Child Psychiatry & Human Development

, Volume 45, Issue 1, pp 119–131

Child–Mother and Child–Father Attachment Security: Links to Internalizing Adjustment Among Children with Learning Disabilities

Authors

    • Special Education Program, Constantiner School of EducationTel-Aviv University
Original Article

DOI: 10.1007/s10578-013-0383-9

Cite this article as:
Al-Yagon, M. Child Psychiatry Hum Dev (2014) 45: 119. doi:10.1007/s10578-013-0383-9

Abstract

The study examined the unique role of children’s attachment with the father and with the mother, in explaining differences in internalizing features (i.e., loneliness, sense of coherence, hope and effort, and internalizing behavior syndrome) among 107 children with learning disabilities (LD) versus 98 children with typical development ages 8–12. Preliminary analyses yielded significant group differences on most measures. SEM analysis indicated high fit between the theoretical model and empirical findings, and different patterns of relations among the model’s components for the two populations. As hypothesized, child–father and child–mother attachment contributed differently to children’s internalizing features for the two subgroups. Discussion focused on understanding unique and complementary roles of attachment relations with fathers versus mothers among children with and without LD.

Keywords

FathersMothersAttachmentLearning disabilitiesLonelinessInternalizing behavior problemsSense of coherence

Introduction

Decades of research have highlighted the role that quality of parent–child relationships may play in a variety of developmental outcomes (e.g., [1, 2]). Many of these researchers have considered Bowlby’s [3, 4] attachment theory as a highly relevant framework for exploring variations in children’s adjustment (e.g., see review in [2]). Whereas most of the research literature on parent–child attachment relationships has primarily investigated mothers (e.g., [5]), a recent upsurge of interest in child–father attachment has underscored its important role for children’s development and later adjustment [69]. Findings from these studies call for further exploration regarding how children form, represent, and utilize attachment with fathers in unique ways that may complement their attachment relationships with mothers. For example, research has suggested that fathering may be more sensitive than mothering to contextual factors such as children’s characteristics (e.g., [10, 11]).

In contrast to the growing literature on attachment relationships’ contribution to typically developing children’s adjustment, relatively few studies have explored this issue for children with learning disabilities (LD). The few attachment studies existing on children with LD, which mainly focused on mothers, revealed these children’s reports of lower attachment security than their peers without LD [1216], and highlighted the role of these children’s insecure attachment relationships as a risk factor for increasing maladjustment [1518]. The current study focused on the role of children’s attachment to each parent in explaining four internalizing adjustment measures that have shown a high prevalence of difficulties among children with LD: loneliness, sense of coherence, hope and effort, and internalizing behavior syndrome. Emergent evidence has suggested that children with LD demonstrate not only academic dysfunction but also co-occurring diverse socioemotional difficulties (e.g., [19, 20]) that may correspond with these children’s internal neurological deficits. Indeed, prior studies have argued that these children’s information-processing disorders, impulsivity, and performance and production deficits may also affect their social and emotional perceptions and interpretations (see [21] for a review), which, in turn, may impair their interpersonal and intrapersonal abilities and skills. As a group, children with LD manifest lower self-concept as well as higher levels of peer rejection, loneliness, depression, and anxiety compared to their nondisabled counterparts (e.g., [19, 22]). The unique role of children’s attachment relationships to each parent may be of particular interest in LD populations, because these variables may act as risk or protective factors in understanding this vulnerable population who manifests a high incidence of maladjustment.

Children’s Attachment Relationships with Father and Mother

Briefly, attachment theory [3, 4] proposed that in their first year infants develop a specific and enduring relationship with their primary caretakers [23]. Attachment theory assumed that infants internalize their interactions with significant others into “internal working models of attachment”—mental representations of significant others and the self. These internal working models of attachment are manifested later as unique “attachment styles” during social interactions through adulthood.

Numerous findings have pinpointed the associations between children’s attachment relationships and their socioemotional functioning (e.g., [1]; see [2] for a review). Data from these studies indicated associations between secure attachment with caregivers and children’s higher levels of psychological well-being and sociability compared to children with insecure (avoidant or anxious) attachment relationships (e.g., [2]). In contrast, children’s insecure attachment was linked with poorer peer relations during childhood and beyond [1, 24]. Although most attachment studies are correlational designs, findings from attachment-based interventions (e.g., [25]) can be cautiously interpreted as implying causality between children’s attachment and socioemotional functioning.

