International Journal of Educational Development 29 (2009) 39–45

Contents lists available at ScienceDirect

International Journal of Educational Development

journal homepage: www.elsevier.com/locate/ijedudev

Impact of school readiness program interventions on children’s learning in

Cambodia§
Yuko Nonoyama-Tarumi a,*, Kurt Bredenberg b
a
United Nations Children’s Fund (UNICEF), Programme Division, Early Childhood Development Unit, Three United Nations Plaza New York, NY 10017, United States
b
Kampuchean Action for Primary Education (KAPE), Kampong Cham Teacher Training College, Kampong Cham Province, Cambodia

A R T I C L E I N F O A B S T R A C T

Keywords: To reduce the high repetition rates in early years of primary school, the government of Cambodia piloted a
International education school readiness program (SRP) in the first two months of Grade 1 of primary school. This study examines
Development whether such intervention has effects on students’ immediate acquisition of school readiness skills as
School readiness
well as students’ longer term achievement of formal curriculum. The study finds that children who
Cambodia
participated in SRP outperform children that did not participate in both outcomes, controlling for pretest
score and background variables. The findings suggest that the intervention may be one alternative model
to preschools in countries where access to pre-primary education is limited.
ß 2008 Elsevier Ltd. All rights reserved.

1. Introduction phase, thereby making them less likely to repeat a grade or drop
out of school. SRP focuses on building foundational skills in
The issue of high repetition and drop out rates in many academic subjects, promoting learning friendly classroom envir-
developing countries is widely known, and data has shown that the onments, and strengthening learner confidence.
rate is especially high in early years of primary school. Cambodia is In addition to the high repetition rate, access to pre-primary
not an exception where time after time, Grade 1 children have had education is extremely limited in Cambodia, due to the intense
the worst rates of promotion of any grade grouping at primary competition in the education sector for scarce resources, as is the
school level (Bredenberg, 2000, 2005). Although the implementa- case in many economically less-developed countries.1 Under such
tion of reforms at the turn of the decade (e.g., summer vacation a situation, a short school readiness course at the beginning of
remediation programs) led to a decline in repetition of about 50%, primary school may be a more realistic immediate option than a
national repetition rates at Grade 1 have since remained relatively drastic expansion of preschool. By introducing a school readiness
static at around 17–23% each year between 2000 and 2003 (EMIS, course in the first two months of child’s formal education, the
2000–2005). Such rates are high even when compared to the Cambodian Ministry of Education, Youth, and Sport sought to
median repetition rates of developing countries, which is about 7% compensate for the lack of formal preschooling and generally poor
(UNESCO, 2007). early childhood development experiences that an under-resourced
In response to this situation, educators in Cambodia have education sector in Cambodia is currently unable to overcome.
explored a new strategy of introducing a school readiness program As part of the SRP implementation in Cambodia, a systematic
(hereafter referred as SRP) in the first two months of the academic monitoring and evaluation was undertaken by the government and
year of Grade 1. The underlying theory behind the SRP is that local partners leading to a series of research reports on the
children with a higher degree of ‘‘readiness’’ will acquire program’s effectiveness (Bredenberg, 2004, 2005). This study
competencies outlined in the formal curriculum more successfully synthesizes these evaluations, highlighting the major findings in
than children who do not pass through a structured readiness observed impacts of the SRP interventions on children’s learning
outcomes during the first year of implementation. Our research
questions in this study are two-folds. What is the immediate effect
§
We appreciate the comments provided by Patrice Engle to earlier version of this
manuscript. The view expressed herein are those of the authors and do not
necessarily reflect the opinions or policies of the UNICEF or any of its affiliated
1
organizations. According to the Global Monitoring Report 2008 (UNESCO, 2007), the average
* Corresponding author. Tel.: +813 5978 5249. gross enrollment ratio for pre-primary education of economically less-developed
E-mail address: [email protected] (Y. Nonoyama-Tarumi). countries is less than half of that of developed countries, 28% and 73%, respectively.

