Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Oct 1.
Published in final edited form as: Attach Hum Dev. 2020 Apr 17;23(5):523–539. doi: 10.1080/14616734.2020.1751989

The mediating role of child-teacher dependency in the association between early mother-child attachment and behavior problems in middle childhood

Robin Neuhaus 1, Meghan McCormick 2, Erin O’Connor 3
PMCID: PMC7572431  NIHMSID: NIHMS1586311  PMID: 32301379

Abstract

The current study builds on past research on mother-child attachment and affective teacher-child relationships by explicitly considering the role of child-teacher dependency in preschool as a pathway through which early mother-child attachment is associated with children’s behavior problems across middle childhood. Data include direct assessments of attachment security and attachment styles, teacher reports of child-teacher dependency, and maternal reports of behavior problems from the NICHD SECCYD (N = 769 children). Children with more secure attachments at 24 months were less likely to exhibit child-teacher dependency at 54 months. Children with ambivalent, controlling, or insecure/other attachments at 36 months had higher levels of child-teacher dependency at 54 months, relative to securely attached children. Results from multi-level models showed that child-teacher dependency at 54 months was associated with higher levels of internalizing behavior problems across middle childhood (first to fifth grade), but not with externalizing behaviors. Child-teacher dependency partially mediated the association between insecure/other mother-child attachment and internalizing behaviors in middle childhood. Supporting teachers to reduce child-teacher dependency in preschool settings may help ameliorate risk for internalizing behaviors posed by insecure/other attachment.

Keywords: teacher-child relationships, attachment, dependency, externalizing behavior, internalizing behavior

Extensive literature supports the link between teacher-child relationship quality in early childhood and internalizing and externalizing behaviors across middle childhood (Hamre & Pianta, 2001; O’Connor, Dearing, & Collins, 2011). In most studies, affective teacher-child relationship quality is conceptualized in terms of closeness and conflict. The third dimension, dependency, is often considered in combination with conflict (e.g., Hamre & Pianta, 2001) or excluded from analyses entirely (e.g., O’Connor, Collins & Supplee, 2012). Though dependency has been examined far less frequently than either closeness or conflict, it has been identified as a risk factor for behavior problems in early and middle childhood (Hughes, Bullock, & Coplan, 2014). In addition, little is known about the factors – such as early mother-child attachment – that predict child-teacher dependency in early childhood, and whether dependency is a mechanism through which mother-child attachment may influence behaviors in middle childhood.

The current study aims to address these limitations in the extant literature. We first examine whether early mother-child attachment relates to child-teacher dependency in preschool. After then, replicating earlier work demonstrating associations between early attachment and behavior problems in middle childhood (O’Connor, Scott, McCormick, & Weinberg, 2014), we test whether child-teacher dependency predicts internalizing and externalizing behaviors from first to fifth grade. Finally, we examine whether child-teacher dependency mediates associations between early mother-child attachment and behavior problems in middle childhood.

Mother-child Attachment and Behavior Problems

In infancy and early childhood, there are four types of attachment styles that vary according to the overall pattern of mother-child interactions: secure, avoidant, ambivalent, and disorganized (Ainsworth, Blehar, Waters, & Wall, 1978; Bowlby, 1969, 1989). Children whose mothers respond to them with warmth and sensitivity tend to form secure mother-child attachments and perceive others as trustworthy. Avoidant attachments are characterized by demanding yet unresponsive maternal behavior that leads children to view their mother as unavailable. For children with ambivalent attachments, mothers’ responses to their children are ambiguous leading children to perceive them as inconsistently available (e.g., Main & Cassidy, 1988). Variation among children with disorganized attachments indicate two subtypes of disorganized attachment: insecure/other and controlling (Moss, Cyr, Bureau, Tarabulsy, & Dubois-Comtois, 2005). Children with controlling attachment patterns tend to perceive their mother as someone they can manipulate, while insecure/other attachment is characterized by a disjointed, unintegrated relationship model (Teti, 1999).

Meta-analytic findings support the link between insecure attachment and behavior problems in middle childhood (Fearon, Bakermans-Kranenburg, van Ijzendoorn, Lapsley, & Roisman, 2010; Madigan, Atkinson, Laurin, & Benoit, 2013). More specifically, avoidance has been linked to internalizing behavior problems (Groh, Roisman, van Ijzendoorn, Bakermans-Kranenburg, & Fearon, 2012), and disorganization overall. Avoidance, and ambivalence have been linked to externalizing behavior problems (Fearon et al., 2010). Groh and colleagues (2012) did not find a significant association between disorganization overall and internalizing behavior problems, yet another study found that insecure/other attachment, but not controlling attachment, was linked to internalizing behavior problems (O’Connor et al., 2014). This discrepancy suggests that combining the insecure/other and controlling subtypes may limit the ability to detect potential associations.

Mother-child Attachment and Child-teacher Dependency

A central tenet of attachment theory is the internalization of early experiences with caregivers and the tendency to extend those internalized models in the formation of subsequent relationships (e.g., Thompson, 2008). This framework is supported by empirical evidence of the link between insecure attachments and impairments in teacher-child relationships (Buyse, Verschueren, & Doumen, 2011; O’Connor & McCartney, 2006; O’Connor et al., 2012). Teacher-child relationship quality from an attachment perspective is generally conceptualized by the relationship’s degree of closeness, conflict, and dependency (Pianta, 2001). Closeness is characterized by warmth and open communication whereas high levels of conflict indicate antagonism and hostility. Dependency, in contrast, refers to a degree of overreliance on a teacher that is not developmentally appropriate given a child’s age (Pianta, 2001).

Sroufe, Fox, and Pancake (1983) theorized that insecure mother-child attachment limits the development of children’s autonomy, thus resulting in overly dependent behavior in early childhood teacher-child relationships. Maladaptive attention seeking strategies developed in the parent-child relationship may hinder a child’s formation of a positive teacher-child relationship and the use of the teacher as a secure base (Sroufe et al., 1983). Ambivalent and avoidant mother-child attachment have indeed been found to relate to higher levels of dependency on teachers (Sroufe, 2005). Sroufe (2005) suggests that dependent behavior may manifest differently such that ambivalent children tend to turn to their teacher immediately when facing even minimally challenging situations. In contrast, avoidant children do not turn to their teacher for support when distressed, but rather tend to hover close to their teacher during quieter times. Little is known about how disorganization and its subtypes relate to dependency.

