Matching parenting

to child
What role do child temperament and parenting style
play in the development of later behavioural and
emotional problems? Recent findings from the Social
Development Project suggest that both factors are
important influences on child adjustment, and that
the “best” style of parenting may differ for children
with different temperaments.

SHERYL HEMPHILL AND ANN SANSON

hildhood behavioural problems have

C been defined in many different ways, but

in the context of this article, they refer to
“acting-out” behaviours such as aggres-
sive behaviour and noncompliance. The
prevalence of children’s behavioural problems is
concerningly high, with rates of about 15-20 per
cent being recorded (Prior et al. 2000; Zubrick et al.
2000). Rates of problems seem to vary by gender,
particularly in middle childhood to adolescence,
with boys exhibiting more behavioural problems INFLUENCES ON EA
than girls. Once entrenched, behavioural problems
are likely to continue through childhood and ado-
lescence and into adulthood. For example, Child temperament
aggressive behaviour in childhood is associated
with adolescent delinquency and criminality (Loe- Child temperament refers to constitutionally-based
ber et al. 1991; Prior et al. 2000). individual differences in behavioral style that are vis-
Behavioural problems have been notoriously dif- ible from early childhood (Rothbart and Bates 1998).
ficult to treat and there is an increasing emphasis Three broad aspects of temperament are gaining
on prevention (that is, intervening before problems wide acceptance. Negative reactivity refers to high-
are detected) and early intervention (that is, inter- intensity negative reactions such as irritability,
vening at early signs of difficulties) to avoid the whining and whingeing. Approach/inhibition
development and entrenchment of such problems. describes the tendency of a child to approach novel
In order to identify children at risk of difficulties, situations and people, or conversely to withdraw and
and to intervene effectively, an understanding of be wary. Persistence refers to the ability to stick at
the factors influencing the development of chil- one task for some time and sustain organised play.
dren’s behavioural problems is vital. Two of the The aspects of child temperament that seem to be
main factors implicated in the development of risk factors for the development of childhood behav-
behavioural problems are child temperament and ioural problems are high negative reactivity and low
parenting. Most research has investigated the influ- persistence (Azar 1995; Prior et al 2000; Rothbart
ence of child temperament and parenting and Bates 1998). Inhibition (a tendency to withdraw
separately but there is increasing evidence that the from novelty) may reduce the likelihood of the devel-
interaction between the two may be particularly opment of childhood aggressive behaviour (Rothbart
important. In the section that follows, these key and Bates 1998). In the study reported here the focus
constructs are defined and their role in children’s is on the influence of negative reactivity and inhibi-
development outlined. tion on the development of behavioural problems.

42 Family Matters No.59 Winter 2001 Australian Institute of Family Studies

temperament

RLY CHILDHOOD BEHAVIOURAL PROBLEMS

Parenting Inductive reasoning refers to parents explaining
consequences of misbehaviour to a child, setting
Three key aspects of parenting relevant to child limits on behaviour, and allowing child input into
behavioural problems are punishment, warmth, disciplinary decisions. High levels of inductive rea-
and inductive reasoning. soning also have been associated with children’s
Punishment refers to the use of harsh, high- positive social adjustment (Chamberlain and Pat-
intensity disciplinary strategies, with the parent terson 1995; Hart et al. 1992), whereas low levels
attempting to assert power over the child through might be expected to predict behavioural problems
direct commands and threats, or physical punish- such as aggressive behaviour and noncompliance.
ment. This style of parenting has been associated
with later child aggressive behaviour and noncom- Match between temperament and parenting
pliance (Booth et al. 1994; Patterson et al. 1989; Research has shown that both a child’s temperament
Pettit and Bates 1989). and the parenting s/he receives are important in the
Parental warmth includes the expression of ver- development of behavioural problems. However, little
bal and physical affection towards the child, as well research has examined whether a particular parent-
as praise and acceptance. Low parental warmth ing style has the same effect on all children, or
(criticism, disapproval and rejection of the child) whether it differs according to the child’s tempera-
has also been associated with childhood aggressive ment. The concept of “goodness of fit” (Thomas et al.
behaviour and noncompliance (Booth et al. 1994; 1968) suggests, for example, that a child with a
Chamberlain and Patterson 1995; Patterson et al. negative, reactive temperament whose parent uses
1989). In contrast, high warmth has been linked to harsh punishment may be more likely to develop
positive adjustment (Baumrind 1966; Booth et al. behavioural problems than the same child with a par-
1994). ent who uses inductive reasoning. For example,

