Help Needed

Family Processes in Child Anxiety: the Long-Term Impact of Fathers and Mothers

Kaela L. Stuart Parrigon1 & Kathryn A. Kerns1

Published online: 23 December 2015 # Springer Science+Business Media New York 2015

Abstract The current study investigated links between parent and child anxiety, and family process factors, over a 9 year period. Specifically, we examined the role of partner conflict, attachment security and parental autonomy granting in explaining changes in child, father, and mother anxiety symp- toms. We utilized data from the NICHD Study of Early Child Care and Youth Development (N=661), from when target children were in first grade (6 years), fifth grade (10 years), and 15 years of age. We tested a longitudinal path model including both fathers and mothers, and found that the model was a good fit for the data. We also found that lower attach- ment security to fathers and a restriction ofmaternal autonomy granting predicted which children maintained anxiety into ad- olescence. Partner conflict explained the association between earlier and later parental anxiety, which is a novel finding in the literature. Together, these findings suggest that fathers play an important long-term role in child anxiety, alongside the role of mothers.

Keywords Anxiety . Fathers . Mothers . Partner conflict .

Attachment security . Autonomy granting

One of the most problematic and prevalent psychological dis- orders for children and adolescents is anxiety (Albano et al. 2003 For example, research has found that children with anx- iety disorders have difficulty with school and peers (Albano et al. 2003) and with parent relationships (McLeod et al. 2007). Experiences in families have been found to predict

child anxiety (e.g., Bögels and Brechman-Toussaint 2006; Kerns et al. 2011), and thus it is important to study child anxiety within the context of the family. The extant research has largely focused on the role of mothers in child anxiety to the relative exclusion of fathers, despite accumulating evi- dence and theory that fathers are likely to have a significant impact on children’s development of internalizing symptoms (Bögels and Phares 2008). Especially absent from the litera- ture is knowledge about the long term role that fathers may play in child anxiety. Although family process models are a relatively common paradigm in the child anxiety literature, and several models of child anxiety incorporating family pro- cess factors have been proposed (e.g., Bögels and Perotti 2011; Bögels and Phares 2008; DeKlyen and Greenberg 2008), many studies have looked at one or two risk factors in isolation (Vasey and Dadds 2001).More recently, multi-risk factors models have been tested, although studies looking at the development and maintenance of child anxiety over time are still few in number.

Given that research combining and testing multiple risk factors is relatively new, in addition to a lack of understanding of the long term role of fathers, our study sought to contribute to these gaps. We examined the roles of both mothers and fathers by proposing and testing a path analysis model of child anxiety (see Fig. 1). The model specifies that parent and child anxiety affect family processes, which in turn can influence subsequent child anxiety. Thus, family processes are concep- tualized as mediating mechanisms that can account for chang- es in child anxiety across childhood and adolescence. The model focused on family processes which have been identi- fied as key factors in models of child psychopathology or anxiety (DeKlyen and Greenberg 2008; Bögels and Brechman-Toussaint 2006): partner conflict, insecure par- ent–child attachment, and parental restriction of autonomy granting. These factors are common in the literature but, to

* Kaela L. Stuart Parrigon [email protected]

1 Department of Psychological Sciences, Kent State University, 600 Hilltop Dr, Kent, OH 44240, USA

J Abnorm Child Psychol (2016) 44:1253–1266 DOI 10.1007/s10802-015-0118-4

our knowledge, have not been combined previously with fathers. In addition, research has begun to explore changes in anxiety over longer time periods, although largely with younger children. For example, Feng et al. (2008) examined anxiety trajectories over an 8 year period, when children were ages 2–10 years. Our study is one of the few to examine how family factors predict changes in anxiety from childhood to adolescence (ages 6 to 15 years). We also include multiple methods of data, which is rare in the current literature.

Parental Anxiety

Evidence has accumulated to suggest that maternal anxiety is a risk factor for child anxiety (Costa and Weems 2005; Kerns et al. 2011). Phares and Compas (1992) reviewed the effects of paternal psychopathology on children, and determined that, in general, paternal psychopathology was Bsufficient, but not necessary^ (p. 407) to influence children’s psychopathology; however, only one study in their review looked at paternal anxiety.More recent evidence demonstrates that anxiety of both parents is a risk factor for child anxiety. For example, a meta- analysis looking anxiety and depression rates in offspring of anxious parents indicated that offspring were nearly four times as likely to meet criteria for an anxiety disorder as compared to offspring of non-anxious parents (Micco et al. 2009).

Subsequent research has made an effort to propose process- es that may account for associations between parental anxiety and child anxiety. For example, Bögels and Phares (2008) suggested that attachment and parenting may account for why children of anxious parents are themselves prone to

experiencing anxiety. One hypothesis in particular is that fa- thers may influence their children’s anxiety more thanmothers as they are often involved in play with children and support children’s exploration away from home, whereas mothers may be more nurturing (fostering comfort). Evidence from a meta- analysis, which included several studies with fathers, demon- strated that parental anxiety was significantly related to ob- served parenting, and in particular autonomy granting, such that anxious parents granted less autonomy to their children (van Der Bruggen et al. 2008). Therefore, we wanted to look not just at paternal anxiety as a predictor of child anxiety, but also at three mechanisms by which fathers and mothers might contribute to child anxiety.

Partner Conflict

It has long been recognized that conflict in the marital or partner relationship affect children’s adjustment (Cummings 1998). For example, Grych et al. (1992) found that when children perceived their parents’ conflicts as stable, intense and poorly resolved, they were more likely to experience in- ternalizing problems such as anxiety. In addition, destructive forms of marital conflict such as aggression withdrawal from the relationship, threats, and child-related conflicts have been found to be particularly harmful for children (Cummings 1998). In this study, we hypothesized that parental anxiety might lead to greater partner conflict, in line with prior re- search suggesting that couples with anxiety are more likely to report lower marital quality (McLeod 1994).

Hypothesized Model of Family Processes and Anxiety

Time 1: 1 st grade Time 2: 5

th grade Time 3: Age 15 years

Fig. 1 Hypothesized model of family processes and anxiety

1254 J Abnorm Child Psychol (2016) 44:1253–1266

A number of mechanisms might account for why marital conflict is associated with child anxiety. The Emotional Secu- rity Hypothesis (Davies and Cummings 1994; Davies and Martin 2014) suggests that destructive forms of marital con- flict may undermine a child’s feelings of emotional security regarding the marital relationship, leaving children vulnerable to difficulties such as anxiety as a product of arousal and hyper-vigilance resulting from observing conflict. Alterna- tively, marital conflict might influence child anxiety by affect- ing the child’s sense of attachment security within parent– child relationships. For example, parents who are burdened by conflict may bemore likely to withdraw from their children (i.e., children may experience lower attachment security). Al- though emotional security about the marital relationship has been studied extensively, few studies have considered how attachment security within parent–child relationships may ex- plain effects associated with exposure to marital conflict (Harold et al. 2004). In addition, parents experiencing partner conflict might be over-controlling and/or absent in providing appropriate boundaries (i.e., autonomy) necessary for optimal development (Cummings and Davies 2002). We therefore hy- pothesized that partner conflict would lead to poorer parent– child attachment perhaps through withdrawal by the parent, and would affect parental autonomy granting as a result of potentially increased stress from the conflict. Specifically, we included attachment and autonomy granting as mediators between partner conflict and later child anxiety in our model (see Fig. 1 for proposed model).

