Content uploaded by Juan Elias Ponce Jarrin
Author content
All content in this area was uploaded by Juan Elias Ponce Jarrin on Feb 23, 2023
Content may be subject to copyright.
https://doi.org/10.1177/14752409221122038
Journal of Research in
International Education
1 –16
© The Author(s) 2022
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/14752409221122038
journals.sagepub.com/home/jri
JRIE
Evaluating the impact of the
International Baccalaureate: an
analysis in Ecuador
Juan Ponce
Ruthy Intriago
Facultad Latinoamericana de Ciencias Sociales, Ecuador
Abstract
This study provides an evaluation of the impact of the International Baccalaureate Diploma Programme
introduced in 2006 to 200 public schools in Ecuador. Using a combination of different methodologies in
order to have a credible identification strategy, we find that the International Baccalaureate Diploma has a
positive impact on academic achievements as well as on secondary school progression.
Keywords
International Baccalaureate, Diploma Programme, Ecuador, Impact Evaluation
Introduction
International evaluations show that many developing countries have serious problems with their
quality of education at all levels. Because of the cumulative nature of the knowledge and skills
involved, it can be difficult to improve the quality of education at high school levels. In recent
years a good deal of emphasis has been placed on improving the quality of education at secondary
level. One of the policies adopted for this purpose is the implementation of programs that have
international recognition, such as the International Baccalaureate (IB) programs and, in particular,
the pre-university IB Diploma Programme (DP). These kinds of programs have been implemented
mostly in developed countries, with their introduction in developing countries only starting rela-
tively recently. In this respect, evaluation of the effect of these programs on school quality is rele-
vant as a policy guide to implementing them in developing countries as an initiative to improve
students’ cognitive achievements. The application of the IB DP in Ecuador offers a unique oppor-
tunity to evaluate the impact of the program on students’ cognitive achievements. As will be seen
in the literature review below, impact evaluation studies, while able to control for many potentially
confounding factors, are not able to control for students’ prior academic ability. In this regard, one
contribution of our research is that it is the first time that the impact of the IB DP on students’
Corresponding author:
Juan Ponce, Facultad Latinoamericana de Ciencias Sociales, Almagro y Pradera, Quito, 170407, Ecuador.
Email: jponce@flacso.edu.ec
1122038JRI0010.1177/14752409221122038Journal of Research in International EducationPonce and Intriago
research-article2022
Original Article
2 Journal of Research in International Education 00(0)
results in a developing country, using baseline and follow up information, has been carried out.
In addition, we evaluate the effects of the program over time by using tests applied during two
consecutive years. We found a positive effect of the IB DP on the cognitive achievements of final
year secondary students in Ecuador, with the effect remaining over time.
This paper is structured as follows: after this introduction we present a literature review regard-
ing the International Baccalaureate programs. In the following section we describe the country
context and the IB Diploma Programme. After that, we introduce the methodology, where we
explain the construction of the databases and the identification strategies used. We then present the
results and a discussion section, prior to the final conclusion.
Literature Review
In this section we introduce studies that analyse IB programs around the world. First we introduce,
as a general reference, studies that provide a big picture of IB programs. In addition, we introduce
studies that analyse IB programs in developed countries. Most of these studies use qualitative
methodologies. Finally, we concentrate on studies directed to analysis of IB programs in develop-
ing countries, with special emphasis on those studies that use impact evaluation techniques. We
concentrate on this type of study since our research is related to evaluating the impact of an IB
program in a developing country.
A review of the impact of IB programs on teaching and learning is made by Dickson, Perry and
Ledger (2018), while a general assessment of the IB experience is available in Bunnell (2008). We
find important qualitative studies that analyse IB programs, including studies that analyse students’
perceptions, in Tarc and Beatty (2012), Hatziconstantis and Kolympari (2016), Shaunessy et al
(2006), Poonoosamy (2016), Hertberg-Davis and Callahan (2008), Culross and Tarver (2007) and
O’Boyle (2009). In addition, studies that analyse the importance of teachers and administrators as
well as their perceptions of the IB programs are found in Savage and Drake (2016), Pendergast,
Dole and Rentoule (2014) and Halicioglu (2008). Studies that evaluate the importance of several
components of the IB programs may be found in Alford et al 2013), Hatziconstantis and Kolympari
(2016), Hayden (2006), Hinrichs (2003), Holman et al (2019), Munro (2003), Alford et al (2013),
Palmer (2016) and Ballantyne and Rivera (2014). Finally, there are studies that analyse the impact
of the IB programs on students’ achievement or performance at university. A study by Green and
Vignoles (2012) found that university students in the United Kingdom who had studied the IB DP
had higher grades in their university studies than did students who had completed the traditional
pathway for university admission. Similarly, Lee et al (2014) found that the IB DP served as a suc-
cessful channel for students in China to enter top-ranking universities around the world. The study
found that 72% of IB DP graduates in China from 2002 to 2012 attended one of the top 500 uni-
versities in the world. 51% of IB DP graduates went to study in US universities. In a previous
study, Mayer (2008) used mixed methods to investigate the relationship between the design of a
rigorous university program, the IB DP, and the socio-economic index of the students. He found
that the IB DP improved the retention rate of economically disadvantaged Afro-Americans,
Hispanics and Native Americans. Also in the US, Saavedra (2014) examined whether enrolment on
the IB DP in public schools in Chicago improved students’ academic attainment, as well as their
likelihood of graduating and enrolling at university. He found positive and significant impacts in
students’ achievement, as well as on secondary graduation rates and university enrolment, with the
highest impact amongst men. Gurumurthy (2016), meanwhile, examined the correlation between
studying in an IB DP school and grades (GPA) in the first semester of university using a sample of
IB DP graduates. He found that graduates of the IB DP attained a higher GPA during their first
semester at university. The same paper shows qualitative results which indicate that, because of the
Ponce and Intriago 3
demanding nature of the program, the students were better prepared in the first semester at univer-
sity in writing skills, problem solving, time management and coping with stress than students who
had not completed the IB DP.
