standing the factors associated with the aetiology and devel-opment of ADHD can help inform targets for intervention toassist in lessening these negative consequences.

One factor that has consistently been associated with thedevelopment of child ADHD is less-than-optimal parenting.Cross-sectional research has demonstrated that children with a

Electronic supplementary material The online version of this article(doi:10.1007/s10802-017-0284-7) contains supplementary material,which is available to authorized users.

* Merrilyn Hooleymerrilyn.hooley@deakin.edu.au

1 School of Psychology, Faculty of Health, Deakin University,Geelong, Australia

2 Deakin University, 221 Burwood Highway, Burwood, VIC 3125,Australia

3 Psychological Sciences, University of Melbourne,Melbourne, Australia

4 University of Melbourne, Grattan Street, Parkville, VIC 3010,Australia

J Abnorm Child Psychol (2018) 46:113–125DOI 10.1007/s10802-017-0284-7

diagnosis of ADHD tend to have parents who are less warmand less involved (Ellis and Nigg 2009; Tripp et al. 2007), lessconsistent in their punishment (Cussen et al. 2012), and/ormore overprotective (Chang et al. 2013) and controlling(Rogers et al. 2009) than the parents of their typically-developing peers. Higher levels of child ADHD symptomsare also related to higher levels of maternal hostility (Haroldet al. 2013), parental stress (Graziano et al. 2011), as well asvarious other aspects of negative parenting (e.g., poormonitoring/supervision, inconsistent discipline and corporalpunishment; Haack et al. 2016).

Despite the wealth of investigations into the relationshipsbetween various aspects of parenting and child ADHD, sev-eral important questions remain unanswered. First, as mostprevious investigations have been cross-sectional in design,the causal direction of the association between parenting andchild ADHD (i.e., whether certain parenting dimensions arepredictive of, or an outcome of, child ADHD, or if a bi-directional relationship exists) remains unclear. Second, de-spite suggestions that child gender may moderate the relation-ship between parenting and child ADHD (Braza et al. 2015;Johnston and Mash 2001), this notion is yet to be formallytested (Lifford et al. 2008).

Two longitudinal studies have explored the prospectiverelationships between parenting and child ADHD (Keown2012; Lifford et al. 2008) in attempts to address the questionof causality. The first, Lifford et al., examined the notion of bi-directionality and found different results for mothers and fa-thers. For mothers, higher levels of child ADHDsymptomology predicted higher levels of maternal rejectionover a 12-month period, while the reverse relationship wasfound for fathers, with higher levels of paternal rejectionpredicting increases child ADHD symptoms in children overthe same period. The second, Keown (2012), examined uni-directional relationships only, and again found different rela-tionships for mothers and fathers. Results demonstrated thathigher levels of paternal sensitivity and positive regard, andhigher levels of maternal positive regard, predicted lowerlevels of ADHD symptoms in sons across a 2 1/2 year period.Further, lower levels of maternal warmth predicted higherlevels of sons’ ADHD symptoms over the same period.

The findings of Lifford et al. (2008) and Keown (2012)contribute to the understanding of the prospective relationshipbetween parenting and child ADHD, however several limita-tions exist with these investigations that necessitate furtherwork in this area. First, both studies called for investigationsutilising larger sample sizes to strengthen the confidence offindings. Second, Keown was limited by investigating onlyuni-directional relationships (earlier parenting predicting laterchild ADHD), which precluded the potential identification ofalternate, or bi-directional relationships. This is a notableomission given common developmental models of ADHD(e.g., Johnston and Mash 2001), and the findings of Lifford

et al., suggest bi-directional relationships likely exist betweenparenting and child ADHD. Third, both studies omit childgender as a potential moderator despite the potential impor-tance in examining its effect on the relationship between par-enting and child ADHD (Braza et al. 2015; Johnston&Mash).

Gender, specifically male gender, is a known risk factor inthe development of ADHD, as demonstrated by the higherprevalence of ADHD in boys when compared to girls(approximately three boys to every one girl in communitysamples; Erskine et al. 2013; Willcutt 2012). However, asADHD research typically involves male-only samples, theway that child gender operates as a risk factor in the develop-ment of the disorder remains unclear (Johnston and Mash2001). This may have important implications in the treatmentof ADHD, particularly if different pathways of disease occurfor boys and girls. For example, parent training is an importantcomponent of most psychosocial treatments for child ADHD(National Health and Medical Research Council 2012) basedon the notion that parenting impacts child ADHD. Therefore,if parenting is differentially related to ADHD for girls andboys then current interventions may not be equally effectivefor both genders if the current ‘one-size-fits-all’ approach totreatment is taken.

Although the notion of whether the prospective relation-ships between parenting and ADHD differ between daughtersand sons is yet to be examined, parallel areas of researchsuggest a gender difference may exist. For example, it hasconsistently been demonstrated that parents commonly en-gage in different parenting behaviours with sons comparedto daughters. Sons tend to receive higher levels of authoritar-ian parenting (e.g., corporal punishment, lack of explanationabout punishment, verbal hostility) and less positive parenting(e.g., less displays of warmth, less aware and responsive tocues from the child) than girls (Barnett and Scaramella 2013;Russell et al. 1998). Sons also receive fewer displays of emo-tional understanding (Fivush et al. 2000), praise, and physicalaffection than daughters, yet receive higher rates of yellingand smacking (Lloyd and Devine 2006). Given that (i) manyof these less-than-optimal parenting behaviours have beenlinked to ADHD, and (ii) child gender is related to both par-enting and ADHD, it might be that child gender also impactsthe relationship between parenting and ADHD. For example,theory (e.g., Keenan and Shaw 1997) as well as previousresearch (e.g., Wright et al. 2013) suggests that thesocialisation of gender often involves a greater tolerance ofinternalising behaviours in girls and externalising behavioursin boys by socialising agents (e.g., parents). These socialisinginfluences potentially attenuate the relationship between neg-ative parenting and externalising behaviours in girls, and inboys potentially contribute to the development of a maladap-tive cycle of less-than-optimal parenting influencingexternalising behaviours, which go on to influence less-than-optimal parenting. It might therefore be expected that less-

114 J Abnorm Child Psychol (2018) 46:113–125

than-optimal parenting might have a greater influence on thedevelopment of externalising behaviours (such as those dem-onstrated in ADHD) in boys than in girls.

The Current Study

The aims of the current study were twofold; first, toexplore the potential bidirectional, prospective relation-ships between parenting dimensions and child ADHD,and second, to explore whether these relationships dif-fered on the basis of child gender, in a large sample ofAustralian nuclear families. Nuclear families were inves-tigated due to previous research often focusing only onmothers’ parenting. This focus, however, overlooks theimportant influence of fathers’ parenting on child behav-iours. As evidence suggests that child developmental out-comes differ depending on whether maternal or paternalinfluences are considered (Braza et al. 2015; Lamb2004), and based on previous longitudinal research inthis area demonstrating that mothers’ and fathers’ parent-ing are differentially associated with child ADHD(Keown 2012; Lifford et al. 2008), it is important toconsider both parents in investigations. The current studytested a number of hypotheses. Bi-directional relation-ships were expected based on arguments by commondevelopmental models of ADHD (e.g., Johnston andMash 2001). Gender differences in these relationshipswere also expected based on previous theory (Keenanand Shaw 1997) that suggests girls might be discouragedfrom presenting their difficulties in an externalizing man-ner. It was hypothesized that:

1. Bi-directional, prospective relationships will exist be-tween parenting dimensions and child ADHD symptoms.Specifically, higher levels of less-than-optimal parenting(i.e., higher scores on Angry Parenting, lower scores onConsistent Parenting, and lower scores on WarmParenting) will be predictive of, and predicted by, higherlevels of ADHD symptoms in boys and girls across time

2. The predictive relationship between parenting dimensionsand child ADHD symptoms will be stronger for boys thangirls as measured by chi square change between gender-specific models

3. Less-than-optimal parenting (i.e., higher scores on AngryParenting, lower scores on Consistent Parenting, and low-er scores onWarm Parenting) at wave 1 will be predictiveof an ADHD diagnosis in boys and girls at wave 3

4. The predictive relationship between parenting dimensionsand ADHD diagnosis will be stronger for boys than girlsas measured by chi square change between gender-specific models

Method

Participants

Data for this study were obtained from waves 1 through 5 ofthe Longitudinal Study of Australian Children (LSAC). Acomprehensive overview of the LSAC sampling design, datacollection methods and measures have been described else-where (Australian Institute of Family Studies 2013; Soloffet al. 2005; Zubrick et al. 2014), thus only a brief outline isprovided here. A two-stage cluster sampling design was usedto recruit two cohorts, an infant cohort (children 3–19 monthsof age at wave 1) and a child cohort (children 4–5 years of ageat wave 1). First, stratification occurred at the state of resi-dence level, and urban versus rural level. Postcodes (exclud-ing the most remote) were then sampled. Second, all childrenfrom sampled postcodes who were born between March 2003and February 2004 (infant cohort), and March 1999 andFebruary 2000 (child cohort), and enrolled in the AustralianMedicare Database, which is the most comprehensive data-base of Australia’s population, were contacted. Final sampleswere 5107 for the infant cohort and 4983 for the child cohort.Only the child cohort was used in the current study due tochild ADHD measures being present from wave 1 forthis cohort but not for the infant cohort. In order toexplore dimensions of both maternal and paternal par-enting, a subsample of the original 4983 child cohortcases was extracted. Cases were included in the currentsample if mother, father and child were present, andparticipated, through all waves, with a final subsampleof 1932; 951 daughters (49.2%) and 981 sons (50.8%).

As causal pathways between non-referred and referredcases are considered fundamentally different, all referred cases(i.e., children with a diagnosis of ADHD) were removed fromthe cross-lagged path analysis and the latent growth curvemodels (LGCMs) as per appropriate epidemiological method-ology (Rothman et al. 2008). However, parallel cross-laggedand LGCM analyses were conducted with the full sample (i.e.,referred and non-referred cases) with no difference in path-ways found between the models. Referred cases were includ-ed in the logistic regression as child ADHD diagnosis was thedependent variable in this analysis.

Measures

ADHD

Child ADHD symptoms were measured in the LSAC datasetvia the strengths and difficulties questionnaire (SDQ;Cronbach’s α = 0.88). Mothers’ SDQ scores were used inthe current study as mothers had lower levels of missing dataon this measure compared to fathers. The SDQ is a brief be-havioural screening questionnaire for children 4- to 16-years

J Abnorm Child Psychol (2018) 46:113–125 115

of age, consisting of 25 items covering behavioural and emo-tional problems, and prosocial behaviours (Goodman 1997),and contains five subscales. The hyperactivity-inattention sub-scale is made up of five items (e.g., BRestless, overactive,cannot stay still for long^ and Beasily distracted, concentrationwanders^). Items on the subscale are rated on a three-pointindex: 0 (not true), 1 (somewhat true), or 2 (certainly true),with a possible range of scores from 0 to 10. Higher scoresindicate higher levels of hyperactive-inattentive symptoms.The validity and sensitivity of the SDQ as a screening toolfor ADHD symptoms has been well established (Goodmanet al. 2000; Stone et al. 2010), including in Australian samples(Hawes and Dadds 2004). For example, Hawes and Daddsused factor analytic techniques to investigate the validity andreliability of the SDQ (parent-report) on an Australian com-munity sample of 1359 boys and girls 4- to 9-years of age.Moderate to strong internal reliability and stability was foundacross all subscales, suggesting the SDQ is a valid and reliablemeasure of both behavioural and emotional symptomology inAustralian children.

Child ADHD diagnosis in the LSAC dataset was deter-mined by a single item parent-report question with a yes orno response (BDoes child have any of these ongoing prob-lems? ADD or ADHD?^). A second question regardingADHD medication use (BHas your child ever taken any med-ication for attention deficit disorder or ADHD?^) was used inthe current study to validate the response (83% of parents whoreported ‘yes’ to ADD or ADHD as an ongoing problem alsoreported their children had been prescribed ADD or ADHDmedication at wave 5).

Parenting Dimensions

The parenting dimensions extracted from the LSAC datasetwere Angry Parenting, Warm Parenting and ConsistentParenting. These dimensions have previously demonstratedassociations with child ADHD (angry: Keown andWoodward 2002; warm: Chang et al. 2013; Keown 2012;consistent: Ellis and Nigg 2009). The construct validity andmeasurement quality of these scales, as they relate to theLSAC dataset, have been published in a technical paper(Zubrick et al. 2014). For the current study, very good internalconsistencies were found for all scales (mothers’ AngryParenting α = 0.86, fathers’ Angry Parenting α = 0.85;mothers’ Consistent Parenting α = 0.88, fathers’ ConsistentParenting α = 0.86; mothers’ Warm Parenting α = 0.88; fa-thers’ Warm Parenting α = 0.89). Each scale contained fiveitems with responses provided on a 5-point scale (1 = never/almost never, 2 = rarely, 3 = sometimes, 4 = often, 5 = always/almost always). Both Angry Parenting and ConsistentParenting were measured using items from scales in theNational Longitudinal Survey of Children and Youth(NLSYC; Statistics Canada 2000), while Warm Parenting

was measured using a modified scale from the ChildrearingQuestionnaire (Patterson and Sanson 1999).

Angry Parenting measured parents’ use of aversive orharsh discipline via items regarding feelings of anger or frus-tration towards the child, as well as emotional reactivity (e.g.,BHow often are you angry when you punish this child?^), withhigher scores on this measure representing less praise, moredisapproval, and more negative emotions directed towards thechild. Parents of children with ADHD have been demonstrat-ed to be more angry and mean in their disciplinary encounterswith their children than parents of non-ADHD children(Keown and Woodward 2002). Further, harsh discipline and/or hostility directed towards children has been shown to influ-ence the development and maintenance of behavioural prob-lems during childhood (Chang et al. 2003), while decreases inthese parenting behaviours leads to positive change in childbehaviour (Sanders et al. 2000).

Warm Parenting measured the amount of warmth andaffection displayed towards the child (e.g., BHow oftendo you hug or hold this child for no particular reason?^),with higher scores on this measure representing more dis-plays of warmth and affection. Children with ADHD reg-ularly receive less warmth (Chang et al. 2013), with lowerlevels of parental warmth predicting poorer developmentaloutcomes for children (Davidov and Grusec 2006) andsubstance use disorders in children with ADHD once theyreach adolescence (Tandon et al. 2014).

Consistent Parenting measured the setting and consistentapplication of age-appropriate rules and expectations (e.g.,BHow often does this child get away with things that you feelshould have been punished?^). Higher scores on this measurerepresented the more consistent setting and application ofrules and expectations. Inconsistent parenting has been shownto strongly contribute to children’s problem behaviours, and isone of the key areas addressed in behavioural family interven-tions (Sanders et al. 2000).