Parent–child Attachment

According to attachment theory (Bowlby 1969, 1973), care- givers can act as both a safe haven to afford their child comfort in times of distress, and a secure base from which to explore the world. Children are hypothesized to be less prone to experiencing anxiety if they form secure attachments, that is, are readily able to use a parent figure as a secure base or safe haven (Bowlby 1973). Although Bowlby’s initial proposal emphasized the role of primary caregivers, who were typically mothers, fathers have been identified as important attachment figures as well, although their role in comparison to mothers is not as well understood (Bretherton 2010).

A connection between mother-child attachment and child anxiety has been established in childhood and adolescence (Brumariu and Kerns 2010; Colonessi et al. 2011; Madigan et al. 2013; Groh et al. 2012). In a narrative review by Brumariu and Kerns (2010), which examined both mother and father attachment, attachment insecurity was related to child anxiety, particularly in adolescence. This effect was clearer for mothers, as few studies examined father-child at- tachment. In contrast, Groh et al. (2012) reported on three studies looking at father-child attachment, concluding that

father-child attachment was not associated with internalizing disorders. However, these reviews differed in the studies in- cluded and the type of attachment measures included (Kerns and Brumariu 2014), thus the importance of father-child at- tachment requires greater exploration. Studies that looked spe- cifically at anxiety rather than internalizing symptoms, and at both mother-child and father-child attachment, provide some support for an association between anxiety and father-child (Verschueren andMarcoen 1999; van Eijck et al. 2012). Thus, father-child attachment relationship may be one of the factors influential in the development and maintenance of child anx- iety as proposed by B gels and colleagues (e.g., Bögels and Brechman-Toussaint 2006; Bögels and Phares 2008), al- though further evidence is needed to bolster this hypothesis.

Parental Control and Autonomy Granting

Parental control is another factor related to child anxiety which has often been studied in isolation. Although moderate levels of control (i.e., discipline) may be positive for children’s development, over-controlling parental behaviors, or a restric- tion of autonomy granting, may result in anxiety (Wood et al. 2003). In a meta-analysis, McLeod et al. (2007) found that parental over-control was related to child anxiety, and in par- ticular found that autonomy granting was the aspect of parent- ing most strongly related to child anxiety, accounting for 18% of the variance in child anxiety. Of note, the majority of stud- ies in the meta-analysis focused on mothers’ parenting. In addition, effects sizes were larger for studies that use observa- tional methods instead of questionnaire measures of parenting.

Several recent studies found that paternal controlling be- havior is an important factor in the development of child anx- iety (Greco and Morris 2002; Bögels and van Melick 2004). Edwards et al. (2010) looked at changes in child anxiety over 1 year in children ages 3–5 years, and found that perceived over-controlling behaviors by both mothers and fathers led to an increase in child anxiety. Verhoeven et al. (2012) studied child and parent reported parental over-control, rejection and autonomy granting in two samples of children, 8-12-year-olds and 13-18-year-olds, and found that child reported paternal over-control predicted adolescent anxiety. Higher levels of autonomy granting by both parents in younger children was related to greater anxiety, suggesting that excessive autonomy may be anxiety provoking for younger children, who require support and guidance. Conversely, restriction of autonomy for adolescents may hinder the growth of independence and con- fidence necessary for adolescents to explore their expanding worlds. In studies utilizing observational data, both paternal and maternal autonomy granting were important factors relat- ed to child anxiety (van Der Bruggen et al. 2008). Bögels et al. (2008) found that observed parental rearing by both parents was related to child anxiety. In particular when families had

J Abnorm Child Psychol (2016) 44:1253–1266 1255

anxious fathers, children had parents who granted less auton- omy, were more controlling, and had fathers who were less supportive of their own partners, and dominated conversations regarding their child’s anxiety. In another observational study, anxious fathers were more involved and controlling than anx- ious mothers (Teetsel et al. 2014), and anxious mothers and fathers also reported engaging in parenting behavior likely to influence their child’s anxiety (e.g., punishment of anxious behavior, modeling coping or fear).

We hypothesize that restriction of autonomy may be im- portant for anxiety during the preadolescent period, as chil- dren are becoming more independent (Kerns 2008). Although Bögels and Phares (2008) suggested parental autonomy granting from fathers may be especially important as they support children’s exploration outside the family, extant liter- ature is still quite mixed as to whether maternal and paternal autonomy granting have different influences, and thus we hy- pothesize that autonomy granting by both parents will be re- lated to child anxiety. Overall, based on prior evidence and proposed models of child anxiety (e.g., Bögels and Perotti 2011), parental autonomy granting is a factor which is likely influenced by parental anxiety, particularly when observation- al data is utilized, and is important to include in a family process model.

Study Goals

The goal of this study was to examine the role of fathers and mothers in the development of child anxiety by testing a lon- gitudinal, multi-risk factor model, incorporating multiple methods of data, with a focus on family processes. This study utilized the National Institute of Child Health and Human Development (NICHD) Study of Early Child Care Database. Parent and child anxiety were measured when the child was in first grade (time 1; 6 years), partner conflict, parent–child attachment and parental autonomy were measured when the child was in fifth grade (time 2; 10 years), and parent and child anxiety were assessed again when the child was 15 years of age (time 3). We chose these ages to examine our questions because middle childhood (around fifth grade) brings impor- tant change to the emerging adolescent’s life (e.g., growing independence, expanding world; Kerns 2008), which may in- fluence attachment and autonomy, and thereby later anxiety. Additionally, research suggests that anxiety in early and mid- dle childhood predicts later anxiety and/or other psychiatric disorders such as depression (Bittner et al. 2007) so it is im- portant to ascertain factors that may be promoting the mainte- nance of anxiety during this time.

Previous studies utilizing the NICHD database have ex- plored several factors related to anxiety. For example, Brumariu and Kerns (2013) found that temperament, peer competence and the ability to manage intense emotions, in

conjunction with attachment history, predicted child anxiety at ages 11–12 years. Other studies looked at the relationship between infant security, aggression, and anxiety in first grade (Dallaire and Weinraub 2007), or predicted toddler’s anxious and depressive symptoms from maternal separation anxiety and depressive symptoms, infant temperament, and maternal sensitivity (Warren and Simmens 2005). Kerns et al. (2011) examined several factors (mother-child relationships, child characteristics, family context) that might account for changes in child anxiety across middle childhood (ages 6 to 12 years), and found that attachment, maternal anxiety, behavioral inhi- bition, and negative life events predicted changes in anxiety. To our knowledge, research using the NICHD database has not examined processes related to adolescent anxiety or in- cluded fathers; thus, our study extends prior findings with the NICHD sample, and, more importantly, the literature as a whole, by exploring how both mothers and fathers may influence anxiety in adolescence after controlling for early anxiety.

Given our goal of understanding the long-term impact of mothers and fathers in the development and maintenance of child anxiety through the testing of a longitudinal multi-factor risk model, we put forth the following hypotheses. We expect- ed that our model (Fig. 1), specifying family process variables, would be a good fit for the data, and that parent and child anxiety at time 1 (when children were 6 years old) would be negatively associated with family processes, which would, in turn, predict higher levels of anxiety at age 15 years. In addition to testing mediational pathways, we conducted a multi-group analysis to see if gender would moderate our outcomes. The literature on gender differences is mixed; some studies show gender moderates associations between parent- ing and child anxiety (e.g., Edwards et al. 2010; van Eijck et al. 2012), but others do not (e.g., Verhoeven et al. 2012; McLeod et al. 2007), and thus we did not have a clear hypoth- esis about gender differences.