In Canada, Poelzer and Feldhusen (1996) compared four groups of students in different pro-
grams: higher level IB DP with subsidiary level (now known as standard level) IB DP, and higher
level IB DP with regular (non-IB DP) students. Achievements in three areas of study were exam-
ined: biology, physics and chemistry. The sample included 708 students from seven schools in
Canada. The authors found that in all three areas of study the achievements of the IB DP students
were above those of students on regular programs, but were equal between those in higher level IB
DP and subsidiary level IB DP.
As noted above, the implementation of programs such as the IB DP is at an early stage in devel-
oping countries, and studies that analyse the IB DP in this context are scarce. Using qualitative
techniques, Saavedra, Lavore and Flores-Ivich (2016) assessed in the context of Mexico the rela-
tionship between studying the IB DP and preparation for university. They used a case study meth-
odology in four schools and interviewed teachers, managers and students in order to establish their
views on the extent to which the IB DP prepares students for university, finding that students
believed that being part of the program would help them to perform satisfactorily at university.
Similarly, teachers believed that the program enabled students to perform better. To complement
the qualitative part of the study, they carried out analysis of the final qualifications of the students
in school and their enrolment at university, finding that most of the IB students were successful
with their university applications and secured places at the best institutions in Mexico.
In China, Wright and Lee (2014) analyzed the potential of schools that offered the IB DP to
respond to the demands from China regarding the skills of their students. Wright & Lee worked
with case studies as their methodology in five Chinese elite schools. They found that the teaching
philosophy of the DP led to the students developing skills relating to Creativity, Action, Service
(CAS) and the Theory of Knowledge (compulsory components of the IB DP). Also in China, Lee
et al (2014) evaluated the impact of the IB DP with a focus on the preparation of students for uni-
versities overseas in terms of admission, preparation and performance. They carried out a quantita-
tive analysis over time with data from schools and university entrance data for students who had
completed the DP. In addition, they compared data from the IB DP examinations with the univer-
sity entrance score. They found that results for the program were a strong predictor of university
results, which suggests that higher IB DP results lead to admission to a better university (after
controlling for variables at household level), and at the same time suggests greater academic suc-
cess in university studies during the first two years.
Ecuador Context And IB Diploma Programme
Ecuador is a middle income country with, in 2018, a per capita GDP of around US$6,200, a
poverty incidence of 25% and extreme poverty incidence of 9% (INEC, 2018). There has been an
important increase in the rate of secondary enrolments in recent years, as can be seen in Table 1. At
the national level, the secondary enrolment rate went from 42% in 1999 to 65% in 2014. However,
sizeable disparities remain between the different groups in the population.
To evaluate cognitive achievements, Ecuador uses the Ser Bachiller test with students in their
final grade at secondary school, in which the areas of mathematics, language, sciences and social
studies are evaluated. As an example, the results of the 2017 tests show a high percentage of young
people with an insufficient or basic performance in mathematics (75%), language (54%), sciences
(77%) and social studies (71%). In this context one of the policies applied with the aim of improv-
ing the quality of secondary education was the IB Diploma Programme. A detailed review of the
4 Journal of Research in International Education 00(0)
implementation of the IB DP in Ecuador is found in Barnett (2013). The IB develops and promotes
knowledge, skills and abilities in students and is aimed at proactive, cooperative and collaborative
work with meaningful, cultural and continuous learning, promoting the training of enterprising
people with a spirit of investigation and innovation, ethical values, moral principles, with a social,
cultural and environmental awareness.
The State of Ecuador through the Ministry of Education, and the International Baccalaureate
organization in Geneva, Switzerland, signed a Memorandum of Mutual Agreement and
Understanding on 9 February 2006 which initiated the “Introduction of the International
Baccalaureate in public schools in Ecuador” project. The project was rolled out nationally in
approximately 200 public schools from its inception in 2010 until 2017. During this period, it was
budgeted to cost the Ministry of Education US$ 29.2 million. It takes two years (maximum three)
for a secondary school to be authorised by the IB to offer the Diploma Programme. The process
starts with an application from the school to participate in the program. Then representatives of the
IB organization make an in situ visit where checks are made that the school meets requirements
related to managers, teachers, IB coordinators, physical resources (learning spaces, library, labora-
tories), as well as implementation of the IB philosophy. On the basis of this visit the IB organiza-
tion makes recommendations to the school which, once complied with, would normally lead to
authorisation to offer the IB DP. The authorisation process for an institution in the region costs US$
54,000.
One of the most important elements of the IB DP is its curricular structure, which consists of six
groups of subjects: language and literature studies in one’s native language; acquisition of lan-
guages (English for public schools in Ecuador); individuals and society (history for public schools
in Ecuador); experimental sciences (choice between biology, physics and/or chemistry for public
schools in Ecuador); mathematics; and arts. As a complement to the six subjects of the IB DP,
students have to undertake a research project (Extended Essay), complete a course in the Theory of
Knowledge, and participate in Creativity, Action, Service (CAS) activities.
