Toward a further understanding of teachers’ reports of early teacher–child relationships: Examining the roles of behavior appraisals and attributions

https://doi.org/10.1016/j.ecresq.2009.03.001Get rights and content

Abstract

This study examined teachers’ reports of early teacher–child relationships by focusing on their assessments of the severity and the causes of children's social behaviors. Eighty-one kindergarten teachers filled out questionnaires about socially inhibited, hyperactive, and average children (n = 237) selected from their own classes. Multilevel analyses indicated that teachers reported less close and more dependent relationships for the inhibited and hyperactive versus the average children, and more conflictual relationships for the hyperactive versus the average children. These differences were largely mediated by teachers’ perceptions of children's personal behavior problems. In addition, we found that the teachers’ control attributions for children's social behaviors increased the link between children's perceived (personal and social) problems and relationship closeness. Results further support the idea that teachers’ relationship reports are personal, evaluative accounts rather than objective measures of teacher–child interactions.

Introduction

Several studies have shown that early teacher–child relationships are important determinants of children's socioemotional and academic functioning (Birch & Ladd, 1997; Hamre & Pianta, 2001; Howes, 2000; Silver, Measelle, Armstrong, & Essex, 2005). Most of the available research either tacitly or explicitly acknowledges that teacher and child both contribute to the bond between them. Yet, teacher–child relationships are more often examined as the product of attributes of children rather than as characteristics of teachers. For instance, studies have focused on the correspondence between mother–child and teacher–child relationships (e.g., Pianta, Nimetz, & Bennett, 1997) or examined the latter as child-specific determinants of classroom adaptation (cf., Birch & Ladd, 1998; Ladd & Burgess, 1999). This focus on the child is quite remarkable, particularly so because most of the available studies have relied on teacher reports.

More recently, researchers have begun to show how personal characteristics of teachers can influence their assessments of the teacher–child relationship (Hamre, Pianta, Downer, & Mashburn, 2008; Kesner, 2000; Mashburn, Hamre, Downer, & Pianta, 2006). The present study aims to expand upon this line of research by examining how teachers’ reports of their relationships with different types of kindergartners (socially inhibited, hyperactive, and average) are related to their personal assessments of the severity and the causes of these children's behaviors. We hypothesized that it is teachers’ appraisals and attributions rather than children's mere behavioral characteristics that affect teachers’ perceptions of the teacher–child relationship. To our knowledge, these hypotheses have not been investigated before.

Teacher–child relationships are viewed as micro-systems with important implications for children's development. They consist of multiple interrelated components that involve characteristics of both child and teacher, and the interactions and communication between them (Pianta, Hamre, & Stuhlman, 2003). Central among these components are the perceptions both partners have about the relationship. Rather than neutral recordings of self–other interactions, these inside perceptions are personal representations colored by feelings, evaluations, beliefs, and expectations. Still, they have important consequences for actual teacher–child interactions, because they are psychologically real and influence the behaviors of each relationship partner (Pianta et al., 2003; Pianta & Nimetz, 1991; Stuhlman & Pianta, 2002).

Most of the research on student-teacher relationships has relied on teachers’ relationship perceptions. A well-known instrument to assess these perceptions is the Student Teacher Relationship Scale (STRS; Pianta, 2001). With the STRS teachers can evaluate their relationships with individual students along the dimensions of closeness, dependency, and conflict. Closeness refers to the degree to which teacher and child interact and communicate in a warm, open, and positive manner. By contrast, dependency and conflict are negative relationship characteristics referring to, respectively, students’ overdependence on their teachers, and the extent to which the relationship is characterized by anger and negativity (Pianta, 2001).

Only a few studies have paid attention to the degree of concordance between teachers’ perceptions of the relationship (assessed with the STRS) and the perception of independent external observers (Howes & Ritchie, 1999; NICHD Early Child Care Research Network, 2003; Stuhlman & Pianta, 2002). With respect to the conflict dimension, the results of these studies consistently showed moderate agreement. However, dependency was not included, and results for closeness were inconclusive. A recent study of Doumen, Verschueren, and Buyse (2008) was the first to reveal moderate convergence between teacher and observer reports for all three relationship dimensions. These findings attest to the subjective nature of teachers’ perspectives on the relationships with their students.

