Vous êtes sur la page 1sur 17

Int Rev Educ (2015) 61:173189

DOI 10.1007/s11159-015-9479-8

The privatepublic literacy divide amid educational


reform in Qatar: What does PISA tell us?

Jehanzeb R. Cheema

Published online: 4 March 2015


Springer Science+Business Media Dordrecht and UNESCO Institute for Lifelong Learning 2015

Abstract The education system in Qatar comprises of both private schools, which
receive money through student fees, and public schools, which are fully govern-
ment-funded. In the mid-2000s, Qatar started its transition towards an independent
school model with the aim of eventually converting all public schools into gov-
ernment-supported independent schools. The idea was to give public schools more
autonomy in terms of hiring decisions, adoption of curriculum and textbooks, and
budget spending, enabling them to emulate some of the private schools strategies
for turning out successful students. This study examines evidence from the 2006
2012 administrations of the Program for International Student Assessment (PISA) in
Qatar in order to evaluate whether or not recent educational reform efforts in this
country have succeeded in bridging the literacy divide between private and public
schools. The results, presented in a number of detailed tables and discussed in the
last part of the article, indicate that there is a significant difference in key literacy
skills between the two types of schools. Private schools were found to outperform
their public counterparts in areas such as mathematics, reading and science, both
before and after controlling for important student-level differences, and this gap has
evidently persisted from 2006 to 2012.

Keywords Public school  Private school  Qatar  Literacy gap  PISA 


ANCOVA

Resume Le fosse public/prive de lalphabetisation durant la reforme educative au


Qatar : que nous revele letude PISA ? Le systeme educatif du Qatar compte a la
fois des etablissements prives qui se financent par les frais de scolarite, et des ecoles
publiques entierement subventionnees par lEtat. Au milieu des annees 2000, le

J. R. Cheema (&)
College of Education, University of Illinois at Urbana-Champaign, 314N EDUC MC-708,
1310 South Sixth Street, Champaign, IL 61820, USA
e-mail: jrcheema@illinois.edu

123
174 J. R. Cheema

Qatar a entame une transition vers un modele detablissement independant dans le


but final de transformer toutes les ecoles publiques en etablissements autonomes
finances par lEtat. Lintention etait dattribuer aux ecoles publiques une plus
grande autonomie dans les decisions de recrutement, ladoption des programmes
detudes et manuels ainsi que dans lutilisation budgetaire, leur permettant ainsi de
sinspirer de certaines strategies des ecoles privees pour former de bons eleves. La
presente etude examine les donnees du Programme international pour le suivi des
acquis des eleves (PISA) applique au Qatar entre 2006 et 2012, en vue devaluer si
les recents efforts de reforme educative dans le pays sont parvenus a combler le
fosse de lalphabetisation entre les ecoles publiques et privees. Les resultats, pre-
sentes dans de nombreux tableaux detailles et analyses dans la derniere partie de
larticle, signalent une difference notable dans les competences de base entre les
deux types detablissement. Les ecoles privees sont plus performantes que leurs
homologues publiques dans les matieres telles que mathematiques, lecture et sci-
ences, a la fois avant et apres prise en compte dimportantes differences de niveaux
scolaires, et ce fosse sest a levidence maintenu entre 2006 et 2012.

Background

Qatar has seen massive reforms in its educational sector in the last decade. Termed
Education for a New Era (ENE), these reforms were initiated in 2002 at the behest
of the then Emir of Qatar, Sheikh Hamad Bin Khalifa Al Thani. The objectives of
ENE reforms include providing a greater variety in educational products offered to
students, more choice for parents in school selection, independence for new schools,
and implementation of a system of standards-based assessment and evaluation for
public schools (for Qatar see Constant et al. 2010 and Gonzalez et al. 2009; for the
Middle East in general see Shafiq 2011 and World Bank 2008).
The education system in Qatar comprises of both private and public schools.
Public schools are fully funded by the government while private schools primarily
support themselves through student fees and donations from benefactors. In the mid-
2000s Qatar started its transition towards an independent school model, similar in
function to the charter school system in the United States, with the aim of eventually
converting all public schools into government-supported independent schools. The
primary feature of such a system is that although the school receives funding from
the government, the school administration has significant control over school
policies including hiring decisions, adoption of curriculum and textbooks, and
budget spending. The rationale for this transition to the new system is that giving
autonomy to public schools allows them to emulate some of the successes of their
private counterparts (Brewer et al. 2007).
The privatepublic school performance debate is neither new nor is it specific to
Qatar. A large number of past studies have looked at private versus public school
differences across several countries with mixed results (Espinoza and Gonzalez
2013; Filer and Munich 2013; Friedman 1955; Kenayathulla 2013; Levy 2012;
Sandefur et al. 2013; Stitzlein 2013; Thapa 2013). Based on empirical findings,
these studies can be divided into two camps. There are those which attribute the

123
The privatepublic literacy divide amid educational reform in Qatar 175

privatepublic school performance differences to the ability of private schools to