Child-teacher Dependency and Behavior Problems

A reciprocal link between child-teacher dependency and internalizing behavior problems among preschoolers has been found in prior work (Roorda, Verschueren, Vancraeyveldt, Van Craeyevelt, & Colpin, 2014). This association may continue into middle childhood in part because of a pattern of negative reinforcement as dependent children rely on their teacher to avoid situations that provoke anxiety, particularly challenging situations with peers that invoke social coping skills (Birch & Ladd, 1998). If children find that dependent behavior alleviates their anxiety, then they may be more inclined to continue such behavior as a means of self-regulation. Though previous studies did not find that child-teacher dependency predicted externalizing behaviors among younger children (Hughes et al., 2014; Roorda et al., 2014), little research has examined dependency as it is associated with externalizing behavior in middle childhood.

The nature of the association between dependency in early childhood and behavior problems may evolve over time, underscoring the importance of examining outcomes through middle childhood. As hypothesized by Birch and Ladd (1998), dependency may become less prevalent as children progress through elementary and middle school, particularly as children transition into novel environments. Consequently, children who continue to be overly dependent may have inadequate social skills and be stigmatized by their peers leaving them vulnerable to the development of behavior problems and peer victimization in late childhood (Troop-Gordon & Kopp, 2011). Peers may become irritated because they may perceive a dependent relationship as favoritism and grow resentful or, if they sense that the teacher is irritated by the child’s behavior, they may begin to feel the same way (Mejia & Hoglund, 2016). The child may then react with elevated levels of internalizing or externalizing behaviors.

Further, a review of the association between mother-child attachment and behavior problems in middle childhood emphasized the need for additional research examining mediating mechanisms (Moss & Lecompte, 2015). Using the NICHD SECCYD sample, O’Connor and colleagues (2012) found that the association between insecure/other mother-child attachment and fifth grade externalizing behaviors was fully mediated by early externalizing behaviors at 36 months and teacher-child closeness and conflict at 54 months. However, given that insecure/other attachment remained a significant predictor of internalizing behaviors even after including teacher-child closeness and conflict as mediating variables, child-teacher dependency may be another factor partially explaining this association.

Current Study

The current study aims to extend the literature by considering child-teacher dependency as a pathway through which mother-child attachment is associated with children’s behavior problems in middle childhood. In our first research question, we ask: What is the association between secure attachment at 24 months and attachment style (avoidant, ambivalent, controlling, or insecure/other, relative to secure) at 36 months and child-teacher dependency at 54 months? As found in previous research (Sroufe, 2005), we hypothesize that avoidance and ambivalence will predict greater dependency. Insecure/other children’s characteristic lack of a coherent attachment model and difficulty using their mothers as a secure base (Moss et al., 2005) may manifest as elevated dependency in teacher-child interactions. Descriptions of controlling children as “excessively helpful” with “forced, ‘over-bright’ affect” in mother-child relationships may translate to dependent behavior in teacher-child relationships (O’Connor, Bureau, McCartney, & Lyons-Ruth, 2011, p. 468).

Our second research question is: What is the relation between child-teacher dependency at 54 months and children’s internalizing and externalizing behaviors across middle childhood? In line with previous findings (Roorda et al., 2014), we hypothesize that dependency will predict more internalizing behavior. If there is a positive association with externalizing behavior, we expect it to be a weaker relationship, considering that previous research did not find this association in early childhood (Hughes et al., 2014). The strength of these associations may increase over time as dependent behavior becomes less prevalent and potentially increasingly stigmatized.

Finally, for our third question we ask: Does child-teacher dependency mediate the association between attachment style (secure, avoidant, ambivalent, controlling, or insecure/other) at 36 months and children’s internalizing and externalizing behaviors across middle childhood? We do not put forward specific hypotheses regarding the mediation model except in the case of insecure/other attachment. Based on O’Connor and colleagues’ (2012) finding that teacher-child closeness and conflict only partially mediated the relation between insecure/other attachment and internalizing behavior, child-teacher dependency may be another factor that mediates this association.

Method

A total of 1,364 families in the U.S. enrolled in the NICHD SECCYD (see NICHD ECCRN, 2001 for recruitment and demographic information). The sample for this study consists of 769 children who had completed the assessment of mother-child attachment patterns at 36 months, had complete data on the child-teacher dependency measure at 54 months, and had at least one time point data on behavior problems at 36 months or in 1st, 3rd, or 5th grade. Because the current study is specifically interested in child-teacher dependency – which was only measured at 54 months – the sample excludes children who were not enrolled in child care outside of their home at 54 months, and thus had a caregiver available to report on this relationship. A series of independent samples t-tests demonstrated that the children included in the current study sample – meaning those children who were enrolled in care at 54 months and thus had teacher-child dependency data – were more likely to be white (82% of the current sample vs. 76% of the broader study sample; t(1362) = 12.85, p < .01). No other statistically significant differences in demographic characteristics were found comparing children in the broader NICHD SECCYD study sample to those in the current analyses. Throughout this paper we use the term “teacher” to refer to the child’s non-parental caregiver at 54 months. See Table 1 for the demographic characteristics of the current study sample.

Table 1.

Means and Standard Deviations for Study Variables

Variable Mean/Percent SD Minimum Maximum
Child female 0.51 - -
Child Black 0.10 - - -
Child Hispanic 0.06 - - -
Income to needs ratio 3.97 3.26 0 6.85
Mother-child attachment security at 24 months 0.30 0.20 0.05 0.54
Attachment style at 36 months1
 Secure attachment 0.62 - - -
 Avoidant attachment 0.05 - - -
 Ambivalent attachment 0.17 - - -
 Controlling attachment 0.09 - - -
 Insecure/other attachment 0.07 - - -
Internalizing behaviors at 36 months 8.42 5.11 0 23
Externalizing behaviors at 36 months 13.13 6.89 0 43
Average internalizing behaviors from 1st - 5th grade 4.96 4.85 0 28.43
Average externalizing behaviors from 1st - 5th grade 7.02 6.23 0 47.82
Child-teacher dependency at 54 months 7.29 2.72 4 18
Child-teacher conflict at 54 months 18.84 6.67 15 55
Child-teacher closeness at 54 months 45.53 6.31 12 49
Open in a new tab

NOTE: N = 769 children with valid child-teacher dependency data at 54 months.

