Handbook of psychology volume 7 educational psychology

Diversity in Early Childhood Education: Individual Exceptionality and Cultural Pluralism

307

dren’s individuality and development. Their work was soon

followed, and expanded upon, by other leading thinkers. The

topic of temperament has consequently played a central role

in the development of new interactionist theories of child de-

velopment and in the formulation of etiological theories in

child and adolescent psychiatry. It has been accompanied by

intensive discussion of basic research and clinical questions,

many of which continue to arise as children’s individuality is

investigated today. It is therefore useful to consider first the

central themes that have recurred as the concept of tempera-

ment has been put to the empirical test and as new conceptual

issues have emerged. Following this, the implications and

empirical investigation of temperament in early childhood

education and day care will be discussed.

The Definition of Temperament and Temperament Traits 

Chess and Thomas defined temperament as the behavioral

style of a person, distinguishing this aspect from the person’s

abilities and motivations (Chess & Thomas, 1987). They

have also characterized it as “a non-motivational factor in the

determination of behavioral patterns” (Chess & Thomas, in

Carey & McDevitt, 1989, p. 26). They clarified this idea by

explaining that an individual responds to an internal or exter-

nal stimulus through the meditating effects of temperament,

along with other factors, such as past events, cognitive level,

subjective feelings, and ideals. Others have characterized

temperament as a mediating variable, representing an indi-

vidual’s response patterns and manner of coping with stress

and adversity (Rutter, 1994). Carey and McDevitt (1995)

stated that although there is still no universal agreement on a

definition, the general usage is in accord with the view of

Thomas and Chess’s original concept of behavioral style.

They noted an important distinction between temperament

and cognitive factors:

Thomas and Chess explain it as the “how” of behavior as con-

trasted with the “what” (abilities or developmental level) and the

“why” (motivations and behavioral adjustment). This conceptual

and empirical separation of temperament and cognitive function

has been demonstrated at various times in childhood including

infancy (Plomin et al., 1990, 1993) and the elementary school

years (Keogh, 1986; Martin, 1989, 1989b). (p. 10)

Although new methods are now available to investigators,

particularly in molecular genetics and in direct studies of

brain function, no new findings have overturned earlier views

that temperament represents an innate attribute of a child and

arises from a combination of genetic, biological, and envi-

ronmental contributors. 

Chess and Thomas identified the nine temperament traits

listed in Table 13.1. Although the traits developed by Chess

and Thomas have formed the basis of several temperament

scales, which in turn have been utilized clinically as well as

in vast numbers of individual studies, others have developed

different formulations of the construct. Drawing on Carey

and McDevitt’s (1995) discussion of these alternatives, a

brief summary is presented in Table 13.2.

The Measurement and the Functional

Significance of Temperament

Although temperament is manifested in behavioral styles,

researchers, clinicians, and parents are aware that similar

behavior can have any number of nontemperamental causes.

It is not surprising, then, that the assessment of specific be-

havior as related to temperament and the measurement of an

individual’s temperamental profile have posed especially

difficult problems for researchers and clinicians alike. Al-

though some investigators have used psychophysiological

TABLE 13.1

The Nine Temperament Traits Identified by Chess

and Thomas

Trait

Definition

Activity level

Motor activity and the proportion of active and 

inactive periods.

Rhythmicity

The predictability or unpredictability of the 

timing in biological functions, such as 

hunger, sleep-wake cycle, and bowel 

elimination.

Approach/withdrawal

The nature of the initial response to a new 

situation or stimulus—a new food, toy, 

person, or place.

Adaptability

Long-term responses to new or altered 

situations. Here the concern is not the nature 

of the initial responses but the ease with 

which they are modified in desired 

directions.

Sensory threshold

The intensity level of stimulation necessary to 

evoke a discernible response, irrespective of 

the specific form the response may take.

Quality of mood

The amount of pleasant, friendly, and joyful 

behavior and mood expression, as contrasted 

with unpleasant crying and unfriendly 

behavior and mood expression.

Intensity of reactions

The energy level of response, positive or 

negative.

Distractibility

The effectiveness of an outside stimulus in 

interfering or changing the direction of the 

child’s ongoing behavior.

Persistence and 

These two categories are usually related.

attention span

Persistence refers to the continuation of an 

activity in the face of obstacle or difficulties. 

Attention span concerns the length of time a 

particular activity is pursued without 

interruption.

Source: Based on Chess & Thomas (1984, pp. 42– 43).

308

Early Childhood Education

instruments to support their claims about biological factors,

most contemporary research has been conducted using paper-

and-pencil checklists of children’s behavioral tendencies as

observed by parents, teachers, investigators, and clinicians.