manage resources relatively more efficiently than their public counterparts
(Caldwell 2010; Greenwald et al., 1996; Levy 2013). And there are those which
maintain that such differences are only a reflection of student-level characteristics;
they argue that once those characteristics are controlled for, the privatepublic
performance differences disappear (Braun et al. 2006; Carbonaro 2006; Lubienski
et al. 2008). Although both sides have presented strong arguments and empirical
evidence in support of their positions, the overall evidence is ambiguous. For
example, thorough reviews of relevant literature (McEwan 2000, 2004; Coulson
2009) have indicated mixed findings which failed to provide convincing support for
either side of the argument.
Notwithstanding the mechanism through which privatepublic performance
differences originate, there is little ambiguity about the fact that privatepublic
performance differences do exist. This makes it important from a policy perspective
to study and understand those differences, since they can provide valuable
information about the success and failure of educational reform efforts. It is in this
context that this study evaluates the changes which have taken place in the
education sector in Qatar over the past decade. Given the weaknesses in the
countrys pre-reform public education system, one objective of the reform efforts
was to improve the quality of public education available to the Qatari student
population. Thus, if there was evidence of an improvement over time in public
school performance relative to the performance of private schools, such evidence
could be an indication of the reforms success, regardless of whether such
improvement is due to increased competitiveness of public schools in resource
management or due to their ability to attract high-performing students. On the other
hand, the continued presence of a performance gap favouring private schools would
be indicative of government failure in providing education of the quality which is
otherwise available through private means in the country.
The case of Qatar is somewhat unique as compared to most countries in several
respects. First, Qatar has abundant proven reserves of natural gas and oil,
commodities which tend to have a continuous global demand and an oligopolistic
supply structure, and a very small population among whom to distribute the revenue
from these resources (Brewer et al. 2007; Gonzalez et al. 2009). These factors make
Qatar one of the richest countries in the world (GDP per capita at current prices:
USD 104,650 in 2013, USD 104,750 in 2012; [IMF 2013]). However, unlike most
other rich countries, especially those situated in Western Europe and North
America, Qatar has acquired this wealth within a period spanning decades rather
than centuries. Qataris who were herding cattle and were dependent on the fishing
industry for their livelihoods just two generations ago today have one of the highest
standards of living in the world (Brewer et al. 2007). Despite occasional efforts at
diversification, Qatars economy has remained almost totally dependent on export of
her natural resources until the turn of the century, with little investment in
alternative infrastructure. The export-fuelled economic growth outpaced and
eclipsed social and political developments to the extent that the education sector,
which in most countries serves as a driver for economic growth, took a back seat

123
176 J. R. Cheema

and never became a major priority for Qatari society until the early 2000s, when
ENE reforms were introduced (Zellman et al. 2009).
Second, the low priority allocated to the education system by the government of
Qatar until recently meant that Qatari students leaving the school system did not
have sufficient skills for them to prosper in higher education and to consequently fill
skilled positions in the labour sector (Brewer et al. 2007; Gonzalez et al. 2009). In
order to deal with this shortfall between demand and supply of skilled labour, and to
fill unskilled positions which Qataris no longer want, Qatar has traditionally relied
heavily on imported labour to such an extent that today only 1520 per cent of the
total population in this country is indigenous (Guarino and Tanner 2012; Khalaf and
Alkobaisi 1999; Modarres 2010). Although on the one hand this makes Qatar
heavily dependent on imported labour, on the other hand, the combination of a well-
developed domestic market for imported labour and an abundance of wealth means
that it is relatively easy for the country to hire some of the best policy makers,
administrators and analysts available to serve her education sector, thus filling
skilled roles and professions that Qatars indigenous labour force cannot handle
with outsiders who have completed their education in another country.
Third, the nature and intensity of educational reforms initiated by the Qatari
government clearly show the long-term resolve and seriousness of these efforts
(Berrebi et al. 2009; Guarino and Tanner 2012; Knight and Morshidi 2011). This
makes it a particularly interesting time to evaluate such changes in Qatar, because
students who were just entering primary school at the time the reforms were
initiated are now just about to exit the K-12 system.1 Examining the same target
population of secondary school students over this time period would allow an
evaluation of the privatepublic performance difference for (1) students who
entered the Qatari school system before the reforms were initiated and thus were not
affected by them, (2) students who were mid-way through their K-12 schooling
when the reforms were initiated and thus were somewhat affected by those reforms
and (3) students who had just entered or were about to enter the K-12 system at the
time the reforms were put into place and who were thus fully exposed to the reform
measures throughout their entire school career.
The main purpose of this study was to determine the magnitude of difference in
performance of private and public schools in Qatar in key literacy areas such as
mathematics, reading and science during the past decade after taking into
consideration external factors such as student-level characteristics, and whether or
not this difference had decreased over time. These differences were evaluated both
before and after controlling for student-level characteristics in order to determine
whether or not those differences persisted. A non-existent or decreasing private
public performance gap would be suggestive of a success of educational reform
efforts in Qatar, while a constant or increasing privatepublic performance gap
would be suggestive of their failure. The choice of key literacy areas as outcome

1
K-12 refers to kindergarten (K) followed by 12 years of primary and secondary school education. It is
common in the United States, Canada, South Korea, Turkey, the Philippines and Australia, and was
introduced in Qatar in 2002.

123
The privatepublic literacy divide amid educational reform in Qatar 177

criteria is important, since literacy reflects the long-term potential of a student


(assessment of what the student can do with his/her knowledge, i.e. practice and
application) as opposed to academic achievement (assessment of what a student has
learned, i.e. concepts and theory), which has only short-term relevance. It is a
students long-term potential in terms of practice and education which is related to
lifelong learning. As Paul Belanger (2009) notes, literacy takes its full meaning
within a lifelong perspective.
The specific research question being evaluated here is whether or not the private
public school performance gap has decreased in Qatar since the initiation of
educational reforms a decade ago. I am not aware of any other study which has
examined the effect of educational reforms on the privatepublic performance gap
over time in Qatar. The next section describes the sample and methodology used in
this study, followed by a presentation of the analytical results. The final section
states conclusions, discusses implications of the findings, and summarises the
limitations of this study along with making suggestions for future research.

