7/30/2019 Curva dell'apprendimento
1/52
Developed by the Economist Intelligence Unit
Published by Pearson
2012 REPORTLESSONS IN COUNTRYPERFORMANCE IN EDUCATION
7/30/2019 Curva dell'apprendimento
2/52
Pearson is the worlds leading learning company. We provide
learning materials and services to educators and students o all ages;
business inormation through the Financial Times and the worlds
best books and ebooks through Penguin.
www.pearson.com
The Economist Intelligence Unit wrote all sections o this report with the exception
o the Foreword.
The Economist Intelligence Unit (EIU) is the worlds leading resource or economic
and business research, orecasting and analysis. It provides accurate and impartial
intelligence or companies, government agencies, financial institutions and academic
organisations around the globe, inspiring business leaders to act with confidence
since 1946. EIU products include its flagship Country Reports service, providing
political and economic analysis or 195 countries, and a portolio o subscription-based data and orecasting services. The company also undertakes bespoke research
and analysis projects on individual markets and business sectors. More inormation
is available at www.eiu.com or ollow us on www.twitter.com/theeiu
Disclosure: The Economist Intelligence Unit is part o The Economist Group.
Pearson owns a 50% stake in The Economist Group.
L , p h j h Th L C w...
www.h.p.
Full report online and in PDF
The Learning Curve databank
Case studies
Country profiles
Video interviews
Data visualisations
7/30/2019 Curva dell'apprendimento
3/52
1
FwBy Sir Michael Barber, Chie education advisor, Pearson
Ow
Rh y
App 1
02
E yA summary o the reports findings and it s conclusions 06
PAn explanation o the research context, objectives and itscontributing experts
04
Education inputs and outputs: its complicatedAssessing the evidence or correlations between educationinputs and outputs as well as socio-economic outcomes
12
Th : income, cultureand education outcomesThe relevance o wealth and cultural at titudes toeducation perormance
18
G h wh fHow does the quality o teaching that children receiveafect their uture prospects?
22
Sh h y:caveat scholacticusWhy expanding choice has positive efects in somecircumstances but not others
26
R h: education, labourmarket and social outcomesHighlighting the key personal, social and economic benefitso a better educated population
33
Tw pCreating a comparative index, to identiy common actorsin countries education success
38
C h yNext steps, and a call or more research
44
MhyMethodology or Quantitative Componentso The Learning Curve programme 46
App 2 S phy 49
7/30/2019 Curva dell'apprendimento
4/52
2
Over the last decade, international benchmarking o
education systems has become ever more prevalent.
More importantly, it has become increasingly influential
in shaping education policy at local, regional and
national levels. As studies by OECD-PISA and TIMSS
become more sophisticated and longitudinal time
sequences develop there is ever more to learn about
what successul education systems look like and how
success can be achieved.
In the early days o international benchmarking,education ministers and other leaders tended to worry
more about the media impact than the implications or
policy. However, once the regular routine o published
PISA results was established, in 2001, this changed.
Germany, or example, ound itsel much urther down
the first PISA rankings than it anticipated. The result was
a proound national debate about the school system,
serious analysis o its flaws and then a policy response
to the challenges that were identified. A decade later,
Germanys progress up the rankings is visible to all.
Now, in act, we are beyond the phase o individual
country reactions. Increasingly what we see is a
continuous dialogue among education ministers and
top ocials around the world about the evidence rom
international benchmarking and the implications or
education reorm. Education ministers in places such
as Singapore are constantly monitoring and visiting
other countries to learn what they might do better.
Arne Duncan organised a series o international
Foreword
dialogues with ellow ministers and union leaders
about the uture o the teaching proession around the
world. Meanwhile Michael Gove, Secretary o State
or Education in the United Kingdom has shown more
interest in international benchmarking than any o
his predecessors.
The continuous benchmarking series also enables
more sophisticated analysis o what works in education,
which leaders rom around the world can draw upon.
I have been involved in a series o three publicationswhich have explored the lessons in depth. The first o
these, written with colleagues at McKinsey, How the
Worlds Best-Performing School Systems Come Out on Top,
examined the lessons rom the most successul school
systems, and highlighted the importance o recruiting,
training and developing great teachers.
The second, also written with colleagues at McKinsey,
How the Worlds Most Improved School Systems Keep
Getting Better, took a new angle and looked not atwhat great systems do, but at how, over time, systems
come to be successul. The third, published earlier
this year with colleagues rom Pearson, Oceans of
Innovation, went a step urther and asked whether
achieving educational success as measured by PISA and
TIMSS was sucient to ensure a country was on track
or economic and social success in the 21st century.