N = 728 children with valid data on behaviors from 1st - 5th grade.

1

The reference group is secure attachment style.

Measures

Child-teacher dependency.

The Student Teacher Relationship Scale (STRS; Pianta, 1992) was used to assess teacher perceptions of child-teacher dependency at 54 months. Using a 5-point Likert scale, teachers rated how applicable statements were to each child. The dependency scale has 5 items, with a higher score indicating more dependency. The sum score was used for the current study, meaning that possible scores ranged from 5 to 25 (α = .84). We also used teacher perceptions of child-teacher closeness and conflict at 54 months as covariates (using sum scores) in the analytic models to assess the unique effects of child-teacher dependency on outcomes of interest (α = .92 for closeness and α = .88 for conflict). Studies have shown reliability and validity evidence on the STRS (e.g., Birch & Ladd, 1997).

Maternal reports of behavior problems.

Internalizing and externalizing behavior problems were assessed at 36 months, first (full age range 5 – 8), third (full age range 7 – 10), and fifth (full age range 9 – 12) grade using the parent version of the Child Behavior Checklist (CBCL; Achenbach, 1991). This measure has good psychometric properties (Achenbach, 1991). The CBCL contains 118 items that describe a range of child behavior problems. Mothers reported how well each item described their child currently or within the last 6 months: 0 = not true, 1 = somewhat/sometimes true, and 2 = very true or often true. Higher scores indicate more behavior problems.

Internalizing behavior problem scores, created by summing mothers’ responses to items on the Withdrawn, Somatic, and Anxious/Depressed scales, measure the child’s social inhibition and anxiety (as outlined by Achenbach, 1992). Scores on the externalizing scale, created by summing parents’ responses to items on the Aggressive and Destructive/Delinquent Behaviors scales, measure the child’s antisocial and disruptive behavior (α = .79 first grade, α = .81 third grade, α = .83 fifth grade for internalizing behaviors; α = .86 first grade, α = .88 third grade, α = .89 fifth grade for externalizing behaviors). Behavior problems at 36 months were used as baseline control variables for corresponding outcomes, and behavior problems at first, third, and fifth grades were time-varying outcome variables in predictive models.

Mother-child attachment security at 24 months.

We assessed attachment security via The Attachment Q-Set (Waters & Deane, 1985) at 24 months. Observers sorted 90 behavioral characteristics of 18- to 24-month-old children into fixed piles from most descriptive to least descriptive. The final sort conforms to a symmetric, unimodal distribution with a specified number of items in each of nine piles. The sorted items provide a profile summary of a child’s behavior as observed during a home visit. Trained research assistants conducted the Attachment Q-Set. Observers received 3-days training, including three practice visits, and were tested using five test tapes. Scores on these tapes were highly correlated (r = .77). During data collection, reliability was assessed within each site; across all research assistants, observers’ scores on the Q-Set were correlated .73 (McCartney, Owen, Booth, Vandell, & Clarke-Stewart, 2004). Due to the scale of the data collection procedure, this level of reliability was identified as adequate by the NICHD ECCRN (1997). The 24-month measure of attachment security is a continuous measure wherein a lower score indicates a lower level of attachment security than a higher score.

Mother-child attachment styles at 36 months.

A modified Strange Situation procedure involving two brief separations was used to assess attachment style at 36 months.1 The children’s behaviors during the assessment were classified as secure, ambivalent, avoidant, controlling, or insecure/other according to the system developed by the MacArthur Working Group on Attachment (Cassidy & Marvin, 1992). Intercoder agreement (before conferencing) was 75.7% (κ = .58) (see McCartney et al., 2004 for more detail). Dichotomous variables for each attachment style were created. The reference group was secure. Inclusion of the attachment measure of 36 months, in addition to the 24 month measure, allows one to examine general attachment security as well as attachment styles and sub-types of disorganized attachment, including the insecure/other group categorization, at a developmental point when attachment patterns tend to stabilize (Thompson, 2008).

Child and Family Controls

In addition to child-teacher closeness and conflict, we used child and family characteristics as covariates. Parent reports on child gender were dummy coded (female = 1; male = 0). Two dummy variables for child race/ethnicity were created (black = 1; white = 0 and Hispanic = 1; white = 0). At 36 months, the ratio of family income to needs was computed by dividing total family income by the poverty threshold for the appropriate family size and included as a covariate to account for differences in behaviors attributed to family income.

Data Analysis Plan

Missing data.

All 769 children were included in analyses for the first research question, while the sample dropped to 728 once we included outcome data. Within the 769 children in the current study sample, results from a series of independent samples t-tests comparing the students with missing and non-missing data suggest that there were no systematic patterns of missingness between the 41 children missing outcome data and the 728 children retained in the full set of analyses. As such, complete case analysis was appropriate for the data analysis conducted to examine our research questions.

Modeling strategy.

We used an OLS regression to examine the first research question using data from two distinct time points. The outcome for research questions 2 and 3 was longitudinal and these analyses had repeated measures nested within children. As such, we used STATA XTMIXED to examine multi-level models, with intercept shifts representing differences between subjects (Diggle, Heagerty, Liang, & Zeger, 2002; Singer & Willet, 2003). This is a mixed model where repeated measures are modeled at Level 1 and time-invariant person-level variables are modeled at Level 2 (Raudenbush & Bryk, 2002).