Between 2011 and 2016 about 4,000 students benefitted from the program which, in Ecuador
public schools, was taught over 3 years. Of this total, however, only 711 gained the IB Diploma
Table 1. Net rate of secondary enrolment.
Secondary enrolment 1999 2006 2014
National 41.7% 51.1% 64.7%
Area
Urban 56.2% 60.2% 69.9%
Rural 20.7% 33.7% 53.5%
Quintile
Q1 (poorest) 15.3% 21.9% 44.8%
Q2 30.3% 35.6% 61.1%
Q3 46.3% 54.5% 66.9%
Q4 57.3% 66.8% 76.7%
Q5 68.0% 84.5% 84.5%
Ethnicity
Indigenous 30.3% 51.5%
Non-Indigenous 52.9% 66.0%
Source: Living Standard Measurement Survey. INEC. Several years. The net enrolment rate is the relation between
those aged 12 to 17 years old and enrolled in secondary school, and those aged 12 to 17 years old.
Ponce and Intriago 5
(Ministry of Education, 2018). The remainder only achieved the country’s standard secondary
diploma and did not achieve the IB Diploma at the end of their IB DP studies. However, in this
article we evaluate the impact of the program comparing all IB DP students with comparable stu-
dents of other public schools in Ecuador.
It should be noted that, unfortunately, during 2020 the IB DP was discontinued in Ecuador pub-
lic schools. We hope that the results of this study may nevertheless be of interest, particularly in
contributing to evaluating the possibility of a potential reintroduction of the program if we find
significant and positive effects.
Methodology
An impact evaluation assesses the changes in well-being of individuals that can be attributed to a
particular program or policy. In this regard, it is key to identify the causal relationship between the
program or policy and the outcomes of interest. Impact evaluations generally estimate average
impacts of a program on the welfare of beneficiaries. In our case, we will estimate the impact of
the IB DP on students’ cognitive achievements, and school progression. The essential idea behind
an impact evaluation study is to compare one group that participates in the program with another
group that does not; both groups need to be similar in order for credible estimates of the impact of
the program to be obtained. In our case we compare schools that participate in the IB DP (treatment
group) with schools that do not participate in the program (control group). In order to guarantee
comparability between the treatment and control groups, we match both groups using information
available prior to the implementation of the program (using the list of variables in Table 2). In this
sense, because schools included in the treatment and control groups are similar before the imple-
mentation of the program, any difference in the outcome variable, after the implementation of the
program, can be attributed to the IB DP.
Database used
For this study a panel database was created, which includes for each educational unit (school) in
Ecuador’s school system information on academic achievements (the result of the application of
official tests by the Ministry of Education), plus information about school enrolment at both the
beginning and the end of the school year. In addition general information was included about the
educational institutions: location (province, canton, parish, district and circuit), funding (state,
private, municipal or state-church), area (urban or rural), school system (sierra: starts classes in
September; costa: starts classes in April), jurisdiction (Spanish or bilingual), mode of study (in
attendance or part-attendance), number of teachers (by gender), if there is a principal, deputy prin-
cipal and supervisor, number of students (at the start and end of each school year, by grade), num-
ber of computers and if there is a computing laboratory. Information is available for each institution
from 2008 to 2016. Table 2 presents a description of each variable used in this paper. Given that the
aim of the study is to analyse the effect on both school progression and cognitive achievement,
information is presented below on the outcome variables to be used.
Information on school progression
As indicated above, the panel database created for each school contains the number of students at
both the start and the end of the school year. By comparing enrolments for the treatment and con-
trol groups at the end of the year for two consecutive years in two consecutive grades, we will have
the impact on school progression, defined as:
6 Journal of Research in International Education 00(0)
PM
M
tg
tg
tg
,
,
,
=
++11
where grade progression Pg, in the year t, is equal to the final enrolment (M) of the year t+1, in
grade g+1, divided by the final enrolment (M) in year t in grade g. We work with the average pro-
gression rate for the final three years of secondary study (the IB DP). We evaluate the impact in the
final educational year available (2015-2016), and use as a control variable the progression rate of
the first year of the data series (2008).
Information on academic achievement
The information on academic achievement included in the database is as follows:
1. ‘Ser Bachiller’ test (2008). This test was applied nationwide as a census in 2008, in all public,
municipal and private schools, in the 4th, 7th and 10th grades at primary level and 3rd grade at
secondary, and evaluates the areas of Mathematics and Language. (To evaluate Social and Natural
Sciences a sample for the 7th and 10th grades of primary was included). A total of 803,065 students
were evaluated. The tests measured the command of specific basic skills in each of the learning
areas evaluated. Appendix 1 shows the nomenclature of all school grades with approximate ages.
Table 2. Description of the variables used.