The work o Eric Hanushek has likewise connected
PISA and TIMSS outcomes to the wider goals o society,
especially GDP growth. Eric has demonstrated a strongcorrelation between the quality o school systems and
economic growth.
Th L C 2012 Fw
7/30/2019 Curva dell'apprendimento
5/52
7/30/2019 Curva dell'apprendimento
6/52
4
This report, published by Pearson and writtenby the Economist Intelligence Unit, is part o a
wide-ranging programme o quantitative and
qualitative analysis, entitled The Learning Curve.It seeks to urther our understanding o what leads to successul educational outcomes both economic and social. The design and execution o the programme hasbenefited rom the ongoing advice o some o the worlds leading educational scholars.
This report itsel outlines the main findings rom analysis o a large body ointernationally comparable education data The Learning Curve Data Bank.It also draws on extensive desk research, as well as in-depth interviews conductedwith 16 experts in education. The research was conducted entirely by the Economist
Intelligence Unit, and the views expressed in the report do not necessarily reflectthose o Pearson. The report was written by Dr Paul Kielstra, and edited byDenis McCauley o the Economist Intelligence Unit.
Preace
Th L C 2012 P
7/30/2019 Curva dell'apprendimento
7/52
7/30/2019 Curva dell'apprendimento
8/52
6 Th L C 2012 E y
The goal o improving education today enjoys
great prominence among policymakers and other
stakeholders in societies worldwide. Although they may
not be able to quantiy it, governments in most countries
recognise a link between the knowledge and skills with
which young people enter the workorce and long-term
economic competitiveness. For this reason, interest is
intense in research which explores the actors that seem
to lead in some countries to outstanding educational
perormance, and ultimately to better qualifiedworkorces.
Executive summary
This report, and the broaderThe Learning Curve
programme o which it is part, is aimed at helping
policymakers, educators, academics and other
specialists to identiy some o these actors. At its
heart is a significant body o quantitative research.
The Learning Curve Data Bank (LCDB), which is
accessible online, brings together an extensive set o
internationally comparable data on education inputs
and outputs covering over 50 countries. This in turn has
enabled a wide-ranging correlation analysis, conductedto test the strength o relationships between inputs,
outputs and various socio-economic outcomes. It also
underpins an initiative to create a comparative index
o educational perormance which, as will become
apparent, is anything but a straightorward exercise.
Educators might hope that this or other similar bodies o
research would yield the holy grail: identification o the
input, or set o inputs, that above all else leads to better
educational results wherever it is applied. Alas, i this
report makes nothing else clear, it is that no such magic
bullets exist at an international level or at least that
they cannot, as yet, be statistically proven. Nonetheless,
our research which is also based on insights gathered
rom experts across the world provides some definite
signposts. Following are its highlights:
Although they may not be able to quantiy it,governments in most countries recognise a linkbetween the knowledge and skills with which
young people enter the workorce and long-termeconomic competitiveness. This report is aimedat helping policymakers, educators, academics andother specialists to identiy some o these actors.
7/30/2019 Curva dell'apprendimento
9/52
7Pearson plc
S hp w w
p p
The research examined a wide range o education
inputs, both quantitative such as spending on pupils
and class size as well as qualitative such as level o
school choice. It also looked at numerous potential
outcomes, ranging rom inculcation o cognitive skills
to GDP growth. A number o inputs show a statistical
link over time with certain outputs, notably betweenincome and results. These are discussed in the chapters
that ollow, but the most striking result o the exercise
is how ew correlations there are. Education remains
very much a black box in which inputs are turned into
outputs in ways that are dicult to predict or quantiy
consistently. Experts point out that simply pouring
resources into a system is not enough: ar more
important are the processes which use these resources.
I , y On the surace, money and education seem to create
a virtuous circle, with rich countries and individuals
buying good education or their children who, in turn,
benefit economically. A closer look, though, indicates
that both higher income levels and better cognitive test
scores are the result o educational strategies adopted,
sometimes years earlier, independently o the income
levels existing at the time. More important than money,
say most experts, is the level o support or education
within the surrounding culture. Although cultural change
is inevitably complex, it can be brought about in order to
promote better educational outcomes.
Th h
Good teachers exercise a proound influence: having a
better one is statistically linked not only to higher income
later in lie but to a range o social results including lower
chances o teenage pregnancy and a greater tendency to
save or their own retirement. The problem is that there
is no agreed list o traits to define or identiy an excellent
teacher, let alone a universal recipe or obtaining them.
That said, successul school systems have a number othings in common: they find culturally efective ways to
attract the best people to the proession; they provide
relevant, ongoing training; they give teachers a status
similar to that o other respected proessions; and the
system sets clear goals and expectations but also lets
teachers get on with meeting these. Higher salaries,
on the other hand, accomplish little by themselves.