Methodology

Sample and participants

The data for this study were obtained from three consecutive administrations of the
Program for International Student Assessment (PISA) between 2006 and 2012.
PISA is a cross-country survey with a three-year cycle that focuses on three key
areas of literacy: mathematics, reading and science. The target population for PISA
is the population of 15-year-old students in participating countries enrolled in high
school grades 7 and above. Although PISA has been in existence since 2000, it was
not possible to use data from 2000 and 2003 administrations, because Qatar did not
participate in those survey cycles. Country-level PISA samples are based on a
multistage stratified random selection of schools and students within each school.
This allows sample-based results to be generalised to the target population (OECD
2009, 2012, 2013). In contrast to this scheme, the Qatari sample is based on a
complete census. However, since these census data were only collected at specific
points in time, they may be treated as cross-sectional samples of the target
population at various points in time (namely 2006, 2009 and 2012) for which data
are available.
The numbers of students in the Qatari sub-samples of PISA 2006, 2009 and 2012
were 6,265, 9,078 and 10,966 respectively. These represented students from grades
712. However, for the purposes of this study I was interested only in students from
grades 9, 10 and 11, because these are the typical grades for 15-year-old students in
Qatar. Since the modal grade for this population is grade 10, the idea was to capture
students in the modal grade as well as advanced (grade 11) and retained (grade 9)
students for that grade. The numbers of students from grades 9, 10 and 11 were
5,771, 8,584 and 10,491 respectively for 2006, 2009 and 2012. After listwise
deletion of cases with missing information on variables used in this study, I was left
with 4,905 students from 2006 (sample attrition rate 15.0%), 7,702 students from

123
178 J. R. Cheema

2009 (sample attrition rate 10.3%) and 9,763 students from 2012 (sample attrition
rate 6.9%).

Variables

Dependent variables

In this study, the dependent variables were literacy scores in mathematics, reading
and science. These variables were item response theory (IRT)-generated plausible
value estimates of students true scores based on their performance on underlying
subject assessments. The literacy scores reported in the PISA datasets were scaled to
have a mean of 500 and standard deviation of 100 for OECD participants, since the
focus was not on OECD countries. In order to aid interpretation, the literacy scores
were rescaled to have a mean of 0 and standard deviation (SD) of 1 for the three
survey years included in this analysis. Although such standardisation is appropriate
within the scope of this study, it should be noted that these rescaled scores were only
suitable for making comparisons within a specific year and not across years.
Similarly, these rescaled scores were not appropriate for comparing Qatar with other
PISA participants.
The literacy score for each subject assessed in PISA is reported as set of five
plausible values. Each of these plausible values represents a random draw from the
probability distribution of all scores which may be attributed to a student. The idea
is similar to the multiple imputation method for imputing missing data with the
objective of modelling the uncertainty inherent in predicting a students true literacy
score (Wu 2005). Since averaging plausible values can artificially reduce variability
in parameter estimates, the analyst should either choose one of those values for
analysis, or repeat the calculations as many times as there are plausible values
(Brown and Mickleright 2004; von Davier et al. 2009). This study used the former
approach.
The rescaled literacy scores in the 2006 sample ranged between -3.17 and 4.51
(M = 0, SD = 1) for mathematics, between -3.10 and 4.02 for reading (M = 0,
SD = 1) and between -3.12 and 4.36 (M = 0, SD = 1) for science. In the 2009
sample, the standardised scores ranged between -3.06 and 4.02 for mathematics
(M = 0, SD = 1), between -3.21 and 3.84 (M = 0, SD = 1) for reading and between
-3.08 and 3.98 (M = 0, SD = 1) for science. In the 2012 sample, the standardised
scores ranged between -3.50 and 3.81 for mathematics (M = 0, SD = 1), between
-3.41 and 3.84 (M = 0, SD = 1) for reading and between -5.10 and 2.69
(M = 0, SD = 1) for science.

Independent variables

School type. This is a nominal variable which takes a value of 0 for private schools
and 1 otherwise. The designation of a school as private or public was based on a
survey question which defined a public school as a school managed directly or
indirectly by a government authority or entity, and a private school as a school

123
The privatepublic literacy divide amid educational reform in Qatar 179

Table 1 Descriptive statistics for PISA 2006: standardised literacy scores in mathematics, reading and
science, and socioeconomic status, by gender, grade and immigrant status, cross-tabulated by school type
Factor n Mathematics Reading Science SES

M SD M SD M SD M SD

a. Private schools
Gender
Boys 509 0.83 1.32 0.44 1.30 0.66 1.36 0.50 0.68
Girls 247 1.14 1.11 1.30 1.00 1.26 1.07 0.39 0.62
Grade
9 64 0.51 1.31 0.29 1.50 0.44 1.42 0.28 0.81
10 486 0.90 1.26 0.65 1.26 0.79 1.27 0.51 0.64
11 206 1.16 1.23 1.02 1.18 1.13 1.30 0.44 0.67
Immigrant status
Native 280 0.26 1.11 *0 1.13 0.06 1.07 0.56 0.73
Second generation 146 0.86 1.09 0.60 1.13 0.75 1.13 0.30 0.72
First generation 330 1.53 1.17 1.39 1.09 1.57 1.14 0.46 0.56
b. Public schools
Gender
Boys 1,660 -0.30 0.89 -0.51 0.80 -0.41 0.82 -0.15 1.01
Girls 2,489 -0.08 0.79 0.12 0.84 0.01 0.82 -0.04 1.03
Grade
9 795 -0.38 0.75 -0.37 0.83 -0.34 0.78 -0.54 1.04
10 2,795 -0.19 0.82 -0.14 0.86 -0.18 0.83 -0.04 1.01
11 559 0.21 0.92 0.27 0.92 0.22 0.89 0.32 0.86
Immigrant status
Native 2,584 -0.30 0.81 -0.26 0.84 -0.30 0.80 -0.02 1.07
Second generation 973 *0 0.82 -0.03 0.88 *0 0.84 -0.30 0.98
First generation 592 1.12 0.85 0.26 0.91 0.20 0.89 -0.03 0.98