0738-0593/$ – see front matter ß 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijedudev.2008.07.003
40 Y. Nonoyama-Tarumi, K. Bredenberg / International Journal of Educational Development 29 (2009) 39–45

of SRP on children’s school readiness skills? What is the long-term 2.3. School readiness
impact of SRP on children’s achievement? In other words, we
investigate to what extent students acquired school readiness This transition period described above is a ‘‘critical period’’ for
skills through participating in the intervention, and to what extent acquisition of ‘‘school readiness,’’ defined as the behaviors and
school readiness skills enhanced students’ understanding of formal skills needed to adjust to the formal learning structure found in
Grade 1 curriculum? An evaluation of an alternative program for school. For example, a qualitative study in Nepal showed that
preschool, such as school readiness programme, will have strong teachers often commented that children who have been to early
policy implications for many countries where a drastic expansion learning centers are ‘‘very different [from other children]; they
of preschool is not a realistic option due to scarce and competing know how to be in a group of others, they are less timid, able to
resources. respond’’ (Arnold, 2004). Today, school readiness is recognized as a
multi-faceted construct (Scott-little et al., 2006). As noted above,
2. Prior research one of these dimensions refers to children’s readiness for school,
which focuses on learning and developmental outcomes. A second
2.1. Importance of early intervention dimension is the schools’ readiness for children, which focuses on
school-level outcomes and practices that foster and support a
The recent explosion of research in neurobiological, behavioral, smooth transition into primary school and promote the learning of
and social sciences has led to major advances in understanding the all children. Finally, a third dimension refers to families’ readiness
importance of early childhood experiences. Based on an extensive for school, which focuses on parental and caregiver attitudes and
review of existing science on early childhood in the Unites States, involvement in children’s early learning as well as development
‘‘From Neurons to Neighborhood’’ (Shonkoff and Phillips, 2000) and transition to school.
underscored the importance of early intervention by synthesizing Although it is widely recognized that school readiness has these
two areas of research that shows the power of environmental three dimensions, much of the research conducted in this area has
influences on early development. Brain development has been focused on children’s and families’ readiness for school. Research in
shown to be exquisitely attuned to environmental inputs that, in the area of assessment has produced heated debate on the domains
turn, shape its emerging architecture. Environment provided by and techniques of assessing children’s readiness for school (National
the child’s first caregivers has also been demonstrated to have Education Goals Panel, 1995; Meisels, 1999; La Paro and Pianta,
profound effects on every domain of child’s early development. The 2000). Numerous studies on parenting programs have highlighted
study suggests that early interventions that facilitate child’s the roles parents can play in children’s school readiness (Rouse et al.,
competence and his/her sense of personal efficacy have positive 2005; Aboud, 2007). However, there is scant research on a schools’
effects on later development. readiness for children, which sheds light on how educational
Although a majority of the evidence on the efficacy and institutions can modify the learning environment in order to better
effectiveness of early interventions comes from economically accommodate children’s diverse level of readiness. In this end, the
developed countries, increasingly, evaluation studies conducted in ready schools (Shore, 2000) was prepared to highlight the
economically less-developed countries have shown similar results. importance of ensuring that schools are ready for children, in
In Cape Verde and Guinea, children who participated in preschool addition to ensuring that children start school ready to learn. The
scored approximately half a standard deviation higher on cognitive report delineates the essential attributes of ‘‘ready school’’ in the
development at age five than children who did not participate in United States, with the transition period defined as kindergarten
preschool (Jaramillo and Tietjen, 2001). In Myanmar, the primary through Grade 3. More research and empirical data on the
school enrollment was 13% higher for children who had attended effectiveness of schools’ readiness is needed, especially in economic-
early childhood development (ECD) centers compared to those ally less-developed countries. The SRP builds upon some of the
who had not attended (Save the Children, 2004). In Nepal, the recommended characteristics of the ‘‘ready school.’’ The program’s
repetition rate in first grade for children who attended ECD centers curriculum and instruction is based on the recognition that self-
was 5.5%, which was one-fifteenth of the national norm (Save the esteem stems from competence, by doing tasks that are engaging
Children, 2003). and with instruction at the appropriate levels of pace and content.
The program also builds upon the strategy of giving teachers time to
2.2. Critical period improve their skills and knowledge. The 14-day teacher training
program to orient new teachers to the program and a regular
Entwisle and Alexander’s (1989) work on the ‘‘critical period’’ is monitoring regimen to support teachers in their implementation is
important in studying the efficacy of SRP, as they highlight that the one of the main components of SRP.
first-grade transition constitutes a ‘‘critical period’’ for children’s This study focuses on schools’ readiness for children by
academic and social development. They argue that events during evaluating a school-based intervention in Cambodia, which
the earliest years of school provide a ‘‘particular’’ kind of introduced a readiness course in the first two months of children’s
stimulation. This refers to the observation that primary schools formal education, Grade 1. By doing so, the study will contribute to
are socially organized in ways children have not previously filling a research gap in schools’ readiness, particularly in the
experienced. An unrelated adult (teacher) is in control, and context of economically less-developed countries, and provide
children meet and remain in often rigid classroom environments some practical implications for the modification of learning
for several hours each weekday. The presence of other children programs to support children’s smooth transition into primary
who are at about the same level of competence provides an school and to promote the learning of all children.
incentive to do well because human beings seek social approval. In
addition, human beings are strongly rewarded by positive 3. Context of Cambodia
reinforcement from an authority figure (i.e., a teacher). All these
‘‘particulars’’ provide a social context that is unique to this life 3.1. Demographic characteristics and economy
stage (Entwisle and Alexander, 1998). In other words, such early
learning opportunities provide a critical transition period from the Cambodia is a relatively small country of about 13,000,000
familiar home to a formal instructional setting, i.e., schools. people in Southeast Asia. The country consists of 24 provinces and
 Y. Nonoyama-Tarumi, K. Bredenberg / International Journal of Educational Development 29 (2009) 39–45 41