Australian Institute of Family Studies Family Matters No.59 Winter 2001 43

 Crockenberg (1987) found that irritable infants with study looks intensively at a smaller sample, with the
angry and punitive adolescent mothers were more aim of elucidating developmental processes and pro-
likely than less irritable infants to be angry and non- viding richer data on the children themselves, as well
compliant at two years of age. And in a study of five to as their interactions with their parents and peers.
six year-olds, Paterson and Sanson (1999) found The Social Development Project aims to investi-
that temperamental inflexibility (including negative gate the role of temperament, parenting and social
reactivity) and punitive parenting predicted parent- context in the development of behavioural and emo-
reported behavioural problems. tional problems (for example, social withdrawal from
peers). A total of 112 two-year-old children (58 boys
and 54 girls) and their primary caregivers (110
mothers, 2 fathers) were recruited from the general
community through child care centres, maternal and
child health centres, kindergartens, playgroups, and
responses to advertisements in local newspapers.
At the time of enrolment into the study, the aver-
age age of the children was two years and two
months, and the average age of the participating
caregiver was 35 years. Participants’ scores on meas-
ures of socioeconomic status, parental daily hassles,
stressful family life events, availability of social sup-
port to parents, and other measures of family stress
indicated that in general the participants were well
functioning. The children and their caregivers were
followed up again at four years of age.
Multiple methods of data collection have been
used, including laboratory observations conducted
at a playroom (the Social Development Laboratory)
at the University of Melbourne, and parent-com-
pleted questionnaires. The laboratory visits were
designed to challenge the child by presenting a
Three key aspects of parenting relevant to child behavioural range of novel situations.
problems are punishment, warmth, and inductive reasoning. In one session at both two and four years of age,
the child and his/her caregiver attended, and the
child encountered situations including the entrance
There is continuing debate about the optimal of a clown, unusual and noisy toys, and a brief sep-
ways of measuring temperament and parenting. aration from the parent. The second session at two
Many studies have relied on parental report of their years involved parent–child same-sex pairs, who
style of parenting and also of their child’s tempera- did not know each other, attending the playroom at
ment. There are obvious opportunities for bias in the same time. In this session we observed how the
such reports. Fine-grained observations of parents children played with toys, whether and how they
and children in situations which elicit a range of approached the other child and adult, and how they
reactions from both of them offer valuable insights responded when asked to do things like tidy away
into the processes of parent–child interaction. In the toys. In the second session at four years, four
this study, questionnaire and detailed observational children of the same sex who did not know one
data were used. another attended together and played with toys,
were asked to give a speech in front of the other
children about their last birthday, were asked to do
Social Development Project various other tasks, and then pack up the toys.
This article describes the results of a study that inves- Each session in the playroom was videotaped and
tigated the influence of the match between child later coded to indicate how the parents interacted
temperament and parenting assessed at two years on with their children (for example, comforts, encour-
child behavioural problems at four years of age. It was ages, controls, plays), as well as the child’s style of
expected that the combination of high negative reac- behaving (for example, mood, activity level, how
tivity and high parental punishment would predict long the child sticks to one activity, how strongly the
behavioural problems such as aggressive, noncompli- child reacts to events, how close the child stays to
ant behaviour. Also explored were other possible parent, how long it takes the child to play with a
combinations, including negative reactivity with low novel toy), and adjustment (for example, how often
parental warmth and with low inductive reasoning, the child disobeys parent, how often the child hurts
and low inhibition with high punishment. another child, whether the child watches other chil-
Data reported here come from a longitudinal com- dren play without joining in). From the playroom
munity study of young children’s social development, sessions and parent report questionnaires, measures
known as the Social Development Project (funded by of child temperament (negative reactivity, inhibi-
the Australian Research Council), which began in tion, and persistence), parenting style (warmth,
1996. In contrast to the Australian Temperament punishment, inductive reasoning and overprotec-
Project (Prior et al. 2000) which collects more tiveness), and child adjustment (behaviour
“broad-brush” data on a large sample of children, this problems, social withdrawal) were obtained.