Statistical Methods

In order to conduct a comprehensive investigation, a series ofanalyses were selected and conducted to test the hypotheses ofthe current study and to address different aspects of prospec-tive relationships between parenting and (i) child ADHDsymptoms, and (ii) child ADHD diagnosis. A cross-laggedpanel model, two latent growth curve models (LGCMs), aswell as a logistic regression, were used to test the hypotheses.All analyses were performed using Mplus version 7.2 andSPSS version 23. Standard cut-offs for fit indices were used:CFI and TLI > 0.95 for excellent fit and >0.90 for acceptablefit and RMSEA <0.05 for good fit (Browne and Cudeck 1993;MacCallum et al. 1996; Steiger 1989). Socio-economic status(SES) was controlled for within all analyses as SES has beenshown to be an important associate for both parenting

116 J Abnorm Child Psychol (2018) 46:113–125

(Friedson 2016; Pinderhughes et al. 2000) and child ADHD(Russell et al. 2015). The SES variable was a single continu-ous score calculated based on (i) combined family income, (ii)educational attained of both parents, and (iii) parents’ occupa-tional status. This score therefore represented the social andeconomic resources available to the family. Rates of missingdata for all variables ranged from approximately 0.1% to 8.3%for individual variables across waves. Dimensions of fathers’parenting generally had the highest rates of missing dataacross all waves. Missing data were imputed using multipleimputation in MPLUS.

Cross-Lag Panel Model

A cross-lagged panel analysis using continuous manifest var-iables was conducted to test Hypothesis 1. Bi-directional ef-fects examine the reciprocal relationship between two con-structs measured across time, and addresses whether a partic-ular construct, measured at a particular time point, is predic-tive of change in another variable at a later time point. In ourmodel, variance in each of the variables of interest (parentingdimensions and child ADHD symptoms) was predicted fromtwo main sources: autoregressive paths and cross-laggedpaths. The autoregressive paths represent the effect of theconstruct on itself across time (e.g., child ADHD symptomsat time 1 predicting child ADHD symptoms at time 2; Seligand Little 2012). The inclusion of autoregressive paths mini-mises bias in the estimation of cross-lagged paths (Cole andMaxwell 2003; Gollob and Reichardt 1987; Selig & Little).The cross-lagged paths represent the relationships of interest(e.g., the variance in child ADHD at time 2 that is predicted byparenting dimensions at time 2). Given the inclusion of theautoregressive paths, the only variance available for predictionby the cross-lagged paths is the residual variance in the out-come variable, thus providing a more sensitive analysis.

Latent Growth Curve Models

Latent growth curve models were also used to test Hypothesis1. LGCMs assess whether initial levels of one variable(known as the intercept), or the change trajectory (known asthe slope) in that variable over time, predicts the change tra-jectory (slope) of another variable over time. The value of theLGCM is that it establishes how different responses emergeacross time, accounting for how earlier decisions influencelater outcomes. Further, this enables an understanding ofhow subtle differences at a given time point may lead to con-siderable differences in outcome (i.e., sensitive dependenceupon initial conditions). Two separate LGCMs were con-structed to assess the relationships between the interceptand slope of a predictor variable on the slope of an out-come variable. In the first LGCM, the intercepts andslopes of mothers’ and fathers’ Consistent, Angry and

Warm Parenting were predictors for the slope of childADHD symptoms. In the second LGCM, the interceptand slope of child ADHD symptoms were the predictorvariables, with the slopes of mothers’ and fathers’Consistent, Angry and Warm Parenting as outcomes.

Logistic Regression

A logistic regression analysis was used to test Hypothesis 3,and examined if mothers’ and fathers’ Consistent, Angry and/or Warm Parenting at Wave 1 predicted a child ADHD diag-nosis at wave three. Wave 1 was selected as the predictionwave as the cross-lagged model demonstrated that this wasthe time point where parenting and child ADHD symptomswere most related. Wave 3 was selected because children atthis age (8- and 9-years of age) are at the average age of childADHD diagnosis (Kessler et al. 2005).

Structural Invariance Testing

Structural invariance testing was conducted for the cross-lagpanel model and LGCMs to test Hypothesis 2, and also for thelogistic regression to test Hypothesis 4 (except where statisti-cal significance was not found in the relationships of interest),to determine whether the predictive relationships in thesemodels were equivalent across child genders. First, an uncon-strained model was tested for boys and girls separately, inwhich parameters were free to vary. Second, the model wasconstrained whereby the regression paths for each genderwere constrained to be equal. The paths for both models werethen compared. Structural invariance (i.e., the models weredifferent based on child gender) was determined if Delta chi-square (Δχ2) significantly differed from zero (p < 0.05).

Results

Preliminary Analyses

Table 1 presents the means and standard deviations (SDs) ofparenting dimensions and child ADHD symptoms for non-referred and referred children separately, and summarises theresults of t-tests (significant differences and Cohen’s D effectsizes) comparing non-referred and referred children on parent-ing dimensions and child ADHD symptoms. Table 2 presentsthe means and SDs of parenting dimensions and child ADHDsymptoms for non-referred girls and boys separately, andsummarises the results of t-tests (significant differences andCohen’s D effect sizes) comparing girls and boys on parentingdimensions and child ADHD symptoms. Mothers’ andfathers’ Angry Parenting was higher in (i) referred childrencompared to non-referred children and (ii) non-referred boyscompared to non-referred girls across all five waves. From

J Abnorm Child Psychol (2018) 46:113–125 117

wave 3, referred children received significantly less WarmParenting from their mothers than non-referred children. Thissame difference was found for fathers’Warm Parenting in waves1 and 3.Mothers’WarmParentingwas higher for daughters thanfor sons in waves 4 and 5, and fathers’ Warm Parenting werehigher for daughters than sons from wave 2 onwards. Referredchildren had significantly higher ADHD symptoms than non-referred children across all waves (see Table 1). Further, non-referred boys had significantly higher child ADHD symptomsthan non-referred girls across all waves (see Table 2).

Table 3 presents the rates of ADHD diagnosis across wavesfor girls and boys, as well as significant differences and effectsizes. Across all waves, boys were significantly more likelythan girls to have an ADHD diagnosis.

Cross-Lagged Panel Model

First, a baseline control model was run containing only theautoregressive pathways: χ2(571, N = 1932) = 2464.38. Thepathways of interest were then added to the model (i.e., the

prospective, bidirectional pathways between parenting dimen-sions and child ADHD symptoms): χ2(523, N = 1932)=2157.99. The addition of the pathways of interest significant-ly improved the model χ2(48, N = 1932) = 306.39, p < 0.001,demonstrating that the pathways of interest were importantadditions to the model and predicted significant variance inthe model over and above the control/autoregressive model.