Method

Participants

Participants were from the National Institute of Child Health and Human Development (NICHD) Study of Early Child Care. Families with infants were recruited from ten different locations in the United States. Data were collected from fam- ilies in multiple waves of data collection from birth to age 15 years. In the present study, we utilized data from first grade (6 years), fifth grade (10 years), and age 15 years, which we refer to as time 1 (first grade), time 2 (fifth grade), and time 3 (age 15). The study began with 1364 families, when the sam- ple was 48 % female, 80.4 % Caucasian, 12.9 % African American, and 6.7 % other ethnicity.

1256 J Abnorm Child Psychol (2016) 44:1253–1266

Due to our interest in the role of partner conflict and fathers, we selected families comprised of a mother and father figure (biological father, adopted father, or step-father) at times 1 and 2. We coded families into two groups, intact mother-father families (families with biological fathers or adoptive fathers) and stepfather families.1 In this selected subsample (N=661, 325 girls and 336 boys; 90.8% of children were Caucasian), at time 1, 91.6 % of families were intact mother-father families and 8.4 % were stepfather families. At time 2, 89.3 % of families were intact mother-father families and 10.7 % were stepfather families. At time 3, 74.6 % of families were intact mother-father families, and 25.5 % were stepfather or single parent families. Thus, at time 3, 18% of families in our sample had become single parent households. We chose to retain fam- ilies that later divorced so as not to exclude the most partner distressed families, although we controlled for household type at time 3 in the analyses (single parent families were grouped with stepfather families). See Table 1 for sample N’s, follow- ing the specific selection criteria described above.

The present sample was not representative of the NICHD dataset as a whole, as we selected only families with two parent households. We also assessed whether participants who were lost to attrition (i.e., missing data at time 3 when outcome variables were measured) were missing data completely at random (MCAR) at time 1 and 2 to determine whether there was bias in the data. While some of the variables met the MCAR assumption (mater- nal anxiety t1, child anxiety reported by both mother and father t1, mother-reported partner conflict t2, autonomy granting by both parents t2, paternal anxiety t3, sex, race, income), others did not (paternal anxiety t1, father- reported partner conflict t2, attachment security with mothers and fathers t2, maternal anxiety t3, child anxiety t3, and household type). For those that did not meet the MCAR assumption, we ran correlations to see if there was practical significance to the estimation bias. As all of the associations between the missing data and the outcome variables were less than r=0.10, per Collins et al. (2001) and described by Graham (2009), we concluded that the variables that did not meet the MCAR assumption did not have a practically significant bias on the results.

Procedure

The NICHD study collected many more measures than were utilized for this study; thus, only the specific procedures used in this study are described here. Fathers completed question- naires about their own anxiety as well as a measure of their children’s anxiety in their home when children were in 1st grade. Mothers completed the same questionnaires in regard to themselves and the target child in the laboratory. At fifth grade, children completed a questionnaire about their security to their fathers and mothers in the laboratory. At home, the children and fathers engaged in a discussion task and planning activity designed to assess autonomy granting. This activity was also done with mothers and children in the laboratory. Additionally during the fifth grade period, information about partner conflict was collected from fathers in the home and mothers in the laboratory. When the children were 15 years of age, fathers andmothers completed a questionnaire about their own anxiety (fathers in their homes, mothers in the laborato- ry), and children completed a self-report questionnaire about their own anxiety (in the laboratory). See Table 1 for informa- tion regarding the following variables.

Measures

Parental Anxiety Symptoms Anxiety was assessed for both fathers and mothers at time 1 and time 3 with the State-Trait Anxiety Inventory (Spielberger et al. 1983; example items include: I was tense, I was worried). Fathers and mother rated how they felt during the past week on a four-point scale, from 1 (not at all) to 4 (very much). A mean score was calculated by summing the items, after reverse scoring was completed, with higher scores reflecting greater anxiety (fathers α= 0.85; mothers α=0.86).

Child Anxiety Symptoms-Parent report Fathers and mothers separately completed the Child Behavior Checklist (CBCL; Achenbach 1991a) in the home about their children at time 1. A twelve item anxiety scale was created from the CBCL, based on previous literature (Bosquet and Egeland 2006; Kerns et al. 2011): 8 items from the Depression/ Anxiety scale (31, 32, 34, 45, 50, 71, 89 and 112) and 4 additional anxiety related items (9, 29, 30 and 66). (Example items include: Worries, Can’t get his/her mind off certain thoughts/obsessions). Ameanwas calculated from these items after fathers andmothers rated items on a scale of 0 (not true of the child) to 2 (very true of the child; father rated child anxiety, α=0.71, mother rated child anxiety, α=0.70).

Child Anxiety Symptoms-Self-reportAt age 15, self-reports of anxiety were utilized, as research has demonstrated that self-reports of internalizing symptoms are more accurate than parents’ reports in adolescence (Achenbach 1991b). Anxiety

1 Household type classifications included traditional nuclear families, step-father families, two-parent extended or extended and augmented families, two-parent augmented families, nontraditional nuclear families, nontraditional step-father families, nontraditional extended or extended and augmented families, and nontraditional augmented families. All fam- ily types that were not step-father families included a father figure who was classified as the child’s biological or adoptive father. We combined these latter families into an Bintact mother-father^ group for the purposes of this study, which contrasts with the Bstep-father^ group. All data in- cluded in the step-father group was reported by or collected about step- fathers.

J Abnorm Child Psychol (2016) 44:1253–1266 1257

was measured using the Youth Self-Report (Achenbach 2001), which included the same 12 items described above (α=0.80).

Partner Conflict Partner conflict was assessed at time 2. The questionnaire, as adapted for the NICHD study, included five items for the parental/partner conflict subscale (Braiker and Kelley 1979), and 13 items from the resolution scale (Kerig 1996). Only the five conflict items were used for this study. (Examples of these items include: How often do you feel angry or resentful to your partner; what is the extent to which you communicate negative feelings to your partner). Items from the parental/partner conflict subscale were rated on a 1 (not at all) to 9 (very much) scale, with higher scores reflecting greater amounts of conflict. The alpha level for the conflict scale for fathers was (α=0.81) and for mothers was (α=0.83) for this sample.

Parent–child Attachment To assess children’s perceptions of father-child and mother-child attachment, the Relatedness and Security Questionnaire was completed by the target child. These questions included aspects of the parent–child relation- ship including proximity seeking, security in relationship and the emotional quality of the relationship. It has 18 items, 11 of which come from the fifteen item Security Scale (Kerns et al. 1996), and 7 items from the Relatedness Questionnaire (Toth and Cicchetti 1996). The measure used in the present study included 16 items total; 11 items from the original Security Scale and 5 items which were similar to the original Security Scale from the Relatedness Questionnaire. (Example items include: It’s easy to count on my dad for help, I think my mom does not listen to me). This combination of items has been used to measure attachment security in a previous study (Kerns et al. 2011). Items were rated by the child on a 4-point scale from 1 (not true at all) to 4 (very true), with higher scores

indicating higher levels of security (i.e., attachment). The re- liability for this aggregated measure was (α = 0.83) for mothers and (α=0.87) for fathers.