Variable Description
Variables at baseline
Mate 2008 Grade in math in the year 2008
Lenguage 2008 Grade in language in the year 2008
Matutina Schools that teach in the morning
Vespertina Schools that teach in the afternoon
Nocturna Schools that teach at night
Matutina y vespertina Schools that teach in the morning and in the afternoon
Matutina y nocturna Schools that teach in the morning and at night
Vespertina y nocturna Schools that teach in the afternoon and at night
Matutina, vespertina y nocturna Schools that teach in the morning, the afternoon and at night
Régimen sierra Schools from the highlands
Sistema hispano Schools from the hispanic system
Presencial Face to face schools
Accesso terrestre Road access to school
Estudiantes 2008 Number of students in the year 2008
Docentes 2008 Number of teachers in the year 2008
Computadoras 2008 Number of computers in the year 2008
Outcome variables
Average progression rate Average rate of progression for the three grades of IB DP
Mate 2015 Grade in math in the year 2015
Mate 2015 Grade in math in the year 2016
Language 2015 Grade in language in the year 2015
Language 2016 Grade in language in the year 2016
Ponce and Intriago 7
2. ‘Ser Bachiller’ test (2014-2016). From the school year 2014-2015 the National Institute for Edu-
cational Evaluation (INEVAL) carried out standardized examinations for all students in 3rd grade
at secondary level who had passed the subjects of the corresponding curriculum in public, munici-
pal and private educational establishments. It was nationwide, applied as a census and aimed at
students who sought to obtain their secondary diploma (Ser Bachiller), that is at the end of second-
ary schooling (3rd grade of secondary school). The subjects evaluated were Mathematics, Lan-
guage, Natural and Social Sciences. Given that the implementation of the IB DP was started in
Ecuador public schools in 2010, the test measuring the achievements of 2008 was taken as the
baseline. Since we have census tests only for the school years 2014-2015 and 2015-2016, the
impact on achievements can only be analysed in the 3rd grade of the secondary diploma. We meas-
ured the impact in Language and Mathematics and worked with standardized tests, using the stand-
ard deviation of the control group for the standardization.
Identification Strategy
In order to evaluate the impact of the IB DP on progression and on cognitive achievement, two
methodological strategies were used, as follows.
Strategy 1. Fixed effects and lagged dependent variable with propensity score
matching
When working with a panel database, in other words observing the same unit several times over
time, there are many advantages with regard to the possibility of finding causality. One of the main
advantages when observing the same unit across time is the possibility of correcting for non-
observable effects that remain fixed in time. As an example, principal motivation or teachers’
enthusiasm are non-observed (we do not have measures of these variables), but we can control for
these variables if they remain fixed in time.
We will use here the following specification:
YXTY
ce
it to it ito itit
ll
10
=+ +++++
ββ
ατ
.. (Equation 1)
where the subscript i refers to the unit (in our case schools) and the subscript t refers to time. Yit is
the result variable (which can be progression or test score), X
lt
.. is a vector of control variables that
change over time or not (we include as control variables those reported in Table 2), ci is a compo-
nent of non-observables at the level of the school that does not change over time, and finally eit is
the idiosyncratic error term.
τ
t is the time effect for all schools in the period t. In addition, the result
variable on the baseline (2008), Yito, is included as a control variable; and Tit is the variable that
indicates whether the school received the treatment or not. The impact of the intervention would be
given by the parameter
α
. In essence the identification strategy is a model with fixed effects and
lagged dependent variable (Angrist and Pischke, 2008).
Furthermore, given that important information is available on the baseline (that is, for the year
2008), the probability of participating in the program can be used in order to match schools on the
basis of observable characteristics. There are many ways of matching: the nearest neighbour, the
five nearest neighbours, using the whole distribution, and so on. Recent developments in the litera-
ture on matching reveal potential problems with the efficiency of the matching estimators when
8 Journal of Research in International Education 00(0)
using a bootstrap to calculate the standard errors (Abadie and Imbens, 2011). Hirano, Imbens and
Ridder (2003) propose a different way of matching in order to obtain totally efficient estimators,
which involves estimating Equation (1) but using weighted least squares, in which the weights
used are 1 for the treated units and TX TX
()
−
()
()
1 for the control units. By using these
weights, the average treatment of the treated units is obtained. To obtain the average treatment for
the whole population the weights are 1TX
()
()
for the treated units and 11−
()
()
TX
for the
control units. (See Hirano, Imbens and Ridder, 2003).
In this case T
is the conditional probability of participation in the program (propensity score)
calculated on the basis of a probit model (selection equation) as follows:
TX Y
ii
titi
t
=
′++
00
πγ υ
(Equation 2)
where Ti is a dummy that takes the value of 1 if the school participates in the program and zero if
not. Xit0 is a vector of baseline variables that includes canton dummies, dummies by school system,
jurisdiction, mode, school day, numbers of students and teachers, number of computers, and a
dummy by area. Yit0 is the baseline result variable. All of the variables are taken with 2008 values.
Finally
υ
it is an error term that follows a normal distribution.
In short, the methodological strategy is based on a combination of a model of fixed effects with
lagged dependent variables, with a matching based on the probability of participating in the pro-
gram. In this way we can be sure to correct both by observables and non-observables (which do not
change over time), in order to avoid possible bias in the estimates.
Strategy 2. Fixed effects and pipeline comparison
The second strategy naturally also starts with use of the panel data. However, in this case the fixed
effects model with lagged dependent variable (Equation 1) is calculated using as control group
only those schools on the waiting list to join the IB DP. In other words, the control group is con-
fined to schools that have already been selected to participate in the IB DP, but this participation
would occur after 2016. This methodological strategy, known as pipeline comparison, allows for
the correction of selection bias estimates (Angrist, 1998). The main idea is that by using the pipe-
line comparison strategy the effect of non-observables in the selection process is corrected, whereas
the use of the fixed effects model corrects for non-observables fixed in time. In addition, we include
a vector on controls (X), as well as the lagged outcome variable.
Results
Cognitive achievement
To begin, we present a comparison of schools that offer the IB DP versus those that do not. We
restrict our analysis only to schools in the public system. (See Table 3).
Statistically significant differences were found in all of the baseline variables, except in the
mode of study variable. In general, those in the treatment group have better test scores in both
Mathematics and Language and have much higher numbers of students, teachers and computers.