44 Family Matters No.59 Winter 2001 Australian Institute of Family Studies

 This article focuses on the match between those have behavioural problems by four years were
aspects of temperament and parenting style most reactive, and these children received much lower
likely to be associated with behavioural problems. For warmth at two years than the other 64 reactive
child temperament these are negative reactivity and children, and the 35 non-reactive children with-
inhibition, and for parenting style these include out behavioural problems. However, it is also
warmth, punishment, and inductive reasoning. The notable that only two non-reactive children
article examines how the temperament and parenting developed behavioural problems but these had
measures drawn from the parent–child session at two high levels of parental warmth. Thus, for reactive
years relate to measures of problems at four years of children only, lower levels of parental warmth at
age, as observed with peers and as reported by par- two years predict later behavioural problems.
ents. The behavioural problem measures include Punishment and reactivity: Figure 3 shows
evidence of the child hitting others, damaging objects, that the six reactive children who were later
disobeying instructions, and having temper tantrums. observed to have behavioural problems had
experienced more punishment than the 64
Levels of adjustment problems and parent–child match other reactive children without such problems,
In general, the levels of behavioural problems in and the 35 non-reactive children without behav-
this sample were relatively low. Boys showed more ioural problems. Again, the two non-reactive
behavioural problems than girls, both as reported children who developed behavioural problems
by parents and observed in the laboratory. Figure 1
shows scores on parent-reported behavioural prob-
lems; the pattern of scores was similar for observed Figure 1 Sex differences in behavioural problems
behavioural problems.
Children were divided into those with and with- 2.6
out behavioural problems at four years. Eight
behavioural problems

2.5
Mean 4-year-old

children (seven of them boys) had behavioural

problems by observation at four years, and 21 (16 of 2.4
them boys) showed behavioural problems by parent 2.3
report. Because there were so few girls who showed
2.2
behavioural problems by parent-report and obser-
vation, it was not possible to analyse by sex, and the 2.1
results for the entire sample are described here. Boys Girls
However, from inspection of the data it appears that Sex of child
Note: * p < .05
somewhat different patterns might emerge for girls Source: Social Development Project, 2001.
if a large enough group was analysed.
The main questions of interest were: for each Figure 2 Warmth and reactivity with observed behavioural
type of child temperament, did the type of parent- problems (BPs)
ing style at two years differ for those with and
without behavioural problems at four years? In 3.2 BP groups
particular, we were interested in the impact of par- Low BPs
3.0 2
enting on those with temperament characteristics
Mean rating of

High BPs
(high reactivity and very low inhibition) which
warmth

2.8 35 64
might predispose them to the development of prob-
2.6
lems. Children were therefore divided into groups
on the basis of their temperament scores when 2.4
observed at two years (no or some negative reactiv- 2.2 6
ity, and low, moderate or high inhibition). There No reactivity Some reactivity
were 74 children (44 boys, 30 girls) showing some Reactivity
negative reactivity in the observation session with Note: The bars indicate levels of parenting, not numbers of children; the numbers of
children are indicated by the numbers in the bar.
their parents at two years. Thirty-eight children (22 Source: Social Development Project, 2001.
boys, 16 girls) were highly inhibited in the par-
ent–child observation session at two years. Thus Figure 3 Punishment and reactivity with observed behavioural
there were more highly reactive boys than girls and problems (BPs)
slightly more inhibited boys than girls.
The average levels of each parenting style in the 1.3 BP groups
parent–child laboratory session were then exam- Low BPs
ined for each of the sub-groups formed. The 1.2 6
Mean rating of