Note n = 4,905

managed directly or indirectly by any organisation other than the government


(OECD 2011).
Demographic controls. The study used several control variables to account for
student-level differences and in order to isolate the effect of school type on literacy
scores. These variables include gender, grade, immigrant status and socioeconomic
status. Gender is a nominal variable with two categories, male and female. Grade is
an ordinal variable with three levels, 9, 10 and 11. Immigrant status categorises a
student as either native Qatari (for students born in Qatar and with at least one
parent also born in Qatar), second-generation Qatari (for students born in Qatar but
whose parents were both born outside Qatar), and first-generation Qatari (for
students born outside Qatar with both parents also born outside Qatar). Socioeco-
nomic status is a PISA-reported measure of a students economic and sociocultural
status based on factors such as parental education, parental income, family wealth,

123
180 J. R. Cheema

Table 2 Descriptive statistics for PISA 2009: standardised literacy scores in mathematics, reading and
science, and socioeconomic status, by gender, grade and immigrant status, cross-tabulated by school type
Factor n Mathematics Reading Science SES

M SD M SD M SD M SD

a. Private schools
Gender
Boys 1,140 0.95 1.06 0.65 0.99 0.82 1.07 0.34 0.70
Girls 1100 0.82 0.94 0.98 0.84 0.91 0.90 0.33 0.66
Grade
9 312 0.51 0.99 0.44 0.95 0.44 0.97 0.17 0.79
10 1,295 0.90 0.99 0.82 0.92 0.88 0.97 0.34 0.65
11 633 1.04 0.97 0.97 0.91 1.04 0.99 0.41 0.66
Immigrant status
Native 445 0.37 1.00 0.31 0.99 0.34 1.00 0.59 0.77
Second generation 480 0.72 0.91 0.70 0.86 0.73 0.90 0.18 0.67
First generation 1,315 1.12 0.95 1.01 0.86 1.09 0.94 0.31 0.63
b. Public schools
Gender
Boys 2,303 -0.38 0.78 -0.58 0.78 -0.51 0.75 -0.13 1.07
Girls 3,159 -0.35 0.71 -0.15 0.80 -0.24 0.74 -0.14 1.08
Grade
9 796 -0.56 0.76 -0.50 0.83 -0.50 0.75 -0.67 1.12
10 3,828 -0.37 0.72 -0.33 0.81 -0.36 0.75 -0.11 1.05
11 838 -0.16 0.76 -0.20 0.84 -0.20 0.78 0.24 0.96
Immigrant status
Native 3,633 -0.47 0.71 -0.48 0.78 -0.49 0.72 -0.07 1.11
Second generation 1,080 -0.26 0.71 -0.15 0.80 -0.20 0.74 -0.42 1.04
First generation 749 *0 0.79 0.12 0.82 0.07 0.79 -0.06 0.84

Note n = 7,702

cultural resources, and availability of educational resources at home (OECD 2009,


2012, 2013). Standardised socioeconomic scores ranged between -3.91 and 2.42
(M = 0, SD = 1) in the 2006 sample, between -4.22 and 2.55 (M = 0, SD = 1) in
the 2009 sample, and between -5.10 and 2.69 (M = 0, SD = 1) in the 2012
sample.

Analytical method

The study used both bivariate and multivariate methods in order to examine the
relationship between literacy scores and school type. For bivariate analysis,
independent sample t-tests were used in order to test for the mean difference
between private and public schools separately for each of the three literacy
variables. In order to see whether or not these differences persisted after controlling
for student-level characteristics I used analysis of covariance (ANCOVA) models

123
The privatepublic literacy divide amid educational reform in Qatar 181

Table 3 Descriptive Statistics for PISA 2012: standardised literacy scores in mathematics, reading and
science, and socioeconomic status, by gender, grade and immigrant status, cross-tabulated by school type
Factor n Mathematics Reading Science SES

M SD M SD M SD M SD

a. Private schools
Gender
Boys 2,070 0.61 1.02 0.31 0.97 0.52 1.00 0.23 0.78
Girls 1,712 0.77 0.88 0.85 0.83 0.83 0.87 0.18 0.75
Grade
9 652 0.30 0.83 0.26 0.93 0.33 0.87 -0.04 0.84
10 2,310 0.68 0.94 0.56 0.93 0.66 0.94 0.22 0.74
11 820 0.99 1.01 0.77 0.95 0.93 0.99 0.39 0.73
Immigrant status
Native 659 0.12 0.93 *0 0.99 0.12 0.96 0.49 0.86
Second generation 614 0.58 0.89 0.45 0.87 0.50 0.90 0.09 0.89
First generation 2,509 0.86 0.92 0.73 0.90 0.84 0.91 0.17 0.69
b. Public schools
Gender
Boys 2,766 -0.53 0.80 -0.73 0.86 -0.62 0.79 -0.13 1.11
Girls 3,215 -0.35 0.70 -0.02 0.72 -0.24 0.72 -0.14 1.10
Grade
9 711 -0.59 0.73 -0.58 0.85 -0.48 0.75 -0.69 1.16
10 4,294 -0.45 0.74 -0.36 0.85 -0.45 0.77 -0.10 1.07
11 976 -0.24 0.78 -0.15 0.87 -0.21 0.79 0.11 1.08
Immigrant status
Native 4,029 -0.58 0.70 -0.49 0.83 -0.54 0.74 -0.09 1.13
Second generation 1,077 -0.22 0.73 -0.14 0.81 -0.28 0.76 -0.40 1.11
First generation 875 -0.03 0.81 0.02 0.89 -0.04 0.83 *0 0.89