Parental Autonomy Granting A discussion task and problem-solving task at time 2 were used to code parent–child interactions for parental autonomy granting. Parents and chil- dren were asked to discuss three topics they disagree on for the discussion task, and to create a bungee device together so a raw egg would bounce but not crack when dropped for the problem-solving task. The activities were videotaped and then coded by two trained observers, and an additional coder coded 28 % of the tapes for reliability. Several qualitative rating scales were utilized, but for this study, respect for autonomy was used as the target variable. When respect for child auton- omy was high, parents engaged in behaviors such as acknowl- edging the child’s perspective, asking their opinion, and ne- gotiating with the child. Respect for Child Autonomy codes were given for the two tasks, as well as an overall rating; we analyzed the latter. Coding was based on a 1 (very low) to 7 (very high) scale, where higher scores indicated greater auton- omy granting (see Egeland and Hiester 1993 and Pianta 1994 for further information). The intra-class correlation coefficient for observer ratings was 0.68 for fathers and 0.65 for mothers.

Results

Analysis Plan

We first examined relations between the demographic vari- ables and our variables of interest to determine whether there was a need to control for any demographic variables within the primary analyses. Then, to test our hypotheses, we ran zero-order correlations and employed path analysis. We tested

Table 1 Means, standard deviations, minimum and maximum values of the main study variables

Variable N M SD Minimum Maximum

Mother anxiety at time 1 661 1.72 0.51 1.00 3.80

Father anxiety at time 1 661 1.68 0.49 1.00 3.60

Child anxiety at time 1, mother report 661 0.19 0.19 0.00 1.25

Child anxiety at time 1, father report 661 0.21 0.21 0.00 1.50

Partner conflict at time 2, mother report 603 3.61 1.48 1.00 8.20

Partner conflict at time 2, father report 559 3.37 1.37 1.00 9.00

Security with mother at time 2, child report 598 3.44 0.42 1.81 4.00

Security with father at time 2, child report 559 3.37 0.48 1.63 4.00

Autonomy granted by mother at time 2 558 5.15 0.85 3.00 7.00

Autonomy granted by father at time 2 538 5.21 0.93 2.00 7.00

Mother anxiety at time 3 584 1.78 0.57 1.00 4.00

Father anxiety at time 3 520 1.67 0.37 1.00 3.70

Child self-reported anxiety at time 3 576 0.42 0.31 0.00 1.67

Higher values reflect higher levels of the given construct

1258 J Abnorm Child Psychol (2016) 44:1253–1266

a single model that included both mothers and fathers, con- trolling for demographic variables, to determine whether our proposed model fit the data well, and to test the proposed direct and mediational pathways. We also tested several alter- native models, including testing mothers and fathers separate- ly, and testing the combined model without the family process links, to determine whether the family process factors contrib- uted significantly to explaining changes in child anxiety. Fi- nally, we conducted multi-group analyses for boys and girls to determine if child gender moderated any of the model pathways.

Preliminary Analyses

Associations among the demographic variables2 (gender, race, household type, income): No significant correlations were found between gender and any other demographic variables. Race was significantly related, at each time point, to house- hold type (time 1 r=0.10, p<0.01, time 2 r=0.18, p<0.01, time 3 r=0.14, p<0.01) and family income (time 1 r=0.13, p<0.01, time 2 r=0.12, p<0.01, time 3 r=0.10, p<0.05), with Caucasian families more likely to be intact families with higher income. Household type and income at time 1 (r=0.16, p<0.01) and 2 (r=0.15, p<0.01) were also significantly re- lated to one another, with intact households reporting higher income. Household type at times 1 and 2 were strongly related to household type at time 3 (r=0.77, p<0.01 and r=0.81, p<0.01 respectively). Similarly, income at times 1 and 2 was strongly related to income at time 3 (r=0.81, p<0.01 and r=0.72, p<0.01 respectively). At time 3, household type and income were not related to one another. Because house- hold type had a high, significant correlation across all 3 time points, as did family income, we only included household type and income at time 3 in the model. Correlations were then calculated to examine how the main study variables were re- lated to the following demographic variables: gender, race, household type (time 3) and income (time 3). Regarding household type and substantive variables, the following dif- ferences were noted between step-father and intact families: mothers in intact families reported lower maternal anxiety at t1, fathers in intact families reported lower child anxiety at time 1, children in intact families reported higher attachment security to both parents, and observers rated mothers and fa- thers in intact families higher on respect for autonomy. All of the substantive variables were correlated with at least one demographic variable, and we therefore controlled for demo- graphic variables in the model tests.

Correlations between the Main Study Variables

Table 2 presents the zero-order correlations among the main study variables, with father and child reported variables below the diagonal, mother and child reported variables above the diagonal, and mother-father correlations on the diagonal. For the father and child reported variables, correlations were gen- erally modest in magnitude, with significant correlations rang- ing from −0.08 to 0.35. Father and child anxiety were corre- lated at time 1 and time 3. Both child and father anxiety were significantly correlated over time, although father anxiety was more stable than was child anxiety. Child anxiety at time 1 was associated with higher partner conflict and less secure father-child attachment at time 2, although neither father nor child anxiety at time 1 were associated with autonomy granting at time 2. Partner conflict at time 2 was related to father anxiety but not to child anxiety at time 3. Partner con- flict was also associated concurrently with lower father-child attachment security. Lower father-child attachment security was associated with greater child anxiety at time 3. Autonomy granting was not related to father or child anxiety at time 3.

Zero-order correlations for the mother and child reported variables were again modest in magnitude, ranging from −0.08 to 0.45. Mother and child anxiety were concurrently correlated at time 1 and time 3. Both child and mother anxiety were significantly correlated over time, although mother anx- iety was more stable than was child anxiety Mother and child anxiety at time 1 both predicted greater partner conflict at time 2. Mother, but not child, anxiety at time 1 predicted lower mother-child attachment security, and lower autonomy granting at time 2. Partner conflict predicted lower mother- child attachment security concurrently, and lower mother anx- iety at time 3, but was not related to child anxiety at time 3. Lower mother-child attachment security predicted more child anxiety at time 3. Greater autonomy granting by mothers also predicted less anxiety for both mother and child at time 3. Regarding correlations between mother-father variables, as shown on the diagonal in Table 2, all associations were sig- nificant, and relationships were modest in size with the excep- tion of attachment security.

Tests of Model and Direct Paths

To test the proposed model, that family processes account for associations between earlier and later anxiety, we ran a path analysis that included both father and mother vari- ables (see Fig. 2) using Mplus (Muthén and Muthén 1998–2011). Full maximum likelihood estimation was used to handle missing data. The initial model results indicated that it would be beneficial to add a path to our original model, specifically a link between partner conflict at time 2 and parental anxiety at time 3, for both fathers and mothers. Because this link seemed highly plausible, it

2 Note: Race was categorized as Caucasian and Other; Household type was categorized as intact mother-father families and mother-stepfather families.

J Abnorm Child Psychol (2016) 44:1253–1266 1259

was added to our model. Demographic variables were controlled for in all analyses, by regressing them onto the substantive predictor variables at times 2 and 3.