municipalities, which are broken down into discrete districts that moving away from traditionally high levels of dependency on
in turn consist of communes. The country’s economy is primarily textbook usage.
agricultural and 84% of the population lives in rural areas (National
Institute for Statistics, 2001). Cambodia’s age structure is rather 4. School readiness program
unique stemming from nearly two decades of constant warfare and
‘‘class cleansing.’’ The high mortality experienced among men and 4.1. Rationale
women (who are now in their 40s) during the 1970s and 1980s has
led to a depressed birth rate among these individuals resulting in The primary aim of school readiness curriculum is to provide a
an unusually smaller age cohort of individuals in the 20–24 age bridge between a child’s state of knowledge at the time of entry
range. Because this is a prime child bearing time for Cambodian into primary school and the Grade 1 curriculum, which in
women, the effects of war have in turn led to a smaller age cohort of Cambodia has been frequently criticized for being too academic.
children aged 0–4 during the current decade. Expected declines in The introduction of this curriculum, therefore, was based on the
school intake after 2004, therefore, present important opportu- assumption that the acquisition of readiness skills would enhance
nities within the education system to shift focus from expanding learning achievement when children encountered the Grade 1
physical capacity to more qualitative concerns such as the SRP. curriculum.
According to the World Food Program (WFP), about 38% of the The SRP was seen as a superior alternative to the reception class,
population lived below the poverty line in 2002, defined as which is another method widely used to provide readiness skills to
expenditure of less than $1 per day (WFP, 2002). Although the children entering primary school. The reception class differs from
country has recorded average GDP growth of about 5.5% during the SRP in that it usually occurs during the summer vacation, which
last decade, GDP per capita continues to be very low at about $300 poses three challenges. First, it requires extra payment to teachers.
per person (National Institute for Statistics, 2001). Low levels of Second, it requires teachers to take the additional load of work
economic development and inadequate means to distribute during the summer vacation, when many teachers are farmers and
national wealth from social elites to rural dwellers continue to are occupied with planting during this season in Cambodia. Third,
present a significant challenge to the government. it assumes that parents would be willing to send their children to
school during the summer, which is clearly not the case in most
3.2. Educational system situations. In addition to overcoming these challenges, SRP has the
advantage of being taught by the same teacher as the one who
Education in Cambodia continues to be the primary realm of the teaches the regular grade 1 class, ensuring that there is a greater
public sector with only 2.5% and 2.2% of preschools and primary continuity between what children learn under the readiness
schools operated privately, respectively (EFA Secretariat, 1999). program and Grade 1.
Funding provisions of the educational system greatly favor
primary education, which absorbed 37.4% of government and 4.2. School readiness program in Cambodia
donor educational investment at the end of the last decade. This is
followed by secondary education, vocational training, and tertiary The SRP implementation included various components, from
education. Preschool education absorbed only 1.1% of total the development of special curricular documentation, a 14-day
investment in 1999 (EFA Secretariat), a situation that has not teacher training program to orient new teachers to the program, a
changed substantially over the past several years. regular monitoring regimen to support teachers in their imple-
The most important change in the education system in recent mentation, physical upgrading of classrooms, to formalized
years has been the introduction in 2001 of a broadly based reform student assessment for monitoring and reporting purposes.
program that provides special funds for interventions designed to Although the program was implemented as a pilot, its initial
promote equity, quality, and efficiency of education financing. The scope was still significant for a small country like Cambodia. 544
reform program is implemented by the government as an Grade 1 teachers and approximately 25,000 children across three
integrated, sector-wide program that is reviewed annually with provinces participated in the pilot in 2004.
the stakeholders. In general, reform efforts have been highly In preparation for the implementation, multiple government
successful in increasing participation rates, particularly among the departments and advisors worked to develop a specialized
poorest quintiles of the population. The net enrollment rate for bridging curriculum that prepares children for eventual exposure
primary education was reported to be 91% in 2005, up from 85% in to the formal curriculum. Learning areas in SRP curriculum include
1996 (EMIS, 2006). Enrollment in pre-primary education has basic language skills (e.g., speaking, listening, and reading), the
doubled since the last decade, but gross enrollment rate is still concept of number, time and space, hygiene, working in groups,
estimated to be 10.6% (UNESCO-IBE, 2006). etc. This bridging curriculum consciously avoids many official
While educational reform has dramatically increased participa- curriculum competencies that tend to be highly premature in their
tion rates among children, efficiency and qualitative indicators presentation. The most conspicuous example is the omission of
continue to pose a major challenge for policy-makers. Dropout ‘‘writing,’’ which in the formal curriculum begins during the first
levels have remained static throughout the decade and repetition week of school. In general, the SRP curriculum was designed to
rates have been increasing, particularly in Grade 1 (EMIS, 2006). have a much stronger focus on both fine and gross motor skills. This
Government reports that national repetition rates at Grade 1 stands in contrast to traditional academic methods, which
reached 23.6% in 2004, whereas such rates were only 17.5% at the emphasize verbal and visual abstractions. The rationale underlying
beginning of the decade (Bredenberg, 2005). To a large extent, these approaches is that psychomotor-based learning modes are
these trends are results of reform efforts to increase enrollment at a especially suited for children in rural areas over abstract and picto-
time when capacity, particularly in the area of teacher numbers, verbal modes that have such prominence in existing textbooks.
has remained largely static. The result has been overcrowded Before the beginning of the school year, teachers received
classrooms, textbook shortages, an overtaxed teaching force, and intensive training in using the bridging curriculum as well as the
declining levels of instructional quality. In this context, the SRP has need for certain desired changes in classroom practice. For
been an important initiative to address quality issues as well as the example, teachers were asked to plan their instruction so that
textbook shortages, as SRP teaching methodologies emphasize each lesson was taught with at least two to four activities that used
42 Y. Nonoyama-Tarumi, K. Bredenberg / International Journal of Educational Development 29 (2009) 39–45

