Handbook of psychology volume 7 educational psychology

218

Relationships Between Teachers and Children

Q-set methods utilize criterion sorts in order to derive

scores for subjects on relevant constructs. Experts knowl-

edgeable regarding a particular construct (e.g., child-teacher

security) are asked to sort Q descriptions according to their

view of what an ideal child would receive as a sort, in the

case of the aforementioned example, a child who was ideally

secure in relationship to the teacher.

In several studies (e.g., Howes, Matheson, et al., 1994;

Pianta et al., 1997) criterion sorts were developed to describe

children in specific types of relationships with teachers—for

example, a child in a secure relationship, a child in a con-

flicted relationship, and a child in a dependent relationship.

Interrater reliability for these sorts tends to be high, sug-

gesting that two observers’ impressions of characteristic be-

haviors can agree for individual children. Also, correlations

between sorts of teacher-child relationships are strongly re-

lated to similar sorts of parent-child relationships for con-

structs such as security, indicating a high level of consensus

among various reporters on the behaviors reflective of a

secure child-adult relationship.

Summary

Teachers’ reports of child-teacher relationships reflect di-

mensions of conflict and closeness. These dimensions repli-

cate across samples that vary by age, ethnicity, and

economic status (see Pianta, 1999), are fairly stable, and

correlate with concurrent and future teacher-reported mea-

sures of adjustment, school achievement, and student moti-

vation (Birch & Ladd, 1997; Howes, 2000b; Pianta et al.,

1995). Children’s reports of relationships with teachers re-

flect dimensions of emotional closeness and support, com-

munication and involvement, and negativity, suggesting

parallels with teachers’ reports (Bracken & Crain, 1994;

Ryan et al., 1994; Wentzel, 1996). It appears that in relation

to student or teacher outcomes, negativity is the most salient

feature of teachers’ reports, whereas a sense of closeness

and support appears most salient from the child’s point of

view.

It is critical to note that these conclusions are qualified by

the fact that the literature is limited in terms of multimethod,

multi-informant longitudinal studies. Missing from this liter-

ature is description of the same child-teacher relationship

from its two participants, as well as the extent to which use of

these two perceptions of the same relationship yields dimen-

sions similar to those reported earlier for single-participant

reports and whether two participants’ perceptions converge

or are concordant with one another.

CORRELATES OF RELATIONAL DIMENSIONS

In this section we review studies that link the aforementioned

child-teacher relationship dimensions to child outcomes and

other correlates. Studies cited in the review assess child-

teacher relationships at the relational unit of analysis. These

findings attempt to address the extent to which a focus on this

unit of analysis is helpful in advancing understanding of

development in school settings.

Over the last 10 years research on child-teacher relation-

ships focused around several lines of inquiry, each resulting in

support for these relationships as salient features of develop-

ment. These lines of inquiry involve child-teacher relationships

and peer relations, parent-child relationships, academic com-

petence, and features of social and emotional adjustment (see

Pianta, 1999). Teacher-child relationships are related to chil-

dren’s competencies with peers in the classroom (e.g., Birch &

Ladd, 1998; Howes, 2000b; Howes, Hamilton, et al., 1994) and

trajectories toward academic success or failure (Birch & Ladd,

1996; Hamre & Pianta, 2001; Pianta et al., 1995; van

IJzendoorn, Sagi, & Lambermon, 1992), as well as with pat-

terns of child-mother relationships (Lynch & Cicchetti, 1992)

and disruptive behavior (Hamre & Pianta, 2001).

Howes and colleagues (see Howes, 2000a, 2000b)

conducted a series of studies relating child-parent and child-

teacher relationships to each other and to early childhood so-

cial outcomes (Hamilton & Howes, 1992; Howes, Hamilton,

et al., 1994; Howes et al., 1998; Howes, Matheson, et al.,

1994). They established a low to moderate degree of continu-

ity in the quality of relationships that children have with moth-

ers and form with teachers (Howes & Matheson, 1992). They

further found that both of these relationships play a role in

children’s peer competencies, although relationships with

teachers are stronger predictors of behavior with peers in

the classroom than are relationships with parents (Howes,

Matheson, et al., 1994). Also, child-teacher relationships show

low to moderate levels of continuity in the early grades of

school—at least through second grade (Howes, Phillipsen, &

Peisner-Feinberg, 2000), echoing Birch and Ladd’s (1998)

contention that children’s relationships demonstrate a coher-

ence across relational figures and across time.

