واکنش والدین به احساسات منفی کودکان: رابطه با تنظیم احساسات در کودکان مبتلا به اختلال اضطراب

Abstract Research has demonstrated that parental reactions to children's emotions play a significant role in the development of children's emotion regulation (ER) and adjustment. This study compared parent reactions to children's negative emotions between families of anxious and non-anxious children (aged 7–12) and examined associations between parent reactions and children's ER. Results indicated that children diagnosed with an anxiety disorder had significantly greater difficulty regulating a range of negative emotions and were regarded as more emotionally negative and labile by their parents. Results also suggested that mothers of anxious children espoused less supportive parental emotional styles when responding to their children's negative emotions. Supportive and non-supportive parenting reactions to children's negative emotions related to children's emotion regulation skills, with father's non-supportive parenting showing a unique relationship to children's negativity/lability

Introduction Difficulty regulating negative emotion is emphasised in etiological and maintenance models of anxiety (Mennin et al., 2005, Southam-Gerow and Kendall, 2000, Suveg et al., 2010 and Thompson, 2001). Evidence of these difficulties has been reported in studies comparing children with an anxiety disorder (AD) to youth with no psychopathology. Specifically, AD children are found to have less understanding of hiding and changing emotions (Southam-Gerow & Kendall, 2000), experience negative emotion more intensely (Carthy et al., 2010 and Suveg and Zeman, 2004), are more dysregulated in their expression of emotion (Suveg & Zeman, 2004), and engage in more maladaptive and fewer problem-solving emotion regulations (ER) strategies (Carthy et al., 2010, Suveg et al., 2008 and Suveg and Zeman, 2004). In addition, emotional awareness, a specific facet of ER, has been shown to correlate with other emotional symptoms (e.g. depression) in children with ADs (Kerns, Comer, & Zeman, 2014). Evidence in support of the relationship between ER difficulties in anxious children and other important life domains is also starting to emerge (e.g. social functioning; Jacob, Suveg, & Whitehead, 2013). Research examining family influences on anxiety disorders has been a steady focus for more than a decade. The study of emotion socialisation and related parenting styles, in particular, have provided insight into the practices that might contribute to AD children's emotion functioning. For example, in studies involving observations of family emotion discussions, parents of AD children tend to discourage their children's emotion discussions, engage in less explanation of emotions and espouse a less positive interaction style than parents of ND children (Suveg et al., 2008 and Suveg et al., 2005). Similarly, observed parental responses to children's affect differ between nonclinical and clinical families, with mothers of AD children behaving more intrusively and with less warmth in response to child negative affect than mothers of ND children (Hudson, Comer, & Kendall, 2008). In the study conducted by Hudson and colleagues (2008), no significant group differences emerged for observed parental responses to positive child affect, suggesting that parents of AD children have particular difficulty coping with their children's negative emotion. This study focused on observed reactions to discrete episodes of emotion that emerged during the experimental session. To date, we have limited information in clinically anxious children about parent's response to negative emotions outside the laboratory setting. 