multiple sensory channels. Teachers acquired a repertoire of Table 1

Characteristics and sample size of experimental and control groups
numerous activities involving songs, role plays, drawing, games,
and other activities for teaching designated lessons. It was believed Background characteristics Experimental group Control group
that providing instruction in this way would prove to be more (school context)
Schools Students Schools Students
engaging for children, build their confidence, and provide positive
reinforcement for being in a school environment over a 4-h period Urban/rural
Urban 2 91 2 31
each day. The use of multiple activities also encouraged teachers to
Rural 8 382 8 427
rely less on textbooks, which was quite a novel experience to many
teachers. In order to facilitate these changes in classroom practice, SES
Less than 30% poor 2 91 2 31
physical classroom environments also underwent a certain 30–50% poor 4 191 3 191
amount of upgrading. This usually included the provision of 50% of more poor 4 191 5 236
copious amounts of stationery and raw materials for the
Total number of school/students 10 473 10 458
production of teaching aids as well as decorations to make
classroom environments more interesting.

It should be noted that dropout and absenteeism led to a loss of

5. Methods
test subjects during the year (n = 931 in pretest, n = 861 in posttest,
n = 851 in terminal test). At terminal achievement test, the
5.1. Research design
experimental group sample had dropped from 473 to 415, a loss
of 12% of the sample, and the control group sample had dropped
The design of this study is a comparison between an
from 458 to 436, a loss of 5%. The higher loss of subjects in the
experimental group, children who participated in the SRP, and a
experimental group, who are also more likely to be the most
control group, children who did not participate in the SRP. Given
vulnerable children, may have contributed to a potential spiraling
the challenges that emerge in implementing an experimental
up of mean scores for this group. This possibility suggests that
design in developing countries, the study is based on a none-
caution is needed in interpreting the results of this study. The
quivalent control group pretest–posttest design that introduces
proportion of boys and girls was approximately equal in each
both between-subjects and within-subjects comparisons.
group.
Analysis of covariance (ANCOVA) is used, with posttest and
terminal test scores as the response, treatment as the design factor,
5.3. Measures
and pretest score as a covariate, to determine whether the
perceived difference in mean test scores between the experimental
A number of formal assessments of SRP activities were
group and control group is statistically significant, controlling for
undertaken since the program was first piloted in 2004. First, a
pretest score differences. Multivariate analysis is also used to
pretest was administered before the school year to provide
control for imperfection in matching of the two groups.
baseline data. Second, a posttest was administered at the end of the
eight-week intervention period to determine the acquisition of
5.2. Sample
school readiness skills. Third, a terminal achievement test was
administered at the end of the school year to determine whether
Three provinces were chosen where there were a small handful
the SRP had any long-term effect. In all the assessments, children
of schools participating in SRP in each province.2 Ten schools
were tested in the core subjects of Khmer Language and
participating in the SRP program were randomly chosen from
Mathematics. For this study, we only report findings from Khmer
various districts in those three provinces in order to include both
Language, because the gap between children’s language skills at
rural and urban schools and represent both affluent and poor
the entry of primary school and the Grade 1 curriculum was seen to
communes. Another 10 schools, where SRP interventions had not