We included a variable for time where Time 1 (T1) refers to first grade, Time 2 (T2) is third grade, and Time 3 (T3) is fifth grade. Importantly, we centered time at the final time point so that time-invariant coefficients in the model would represent the effect of the predictors on behaviors in fifth grade. We also included fixed effects for site in all predictive models. All models include a random intercept for student and a random slope for time. We did not find that allowing a correlation between the random intercept and slope improved model fit. We standardized all predictors prior to running predictive analyses. We ultimately decided to fit a linear time trend, as models including quadratic effects were not statistically significant in the unconditional growth models. Additional information relevant to understanding the modeling approach is discussed in the results below.

Results

Descriptive Statistics

Means and standard deviations for continuous variables and percentages for dichotomous variables are presented in Table 1. Attachment distributions were comparable to those found in other studies (van IJzendoorn, Goldberg, Kroonenberg, & Frankel, 1992). At 36 months, the largest attachment group was secure (62%), followed by ambivalent (17%), controlling (9%), insecure/other (7%), and avoidant (5%). Bivariate correlations between key variables of interest are included in Table 2. Here, we examined the correlations between the time-varying measures of internalizing and externalizing behaviors assessed between first and fifth grades (the time-varying outcomes) and the study variables of interest. The time-varying measures are disaggregated from the values of these variables assessed at 36 months, which are considered as covariates in the analysis. As illustrated, there was a small and statistically significant correlation between child-teacher dependency at 54 months and internalizing behaviors from first to fifth grade (r = .17, p < .01). There were also small and significant correlations between child-teacher dependency at 54 months and ambivalent (r = .17, p < .05), controlling (r = .13, p < .05), and insecure/other (r = .11, p < .05) attachment styles measured at 36 months.

Table 2.

Bivariate Correlations Between Study Variables of Interest

Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
1. Internalizing behaviors in 1st - 5th grade -
2. Externalizing behaviors in 1st - 5th grade .57 ** -
3. Child female .03 .06 -
4. Child Hispanic −.01 .02 −.01 -
5. Child Black .02 .14 .01 −.08 -
6. Income to needs ratio at 36 months −.04 −.14 * .04 −.09 −.21 ** -
7. Attachment security at 24 months −.03 −.11 * .12 * −.06 −.15 * .09 -
8. Internalizing behaviors at 36 months .31 ** .32 ** .05 .05 .07 −.14 ** −.16 ** -
9. Externalizing behaviors at 36 months .28 ** .41 ** −.05 .04 .03 −.15 ** −.21 ** −.69 ** -
10. Avoidant attachment at 36 months .04 −.01 −.03 −.01 .02 −.04 −.01 −.01 −.05 -
11. Ambivalent attachment at 36 months .01 .02 .05 .01 −.02 −.04 −.07 .09 .02 −.09 -
12. Controlling attachment at 36 months .01 .01 .07 −.03 .05 .05 −.06 .02 .07 −.07 −.14 * -
13. Insecure/other attachment at 36 months .06 .02 .04 −.01 .09 −.08 −.06 .06 .07 −.06 −.12 * −.09 -
14. Child-teacher closeness at 54 months .11 * .14 * −.03 .04 .09 −.06 −.05 .13 * .22 ** −.05 .07 .02 .08 -
15. Child-teacher conflict at 54 months −.08 −.05 .09 .05 .02 .06 .03 −.11 * −.14 * −.01 −.03 −.01 −.07 −.29 ** -
16. Child-teacher dependency at 54 months .17 ** .13 .07 .03 .21 ** −.01 −.05 .13 * .05 −.01 .17 * .13 * .11 * .34 ** .16 *
Open in a new tab

Note.

**

p < .01,

*

p < .05

Multi-level Models

OLS regression analysis.

The first model regressed child-teacher dependency at 54 months on attachment security at 24 months and mother-child attachment styles (avoidant, ambivalent, controlling, insecure/other; secure attachment as the reference groups) at 36 months. All models adjusted for the set of controls discussed above – including child-teacher closeness and conflict at 54 months – as well as fixed effects for site. Results (Table 3) revealed that children with more secure attachments at 24 months had lower levels of child-teacher dependency at 54 months, on average (β = −.08, p = .03). In addition, children with ambivalent (β = .28, p = .04), controlling (β = .35, p = .03), and insecure/other attachments (β = .17, p = .04) at 36 months were more likely to demonstrate dependency in the teacher-child relationship at 54 months, relative to children with secure attachments. We did not find differences in child-teacher dependency at 54 months for children with avoidant attachments, relative to secure attachments. Finally, Black children were also more likely to be dependent on teachers at 54 months, relative to White children (β = 1.55, p < .01).

Table 3.

OLS Regression Predicting Child-Teacher Dependency from Attachment Styles

Fixed effects β SE
Constant −1.28 ** 0.42
Child Black 1.55 ** 1.62
Child Hispanic 0.27 0.21
Income to needs ratio at 36 months 0.03 * 0.01
Mother-child attachment security at 24 months −0.08 * 0.03
Internalizing behaviors at 36 months 0.02 0.01
Externalizing behaviors at 36 months −0.04 ** 0.01
Avoidant attachment at 36 months1 0.14 0.22
Ambivalent attachment at 36 months 0.28 * 0.12
Controlling attachment at 36 months 0.35 * 0.16
Insecure/other attachment at 36 months 0.17 * 0.08
Child-teacher closeness 0.12 ** 0.01
Child-teacher conflict 0.18 ** 0.01
Open in a new tab

Note. N = 769 students;

**

p < .01,

*

p < .05

1

The reference group is secure attachment style.

Longitudinal analysis.

We first ran a baseline model to examine how time and demographic predictors related to behavior problems. We anticipated that the baseline model would help us determine how well the variables of interest – attachment security, mother-child attachment styles, and child-teacher dependency – explained additional variation in the outcomes, over and above demographic shifts and normative developmental trajectories. A random intercept and slope picked up unexplained differences between the level and slope of individuals’ trajectories. We examined whether there were effects of the predictors of interest on both the intercepts and the slopes.