Pianta and colleagues reported links between teachers’ re-

ports of relationships with children and a range of school

outcomes in the early grades. In one such study (Pianta et al.,

1995), kindergarten teachers’ reports of the degree to which

children displayed security toward them was related to first-

grade teachers’ reports of the children’s competence. In a se-

ries of descriptive studies, Pianta and Steinberg (1992) and

Pianta (1994) showed that teacher-child relationships are also

Correlates of Relational Dimensions

219

fairly stable across the period from kindergarten to second

grade and correlate with concurrent and future teacher-

reported measures of adjustment, grade retention, and special

education referrals (Birch & Ladd, 1997; Pianta et al., 1995).

Furthermore, changes in student adjustment from year to year

were correlated in expected directions with these dimensions

(Pianta et al., 1995): Downward deflections are correlated

with child-teacher conflict, whereas upward deflections are

related to child-teacher closeness. Finally, there is evidence

that child-teacher relationships operate as a protective factor

against risk: Children at high risk for retention or referral for

special education who are not referred or retained are

reported to be more close to their teachers, whereas their

retained/referred counterparts are in greater conflict with

teachers (Pianta et al., 1995).

Within a group of children designated on the basis of low

kindergarten screening scores as high risk for referral for

special education or retention, those who ultimately did get

retained or referred were compared with those who, despite

being high risk, were promoted or not referred (Pianta et al.,

1995). The children who, despite predictions of retention or

referral, were ultimately promoted or not referred had far

more positive relationships with their teachers than did their

high-risk peers who were retained or referred. Significantly,

this successful high-risk group was notable for its lack of

conflict and high degree of open communication. In short, it

appeared that there was a buffering effect of the relationship

between the child and teacher (Pianta et al., 1995).

Hamre and Pianta (2001) extended analysis of the longitu-

dinal relations between early child-teacher relationships

(in kindergarten) and child school outcomes through eighth

grade. Controlling for kindergarten-entry cognitive ability

and problem behavior, negativity in the child-teacher rela-

tionship reported by the child’s kindergarten teacher pre-

dicted achievement test scores, disciplinary infractions, and

school suspensions through either grade. The effects on

eighth-grade achievement scores appeared largely mediated

by effects of the kindergarten child-teacher relationship on

achievement in early elementary school. Furthermore, effects

on disciplinary infractions were most pronounced for chil-

dren who had problems in kindergarten adjustment. This was

the first study to report longitudinal findings for early child-

teacher relationships extending into middle and junior high

school, and in addition the study supports the conclusion of

other investigations that the quality of these relationships

appears particularly important for children who might other-

wise have adjustment problems.

The work of several other investigators also supports the

child-teacher relationship as a key context in which early

school outcomes are developed. Van IJzendoorn et al. (1992)

demonstrated that child-caregiver security added unique vari-

ance over and above that contributed by the child-mother

relationship in the prediction of a range of developmental

status and school readiness variables. Studies have also used

children’s reports of their relationship with teachers, with

findings similar to those using teacher perceptions. Wentzel

(1996) reported that middle school students benefited from

relationships with teachers characterized by open communi-

cation and a sense of closeness, suggesting that this is a rela-

tional context with salience for children beyond the early

grades and preschool years. Similarly, Lynch and Cicchetti

(1992) established that maltreated children, as a result of

experiences with parents, are sensitized to seek certain rela-

tional experiences with teachers; they are less likely to form

optimal relational patterns and seek psychological proximity

and support from teachers.

Birch and Ladd (1996) studied teacher-child relationships

extensively in early elementary classrooms and suggested that

children have a generalized interpersonal style (moving to-

ward, moving against, moving away) that characterizes their

interactions with peers and teachers. Presumably, this style is a

product of interactions with parents. This relational style of the

child is related in predictable ways to the quality of relation-

ships that children form with teachers and peers in the class-

room (Birch & Ladd, 1998; Ladd & Burgess, 1999). Those

children who displayed moving against behaviors in kinder-

garten, such as verbal and physical aggression toward teacher

and peers, were more likely to form negative relationships with

teachers in first and second grade (Ladd & Burgess, 1999).

Children who tended to move away from others in kinder-

garten were more likely to be rated as overly dependent by

first-grade teachers, although there was less stability in these

behaviors than in aggressive behaviors (Birch & Ladd, 1998).

Observed conflict in the child-teacher relationship is related to

less classroom participation and lower achievement over the

first half of kindergarten (Ladd et al., 1999). Children’s moving

toward, or prosocial, behavior in kindergarten was not related

to aspects of children’s relationships with first-grade teachers.

However, kindergarten teachers’ reports of the quality of rela-

tionships with students accounted for significant variance in

children’s social behavior in first grade after controlling for

gender and kindergarten social behavior (Birch & Ladd, 1998).