be the most problematic and of high concerns for policy-makers in
occurred, were assigned to the control group and matched to SRP
Cambodia.4 For both pretest and posttest, the selection of
schools according to demographic setting and poverty ratings3, as
competencies was guided by the SRP curriculum, whereas for
shown in Table 1. Thus, the total number of schools participating in
the terminal achievement test, the selection of competencies was
the sample was 20.The number of children selected in each school
guided by the formal Grade 1 curriculum. The assessment used an
was proportional to the school size. The sample was not matched
interview method, one proctor per student. Although this
by school size, which resulted in larger number of children from
assessment approach proved to be highly labor-intensive, it
rural and poor area in the control group. Therefore, in our analyses,
enabled the research team to avoid the use of written tests, which
we conduct multivariate analyses to control for these demographic
would have been highly inappropriate at the very young age of test
background differences between the treatment and control groups.
subjects.
The actual selection of children in each test site was not
Khmer Language covered 10 sub-topical areas outlined in the
determined by school staff, but by proctors using a simple random
curriculum (see Table 2). The breakdown of skill areas borrowed
sampling technique. These samples were constructed on site from
heavily from the typology developed by Bloom and his associates.
name lists provided by schools. These safeguards were put in place
The items consisted primarily of tasks requiring oral responses or
to minimize possible bias in the selection of children.
psychokinetic manipulations of test materials, such as letter cards.
2 These characteristics allowed the test content to be covered
It should be noted that there are several hundred schools in each province.
Therefore, the SRP schools and non-SRP schools were usually not in proximity and quickly. Test developers tried to formulate questions in a way that
we did not anticipate spill-over effects. both isolated specific skills and also minimized the confounding
3
The poverty rating was based on the National Census data, in which poverty was influence of associated skills that might block or hinder assessment
defined as expenditure of less than $1 per day. In order to assign schools to a poverty of the target area. For example, students were asked to spell out
ranking, three tiers were created which defined the number of people living in
poverty in a particular commune. Poverty tiers were designated as follows: Tier
4
1 = under 30% living in poverty; Tier 2 = 30–50% living poverty; 3 = over 50% living The main findings are similar for mathematics, although in general, the score for
in poverty. language was lower than mathematics for both groups.
 Y. Nonoyama-Tarumi, K. Bredenberg / International Journal of Educational Development 29 (2009) 39–45 43

Table 2
Test content specifications in Khmer Language

Topical area Sub-topical area Weighting in post-test (%) Weighting in terminal test (%)

Listening and speaking Counting syllabus

Syntax 28 28
Oral word usage

Reading Word meanings 33 32

Sound letter discrimination
Reading aloud
Reading comprehension

Writing Spelling 39 40
Writing words
Sentence composition

Table 3
Difference in language scores by topical areas between experimental group and control group

Experimental group Control group

Topical area Posttest (n = 473) Posttest (n = 430) Terminal test (n = 415) Pretest (n = 458) Posttest (n = 431) Terminal test (n = 436)

Listening and Speaking 10% (22%) 48% (37%) 40% (27%) 8% (17%) 23% (31%) 26% (21%)
Reading 5% (19%) 25% (39%) 49% (31%) 3% (12%) 20% (35%) 37% (28%)
Writing 5% (18%) 25% (35%) 24% (29%) 8% (20%) 15% (30%) 13% (21%)

Total 8% (18%) 40% (30%) 37% (26%) 6% (14%) 21% (29%) 24% (20%)

Note: Standard deviations are shown in parentheses.