These scales are usually based on the nine dimensions devel-

oped by Chess and Thomas. These instruments are designed

to tap children’s behavioral dispositions by asking parents and

other observers questions about the child’s typical behavior in

day-to-day situations. These scales, and modifications of

them, have been validated on large numbers of children inter-

nationally (Carey & McDevitt, 1995). Currently, several

scales are available that are appropriate for use by ECE per-

sonnel, often in combination with parent ratings and observa-

tions (McDevitt, in Andersen & McDevitt, 2000). These

include a new questionnaire for caregivers and preschool

teachers, the Teacher and Caregiver Temperament Inventory

for Children (TACTIC), which measures temperament, atten-

tion, emotions, and conduct in 2- to 7-year-old children, and

the Basic Behavioral Assessment Scale (BBAS) by Carey and

McDevitt, which measures behavioral adjustment in 4- to 14-

year-old children in the areas of behavior, achievement, self-

relations, internal state, and coping (S. C. McDevitt, personal

communication, November 2001). 

Although there appears to be a growing consensus on

many issues raised in the earlier years, more recent discus-

sions of temperament continue to reflect earlier concerns

about measurement. In response to Clarke-Stewart, Fitz-

patrick, Allhausen & Goldberg (2000) presentation of a short

and easy measure of infant temperament, Carey (2000) has

argued for caution when hoping that there could be a brief

and simple way to capture a complex phenomenon.

Although the work of Chess and Thomas and their follow-

ers was published as early as the mid-1960s, the findings on

the predictive validity of children’s individual differences in

emotional disposition and behavioral style did not win accep-

tance until the early 1980s. This was in spite of the fact that

several independent lines of research had essentially repli-

cated the early findings (Graham, Rutter, & George, 1973;

Barron & Earls, 1984; Maziade, Caron, Cote, Boutin, &

Thivierige, 1990). Fears of a deterministic view of human

behavior may have played a role in the reluctance of devel-

opmentalists to accept the notion that biologically based dif-

ferences exist in children’s emotional and behavioral

dispositions. Carey and McDevitt (1995) suggested that in

clinical circles another central reason for the slow acceptance

of temperament concepts may lie in the fact that the risks

associated with temperament are seen to lie not in the

temperament itself but in the lack of goodness of fit between

the child’s temperament and the expectations and values

of the environment. Clinicians may have found it difficult to

consider the question of fit because this notion is not in keep-

ing with the habit of looking for problems within either the

child or the environment. However, such a contextual ap-

proach has become not only acceptable, but fully in keeping

with current theoretical understandings of the interplay be-

tween children’s dispositions, abilities, and interests and the

complex factors that interact with their individuality at all

levels of the ecology. 

Despite the early resistance to the concept and despite the

clinical world’s slowness to adopt complex etiological mod-

els, accumulating evidence has left little room for doubt about

the reality and importance of temperament. Research in tem-

perament has now involved the efforts of hundreds of scien-

tists in numerous areas, and the second wave of temperament

research has adopted stringent methodological standards and

extended its reach across cultures and into the genome, utiliz-

ing research methods ranging from ethnographic studies to

the most recent technology such as brain scans and electro-

physiological measurement. Clinician-researchers have been

appraising the role of temperament across culture and socio-

economic differences, exploring the interaction between tem-

perament and physical status, and behavioral-geneticists and

neuroscientists continue to study the relative contribution of

genes and the environment, including the environment of the

womb. Twin and adoption studies have revealed that there is a

substantial genetic contribution to temperament—about 50%

on average. (Contrary to previous views, there is compelling

evidence that temperament in infancy is not highly under

genetic control and that genetic influences are stronger in the

postinfancy years. In addition, temperament is not as stable or

continuous as was once proposed.) Investigation of the role of

maternal hormones, stress, prenatal infections, and perinatal

stress has reminded developmentalists that environmental

contributors to temperament can be biological (Carey &

TABLE 13.2

Six Alternative Definitions of Temperament

Theorist

Definition of Temperament

Buss and Plomin

The EASI formulation: The four traits of 

emotionality, activity, sociability, and 

impulsivity, although impulsivity was later 

withdrawn when it was not found to be 

heritable.

Eysenck

Features of personality: Extraversion-

introversion, neuroticism, and 

psychoticism.

Goldsmith and Campos

Temperament is limited to the emotional 

sphere.

Rothbart and Derryberry

Dimensions of reactivity and self-regulation.

Strelau

Regulative theory of temperament, including 

components of energy and temporal traits.

Zuckerman’s

Description of sensation seeking.

