29-08-2016, 04:23 PM
1451551771-57.pdf (Size: 3.32 MB / Downloads: 389)
Introduction
Schooling has direct effects on children's educational achievement, their acquisition
of literacy, numeracy and scientific knowledge. These basic skills provide the
foundation for later "subjects" such as geography, physics and foreign languages.
Formal educational qualifications are the key to a child's entry into higher education
or training and also employment. The learning of specific knowledge and skills is
a direct effect of classroom teaching (Good & Brophy, 1986b). However, social
cognitions and feelings are also influenced by school and these may be just as powerful
in predicting later outcome as intelligence or school curriculum. Such indirect effects
of school are more elusive because they are mediated by children's motivation to learn
or avoid learning, their conception of themselves as pupils, and the attributions they
create for explaining success and failure. Cognitive and motivational mediators of
indirect effects continue to exert influence on individual development outside and
beyond school.
This selective review considers the evidence concerning direct and indirect effects
of school on children's development. Section 1 examines the evidence on the effect
of pre-school education on children's academic attainment, social behaviour and
cognitions. There are several well designed experimental studies of the impact of preschool
education which have included follow-up through young adulthood. These
landmark studies employed randomised designs which contrasted the development
of children who had and had not experienced pre-school education, thus allowing
causal models to be devised which suggest lasting benefits of pre-school education,
especially for those from disadvantaged backgrounds.
Section 2 outlines a few of the major studies on the effect of primary schooling.
Research on the effects of primary and secondary education does not withhold
education from children, thereby necessitating either natural experiments or
correlational designs employing sophisticated statistical techniques, over time, to
Keywords: School effects, pupil cognitions, pupil attributions, academic self concept
Accepted manuscript received 29 July 1993
Requests for reprints to: Dr K. Sylva, Department of Child Development and Primary Education, Institute
of Education, University of London, 20 Bedford Way, London WCIH OAL, U.K.
© Crown copyright.
135
136 K. Sylva
unravel those experiences associated with better outcomes for different groups of
children. Methodological weaknesses are discussed and the characteristics of "effective
schools" are outlined.
The remainder of the paper examines the means by which schools influence
development. It examines mechanisms which might underlie the influence of schools.
Section 3 examines the ways education shapes pupils' attributions, motivations and
beliefs about self. Section 4 looks at social participation and social responsibility
in terms of its influence on attainment. Section 5 examines teacher influences and
Section 6 looks at the ways classroom size and organization contribute to students'
outcome. Thus, the extent of school effects are presented first, followed by consideration
of mediating factors (cognitive, motivational, organisational) which might underpin
them. The final section explores the direct and indirect effects of school and considers
developmental trends.
The review focuses primarily on pre-school and primary school evidence, mentioning
secondary research only to fill in gaps or explore methodology. It concentrates as
much on "development" as on speciflc educational outcomes such as subject knowledge
or skill. The review is limited to research on "mainstream" pupils because inclusion
of those with special needs would have made the task unmanageable.
1. The Effects of Pre-school Education on Children's Development
Since education often begins before compulsory school, this review begins by
examining the evidence on the effects of pre-school education.
Early programmes of "compulsory education"
The American project Head Start, a legacy of Lyndon Johnson's War on Poverty
(Valentine, 1979) has received government funding for two decades in the hope that
it would "break the cycle of poverty". A simple comparison design was used in early
studies on the impact of Head Start. Typically, I.Q. or attainment test scores of
pre-school "graduates" were compared to scores of control children who had no
pre-school experiences. Initial evaluations seriously underestimated the value of the
programme (Campbell & Erlebacher, 1970; Smith & Bissell, 1970) by focusing on
measures of intelligence as the main outcome. Sadly they found that early I.Q. gains
quickly washed out, leaving Head Start children no different from controls.
The British research mirrors that for the US with interventions during the 1970s
aimed at "closing the poverty gap" leading to disappointment (Smith & James, 1977;
Tizard, 1974) when the initial gains tended to "wash out".
More recent evaluations have employed sophisticated research methods and looked
at a wider array of child outcomes. In 1985, a meta-analysis of research findings on
the American Head Start programme was published (McKey et al., 1985). This
included the results of 210 studies evaluating the impact of Head Start. To enable
comparison amongst the studies, findings were converted to statistical "effect sizes"
and comparisons were made across different sites, target groups and tests on children.