Although attachment theory did reveal an interest in fathers’ contribution, the historical view of mothers as primary caregivers may account for the literature’s almost sole focus on children’s attachment with mothers [6, 10, 26]. Even in today’s society, mothers continue to act as primary caregivers during early childhood, but the father’s influence on children’s behaviors and development becomes equivalent to the mother’s nearer middle childhood [27]. Research on the role of father–child attachment relationships for children’s adjustment in social and emotional domains has yielded inconsistent findings. Some studies linked secure attachment with fathers to children’s positive interactions with friends, but other studies failed to support such a link (see [28] for a review). Grossman et al. [29] argued that father and mother attachment relations derive from different sets of early social experiences, where mothers act as a secure base in times of distress and fathers act as a challenging but reassuring play partner. Thus, rather than regarding fathers as less influential or as subsidiary figures in an attachment hierarchy, the current study assumed that fathers fulfill different but equally influential roles.

Several studies have supported these assumptions and also explored the possible unique role of children’s attachment with fathers versus mothers (e.g., [9, 10, 3033]). For instance, Verschueren and Marcoen [33] reported that secure attachment with mothers may best predict children’s functioning in more intimate small groups or dyadic interactions, whereas secure attachment to fathers may best predict peer acceptance. Furthermore, George et al. [30] indicated that parenting mechanisms associated with insecure attachment differ. Thus, fathers’ low level of responsiveness was linked with children’s insecure-avoidant attachment, whereas mothers’ low level of responsiveness was linked with insecure-ambivalent attachment. Richaud de Minzi [9] also revealed that security with mothers and security with fathers were unique in predicting children’s socioemotional outcomes and in their hierarchy. Specifically, availability of and reliance on the father outweighed attachment to mother in predicting their children’s lower depression scores, whereas availability of and reliance on the mother outweighed father attachment in explaining differences in children’s loneliness. Following these unique findings for mothers and fathers, the current study adopted the independence conceptual model (e.g., [6]), whereby child–father and child–mother attachment comprise an equal focus in exploring differences in socioemotional developmental outcomes among children with LD. The independence model proposes that all attachment relationships are equally important, but each contributes to development in distinct domains of internalizing features among children with LD.

Children’s Internalizing Socioemotional Adjustment

This section focuses on the four internalizing measures that were conceptualized as dependent variables in the current study (presented in the right column of Figs. 1, 2, 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs10578-013-0383-9/MediaObjects/10578_2013_383_Fig1_HTML.gif
Fig. 1

Modified model for both groups. Exogenous factors (i.e., independent variables) are presented within ovals in the left column. Endogenous/latent factors (i.e., dependent variables) are presented within ovals in the right column. NFI = .997, CFI = 1.00, GFI = .998, and RMSEA = .00

https://static-content.springer.com/image/art%3A10.1007%2Fs10578-013-0383-9/MediaObjects/10578_2013_383_Fig2_HTML.gif
Fig. 2

Modified model for children with LD. Exogenous factors (i.e., independent variables) are presented within ovals in the left column. Endogenous/latent factors (i.e., dependent variables) are presented within ovals in the right column

https://static-content.springer.com/image/art%3A10.1007%2Fs10578-013-0383-9/MediaObjects/10578_2013_383_Fig3_HTML.gif
Fig. 3

Modified model for children with typical development. Exogenous factors (i.e., independent variables) are presented within ovals in the left column. Endogenous/latent factors (i.e., dependent variables) are presented within ovals in the right column

Loneliness

Loneliness reflects internal, subjective, and emotional aspects of children’s social lives, providing a global indicator of dissatisfaction from the quality and/or quantity of social interrelations [34, 35]. As suggested by Asher and Paquette [34], loneliness is usually defined by researchers as involving cognitive awareness of a deficiency in one’s social and personal relationships as well as affective reactions such as sadness, emptiness, or longing Findings from large numbers of studies have indicated that feelings of loneliness and social dissatisfaction in childhood are stable over time and are connected with later maladjustment like depression, poor self-concept, and psychosomatic problems (e.g., [36, 37]). Studies on children with LD often yielded higher loneliness levels compared to children without LD [12, 15, 38].

Sense of Coherence

Children’s sense of coherence (SOC)—provided an index of children’s coping resources and personal strength [39, 40]. Antonovsky [41] defined the construct of SOC as a global orientation or enduring tendency to see the world as more or as less comprehensible, manageable, and meaningful. Findings from previous studies have demonstrated that, as a group, children with LD manifest lower SOC compared to children without LD (e.g., [15, 38]).

Hope and Effort

In general, Snyder [42, 43] suggested that the construct of hope is a cognitive set comprises “pathway thinking”, the perceived capacity to generate strategies for attaining goals, and “agency thinking”, those perceptions involving one’s capacity to initiate and sustain movement along the chosen pathways. Effort refers to the individual’s level of investment, intensity, and persistence in task accomplishment [22, 44]. Prior studies indicated that children and adolescents with LD reported lower appraisals of hope and effort compared to their nondisabled peers [12, 22, 44].

Internalizing Behavior Problems

A large body of research (e.g., [45, 46]) indicates that maladaptive functioning in childhood falls into two categories of disorders: internalizing and externalizing. The current study focused on internalizing maladjustment syndrome, which includes problems such as anxiety, depression, and social withdrawal. Prior studies on children with LD reported higher levels of internalizing behavior difficulties than in typically developing children (e.g., [19, 22]).