1.1. Emotion socialisation within the family There is substantial empirical evidence to support the notion that parental coping with children's emotions relate to children's overall emotion socialisation and the quality of their emotional competencies (Denham et al., 1997, Denham and Kochanoff, 2002, Fabes et al., 2002 and Hooven et al., 1995). Amongst nonclinical populations, research demonstrates a significant relationship between parental reactions and children's ER skills and coping (e.g. Davidov and Grusec, 2006 and Eisenberg et al., 1996), with a strong emphasis on emotion socialisation practices that involve emotion-discussion, validation and problem solving (Gottman, 1997). Research further suggests a number of pathways through which emotion socialisation processes can occur. These include direct pathways (e.g. emotion discussion, coaching) and indirect pathways (e.g. modelling). It is considered that children's ER, an important skill underlying emotional competence, also develops through these pathways (Morris et al., 2007 and Saarni, 1999) and is fundamental to healthy psychological adjustment (Cicchetti, Ackerman, & Izard, 1995). According to Eisenberg, Cumberland, & Spinrad (1998), there are three main ways by which parents can socialise their children's emotions: emotion discussion, parent reactions to children's emotions and family expressiveness. With regard to parent–child discussion of emotion, both direct and indirect mechanisms are purported to assist children's development of emotion-related knowledge, language and skills (Denham, 1998, Dunn et al., 1991, Eisenberg et al., 1998 and Gottman et al., 1997). When parents are available to discuss emotions with their children and use these opportunities to impart both knowledge and ways to manage them, children's developing emotional awareness and regulation appears to benefit. Similarly, parent reactions to children's emotions can directly influence children's developing emotion management styles, such that supportive responses tend to facilitate appropriate emotional expression, communication and ER, and non-supportive responses to children's greater use of emotional inhibition and dysregulated affect (Eisenberg and Fabes, 1992, Eisenberg and Fabes, 1994, Eisenberg et al., 1996 and Gottman, 1997). Finally, the frequency, intensity and type of emotional expression that occurs in families is suggested to contribute to children's developing emotion-related schemas, such as which emotions to express or inhibit, when to express them and the manner in which to regulate them (for a review see Dunsmore & Halberstadt, 1997). Finally, positive family expressivity has been associated with better ER strategies and fewer negative emotional displays in children (Garner, 1995 and Garner and Power, 1996). 1.2. Parental emotional styles and child emotion regulation In regard to the direct pathways of emotion socialisation, research to date has correlated specific parental reactions to some important child outcomes. For instance, children whose parents react in non-supportive ways (e.g. punitive, dismissing or minimising) tend to display more maladaptive, avoidant or inappropriate methods of ER and coping (Eisenberg et al., 1992 and Eisenberg et al., 1996) and tend to exhibit lower levels of socio-emotional competence (Jones, Eisenberg, Fabes, & MacKinnon, 2002). In contrast, children whose parents react in supportive ways (e.g. emotion-and-problem-focused and encourage emotional expression) tend to be higher in levels of socio-emotional competence (e.g., Eisenberg and Fabes, 1994, Eisenberg et al., 1996 and Roberts and Strayer, 1987). Similarly, Gottman (1997; Gottman and colleagues, 1997) proposes that parents who respond to their children's emotions in an accepting, sensitive and supportive manner will enhance the development of ER skills in their children. Research on this emotion-coaching parental style has yielded positive outcomes in relation to children's ER and other socio-emotional areas of competence, such as self-esteem and peer relationships ( Gottman, 1997 and Gottman et al., 1997). In contrast, parents who ascribe an emotion-dismissing parental style tend to offer little guidance regarding emotions and refrain from using emotional experiences as opportunities to bond or problem-solve with their child. This latter style has been associated with poorer outcomes for children, such as poorer ER skills, poorer academic coping and lower levels of socio-emotional competence ( Gottman, 1997 and Gottman et al., 1997). Amongst other clinical populations, research also shows the benefits of providing children with emotionally sensitive and supportive parenting. For instance, in children diagnosed with oppositional