Source: Based on Carey & McDevitt (1995, pp. 15–18).

Diversity in Early Childhood Education: Individual Exceptionality and Cultural Pluralism

309

McDevitt, 1995, pp. 18–24). At the same time, interactions of

parenting style, family stress factors, and social pressures

continue to be seen as important contributors to outcome.

In the past two decades temperament has been shown to

affect a wide range of areas of children’s functioning, no less

so in the preschool and school years than in infancy. Current

knowledge about the functional significance of temperament

was recently summarized by Carey (1998, pp. 27) as cited by

Andersen and McDevitt (2000) as

fundamental part of the parent-child relationship;

significant factor in patterns of growth and feeding;

predisposition to prolonged crying in infancy;

possible contributor to sleep problems;

reason some children are hard to discipline;

major risk factor for social behavior problems;

substantial component of school performance;

factor in physical conditions;

one factor in recurrent pain;

partial determinant of response to crises.

Although temperament has been shown to play an impor-

tant role in many areas of development, perhaps the most sig-

nificant area for application in early childhood education is

the role of temperament in children’s psychological adjust-

ment and educational achievement. It is useful, therefore, to

review the major findings about this relationship. One of the

most important and replicated findings of the early tempera-

ment research was the identification of specific clusters of

traits that differentiated parents’ experiences with children.

Forty percent of the sample of children who were rated by

their parents as adaptable, approaching, mildly active, mildly

intense, and regular in biological rhythms were described by

their middle-class North American parents as “easy.” These

children fitted well with the demands of family life and cul-

tural expectations and were found to be at low risk for (but

not immune to) the development of psychological and educa-

tional problems (Chess & Thomas, 1987). Ten percent of the

sample were, by contrast, described as “difficult.” These chil-

dren were low in adaptability, predominantly negative in

mood, irregular, withdrawn in new situations, and intense in

reactivity. A third group of children, comprising 15% of the

sample, were mildly intense in response, mildly negative in

mood, and slow to adapt to new situations. This group was

described as “slow-to-warm-up.”

The most important finding of the NYLS was the discov-

ery that “difficult” children were at significantly higher risk

of developing problems in behavioral and emotional adjust-

ment. However, adjustment problems were not inevitable.

Chess and Thomas invoked the concept of goodness of fit to

describe a compatible relationship between parental expecta-

tions and a child’s temperament; when the fit was poor, the

child experienced excessive stress and developed reactive be-

havioral and emotional problems. The finding that tempera-

mentally difficult children were at higher risk of developing

psychological problems has been replicated in subsequent re-

search (Earls & Jung, 1987; Graham et al., 1973; Maziade,

Cote, Bernier, Boutin, & Thivierge, 1989; Maziade et al.,

1990). However, fit has been shown to be context dependent.

Consequently, Carey and McDevitt (1989) proposed the term

temperament risk factors:

Temperament risk factors are any temperament characteristic

predisposing a child to a poor fit (incompatible relationship)

with his or her environment, to excessive interactional stress and

conflict with his or her caretakers, and to secondary problems in

the child’s physical health, development and behavior. These

characteristics are usually perceived as hard to manage, but may

not be. The outcome depends on the strength and durability of

the characteristics and the environmental stresses and supports.

(p. 195)

Temperament Research in Early Childhood Education 

Despite the consensus that now exists about the importance

of the role of children’s individual differences in tempera-

ment as a significant risk factor in specific contexts, applica-

tions of this knowledge in all levels of early childhood

education have been few (Andersen, 1990). In 1995 Carey

and McDevitt published an extensive review of the extant

studies of temperament. In their chapter on day care, they de-

cried the failure of the prevention, intervention, and caregiv-

ing communities to apply this important knowledge. In their

chapters on infants and young children, they noted that

“behavioral investigation in the area [of child day care] has

almost completely ignored the impact of children’s tempera-

ment on the experience” (p. 83). They summarized the little

research there was at the time. However, what little research

there is suggests that further investigation is likely to be fruit-

ful (Anderson-Goetz & Worobey, 1984; Field & Greenberg,

1982; Keogh & Burstein, 1988; Palison, 1986). 