The model was a good fit for the data (see Fig. 2 for path coefficients). Although the Chi-square test of model fit for the model was significant, χ2 (56)=117.88, p<0.001, significant

Table 2 Zero-order correlations among father and mother reported variables

Final model controlling for demographic variables: Significant standardized (unstandardized) paths are shown (p<.05) Time 1 Time 2 Time 3

1260 J Abnorm Child Psychol (2016) 44:1253–1266

Parent anxiety T1

Child anxiety T1

Partner conflict

Attachment security

Autonomy granting

Parent anxiety T3

Child anxiety T3

Time 1

Parent anxiety

0.18** 0.24** 0.21** −0.12** −0.13** 0.45** 0.12**

Child anxiety

0.24** 0.29** 0.12** −0.07 −0.06 0.12** 0.19**

Time 2

Partner conflict

0.23** 0.10* 0.39** −0.08* 0.003 0.20** 0.06

Attachment security

−0.08* −0.12** −0.09* 0.73** 0.12** −0.05 −0.18**

Autonomy granting

0.03 −0.03 −0.04 0.09* 0.11* −0.10* −0.10*

Time 3

Parent anxiety

0.35** 0.16** 0.25** −0.07 −0.07 0.23** 0.16**

Child anxiety

0.18** 0.13** 0.07 −0.23** 0.02 0.16** –

Mother correlations are reported on the top half of the table. Father correlations are reported on the bottom half of the table, mother correlations in the top half, and mother-father correlations on the diagonal which are italicized. *p< 0.05; **p < 0.01

Fig. 2 Final model controlling for demographic variables: Significant standardized (unstandardized) paths are shown (p < 0.05)

Chi square values are not uncommon with large samples. The Comparative Fit Index (CFI) for our data was 0.94, and the RMSEA (Root Mean Square Error of Approximation) value was 0.04 (with a 90 % confidence interval from 0.03 to 0.05). Thus, we cannot reject the close fit hypothesis, which indi- cates our model likely closely fits our data. The SRMR (Stan- dardized Root Mean Square Residual) estimate was 0.04, in- dicating that there was not a large overall difference between the observed and predicted correlations for our data. Overall, these fit indices suggest that our proposed model fits the data well. A baseline model, where family process factors (partner conflict, attachment, and autonomy granting) were modeled in the analysis but were not used to predict any of the outcomes, provided a worse fit, based on a Chi-square difference test, χ2

(94) =306.98, p< 0.00, with model values of CFI = 0.80, RMSEA=0.06, SRMR=0.08, suggesting that inclusion of the family process factors significantly improved the model. We also ran our conceptual model separately for mothers and fathers, and results were similar to the combined model re- sults. We chose to present the combined model for parsimony and to allow for examining unique prediction of mother and father variables.

Several direct paths were of note, including the following significant paths: Father and (father rated) child anxiety, and mother and (mother rated) child anxiety were related to one another at time 1. Father and mother anxiety at time 3 were also related to child reported anxiety at time 3. Paternal anx- iety at time 1 was related to greater partner conflict at time 2, and partner conflict at time 2 predicted greater paternal anxiety at time 3. Similarly, maternal anxiety at time 1 was related to greater partner conflict at time 2, which predicted greater ma- ternal anxiety at time 3. Lower attachment security with fa- thers at time 2 was related to greater child anxiety at time 3, and less autonomy granting by mothers at time 2 was also related to greater child anxiety at time 3. Mother rated child anxiety at time 1 predicted child anxiety at time 3, and mother and father self-rated anxiety at time 1 predicted mother and father anxiety at time 3, respectively.3

To determine whether the parent pathways in our model demonstrated unique effects (i.e., is there a stronger relation- ship for one parent over the other), we conducted a series of analyses where we constrained each pair of mother and father pathways to be equal, and compared the model with constrained pathways to the model without constraints using Chi-square difference tests. There was only one path that dem- onstrated a significant unique effect, which was the link from maternal autonomy granting to child anxiety at time 3,

suggesting that a restriction of autonomy granting by mothers might play a unique role in increasing child anxiety over time.

Mediated Pathways

To test for mediation effects, we estimated multiple mediation pathways within our model separately using Mplus. There were a total of 16 mediational pathways within the model, 8 of which we tested for father variables and 8 of which we tested for mother variables. We found one significant media- tion pathway for the fathers, between paternal anxiety at time 1 and paternal anxiety at time 3 via partner conflict (β=0.04, p<0.001). The same mediational path was also significant for mothers (β=0.16, p<0.05). Thus, partner conflict at time 2 accounted for the association between earlier and later paren- tal anxiety for both mothers and fathers.

Gender Differences

We also tested whether the model pathways were different for boys and girls, using multi-group analysis. We ran a series of tests to determine potential differences in path coefficients between boys and girls, and then evaluated each path using the Chi-square difference test. Results indicated there were no significant path coefficient differences between boys and girls (all Chi-square tests were non-significant), suggesting our findings do not differ for boys and girls.

Discussion

The present longitudinal study sought to understand the rela- tions between parent and child anxiety via family processes. We included multiple family process risk factors—partner conflict, attachment security, and parental autonomy granting—and examined these risk factors over a 9 year peri- od. This study extended earlier work by including fathers, who have often been overlooked in research, especially with regards to their long-term impact on child anxiety. Although child anxiety was moderately stable over a 9 year period, from middle childhood into adolescence, we did identify family process variables that accounted for later anxiety, with the proposed family process model showing a good fit for the data. Further, we identified specific family processes—lower father-child attachment security, and restricted maternal au- tonomy granting—that predicted higher child anxiety in ado- lescence after controlling for early anxiety, and one process (partner conflict) that predicted parent anxiety over 9 years after controlling for initial anxiety. Our results point to the importance of family process variables and to an important long-term impact of fathers in child anxiety, although we did not find evidence that fathers play a unique role in child anx- iety, as proposed by Bögels and Phares (2008).

3 Although we did not include a direct link in the model from partner conflict to child anxiety at time 3 as we were examining mediated path- ways from partner conflict to child anxiety, upon the suggestion of a reviewer we tested this direct link given its significance in prior research, and found that it was not a significant direct path, nor did adding it improve model fit.

J Abnorm Child Psychol (2016) 44:1253–1266 1261

One important finding was the role that partner conflict played in parents’ anxiety. For both fathers and mothers, higher parental anxiety, predicted greater partner conflict 4 years later, which in turn predicted higher anxiety 5 years later, with partner conflict mediating the link between earlier and later anxiety. Thus, it is likely that partner conflict is a factor that both maintains and strengthens anxiety for both mothers and fathers over time. Previous literature has noted concurrent links between parental anxiety, marital quality (McLeod 1994), and marital adjustment (Dehle and Weiss 2002). Other research has looked at anxiety and partner con- flict over time. For example, El-Sheikh et al. (2013b) studied the relationship between inter-partner psychological conflict (IPC), anxiety, and sleep over a week’s time. Anxiety symp- toms mediated the relationship between IPC and poor sleep quality for women, while depressive symptoms mediated this relationship for men. Another study looked at marital discord, neuroticism and stress as potential factors related to internal- izing problems for couples over 6 years (Brock and Lawrence 2014). Although Brock and Lawrence (2014) focused mostly on depression, anxiety symptomswere also included, and they found, for women only, that poor conflict management was related to internalizing symptoms. Dehle and Weiss (2002) examined marital adjustment and state anxiety between married couples over a 3 month period, and found that greater anxiety symptoms reported by husbands, but not wives, was related to poorer marital quality for both wives and husbands. Dehle and Weiss (2002) also explored bidirec- tional links between anxiety and marital quality, but did not find that anxiety predicted marital quality. While these studies establish important connections between anxiety in couples and marital difficulties, our results add to this literature through the inclusion of a 9 year longitudinal design, the dem- onstration of a mediational link between parental anxiety and partner conflict, and the finding that this relationship operates similarly for both mothers and fathers.