defiant disorder (ODD), parental emotion-coaching behaviours are related to greater child ER and more adaptive behaviours (Dunsmore, Booker, & Ollendick, 2013) and to better peer relations (Katz & Windecker-Nelson, 2004). In a sample of depressed adolescents, youth whose mothers held more proactive, coaching and insightful emotion beliefs, tended to have more adaptive emotion beliefs themselves (Hunter et al., 2011). In addition, younger children with symptoms of attention-deficit hyperactivity disorder (ADHD) showed improvements to their ER skills and hyperactivity when mothers were taught emotion socialisation skills, such as emotion coaching (Herbert, Harvey, Roberts, Wichowski, & Lugo-Candelas, 2013). Thus, whilst parents of clinically disordered children appear to espouse less optimal emotion socialisation practices than parents of non-disordered children (e.g. Katz and Windecker-Nelson, 2004, Katz et al., 2014, Suveg et al., 2005 and Suveg et al., 2008), evidence indicates that for children high in emotional lability, supportive parental emotional styles may reduce the risk of worsening emotional and behavioural difficulties (see Dunsmore et al., 2013) and may also help to attenuate symptoms. As such, a greater focus on emotion-related responses of parents may serve to guide intervention programmes for anxious children. Indeed, recent preliminary evidence from a study that coached parents to model effective ER strategies and respond adaptively to children's negative emotion showed improvements in clinical outcomes of AD children (Lebowitz, Omer, Hermes, & Scahill, 2014). In sum, findings on ER in anxious children indicate they have fundamental difficulties managing negative emotions, over and above that reported by non-anxiety disordered children. The study of parent-reported reactions to children's negative emotions more generally in daily life is yet to receive full attention in the study of childhood anxiety disorders. Parent reactions have otherwise been documented in the normative literature (e.g. see Eisenberg et al., 1998, Fabes et al., 2001, Fabes et al., 2002 and Morris et al., 2007) and in studies of at-risk children (Shaffer et al., 2012 and Suveg et al., 2011). Thus, further research is needed to incorporate the emotion-related variables of supportive (problem-focused, emotion-focused and encouragement of emotional expression) and non-supportive (minimisation, punitive and distress reactions) parental reactions in clinically anxious children. These variables further stem from emotion socialisation theory and research, having shown links with children's emotion functioning (Eisenberg et al., 1998, Fabes et al., 2001 and Fabes et al., 2002). 1.3. The current study Parent-report of reactions to children's negative emotions has not yet been assessed in a sample of children diagnosed with an anxiety disorder. Previous work has employed observation methods to capture parental behaviours and interaction/communication styles in the context of child emotion (e.g. Hudson et al., 2008, Suveg et al., 2005 and Suveg et al., 2008), but have not directly asked parents about their typical responses to child emotion. This study attempted to fill this gap by comparing self-reported parental responses of clinically anxious children to children with no psychopathology. Given the empirical links between ER difficulties and psychopathology (Casey, 1996, Southam-Gerow and Kendall, 2000, Suveg and Zeman, 2004 and Zeman et al., 2002) and that anxious children are among the clinical groups to be especially at risk of ER difficulties (Suveg & Zeman, 2004), it was of interest to determine whether parenting practices relate to ER using a sample of clinically anxious children, in particular, the role of ‘supportive’ and ‘non-supportive’ parental emotional styles, in response to children's negative emotions (e.g., fear, sadness and anger). It was hypothesised that in contrast to parents of ND children, parents of AD children would display greater use of non-supportive parental reactions and less use of supportive strategies. Consistent with previous findings, it was also expected AD children would display poorer ER skills and higher levels of dysregulated emotion than children without an anxiety disorder. Finally, it was expected that parental emotional styles involving supportive reactions would relate to and predict better ER skills in children.