Among the most important preliminary findings cited by

Carey and McDevitt (1995) was the (counterintuitive to

some) discovery that mothers of children with difficult tem-

peraments were less likely to go back to work and to place

their children in day care, although the authors commented

that this study was conducted in the late 1950s and early

1960s, when “the pressure to seek employment outside the

home might not have been felt quite as strongly by middle-

class mothers as it is today” (p. 85). More recently, Canadian

310

Early Childhood Education

researchers McKim, Cramer, Stuart, and O’Connor (1999)

examined the family and child factors associated with child

care decisions and found that mothers who preferred to stay

home were more depressed and that their children were more

likely to experience unstable care than were those who were

working and wanted to work. The age at which the child en-

tered day care and aspects of the care such as quality were not

related to attachment measures. However, infants with diffi-

cult temperaments were at higher risk of being rated as inse-

curely attached. Attending day care appeared to reduce the

strength of this relationship. This study, though preliminary,

lends some support to the notion that enrolling a difficult in-

fant in day care might have beneficial effects for the infant

and possibly also for depressed mothers.

Another important question is how children’s individual

temperaments affect their socioemotional adjustment to, and

functioning in, early childhood group environments. In an

early study of 2.5- to 3.5-year-old children entering day care,

Billman and McDevitt (1980) obtained mothers’ and ob-

servers’ temperament ratings on a sample of 78 predomi-

nantly White middle-class children ranging in age from 34 to

64 months and attending two nursery schools in a small

Midwestern American community. There were 40 girls and

38 boys in the study. The researchers set out to clarify the

relationship between home-rated temperament and social be-

havior at nursery school, to investigate any relationship be-

tween difficult temperament and peer interaction, and to

assess the consistency of temperament ratings made by dif-

ferent raters in different settings. Parental ratings of the tem-

perament characteristic of low approach predicted teacher

evaluations of slower adjustment. It is interesting to note that

the “difficult child” cluster of traits emerged as a significant

predictor of peer interactions. Although convergence be-

tween the two temperament ratings was moderate (.18–.46),

correlations were statistically significant for all dimensions

except mood. Activity level, approach-withdrawal, and sen-

sory threshold were significantly related to peer interaction.

Very active children were both more sociable than inactive

children and more often involved in conflictual interactions.

More rhythmic children spoke more to peers and got on

with their tasks more effectively. When the children were as-

signed to the temperament clusters identified by Chess and

Thomas, the difficult child was found to engage in more

wrestling, hitting, jumping, pushing, and beating.

More recently, Harden et al. (2000) used Bronfenbrenner’s

ecological framework to explore externalizing behavior prob-

lems among children enrolled in a suburban Head Start

program and compared them with a subgroup of children

with behavior problems in the clinical or borderline range.

Children’s externalizing behavior was positively associated

with internalizing behavior, parent psychological adjustment,

child temperament, family environment, and exposure to

community violence.

Yen and Ispa (2000) tested the hypothesis that curriculum

type (Montessori and constructivist) moderates the impact of

temperament (specifically, activity level and attention span

and persistence) on the classroom behavior of 3- to 5-year-

old children. A near-significant trend suggested that tempera-

mentally active boys were more likely to be perceived by

their teachers as having behavior problems if they were en-

rolled in Montessori programs than if they were enrolled in

constructivist programs. An interesting finding was that at-

tention span and persistence did not have any effect on the

impact of the type of curriculum on children’s behavior. This

is one of the few studies that explores the important question

of the goodness of fit between specific types of curriculum

and children’s temperament.

Stansbury and Harris (2000) conducted a study to ascer-

tain whether standardized peer entry paradigm would pro-

duce stress responses in 38 3-year-olds and 25 4-year-olds

and how such stress responses might correlate with tempera-

ment, approaches to peers, and peer competence as perceived

by the children themselves. Four-year-olds were significantly

less avoidant and were rated higher on the temperament trait

of “approach.” They showed larger Hypothalamic-Pituitary-

Adrenal Cortical Axis (HPA) stress responses to new peers,

and the disparity between self-reported peer competence and

behavior in the peer-entry situation was associated with

greater stress responses on a physiological measure. The re-

searchers stressed the importance of examining discrepancies

between self-perception and action in research on stress.

The need for the education of early childhood profession-

als about temperament was acknowledged in a study by

Franyo and Hyson (1999) of the effectiveness of tempera-

ment training of early childhood caregivers. This study pro-

vided information about caregivers’ preexisting knowledge

of temperament concepts and investigated the effect of edu-

cational workshops about temperament concepts. The find-

ings were that without training, caregivers had heard about

temperament but knew little about the specifics of the empir-

ical findings. Without training, caregivers did not appear to

have many ideas about how to achieve a goodness of fit by

using behavioral management techniques, although this was

an area in which they showed special interest. Encouragingly,

caregivers were very accepting of the concept of tempera-

ment, and training sessions were effective in increasing their

knowledge about temperament concepts. However, the in-

vestigators reported that there was no statistically significant

evidence that the training was effective in improving the

caregivers’ acceptance of children’s behaviors and feelings.