Although partner conflict was related to later parental anx- iety, it did not predict later child anxiety via our proposed mediational pathways of parent–child attachment security or autonomy granting, nor was a direct path from partner conflict to child anxiety at time 3 significant. This was surprising, given that research in this area has previously found evidence that exposure to inter-parental conflict puts children at risk for psychopathology, including anxiety (Davies et al. 2002), es- pecially for girls (El-Sheikh et al. 2013a). In a similar vein, Katz and Gottman (1996) suggested that there may be a Bspillover^ effect (i.e., the transfer of mood, emotions, or behaviors from one setting to another) for partner conflict, where specific types of conflict experienced by both fathers and mothers influences their children’s well-being. The cur- rent study is not consistent with prior findings, although our study does differ in an important way from previous work, specifically, our consideration of conflict globally as opposed

to its more specific forms, and the use of parent report (rather than child report) of partner conflict. In addition, it is possible that in selecting families still intact when children were in fifth grade may have resulted in testing these questions in a sample with relatively low levels of marital distress. By contrast, our findings did suggest that the relationships in which the child directly participates (e.g., the parent–child relationship) have a more direct effect on children’s well-being.

We tested two ideas regarding how attachment security might play a role in the development of anxiety. One hypoth- esis was that insecure parent–child attachment might mediate the link between partner conflict and child anxiety. This was based on an extension of the Emotional Security Hypothesis (Davies and Cummings 1994), which proposes that destruc- tive forms of partner conflict lead to children feeling insecure about the marital relationship, which in turn contributes to the development of internalizing and externalizing problems. Substantial research has found that emotional security related to parental conflict uniquely predicts children’s adjustment problems (e.g., Sturge-Apple et al. 2008; Davies et al. 2002). As noted by Davies and Martin (2014) however, emo- tional security and attachment security are distinct constructs; emotional security refers to security in relation to the parents’ relationship, whereas attachment security refers to security within the parent–child relationship. We did not find that low- er attachment security in parent–child relationships explained the link between partner conflict and child anxiety. Our find- ings underscore the importance of Davies and Martin’s (2014) distinction between attachment security and emotional securi- ty based on the parental or marital relationship, as it may be that the latter is more influenced by conflict in the marital relationship. Studies assessing both parent–child attachment security and emotional security in relation to the marital rela- tionship are needed to test this proposal.

A second idea that our study tested regarding attachment security was empirically supported. We found that both mother-child and father-child attachment were correlated with later child anxiety, a finding that is consistent with previous research showing that insecure attachment is related to child anxiety both concurrently and as a predictor of later anxiety (e.g., Brumariu and Kerns 2010; Colonessi et al. 2011). When studied jointly however, only lower attachment security to fathers at time 2, controlling for child anxiety at time 1, pre- dicted child anxiety at time 3. With the addition of fathers, we incorporated a family influence that has previously been little accounted for, and it is possible that fathers’ relationships may account for some of the variance often attributed to mothers. However, these findings should be interpreted with caution, as results did not indicate that this link was unique to fathers, and our findings might be influenced by the high correlations be- tween mother-child and father-child attachment. Our study is one of the first to test whether attachment security in child- hood can explain anxiety across childhood and into

1262 J Abnorm Child Psychol (2016) 44:1253–1266

adolescence; thus, our findings contribute to an understanding of the lasting impact of childhood attachment relationships into adolescence. Finally, our results lend support to the idea that fathers as well as mothers do matter for children’s internalizing symptoms.

Autonomy granting was another potential risk factor for child anxiety that differed for mothers and fathers in this study. Low autonomy granting and high parental control are established correlates of child anxiety based both on observa- tional and questionnaire data (McLeod et al. 2007; van Der Bruggen et al. 2008), although the majority of studies have only included mothers, and the directionality of findings (i.e., are low autonomy granting and high parental control a re- sponse to or the cause of child anxiety) are currently unclear. Studies that have attempted to parse out directionality have provided mixed evidence regarding how child anxiety and parenting are related over time. In a longitudinal study of child anxiety and parental over-involvement (Hale et al. 2013), child anxiety predicted increases in perceived parental over- involvement rather than parenting predicting changes in anx- iety. Similarly to Edwards et al. (2010) however, we found that a restriction of maternal autonomy granting was related to increases in child anxiety, but was not related to prior child anxiety. Additionally, we found a significant unique effect for observed maternal restriction of autonomy granting on child anxiety, suggesting that the control mothers exert over their children is likely to be more influential than that of fathers for child anxiety. Although a restriction of autonomy granting by both parents has been found to contribute to later child anxiety based on both self-report and observational measures (e.g., Verhoeven et al. 2012), and some evidence and theory have suggested that fathers may be more important in this domain (e.g., Teetsel et al. 2014; Bögels and Perotti 2011), we did not find an association between paternal autonomy granting and child anxiety in this study. As with the attachment findings, however, it is important to interpret these results with caution, as we did not assess autonomy granting at younger ages. More research is needed to further establish both the directionality of effects, as well as gaining a better understanding of the potential differences between fathers and mothers, and to eval- uate how fathers’ changing involvement in their children’s lives may influence the role of autonomy granting in child anxiety.

This study possesses several strengths, including the large sample size, multi-method approach, and longitudinal design. Although the magnitude of the effects were small, we were able to detect important relationships between our variables over a 9 year period. We also examined both mother-child and father-child relationships, which is a strength as father data are not as commonly utilized in the literature. Despite these strengths, the use of the NICHD Study of Early Child Care data did introduce limitations. As the data were already col- lected, the authors had no control over the measures used, and

a few of the measures were not optimal to the study’s ques- tions (e.g., State-Trait Anxiety assessed over the past week instead of an explicit trait measure of anxiety; attachment se- curity was only measured via questionnaire, and did not in- clude assessments of insecure attachment patterns). Another limitation was that parental conflict, attachment security, and autonomy granting were assessed at the same time point. Additionally, most participants reported subclinical levels of anxiety, and thus, our results may not generalize to a popula- tion of clinically anxious individuals. It is also important to note that path models do not indicate causality, and can only inform whether the model is a good fit for the data. It is possible that alternative models may better account for the role of parents in child anxiety. For example, factors such as behavioral inhibition, environmental risk factors such as peer relationships and school experiences (e.g., Degnan et al. 2010), or child characteristics such as emotion regulation (e.g., Brumariu and Kerns 2013) could enhance the model tested here. Finally, the results were based on a primarily Caucasian, upper-to-middle class, two parent household sample; thus, they may not be generalizable to other populations.

The present findings indicate a few potential clinical con- siderations. Current literature questions the efficacy of includ- ing parents in their children’s treatment, as results of a meta- analysis suggest that the harm or benefit in including parents are inconclusive (Barmish and Kendall 2005). Further, efforts to date suggest that treating the parent–child relationship may have only limited effects on child anxiety (Silverman et al. 2009). The results of our study do suggest that attachment security and autonomy granting are related to child anxiety, and therefore it would be important for a clinician to assess these areas to more comprehensively understand children’s clinical presentations. Our study also suggests implications for couples counseling. For example, it may be helpful for couples attending therapy who have high levels of conflict to be screened for anxiety, and either or both partners in the dyad may benefit from individual treatment for anxiety disor- ders if needed.

Overall, the present study adds to the field in several critical ways. First, it provides data on fathers, in addition to combin- ing the influences of mothers and fathers in onemodel, both of which are rare in the extant literature. This allows insights into potential unique influences of mothers and fathers on chil- dren’s anxiety, and the findings suggest that etiological models of child anxiety will benefit from considering the po- tential unique roles mothers and fathers may play. . Further, the present study involves several design strengths which are much needed in the extant research, including longitudinal design, multi-method approach, and large sample size. With this large sample, we showed that maternal autonomy granting and paternal attachment each uniquely predict later child anx- iety, even after controlling for several other risk factors for anxiety. By contrast, despite the large sample and ample

J Abnorm Child Psychol (2016) 44:1253–1266 1263

power, we did not replicate prior findings that parent conflict predicts later anxiety through its effects on parent–child rela- tionships, although we did find that parent conflict predicted later parent anxiety. Although negative findings can be diffi- cult to interpret, the present results add to the literature given recent calls for the need to replicate prior findings (Open Source Collaboration 2012). Collectively, the results support the interpretation that the relationships children have with their parents likely contributes to child anxiety more so than factors outside of direct parent–child relationships (e.g., part- ner conflict).