The Current Study

The literature reviewed here has highlighted some important questions calling for additional exploration regarding the possible unique roles played by children’s attachments with fathers and mothers. As described above, studies on attachment have rarely extended their investigation to children with LD or to fathers, despite growing awareness about the importance of close relationships with fathers for children’s adjustment and about the heightened vulnerability of the LD population for insecure attachment as well as for a high incidence of maladjustment (e.g., [14, 18, 20]). The current study aimed to further explore the relevance and validity of the conceptual “independence model,” which assumed that all attachment relationships are equally important, but each contributes to development in distinct domains. Thus, the present study examined the contribution of children’s attachment with fathers and mothers to the explanation of differences in internalizing socioemotional adjustment—namely, loneliness, SOC, hope and effort, and internalizing behavior syndrome—among children with LD versus children with typical development.

Corresponding with these objectives, this study tested the following two empirical hypotheses among school-age Israeli children with LD who attended regular classes and a comparison group of typically developing children. First, in accordance with previous studies on typically developing children (e.g., [47]), in both groups, children’s higher attachment security scores were expected to contribute to lower levels of internalizing difficulties, but differences were expected in line with the unique roles of children’s attachment toward fathers versus mothers (e.g., [10]). Second, similar to prior studies that highlighted the vulnerability of children with LD to risk factors at the individual and family levels (e.g., [12]), children with LD were expected to reveal more vulnerability to insecurity in close relationships; therefore, they were expected to demonstrate a greater number of significant paths between attachment and the internalizing measures, compared to children without LD.

Methods

Participants

The sample consisted of 205 children in Grades 3—(ages 8–12 years; M = 9.98, SD = 1.08) from seven public elementary schools in urban areas of central Israel, comprising 107 children with LD (52 girls, 55 boys) and 98 children with typical development (56 girls, 42 boys). Analysis of variance (ANOVA) showed no significant differences between the two groups regarding children’s age or sex.

LD Group (n = 107)

These children had been diagnosed with LD through previous psycho-educational evaluations. Corresponding to the Israeli Ministry of Education’s educational policy, in line with the DSM-IV-TR [48] diagnostic features, all children receiving the LD diagnosis revealed an average IQ level as well as significantly lower achievements on standardized tests (in reading, writing, and/or mathematics) than those expected for age, schooling, and intelligence level. Parents reported on: (a) neuropsychological, psychodidactic, and other diagnostic evaluations undergone by the child; (b) testing accommodations obtained from school psycho-educational staff; and (c) specific treatments obtained from school and extra-school resources. According to parental reports on extra-school interventions, 57 children (53.3 %) received educational treatment outside school, 14 children (13.1 %) received psychological treatment, 36 children (33.6 %) received medical (pharmacological) treatment, and 4 children (3.7 %) received linguistic-communication intervention.

Regarding demographic characteristics of families of children with LD, 101 of these children had married parents, and 6 had divorced parents. Mothers’ ages were 31–52 years (M = 38.26, SD = 4.40), with 11–22 years of education (M = 14.71, SD = 2.34). Regarding maternal work status, 72 worked full time, 23 worked part time, and 12 were not employed. Fathers’ ages were 34–72 years (M = 41.09, SD = 5.84), with 9–25 years of education (M = 14.57, SD = 2.84). Regarding fathers’ work status, 96 worked full time, 4 worked part time, and 7 were not employed.

Comparison Group (n = 98)

According to parent reports, these children had typical development, without difficulties in academic functioning or in social, behavioral, or emotional functioning. Specifically, these children in the comparison group did not demonstrate reading, writing, or mathematics problems, and none had received any diagnostic evaluation or special assistance from school staff or other professionals (including medical treatments).

Regarding demographic characteristics of families of children with typical development, 94 of these children had married parents and 4 had divorced parents. Mothers’ ages were 31–54 years (M = 39.37, SD = 4.88), with 11–22 years of education (M = 15.02, SD = 2.42). Regarding maternal work status, 71 worked full time, 19 worked part-time, and 8 were not employed. Fathers’ ages were 32–69 years (M = 42.44, SD = 5.82), with 9–25 years of education (M = 15.37, SD = 2.84). Regarding fathers’ work status, 89 worked full time, 4 worked part time, and 5 were not employed.

A set of t test analyses as well as Chi square tests revealed no significant differences between the LD and typically developing groups regarding children’s age or sex, or regarding parents’ education, marital status, or work status.

Instruments

Children’s internalizing variables were assessed via two information sources: four self-report measures and one measure completed by parents.