3. Results 3.1. Descriptive measures There were no differences in mean age between anxious children and nonclinical children, t(132) = −0.11, p > .05 (anxious M = 9.59 years, SD = 1.84, nonclinical M = 9.63 years, SD = 1.91). A Chi-square test for independence (with Yates Continuity Correction, used to reduce the error in approximation) indicated that children's gender did not differ between the clinical and nonclinical groups, χ2(1, N = 134) = .0, p > .05 (clinical = 48% male, 52% female; nonclinical = 46% male, 35% female). There were also no differences in family income between the clinical and nonclinical groups, χ2(3, N = 126) = .54, p > .05 (clinical = 75% of families earn over $80, 000; nonclinical = 78% of families earn over $80,000). The mean scores for both child and parent measures of symptomatology for the clinical and nonclinical groups are presented in Table 1 and Table 2, respectively. Anxious children were found to have significantly higher scores on the SCAS and the SDQ than nonclinical children. For the DASS, mothers and fathers of anxious children had significantly higher stress scores than mothers of nonclinical children. In addition, fathers of anxious children reported significantly higher symptoms of depression than fathers of nonclinical children. Differences between groups on the symptom measures provide support for the distinction between the clinical and nonclinical groups. Table 1. Means and standard deviations for measures of symptomatology across groups. Clinical Nonclinical Questionnaire M SD M SD SCAS – Mother 34.61a 14.35 9.27b 6.04 SCAS – Father 32.17a 13.22 10.02b 16.21 SDQ – Mother 15.08a 6.80 5.43b 4.13 SDQ – Father 15.13a 7.20 5.10b 3.50 Clinical Nonclinical M SD M SD Mothers Depression 7.2a 6.48 5.1a 5.72 Anxiety 5.46a 5.82 4.64a 6.28 Stress 14.28a 9.46 10.26b 6.9 Fathers Depression 7.94a 8.9 4.86b 5.38 Anxiety 3.9a 5.84 3.12a 4.14 Stress 14.26a 8.44 9.88b 6.02 Note. Means sharing superscripts are not significantly different at the critical alpha (p < .01). SCAS = Spence Child Anxiety Scale Scales; SDQ = Strengths and Difficulties Questionnaire (Total Difficulties). Table options Table 2. Means and standard deviations for parent reactions across groups. Clinical Nonclinical CCNES Subscale M SD M SD Mothers Punitive Reactions 29.76a 8.61 28.39a 7.92 Minimisation Reactions 30.71a 11.34 32.66a 10.69 Emotion-Focused 64.86a 9.22 70.45b 8.30 Problem-Focused 67.09a 8.98 72.95b 5.73 Expressive Encouragement 55.88a 12.38 60.60a 10.25 Fathers Punitive Reactions 30.31a 7.95 27.40a 8.75 Minimisation Reactions 36.64a 10.96 35.49a 12.45 Emotion-Focused 62.47a 10.19 65.28a 11.26 Problem-Focused 63.29a 9.39 67.00a 10.80 Expressive Encouragement 50.90a 14.00 48.35a 16.80 Note. Means sharing superscripts are not significantly different at the critical alpha (p < .01). CCNES = Coping With Children's Negative Emotions Scale. Table options 3.2. Parental reactions to children's negative emotions Examination of maternal reactions revealed significant differences between mothers of anxious children and mothers of nonclinical children on the Emotion Focused (EF) subscale, t(128) = −3.60, p < .01; d = 0.6 and the Problem Focused (PF) subscale, t(126) = −4.34, p < .01; d = 0.8. Results indicated that mothers of non-anxious children reported using more Emotion- and Problem-Focused Reactions than mothers of anxious children. Comparisons of fathers between the clinical and nonclinical groups revealed no significant differences on any of the subscales of the CCNES. The mean scores for both mothers and fathers on the subscales of the CCNES are presented in Table 2. Supportive and non-supportive parenting variables were also examined to compare maternal and paternal responses. Results within-groups revealed that mothers (M = 194.24, SD = 25.09) reported significantly more supportive reactions than fathers (M = 178.83, SD = 31.72), t(98) = 4.02, p < .0005 (two-tailed); d = 0.8. No difference was found between mothers and fathers for non-supportive parenting. A mixed between-within