We then examined the association between attachment security at 24 months and attachment styles at 36 months and children’s internalizing and externalizing behaviors across first, third, and fifth grade. As illustrated in Table 4 (internalizing behaviors), and Table 5 (externalizing behaviors), results demonstrated statistically significant relationships between avoidant attachment style and internalizing behaviors (β = 1.06, p = .04; Table 4, Model 1). In addition, children with insecure/other attachment styles were more likely to exhibit internalizing behaviors in middle childhood (β = .69, p = .04; Table 4, Model 1) than securely attached children. Children with more secure attachments at 24 months had lower levels of externalizing behaviors in middle childhood, on average (β = −.14, p = .03; Table 5, Model 1). In addition, there was no evidence that attachment security at 24 months or attachment styles at 36 months were associated with change in behavior problems across middle childhood.2

Table 4.

Individual Growth Models Predicting Internalizing Behavior Problems

Model 1 (c path) Model 2 (b path)
Fixed effects β SE β SE
Constant 1.89 ** 0.43 1.40 1.19
Time 0.10 * 0.04 0.04 0.05
Female 0.28 0.20 0.23 0.24
Child Black −0.30 * 0.35 0.25 0.44
Child Hispanic −0.34 0.45 −0.41 0.57
Income to needs ratio at 36 months 0.03 0.03 0.03 0.04
Mother-child attachment security at 24 months 0.02 0.11 0.15 0.13
Internalizing behaviors at 36 months 0.19 ** 0.02 0.22 ** 0.03
Externalizing behaviors at 36 months 0.09 ** 0.02 0.08 ** 0.01
Avoidant attachment at 36 months 1.06 * 0.48 - -
Ambivalent attachment at 36 months −0.17 0.28 - -
Controlling attachment at 36 months 0.08 0.36 - -
Insecure/other attachment at 36 months 0.69 * 0.31 - -
Child-teacher closeness - - −0.03 0.02
Child-teacher conflict - - 0.02 0.02
Child-teacher dependency - - 0.08 * 0.04
Random effects
Intercept 6.21 ** 0.48 6.01 ** 0.56
Grade slope 0.03 0.02 0.02 ** 0.01
Residual variation 15.24 ** 0.43 14.98 ** 0.49
AIC 20974.77 14875.33
BIC 21129.48 15015.67
Open in a new tab

Note. N = 728 students;

**

p < 0.01,

*

p < 0.05,

†

p < 0.1

Table 5.

Individual Growth Models Predicting Externalizing Behavior Problems

Model 1 (c path) Model 2 (b path)
Fixed effects β SE β SE
Constant 4.20 ** 0.52 −0.13 1.42
Time −0.40 ** 0.06 −0.41 ** 0.06
Female −0.61 * 0.25 −0.50 † 0.29
Child Black 0.35 0.43 0.76 0.54
Child Hispanic 0.05 0.56 −0.18 0.67
Income to needs ratio at 36 months −0.08 † 0.04 −0.08 † 0.05
Mother-child attachment security at 24 months −0.14 * 0.05 −0.07 0.15
Internalizing behaviors at 36 months 0.04 0.03 0.08 † 0.04
Externalizing behaviors at 36 months 0.32 ** 0.03 0.31 ** 0.03
Avoidant attachment at 36 months −0.13 0.59 - -
Ambivalent attachment at 36 months 0.07 0.35 - -
Controlling attachment at 36 months −0.21 0.44 - -
Insecure/other attachment at 36 months 0.12 0.50 - -
Child-teacher closeness - - 0.05 * 0.02
Child-teacher conflict - - 0.11 ** 0.03
Child-teacher dependency - - −0.06 0.06
Random effects
Intercept 8.89 ** 0.73 7.99 ** 0.8
Grade slope 0.02 0.02 0.02 0.02
Residual variation 24.15 ** 0.67 23.55 ** 0.78
AIC 22571.23 15965.72
BIC 22725.94 16106.06
Open in a new tab

Note. N = 728 students;

**

p < 0.01,

*

p < 0.05,

†

p < 0.1

Finally, we tested whether there were associations between child-teacher dependency at 54 months and behaviors in middle childhood. We found that child-teacher dependency at 54 months was associated with more internalizing behavior problems across middle childhood (β = .08, p = .04; Table 4, Model 2), but not with externalizing behaviors (Table 5). Although we found that inclusion of a random slope was warranted in our models, we did not find a significant slope effect for either internalizing or externalizing behaviors, meaning that there was no evidence that the effect of child-teacher dependency at 54 months affected change in behavior problems across time.

Mediation tests.

Analyses conducted thus far established evidence for three statistically significant associations suggesting mediation: 1) insecure/other attachment style at 36 months and internalizing behaviors in middle childhood (c path); 2) insecure/other attachment style at 36 months and child-teacher dependency at 54 months (a path); and 3) child-teacher dependency at 54 months and internalizing behaviors in middle childhood (b path). We used the framework for multi-level mediation recommended by Zhang, Zyphur, and Preacher (2009) to test whether child-teacher dependency at 54 months – a Level 2 variable – mediated associations between insecure/other attachment style at 36 months (another Level 2 variable) and children’s internalizing behaviors across middle childhood (a Level 1 variable). Results revealed that child-teacher dependency at 54 months partially mediated the association between insecure/other attachment (β = .11, p = .04) and internalizing behaviors in middle childhood (see Figure 1). We determined that there was evidence of partial mediation because the effect of insecure/other was still statistically significant in the model after we had adjusted for the mediator – child-teacher dependency at 54 months. We calculated an indirect effect of .05 standard deviations by multiplying the coefficients on the a and b paths and used Sobel’s test (using the calculator developed by Preacher & Hayes, 2004), to determine that the indirect effect was statistically significant (z = 2.79, p < .01).

Figure 1.

Figure 1.

Open in a new tab

Standardized estimates of child-teacher dependency as a mediator for the association between insecure/other attachment and internalizing behavior problems.

Discussion

According to attachment theory (Bowlby, 1969, 1989), early mother-child attachments shape children’s approach to subsequent affective relationships and social-emotional development. Extensive research has explored associations between early attachment and behaviors across middle childhood, yet few studies have examined mediating mechanisms. The longitudinal data from the NICHD SECCYD provide a large enough sample size to allow for examination of specific attachment styles in early childhood, as well as their links to mediators and outcomes across middle childhood. The current study adds to the growing body of research on mediational processes in attachment research with evidence of child-teacher dependency as a mechanism partially explaining the association between mother-child attachment and internalizing, but not externalizing, behavior in middle childhood. Moreover, a key strength of the current study is its focus on child-teacher dependency while controlling for teacher-child closeness and conflict.