Good fit for the cross-lagged model examining the bi-directional effects of parenting dimensions and child ADHDwas found (RMSEA = 0.04, CFI = 0.96, TFI = 0.94). Forsimplicity, Fig. 1 presents just the significant findings regard-ing the relationships of interest (i.e., the bi-directional pro-spective relationships between parenting dimensions andchild ADHD symptoms) as well as the significantautoregressive pathways of the connecting time points. Thefull table of results is available by request from the corre-sponding author. Overall, mothers’ and fathers’ AngryParenting and child ADHD symptoms demonstrated a pro-spective, bi-directional relationship; however this was onlysignificant in the early years (ages 4- to 7-years). Further, this

Table 1 Means and standard deviations for, and significant differences and effect sizes (Cohen’s D) between, non-referred and referred children acrosswaves

Non-referred children Mean(SD) Referred children Mean(SD) [Cohen’s D effect size for non-referred versusreferred children t-test]

Wave 1 Wave 2 Wave 3 Wave 4 Wave 5 Wave 1 Wave 2 Wave 3 Wave 4 Wave 5

Mother Angryparenting

2.15(0.57)

2.13(0.55)

2.12(0.60)

2.10(0.62)

2.08(0.64)

2.72(0.67)*[0.92]

2.76(0.68)**[1.02]

2.78(0.70)***[1.01]

2.78(0.63)***[1.09]

2.57(0.83)***[0.66]

Father Angryparenting

2.26(0.59)

2.06(0.58)

2.13(0.60)

2.16(0.61)

2.10(0.63)

3.14(0.74)**[1.31]

2.59(0.71)**[0.82]

2.82(0.67)***[1.08]

2.64(0.56)***[0.82]

2.74(0.61)***[1.03]

Mother Warmparenting

4.42(0.43)

4.43(0.47)

4.32(0.53)

4.27(0.85)

4.18(0.62)

4.45(0.33)[ns]

4.43(0.55)[ns]

4.11(0.52)*[0.40]

4.02(0.47)**[0.47]

3.98(0.61)*[0.32]

Father Warmparenting

4.07(0.53)

4.13(0.59)

4.05(0.61)

3.94(0.62)

3.82(0.68)

3.57(0.50)*[0.97]

3.92(0.78)[ns]

3.83(0.62)*[0.36]

3.81(0.68)[ns]

3.67(0.60)[ns]

Motherconsistentparenting

4.20(0.61)

4.27(0.55)

4.26(0.58)

4.27(0.59)

4.22(0.62)

3.92(0.10)[ns]

4.07(0.66)[ns]

4.12(0.57)[ns]

4.02(0.68)**[0.39]

4.21(0.58)[ns]

Fatherconsistentparenting

4.04(0.65)

4.15(0.63)

4.15(0.63)

4.11(0.62)

4.08(0.63)

3.82(0.68)[ns]

3.86(0.65)[0.45]

3.90(0.65)*[0.39]

4.10(0.71)[ns]

3.94(0.69)[ns]

ADHDsymptoms

3.14(2.20)

2.92(2.15)

2.79(2.14)

2.78(2.16)

2.53(2.11)

7.90(1.52)***[2.52]

8.55(1.90)***[2.77]

7.69(1.75)***[2.50]

7.80(2.28)***[2.26]

7.19(2.29)***[2.12]

t-tests compared non-referred children to referred children for wave and parenting dimension. E.g., Wave 1 mother angry parenting for non-referredchildren compared to Wave 1 mother angry parenting for referred children. Significances and effect sizes relate to these comparisons

ns = t-test non-significant therefore no effect size calculated

*p < 0.05

**p < 0.01

***p < 0.001

118 J Abnorm Child Psychol (2018) 46:113–125

relationship appeared more continuous over time for fathersthan mothers. In terms of uni-directional effects, higher levelsof child ADHD symptoms predicted higher levels of mothers’and fathers’ Angry Parenting throughout all waves. In addi-tion, higher levels of child ADHD symptoms at wave 1 pre-dicted decreases in Warm Parenting at wave 2 for fathers andat wave 3 for both mothers and fathers. Higher child ADHDsymptoms at preceding waves also predicted less ConsistentParenting at wave four for mothers and wave five for fathers.Increases in mothers’ Consistent Parenting at wave three alsopredicted decreases in child ADHD symptoms at wave four.Structural invariance testing indicated the cross-lag model didnot differ based on the gender of the child (Δχ2 = p > 0.05).

Latent Growth Curve Models

Two separate LGCMs were conducted. The paths specified inthe LGCMs were identical to the cross-lagged path modelhowever also included the intercept and slopes for ADHDsymptoms and each parenting dimension for both mothersand fathers. Model one, where the intercept and slope ofmothers’ and fathers’ Consistent, Angry and WarmParenting were predictors, demonstrated good fit(RMSEA = 0.02, CFI = 0.99, TFI = 0.99). Model one resultsdemonstrated that no parenting dimension significantly pre-dicted change in child ADHD symptoms over time. Due tothis null finding, no invariance testing for child gender was

Table 2 Means and standard deviations for, and significant differences and effect sizes (Cohen’s D) between, non-referred girls and boys across waves

Non-referred girls Mean(SD) Non-referred boys Mean(SD) [Cohen’s D effect size for girls versus boys t-test]

Wave 1 Wave 2 Wave 3 Wave 4 Wave 5 Wave 1 Wave 2 Wave 3 Wave 4 Wave 5

Motherangryparenting

2.10(0.53)

2.08(0.53)

2.05(0.58)

2.02(0.58)

2.00(0.62)

2.19 (0.61)**[0.16]

2.17 (0.57)**[0.16]

2.18 (0.61)***[0.22]

2.17 (0.64)***[0.24]

2.16 (0.65)***[0.25]

Father angryparenting

2.19(0.57)

2.00(0.57)

2.05(0.57)

2.07(0.59)

2.02(0.62)

2.31 (0.59)***[0.21]

2.12 (0.57)***[0.21]

2.20 (0.61)***[0.27]

2.23 (0.60)***[0.27]

2.18 (0.63)***[0.25]

Motherwarmparenting

4.43(0.41)

4.45(0.47)

4.35(0.52)

4.32(0.56)

4.22(0.60)

4.41 (0.44)[ns]

4.42 (0.47)[ns]

4.30 (0.55)[ns]

4.23 (0.59)**[0.14]

4.14 (0.63)**[0.13]

Father warmparenting

4.09(0.52)

4.18(0.58)

4.12(0.60)

4.00(0.61)

3.87(0.67)

4.06 (0.54)[ns]

4.08 (0.60)**[0.17]

3.98 (0.61)***[0.23]

3.89 (0.62)***[0.18]

3.76 (0.69)**[0.16]

Motherconsistentparenting

4.21(0.51)

4.28(0.55)

4.27(0.59)

4.29(0.59)

4.24(0.61)

4.18 (0.61)[ns]

4.25 (0.55)[ns]

4.24 (0.57)[ns]

4.25 (0.60)[ns]

4.20 (0.62)[ns]

Fatherconsistentparenting

4.04(0.64)

4.14(0.63)

4.14(0.63)

4.13(0.61)

4.08(0.61)

4.04 (0.66)[ns]

4.16 (0.63)[ns]

4.16 (0.62)[ns]

4.11 (0.63)[ns]

4.09 (0.64)[ns]

ADHDsymptoms

2.61(1.98)

2.40(1.90)

2.19(1.84)

2.20(1.93)

1.95(1.80)

3.53 (2.24)***[0.48]

3.31 (2.19)***[0.44]

3.30 (2.20)***[0.55]

3.26 (2.18)***[0.51]

3.06 (2.22)***[0.55]

t-tests compared girls to boys for wave and parenting dimension. E.g., Wave 1 mother angry parenting for girls compared to Wave 1 mother angryparenting for boys. Significances and effect sizes relate to these comparisons

ns = t-test non-significant therefore no effect size calculated

*p < 0.05

**p < 0.01

***p < 0.001

Table 3 Rates of ADHD diagnosis, significant differences and effect size (phi), for girls and boys across waves

Wave 1 Wave 2 Wave 3 Wave 4 Wave 5

Girls 0 (0.00%) 3 (0.31%) 3 (0.31%) 5 (0.52%) 9 (0.95%)