Having carried out the matching, using the method of Hirano, Imbens and Ridder (2003), the
comparison of means between treatment and control groups succeeds in matching the groups
(although not completely) in the baseline variables. In order to see this, we ran a regression where
the dependent variable is the treatment variable, and the regressors are the variables reported in
Ponce and Intriago 9
Table 2. The regression was weighted in accordance with Hirano, Imbens and Ridder (2003).
Results are presented in Table 4.
Working at conventional levels of 5% significance, there remain significant differences only in
the number of teachers and computers. In the case of baseline test scores, the difference is signifi-
cant at 10%, but the coefficient is negative, which means that the treatment group has lower base-
line test scores than the control group. Something similar occurs in the case of size of school. The
treatment units after matching are smaller on average than the control units. In any case, all these
variables are introduced as control variables in the econometric models in order to calculate the
impact of the program.
We ran the fixed effects model as defined in Equation (1), using weighted ordinary least square
(OLS) as defined in the methodology, for both the 2015 and the 2016 tests. In all cases three speci-
fications were used. Specification 1 only uses as control variables the 2008 test. Specification 2
includes, in addition, all the control variables at baseline as shown in Table 2. Finally, Specification
3 includes canton dummies. The results for 2015 are shown in Table 5. In every case the coefficient
and the standard error are presented. There is a positive impact in the 2015 tests of 0.09 standard
deviation in Language and 0.19 standard deviation in Mathematics.
Table 6 shows the results for the 2016 tests. In this case there is also a positive impact both in
Language and in Mathematics, of 0.16 and 0.17 standard deviations respectively.
A potential problem with the results encountered thus far is that we are not correcting for vari-
ables, especially non-observable ones, that may have had an influence on the selection process of
the schools to participate in the program. One way of correcting this is to use the schools that had
already been authorised to offer the IB DP but had not yet implemented it. To this end we used as
a control group the schools authorised in 2016 to offer the IB DP, and those that were to be author-
ised in 2017.
Table 7 shows the comparison of means between the treatment and control groups previously
defined. There are significant differences between treatment and control groups in test scores,
Table 3. Comparison of baseline means between treatment and control schools.
Variable Controls SD Treated SD p-value
Mathematics 2008 0.017 (0.863) 0.490 (0.754) [0.000]
Language 2008 0.031 (0.808) 0.562 (0.687) [0.000]
Mornings 0.601 (0.490) 0.272 (0.446) [0.000]
Evenings 0.069 (0.254) 0.000 (0.000) [0.000]
Nights 0.005 (0.074) 0.000 (0.000) [0.000]
Mornings and evenings 0.174 (0.380) 0.329 (0.471) [0.000]
Mornings and nights 0.069 (0.253) 0.154 (0.362) [0.000]
Evenings and nights 0.007 (0.081) 0.000 (0.000) [0.000]
Mornings, evenings and nights 0.074 (0.262) 0.244 (0.430) [0.000]
Sierra regime 0.432 (0.495) 0.537 (0.500) [0.001]
Hispanic system 0.998 (0.039) 1.000 (0.000) [0.025]
In attendance 0.996 (0.060) 0.996 (0.064) [0.918]
Physical access 0.993 (0.081) 1.000 (0.000) [0.000]
Students 2008 547 –673 1.580 –1.136 [0.000]
Teachers 2008 29.600 –30.425 78.341 –49.492 [0.000]
Computers 2008 16.442 –22.481 37.224 –33.095 [0.000]
N 3302 246
10 Journal of Research in International Education 00(0)
mornings, Sierra region and number of computers. In the remaining variables the schools are simi-
lar. With regard to test scores, the treatment group have lower scores than the control group.
Table 8 shows the impact results on test scores in 2015. There is a significant impact in both
Language and Mathematics, of 0.34 and 0.32 standard deviations respectively. This result is sig-
nificant at 90%.
Table 9 shows the results for 2016. There is a significant impact both in Language and in
Mathematics of 0.24 and 0.35 standard deviations respectively.
Table 5. Impact of IB DP on test scores 2015.
Variable Language Mathematics
Spec.1 Spec.2 Spec.3 Spec.1 Spec.2 Spec.3
bach_intern 0.098 0.095 0.096 0.164 0.186 0.187
Std. err 0.059 0.049 0.031 0.068 0.063 0.042
leng_se2009 0.501 0.569 0.506
Std. err 0.041 0.040 0.035
mat_se2009 0.458 0.438 0.476
Std. err 0.044 0.050 0.045
N 1829 1823 1823 1829 1823 1823
r20.177 0.386 0.713 0.168 0.205 0.640
Note: Specification 1 only uses as control variable the test for 2008. Specification 2 includes, in addition, all the control
variables at baseline as defined in Table 2. Finally, specification 3 includes canton dummies.
Table 4. Comparison of means between treatment and control in baseline variables, after matching.
Variable Coefficient Stat. error P-value No. r2
Mathematics 2008 –0.156 0.081 0.054 1960 0.009
Language 2008 –0.107 0.063 0.086 1960 0.006
Mornings 0.030 0.036 0.407 1960 0.001
Evenings 0.000 0.000 . 1960 .
Nights 0.000 0.000 . 1960 .
Mornings and evenings 0.037 0.043 0.383 1960 0.002
Mornings and nights –0.006 0.034 0.852 1960 0.000
Evenings and nights 0.000 0.000 . 1960 .