McKey and his colleagues concluded that Head Start had immediate, positive effects
on children's cognitive ability but that cognitive gains were no longer apparent after
School influences 137
the end of the second year at school. Head Start also had short-term positive effects
on children's self-esteem, scholastic achievement, motivation and social behaviour,
but these advantages disappeared by the end of the third year. The authors point
out that the studies were designed so differently and varied so widely in terms of
rigour that it was impossible to come to firm conclusions on many questions, including
those concerning the role of parental involvement. Most studies included in this
research synthesis did not control adequately for pre-intervention differences in
children's ability, many studies were on one site only, and a few had no "control"
groups.
The smaller, better controlled studies of the effects of Head Start have yielded more
robust findings. A well designed study by Lee, Brooks-Gunn and Schnur (1988)
compared the outcomes of 969 disadvantaged children who had experienced three
different pre-school environments: Head Start, some other pre-school programme
and no pre-school. Large, initial differences on a wide range of outcomes were found
at school entry, with Head Start children lower on almost all measures. After adjusting
for initial scores (because those in the Head Start sample were lower). Head Start
children showed larger gains on measures of social and cognitive functioning compared
with children in the other two groups. Often children in Head Start begin school
with lower levels of functioning since many researchers have found children
participating in Head Start come from families of serious social disadvantage (Seitz,
Abelson, Levine & Zigler, 1985) includirig lower levels of income and education
(Hebbler, 1985). Thus, in Lee's study. Head Start was effective in "closing the gap"
but did not succeed in doing so completely because its children began at greater levels
of disadvantage.
Notable in Lee's study were the large gains made by black children in Head Start.
"Pre-school intervention is particularly effective for the most economically
disadvantaged children" (Zigler, 1987). Lee et al. reported that black children gained
more than white children, even when controlling for initial levels of ability. Further,
black students of below average ability gained more than their counterparts of average
ability. Lee affirmed the effectiveness of Head Start:
not only were those students most in need of pre-school experience likely to be in Head
Start programs, but also that those Black students who exhibited the greatest cognitive
disadvantage at the outset appeared to benefit most from Head Start participation. (Lee
etal., 1988, p. 219).
Research studies designed as experiments
The failure to find a long-term impact of early education has not been confined to
Head Start (see Porter, 1982, on Australia and Myers, 1992, on the developing world).
However, there is cause for optimism when examining research on programmes which
are not part of large government initiatives. Lazar and Darlington (1982) carried out
a meta-analysis of the effects of pre-school education programmes which were part
of research projects rather than government programmes. They ignored the gardenvariety
programmes (which included Head Start) to focus on projects of good curricular
quality, mostly aimed at disadvantaged groups, which employed rigorous research
designs. The meta-analysis was limited to pre-school projects with sample sizes greater
than 100 children, which used norm-referenced assessment tests, comparison/control
138 K. Sylva
groups, and followed up children well beyond school entry. By these strict criteria
11 carefully monitored programmes were chosen for meta-analysis, half using random
assignment. The researchers located approximately 2000 pre-school "graduates" and
their matched controls in order to document educational and employment histories.
In addition they interviewed the youth, all aged 19, and their families.
Results from the 11 studies showed that attendance at excellent, cognitively oriented
pre-school programmes was associated with later school competence. Pre-school
graduates were less likely to be assigned to "special" education or to be held back
in grade while their peers moved up, and where data were available, were more likely
to be in employment. Interviewers found that pre-school graduates were more likely
than the control group to give achievement-related answers to the invitation "tell
me something you've done that made you feel proud of yourself". When parents
were interviewed about attitudes towards school performance, mothers of pre-school
graduates expressed more satisfaction with their children (even after controlling for
the actual performance of each child). The greater satisfaction with school work appears
related to the mother's aspirations for their children's employment. In answer to the
question "What kind of job would you like (your child) to have later in life?", mothers
of children who attended pre-school replied with skilled or managerial jobs.
Taken together with the "harder" outcomes, the attitudinal fmdings suggest that
early education changed the achievement orientation of the family. Mothers whose
children attended pre-school expected more from their children, and these expectations
were fulfilled. The children themselves showed more pride in their achievements.
Lazar and Darlington have constructed a theoretical model to explain their fmdings.
. . . it seems possible that mutual reinforcement processes occurred between the early
education participants and their parents. Perhaps the children's participation raised the
mothers' hopes and expectations for their children . . . Perhaps children interpreted these
parental attitudes as a belief in and support of their efforts, and it served to spur them
on. Along these lines .. . (it is) hypothesized that the early education experience may change
children from passive to active learners who begin to take the initiative in seeking
information, help, and interaction with others. When this increased motivation to learn
is met by a positive response at home and at school, long-term gains on outcome measures
of cognitive development can result. (Lazar & Darlington, 1982, p. 63).