words not by writing out the words but by arranging letter cards in week SRP course, the difference between the experimental group
meaningful strings. Thus, even if a child had limited writing skills, and control group varies substantially by topical areas, whereas
s/he could still arrange letters in meaningful groupings simply by the difference becomes similar across the three topical areas by the
manipulating the letter cards provided. end of the year. For example, the difference between the
The items were field-tested on a sample of children who had experimental group and control group is 25% for listening and
completed Grade 1 the previous year. The outcomes were analyzed speaking, 10% for writing, and 5% for reading in posttest, whereas
with respect to levels of difficulty and discrimination using the difference is 14%, 11%, and 12%, respectively in terminal test.
classical true score test theory models. The final tests retained The largest difference in listening and speaking in posttest may be
items of moderate difficulty and high discrimination. Tests were a reflection of the nature of the SRP programme with its de-
also reviewed in terms of administration time for each child, the emphasis of traditional pedagogy. The consistent difference across
clarity of directions, and the effectiveness of examples. topical areas in terminal test may suggest that the school readiness
Interviewers were recruited locally to administer subject tests skills, once acquired, can be beneficial across all topical areas of the
in each provincial site. All proctors received one-day training from formal curriculum. The smallest gain in writing also reflects the
the visiting research team in an explicit behavioral protocol to fact that the SRP curriculum purposefully avoids the inclusion of
ensure standardized testing conditions in each site. This protocol writing tasks to evade perceptual overload for children in the first
included guidelines on the set-up of test stations that were suitably two months of school. It appears, however, that the rigorous
separate from each other to prevent cuing, greetings to be used emphasis on writing skills in the official curriculum presents a
with children to put them at ease, guidelines that prohibited any serious disjuncture between SRP and what children must
commentary on student performance that might inhibit future eventually face when they are introduced to the Grade 1 textbook
responding, and other behaviors that might affect children’s ability later in the school year.
to answer to the best of their knowledge. Adherence to protocols Next, we conduct analysis of covariance (ANCOVA), with the
was monitored by spot checks during the testing by the research pretest as a covariate, to determine whether the perceived
team. Each subject test lasted for 15 min requiring a total of 30 min difference in mean test scores between the experimental group
of testing for each child. and control group is significant at p < .05, controlling for pretest

Table 4
6. Findings
Analysis of covariance (ANCOVA) of posttest scores and terminal test scores as a
function of treatment, with pretest scores as covariate
Overall, children in the experimental group significantly
Source d.f. MS F h2
outperformed children in the control group in the language test.
First, we eyeball the difference in the scores between the Posttest
experimental group and control group for each topical area Pretest (covariate) 1 10.93 142.85*** 0.13
Treatment 1 6.38 83.42*** 0.08
(Table 3).5 In all topical areas, the experimental group performed
Error 858 0.08
higher than the control group in both posttest and terminal test. In
the posttest, which was administered immediately after the eight- Terminal test
Pretest (covariate) 1 1.31 25.27*** 0.03
5
Treatment 1 3.10 59.75*** 0.06
We do not conduct a formal significance test by each topical area, because Error 848 0.05
multiple dependent variables may increase the probability of Type I error, and our
primary interest is to examine the difference in the overall language score. *p < .05; **p < .01; ***p < .001.
44 Y. Nonoyama-Tarumi, K. Bredenberg / International Journal of Educational Development 29 (2009) 39–45