Mother-child Attachment and Child-teacher Dependency

Our findings indicate that children with ambivalent, controlling, or insecure/other attachments at 36 months had elevated levels of dependency on their teacher at 54 months relative to securely attached children. In contrast with previous research (Sroufe, 2005), we did not find an association between avoidance and dependency. This may be due to sample differences and/or when attachment was assessed. Sroufe (2005) studied a sample of mothers identified to be at moderate risk of parenting difficulties, whereas the NICHD sample is relatively more representative.

Ambivalent children’s overreliance on their teacher is similar to the clingy behaviors they tend to use in response to inconsistent maternal care. This suggests that they may transfer their behavior with their mother to the interactions with their teacher. A similar transfer of behavior may have occurred among controlling children. Controlling children’s elevated dependency may be a reflection of the combination of excessively controlling and overly solicitous behavior observed in mother-child interactions in previous studies (O’Connor, Bureau et al., 2011). Insecure/other children’s characteristic lack of a consistent pattern of dyadic behavior may make them vulnerable to developing dependent relationships with preschool teachers. Given the association between insecure/other attachment and negative outcomes, such as greater difficulties at home and anxious behaviors (Hughes et al., 2014), teachers may be more tolerant of dependent behavior or even inadvertently reinforce dependence by spending more time with such children to support them in engagement with their peers and learning context (Coplan & Prakash, 2003). For insecure/other, ambivalent, and controlling children, their preschool teachers may be the first caregivers with whom they have consistently positive, warm interactions. They may turn to their teacher for attention, reassurance, and comfort, aspects of their mother-child relationship that may be lacking or inconsistent. Yet, they continue to have difficulty using their teacher as a secure base from which to explore their broader environment.

Child-teacher Dependency and Behavior Problems

This study also identified an association between dependency at 54 months and internalizing behavior problems across middle childhood. Greater dependency on teachers may prevent children from learning to effectively manage their internalizing behavior problems themselves if they use their teacher to avoid anxiety provoking situations (Birch & Ladd, 1998). Dependency was not associated with externalizing behaviors across middle childhood. This is consistent with findings from previous studies showing that the behaviors of overly dependent children were more likely to be perceived as anxious rather than as aggressive or hostile (e.g., Hughes et al., 2014; Roorda, Zee, & Koomen, 2020).

Mediating Mechanisms

The association between mother-child attachment and later internalizing and externalizing behavior has received substantial attention, yet relatively little is known about the processes through which attachment is associated with behavior problems. Meta-analyses have called for mediation studies in attachment research to explain these associations (Moss & Lecompte, 2015; Thompson, 2008). Results from the current study indicate that child-teacher dependency partially mediate associations with internalizing behaviors for children with insecure/other attachment. Insecure/other children tend to be fearful and uncertain (Teti, 1999), and if they are overly dependent on preschool teachers, then they may not have opportunities to develop the self-reliance and social skills that would help to prevent internalizing behavior problems later in childhood.

Our finding that insecure/other children exhibited higher levels of internalizing behaviors than securely attached children contradicts the findings of two meta-analyses (Groh et al., 2012; Madigan et al., 2013) which did not find evidence for an association between disorganization and internalizing problems. This discrepancy may relate to age given that analyses combining internalizing behavior in early and middle childhood may mask the link between disorganization and internalizing behavior in middle childhood (Moss & Lecompte, 2015). It may also be due to varying trajectories of disorganized subtypes. As suggested previously (e.g., McCormick, O’Connor, & Barnes, 2016; O’Connor et al., 2014), examining disorganization as a single classification may inhibit researchers’ ability to examine the implications of each of its subtypes. The inclusion of controlling and insecure/other subtypes of disorganization is therefore a core strength of the current study.

Limitations and Future Research

Additional studies are necessary to understand pathways through which mother-child attachment relates to psychological and behavioral outcomes in middle childhood. Examples of further potential mediating mechanisms include peer relations, parent-child interactions, and child emotion regulation (for a review, see Moss & Lecompte, 2015). The current study also has a few key limitations that future work should aim to address. First, findings from this study cannot be interpreted causally as the data are observational and not experimental. In addition, there is missingness in the child-teacher dependency variable at 54 months. As such, findings are only generalizable to the group of students for whom the variable was available. More contemporary samples are more likely to reflect a broader sample of children enrolled in non-parental care at 54 months. Future studies should consider collecting information on child-teacher dependency on a larger more contemporary sample in order to address 2020 of missingness in the NICHD SECCYD. This study also found that Black children were at heightened risk for child-teacher dependency at 54 month. This result may be indicative of the teacher-reported nature of the measure and it may be important to use other modes of data collection to understand how reports of dependency are affected by children’s racial/ethnic and demographic characteristics. More work specific to this heightened risk for Black children is needed in particular, as it was not originally theorized, although race/ethnic differences in teacher-child closeness and conflict have been considered before. Moreover, future work should consider further exploring the roles that closeness, conflict, and dependency jointly play in predicting future behavior problems. Although we sought to use a variety of measures in this study – assessments, teacher reports, parent reports – in order to avoid mono-method biases, it is also true that we could consider more objective measures of dependency and behaviors not subject to reporter biases such as a direct observation conducted by an outside observer (see Verschueren & Spilt, 2020).

Conclusion

The current study contributes to our understanding of the role of teacher-child relationships in explaining associations between early mother-child attachment quality and behavior problems in middle childhood. Findings add to the relatively limited research on child-teacher dependency by examining how this understudied construct relates to both mother-child attachment styles and to internalizing and externalizing behaviors. Findings demonstrating that child-teacher dependency mediates the relationship between early mother-child attachment and internalizing behaviors in middle childhood suggest that children with insecure/other attachments may benefit from additional relational support at school in early childhood.