Children’s Self-Report Instruments

Attachment Security Style [49]

This 15-item Hebrew adaptation [50] assessed children’s perceptions of security in parent–child relationships using Harter’s [51] 4-point “Some kids … other kids” format (α = .71). Scores range from 15 to 60, with a categorical cut-off point of 45 distinguishing secure from insecure child–parent attachment [49]. The scale was administered twice, once about mothers and once about fathers. In the current study, Cronbach alphas were .77 for the child–mother attachment scale and .79 for the child–father attachment scale.

Loneliness and Social Dissatisfaction Questionnaire [52]

This questionnaire (Hebrew adaptation: [53]) included 16 primary items tapping children’s feelings of loneliness (e.g., “I have nobody to talk to in my class,” “I am lonely”) rated on a 5-point scale from Never (1) to Always (5). Per Asher et al. [52], computation of a single score tapped global sense of loneliness. In the current study, Cronbach alpha was .92.

Children’s Sense of Coherence Scale (SOC; [54])

This scale included 16 items tapping three dimensions of children’s SOC in the world—comprehensibility, manageability, and meaningfulness (e.g., “I feel that I don’t understand what to do in class;” “I have trouble with most of the things I try to do”)—rated on a 4-point scale from Never (1) to Always (4). Per Antonovsky [41], computation of a single total score tapped global SOC. In the current study, Cronbach alpha was .84.

Children’s Hope and Effort Scales (Ages 8–16; Hope: [55]; Effort: [22])

This scale assessing hope and effort (Hebrew adaptation: [22]) included 3 agency items referring to goal-directed energy (e.g., “I meet the goals that I set for myself”), 3 pathway items referring to planning to meet goals (e.g., “I can think of many ways to get the things in life that are important to me”), and 4 items tapping children’s investment and effort levels (e.g., “I don’t give up even when it is difficult for me”), rated on a 6-point scale from None of the time (1) to All of the time (6). In the current study, Cronbach alpha for the overall scale was .90.

Instrument Completed by Parents

Internalizing Syndrome Scales from the Standardized Child Behavior Checklist (CBCL; [45])

This standardized instrument for rating children’s behavior (Hebrew adaptation: [56]) included 112 behavioral items scored on a 3-point scale from Nottrue (0) to Very/Oftentrue (2). For each child, parents completed the CBCL separately. Achenbach’s [45] principal components analyses yielded eight narrow-band syndrome scales and two broad-band syndrome scales: internalizing and externalizing. The current study used only the internalizing broad-band syndrome scale. Thus, withdrawal, somatic complaints, and anxiety/depression formed the broad-band internalizing syndrome (fathers’ α = .84; mothers’ α = .86). Children’s internalizing scores were computed by averaging fathers’ and mothers’ evaluations (α = .90).

Procedure

After receiving approval from the Israel Ministry of Education and the seven school principals, research team members entered each classroom to briefly explain the study and distribute letters for children to take home to their parents. Parents who consented for their families to participate returned the signed consent form to the teacher. Research team members then contacted parents to furnish additional information on study objectives and procedures, and to coordinate data collection. After obtaining parental consent and collecting data from all of the children with LD, the research team then collected data only from those families of children with typical development who matched the children with LD on age and sex.

Trained graduate students of educational counseling entered schools at the recruitment stage and administered the test battery individually to parents and children during home visits. Children completed the set of four self-reports alone in a quiet room. The examiner read sample items aloud for each questionnaire to ensure children’s understanding and supplied further help if necessary to children with LD. Next, the examiner explained the instrument to the parents, who each completed it separately in a quiet room.

Data Analysis

The first set of analyses, focusing on descriptive statistics, explored group differences: children with LD versus without LD. Second, analyses were conducted via the structural equation modeling (SEM) method (Amos program). These analyses comprised the estimation of the model examined in the current study, simultaneously for the two groups: children with and without LD. This multigroup SEM approach (MSEM) measures the contribution of the interaction between the group and the model’s components. Variable fitness to the base model simultaneously considers the relations between the correlation and the existence of two groups [57].

Results

Descriptive Statistics for Children with LD Versus with Typical Development

To examine group differences (children with LD vs. children with typical development) in children’s internalizing and attachment measures and to decrease the chance of Type 1 errors, a multivariate analysis of variance (MANOVA) was conducted, with the following six variables: children’s measures (self-rated loneliness, SOC, hope and effort, and attachment toward father; attachment toward mother; and parent-rated internalizing child behavior). The MANOVA yielded a significant main effect for study group, F(6, 198) = 11.17, p < .001, η2 = .25. Table 1 presents the means, standard deviations, and F scores for the univariate ANOVAs of all the study variables for the two groups.
Table 1

Means, SD, and F scores of child variables according to study group

Child measures

Children with learning disabilities (n = 107)

Children with typical development (n = 98)

F (1, 204)

η2

M

SD

M

SD

Attachment with father

45.20

7.81

48.83

6.04

13.70**

.07

Attachment with mother

48.63

7.08

50.01

6.21

2.17

.01

Loneliness

34.00

14.28

26.53

9.97

18.52**

.09

Sense of coherence

47.20

6.33

52.10

5.68

33.70**

.14

Hope and effort

40.50

9.53

48.00

6.57

42.01**

.17

Internalizing problems

5.95

5.21

3.36

6.21

16.13**

.08

** p < .001

Significant intergroup differences emerged on most of the children’s internalizing measures. As a group, the children with LD reported significantly higher loneliness, lower SOC, lower hope and effort, and higher level of internalizing problems compared to their typically developing peers. In addition, children with LD reported lower attachment security scores toward the father than their typically developing peers. However, no significant differences emerged regarding children’s attachment scores toward the mother.