Research on teachers’ and classmates’ effects on adoles-

cents’ motivation, self-esteem, and ability to express their

opinion are reviewed in Harter (1996). Harter discussed how

relationships with teachers change from elementary to junior

high school (relationships between teachers and students be-

come less personal, more formal, more evaluative, and more

220

Relationships Between Teachers and Children

competitive). These changes can lead to more negative self-

evaluations and attitudes toward learning because the imper-

sonal and evaluative nature of the relational context in junior

high does not match well with the children’s relational needs

at that age. Harter (1996) found that this model applies

particularly to students who have lower levels of intrinsic moti-

vation. In this way, teacher-child relationships (which are typi-

cally viewed as potential resources for amelioration of risk) can

actually exacerbate risk if they either are not positive or do not

match with the developmental needs of the child. Harter (1996)

also reported that classmate support and teacher approval are

associated with self-esteem in middle-school-aged popula-

tions. Teacher support can be particularly salient in students

who have low levels of parent support (i.e., teacher and parent

support may have additive effects on student self-esteem).

Consistent with this view of middle schoolers’ ongoing

needs for support from adult figures, teacher support has been

found to be related to sixth-grade children’s school- and

class-related interests and to their pursuit of social goals

(Wentzel, 1998). These self-beliefs and motivations in sixth

grade in turn predicted pursuit of social goals and grades in

seventh grade (Wentzel, 1998). It is important to note that the

support that youth receive from their parents, peers, and

teachers seemed to have additive, and thus fairly indepen-

dent, effects. Support from teachers was uniquely related to

classroom functioning (Wentzel, 1998). Wentzel (1998) sug-

gested the possibility that support in teacher-child relation-

ships may be particularly salient at transition points, such as

the transition from elementary to middle school.

In young children (kindergarten, first, and second grade),

a teacher’s feedback about a child’s behavior also has a sig-

nificant impact on how peers perceive that child (Hughes,

Cavell, & Willson, in press; White & Kistner, 1992). When a

teacher characterizes a child’s behavior as positive, other

children report increased preferences for that child and are

more likely to characterize the child’s behavior as positive.

When a teacher characterizes a child’s behavior as negative,

an impact on peer preferences was not found; but if the

teacher was derogatory toward the child, peers demonstrated

more negative views of that child (White & Kistner, 1992).

Implications of these findings include the possibility that

teachers can play an active role in changing peers’ percep-

tions of rejected children by sensitizing the class to the po-

sitive behaviors that the child engages in (Hughes et al., in

press; White & Kistner, 1992).

Summary

There is ample evidence to demonstrate that the qualities of

child-teacher relationships are related in expected ways to child

outcomes throughout the school-age years. Although there are

expected developmental transformations in the extent to which

these qualities are manifest in highly proximal or concrete

forms with age, the degree and form of child-teacher engage-

ment or involvement and the affective quality of that involve-

ment describe a wide range of variation in individual and group

differences in child-teacher relationships. Variable-focused or

individual-focused analyses of these dimensions consistently

show that various parameterizations of these two dimensions

relate to children’s engagement in learning, motivation and

self-esteem, attitudes and engagement with the goals of school,

and behavior toward one another and the teacher. As character-

ized by Hamre and Pianta (2001), these findings reveal that the

quality of child-teacher relationships is an indicator of the

extent to which the child is benefiting from the resources of

schooling. This general conclusion is consistent with the theo-

retical framework of developmental systems theory outlined

earlier, in which the qualities of child-adult relationships are

key developmental resources for children.

It is critical to emphasize that in several of the investiga-

tions described earlier (e.g., Birch & Ladd, 1998; Hamre &

Pianta, 2001; Howes, 2000b; Stuhlman & Pianta, in press), re-

lations were reported between child outcomes and qualities of

the child-teacher relationship controlling for aspects of child

behavior considered principle predictors of the outcomes as-

sessed. For example, Hamre and Pianta (2001) controlled for

kindergarten teachers’ reports of children’s problem behavior

when predicting problem behavior outcomes in later elemen-

tary and middle school using kindergarten child-teacher rela-

tional negativity as a predictor. Similarly, Stuhlman and

Pianta (in press) controlled for observed child competence

when examining relations between teachers’ representations

and observed sensitivity. Relational dimensions provide

unique prediction of child outcomes independent of attributes

of the child (e.g., Birch & Ladd, 1998) and teacher (Stuhlman

& Pianta, in press). This focus on this relational unit of analy-

sis, rather than on discrete characteristics of the individuals

themselves, provides considerably more conceptual power

for the purposes of understanding behaviors in settings and

the influence that such settings have on developmental

processes.