Table 5 perilously low in ‘writing.’ The SRP curriculum excludes writing

Regression analyses for posttest scores and terminal test scores
tasks to avoid an overload for children at the beginning of their
Model 1: posttest Model 2: terminal test school life, but it needs to be highlighted that this might present a
(n = 858) (n = 849) disjuncture when they are eventually introduced to the Grade 1
B S.E. B S.E. textbook.
Although the findings described above clearly demonstrate a
Constant 0.19** (.068) 0.05 (.055)
long-term impact on learning among children who attend SRP
Treatment 0.17*** (.019) 0.11*** (.015)
Pretest 0.65*** (.057) 0.24*** (.048) classes, it is important to consider these findings in light of several
Male 0.00 (.011) 0.02 (.015) constraints in the way that this study unfolded. This refers to the
Age 0.04*** (.009) 0.03*** (.008) loss of more subjects in the experimental group than in the control
Medium SES 0.11*** (.020) 0.15*** (.024)
group as a result of unforeseen circumstances, which affected the
High SES 0.11*** (.029) 0.14*** (.016)
comparability of sample groupings in the terminal achievement
*p < .05; **p < .01; ***p < .001. tests. As noted earlier, the experimental group exhibited a loss of
subjects that was almost double of that of the control group. This
loss comprised of both dropouts and absences among students.
score differences (Table 4). The results show that the experimental There is a high probability that the students who were lost from the
group significantly outperforms the control group in posttest [F(1, sample probably conform to a national profile of children who are
858) = 83.42, p = .000]. Furthermore, this significant difference poor, vulnerable, and weaker academically. An examination of
remains after a year [F(1, 848) = 59.75, p = .000]. That is, the SRP is ‘lost’ sample subjects within the experimental group found that
not only successful in transmitting the school readiness skills as 66% came from two schools within the same province. These
outlined in the programme, but also enables students to use those schools were described by assisting program staff as poorly
skills to better understand and pursue the formal Grade 1 managed schools. Of the students who were absent in these
curriculum. Using eta square as the measure of effect size, the schools on the day of the test, about a quarter had dropped out
treatment accounted for 8% of total variability in posttest score, completely while the others were still enrolled but absent. Given
and 6% of total variability in terminal test score. the greater rate of loss in the experimental group and the
Lastly, to account for the imperfection of our sample, that is, the likelihood that many of these were indeed more vulnerable
sample not being matched by school size and resulting in larger children, this may have resulted in introducing sample bias,
number of children from rural and poor area in the control group, leading to some relative advantage for the experimental group.
we conduct multivariate analyses to control for these demographic Another constraint that needs to be taken into account is the
background differences. We employ an ordinary least squares school selection procedures that may have missed management
(OLS) analyses, and control for SES, age and gender, in addition to considerations. As noted above, research team members reported
pretest, to determine whether the abovementioned difference still that two of the schools included in the experimental group were
remains.6 Table 5 shows that the effect of treatment is statistically driven by internal conflicts leading to poor management of
significant in both posttest and terminal test. In other words, the teachers. Indeed, mean scores in these two schools were decidedly
SRP students outperform the non-SRP students in both posttest lower than the two schools in the control group with which they
and terminal test, controlling for key background differences were paired. Surprisingly, however, these problems did not appear
between the two groups. The adjusted R-square is .26 for Model 1 to affect overall experimental group performance in a way contrary
and .19 for Model 2; the larger percentage explained to predict the to what we had expected, and indeed may have compensated for
variance in posttest is expected due to the high correlation sample bias stemming from subject loss.
between pretest and posttest. The standardized coefficient of
treatment is .27 for Model 1 and .23 for terminal test. Although this 7.2. Policy implications
effect size is small,7 it is important to note that the effect is larger
than that of SES, and that the reduction in the effect size during the Research has repeatedly shown that children perform better in
year is minimal. schools when they participate in pre-primary education, but the
dearth of resources in many economically less-developed coun-
tries ensures that only a small minority of children has such
7. Conclusions opportunities. This situation contributes to poor school readiness
among children when they first enter school, and hinders their
7.1. Summary of findings smooth transition of the ‘critical period,’ which in turn leads to
high rates of grade repetition.
Overall, the findings show that children who received the SRP The findings from this study suggest that specialized interven-
intervention performed significantly higher than children who did tions which provide school readiness skills to children early in the
not receive the intervention in both school readiness skills and school year, enhances children’s learning performance in core
achievement of formal curriculum. Although the effect size was curriculum areas such as Language when they move into more
small, approximately a quarter of a standard deviation, once formalized schooling during their first year in school. School
controlling for pretest score and background variables, it is readiness interventions of this nature are to be distinguished from
important to emphasize that children maintained their learning reception classes in that they are embedded in the primary school
advantage after a year. Children who participated in the SRP cycle at the beginning of the school year (for a period of several
intervention scored better than their peers in the control group in weeks), with the framework of making schools ready for children
all topical areas of language, but the absolute scores were through emphasis on building children’s self-esteem, engaging
children in tasks, and instructing at appropriate levels of pace and
6
Since all the high SES schools were in urban area, only the SES variable was used content. In addition, the curricular design of such programs are
in our multivariate model.
7
Effect size at or above .5 standard deviation (S.D.) is considered as ‘‘large,’’ .3–.5
directly linked to the official curriculum so that they act as a bridge
S.D. as ‘‘moderate,’’ and .1–.3 S.D. as ‘‘small,’’ and those below.1 S.D. as ‘‘trivial’’ between children’s ‘before’ school experiences and the formal
(Rosenthal and Rosnow, 1984). Grade 1 curriculum, which in many countries, such as Cambodia,
 Y. Nonoyama-Tarumi, K. Bredenberg / International Journal of Educational Development 29 (2009) 39–45 45