Mornings, evenings and nights –0.061 0.055 0.262 1960 0.005
Sierra region –0.074 0.046 0.109 1960 0.006
Hispanic system 0.000 0.000 1.000 1960 0.000
In attendance –0.001 0.004 0.795 1960 0.000
Physical access 0.000 0.000 1.000 1960 0.000
Students 2008 –524.436 282.324 0.063 1960 0.028
Teachers 2008 –25.295 11.461 0.027 1960 0.032
Computers 2008 –9.851 4.107 0.017 1960 0.018
Note: The regression is computing using weighted OLS. The dependent variable is a dummy for treatment
(1 Baccalaureate, 0 otherwise). Using the propensity score of program participation ( TX
()
), the weights used are 1 for
the treated units and TX TX
()
−
()
()
1 for the control units. A significant coefficient means that there are differences
at baseline.
Ponce and Intriago 11
Impact on School Progression
As indicated in the methodology section, we have calculated the rate of progression by school year
in order to analyse the impact on enrolment. A synthesis indicator is the rate of average progression
for the three years of the IB Diploma. The fixed effects model from Equation (1) is estimated by
running a weighted OLS model with the weights defined by Hirano, Imbens and Ridder (2003), as
explained in the methodology.
Table10 shows the summary results for the three years of the IB DP. There is a positive impact
on the progression rate of 14%. When this information is disaggregated by school year, it can be
seen that the major impact is on the first year of the IB DP. See Appendix 2.
Table 6. Impact of IB DP on test scores 2016.
Variable Language Mathematics
Spec.1 Spec.2 Spec.3 Spec.1 Spec.2 Spec.3
bach_intern 0.090 0.133 0.164 0.124 0.156 0.171
Std. err 0.050 0.049 0.029 0.063 0.063 0.038
leng_se2009 0.563 0.459 0.471
Std. err 0.049 0.058 0.039
mat_se2009 0.413 0.354 0.392
Std. err 0.031 0.046 0.044
N 1820 1814 1814 1820 1814 1814
r20.260 0.315 0.702 0.150 0.173 0.649
Note: Specification 1 only uses as control variables the test for 2008. Specification 2 includes, in addition, all the control
variables at baseline as defined in Table 3. Finally, specification 3 includes canton dummies.
Table 7. Comparison of means between baseline treatment and control using the waiting list as the
control group.
Variable Controls SD Treated SD p-value
Mathematics 2008 0.819 (0.707) 0.443 (0.761) [0.001]
Language 2008 0.857 (0.582) 0.513 (0.697) [0.000]
Mornings 0.163 (0.373) 0.300 (0.460) [0.027]
Evenings 0.000 (0.000) 0.000 (0.000) [.]
Nights 0.000 (0.000) 0.000 (0.000) [.]
Mornings and evenings 0.388 (0.492) 0.315 (0.465) [0.342]
Mornings and nights 0.143 (0.354) 0.150 (0.358) [0.895]
Evenings and nights 0.000 (0.000) 0.000 (0.000) [.]
Mornings, evenings and nights 0.306 (0.466) 0.235 (0.425) [0.324]
Sierra region 0.857 (0.354) 0.479 (0.501) [0.000]
Hispanic system 1.000 (0.000) 1.000 (0.000) [.]
In attendance 1.000 (0.000) 0.995 (0.069) [0.319]
Physical access 1.000 (0.000) 1.000 (0.000) [.]
Students 2008 1.809 –1.257 1.578 –1.144 [0.238]
Teachers 2008 92.959 –59.133 78.000 –49.869 [0.100]
Computers 2008 51.755 –35.961 34.526 –31.152 [0.002]
N 49 213
12 Journal of Research in International Education 00(0)
Discussion
As noted in the literature review, we were not able to find impact evaluation studies of IB DP pro-
grams on students’ cognitive achievements for developing countries. In this sense, our research
makes an important contribution to understanding of these issues, and can be used as the basis of
education policy recommendations relating to the implementation of the IB Diploma Programme
in other developing countries.
Unfortunately, the IB DP was withdrawn from public schools in Ecuador in the year 2020. We
hope, however, that the positive impact found in this study may be taken into account in consider-
ing a possible reintroduction of the program.
Dividing the annual budget of the program by the total number of participating students gives
the annual per capita cost of the program, which is around US$ 30. As noted in the results, the
impact of the program, in the second year, was of 0.16 and 0.17 standard deviations in Language
and Mathematics respectively. In this sense, the cost of increasing 0.1 standard deviations, on aver-
age, is around US$ 17.5 per capita. Compared with other interventions directed to improvement of
academic outcomes in public secondary schools in Ecuador, the cost-benefit relation of the IB DP
Table 9. Impact of IB on test scores in 2016.
Variable Language Mathematics
Spec.1 Spec.2 Spec.3 Spec.1 Spec.2 Spec.3
Int_bac 0.193 0.259 0.241 0.403 0.424 0.348
Std. Err 0.084 0.093 0.101 0.120 0.138 0.153
leng_se2009 0.493 0.401 0.468
Std. err 0.084 0.091 0.079
mat_se2009 0.522 0.488 0.526
Std. Err 0.052 0.068 0.075
N 262 262 262 262 262 262
r20.195 0.288 0.541 0.202 0.265 0.606
Note: Specification 1 only uses as control variables the test for 2008. Specification 2 includes, in addition, all the control
variables at baseline as defined in Table 3. Finally, specification 3 includes canton dummies.
Table 8. Impact of the IB DP on test scores in 2015.