EDUCATIONAL AND PSYCHOLOGICAL

APPLICATIONS RELATED

TO CHILD-TEACHER RELATIONSHIPS

Evidence that qualities of child-teacher relationships pre-

dict child outcomes and are related to features of school cli-

mate, teacher characteristics, child attributes, and classroom

Educational and Psychological Applications Related to Child-Teacher Relationships

221

variables provides ample support for examining how this

information can be used to create more developmentally

supportive school environments (see Battistich et al., 1997;

Hughes & Cavell 1999; Pianta, 1999). Consistent with the

prevention-oriented bias in applications informed by develop-

mental systems theory (Magnusson & Stattin, 1998), a com-

prehensive approach to prevention suggests that applications

should have long-term implementation and be aimed at

changing institutions as well as people (Weissberg & Bell,

1997). More specifically, Weissberg and Bell (1997) outlined

four different foci for application of techniques or resources,

including a focus on changing the child, changing the imme-

diate environment, changing multiple components of the

environment that affect adults who are working with children,

and changing structure or policy, each of which intersects with

the research base available on child-teacher relationships.

Issues in Prevention-Oriented Applications Involving

Child-Teacher Relationships

In thinking about applications of knowledge about child-

teacher relationships across the many levels of organization

and processes in schools, we approach the task with a bias to-

ward the deployment of resources (or applications of tech-

niques) prior to emergence of problems, with the distinct goal

of enhancing wellness and strengthening developmental

competencies (Cowen, 1999). Several principles inform this

analyses: an emphasis on application in context, the extent to

which an application embraces conceptualizations of devel-

opmental pathways in its design and execution (Loeber,

1990), emphasis on standardized protocols and theoretically

driven decision making, and focus on risk reduction or well-

ness promotion.

Applications in Context

Intervention in educational and psychological processes with

children and teachers most often involves rearranging con-

textual inputs to achieve a desired outcome (Nastasi, 1998).

Interventions applied in the contexts in which the concern

arises and is manifest can be more effective agents of change

than efforts at change that take place in a context remote

to the problem at hand (e.g., Henggeler, 1994; Henggeler,

Schoenwald, Borduin, Rowland, & Cunningham, 1998;

Nastasi, 1998). The design of treatment plans for child and

adolescent problem behavior ideally recognizes distributed

competence and the related concept of contextual affordance

and produces change as a function of manipulating contextual

properties (Adelman, 1996; Henggeler et al., 1998; Roberts,

1996). Unfortunately, due to the inherent asymmetries in

child and adult relationships, it is usually the case that prob-

lem identification and remediation focus on the child as a

locus of the “problem” (Adelman, 1996; Henggeler et al.,

1998; Johnson, Malone, & Hightower, 1996; Nastasi, 1998).

A specific, dedicated focus on the relational unit of analysis

inherent in child-teacher relationships supports a view of

bidirectionality and reciprocity, which can enhance the extent

to which contextualized, comprehensive approaches to inter-

vention can be designed.

Developmental Time: Pathways

One lesson learned from developmental research is that there

is no single, linear, one-to-one mapping of early risk (or non-

risk) status onto problem (or competent) outcomes. Instead,

many possible outcomes are possible from a given starting

point (Egeland, Pianta, & O’Gawa, 1996). The success of

risk reduction and competence enhancement efforts depends

on understanding the processes that shape developmental

pathways. Targeting these processes for intervention could be

key to interrupting the relation between risk and later prob-

lems (Loeber, 1990) by creating alternate routes along devel-

opmental pathways to positive outcomes.

Intentional efforts to reduce risk and enhance competence

through the application of psychological and educational

interventions vary widely in the extent to which they embody

principles of developmental pathways and a longitudinal

focus in their design and execution (Durlak & Wells, 1997;

McConaughy, Kay, & Fitzgerald, 1999). It is in fact the norm

that resources are deployed to children in short-term bursts of

six-week groups, semester-long mentoring, or placements

that last a school year (Durlak & Wells, 1997). Most efforts

are short-term in focus, with pressure increasing to deliver

positive effects in shorter and shorter time frames. In addition

to the problem of a short-term focus, interventions rarely

conceptualize their effects or efforts in terms of developmen-

tal pathways that link subordinate outcomes or processes to

the goals that the interventions embrace. For example, a large

number of children are enrolled in programs designed to

reduce antisocial behavior and teach social skills (Durlak

& Wells, 1997). It is a laudable goal to accomplish such

significant developmental changes in patterns of maladapta-

tion and skill deficit in (usually) the short time frame of

6 weeks, particularly for children for whom these have a

long-standing history and status as concerns. However, rather

than conceptualizing intervention success as a return to

health or normative functioning, interventionists might ex-

amine intermediate or subordinate outcomes or processes

that signal developmental change in increments that, al-

though smaller than ultimately desired, may be more realistic