Boys 10** (1.02%)phi = −0.07

17** (1.73%; ratio = 5.67:1)phi = −0.07

34*** (3.46%; ratio = 11.34:1)phi = −0.11

35*** (3.57%; ratio = 7:1)phi = −0.11

45*** (4.59%; ratio = 5:1)phi = −11

Ratio = boy:girl ratio

*p < 0.05

**p < 0.01

***p < 0.001

J Abnorm Child Psychol (2018) 46:113–125 119

conducted. Model two, where the intercept and slope of childADHD symptoms were the predictors of the slope of mothers’and fathers’ Consistent, Angry and Warm Parenting were out-comes, also demonstrated good fit (RMSEA = 0.02,CFI = 0.99, TFI = 0.99). Model two results indicated thatincreases in child ADHD symptoms over time significantlypredicted increases in both mothers’ and fathers’ AngryParenting, as well as decreases in mothers’ Warm Parenting,over time. Child ADHD did not significantly predict any otherparenting dimensions. Structural invariance testing demon-strated these relationships were not significantly differentbased on the gender of the child (Δχ2 = p > 0.05).

Logistic Regression

The results of the logistic regression demonstrated thatmothers’ and fathers’ Angry Parenting at wave one sig-nificantly predicted a child ADHD diagnosis at wavethree. No other parenting styles were significant predic-tors of a child ADHD diagnosis. Structural invariancetesting demonstrated these relationships were not signif-icantly different based on the gender of the child(Δχ2 = p > 0.05). The full table of results can befound in the online supplementary material.

Discussion

The aim of the current study was to explore the potentiallybidirectional prospective relationships between parenting andchild ADHD, and to examine if these relationships varied as afunction of child gender. Partial support was found forHypothesis 1, that predicted bi-directional prospective relation-ships between less-than-optimal parenting and child ADHDsymptoms, with significant prospective, bi-directional relation-ships found between mothers’ and fathers’ Angry Parentingand child ADHD symptoms in our cross-lagged panel models,however these bi-directional relationships appeared confinedto early years (children aged 4- to 7-years) and were morecontinuous for fathers than for mothers. Although AngryParenting no longer predicted child ADHD symptoms fromwave three, higher child ADHD symptoms continued to pre-dict higher levels of Angry Parenting throughout all waves in auni-directional manner. Higher child ADHD symptoms atwave 1 (children age 4- and 5-years) also appeared to lead toan overall increase in less-than-optimal parenting (i.e., higheranger, lower warmth and lower consistency) when childrenwere 6- and 7-years and 10- and 11-years of age, howeverthese patterns were not identical. Child ADHD symptoms alsopredicted the overall increase in Angry Parenting across allwaves, as evidenced in LGCM two. Hypothesis 2, that the

Fig. 1 Results of the cross-lagged model.MConMothers consistent parenting,MWarMothers warm parenting,MAngMothers angry parenting, FConFathers consistent parenting, FWar Fathers warm parenting, FAng Fathers angry parenting, ADHD ADHD scores as measured on SDQ, W Wave

120 J Abnorm Child Psychol (2018) 46:113–125

bi-directional prospective relationships between less-than-optimal parenting and child ADHD symptomswould be stronger for sons than daughters, was notsupported as the modelled relationships did not differas a function of child gender.

Partial support was also found for Hypothesis 3 that pre-dicted less-than-optimal parenting (i.e., higher scores onAngry Parenting, lower scores on Consistent Parenting, andlower scores on Warm Parenting) at wave one would be pre-dictive of a child ADHD diagnosis at wave three. Onlymothers’ Angry Parenting and fathers’ Angry Parenting atwave one were predictive of a child’s ADHD diagnosis atwave three. Hypothesis 4 was not supported as these predic-tive relationships between parenting and children’s’ ADHDdiagnosis did not differ between sons and daughters.

Mothers and fathers in the current sample appear similar toexisting literature in regards to their parenting of sons anddaughters. For example, the temporal stability of theautoregressive paths for parenting dimensions in the cross-lagged panel model demonstrate that parenting remains some-what stable over time (Holden and Miller 1999; Landry et al.2001). Further, as is commonly found in parenting literature(e.g., Barnett and Scaramella 2013; Russell et al. 1998), non-referred sons and daughters in the current sample receiveddifferent levels of parenting dimensions, with sons receivinghigher levels of less-than-optimal parenting compared to girls(e.g., lower levels of warmth and higher levels of anger).

Prospective Relationships between Parenting and ChildADHD

The results of our cross-lagged panel model, LGCM two, andlogistic regression add to the existing evidence regarding theprospective relationships between parenting and child ADHD.These results provide some support for claims (e.g., Johnstonand Mash 2001) that a bi-directional relationship exists be-tween child behaviour and parenting. Our results, however,suggest this bi-directional relationship is confined to earlierin childhood and may be more consistent for fathers thanmothers, while a uni-directional relationship whereby childADHD predicts certain parenting dimensions (AngryParenting in the current sample), continues throughout child-hood and into early adolescence. Therefore, our results alsosupport arguments that child behaviour may actually be astronger or more consistent predictor of parenting across timethan the reverse relationship (Barkley 1988; Singh 2003).

As discussed, the most consistent findings that arose fromthe current analyses relates to the relationship between childADHD and mothers’ and fathers’ Angry Parenting. Our mea-sure of Angry Parenting related to parental feelings of angerand annoyance towards the child, the level of praise/approvaldirected towards their child, and the difficulty parents experi-enced in managing their child’s behaviours. Our findings that

Angry Parenting shares a bi-directional relationship with childADHD symptoms in early years before converting to a uni-directional prospective relationship (child ADHD predictingangry parenting) in middle childhood, is in line with previoussuggestions that parenting may become less influential onADHD at later stages of development (Carlson et al. 1995;Johnston and Mash 2001). The continued influence of childADHD symptoms on Angry Parenting may have flow-on ef-fects in other domains. For example, parent anger and hostilityhas been linked to children’s perceptions of parental detach-ment (Domitrovski and Bierman 2001), the adoption of amore authoritarian parenting approach (Coplan et al. 2002),parental conflict (Krishnakumar and Buehler 2000) and thedevelopment of depression in adolescence (Kaitainen et al.1999). Further, Angry Parenting behaviours are transmittedgenerationally, influencing the level of hostility children willlater display towards their offspring (Scaramella and Conger2003). Thus, ADHD treatments that lower ADHD symptomsmay assist in also lowering rates of Angry Parenting and thenegative outcomes associated with angry parenting within,and across, generations.

Our results extend the findings of previous longitudinalstudies regarding parenting and ADHD behaviours; forexample, the studies by Keown (2012) and Lifford et al.,(2008), by considering (i) bi-directional relationships, and(ii) invariance in these relationships across child genders.However, some differences are notable. Keown found less-than-optimal parenting (i.e., lower positive regard, sensitivityand warmth) was predictive of higher child ADHD symptomsacross children 4- to 7-years of age.Within the same age rangein the current sample, we found a bi-directional relationshipfor Angry Parenting, and uni-directional influences of childADHD symptoms predicting less-positive parenting (i e., low-er Warm and Consistent Parenting). Further, across a 12-month period, Lifford et al. investigated the cross-lagged re-lationships between parental rejection and ADHD symptomsin 11- to-12 year old children. Lifford et al. found higher levelsof paternal rejection predicted higher child ADHD symptoms,while higher rates of child ADHD symptoms predicted highermaternal rejection. This suggests different relationships existbetween parenting and child ADHD symptoms depend-ing on whether mothers or fathers are considered.Comparison of the findings of waves four and five ofthe current study (where children were identical in ageto the Lifford et al. study) supports their finding thatADHD symptoms predict fathers’ Angry Parenting,however extends this same finding to mothers.