High BPs
punishment

findings are presented in Figures 2–5. Note that the

bars indicate levels of parenting, not numbers of 1.1
children; the numbers of children are indicated by 35
1.0 64
the numbers inside the bars. Results for behav- 2
ioural problems as assessed by questionnaire and 0.9
observation were often similar, so graphs of only No reactivity Some reactivity
one method of assessment are presented here for Reactivity groups
each combination of parenting and temperament. Note: The bars indicate levels of parenting, not numbers of children; the numbers of
Warmth and reactivity: It can be seen in Figure 2 children are indicated by the numbers in the bar.
Source: Social Development Project, 2001.
that six of the eight children who were observed to

Australian Institute of Family Studies Family Matters No.59 Winter 2001 45

 were receiving low levels of punishment at two effects for children who were moderate or high on
years. However, in general these results show that inhibition. A similar pattern was found for observed
the combination of reactivity and high levels of behavioural problems.
punishment at two years was the most common
precursor to behavioural problems at four years. Implications for parenting education
Inductive reasoning and reactivity: Of the 21 In summary, it appears that parenting, in some cir-
children who had parent-reported behavioural prob- cumstances, has a different impact on child
lems at four years, 19 were highly reactive at two adjustment depending on the temperament of the
years and had experienced lower levels of inductive child. A highly reactive child is at risk of developing
reasoning than the other 55 reactive children with- behavioural problems if the parenting s/he receives
out behavioural problems and the 35 non-reactive is higher on punishment, but is protected from this
children without problems (Figure 4). Again, the two outcome if parents do not use high levels of punish-
non-reactive children who developed behavioural ment. A highly reactive child is also more likely to
problems showed a different pattern, with high lev- develop behavioural problems if the parent shows
els of inductive reasoning. With these exceptions, lower levels of warmth during interactions with
the most common pattern for children whose par- him/her. Parental inductive reasoning seemed to
ents reported behavioural problems at four years of interact with temperament less often than punish-
age was for them to be reactive and to receive lower ment and warmth, however it appeared that low
levels of inductive reasoning as toddlers. levels of inductive reasoning posed a risk specifi-
Punishment and inhibition: Figure 5 shows that cally for reactive children.
six of the 21 children with parent-rated behavioural A highly outgoing child who experiences high
problems at four years were low on inhibition parental punishment appears more likely than
(highly uninhibited or outgoing), and they received other children to develop behavioural problems.
more punishment in the parent–child session than But if the same highly outgoing child receives par-
the other 32 low-inhibited children who did not enting which channels the child’s energy and
develop behavioural problems. The other 15 chil- exuberance in positive ways, the child is likely to be
dren with four-year-old parent-rated behavioural well-adjusted. A more inhibited child is unlikely to
problems did not differ from their non-behavioural develop “acting-out” behavioural problems of the
problems counterparts in the levels of punishment. sort assessed here, irrespective of parenting,
These results show that for uninhibited toddlers, although this study and others (Rubin et al. 2001)
higher levels of punishment at two years appear to suggest that they can be susceptible to social with-
contribute to later behavioural problems. However, drawal and later anxiety and fearfulness, given
levels of punishment did not appear to have different exposure to particular parenting styles such as
overprotectiveness.
Figure 4 Inductive reasoning and reactivity with reported behavioural Of course, a highly sociable (possibly risk-taking)
problems (BPs) child and a highly reactive, intense, and irritable
child are more demanding to parent, and may elicit
4.0 BP groups exactly the sort of parenting that appears worst for
3.8 Low BPs them. Patterson et al. (1989) have argued that chil-
inductive reasoning