subjects analysis of variance was also conducted to examine the influence of group (Anxious, Non-Anxious) on parental reactions, across mothers and fathers. For supportive parenting, there was a substantial main effect, Wilka Lambda = .85, F(1, 97) = 17.81, p < .00005, partial eta squared = .16, with both groups showing less supportive parenting from fathers. The main effect comparing groups was also significant, F(1, 97) = 5.962, p < .05, partial eta squared = .058, suggesting higher levels of overall parental support to non-anxious children compared to anxious children. No significant findings emerged for non-supportive parenting. Effects of child sex on parent reactions for mothers and fathers were also examined. No significant differences emerged between girls and boys for maternal supportive and non-supportive parenting, F(1, 83) = 1.64, p > .05, partial eta squared = .02 and F(1, 83) = 1.47, p > .05, partial eta squared = .02, respectively. There were also no significant differences between girls and boys for paternal supportive and non-supportive parenting, F(1, 83) = 2.18, p > .05, partial eta squared = .03 and F(1, 83) = .683, p > .05, partial eta squared = .01, respectively. 3.3. Children's emotion regulation The mean scores and standard deviations for both parent- and self- reported ER skills are seen in Table 3. For the Regulation subscale on the ERC, both mothers and fathers in the clinical group rated their children as having greater difficulty regulating their emotions than mothers and fathers in the nonclinical group, t(134) = −7.45, p < .01; d = 1.29 and t(115) = −5.82, p < .01; d = 1.09, respectively. On the Negativity/Lability subscale, mothers and fathers of anxious children rated their children as being more inflexible, labile and emotionally negative than parents in the nonclinical group, t(121) = 7.12, p < .01; d = 1.29 and t(112) = 7.01, p < .01; d = 1.32, respectively. For the EESC, anxious children rated themselves as significantly less aware of their emotions on the Poor Awareness subscale, t(127) = 6.82, p < .01; d = 1.21. Anxious children were also less likely to express their emotions on the Expressive Reluctance subscale compared to ratings of non-anxious children, t(127) = 4.59, p < .01; d = 0.81. Using the CEMS, anxious children rated themselves on the Regulation subscales as having greater difficulty regulating feelings of both sadness and anger than non-anxious children, t(129) = −3.84, p < .01; d = 0.68 and t(129) = -5.94, p < .01; d = 1.05, respectively. For the Dysregulated subscales, anxious children rated themselves as significantly more dysregulated in their expression of both sadness and anger compared to non-anxious children, t(128) = 3.35, p < .01; d = 0.59 and t(134) = 2.95, p < .01; d = 0.51, respectively. No differences on the Inhibition subscales for sadness and anger were found between anxious and non-anxious children, t(129) = 1.63, p > .01; d = 0.29 and t(131) = −1.28, p > .01; d = 0.22, respectively. Table 3. Means and standard deviations for child emotion regulation across groups. Clinical Nonclinical Questionnaire/Subscale M SD M SD ERC – Mother Emotion Regulation 24.42a 3.66 28.63b 2.83 Lability/Negativity 31.15a 6.86 23.48b 5.03 ERC – Father Emotion Regulation 24.94a 3.50 28.35b 2.60 Lability/Negativity 31.38a 5.73 24.04b 5.29 EESC – Child Poor Awareness 20.73a 6.15 13.82b 5.28 Expressive Reluctance 20.91a 5.55 16.37b 5.67 CEMS – Child Sadness Inhibition 7.66a 2.10 7.05a 2.17 Anger Inhibition 6.89a 2.07 7.38a 2.33 Sadness Regulation 9.49a 1.91 10.95b 2.37 Anger Regulation 7.38a 2.05 9.45b 1.88 Sadness Dysregulation 5.66a 1.60 4.73b 1.55 Anger Dysregulation 5.45a 1.73 4.60b 1.60 Note. Means sharing superscripts are not significantly different at the critical alpha (p < .01). ERC = Emotion Regulation Checklist; EESC = Emotion Expression Scale for Children; CEMS = Child Emotion Management Scales. Table options 3.4. Analysis of parent reactions to children's negative emotions in relation to child emotion regulation skills Since mothers and fathers of anxious children respectively reported higher stress and depression scores on the DASS than parents of non-anxious children, correlations between parental