Variable Language Mathematics
Spec.1 Spec.2 Spec.3 Spec.1 Spec.2 Spec.3
Int_bac 0.459 0.243 0.343 0.273 0.257 0.322
Std. Err 0.132 0.127 0.139 0.155 0.162 0.183
leng_se2009 0.532 0.590 0.473
Std. err 0.066 0.060 0.066
mat_se2009 0.536 0.541 0.422
Std. Err 0.057 0.069 0.090
N 262 262 262 262 262 262
r20.201 0.398 0.626 0.182 0.243 0.459
Note: Specification 1 only uses as control variables the test for 2008. Specification 2 includes, in addition, all the control
variables at baseline as defined in Table 3. Finally, specification 3 includes canton dummies.
Ponce and Intriago 13
is quite positive (Ponce, 2018). In this sense, as a policy recommendation, we argue that the pro-
gram should be reintroduced in Ecuador, as well as being implemented more widely in other devel-
oping countries.
Conclusion
The International Baccalaureate Diploma Programme has as its principal aim the improvement of
the quality of education in schools through a process of authorisation which enables them to offer
an internationally recognized qualification. Around 200 authorised public schools in Ecuador
offered the IB DP to around 4,000 students with a committed budget since the start of the program
in 2010 of US$ 29.2 million. Using the “Ser Bachiller” tests of 2015 and 2016, we assessed the
impact on academic achievements, as well as the impact on student progression. We used two
impact evaluation methodologies. First, we combined a model of fixed effects with lagged
dependent variable and matching based on the probability of participating in the program. The
results, using this methodological strategy, show a positive and significant impact in cognitive
achievements in Language and Mathematics of around 0.09 and 0.19 standard deviations in the
2015 tests. This is equivalent to an increase of 7 and 16 marks out of 100 respectively in scores in
the tests. There is also a positive impact in the 2016 tests in Language and Mathematics of 0.16
and 0.17 standard deviations respectively, equivalent to an increase of 12 to 16 marks out of 100
respectively in the test scores.
The second methodological strategy to be used combined a fixed effect model with lagged
dependent variable and a pipeline comparison design. This means that the schools in the waiting
list to be authorised after 2016 were used as a control group. This strategy enabled us to correct for
possible selection bias. In this case there were also positive and significant effects in both Language
and Mathematics. In the 2015 tests the impact was 0.24 and 0.32 standard deviations respectively,
while in 2016 the impact was 0.24 and 0.35 standard deviations respectively.
We also assessed the impact on the rate of student progression, finding an increase in progres-
sion of 15% on average in the three years of IB DP study. The major effect was in the first year of
the three.
The cost-benefit relation of the program is quite positive, compared with other interventions
directed at improving academic outcomes in public secondary schools of Ecuador. Although,
unfortunately, the IB DP was withdrawn during the year 2020, we hope our results might be taken
into account in any consideration of a possible reintroduction of the IB DP in Ecuador.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Table 10. Impact of the IB DP on average school progression.
Variable Bacc_pro_1 Bacc_pro_2 Bacc_pro_3
Intern_Bacc 0.139 0.147 0.145
Std. Err 0.054 0.057 0.049
N 5498 5495 5495
r20.018 0.058 0.197
Note: Specification 1 only uses as control variables the test for 2008. Specification 2 includes, in addition, all the control
variables at baseline as defined in Table 3. Finally, specification 3 includes canton dummies.
14 Journal of Research in International Education 00(0)
References
Abadie A and Imbens G W (2011) Bias-corrected matching estimators for average treatment effects. Journal
of Business and Economic Statistics. 29(1): 1–11.
Alford B L et al (2013) Observing classroom instruction in schools implementing the International
Baccalaureate programme. Current Issues in Education. 16(2): 1-14
Angrist J D (1998) Estimating the Labor Market Impact of Voluntary Military Service Using Social Security
Data on Military Applicants. Econometrica. 66(2): 249–288.
Angrist J D and Pischke J S (2008) Mostly Harmless Econometrics: An Empiricist’s Companion. Princeton:
Princeton University Press
Ballantyne K and Rivera C (2014) Language Proficiency for Academic Achievement in the International
Baccalaureate Diploma Program. Available at: www.ceee.gwu.edu.
Barnett E (2013) Research on the Implementation of the Diploma Programme in Ecuador’s State Schools.
New York: Columbia University. Available at: http://www.ibo.org/research/policy/programmevalida-
tion/diploma/documents/ImplementationofDPinEcuadorStateSchools_Finalreport.pdf.
Bunnell T (2008) The global growth of the International Baccalaureate Diploma Programme over the first 40
years: A critical assessment. Comparative Education. 44(4): 409–424.
Culross R R and Tarver E T (2007) Teacher and student perceptions of the International Baccalaureate
Program: A first year perspective. Journal of School Choice. 1(4): 53–62.
Dickson A, Perry L B and Ledger S (2018) Impacts of International Baccalaureate programmes on teaching
and learning: A review of the literature. Journal of Research in International Education. 17(3): 240–261.
Green F and Vignoles A (2012) An empirical method for deriving grade equivalence for university entrance
qualifications: An application to A levels and the International Baccalaureate. Oxford Review of
Education. 38(4): 473–491.
Gurumurthy U (2016) International Baccalaureate: a Study in College Readiness. Texas A&M University-
Corpus Christi.
Halicioglu M (2008) The IB Diploma programme in national schools: The case of Turkey. Journal of Research
in International Education. 7(2): 164–183.
Hatziconstantis C and Kolympari T (2016) Student perceptions of Academic Service Learning: Using mixed
content analysis to examine the effectiveness of the International Baccalaureate Creativity, Action,
Service programme. Journal of Research in International Education. 15(3): 181–195.
Hayden M (2014) The International Baccalaureate (IB) and International Education. In T Pound (ed) The
International Baccalaureate Diploma Programme: An Introduction for Teachers and Managers.