In conclusion, the present study provides evidence that family processes contribute to the maintenance of both parent and child anxiety over time. Our results indicate that fathers andmothers may both have long-term influences on children’s anxiety, although this occurs via different mechanisms (fa- thers, parent–child attachment; mothers, autonomy granting). In addition to findings regarding child anxiety, the mediated relationship between parental anxiety, partner conflict, and later parental anxiety is a novel finding that provides new knowledge regarding the effects of partner conflict for parents. While this study made progress towards the identification of the long-term impact of fathers in child anxiety, important next steps include examining the current study’s risk factors utiliz- ing different measures, including additional risk factors that may further explain child anxiety, and examining interaction effects between mothers and fathers that may contribute to child anxiety.

Compliance with ethical standards

Conflict of Interest The authors declare that they have no conflict of interest.

References

Achenbach, T. M. (1991a). Manual for the child behavior checklist/4-18 and profile. Burlington: University of Vermont, Department of Psychiatry.

Achenbach, T. M. (1991b). Manual for the youth self-report and 1991 profile. Burlington: University of Vermont, Department of Psychiatry.

Achenbach, T. M. (2001). Child behavior checklist for ages 6 to 18. Burlington: University of Vermont, Research Center for Children, Youth, and Families.

Albano, A. M., Chorpita, B. F., & Barlow, D. H. (2003). Childhood anxiety disorders. Child Psychopathology, 2, 279–329.

Barmish, A. J., & Kendall, P. C. (2005). Should parents be co-clients in cognitive-behavioral therapy for anxious youth? Journal of Clinical Child and Adolescent Psychology, 34, 569–581.

Bittner, A., Egger, E., Erkanli, A., Costello, E. J., Foley, D., & Angold, A. (2007). What do childhood anxiety disorders predict? Journal of Child Psychology and Psychiatry, 48, 1174–1183.

Bögels, S. M., & Brechman-Toussaint, M. L. (2006). Family issues in child anxiety: attachment, family functioning, parental rearing and beliefs. Clinical Psychology Review, 26, 834–856.

Bögels, S. M., & Perotti, E. C. (2011). Does father know best? A formal model of the paternal influence on childhood social anxiety. Journal of Child and Family Studies, 20, 171–181.

Bögels, S., & Phares, V. (2008). Fathers’ role in the etiology, prevention and treatment of child anxiety: a review and new model. Clinical Psychology Review, 28, 539–558.

Bögels, S. M., & van Melick, M. (2004). The relationship between child- report, parent self-report, and partner report of perceived parental rearing behaviors and anxiety in children and parents. Personality and Individual Differences, 37, 1583–1596.

Bögels, S. M., Bamelis, L., & van der Bruggen, C. (2008). Parental rearing as a function of parent’s own, partner’s, and child’s anxiety status: fathers make the difference. Cognition & Emotion, 22, 522– 538.

Bosquet, M., & Egeland, B. (2006). The development and maintenance of anxiety symptoms from infancy through adolescence in a longi- tudinal sample. Development and Psychopathology, 18, 517–550.

Bowlby, J. (1969). Attachment and loss: Vol. 1. Attachment. NY: Basic Books.

Bowlby, J. (1973). Attachment and loss: Vol. 2. Separation anxiety and danger. New York: Basic Books.

Braiker, H., & Kelley, H. (1979). Conflict in the development of close relationships. In R. Burgess & T. Huston (Eds.), Social exchange and developing relationships. NY: Academic.

Bretherton, I. (2010). Fathers in attachment theory and research: a review. Early Child Development and Care, 180, 9–23.

Brock, R. L., & Lawrence, E. (2014). Martial Processes, neuroticism, and stress as risk factors for internalizing symptoms. Couple and Family Psychology: Research and Practice, 3, 30–47.

Brumariu, L. E., & Kerns, K. A. (2010). Parent–child attachment and internalizing symptoms in childhood and adolescence: a review of empirical findings and future directions. Development and Psychopathology, 22, 177–203.

Brumariu, L. E., & Kerns, K. A. (2013). Pathways to anxiety: contribu- tions of attachment history, temperament, peer competence, and ability to manage intense emotions. Child Psychiatry and Human Development, 44, 504–515.

Collins, L., Schafer, J., & Kam, C.-M. (2001). A comparison of inclusive and restrictive strategies in modern missing data procedures. Psychological Methods, 6, 330–351.

Colonessi, C., Draijer, E. M., Stams, G. J. J., Van der Bruggen, C. O., Bögels, S.M., & Noom,M. J. (2011). The relation between insecure attachment and child anxiety: a meta-analytic review. Journal of Clinical Child and Adolescent Psychology, 40, 630–645.

Costa, N. M., & Weems, C. F. (2005). Maternal and child anxiety: do attachment beliefs or children's perceptions of maternal control me- diate their association? Social Development, 14, 574–590.

Cummings, E. M. (1998). Children exposed to marital conflict and vio- lence: Conceptual and theoretical directions. In G. Holden, B. Geffner, & E. Jouriles (Eds.), Children exposed to marital violence: Theory, research and applied issues (pp. 55–94).Washington, D. C.: American Psychological Association.

Cummings, E. M., & Davies, P. T. (2002). Effects of marital conflict on children: recent advances and emerging themes in process-oriented research. Journal of Child Psychology and Psychiatry, 43, 31–63.

Dallaire, D. H., & Weinraub, M. (2007). Infant–mother attachment secu- rity and children's anxiety and aggression at first grade. Journal of Applied Developmental Psychology, 28, 477–492.

Davies, P. T., & Cummings, E. M. (1994). Marital conflict and child adjustment: an emotional security hypothesis. Psychological Bulletin, 116, 387–411.

1264 J Abnorm Child Psychol (2016) 44:1253–1266

Davies, P. T., & Martin, M. (2014). Children’s coping and adjustment in high-conflict homes: the reformulation of emotional security theory. Child Development Perspectives, 8, 242–249.

Davies, P. T., Harold, G. T., Goeke-Morey, M. C., Cummings, E. M., Shelton, K., Rasi, J. A., & Jenkins, J. M. (2002). Child emotional security and interparental conflict. Monographs of the Society for Research in Child Development, 51, i-127.

Degnan, K. A., Almas, A. N., & Fox, N. A. (2010). Temperament and the environment in the etiology of childhood anxiety. Journal of Child Psychology and Psychiatry, 51, 497–517.

Dehle, C., & Weiss, R. L. (2002). Associations between anxiety and marital adjustment. The Journal of Psychology, 136, 328–338.

DeKlyen,M., & Greenberg, M. (2008). Attachment and psychopathology in childhood. Handbook of attachment: Theory, research, and clin- ical applications (pp. 637–665). New York: Guilford Press.

Edwards, S. L., Rapee, R. M., & Kennedy, S. (2010). Prediction of anx- iety symptoms in preschool‐aged children: examination of maternal and paternal perspectives. Journal of Child Psychology and Psychiatry, 51, 313–321.