psychopathology and the CCNES were examined to determine if it was an appropriate covariate. Results of the bivariate correlations between the Stress and Depression subscales of the DASS and the six subscales of the CCNES revealed no significant relationships. However, it was decided to include parental psychopathology in order to control for the influence of parent symptomatology on children's ER. In regards to parent reactions on the CCNES, correlation analyses showed that only three of the subscales were correlated between mother and father responses: Distress Reactions, Minimisation Reactions and Emotion-Focused Reactions. Due to a lack of agreement between mothers and fathers on every scale of the CCNES, it was decided to examine each GLM model separately for mothers and fathers. 3.5. GLM models examining parent reactions with parent-reported child emotion regulation The Emotion Regulation and Negativity/Lability subscales from the ERC were used as the dependent variables. For the models examining maternal responses, the R squared values were .39 and .45 for Emotion Regulation and Negativity/Lability, respectively. For the models examining paternal responses, the R squared values were .32 and .43 for Emotion Regulation and Negativity/Lability, respectively. Only maternal supportive parenting significantly predicted children's Emotion Regulation, b = .038, t(98) = 2.84, p < .01, partial eta squared = .07, such that higher levels of supportive parenting was associated with higher children's ER skills. Father's non-supportive parenting significantly predicted both child's Negativity/Lability, b = 0.067, t(85) = 2.80, p < .01, partial eta squared = .079 and Emotion Regulation, b = −.03, t(88) = −1.98, p < .05, partial eta squared = .04, respectively. However, mother's non-supportive parenting did not predict either children's Negativity/Lability or Emotion Regulation. Group was also a significant predictor for maternal-reported ER on the Emotion Regulation subscale, b = −3.36, t(98) = −3.67, p < .0005, partial eta squared = .12 and Lability/Negativity subscales, b = .8.75, t(98) = 4.88, p < .0005, partial eta squared = .22. For paternal-reported child ER, group was a significant predictor for both the Emotion Regulation subscale, b = −3.74, t(88) = −4.59 p < .0005, partial eta squared = .20 and Lability/Negativity subscales, b = 6.93, t(85) = 4.14, p < .0005, partial eta squared = .18. For parent psychopathology, both maternal and paternal symptoms were found to be a significant predictor for the Lability/Negativity subscales, b = 0.18, t(98) = 2.34, p < .05, partial eta squared = .03 and b = 0.15, t(85) = 2.15, p < .05, partial eta squared = .07, respectively, such that higher scores on the DASS related to higher levels of child dysregulated emotion. Parent psychopathology was not found to be a significant predictor for the Emotion Regulation subscale. Age and sex were not found to be significant predictors for the ERC subscales (ps > .05). 3.6. GLM models examining parent reactions with child-reported emotion regulation Results for supportive and non-supportive parental reactions on children's self-reported ER skills are presented in Table 4 for mothers and Table 5 for fathers. The individual subscales were examined separately in the GLM analyses, with the exception of EESC that provides an interpretable total scale score. Table 4. Summary of GLM analyses for child emotion regulation with maternal supportive and non-supportive reactions, group, age and parent psychopathology. Supportive parenting Non-supportive parenting Group Age Parent psychopathology Dependent variables B t p ηp2 B t p ηp2 B t p ηp2 B t p ηp2 B t p ηp2 R2 CEMS Sadness Inhibition −.02 −2.30 .02* .06 −.01 −1.30 .20 .02 .06 .09 .93 .00 .15 −1.17 .24 .03 −.02 −.70 .49 .01 .10 Anger Inhibition −.01 −.63 .53 .00 −.01 −1.40 .16 .02 −.87 −1.82 .20 .02 .35 2.75 .01** .08 −.02 −.85 .40 .01 .13 Sadness Regulation .01 .99 .33 .01 −.03 −3.22 .00*** .12 −1.36 −2.18 .03* .05 .25 2.04 .04* .05 −.03 −1.2 .23 .02 .28 Anger Regulation .01 1.45 .15 .02 −.01 −1.41 .16 .02 −2.26 −3.75 .00*** .14 .22 1.79 .08 .04 −.01 −.45 .66 .04 .33 Sadness Dysregulation −.01 −.95 .35 .01 .01 1.10 .28 .01 .86 1.81 .07 .04 −1.1 −1.43 .16 .02 .01 .75 .45 .01 .17 Anger Dysregulation −.02 −2.31 .02* .06 .00 .05 .96 .00 .46 .96 .34 .01 .01 .09 .93 .00 .04 2.4 .02* .06 .20 EESC −.11 −2.45 .02* .07 −.05 −.88 .38 .01 9.45 3.04 .00*** .10 .32 .54 .59 .00 .04 .40 .69 .00 .30 Note. CEMS = Child Emotion Management Scales; EESC = Emotion Expression Scale for Children. Overall R2 also includes Sex in the model. * p < .05. ** p < .01. *** p < .001. Table options Table 5. Summary of GLM analyses for child emotion regulation with paternal supportive and non-supportive reactions, group, age and parent psychopathology. Supportive parenting Non-supportive parenting Group Age Parent psychopathology Dependent variables B t p ηp2 B t p ηp2 B t p ηp2 B t p ηp2 B t p ηp2 R2 CEMS Sadness Inhibition .00 .22 .83 .00 −.00 −.40 .69 .00 .92 1.44 .16 .02 .18 1.45 .15 .04 .02 .52 .61 .00 .07 Anger Inhibition −.00 −.30 .77 .00 .00 .40 .74 .00 −.90 −1.33 .19 .02 .31 2.43 .02* .07 −.04 1.20 .24 .02 .12 Sadness Regulation −.01 −.75 .46 .01 −.03 −2.44 .02* .07 −.90 −1.38 .17 .02 .26 2.40 .06 .45 .02 .53 .60 .00 .21 Anger Regulation −.01 −1.88 .08 .04 −.00 −.12 .91 .00 −1.84 −2.80 .01** .1 .21 1.59 .12 .03 −.00 −.1 .92 .00 .22 Sadness Dysregulation .00 .32 .75 .00 .01 .98 .33 .01 .09 .19 .85 .00 −.14 −1.65 .10 .03 −.01 −.35 .73 .03 .09 Anger dysregulation .03 .00 .54 .59 .00 −.02 −.04 .97 .00 −.06 −.55 .59 .00 .04 1.52 .13 .00 .06 EESC .03 .66 .51 .01 −.00 −.02 .98 .00 10.78 3.28 .00*** .13 .96 1.52 .13 .03 .34 2.28 .03* .07 .31 Note. EESC = Emotion Expression Scale for Children; CEMS = Child Emotion Management Scales. Overall R2 also includes Sex in the model. * p < .05. ** p < .01. *** p < .001. Table options In the GLM models examining maternal supportive and non-supportive reactions, a number of significant predictors emerged for children's self-reported ER (see Table 4). In all cases, child sex was not found to be a significant predictor (p < .05). For Sadness Inhibition (CEMS-SI), supportive parenting was a significant predictor such that higher levels of supportive parenting were associated with lower inhibition of sadness. For Anger Inhibition (CEMS-AI), age was a significant predictor, suggesting reduced inhibition of anger with increasing age. For Sadness Regulation (CEMS-SR), non-supportive parenting, group and age were significant predictors such that higher levels of non-supportive parenting and clinical group membership related to less regulation for sadness, whereas increasing age related to higher regulation for sadness. For Anger Regulation (CEMS-AR), group was found to be the only significant predictor. There were no significant predictors for Sadness Dysregulation (CEMS-SD). For Anger Dysregulation (CEMS-AD), supportive parenting and psychopathology were significant predictors such that higher levels of supportive parenting related to less dysregulated anger, whereas a higher DASS score was associated with increased dysregulation for anger. For children's awareness of and willingness to express emotion (EESC), supportive parenting and group were significant predictors, relating to better awareness and expression of emotions for supportive parenting and less awareness and expression of emotions for clinical group membership. For the GLM models examining paternal reactions, non-supportive parenting was found to be significant predictor for Sadness Regulation (CEMS-SR) such that increasing levels of non-support related to poorer regulation for sadness. For inhibition of sadness (CEMS-IS), neither supportive or non-supportive, nor group, age, sex or psychopathology were significant predictors. For inhibition of anger (CEMS-IA), age was the only significant predictor such that inhibition for anger increased with age. For Anger Regulation (CEMS-AR), group was a significant predictor such that clinical group membership related to poorer regulation. For Sadness Dysregulation (CEMS-SD) and Anger Dysregulation (CEMS-AD), there were no significant predictors. For children's awareness of and willingness to express emotion (EESC), group and parent psychopathology were significant predictors, relating to less awareness and expression of emotions for higher paternal DASS scores and clinical group membership.