London: Routledge. [pp 25–43].
Hertberg-Davis H and Callahan C M (2008) A narrow escape: Gifted students’ perceptions of Advanced
Placement and International Baccalaureate programs. Gifted Child Quarterly. 52(3): 199–216.
Hinrichs J (2003) A comparison of levels of international understanding among students of the International
Baccalaureate Diploma and Advanced Placement programs in the USA. Journal of Research in
International Education. 2(3): 331–348.
Hirano K, Imbens G W and Ridder G (2003) Efficient estimation of average treatment effects using the esti-
mated propensity score. Econometrica. 71(4): 1161-1189.
Holman A C et al (2019) Developing academic persistence in the International Baccalaureate Diploma
Programme: Educational strategies, associated personality traits and outcomes. International Journal of
Educational Psychology. 8(3): 270–297.
Lee M et al (2014) A study of the International Baccalaureate Diploma Programme in China: Impact on
student preparation for university studies abroad. Available at: https://ibo.org/globalassets/publications/
ib-research/dp/chinasummaryinenglishweb.pdf.
Mayer A P (2008) Expanding Opportunities for High Academic Achievment: An International Baccalaureate
Diploma Program in an Urban High School. Journal of Advanced Academics. 19(2): 202–235.
Ministry of Education (2018) Informe para proceso de transición BI. Gobierno de Ecuador.
Munro J (2003) The Influence of Student Learning Characteristics on Progress Through the Extended Essay.
Journal of Research in International Education. 2(1): 5–24.
Ponce and Intriago 15
O’Boyle É (2009) Whispers from within: Students’ perceptions of the first year of the International
Baccalaureate Middle Years Programme in an international school. Journal of Research in International
Education. 8(1): 30–56.
Palmer N (2016) Seeing the forest for the trees: The International Baccalaureate Primary Years Programme
exhibition and Global Citizenship Education. Journal of Research in International Education. 15(3):
208–223.
Pendergast D, Dole S and Rentoule D (2014) Teacher dependent discourse patterns: variations in student
discourse roles in an International Baccalaureate school: a mathematics focus. International Schools
Journal. 33(2): 16–25.
Poelzer G and Feldhusen J F (1996) An Empirical Study of the Achievements of International Baccalaureate
Students in Biology, Chemistry, and Physics in Alberta. Journal of Secondary Gifted Education. 8(1):
28–40.
Ponce J (2018) Síntesis de las Evaluaciones de Impacto del MINEDUC: 2017-2018. Quito, Ecuador:
MINEDUC.
Poonoosamy M (2016) Aspirations and tensions in developing international mindedness: case study of two
students in an IB school in an Indian Ocean Island Nation. Asia Pacific Journal of Education. 36(4):
583–598.
Saavedra A R (2014) The academic impact of enrollment in International Baccalaureate Diploma Programs:
A case study of Chicago public schools. Teachers College Record. 116(4): 1–40.
Saavedra A R, Lavore E and Flores-Ivich G (2016) The International Baccalaureate Diploma Programme in
Mexico as preparation for higher education. Compare. 46(3): 344–368.
Savage M J and Drake S M (2016) Living transdisciplinary curriculum: Teachers’ experiences with the
International Baccalaureate’s Primary Years Programme. International Electronic Journal of Elementary
Education. 9(1): 1–20.
Shaunessy E et al (2006) School Functioning and Psychological Well-Being of International Baccalaureate
and General Education Students: A Preliminary Examination. Journal of Advanced Academics. 17(2):
76–89.
Tarc P and Beatty L (2012) The emergence of the International Baccalaureate Diploma in Ontario: Diffusion,
pilot study and prospective research. Canadian Journal of Education. 35(4): 341–375.
Wright E and Lee M (2014) Developing skills for youth in the 21st century: The role of elite International
Baccalaureate Diploma Programme schools in China. International Review of Education. 60(2):
199–216.
Author biographies
Juan Ponce is a Professor in the Department of Economics at the Facultad Latinoamericana de Ciencias
Sociales, Flacso-Ecuador.
Ruthy Intriago is a research assistant at the Facultad Latinoamericana de Ciencias Sociales, Flacso-Ecuador.
Appendix 1. School grades and approximate ages.
Education level School grades Approximate ages
Preschool (not mandatory) Initial I
Initial II
3-5 years old
Basic General Education Preparatory (1st grade) 5 years old
Basic elementary (2nd, 3rd, 4th grade) 6-8 years old
Middle elementary (5th, 6th, 7th grade) 9-11 years old
Upper elementary (8th, 9th, 10th grade) 12-14 years old
Secondary schooling (high school) 1st, 2nd and 3rd of secondary schooling 15-17 years old
16 Journal of Research in International Education 00(0)
Appendix 2. Impact of the IB DP on school progression by grade.
First grade of bac
Variable Pro_1b_1 Pro_1b_2 Pro_1b_3
Intern_bacc 0.366 0.378 0.380
Std. err 0.142 0.154 0.134
No. 5498 5495 5495
r20.013 0.036 0.158
2nd bacc
Variable Pro_2b_1 Pro_2b_2 Pro_2b_3
Intern_bacc 0.019 0.021 0.019
Std. err 0.018 0.019 0.016
No. 5498 5495 5495
r20.173 0.241 0.426
3rd bacc
Variable Pro_3b_1 Pro_3b_2 Pro_3b_3
Intern_bacc 0.030 0.041 0.035
Std. err 0.029 0.026 0.019
No. 5498 5495 5495
r20.002 0.113 0.361