Egeland, B. & Hiester, M. (1993). Teaching task rating scales. Institute of Child Development, University of Minnesota.

El-Sheikh, M., Keiley, M., Erath, S., & Dyer, W. J. (2013a). Marital conflict and growth in children's internalizing symptoms: the role of autonomic nervous system activity. Developmental Psychology, 49, 92–108.

El-Sheikh, M., Kelly, R., & Rauer, A. (2013b). Quick to berate, slow to sleep: interpartner psychological conflict, mental health, and sleep. Health Psychology, 32, 1057–1066.

Feng, X., Shaw, D. S., & Silk, J. S. (2008). Developmental trajectories of anxiety symptoms among boys across early and middle childhood. Journal of Abnormal Psychology, 117, 32–47.

Graham, J. W. (2009). Missing data analysis: making it work in the real world. Annual Review of Psychology, 60, 549–576.

Greco, L. A., & Morris, T. L. (2002). Paternal child-rearing style and child social anxiety: investigation of child perceptions and actual father behavior. Journal of Psychopathology and Behavioral Assessment, 24, 259–267.

Groh, A. M., Roisman, G. I., van IJzendoorn, M. H., Bakermans‐ Kranenburg, M. J., & Fearon, R. (2012). The significance of inse- cure and disorganized attachment for children’s internalizing symp- toms: a meta‐analytic study. Child Development, 83, 591–610.

Grych, J. H., Seid, M., & Fincham, F. D. (1992). Assessing marital con- flict from the child's perspective: the children's perception of interparental conflict scale. Child Development, 63, 558–572.

Hale III, W. W., Klimstra, T., Branje, S., Wijsbroek, S., & Meeus, W. (2013). Is Adolescent Generalized Anxiety Disorder a Magnet for Negative Parental Interpersonal Behaviors?Depression and Anxiety, 0, 1–8.

Harold, G. T., Shelton, K. H., Goeke‐Morey, M. C., & Cummings, E. M. (2004). Marital conflict, child emotional security about family rela- tionships and child adjustment. Social Development, 13, 350–376.

Katz, L. F., & Gottman, J. M. (1996). Spillover effects of marital conflict: in search of parenting and coparenting mechanisms. New Directions for Child and Adolescent Development, 1996, 57–76.

Kerig, P. K. (1996). Assessing the links between interparental conflict and child adjustment: the conflicts and problem-solving scales. Journal of Family Psychology, 10, 454–473.

Kerns, K. A. (2008). Attachment in middle childhood. In J. Cassidy & P. Shaver (Eds.),Handbook of attachment: Theory, research, and clin- ical applications (2nd ed., pp. 366–382). New York, NY: Guilford Press.

Kerns, K. A., & Brumariu, L. E. (2014). Is insecure parent–child attach- ment a risk factor for the development of anxiety in childhood or adolescence? Child Development Perspectives, 8, 12–17.

Kerns, K. A., Klepac, L., & Cole, A. (1996). Peer relationships and preadolescents’ perceptions of security in the child-mother relation- ship. Developmental Psychology, 32, 457–466.

Kerns, K. A., Siener, S., & Brumariu, L. E. (2011). Mother-child rela- tionships, family context, and child characteristics as predictors of anxiety symptoms in middle childhood. Development and Psychopathology, 23, 593–604.

Madigan, S., Atkinson, L., Laurin, K., & Benoit, D. (2013). Attachment and internalizing behavior in early childhood: a meta-analysis. Developmental Psychology, 49, 672–689.

McLeod, J. D. (1994). Anxiety disorders and marital quality. Journal of Abnormal Psychology, 103, 767–776.

McLeod, B. D., Wood, J. J., & Weisz, J. R. (2007). Examining the asso- ciation between parenting and childhood anxiety: a meta-analysis. Clinical Psychology Review, 27, 155–172.

Micco, J., Henin, A., Mick, E., Kim, S., Hopkins, C., Biederman, J., & Hirshfeld-Becker, D. (2009). Anxiety and depressive disorders in offspring at high risk for anxiety: a meta-analysis. Journal of Anxiety Disorders, 23, 1158–1164.

Muthén, L. K., &Muthén, B. O. (1998–2011). Mplus. Statistical analyses with latent variables. User’s guide, 3.

Open Science Collaboration. (2012). An open, large-scale, collaborative effort to estimate the reproducibility of psychological science. Perspectives on Psychological Science, 7, 657–660.

Phares, V., & Compas, B. E. (1992). The role of fathers in child and adolescent psychopathology: make room for daddy. Psychological Bulletin, 111, 387–412.

Pianta, R. C. (1994). Rating scales for parent–child interaction in pre- schoolers. University of Virginia.

Silverman, W. K., Kurtines, W. M., Jaccard, J., & Pina, A. A. (2009). Directionality of change in youth anxiety treatment involving par- ents: an initial examination. Journal of Consulting and Clinical Psychology, 77, 474–485.

Spielberger, C. D., Gorsuch, R. L., Lushene, R., Vagg, P. R., & Jacobs, G. A. (1983). Manual for The State-Trait Anxiety Inventory (Form Y). Palo Alto: Consulting Psychologists Press.

Sturge-Apple, M. L., Davies, P. T., Winter, M. A., Cummings, E. M., & Schermerhorn, A. (2008). Interparental conflict and children's school adjustment: the explanatory role of children's internal repre- sentations of interparental and parent–child relationships. Developmental Psychology, 44, 1678–11690.

Teetsel, R., Ginsburg, G., & Drake, K. (2014). Anxiety promoting par- enting behaviors: a comparison of anxious mothers and fathers. Child Psychiatry and Human Development, 45, 133–142.

Toth, S. L., & Cicchetti, D. (1996). The impact of relatedness with mother on school functioning in maltreated children. Journal of School Psychology, 34, 247–266.

vanDer Bruggen, C. O., Stams, G. J. J., & Bögels, S.M. (2008). Research review: the relation between child and parent anxiety and parental control: a meta‐analytic review. Journal of Child Psychology and Psychiatry, 49, 1257–1269.

van Eijck, F. E., Branje, S. J., Hale, W. W., III, & Meeus, W. H. (2012). Longitudinal associations between perceived parent-adolescent at- tachment relationship quality and generalized anxiety disorder symptoms in adolescence. Journal of Abnormal Child Psychology, 40, 871–883.

Vasey, M. W., & Dadds, M. R. (2001). The developmental psychopathol- ogy of anxiety. New York: Oxford University Press.

Verhoeven, M., Bögels, S. M., & van der Bruggen, C. C. (2012). Unique roles of mothering and fathering in child anxiety; moderation by child’s age and gender. Journal of Child and Family Studies, 21, 331–343.

Verschueren, K., & Marcoen, A. (1999). Representation of self and socioemotional competence in kindergartners: Differential and com- bined effects of attachment to mother and to father. Child Development, 70, 183–201.

J Abnorm Child Psychol (2016) 44:1253–1266 1265

Warren, S. L., & Simmens, S. J. (2005). Predicting toddler anxiety/ depressive symptoms: effects of caregiver sensitivity on tempera- mentally vulnerable children. Infant Mental Health Journal, 26, 40–55.

Wood, J. J., McLeod, B. D., Sigman, M., Hwang, W. C., & Chu, B. C. (2003). Parenting and childhood anxiety: theory, empirical findings, and future directions. Journal of Child Psychology and Psychiatry, 44, 134–151.

1266 J Abnorm Child Psychol (2016) 44:1253–1266

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.