Several factors may account for the differences between thepresent results and those of Lifford et al., (2008). First, withboth mothers and fathers included in a single, bi-directionalmodel in the current study, as opposed to separate models formothers and fathers in Lifford et al., less variance was avail-able in each outcome variable for significant prediction in our

J Abnorm Child Psychol (2018) 46:113–125 121

model (i.e., significant relationships were harder to find).However, given the models in the current study were strin-gently controlled with autoregressive paths and for SES, ourfindings were sensitive to predictive relationships. Second,our models extended over a longer period (children aged4- to 13-years), compared to Lifford et al. (children agedfrom 11- to 13-years). As such, we were able to examinepredictive relationships across a broader range of develop-mental periods. Further, by the ages of 11- to 13-years it ispossible the behaviours of the children in Lifford et al.were well established, with parenting having only an inci-dental, rather than predictive, effects.

The Influence of Child Gender on the ProspectiveRelationships between Parenting Dimensions and ADHD

Gender differences were found in several aspects of the cur-rent results. First, across all waves, boys had significantlyhigher ADHD symptoms than girls, and were significantlymore likely to have a diagnosis of ADHD, consistent withprevious findings (Erskine et al. 2013; Willcutt 2012).Gender differences were also found in relation to parenting,with sons more likely to receive higher levels of angry parent-ing than daughters from both mothers and fathers across allwaves. However, no gender differences were found in thepredictive relationships examined. Taken together with thefinding that sons had higher ADHD symptoms than daughtersacross all waves, this suggests boys and girls are not parenteddifferently as a function of their gender, but rather, differentialparenting may occur as a function of child behaviours. Thisidea requires further investigation and suggests that futureresearch into the relationship between parenting and gendershould consider child behaviour as an influential associate.

No differences between sons and daughters werefound in the prospective relationships between parentingand child ADHD in the current study. Thus, althoughgender differences existed in the mean levels of ADHDsymptoms and parenting dimensions, as well as in therate of diagnosis, it appears ADHD in both boys andgirls share the same relationships with mothers’ andfathers’ Angry, Warm and Consistent Parenting. Onepossibility that may inform future research is that gen-der differences may arise when comparing boys andgirls who receive similar mean levels of parenting and/or have similar levels of ADHD symptoms.

It may be that child gender more closely impacts therelationship between parenting dimensions, other thanthose investigated here, and child ADHD symptomology(e.g., autonomy support, coercion, control, responsive-ness; Holden and Miller 1999; Skinner et al. 2005).Further research is needed. It may also be that childgender impacts the relationship between parenting styles(i.e., the more stable aspects of parenting and the

emotional climate in which parenting occur; Darlingand Steinberg 1993) and child ADHD more so than therelationship between parenting dimensions and childADHD. As parenting dimensions appear to be more in-fluenced by child behaviours than are parenting styles, itmay be that parenting styles are more influenced by childgender. For example, the authoritative and/or authoritari-an styles may hold significant differences in their rela-tionships to child ADHD based on child gender than domore transient and reactive parenting dimensions.

It has also been suggested that child gender may be astronger moderator when considering ADHD subtypesrather than ADHD-combined (Bauermeister et al.2007). Thus, the current study may have failed to findsignificant child gender differences given our measureof ADHD assessed the combined subtype (i.e., includedquestions regarding both the inattentive and hyperactive/impulsive). It may be fruitful for future research to ex-amine ADHD subtypes when exploring child genderdifferences in prospective relationships between parent-ing and child ADHD.

Limitations

The authors acknowledge limitations of the current study. It ispossible the mother-report measurement of child ADHDsymptomology on the SDQ may have overestimated/underestimated symptoms. Despite the established validityand reliability of the SDQ, previous research has demonstrat-ed some inconsistency between SDQ ratings between infor-mants (e.g., mother, father, teacher, self; Stokes et al. 2014).Thus, different relationships may have been found in ourmodels had different ADHD informants been used. Further,ratings of parenting dimensions and child ADHD symptomsrelied on self-report and informant-report, respectively. Thismay have impacted the validity and reliability of ratings com-pared to a more objective measure of these variables, such asbehavioural observation. Identification of the clinical samplewas also based on parent report, and although a further ques-tion regarding medication use was used in an attempt to verifyclinical status, it is not possible to determine reliability of self-reported diagnosis.

There is also Type 1 error potential given the multi-ple analyses and pathways in the current study.However, confidence that true relationships were foundis increased by (i) the relationships being in the expect-ed directions, and (ii) the fact significant results werefound despite the comprehensive autoregressive/controlmodel. In order to examine both maternal and paternalparenting styles, non-nuclear families (e.g., single-parentfamilies, same-sex parents) were excluded from the cur-rent analysis, thus precluding generalisation of our find-ings to non-nuclear families. Unfortunately, limitations

122 J Abnorm Child Psychol (2018) 46:113–125

existed in the measurement of culture in the currentdata, and thus it was unable to be included in the anal-ysis. However, it will be important for future studies toconsider the role cultural plays in this area (at the leastas a control variable), particularly in countries with highcultural variability.

Conclusion

The results of the current study suggest implications forthe psychosocial interventions for child ADHD. In par-ticular, they suggest that treatment should occur withina systemic framework; at a minimum, at the child andparent levels; should involve both parents in dual-familyhouseholds; and that similar interventions may be effec-tive for both boys and girls. Further, in the early child-hood years (between 4- to 7-years of age), addressinglevels of Angry Parenting may help to lessen ADHDsymptoms in children. Ensuring behaviours associatedwith angry parenting are addressed in parent trainingprograms for child ADHD, particularly given the possi-ble flow-on effects of angry parenting into other do-mains of family functioning, may be beneficial.Despite the null findings of the current study in termsof the influence of child gender on the relationship be-tween parenting and child ADHD, the strong associationin existing literature between child gender and ADHD,and child gender and parenting, suggests effects arelikely occurring between these constructs, though possi-bly not in the specific parenting dimensions investigatedin the current study. Continued investigations are neededinto this notion in order to help understand theaetiology and maintenance factors of child ADHD.

Acknowledgements None.

Compliance with Ethical Standards

Funding There was no funding for this study.

Conflict of Interest David H. Demmer declares he has no conflict ofinterest. Francis Puccio declares he has no conflict of interest. Mark A.Stokes declares he has no conflict of interest. Jane A. McGillivray de-clares she has no conflict of interest. Merrilyn Hooley declares she has noconflict of interest.

Ethical Approval All procedures performed in studies involving hu-man participants were in accordance with the ethical standards of theinstitutional and/or national research committee and with the 1964Helsinki declaration and its later amendments or comparable ethicalstandards.

Informed Consent Informed consent was obtained from all individualparticipants included in the study.

References

American Psychiatric Association. (2013). Diagnostic and statisticalmanual of mental disorders (5th ed.). Arlington: AmericanPsychiatric Publishing.

Australian Institute of Family Studies. (2013). The longitudinal study ofAustralian Chilren: 2011–12 annual report. Melbourne: AustralianInstitute of Family Studies Retrieved from: https://www.dss.gov.au/our-responsibilities/families-and-children/publications-articles/growing-up-in-australia-the-longitudinal-study-of-australian-children-2011-12-annual-report.