Note n = 9,763

with literacy scores as dependent variables, school type, gender, grade and
immigrant status as fixed factors, and socioeconomic status as covariate. Pearson
product moment coefficients of correlation were also computed, in order to examine
the association between socioeconomic status and the three literacy scores. The data
were not pooled across the three years, because the individuals in each survey year
would then have been related. This would have violated one of the key assumptions
of ordinary least squares (OLS) regression, which is that the error terms should be
independent of each other.
R-squared was used as a measure of proportion of total variation in literacy
scores explained by each multiple regression model, and partial eta squared as a
measure of proportion of unique contribution of each independent variable to the
total explained variation. All t-test and ANCOVA assumptions were evaluated, and
standardised residuals from each estimated model were inspected. SPSS 20.0

123
182 J. R. Cheema

Table 4 Pearson productmoment coefficients of correlation for literacy scores in mathematics, reading
and science, and socioeconomic status; PISA 20062012
Survey year Variable r*

1 2 3 4

2006 1. Literacy in mathematics


2. Literacy in reading 0.74 (L)
3. Literacy in science 0.80 (L) 0.86 (L)
4. Socioeconomic status 0.22 (M) 0.14 (S) 0.15 (S) -
2009 1. Literacy in mathematics
2. Literacy in reading 0.84 (L)
3. Literacy in science 0.86 (L) 0.91 (L)
4. Socioeconomic status 0.27 (M) 0.20 (S) 0.22 (M)
2012 1. Literacy in mathematics
2. Literacy in reading 0.84 (L)
3. Literacy in science 0.89 (L) 0.84 (L)
4. Socioeconomic status 0.22 (M) 0.21 (M) 0.22 (M)

Note n = 4,905, 7,702 and 9,763 for 2006, 2009 and 2012 respectively. Cohens (1992) effect size
interpretation in parentheses: S = small, M = medium, L = large
*p \ .001 for all reported correlations

software was used for all statistical computations, tests of hypotheses were
evaluated at 5 per cent level of significance, normalised sampling weights were
used, and effect size interpretations were based on the method devised by Jacob
Cohen (1992).

Results

Bivariate results

The descriptive statistics for literacy, gender, grade, immigrant status and
socioeconomic status are presented by school type in Tables 1, 2 and 3. For all
areas of literacy these statistics show, with few exceptions, a general trend of higher
mean scores for private schools (as compared to public schools), girls (as compared
to boys), first-generation students (as compared to second-generation students),
second-generation students (as compared to native students), grade 11 students (as
compared to grade 10 students) and grade 10 students (as compared to grade 9
students). The correlation matrix for literacy scores and socioeconomic status is
presented in Table 4. For each survey year, the correlation coefficients suggest
strong associations among the three literacy scores, and weak to medium
correlations between socioeconomic status and literacy scores.
Results from the independent sample t-tests are presented in Table 5. These
results suggest that for the 2006 sample, there was a significant difference in mean

123
The privatepublic literacy divide amid educational reform in Qatar 183

Table 5 Independent sample t-test results for mean difference between private and public schools for
literacy scores in mathematics, reading and science; PISA 20062012
Survey Literacy score t* df DM SE(DM) 95% CI Effect size

year Lower Upper d Interpretation

2006 Mathematics 23.78 996.51 1.08 0.05 0.99 1.17 1.16 Large
Reading 17.77 1005.06 0.82 0.05 0.73 0.91 0.86 Large
Science 21.20 986.92 0.99 0.05 0.90 1.08 1.05 Large
2009 Mathematics 54.11 3596.47 1.24 0.02 1.19 1.28 1.50 Large
Reading 50.97 4105.21 1.13 0.02 1.09 1.18 1.32 Large
Science 52.77 3675.00 1.21 0.02 1.16 1.25 1.44 Large
2012 Mathematics 60.79 6683.22 1.12 0.02 1.08 1.15 1.33 Large
Reading 47.66 7501.61 0.91 0.02 0.87 0.94 1.01 Large
Science 58.25 6880.63 1.08 0.02 1.04 1.11 1.27 Large

Note n = 4,905, 7,702 and 9763 for 2006, 2009 and 2012 respectively. n may not correspond with reported
df values due to t-test adjustments for unequal variances. Effect size interpretation based on Cohen (1995)
*p \ .001 for all reported t values

literacy between private and public schools in mathematics, reading and science,
p\.001. In the 2009 and 2012 samples, there were similar significant differences in
mean literacy between private and public schools, p \ .001. Cohens d indicated a
large effect of school type on mean differences in literacy across the three areas and
for all three survey years.

ANCOVA results

ANCOVA results are presented in Table 6, with corresponding parameter estimates


presented in Table 7. For the 2006 sample, ANCOVA results indicated a significant
effect of school type, gender, grade, immigrant status and socioeconomic status on
literacy scores in mathematics, reading and science. Based on eta-squared estimates,
the largest contribution to total explained variation came from school type and
immigrant status for maths scores; gender and immigrant status for reading scores;
and immigrant status and school type for science scores. Parameter estimates from
ANCOVA models suggested that as compared to public schools the mean literacy in
private schools was 0.82 standard deviations higher for mathematics, 0.69 standard
deviations higher for reading, and 0.80 standard deviations higher for science,
p \ .001.
In addition to the main effect of school type, the effect of gender was significant
with girls outperforming boys in mathematics by 0.24 standard deviations, in
reading by 0.68 standard deviations, and in science by 0.46 standard deviations,
p\.001; the effect of grade was significant with students in grade 11 outperforming
those in grade 9 and grade 10 in all three areas of literacy, p \ .001; the effect of
immigrant status was significant with second-generation students outperforming
native students and first-generation students in all three areas of literacy, p \ .001;
and the effect of socioeconomic status was significant with an increase of one