Researchers have used the STRS to show that the quality of the teacher–child relationship depends on a wide variety of child characteristics, including children's behavioral orientation (Birch & Ladd, 1998; Ladd & Burgess, 1999), relationships with parents (Howes & Matheson, 1992; Rydell, Bohlin, & Thorell, 2005), temperament and language skills (Rudasill, Rimm-Kaufmann, Justice, & Pence, 2006), second language command (Fumoto, Hargreaves, & Maxwell, 2007), and intellectual disability (Eisenhower, Blaker, & Blacher, 2007). Less is known, however, about teachers’ contributions to the teacher–child relationship. Recent research by Mashburn and colleagues (2006) revealed systematic differences between teachers in their reports of relationships with individual students. Using multilevel modeling, these authors found that 33% of the variance in closeness in their sample was at the teacher level. Accordingly, significant relations have been reported between STRS ratings and individual teacher characteristics, including self-efficacy, depressive feelings, teaching experience, and personal attachment history (Hamre et al., 2008, Kesner, 2000, Mashburn et al., 2006).

In this study, we examined teachers’ contributions to their reports of relationships with individual students, by focusing on their unique and subjective assessments (appraisals and attributions) of each student's social behaviors. Thus, unlike the aforementioned studies, our main focus was not so much on the differences between teachers, but on the variation within their relationship ratings. We wanted to obtain variation within teachers, but were hesitant to ask too much of their time. Hence, rather than examining teachers in relation to all their students, we questioned them about three different types of children: kindergartners selected as either socially inhibited, hyperactive, or average (i.e. noninhibited and nonhyperactive). Social inhibition and hyperactivity are representative of the two broadband categories of problem behavior, i.e. internalizing and externalizing behaviors (Wenar & Kerig, 2000). We focused on these subcategories to facilitate the selection of children with non-overlapping behaviors. Data from our previous research (Thijs, Koomen, de Jong, van der Leij, & van Leeuwen, 2004) indicated that, unlike internalizing and externalizing behaviors, hyperactivity and social inhibition were negatively related. Socially inhibited children tend to be fearful and wary in challenging social situations, and may go unnoticed in classroom situations (Asendorpf, 1993; Rubin & Burgess, 2001). By contrast, hyperactive children show motor restlessness and are frequently off task (Hinshaw, 1987; Wenar & Kerig, 2000). Although contradictory findings have been reported for conflict, there are indications that teachers perceive more dependent and especially less close relationships with inhibited children compared to noninhibited children (Birch & Ladd, 1998; Ladd & Burgess, 1999; Rudasill et al., 2006; Rydell et al., 2005). It also appears that teachers rate their relationships with hyperactive children as less close, more dependent, and especially more conflictual (Birch & Ladd, 1998; Ladd & Burgess, 1999). In keeping with these findings, we expected to find large differences in teachers’ relationship ratings for the three types of children.

In clinical child psychology, behaviors are often evaluated in terms of their consequences. According to Rutter (1975), child behavior can be considered problematic if it causes a child's suffering, impedes in his or her socioemotional functioning, interrupts typical development, and/or negatively affects others. One assumption is that professional educators and teachers are attentive to the personal and social consequences of young children's behaviors (see Chazan, Laing, & Harper, 1987). Thus, the present research used Rutter's criteria to examine teachers’ appraisal of social behaviors of inhibited, hyperactive, and average children. Research suggests that childhood social inhibition and hyperactivity generally fulfill the first three criteria. Socially inhibited children tend to experience anxiety and insecurity (Rubin & Asendorpf, 1993), and may develop depressive symptoms (Gazelle & Ladd, 2003) and feelings of social incompetence (Hymel, Rubin, Rowden, & LeMare, 1990). Likewise, there are indications that hyperactive children are more depressed (Treuting & Hinshaw, 2001), less socially skilled (DeWolve, Byrne, & Bawden, 2000; Merrell & Wolfe, 1998), and have more risk for developing learning problems and behavior difficulties (McGee, Partridge, Williams, & Silva, 1991). Although teachers may sometimes rate internalizing behaviors as less serious than externalizing behaviors (Chang & Sue, 2003), they are probably aware of these negative personal problems for both the inhibited and the hyperactive child. However with respect to negative social problems, Rutter's (1975) fourth criterion, we can anticipate a clear difference between teachers’ appraisals of socially inhibited versus hyperactive children. Due to their quiet and restrained behaviors, inhibited children are unlikely to annoy or bother their classmates and teachers (cf., Coplan & Rubin, 1998; Rubin & Asendorpf, 1993). By contrast, hyperactive children tend to have a negative influence on others as they often display disturbing behaviors that frequently go together with conduct problems or aggression (Hinshaw, 1987; Klein & Mannuzza, 1991).