Barbaresi, W. J., Katusic, S. K., Colligan, R. C., Weaver, A. L., &Jacobsen, S. J. (2007). Long-term school outcomes for children withattention-deficit/hyperactivity disorder: a population-based perspec-tive. Journal of Developmental & Behavioral Pediatrics, 28, 265–273.

Barkley, R. A. (1988). The effects of methylphenidate on the interactionsof preschool ADHD children with their mothers. Journal of theAmerican Academy of Child & Adolescent Psychiatry, 27, 336–341.

Barnett, M., & Scaramella, L. (2013). Mothers’ parenting and child sexdifferences in behaviour problem among African American pre-schoolers. Journal of Family Psychology, 27, 773–783.

Bauermeister, J. J., Shrout, P. E., Chávez, L., Rubio-Stipec, M., Ramírez,R., Padilla, L., et al. (2007). ADHD and gender: are risks and se-quela of ADHD the same for boys and girls? Journal of ChildPsychology and Psychiatry, 48, 831–839.

Biederman, J., Monuteaux, M. C., Mick, E., Spencer, T., Wilens, T. E.,Silva, J. M., et al. (2006). Young adult outcome of attention deficithyperactivity disorder: a controlled 10-year follow-up study.Psychological Medicine, 36, 167–179.

Braza, P., Carreras, R., Muñoz, J. M., Braza, F., Azurmendi, A., Pascual-Sagastizábal, E., et al. (2015). Negative maternal and paternal par-enting styles as predictors of children’s behavioral problems: mod-erating effects of the child’s sex. Journal of Child and FamilyStudies, 24, 847–856.

Browne, M.W., & Cudeck, R. (1993). Alternate ways of assessing modelfit. In K. Bollen & J. Long (Eds.), Testing structural equationmodels (pp. 136–162). Newbury Park: Sage.

Burke, J. D., Loeber, R., Lahey, B. B., & Rathouz, P. J. (2005).Developmental transitions among affective and behavioral disordersin adolescent boys. Journal of Child Psychology and Psychiatry,and Allied Disciplines, 46, 1200–1210.

Carlson, E. A., Jacobvitz, D., & Sroufe, L. A. (1995). A developmentalinvestigation of inattentiveness and hyperactivity. ChildDevelopment, 66, 37–54.

Chang, L., Schwartz, D., Dodge, K. A., & McBride-Chang, C. (2003).Harsh parenting in relation to child emotion regulation and aggres-sion. Journal of Family Psychology, 17, 598–606.

Chang, L. R., Chiu, Y. N., Wu, Y. Y., & Gau, S. S. F. (2013). Father’sparenting and father–child relationship among children and adoles-cents with attention-deficit/hyperactivity disorder. ComprehensivePsychiatry, 54, 128–140.

Cole, D. A., & Maxwell, S. E. (2003). Testing mediational models withlongitudinal data: Questions and tips in the use of structural equationmodeling. Journal of Abnormal Psychology, 112, 558–577.

Coplan, R. J., Hastings, P. D., Lagacé-Séguin, D. G., & Moulton, C. E.(2002). Authoritative and authoritarian mothers’ parenting goals,attributions, and emotions across different childrearing contexts.Parenting, 2, 1–26.

Cussen, A., Sciberras, E., Ukoumunne, O. C., & Efron, D. (2012).Relationship between symptoms of attention-deficit/hyperactivitydisorder and family functioning: a community-based study.European Journal of Pediatrics, 171, 271–280.

Darling, N., & Steinberg, L. (1993). Parenting style as context: an inte-grative model. Psychological Bulletin, 113, 487–496.

J Abnorm Child Psychol (2018) 46:113–125 123

Davidov, M., & Grusec, J. E. (2006). Untangling the links of parentalresponsiveness to distress and warmth to child outcomes. ChildDevelopment, 77, 44–58.

Domitrovski, C. E., & Bierman, K. L. (2001). Parenting practices andchild social adjustment. Merrill-Palmer Quarterly, 47, 235–263.

Ellis, B., & Nigg, J. (2009). Parenting practices and attention-deficit/hy-peractivity disorder: new findings suggest partial specificity of ef-fects. Journal of the American Academy of Child & AdolescentPsychiatry, 48, 146–154.

Erskine, H., Ferrari, A., Nelson, P., Polanczyk, G., Flaxman, A., Vos, T.,et al. (2013). Research review: epidemiological modelling of atten-tion-deficit/hyperactivity disorder and conduct disorder for the glob-al burden of disease study 2010. Journal of Child Psychology andPsychiatry, 54, 1263–1274.

Fivush, R., Brotman, M., Buckner, J., & Goodman, S. (2000). Genderdifference in parent-child emotion narratives. Sex Roles, 42, 233–253.

Friedson, M. (2016). Authoritarian parenting attitudes and social origin:the multigenerational relationship of socioeconomic position tochildrearing values. Child Abuse & Neglect, 51, 263–275.

Gollob, H. F., & Reichardt, C. S. (1987). Taking account of time lags incausal models. Child Development, 58, 80–92.

Goodman, R. (1997). The strengths and difficulties questionnaire: a re-search note. Journal of Child Psychology and Psychiatry, 38, 581–586.

Goodman, R., Ford, T., Simmons, H., Gatward, R., &Meltzer, H. (2000).Using the strengths and difficulties questionnaire (SDQ) to screenfor child psychiatric disorders in a community sample. The BritishJournal of Psychiatry, 177, 534–539.

Graziano, P. A., McNamara, J. P., Geffken, G. R., & Reid, A. (2011).Severity of children’s ADHD symptoms and parenting stress: a mul-tiple mediation model of self-regulation. Journal of Abnormal ChildPsychology, 39, 1073–1083.

Haack, L. M., Villodas, M. T., McBurnett, K., Hinshaw, S., & Pfiffner, L.J. (2016). Parenting mediates symptoms and impairment in childrenwith ADHD-inattentive subtype. Journal of Clinical Child &Adolescent Psychology, 45, 155–166.

Harold, G. T., Leve, L. D., Barrett, D., Elam, K., Neiderhiser, J. M.,Natsuaki, M. N., et al. (2013). Biological and rearing mother influ-ences on child ADHD symptoms: revisiting the developmental in-terface between nature and nurture. Journal of Child Psychologyand Psychiatry, 54, 1038–1046.

Hawes, D. J., & Dadds, M. R. (2004). Australian data and psychometricproperties of the strengths and difficulties questionnaire. Australianand New Zealand Journal of Psychiatry, 38, 644–651.

Holden, G. W., & Miller, P. C. (1999). Enduring and different: a meta-analysis of the similarity on parents’ child rearing. PsychologicalBulletin, 125, 223–254.

Hoza, B. (2007). Peer functioning in children with ADHD. Journal ofPediatric Psychology, 32, 655–663.

Johnston, C., & Mash, E. (2001). Families of children with attention-deficit/hyperactivity disorder: review and recommendations for fu-ture research. Clinical Child and Family Psychology Review, 4,183–207.

Kaitainen, S., Raeikkoenen, K., Keskivaara, P., & Keltikangas-Jaervinen,L. (1999). Maternal child-rearing attitudes and role satisfaction andchildren’s temperament as antecedents of adolescent depressive ten-dencies: follow-up study of 6- to 15-year-olds. Journal of Youth andAdolescence, 28, 139–163.