Footnotes

Footnotes

1

After 3 minutes together in the laboratory, the mother was signaled to leave. The first separation lasted 3 minutes, unless the child was overly distressed. After a 3-minute reunion, the mother left again. The second separation lasted 5 minutes.

2

For the sake of brevity, we do not include the non-significant slope effects in the predictive models because they are close to 0.

Contributor Information

Robin Neuhaus, New York University.

Meghan McCormick, MDRC.

Erin O’Connor, New York University.

References

  1. Achenbach TM (1991). Manual for Child Behavior Checklist/ 4–18 and 1991 Profile. Burlington, VT: University of Vermont, Dept. of Psychiatry. [Google Scholar]
  2. Achenbach TM (1992). Manual for Child Behavior Checklist/ 2–3 and 1992 Profile. Burlington, VT: University of Vermont, Dept. of Psychiatry. [Google Scholar]
  3. Ainsworth MDS, Blehar MC, Waters E, & Wall S (1978). Patterns of attachment: A psychological study of the strange situation. Hillsdale, NJ: Erlbaum. [Google Scholar]
  4. Birch SH, & Ladd GW (1997). The teacher-child relationship and children’s early school adjustment. Journal of School Psychology, 35(1), 61–79. doi: 10.1016/S0022-4405(96)00029-5 [DOI] [Google Scholar]
  5. Birch SH, & Ladd GW (1998). Children’s interpersonal behaviors and the teacher-child relationship. Developmental Psychology, 34, 934–946. doi: 10.1037/0012-1649.34.5.934 [DOI] [PubMed] [Google Scholar]
  6. Bowlby J (1969). Attachment and Loss: Vol. 1. Attachment. New York, NY: Basic Books. [Google Scholar]
  7. Bowlby J (1989). The role of attachment in personality development and psychopathology. In Greenspan SI & Pollock GH (Eds.), The course of life: Vol. 1 (pp. 229–270). Madison, CT: International Universities Press. [Google Scholar]
  8. Buyse E, Verschueren K, & Doumen S (2011). Preschoolers’ attachment to mother and risk for adjustment problems in kindergarten: Can teachers make a difference? Social Development, 20, 33–50. doi: 10.1111/j.1467-9507.2009.00555.x [DOI] [Google Scholar]
  9. Cassidy J, & Marvin RS (1992). Attachment organization in preschool children: Procedures and coding manual. Unpublished manuscript. Pennsylvania State University and University of Virginia. [Google Scholar]
  10. Coplan R, & Prakash K (2003). Spending time with teacher: Characteristics of preschoolers who frequently elicit versus initiate interactions with teachers. Early Childhood Research Quarterly, 18, 143–158. doi: 10.1016/S0885-2006(03)00009-7 [DOI] [Google Scholar]
  11. Diggle P, Heagerty P, Liang K, & Zeger S (2002). Analysis of longitudinal data (2nd ed.). Oxford: Oxford University Press. [Google Scholar]
  12. Fearon RP, Bakermans-Kranenburg MJ, van Ijzendoorn MH, Lapsley A-M, & Roisman GI (2010). The significance of insecure attachment and disorganization in the development of children’s externalizing behavior: A meta-analytic study. Child Development, 81, 435–456. doi: 10.1111/j.1467-8624.2009.01405.x [DOI] [PubMed] [Google Scholar]
  13. Groh AM, Roisman GI, van Ijzendoorn MH, Bakermans-Kranenburg MJ, & Fearon RP (2012). The significance of disorganized attachment for children’s internalizing symptoms: A meta-analytic study. Child Development, 83, 591–610. doi: 10.1111/j.1467-8624.2011.01711.x [DOI] [PubMed] [Google Scholar]
  14. Hamre BK, & Pianta RC (2001). Early teacher–child relationships and the trajectory of children’s school outcomes through eighth grade. Child Development, 72(2), 625–638. doi: 10.1111/1467-8624.00301 [DOI] [PubMed] [Google Scholar]
  15. Hughes K, Bullock A, & Coplan RJ (2014). A person-centred analysis of teacher–child relationships in early childhood. British Journal of Educational Psychology, 84, 253–267. doi: 10.1111/bjep.12029 [DOI] [PubMed] [Google Scholar]
  16. Madigan S, Atkinson L, Laurin K, & Benoit D (2013). Attachment and internalizing behavior in early childhood: A meta-analysis. Developmental Psychology, 49(4), 672–689. doi: 10.1037/a0028793 [DOI] [PubMed] [Google Scholar]
  17. Main M, & Cassidy J (1988). Categories of response to reunion with the parent at age 6: Predictable from infant attachment classifications and stable over a 1-month period. Developmental Psychology, 24(3), 415–426. doi: 10.1037/0012-1649.24.3.415 [DOI] [Google Scholar]
  18. McCartney K, Owen MT, Booth CL, Vandell DL, & Clarke-Stewart A (2004). Testing a maternal attachment model of behavior problems in early childhood. Journal of Child Psychology and Psychiatry and Allied Disciplines, 45(4), 765–778. doi: 10.1111/j.1469-7610.2004.00270.x [DOI] [PubMed] [Google Scholar]
  19. McCormick MP, O’Connor E, & Barnes SP (2016). Mother-child attachment styles and math and reading skills in middle childhood: The mediating role of children’s exploration and engagement. Early Childhood Research Quarterly, 36, 295–306. doi: 10.1016/j.ecresq.2016.01.011 [DOI] [Google Scholar]
  20. Mejia T, & Hoglund WLG (2016). Do children’s adjustment problems contribute to teacher-child relationship quality? Support for a child-driven model. Early Childhood Research Quarterly, 34, 13–26. doi: 10.1016/j.ecresq.2015.08.003 [DOI] [Google Scholar]
  21. Moss E, Cyr C, Bureau JF, Tarabulsy GM, & Dubois-Comtois K (2005). Stability of attachment during the preschool period. Developmental Psychology, 41, 773–783. doi: 10.1037/0012-1649.41.5.773 [DOI] [PubMed] [Google Scholar]
  22. Moss E, & Lecompte V (2015). Attachment and socioemotional problems in middle childhood. New Directions for Child and Adolescent Development, 148, 63–76. doi: 10.1002/cad.20095 [DOI] [PubMed] [Google Scholar]