The most carefully controlled of the 11 programmes reviewed by Lazar was the Perry
Pre-school Project, later known as High/Scope. This active-learning curriculum
includes a complex training scheme for staff, and parent participation. Research on
it employed random assignment to experimental and control groups and has been
carried out for almost 30 years. Although an initial I.Q. advantage for pre-school
graduates disappeared by entry to secondary school, there were startling differences
in outcome between the 65 children who attended the half-day educational programme
over two years and the control group of 58 children who had remained at home
(Berrueta-Clement, Schweinhart, Barnett, Epstein & Weikart, 1984). By the age of
27, the High/Scope "graduates" had
• significantly higher monthly earnings at age 27 (29% vs 7% earning $2000 or more
per month)
• significantly higher percentage of home ownership (36% vs 13%) and second car
ownership (30% vs 13%)
School influences 139
• a significandy higher level of schooling completed (71 % vs 54% completing 12th
grade or higher)
• a significantly lower percentage receiving social services at some time in the past
10 years (59% vs 80%)
• significantly fewer arrests by age 27 (7% vs 35% with 5 or more), including
significantly fewer arrested for crimes of drug taking or dealing (7% vs 25%).
Schweinhart and Weikart (1993) speculate on the mechanisms which brought about
such lasting change in disadvantaged children.
The essential process connecting early childhood experience to patterns of improved success
in school and the community seemed to be the development of habits, traits, and dispositions
that allowed the ehild to interact positively with other people and with tasks. This process
was based neither on permanently improved intellectual performance nor on academic
knowledge. (Schweinhart & Weikart, 1993, p. 4).
Despite the attraction of so persuasive a theory, Weikart and his colleagues present
litde evidence from primary school to demonstrate the emergence of altered cognitions
and motivations which they suggest underpin the "virtuous career" of children who
had attended the High/Scope nursery school. Still, Schweinhart and Weikart (1993)
present a cost-benefit analysis which shows that for every $1000 that was invested
in the pre-school programme, at least $7160 (after adjustment for inflation) has been
or will be returned to society. These calculations were based on the financial cost
to society of juvenile delinquency, remedial education, income support, and
joblessness—set against the running costs of a well-resourced pre-school programme.
The economic analysis also estimates the return to society of taxes from the higher
paid pre-school graduates.
There have been two other cost-benefit analyses carried out on pre-school
interventions, both in the US. Barnett and Escobar (1990) present American data
from a pre-school language intervention curriculum studied by Weiss and a
comprehensive early day care programme for disadvantaged families studied by Seitz
et al. Both showed that the costs of the pre-school interventions were more than offset
by the savings later on in the children's educational career and medical history.
Although American studies do not suggest that all pre-school programmes will bring
lasting benefits, they demonstrate that early education can change the course of
children's lives, especially those from disadvantaged backgrounds. Although samples
are often small, the experimental designs and longitudinal follow-up give strong support
to the claim that the pre-school experiences actually caused the beneficial and costeffective
outcomes.
In a later study, Weikart and his colleagues compared the effects of three different
curricula on randomly assigned children (Schweinhart, Weikart & Larner, 1986).
They found that children from a High/Scope programme, a "free play" programme
and also a formal skills-based curriculum all had increased I.Q.s at school entry.
However, follow up at the age of 15 showed that children who had attended the formal
programme engaged more in anti-social behaviour and had lower commitment to
school than those who attended the two programmes based on active learning/play.
Thus, raised I.Q,. at school entry does not necessarily give children a good start
to schooL Only the children who experienced active learning programmes before
140 K. Sylva
school retained the advantage of their early education, an advantage they showed
by pro-social behaviour and higher confidence in adolescence.
There is considerable debate about why programmes such as High/Scope are so
effective. One causal model supported by path analysis suggests that pre-school
education promotes cognitive and social skills that result in greater school readiness
and a smoother transition to school (Berrueta-Clement et al., 1984). Children leave
the nursery "ready to learn" and are easily recognised by teachers, who show positive
expectations and treatment, and this, in turn, fosters improved student attitudes
towards school and better school behaviour (called "school commitment" by Weikart
and colleagues). These serve as protective factors against the later risk of maladjustment
and delinquency. Schweinhart and Weikart (1993) argue that well resourced, cognitive
orientated pre-school programmes such as High/Scope should be effective outside
the original setting in which the research took place.