Keenan, K., & Shaw, D. (1997). Developmental and social influences onyoung girls’ early problem behavior. Psychological Bulletin, 121,95–113.

Keown, L. (2012). Predictors of boys’ ADHD symptoms from early tomiddle childhood: the role of father-child and mother-child interac-tions. Journal of Abnormal Child Psychology, 40, 569–581.

Keown, L. J., &Woodward, L. J. (2002). Early parent–child relations andfamily functioning of preschool boys with pervasive hyperactivity.Journal of Abnormal Child Psychology, 30, 541–553.

Kessler, C., Berglund, P., Demler, O., Jin, R., Merikangas, K., &Walters,E. (2005). Lifetime prevalence and age-of-onset distributions ofDSM-IV disorders in the national comorbidity survey replication.Archives of General Psychiatry, 62, 593–602.

Krishnakumar, A., & Buehler, C. (2000). Interparental conflict and par-enting behaviours: a meta-analysis. Family Relations, 49, 25–44.

Lamb, M. E. (2004). The role of the father in child development (4th ed.).Hoboken: John Wiley & Sons.

Landry, S. H., Smith, K. E., Swank, P. R., Assel, M. A., & Vellet, S.(2001). Does early responsive parenting have a special importancefor children’s development or is consistency across time early child-hood necessary? Developmental Psychology, 37, 387–403.

Lifford, K. J., Harold, G. T., & Thapar, A. (2008). Parent-child relation-ships and ADHD symptoms: a longitudinal analysis. Journal ofAbnormal Child Psychology, 36, 285–296.

Lloyd, K., & Devine, P. (2006). Parenting practices in northern England:evidence from the northern England household panel survey. ChildCare in Practice, 12, 365–376.

MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Poweranalysis and determination of sample size for covariance structuremodeling. Psychological Methods, 1, 130–149.

Meinzer, M. C., Pettit, J. W., Waxmonsky, J. G., Gnagy, E., Molina, B. S.,& Pelham, W. E. (2016). Does childhood attention-deficit/hyperac-tivity disorder (ADHD) predict levels of depressive symptoms dur-ing emerging adulthood? Journal of Abnormal Child Psychology,44, 787–797.

National Health and Medical Research Council. (2012). Clinical practicepoints on the diagnosis, assessment and management of ADHD inchildren and adolescents. Retrieved from: https://www.nhmrc.gov.au/guidelines-publications/mh26.

Patterson, G., & Sanson, A. (1999). The association between behaviouraladjustment to temperament, parenting and family characteristicsamong 5 year-old children. Social Development, 8, 293–309.

Pinderhughes, E. E., Dodge, K. A., Bates, J. E., Pettit, G. S., & Zelli, A.(2000). Discipline responses: influences of parents’ socioeconomicstatus, ethnicity, beliefs about parenting, stress, and cognitive-emotional processes. Journal of Family Psychology, 14, 380.

Polanczyk, G. V., Willcutt, E. G., Salum, G. A., Kieling, C., & Rohde, L.A. (2014). ADHD prevalence estimates across three decades: anupdated systematic review and meta-regression analysis.International Journal of Epidemiology, 43, 434–442.

Reid, R., Gonzalez, J. E., Nordness, P. D., Trout, A., & Epstein, M. H.(2004). A meta-analysis of the academic status of students withemotional/behavioral disturbance. Journal of Special Education,38, 130–143.

Rogers, M. A., Wiener, J., Marton, I., & Tannock, R. (2009). Parentalinvolvement in children's learning: comparing parents of childrenwith and without attention-deficit/hyperactivity disorder (ADHD).Journal of School Psychology, 47, 167–185.

Rothman, K. J., Lash, T. L., & Greenland, S. (Eds.). (2008). Modernepidemiology. Philadelphia: Lippincott Williams & Wilkins.

Russell, A., Aloa, V., Feder, T., Glover, A., Miller, H., & Palmer, G.(1998). Sex-based differences in parenting styles in a sample withpreschool children. American Journal of Psychology, 50, 89–99.

Russell, A. E., Ford, T., Williams, R., & Russell, G. (2015). The associ-ation between socioeconomic disadvantage and attention deficit/hyperactivity disorder (ADHD): a systematic review. ChildPsychiatry & Human Development, 47, 1–19.

Sanders, M. R., Gooley, S., & Nicholson, J. (2000). Early intervention inconduct problems in children. In R. Rosky, A. O’Hanlon, G. Matin,& C. Davis (Eds.),Clinical approaches to early intervention in childand adolescent mental health (Vol. 3). Bedford Park: AustralianEarly Intervention Network for Mental Health in Young People.

124 J Abnorm Child Psychol (2018) 46:113–125

Scaramella, L. V., & Conger, R. D. (2003). Intergenerational continuity ofhostile parenting and its consequences: the moderating influence ofchildren’s negative emotional reactivity. Social Development, 12,420–439.

Selig, T. D., & Little, J. P. (2012). Autoregressive and cross-lagged panelanalysis for longitudinal data. In B. Laursen, T. D. Little, & N. A.Card (Eds.), Handbook of developmental research methods. NewYork: The Guilded Press.

Singh, I. (2003). Boys will be boys: fathers’ perspectives on ADHDsymptoms, diagnosis and drug treatment. Harvard Review ofPsychiatry, 11, 308–316.

Skinner, E., Johnson, S., & Snyder, T. (2005). Six dimensions of parent-ing: amotivational model. Parenting: Science and Practice, 5, 175–235.

Soloff, C., Lawrence, D., & Johnstone, R. (2005). LSAC technical paperno. 1: sample design. Melbourne: Australian Institute of FamilyStudies.

Statistics Canada. (2000). National longitudinal survey of children andyouth (NLSCY) cycle 3 survey instruments: parent questionnaire.Ottawa: Author.

Steiger, J. H. (1989). Causal modelling: a supplementary module forSYSTAT and SYGRAPH. Evanston: SYSTAT.

Stokes, M., Mellor, D., Yeow, J., & Hapidzal, N. (2014). Do parents,teachers and children use the SDQ in a similar fashion? Qualityand Quantity, 48, 983–1000.

Stone, L. L., Otten, R., Engels, R. C., Vermulst, A. A., & Janssens, J. M.(2010). Psychometric properties of the parent and teacher versionsof the strengths and difficulties questionnaire for 4- to 12-year-olds:A review. Clinical Child and Family Psychology Review, 13, 254–274.

Tandon, M., Tillman, R., Spitznagel, E., & Luby, J. (2014). Parentalwarmth and risks of substance use in children with attention-defi-cit/hyperactivity disorder. Addiction Research & Theory, 22, 239–250.

Tripp, G., Schaughency, E. A., Langlands, R., & Mouat, K. (2007).Family interactions in children with and without ADHD. Journalof Child and Family Studies, 16, 385–400.

Willcutt, R. (2012). The prevalence of DSM-IVattention-deficit/hyperac-tivity disorder: a meta-analytic review. Neurotherapeutics, 9, 490–499.

Wright, A. W., Parent, J., Forehand, R., Edwards, M. C., Conners-Burrow, N. A., & Long, N. (2013). The relation of parent and childgender to parental tolerance of child disruptive behaviors. Journal ofChild and Family Studies, 22, 779–785.

Zubrick, S. R., Lucas, N., Westrupp, E. M., & Nicholson, J. M. (2014).Parenting measures in the longitudinal study of Australian children:construct validity and measurement quality, waves 1 to 4. Canberra:Department of Social Services.

J Abnorm Child Psychol (2018) 46:113–125 125