2
dren with particular temperamental traits are more
Mean rating of

High BPs
3.6
likely to have parents who use high levels of punish-
3.4 35
ment, and have shown that the parent–child
3.2 55 interactions in such cases often develop into “coer-
3.0 19 cive cycles” of mutually antagonistic behaviour,
2.8 with the longer-term result that children develop
No reactivity Some reactivity aggressive behaviour (Patterson et al. 1989). Simi-
Reactivity larly, Scarr and McCartney (1984) have postulated
Note: The bars indicate levels of parenting, not numbers of children; the numbers of “evocative” gene-environment interactions, where
children are indicated by the numbers in the bar.
Source: Social Development Project, 2001. temperamental traits (which are partly genetically
determined) elicit a particular style of parenting,
Figure 5 Punishment and inhibition with reported behavioural which then results in particular outcomes.
problems (BPs) A major implication of these findings is that par-
ents need help in understanding the unique nature
1.4 BP groups of their child, and in finding appropriate ways of par-
Low BPs enting that child. It should be noted that the
1.3
majority of children in the present study, whatever
Mean rating of

6 High BPs
punishment

1.2 their temperament (high or low reactivity, high or

low inhibition), did not have behavioural problems.
1.1
31
Therefore it is inappropriate to regard high reactivity
1.0 32 7 and uninhibited styles as necessarily “difficult” traits
27 8
0.9 – the challenge is to find the best fit between these
Low inhibition High inhibition High inhibition traits and parenting style. Levels of child negative
Inhibition reactivity observed in the laboratory were also rela-
Note: The bars indicate levels of parenting, not numbers of children; the numbers of tively low, which is not surprising for this sample.
children are indicated by the numbers in the bar. It should also be noted that most children in this
Source: Social Development Project, 2001.
study had experienced fairly positive parenting.