Further Chi square analysis examined the link between study group and attachment classification toward father and mother. Children were assigned either a secure or an insecure child–mother attachment classification as well as either a secure or an insecure child–father attachment classification, based on the categorical cutoff point (score of 45) on the children’s self-reported Attachment Security Style [49].

A significant association emerged between study group (with/without LD) and attachment classification toward father, χ2(1) = 7.07, p = .01. Whereas 75.5 % (n = 74) of the typically developing children in Grades 3–6 received a secure attachment style classification toward the father, only 58 % (n = 62) of the children with LD were classified as securely attached. In contrast, Chi square analysis showed no significant association between study group and attachment classification toward the mother, χ2(1) = 2.10, p = .09. Thus, 81 % (n = 79) of the typically developing children and 72 % (n = 77) of the children with LD received a secure attachment classification toward the mother.

Estimations of the Model

This section describes the estimations of the model tested in the present study via the SEM (Amos program) and the MSEM approaches, which measured the contribution of the interaction between group and the model’s components. Thus, examination of the data’s fitness to the base model takes into account the relations between the correlation and the existence of two groups [57].

Testing of the current model was conducted in three steps of estimation: (1) the base model, (2) modified models consisting of the significant pathways that emerged in the first step, and (3) group comparisons.

Step 1: Estimation of the Base Model

The first step of analysis attempted to estimate the base model simultaneously for the two groups (children with and without LD) by examining variable fitness to the base model. Inasmuch as this approach simultaneously considers the relations between the correlation and the existence of two groups, fitness is therefore reported once.

This base model assumed that the exogenous variables of child attachment relations toward the mother and the father (i.e., the independent variables; see left columns of Figs. 1, 2, 3) would contribute to the endogenous internalizing socioemotional variables (i.e., the dependent variables presented in the figures’ right columns). Thus, children’s attachment toward the father/mother was assumed to contribute to each of the children’s internalizing measures. Second, the model assumed a correlation between the two exogenous variables themselves (i.e., among children’s attachment toward the father and mother). Third, the model assumed correlations among the four endogenous internalizing variables (i.e., loneliness, SOC, hope and effort, and internalizing behavior syndrome).

This examination of the base model demonstrated a good fit, shown by the nonsignificant Chi square test, χ2(2, 205) = 1.22, p > .05, and by the high indices-of-fit values: NFI = .997, CFI = 1.00, GFI = .998, and RMSEA = .00. Thus, the first step of analysis revealed a high fit between the theoretical model and the empirical data.

Step 2: Estimation of the Modified Model

To design more parsimonious models for the empirical data, only paths (relations) with significance higher than t > |2.00| were considered. Thus, in estimating the modified model, the following two non-significant paths were omitted: the path between children’s attachment toward mother to hope and effort and the path between children’s attachment toward father to the internalizing syndrome. Therefore, the modified model comprised only the significant remaining paths. The current Chi square result was nonsignificant, χ2(4, 205) = 1.23, p > .05, and the indices-of-fit values demonstrated a high fit: NFI = .997, CFI = 1.00, GFI = .998, and RMSEA = .00. Thus, this step of estimation indicated a high fit between the modified model in both groups and the empirical findings (see Fig. 1).

Step 3: Group Comparison

In this step, a group comparison was conducted in order to estimate the contribution of the interaction between the two groups and the model’s components. This analysis showed significant group differences between the modified model for the children with LD versus the modified model for the children with typical development: CMIN (i.e., minimum value of the Chi square discrepancy) = 20.644, p = .023. The outcome of the Chi square test was nonsignificant, χ2(10, 205) = 5.92, p > .05, and the indices-of-fit values demonstrated a high fit: NFI = .986, CFI = 1.00, GFI = .991, and RMSEA = .00. Figures 2 and 3 present the modified models for the children with and without LD, respectively.

Modified Model for Children with LD

Contribution of Exogenous Variables to Endogenous Variables

As seen in Fig. 2, in the model modified for the children with LD, the exogenous variable attachment toward mother significantly contributed to the explanation of three out of four endogenous variables: loneliness (B = −.67), SOC (B = .52), and internalizing behavior syndrome (B = −.29). Children with LD who felt more securely attached with the mother reported lower feelings of loneliness and higher SOC, and their parents rated them as demonstrating lower levels of internalizing behavior problems, compared to children with LD who felt less securely attached with the mother.