  23. NICHD Early Child Care Research Network (1997). The effects of infant child care on infant–mother attachment security: Results of the NICHD study of early child care. Child Development, 68, 860–879. doi: 10.1111/j.1467-8624.1997.tb01967.x. [DOI] [PubMed] [Google Scholar]
  24. NICHD Early Child Care Research Network (2001). Nonmaternal care and family factors in early development: An overview of the NICHD Study of Early Child Care. Journal of Applied Developmental Psychology, 22, 559–579. doi: 10.1016/S0193-3973(01)00092-2 [DOI] [Google Scholar]
  25. O’Connor E, Bureau J-F, McCartney K, & Lyons-Ruth K (2011). Risks and outcomes associated with disorganized/controlling patterns of attachment at age three years in the National Institute of Child Health & Human Development Study of Early Child Care and Youth Development. Infant Mental Health Journal, 32, 450–472. doi: 10.1002/imhj.20305 [DOI] [PubMed] [Google Scholar]
  26. O’Connor E, Collins B, & Supplee L (2012). Behavior problems in late childhood: the roles of early maternal attachment and teacher-child relationship trajectories. Attachment & Human Development, 14(3), 265–288. doi: 10.1080/14616734.2012.672280. [DOI] [PubMed] [Google Scholar]
  27. O’Connor E, Dearing E, & Collins BA (2011). Teacher-child relationship and behavior problem trajectories in elementary school. American Educational Research Journal, 48(1), 120–162. doi: 10.3102/0002831210365008 [DOI] [Google Scholar]
  28. O’Connor E, & McCartney K (2006). Testing associations between young children’s relationships with mothers and teachers. Journal of Educational Psychology, 98(1), 87–98. doi: 10.1037/0022-0663.98.1.87 [DOI] [Google Scholar]
  29. O’Connor E, Scott MA, McCormick MP, & Weinberg SL (2014). Early mother–child attachment and behavior problems in middle-childhood: The role of the subsequent caregiving environment. Attachment & Human Development, 16, 590–612. doi: 10.1080/14616734.2014.937817 [DOI] [PubMed] [Google Scholar]
  30. Pianta RC (1992). Conceptual and methodological issues in research on relationships between children and nonparental adults. In Pianta Robert (Ed.), New directions for child development: Vol. 57. Beyond the parent: The role of other adults in children’s lives. (pp. 121–129). San Francisco: Jossey-Bass Inc. doi: 10.1002/cd.23219925709 [DOI] [Google Scholar]
  31. Pianta RC (2001). STRS, Student–Teacher Relationship Scale, Professional manual. Lutz, FL: Psychological Assessment Resources. [Google Scholar]
  32. Preacher KJ, & Hayes AF (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments, & Computers. 36(4), 717–731. doi: 10.3758/BF03206553 [DOI] [PubMed] [Google Scholar]
  33. Raudenbush SW, & Bryk AS (2002). Hierarchical linear models: Applications and data analysis methods (2nd ed.). Thousand Oaks, CA: Sage Publications Inc. [Google Scholar]
  34. Roorda DL, Verschueren K, Vancraeyveldt C, Van Craeyevelt S, & Colpin H (2014). Teacher-child relationships and behavioral adjustment: Transactional links for preschool boys at risk. Journal of School Psychology, 52(5), 495–510. doi: 10.1016/j.jsp.2014.06.004 [DOI] [PubMed] [Google Scholar]
  35. Roorda Zee, & Koomen, 2020
  36. Singer JD, & Willett JB (2003). Survival analysis. In Velicer JA, Schinka WF (Eds.), Handbook of psychology, Volume 2: Research methods in psychology (pp. 555–580). New York, NY: John Wiley. doi: 10.1002/0471264385.wei0222 [DOI] [Google Scholar]
  37. Sroufe LA (2005). Attachment and development: A prospective, longitudinal study from birth to adulthood. Attachment & Human Development, 7(4), 349–367. doi: 10.1080/14616730500365928 [DOI] [PubMed] [Google Scholar]
  38. Sroufe LA, Fox N, & Pancake VR (1983). Attachment and dependency in developmental perspective. Child Development, 54(6), 1615–27. doi: 10.2307/1129825 [DOI] [Google Scholar]
  39. Teti DM (1999). Conceptualizations of disorganization in the preschool years: An integration. In Solomon J & George C (Eds.), Attachment disorganization (pp. 213–242). New York, NY: Guilford Press. [Google Scholar]
  40. Thompson RA (2008). Measure twice, cut once: Attachment theory and the NICHD Study of Early Child Care and Youth Development. Attachment & Human Development, 10(3), 287–297. doi: 10.1080/14616730802113604 [DOI] [PubMed] [Google Scholar]
  41. Troop-Gordon W, & Kopp J (2011). Teacher-child relationship quality and children’s peer victimization and aggressive behavior in late childhood. Social Development, 20, 536–561. doi: 10.1111/j.1467-9507.2011.00604.x [DOI] [Google Scholar]
  42. van IJzendoorn MH, Goldberg S, Kroonenberg PM, & Frankel OJ (1992). The relative effects of maternal and child problems on the quality of attachment: A meta-analysis of attachment in clinical samples. Child Development, 63, 840–858. doi: 10.1111/j.1467-8624.1992.tb01665.x [DOI] [PubMed] [Google Scholar]
  43. Verschueren & Spilt, 2020
  44. Waters E, & Deane KE (1985). Defining and assessing individual differences in attachment relationships: Q-methodology and the organization of behavior in infancy and early childhood. Monographs of the Society for Research in Child Development, 50(1–2), 41–65. doi: 10.2307/3333826 [DOI] [Google Scholar]
  45. Zhang Z, Zyphur MJ, & Preacher KJ (2009). Testing multilevel mediation using hierarchical linear models: Problems and solutions. Organizational Research Methods, 12(4), 695–719. doi: 10.1177/1094428108327450 [DOI] [Google Scholar]

RESOURCES