123
184 J. R. Cheema

Table 6 ANCOVA results for literacy scores in mathematics, reading and science, PISA 20062012a
Survey Source of variation Mathematics Reading Science

year F g2p F g2p F g2p

2006 School type 527.99* .09 403.92* .07 519.07* .09


Gender 91.86* .02 799.45* .14 354.08* .07
Grade 63.01* .02 81.46* .03 69.67* .03
Immigrant status 212.51* .08 302.52* .11 314.87* .11
Socioeconomic status 107.31* .02 18.45* *0 27.92* .01
2009 School type 1458.25* .16 1213.39* .13 1406.61* .15
Gender 0.12 *0 572.26* .07 175.17* .02
Grade 67.95* .02 53.65* .01 63.11* .02
Immigrant status 311.65* .07 418.68* .10 394.64* .09
Socioeconomic status 249.58* .03 89.83* .01 115.25* .01
2012 School type 1440.71* .13 910.46* .09 1392.51* .12
Gender 99.60* .01 1614.41* .14 476.40* .05
Grade 156.40* .03 109.47* .02 106.43* .02
Immigrant status 528.25* .10 460.98* .09 425.66* .08
Socioeconomic status 251.12* .03 294.41* .03 248.37* .02

Note n = 4,905, 7,702 and 9,763 for 2006, 2009 and 2012 respectively
*p \ .001
a
ANCOVA R2 values were .26, .40, and .40 for mathematics; .31, .39 and .39 for reading; and .29, .40
and .39 for science, for survey years 2006, 2009 and 2012 respectively

standard deviation in socioeconomic status associated with an increase of 0.13


standard deviations in mathematics, 0.05 standard deviations in reading literacy, and
.07 standard deviations in science literacy, p \ .001.
For survey years 2009 and 2012, the pattern of parameter estimates and their
significance in the ANCOVA models was similar to that for 2006. The only
exception was the effect of gender on mathematics literacy scores in 2009, which
was statistically not different from zero, |t| = 0.34, p = .731. R2 values from
ANCOVA models for 2009 and 2012 were respectively .40 and .40 for
mathematics; .39 and .39 for reading and .40 and .39 for science, with school
type, immigrant status and gender generally being the largest contributors to R2
values.

Discussion

Over the past decade, the education sector in Qatar has witnessed massive
government-initiated reforms directed at public schools. These include a complete
overhaul of assessment and evaluation policies at the K-12 level, and conversion of

123
The privatepublic literacy divide amid educational reform in Qatar 185

Table 7 Parameter estimates from ANCOVA for literacy scores in mathematics, reading and science,
PISA 20062012
Survey Source of variation Mathematics Reading Science

year b t b t b t

2006 School type -0.82 -22.98* -0.69 -20.10* -0.80 -22.78*


Gender 0.24 9.58* 0.68 28.27* 0.46 18.82*
Grade (9) -0.46 -10.50* -0.50 -11.93* -0.46 -10.79*
Grade (10) -0.33 -9.80* -0.36 -11.15* -0.35 -10.61*
Immigrant status (native) -0.64 -19.11* -0.78 -24.15* -0.78 -24.09*
Immigrant status (first generation) -0.26 -6.70* -0.46 -12.34* -0.39 -10.50*
Socioeconomic status 0.13 10.36* 0.05 4.30* .07 5.28*
2009 School type -0.87 -38.19* -0.80 -34.83* -0.86 -37.50*
Gender *0 0.34 0.43 23.92* 0.23 13.24*
Grade (9) -0.37 -11.64* -0.33 -10.35* -0.35 -11.21*
Grade (10) -0.17 -7.18* -0.15 -6.38* -0.16 -7.13*
Immigrant status (native) -0.59 -24.75* -0.68 -28.03* -0.66 -27.50*
Immigrant status (first generation) -0.31 -11.54* -0.28 -10.22* -0.30 -11.10*
Socioeconomic status 0.15 15.80* 0.09 9.48* 0.10 10.73*
2012 School type -0.75 -37.96* -0.60 -30.17* -0.74 -37.32*
Gender 0.16 9.98* 0.64 40.18* 0.35 21.83*
Grade (9) -0.49 -17.39* -0.42 -14.77* -0.39 -13.58*
Grade (10) -0.26 -12.76* -0.19 -9.28* -0.26 -12.25*
Immigrant status (native) -0.66 -31.47* -0.62 -29.11* -0.61 -28.86*
Immigrant status (first generation) -0.22 -9.13* -0.18 -7.47* -0.27 -11.07*
Socioeconomic status 0.13 15.85* 0.14 17.16* 0.13 15.76*

Note n = 4,905, 7,702 and 9,763 for 2006, 2009 and 2012 respectively
*p \ .001
a
Reference category was private for school type, male for gender, 11 for grade and second generation for
immigrant status

public schools into independent entities similar in function to charter schools in the
United States. The aim of these reform efforts is to improve academic achievement,
to provide students with a holistic educational experience, and to raise the profile of
Qatar in international standardised assessments. I examined evidence from the
20062012 administrations of the Program for International Student Assessment
(PISA) in order to evaluate whether or not these reform efforts have succeeded in
bridging the literacy divide between private and public schools in the country.
The statistical results indicate a significant difference in literacy between the two
types of schools, with private schools outperforming their public counterparts in key
areas such as mathematics, reading and science, both before and after controlling for
important student-level differences. These differences were observed for all three