Woodhead (1985, 1989) questions this; he suggests that the explanation for lasting
benefit in the longitudinal American studies is differential retention at grade or referrals
to special education. Children in the pre-school group in Lazar's meta-analysis were
less likely to be held back while the year-group moved up. If avoidance of grade
retention, rather than the emergence of "bright and ready" graduates is what sets
children on different pathways, then the positive results seen in these studies may
not generalise to other settings where grade retention is not practised.
The next study to be reviewed is a quasi-experiment conducted by Jowett and Sylva
(1986) on 90 working class children in Britain. This research examined the impact
on the first year of school on two groups of children, one coming from well resourced
state nursery education and the other from voluntary playgroups run by parents.
Although the "quasi-experiment" did not employ random assignment to treatment
condition, many background variables which might affect outcome in the two
groups of children were carefully controlled (e.g. parental occupation, family structure,
type of housing). Parent choice was eliminated to some extent because children were
drawn from neighbourhoods where there was only one form of half-day education:
either playgroup or nursery education. Previous research on provision in the same
education authority (Sylva, Roy & Painter, 1980) showed curricular variation with
the playgroups run along "free play" lines and nursery classes focusing on "guided
play" and "extension" of children's own activities by highly trained adults. The
researchers hypothesised that children from the better resourced provision would
be recognisable by their teacher when they entered school as more "learning
orientated".
Results confirmed that the children who had attended nursery engaged in more
purposeful and complex activity in the first year of school than did the children who
attended playgroup. During "free choice" sessions in the reception class, the nursery
"graduates" chose more educational activities while the playgroup children spent
more time engaging in non-demanding play. Nursery children were more likely than
the playgroup children to initiate contacts with the teacher that were "learning
orientated" whUe the playgroup children approached teachers for help. Finally, nursery
"graduates" were more persistent and independent when they encountered obstacles
in their learning. This study shows that the kind of pre-school education a child
experiences affects the ease with which children begin their school careers.
School influences 141
Furthermore, it provides evidence for Weikart's hypothesis that children characterised
by learning orientation can be easily recognised by their classroom behaviours.
Education in day care settings
There is still disagreement amongst scholars as to whether early entrance into day
care, say before the age of one, is detrimental to children's later development (ClarkeStewart,
1989; McGurk, Caplan, Hennessy & Moss, 1993). Hennessy and Melhuish
(1991) reviewed the effects of^day care on school adjustment and found mixed results;
often day care was associated with better grades and social acceptance (Field, 1991)
but sometimes with worse.
A longitudinal study by Osborn and Milbank (1987) on more than 8400 children
born in the UK during 1970 showed a clear association between pre-school attendance
and educational outcomes (reading, maths) and social ones (behaviour problems) at
the age of 10. The authors argued that pre-school attendance brought about the better
cognitive performance seen in children who had attended pre-school "education"
but not those who attended "care". But is the causal evidence firm? Birth cohort
studies do not randomly assign children to different pre-school experiences; these
researchers ustApost hoc statistical analysis for differential intake. Clark (1988) queried
whether the statistical adjustment could completely eliminate factors such as family
choice or parental interaction. In the '80s we have been taught to measure the "value
added" component (see Section 2). However, the longitudinal study shows clearly
that many day care children fare worst of all in later life. Explanation for their poor
outcome may lie in identifying those families who managed to obtain one of the places
in government day care. McGuire and Richman (1986) found that children attending
day care centres run by UK Social Services Departments had 10 times more behaviour
and emotional problems as children in the same neighbourhood attending playgroups.
The educational outcome of day care may rest on social balance in enrolment and
perhaps on age of entry.
Swedish longitudinal research by Andersson (1989, 1992) and Cochran and
Gunnarsson (1985) found day care experience gave children a better start in school.
Andersson examined the development of 128 children who attended neighbourhood
day care centres in Gothenburg where both low and middle income families routinely
sent their children. Progress was monitored from the children's first year in day care
to the age of 13. No developmental disadvantage was found in the day care group
compared to children who had stayed at home. In fact, the highest performance in
school tests and the best emotional adjustment was found in the children who had
experienced the most day care, even before the age of one year. This study is important
because it shows that the benefits of pre-school education are not confined to those
from disadvantaged backgrounds.
An interesting study in the US by Howes (1990) focused on 80 children in deliberately
contrastive care. Half were enrolled in excellent centres and half in poor ones. "High
quality" centres were characterised by the following: (1) stable child care arrangements
such that children interacted with just a few primary caregivers in any one day; (2)
low staff turnover so that children were cared for by the same individuals over several
years; (3) good staff training in child development; and (4) low adult: staff ratios