tends to be somewhat academic in nature. The SRP interventions, Entwisle, D., Alexander, K., 1989. Early Schooling as a ‘‘Critical Period’’ Phenomenon.
In: Namboodiri, K., Corwin, R. (Eds.), Sociology of Education and Socialization.
however, should not be treated as a panacea for the repetition JAI, Greenwich, CT, pp. 27–55.
problem in school systems. In many countries, as in Cambodia, Entwisle, D., Alexander, K., 1998. Facilitating the transition to first grade: the nature
repetition stems from a complex interplay of many factors that of transition and research on factors affecting it. The Elementary School Journal
98 (4), 351–364.
include crowded classrooms, poor parental supervision, direct Jaramillo, A., Tietjen, K., 2001. Early childhood development in Africa: can we do
educational costs, teacher shortages, and distance to schools. more or less? In: A Look at the Impact and Implications of Preschools in Cape
Nevertheless, such interventions may be effective in enhancing Verde and Guinea, World Bank, Washington, D.C..
La Paro, K., Pianta, R., 2000. Predicting children’s competence in the early school
children’s learning potential by providing them with readiness
years: a meta-analytic review. Review of Educational Research 70 (4), 443–484.
skills as well as in addressing school quality issues by modifying Meisels, S., 1999. Assessing readiness. In: Pianta, R.C., Cox, M. (Eds.), The Transition
classroom practices. to Kindergarten: Research, Policy, Training, and Practice. Paul H. Brookes,
Baltimore.
Another important policy implication from this study is the
National Education Goals Panel, 1995. National Education Goals Report Executive
potential of SRP interventions to help bridge the gap in the Summary: Improving Education through Family-School-Community Partner-
provision of preschool opportunities. The SRP may be one way to ships. National Education Goals Panel, Washington, D.C.
compensate for children’s inability to participate in preschools or National Institute for Statistics, 2001. Cambodia Statistical Yearbook. Ministry of
Planning, Phnom Penh.
ECD centers before they start primary school, especially in societies Rosenthal, R., Rosnow, R., 1984. Essentials of Behavioral Research: Methods and
where the prevalence of pre-primary education is still extremely Data Analysis. McGraw-Hill, New York.
low. SRP interventions may, therefore, serve as a surrogate for Rouse, C., Brooks-Gunn, J., Mclanahan, S., 2005. Introducing the issue. The Future of
Children 15 (1), 5–14.
preschool attendance in some context. Save the Children., 2004. Early Childhood Care and Development: A Positive Impact,
Myanmar. Save the Children Myanmar Field Office, Yangon.
References Save the Children, 2003. What’s the Difference? The Impact of Early Childhood
Development Programme. A Study from Nepal of the Effects for Children,
Aboud, F., 2007. Evaluation of an early childhood parenting program in rural their Families and Communities. Save the Children Nepal Field Office, Kath-
Bangladesh. Journal of Health, Population and Nutrition 25 (1), 3–13. mandu.
Arnold, C., 2004. Positioning ECCD in the 21st century. Coordinators Notebook 28, Scott-Little, C., Kagan, S.L., Stebbins, Frelow, V., 2006. Conceptualization of readiness
1–34. and the content of early learning standards: The intersection of policy and
Bredenberg, K., 2000. Student Repetition in Cambodia: Causes, Consequences, and research? Early Childhood Research Quarterly 21 (2), 153–173.
its Relationship to Learning. UNICEF/Swedish International Development Shonkoff, J., Phillips, D. (Eds.), 2000. From Neurons to Neighborhoods: The Science
Agency, Phnom Penh. of Early Childhood Development. National Academy Press, Washington, D.C..
Bredenberg, K., 2004. School Readiness: Assessment Report. UNICEF/Swedish Inter- Shore, R., 2000. Ready Schools: A Report of the Goal 1 Ready Schools Resource
national Development Agency, Phnom Penh. Group. National Education Goals Panel, Washington, D.C.
Bredenberg, K., 2005. School Readiness Program Evaluation Follow-Up: Terminal UNESCO, 2007. EFA Global Monitoring Report 2008. Education for All by 2015—Will
Achievement Testing. UNICEF/Swedish International Development Agency, We Make It? UNESCO, Paris.
Phnom Penh. UNESCO-IBE, 2006. Country Profile Prepared for the EFA Global Monitoring Report
Educational Management and Information System, 2000–2006. Educational Sta- 2007. Strong Foundations: Early Childhood Care and Education. Cambodia:
tistics and Indicators. Ministry of Education, Youth, and Sport, Phnom Penh. Early Childhood Care and Education (ECCE) Programmes. UNESCO-IBE, Geneva.
EFA Secretariat, 1999. The Year 2000 Assessment. Ministry of Education, Youth, and World Food Program, 2002. Estimation of Poverty Rates at Commune Level in
Sport, Phnom Penh. Cambodia. Ministry of Planning.