In contrast, the exogenous variable attachment toward father significantly contributed to two of the four endogenous variables, and only one of these—SOC—replicated the pattern for child–mother attachment. Thus, father–child attachment significantly contributed only to the explanation of children’s SOC (B = .32) and children’s self-reported hope and effort (B = .52). Children with LD who felt more securely attached with the father reported higher hope and efforts and higher SOC than children with LD who felt less securely attached with the father.

Modified Model for Children without LD

Contribution of Exogenous Variables to Endogenous Variables

Overall, the model modified for the children with typical development (seen in Fig. 3) showed fewer significant paths than the model modified for children with LD. Thus, the exogenous variable attachment toward mother significantly contributed only to the explanation of children’s SOC (B = .87), and the exogenous variable attachment toward father significantly contributed only to the explanation of children’s hope and effort (B = .32). Children with typical development who felt more securely attached with the mother thus reported higher feelings of SOC, compared to children with typical development who felt less securely attached with the mother. In turn, children with typical development who felt more securely attached with the father reported higher hope and efforts than children who felt less securely attached with the father.

Correlations Between Model Variables

Overall, a similar pattern of correlations emerged between the model’s components for the two populations studied (see Figs. 2, 3). In each group, a high and significant correlation emerged between children’s attachment toward the father and toward the mother, with a slightly higher correlation in the group of children with typical development than children with LD (r = .85, p < .001 and r = .75, p < .001, respectively). Additionally, in both groups, children’s high SOC was significantly associated with a high level of hope and efforts (r = .72, p < .001 for LD and r = .90, p < .001 for typical development) and with a low level of loneliness (r = −.72, p < .001 for LD and r = −.44, p = .01 for typical development). Moreover, children’s high level of hope and efforts was significantly associated with children’s low level of loneliness in both groups (r = −.32, p = .01 for LD and r = −.28, p < .05 for typical development). Also, in neither group did a significant correlation emerge between parents’ ratings of their children’s internalizing behavior problems and any of the children’s three self-reported internalizing measures.

Discussion

Based on the conceptual models suggested by prior studies on the possible links between children’s attachment with parental figures and subsequent developmental outcomes (e.g., [6, 58]), the current study exploring the “independence model” assumed that all attachment relationships are equally important but each contributes to development in distinct domains. Specifically, the present study investigated the unique roles of children’s attachment with the father and with the mother, in explaining differences in internalizing socioemotional adjustment (i.e., loneliness, SOC, hope and effort, and internalizing behavior syndrome), among children with LD and/or with typical development.

Overall, the findings supported the study’s hypotheses, indicating a high fit between the theoretical model and the empirical findings. As hypothesized, findings showed the unique and complementary role of children’s attachment with the father and with the mother for explaining children’s internalizing variables, with a different pattern of relations between the model’s components emerging for the two populations as well as differences for the contribution of child–father attachment versus child–mother attachment. Before addressing the results of the SEM analyses, the following section first briefly discusses the findings yielded by the descriptive statistics.

LD Versus Typical Development Groups

The current descriptive outcomes clearly demonstrated that children with LD manifested more internalizing difficulties than did their typically developing peers. Thus, as expected, these children with LD as a group reported higher feelings of loneliness, lower levels of SOC, and lower levels of hope and effort compared to their peers without LD. Children with LD were also rated by their parents as manifesting higher levels of internalizing behavior problems such as depression, anxiety, and social withdrawal, compared to parent ratings of the children with typical development. These findings resemble prior studies highlighting the vulnerability of children with LD to a variety of difficulties in the social, emotional, and behavioral domains (e.g., [19, 59, 60]).

Group differences also emerged on the children’s attachment classifications with fathers, which rarely have been explored among children with LD and may therefore expand knowledge regarding two major aspects of these children’s close relationships with fathers. First, the current findings showed that children with LD reported a significantly lower incidence of attachment security toward their fathers than did their typically developing peers. Second, similar to the few prior studies among children with typical development [33], children with LD reported feeling more secure with their mothers than with their fathers.

However, unexpectedly, the current descriptive findings revealed no significant group differences regarding children’s attachment with mothers. These results differed from those reported by prior studies, where children with LD revealed a lower incidence of secure attachment with mothers than children with typical development [12, 15, 16]. One possible explanation for the current surprising outcome may be, perhaps, that the current sample’s high incidence of extra-school interventions (i.e., psychological, educational, medical, and linguistic-communication) may have influenced these children’s relationships with their mothers. For example, such interventions often offer parental supervision, which may decrease maternal feelings of burden that may interfere with secure attachment relations. Among other directions of research, future studies should examine the possible contribution of such interventions to maternal level of stress and feelings of burden associated with children’s disabilities, which, in turn, may affect the mother’s ability to provide securely attached caregiving through available, supportive, and responsive care.