46 Family Matters No.59 Winter 2001 Australian Institute of Family Studies

This means that we do not know whether the pat- Loeber, R., Stouthamer-Loeber, M. & Green, S. M. (1991), “Age
of onset of problem behaviours in boys, and later disruptive
tern of results found here would hold for very high and delinquent behaviours”, Criminal Behaviour and
negative reactivity and high levels of parental pun- Mental Health, vol. 1, pp. 229-246.
ishment. But the detailed observational data reveals Paterson, G. & Sanson, A. (1999), “The association of behav-
that relatively small differences in the levels of ioural adjustment to temperament, parenting and family
characteristics among five-year-old children”, Social
warmth, punishment and inductive reasoning have Development, vol. 8, pp. 293-309.
a significant impact for children who are tempera- Patterson, G.R., DeBaryshe, B. & Ramsey, E. (1989), “A devel-
mentally at risk. The implication remains that opmental perspective on antisocial behaviour”, American
“recipe book” approaches in parenting programs, Psychologist, vol. 44, pp. 329-335.
promoting “the right way” to parent, may miss the Pettit, G.S. & Bates, J. E. (1989), “Family interaction patterns
and children’s behaviour problems from infancy to four
mark for many children, and parents need to be years”, Developmental Psychology, vol. 25, pp. 413-420.
given the confidence and the skills to adapt their Prior, M., Sanson, A., Smart, D. & Oberklaid, F. (2000),
parenting as appropriate. Pathways from Infancy to Adolescence: Australian
A research implication from this study is that a Temperament Project 1983–2000, Australian Institute of
Family Studies, Melbourne.
fine-grained approach to research, using painstak-
ing coding of observational data which necessitates Rothbart, M.K. & Bates, J. E. (1998), “Temperament”, in W.
Damon (series ed.) & E. Eisenberg (volume ed.) Handbook
smaller samples than surveys, is very useful for of Child Psychology: Vol. 3. Social, Emotional and
increasing understanding of the actual processes of Personality Development, (5th edn), Wiley, New York.
child development. Rubin, K. H., Cheah, C. S. L., & Fox, N. (2001), “Emotion reg-
The children will continue to be followed to ulation, parenting and display of social reticence in
preschoolers”, Early Education and Development, vol. 12,
seven years of age to see whether the trends are pp. 97-115.
consistent with those described here. Scarr, S. & McCartney, K. (1984), “How people make their own
environments: a theory of genotype-environment effects”,
References Annual Progress in Child Psychiatry and Child
Development, pp. 98-118.
Azar, B. (1995), “Timidity can develop in the first days of life”,
Monitor, vol. 26, no. 23, November. Thomas, A., Chess, S. & Birch, H.G. (1968), Temperament
and Behaviour Disorders in Children, New York University
Baumrind, D. (1966), “Effects of authoritative parental control Press, New York.
on child behaviour”, Child Development, vol. 37, pp. 887-907.
Zubrick, S.R., Silburn, S.R., Burton, P. & Blair, E. (2000),
Booth, C. L., Rose-Krasnor, L., McKinnon, J. & Rubin, K. H. “Mental health disorders in children and young people:
(1994), “Predicting social adjustment in middle childhood: scope, cause and prevention”, Australian and New
The role of preschool attachment security and maternal Zealand Journal of Psychiatry, vol. 34, pp. 570-578.
style”, Social Development, vol. 3, pp. 189-204.
Chamberlain, P. & Patterson, G. R. (1995), “Discipline and Sheryl Hemphill is a Research Fellow in the Department of
child compliance in parenting”, in M.H. Bornstein (ed.) Psychology at the University of Melbourne. Ann Sanson is
Handbook of Parenting: Vol. 4: Applied and Practical Deputy Director (Research) at the Australian Institute of
Parenting, Lawrence Erlbaum, NJ. Family Studies.
Crockenberg, S. (1987), “Predictors and correlates of anger Some of the data in this article were presented at a collo-
toward and punitive control of toddlers by adolescent moth- quium at the School of Psychology, La Trobe University in
ers”, Child Development, vol. 58, pp. 964-975. August 2000. The research described here has been funded
Hart, C.H., DeWolf, M., Wozniak, P. & Burts, D. C. (1992), by two Australian Research Council grants to Associate Pro-
“Maternal and paternal disciplinary styles: relations with fessor Ann Sanson. The authors would like to thank the
preschoolers’ playground behavioural orientations and peer participants and staff of the Social Development Project
status”, Child Development, vol. 63, pp. 879-892. who have made this study possible.

F u t u r e d i r e c t i o n s

S o c i a l D e ve l o p m e n t P r o j e c t
The data reported in this Family Matters article came from a In another extension of the study, the authors are collaborat-
community sample in which both parents and children were ing with Professor Ken Rubin from the University of
generally functioning well. The risk for the development of Maryland, USA, in conducting cross-cultural comparisons
adjustment problems among children is much higher if par- between groups of Canadian, Chinese, Italian and Australian
ents have difficulties of their own. children, on all of whom similar observational and parent-
report data have been collected.
In 1999, the authors, along with Jordana Bayer, received a
VicHealth grant to replicate the procedures of the Social Devel- Analyses of data on the two-year-olds suggests that the Ital-
opment Project with a sample of 50 children whose parents ian children are the most outgoing, followed by the
reported having difficulties with anxiety and/or depression. So Australians, and then the Canadians. The Chinese children
far, two-year-old data have been collected and analysed, and are the most shy. Parenting behaviour in each country differs,
results suggest that these children are indeed showing higher with Australian parents exhibiting the most warmth and pro-
levels of adjustment difficulties. There is also some suggestive tectiveness.
evidence that the pathways to the development of behavioural
and emotional problems in this group differ from those for the As the children are followed forwards, the authors hope to
community sample. These children are currently attending the unravel further the influence of culture on both the children’s
lab for their four-year-old sessions. and the parents’ behaviours.

Australian Institute of Family Studies Family Matters No.59 Winter 2001 47