123
186 J. R. Cheema

PISA assessments between 2006 and 2012. The persistence of this privatepublic
literacy gap suggests that Qatar still has a long way to go before meeting the
objectives of its educational reform efforts.
The findings have several implications. First, although a privatepublic literacy
gap can be seen in Qatar, the size of the gap does not seem to be changing
significantly. The magnitude of literacy gaps in mathematics, reading and science
increased from their initial values in 2006 to slightly higher values in 2009, before
falling back in 2012 to values lower than their starting values of 2006. Although the
decrease in the privatepublic literacy gap between 2009 and 2012 has been large
(0.20 SD for reading; 0.12 SD for mathematics and science), the change in
magnitude between 2006 and 2012, and between 2006 and 2009 has remained
modest. This gives the overall impression that if the privatepublic literacy gap is
not definitely shrinking, then it is not definitely expanding either. This is still bad
news for the Qatari government considering that a major objective of public
education reform efforts is to improve the standard of public education and to ensure
that all students have access to the best quality of education available in the country.
The existence of a literacy gap in favour of private schools suggests that the Qatari
government is currently unable to offer education of the quality that is available
through private enterprise in the country.
Second, the results indicate that the effect of school type varies across the three
literacy areas, with stronger effects observed for mathematics and science, and a
relatively weaker effect observed for reading. This suggests that although private
schools enjoy a comparative advantage in all three literacy areas over their public
counterparts, this advantage is relatively stronger in mathematics and science as
compared to reading. There can be any number of reasons for this. Without
endorsing a particular reason, possibilities such as private schools ability to hire
relatively more qualified teachers, the superiority of their curriculum, or the
availability of a better infrastructure seem plausible.
Third, student-level variables such as gender, grade, immigrant status and
socioeconomic status had a significant effect on literacy. In terms of grade, after
controlling for the effect of all other variables, there was a universal positive
association between grade level and literacy score in all survey years for
mathematics, reading and science. In general, after controlling for the effect of
all other variables, girls outperformed boys in all three areas of literacy. However,
the gender literacy gap was relatively more pronounced in reading, as compared to
mathematics and science.
The partial effect of immigrant status on literacy was also significant, with the
average performance of first-generation students being the highest, followed by
second-generation students and then native students. Although it is known that in
Qatar native students tend to have lower scores as compared to their first- or second-
generation counterparts (Cheema 2014), what is interesting to note here is that the
effect of immigrant status persists even after controlling for the effect of school
type. Most immigrants in Qatar tend to attend a private school, and since students in
private schools tend to perform relatively better than students in public schools, the
effect of immigrant status on literacy is expected. However, if this difference is only
due to a relatively higher proportion of immigrant students attending private

123
The privatepublic literacy divide amid educational reform in Qatar 187

schools, then it should disappear once we control for school type. The persistent
effect of immigrant status on literacy even after controlling for school type and
other variables suggests that superior performance of non-native Qatari students is
due to reasons other than the quality of the education they receive. Prior research
has suggested that lower performance of native students could be due to the very
generous welfare system offered by the Qatari government to its citizens, including
guaranteed public-sector employment as soon as they complete their education.
Such incentives could be responsible for lowering competition among native
students and for negatively affecting their motivation to do better at school as
compared to their non-native counterparts (Berrebi et al. 2009; Dargin 2007).
Finally, the partial effect of socioeconomic status on literacy was significant but
small with this variable accounting for 13 per cent of the total variation in literacy,
suggesting that once variables such as school type, gender, grade and immigrant
status are accounted for, socioeconomic status tends to have a relatively small effect
on literacy scores of students.
Although this study revealed several significant results, it has some limitations.
The results of the statistical analyses presented in this study are generalisable only
to the population of 15-year-old students enrolled in grades 9, 10 and 11 in private
and public schools in Qatar. Furthermore, some variables such as socioeconomic
status and immigrant status were not directly observed but were based on voluntary
student responses. Since no one has any control over the accuracy of what students
chose to report on the survey, these analytical results are reliable only to the extent
to which the underlying survey information is accurate. Another limitation of this
data source is that it has missing data on this studys particular variables of interest,
even though the proportion of such missing data in any given year remains at 15 per
cent or less. Despite my careful examination of the data, I found little evidence to
suggest that reported attrition rates may have generated bias in parameter estimates
or their standard errors under missing at random (MAR) and missing completely at
random (MCAR) assumptions; the data may be missing not at random (NMAR), a
possibility which is always present no matter how large a dataset is. However, little
can be done about this issue, because (1) the data have already been collected and
(2) identification of the exact nature of missingness is unlikely because missing data
are not observed in the first place. We further note that the missing data in this
sample are mostly due to missing values on socioeconomic status and immigrant
status, and not due to missing data on school type or literacy scores. Finally, since
this study examined literacy scores only in mathematics, reading and science, its
results should not be generalised to other areas of literacy. Researchers interested in
this study are encouraged to replicate this method for additional areas of literacy,
and for grades and age groups other than those examined in this study.

References

Belanger, P. (2009). Literacy and lifelong learning. In P. Jarvis (ed.), The Routledge international
handbook of lifelong learning (pp. 8090). New York, NY: Routledge.