Estimation of the Model

Group Differences

As hypothesized, the current study highlighted the significant role of children’s attachment relationships with mothers and fathers, in explaining differences in internalizing difficulties between the two groups. Nevertheless, the present findings clearly showed a different pattern of relations between the model’s components for the two populations studied. Thus, a greater number of significant paths emerged between attachment with parents and the internalizing adjustment measures for the LD group than for the typically developing group, suggesting that children with LD may be more vulnerable to insecurity in close relationships. These findings may extend knowledge regarding the role of insecure attachment as a risk factor in increasing maladjustment in children with LD [15, 18] and may also support prior outcomes that showed less evidence for a main effect of insecure attachment in explaining behavior problems among low risk samples (e.g., [32]).

Of particular importance, the present results raise questions regarding the contribution of specific internal neurological difficulties characterizing children with LD, such as information-processing disorders, impulsivity, and performance and production deficits, which may affect their vulnerability to insecure relationships with parents. As argued by recent studies, children may vary in their responsivity to the qualities of their environments, including their childrearing experiences (see [61] for a review). Such ideas are generally framed in diathesis-stress or “dual risk” terms (e.g., [61, 62]), suggesting that some individuals are more vulnerable whereas others may be relatively resilient to their own dual risk: on the one hand, to biological/neurological/genetic factors and/or behavioral characteristics (i.e., “stress” or “Risk 1”) and on the other hand, to the adverse effects of negative experiences and environmental qualities (i.e., “diathesis” or “Risk 2”). Thus, future studies may do well to explore the possible contribution of such neurological difficulties as well as these children’s difficulties in social information processing (see [21] for a review) as impacting their vulnerability to insecure relationships with parents.

Attachment with Fathers Versus Attachment with Mothers

The current results supported the study’s hypothesis regarding the unique and commentary roles played by children’s attachment to each parent, among the group of children with LD. The model for the children with typical development revealed fewer significant paths for both attachment figures, as discussed above.

In particular, in the model modified for children with LD, a greater number of significant paths emerged between child–mother attachment relationships and internalizing measures than for child–father attachment. Data also showed that attachment with fathers contributed mainly to children’s coping resources (i.e., SOC and hope and effort), whereas attachment with mothers contributed to a broader range of internalizing adjustment measures including not only SOC but also self-reported loneliness and parent-rated internalizing problems.

In other words, regarding attachment to fathers, those children with LD who viewed themselves as more securely attached with the father reported a higher tendency to see the world as comprehensible, manageable, and meaningful (i.e., higher SOC per Antonovsky [41]) compared to children with LD who viewed themselves as less securely attached to the father. The children who viewed themselves as more securely attached with the father also reported a higher level of hope and effort, which refers to higher pathway thinking (i.e., the perceived capacity to generate strategies for attaining goals), higher agency thinking (i.e., perceptions involving one’s capacity to initiate and sustain movement along the chosen pathways), and a higher level of investment, intensity, and persistence in task accomplishment [22, 42, 43, 63], compared to children with LD who viewed themselves as less securely attached to the father.

In contrast, regarding attachment to mothers, those children with LD who viewed themselves as more securely attached with the mother reported lower feelings of loneliness and were also rated by their parents as manifesting fewer internalizing behavior problems such as anxiety, depression, and social withdrawal, compared to children with LD who viewed themselves as less securely attached to the mother. Similarly to attachment with fathers, children who viewed themselves as more securely attached with the mother reported having higher coping resources (i.e., SOC) compared to children with less secure relationships with the mother.

The current findings based on comparisons of these two groups are preliminary; yet, these initial evidences may expand knowledge regarding the role of attachment relationships among children with and without LD in several ways. First, these outcomes concerning fathers are of particular interest in light of the paucity of studies examining the role of insecure child–father attachment as a risk factor for the LD population and its possible contribution to these children’s maladjustment. Furthermore, the group differences in child–father attachment classification between children with and without LD raise some important questions calling for additional investigation in light of previous research. For instance, prior studies have argued that fathering tends to be more sensitive to contextual factors such as children’s characteristics (e.g., [11]). Additionally, studies on families of children with various disabilities have described a maternal tendency toward over involvement with children and a paternal tendency toward emotional or physical withdrawal behaviors (e.g., [64]). Thus, one may speculate that fathers of children with LD may demonstrate low involvement and availability in their children’s lives, which in turn may contribute to a high incidence of insecure father–child relationships. Future empirical study should deepen the understanding of these important attachment relationships.