In the present study, we evaluated the hypothesis that teachers’ appraisals of children's behavior mediate the differences in the quality of the relationship reported for different types of children. We anticipated, for example, that teachers would report sharing less close relationships with inhibited versus average children due to the larger (perceived) personal problems of the former. We had two reasons for this until now unexamined expectation. First, as personal representations, teachers’ relationship perceptions involve subjective evaluations, rather than neutral observations of children and their behaviors (see Mashburn et al., 2006, Pianta et al., 2003). To the extent that teachers’ appraisals are also subjectively biased, they are likely to explain the link between children's behavior characteristics and teachers’ relationship reports. Second, insofar as teachers’ appraisals provide accurate descriptions of the problematic aspects of children's behavior characteristics, these appraisals may have a greater impact on teacher–child interactions than these characteristics themselves. Children who experience more serious personal suffering and a greater lack of social competence are probably hindered in forming and maintaining favorable relationships with their teacher. Moreover, teachers may have difficulties interacting with children whose behaviors are disturbing to others.

To further examine the subjective nature of teachers’ reports of the teacher–child relationship we also assessed their attributions for children's social behaviors. There is clear evidence that causal attributions (or lay explanations) are important for the perceptions and quality of intimate adult relations (see e.g., Bradbury & Fincham, 1990; Collins, Ford, Guichard, & Allard, 2006). However, little is known about the impact of attributions on teachers’ representations of the teacher–child relationship. According to interpersonal attribution theory (Weiner, 1995, Weiner, 2000), people's causal attributions for the outcomes received by others predict their emotional reactions toward them. Interpersonal attributions can be characterized into three dimensions: locus, involving the question whether the perceived cause lies within or outside the other; stability, the extent to which the perceived cause is stable over time; and control, the extent to which the other is presumed to have control over his or her outcomes. The last of these dimensions appears to be most important for people's interpersonal judgments. When teachers attribute students’ negative outcomes (e.g., school failure) to controllable factors (e.g., a lack of effort) they will hold these students responsible and hence react negatively toward them (e.g., by showing anger). However, when the same outcomes are attributed to uncontrollable factors (e.g., a lack of ability) positive emotional reactions such as sympathy or pity are more likely (Weiner, 1995, Weiner, 2000). Although few studies have examined teachers’ attributions toward children's classroom behaviors (Arbeau & Coplan, 2007), the available evidence indicates that teachers who think that students have control over their problem behaviors show negative emotional and behavioral reactions toward them (see for a review, Ho, 2004).

To date, no studies have considered how teachers’ perceptions of the teacher–child relationship are affected by their attributions for children's social behaviors. However, as subjective representations, these perceptions reflect teachers’ feelings about their students and their interactions with them (see Pianta et al., 2003). Hence, given the premises of interpersonal attribution theory, it is reasonable to expect that teachers’ perceptions of control increase the impact of their appraisals of children's behavior on their relationship reports. It can be anticipated, for instance, that teachers will report less close relationships for children with more (perceived) personal behavior problems, but particularly so when these children are perceived to be in control of their own behaviors. In the present research, we evaluated this possibility by testing the interactions between teachers’ appraisals of children's behavior characteristics and their attributions of control. For exploratory purposes and reasons of completeness, the other attribution dimensions, locus and stability, were also included in the analyses.

The goal of this study was to investigate how teachers’ reports of early teacher–child relationships are based upon their personal appraisals and attributions for children's social behaviors. Three main hypotheses were evaluated and summarized in Fig. 1. First, based upon the literature and previous research including teachers and children from the present study (Thijs, Koomen, & van der Leij, 2008), we anticipated that teachers would report less close and more dependent relationships for the inhibited and hyperactive children versus the average children, and more conflicted relationships for the hyperactive versus the average children. Unlike, the present study, the previous research examined how teachers’ relationship reports were associated with their self-reported pedagogical practices toward a larger group of kindergartners.

Second, we tested whether these anticipated differences in reported relationship quality were mediated by teachers’ appraisals of children's social behaviors, i.e. their perceptions of children's personal and social problems. Mediation requires significant relations between the mediator and the independent and dependent variable (see Baron & Kenny, 1986). Accordingly, we tested the specific assumptions (a) that teachers would perceive more negative personal problems for inhibited and hyperactive versus average children, and more negative social problems for hyperactive versus average children, and (b) that teachers would report less favorable relationships for children whose behaviors they negatively appraised. The third main hypothesis was based on interpersonal attribution theory. We expected that the impact of perceived problems on the quality of the teacher–child relationship was moderated (i.e. increased) by teachers’ attributions of control.