123
188 J. R. Cheema

Berrebi, C., Martorell, F., & Tanner, J. (2009). Qatars labor markets at a crucial crossroad. The Middle
East Journal, 63(3), 421442.
Braun, H., Jenkins, F., & Grigg, W. (2006). Comparing private schools and public schools using
hierarchical linear modeling (NCES 2006461). U.S. Department of Education, National Center for
Education Statistics, Institute of Education Sciences. Washington, DC: U.S. Government Printing
Office.
Brewer, D., Augustine, C., Zellman, G., Ryan, G., Goldman, C., Satsz, C., & Constant, L. (2007).
Education for a new era: Design and implementation of K12 education reform in Qatar. Santa
Monica, CA: RAND Corporation.
Brown, G., & Micklewright, J. (2004). Using international surveys of achievement and literacy: A view
from the outside. Montreal, Canada: UNESCO Institute for Statistics.
Caldwell, B. (2010). Is private schooling becoming the preferred model of school choice in Australia?
Journal of School Choice, 4(4), 378397.
Carbonaro, W. (2006). Publicprivate differences in achievement among kindergarten students:
Differences in learning opportunities and student outcomes. American Journal of Education,
113(1), 3165.
Cheema, J. (2014). The migrant effect: An evaluation of native academic performance in Qatar. Research
in Education, 91, 6577.
Cohen, J. (1992). Quantitative methods in Psychology: A power primer. Psychological Bulletin, 112(1),
155159.
Constant, L., Goldman, C.A., Zellman, G.L., Augustine, C.H., Galama, T., Gonzalez, G., Guarino, C.A.,
Karam, R., Ryan G.W., & Salem, H. (2010). Promoting quality and variety through the public
financing of privately operated schools in Qatar. Journal of School Choice, 4(4), 450473.
Coulson, A. (2009). Comparing public, private, and market schools: The international evidence. Journal
of School Choice, 3(1), 3154.
Dargin, J. (2007). Qatars natural gas: The foreign-policy driver. Middle East Policy, 14(3), 136142.
Espinoza, O., & Gonzalez, L. E. (2013). Access to higher education in Chile: A public vs. private
analysis. Prospects, 43(2), 199214.
Filer, R., & Munich, D. (2013). Responses of private and public schools to voucher funding. Economics of
Education Review, 34, 269285.
Friedman, M. (1955). The role of government in education. In R. Solo (ed.), Economics and the public
interest. New Brunswick, NJ: Rutgers University Press.
Gonzalez, G., Le, V., Broer, M., Mariano, L., Froemel, J., Goldman, C., & DaVanzo, J. (2009). Lessons
from the field: Developing and implementing the Qatar student assessment system, 20022006.
Santa Monica, CA: RAND Corporation.
Greenwald, R., Hedges, L., & Laine, R. (1996). The effect of school resources on student achievement.
Review of Educational Research, 66(3), 361396.
Guarino, C., & Tanner, J. (2012). Adequacy, accountability, autonomy and equity in a Middle Eastern
school reform: The case of Qatar. International Review of Education, 58(2), 221245.
IMF (International Monetary Fund) (2013). Transitions and tensions. World economic outlook, October
2013. Washington, DC. International Monetary Fund.
Kenayathulla, H. (2013). Higher levels of education for higher private returns: New evidence from
Malaysia. International Journal of Educational Development, 33(4), 380393.
Khalaf, S., & Alkobaisi, S. (1999). Migrants strategies of coping and patterns of accommodation in the
oil-rich gulf societies: Evidence from the UAE. British Journal of Middle Eastern Studies, 26(2),
271298.
Knight, J., & Morshidi, S. (2011). The complexities and challenges of regional education hubs: Focus on
Malaysia. Higher Education, 62(5), 593606.
Levy, D. (2012). How important is private higher education in Europe? A regional analysis in global
context. European Journal of Education, 47(2), 178197.
Levy, D. (2013). The decline of private higher education. Higher Education Policy, 26, 2542.
Lubienski, S., Lubienski, C., & Crane, C. (2008). Achievement differences and school type: The role of
school climate, teacher certification, and instruction. American Journal of Education, 115(1),
97138.
McEwan, P. (2000). The potential impact of large-scale voucher programs. Review of Educational
Research, 70(2), 103149.
McEwan, P. (2004). The potential impact of vouchers. Peabody Journal of Education, 79(3), 5780.

123
The privatepublic literacy divide amid educational reform in Qatar 189

Modarres, Ali. (2010). Migration and the Persian Gulf: Demography, identity and the road to equitable
policies. Anthropology of the Middle East, 5(1), 117.
OECD (Organisation for Economic Co-operation and Development) (2009). PISA 2006 technical report.
Paris, France: OECD Publishing.
OECD (2011). School questionnaire for PISA 2012. Paris, France: OECD Publishing.
OECD (2012). PISA 2009 technical report. Paris, France: OECD Publishing.
OECD (2013). PISA 2012 assessment and analytical framework. Paris, France: OECD Publishing.
Sandefur, J., Watkins, K., & Green, D. (2013). Public versus private education: What the debate in the
developing world teaches us. Solutions, 4(5). Accessed 9 February 2015 from http://www.
thesolutionsjournal.com/node/24108.
Shafiq, M. (2011). Do school incentives and accountability measures raise skills in the Middle East and
North Africa? The cases of Jordan and Tunisia. Review of Middle East Economics and Finance 7(2),
128.
Stitzlein, S. (2013). For-profit charter schools and threats to the publicness of public schools.
Philosophical Studies in Education, 44, 8899.
Thapa, A. (2013). Does private school competition improve public school performance? The case of
Nepal. International Journal of Educational Development, 33(4), 358366.
von Davier, M., Gonzalez, E., & Mislevy, R. J. (2009). What are plausible values and why are they
useful? IERI Monograph Series: Issues and Methodologies in Large-Scale Assessments, 2, 936.
World Bank (2008). The road not traveled: Education reform in the Middle East and North Africa.
MENA Development report. Washington, DC: International Bank for Reconstruction and
Development/World Bank.
Wu, M. (2005). The role of plausible values in large-scale surveys. Studies in Educational Evaluation,
31(2), 114128.
Zellman, G.L., Ryan, G.W., Karam, R., Constant, L., Salem, H., Gonzalez, G., Orr, N.,Goldman C.A.,
Al-Thani, H., & Al-Obaidli, K. (2009). Implementation of the K12 education reform in Qatars
schools. Santa Monica, CA: RAND Corporation.

The author

Jehanzeb R. Cheema is a Clinical Assistant Professor at the College of Education, University of Illinois
at Urbana-Champaign. He received his doctorate in Education in May 2012 from George Mason
University and a doctorate in Economics in December 2006 from the University of Wisconsin-
Milwaukee. His research interests include the examination of privatepublic school differences, the
relationship between self-efficacy and academic achievement, and the treatment of missing data in
surveys.

123

Vous aimerez peut-être aussi