Second, the present outcomes suggest possibly unique yet complementary roles played by children’s attachment to each parent. Thus, a fewer number of significant paths emerged between child–father attachment and children’s internalizing measures, focusing mainly on children’s coping resources, whereas the greater number of significant paths that emerged between child–mother attachment indicated links with a broader range of internalizing adjustment measures. These data supported the conceptual “independence model” of links between child–parent attachment and children’s developmental outcomes [6, 58]. In line with this model, child–mother and child–father attachment relationships each contributed to development in distinct domains. As argued by recent studies (e.g., [10, 29, 32]), rather than regarding fathers as less influential or as subsidiary figures in an attachment hierarchy, the present study demonstrated that fathers fulfill different but equally influential roles.

The particular links found between child–father attachment and children’s coping resources and the links between child–mother attachment and a broader range of internalizing adjustment measures call for further scrutiny and replication. As discussed earlier, attachment relationships with fathers have been less studied and showed inconsistent findings regarding the specific direction by which children’s attachment with each parental figure may contribute to adjustment in social and emotional domains. For example, Grossmann et al. [29] initially suggested that attachment with fathers contributed to children’s exploration behaviors, whereas attachment with mothers contributes to emotional regulation. Yet, more recently, Grossmann et al. [65] highlighted that both parents can contribute to secure attachment and exploration skills.

Limitations, Directions for Future Study, and Implications

Several limitations of this study call for further research. First, the present data were collected at one point in time and did not indicate causality. To facilitate validation and generalization of these preliminary evidences, as well as to promote greater understanding of the possibly unique role of attachment with each parental figure, future studies should examine the longevity of such perceptions over time in different developmental phases (i.e., preschool, adolescence) and should investigate working models of attachment with additional significant others such as peers and teachers. Also, future research may utilize qualitative interview methods to elaborate on these children’s structured self-reports. Further studies may also investigate the possible role of attachment relations in explaining children’s externalizing factors such as aggressive and antisocial behaviors.

Second, it should be noted that the current sample size did not permit examination of personal, familial, and cultural characteristics such as parental marital status, parents’ own patterns of attachment, family life stressors, childrearing practices, as well as Israeli cultural factors (e.g., the concurrence of individualistic and collectivistic values). For example, the present sample showed a high incidence of intact families, indicating that the present findings should be interpreted with caution to avoid generalizing the findings to divorced or separated families, which may be characterized by fathers’ low levels of availability. In this context, future studies should also focus on the individual characteristics of children with LD such as their specific disabilities (e.g., reading, writing, and math) and comorbidity of attention deficit—hyperactivity disorder.

Third, the present study utilized the well-known and well-validated attachment scale that classifies attachment by differentiating secure from insecure styles [49]. To further explore children’ patterns of attachment, future research may do well to investigate, too, the possible unique contribution of two insecure attachment subclassifications—insecure avoidant style and insecure anxious style. And fourth, conceptual matters also deserve a word of caution regarding the assessment measures. Although in an effort to enhance validity, the present study included parental assessments of children’s internalizing measures and not only self-reports as recommended by Ronen [66], one may also conjecture about the possible links between child–parent attachment and parental abilities to accurately assess their offspring’s difficulties.

Finally, the current results indicated a strong predictive path between children’s SOC and their hope and effort, which calls for further empirical scrutiny of these measures’ relationship. These two constructs are similar yet differ in several aspects. For example, whereas SOC taps global cognitive, emotional, and behavioral responses to various types of stressful situations, the hope and effort scale taps a cognitive set only, focusing on selecting, initiating, and sustaining movement along chosen pathways toward specific goals.

Overall, the implications of the current findings emphasize the possible protective role of attachment relationships with both mothers and fathers for children with LD, especially when validated by further research. The current results clearly revealed the LD group’s heightened vulnerability to insecure attachment relationships with significant others, compared to the typically developing group. In accordance with Bowlby’s [4] clinical instructions, as well as prior studies [25, 67], interventions may do well to target parenting behaviors such as parents’ abilities to accurately interpret children’s needs for comfort and exploration and to respond contingently. Specifically for fathers, such interventions may help increase paternal levels of involvement, availability, and support, to provide more optimal care and a secure base for children with and without LD [64].

Summary

This study utilized an “independence model” that conceptualized all attachment relationships as equally important but each as contributing distinctly to development. As hypothesized, the current results indicated the unique and complementary roles played by children’s attachment with the father and attachment with the mother for explaining children’s internalizing socioemotional variables. Moreover, a different pattern of relations between the model’s components emerged for the two populations of children with LD and children with typical development. These outcomes clearly revealed the LD group’s heightened vulnerability to insecure attachment relationships with significant others, compared to the typically developing group. Results suggested the importance of developing empirically based interventions aiming to enhance parents’ caregiving behaviors such as the ability to accurately interpret children’s needs and exploration and to respond contingently. Specifically for fathers, such interventions may help increase paternal levels of involvement, availability, and support, to provide more optimal care and a secure base for children with and without LD.

Acknowledgments

The author would like to express her appreciation to Dee B. Ankonina for her editorial contribution.

Copyright information

© Springer Science+Business Media New York 2013