Section snippets

Participants

For our study, we examined 81 teachers (77 females; Mage = 41.5 years, SD = 10.2) in relation to 237 children (120 females; Mage = 70.0 months, SD = 6.8) from regular kindergarten classes. To ensure that relationships had sufficiently developed, we questioned the teachers in the spring. In the Netherlands, kindergarten has two grades (K1 and K2) and starts on a child's fourth birthday. Less than 10% of the children were 4 years old, which implied that the large majority had at least 1 year of

Preliminary analyses

Before testing our hypotheses, we performed three preliminary analyses. First, we inspected the intercorrelations of all continuous variables (shown in the left part of Table 2). The two measures for perceived problems were positively related. Moreover, they showed similar relations to five of the six other measures: Teachers attributed less control to children with personal and/or social behavior problems, and teachers reported less close, and more dependent and conflictual relationships for

Discussion

In this study, we examined the notion that teachers’ reports of their relationships with individual students are based upon their personal assessments of the severity and causes of children's social behaviors. In doing so, we expanded upon other studies indicating that teachers’ relationship ratings should be considered as personal accounts rather than neutral descriptions of teacher–child interactions (Hamre et al., 2008, Kesner, 2000, Mashburn et al., 2006).

We examined three main hypotheses.

Acknowledgments

This research was supported by grant 41121203 from the Netherlands Organization for Scientific Research.

We thank Sylvie Poirier for proofreading the manuscript.

References (57)

  • R.M. Baron et al.

    The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations

    Journal of Personality and Social Psychology

    (1986)
  • S.H. Birch et al.

    Children's interpersonal behaviors and the teacher–child relationship

    Developmental Psychology

    (1998)
  • T.N. Bradbury et al.

    Attributions in marriage: Review and critique

    Psychological Bulletin

    (1990)
  • D.B. Bugental et al.

    Measuring parental attributions: Conceptual and methodological issues

    Journal of Family Psychology

    (1998)
  • D.F. Chang et al.

    The effects of race and problem type on teachers’ assessments of student behavior

    Journal of Consulting and Clinical Psychology

    (2003)
  • M. Chazan et al.

    Teaching five to eight year-olds

    (1987)
  • N.L. Collins et al.

    Working models of attachment and attribution processes in intimate relationships

    Personality and Social Psychology Bulletin

    (2006)
  • R.J. Coplan et al.

    Exploring and assessing nonsocial play in the preschool: The development and validation of the Preschool Play Behavior Scale

    Social Development

    (1998)
  • N.A. DeWolve et al.

    ADHD in preschool children: Parent-rated psychosocial correlates

    Developmental Medicine & Child Neurology

    (2000)
  • S. Doumen et al.

    Observed teacher–child interactions and teacher perceived relationship quality: Their concordance and associations with school engagement.

  • H. Fumoto et al.

    Teachers’ perceptions of their relationships with children who speak English as an additional language in early childhood settings

    Journal of Early Childhood Research

    (2007)
  • H. Gazelle et al.

    Anxious solitude and peer exclusion: A diathesis-stress model of internalizing trajectories in childhood

    Child Development

    (2003)
  • B.K. Hamre et al.

    Early teacher–child relationships and the trajectory of children's school outcomes through eighth grade

    Child Development

    (2001)
  • B.K. Hamre et al.

    Teachers’ perceptions of conflict with young students: Looking beyond problem behaviors

    Social Development

    (2008)
  • P.D. Hastings et al.

    Predicting mothers’ beliefs about preschool-aged children's social behavior: Evidence for maternal attitudes moderating child effects

    Child Development

    (1999)
  • S.P. Hinshaw

    On the distinction between attentional deficits/hyperactivity and conduct problems/aggression in child psychopathology

    Psychological Bulletin

    (1987)
  • I.T. Ho

    A comparison of Australian and Chinese teachers’ attributions for student problem behaviors

    Educational Psychology

    (2004)
  • C. Howes

    Social-emotional classroom climate in child care, child–teacher relationships and children's second grade peer relations

    Social Development

    (2000)
  • Cited by (43)

    • Teacher-child interactions and kindergartners' task behaviors: Observations based on interpersonal theory

      2017, Journal of Applied Developmental Psychology
      Citation Excerpt :

      Teachers rated children's behaviors on a four-point scale. Previous studies have reported high internal consistencies for the BQTSYO-M (Cronbach's alpha's ≥ 0.81 and 0.91 for Internalizing and Externalizing Behavior, respectively) and supported the validity of both scales (Thijs & Koomen, 2009; Thijs, Koomen, & van der Leij, 2008). Four covariates were included in the analyses.

    • The socio-behavioral development of children with symptoms of attachment disorder: An observational study of teacher sensitivity in special education

      2016, Research in Developmental Disabilities
      Citation Excerpt :

      The majority of studies that did employ a dyadic approach have been based on teacher perceptions of close teacher-child relationships. Although teachers are important informants of their own relationships with children, teacher reports are also subjective accounts that are susceptible to perceptual biases (Thijs and Koomen, 2009). There is therefore a strong need for observational research on teacher sensitivity as a dyadic construct to complement and expand the existing literature base on the protective role of teachers for at-risk children.

    View all citing articles on Scopus
    View full text