Wh h h,
Recent research indicates that countries with greater
choice o schools have better education outcomes.
Presumably, allowing parents to choose the best
schools rewards higher quality and leads to overall
improvement. In practice, however, finding the
mechanism to make this happen is dicult. Extensive
studies o voucher programmes and charter schools
in the United States indicate that, while both can be
beneficial, neither is a magic ormula. On the other
hand, or-profit private education is providing students
in some o the least developed areas o the world
an alternative to poor state provision and showing
the potential benefits o choice and accountability.
Ultimately, as in any market or quasi-market, the real
value o choice comes rom people having the right
inormation to select the option that is truly superior.
7/30/2019 Curva dell'apprendimento
10/52
7/30/2019 Curva dell'apprendimento
11/52
7/30/2019 Curva dell'apprendimento
12/52
7/30/2019 Curva dell'apprendimento
13/52
7/30/2019 Curva dell'apprendimento
14/52
12
Education has always mixed the local and the global.The survival o Latin in Europe as a language o learning,
long ater its disappearance almost everywhere
else in society, reflected an ideal o the universality
o knowledge. On the other hand, state education
provision has long been closely associated with local
needs and the preservation o local cultures: in many
ederal systems, it alls to the state or province rather
than the national government. As currently delivered,
says Andreas Schleicher, the OECDs Deputy Director
or Education, education is very inward looking, a very
local activity. A lot o walls exist between countries.
Since the 1990s, the interaction between the parochial
and the international has taken on a new orm.
Comparative tests such as Progress in International
Reading Literacy Study (PIRLS), Trends in International
Mathematics and Science Study (TIMSS), and the
Programme or International Student Assessment
(PISA) maniest a growing emphasis on benchmarking
the perormance o diferent systems and onunderstanding what sets apart the highest achievers.
In Proessor Schleichers words, education debates are
no longer about improvement by national standards.
Best perorming countries now set the tone.
He also believes that PISA has undamentally challengedthe idea that education should be valued largely on the
volume o spending and other inputs, and the premise
that more investment is always better. The shit rom
inputs to outcomes [as the ocus o study] has been a
significant impact o the tests, he says. Such research
has also made clear that, or policymakers, more than
childrens grades are at stake: economists have ound
a close relationship between economic growth and
certain population-wide outputs o education such as
cognitive skills.1
Th D B wh
The Learning Curve Data Bank (LCDB) created
by the Economist Intelligence Unit as part o the
broader Learning Curve programme is an efort
to advance study in this area. It is a purpose-built,
substantial collection o data which includes more
than 60 comparative indicators gathered rom over
50 countries. Many o these indicators in turn relyon multiple pieces o inormation, so that, even with
some inevitable gaps, the LCDB encompasses over
2,500 individual data points. These go well beyond
traditional education metrics, such as teacher-student
ratios and various spending metrics, to cover a broad
range o educational inputs and possible outputs, rom
the degree to which parents demand good results
o schools to the proportion o adults who end up in
jail. The appendix to this report describes the LCDB
and the rest o the Quantitative Component, and themethodology behind it, in detail.
Th L C 2012 E p p: p...
Education inputs and outputs:its complicated...
1 Eric A. Hanushek and Ludger Woessmann,
Education and Economic Growth,
in Dominic J. Brewer and Patrick J. McEwan, eds.
Economics of Education (2010).
7/30/2019 Curva dell'apprendimento
15/52
7/30/2019 Curva dell'apprendimento
16/52
14
Beyond providing a useul tool or researchers, agoal o the Quantitative Component and Data Bank
has been to make possible a search or correlations
between inputs and outputs that endure over time.
The ultimate hope is to uncover, where possible, any
interventions which might have a positive efect not only
on the development o cognitive skills and scholastic
achievement, but also on societal outcomes such as
higher employment. The methodology appendix also
describes how these correlations have been sought.
The data suggest a small handul o strong links.
Two correlations show a connection between national
income and aspects o academic success: higher GDP
seems related to better Grade 8 PISA results; and a
better score on the Human Development Index (o
the United Nations Development Program UNDP)
and its Income Index are associated with higher upper
secondary graduation rates. LCDB data also suggest
a link between more years in school on average and
higher labour productivity in a country. (One apparentlystrong link that the higher a countrys average school
lie expectancy, the greater the proportion o students
will graduate is almost tautological given the time
requirements involved in most diplomas and degrees.)
S
These findings will be discussed in the chapters that
ollow, but the most striking result o the search or
correlations is the overall paucity o clear linkages.
In this, our study is not alone. Ludger Woessmann,
Proessor o Economics at the University o Munich,explains that a lack o any relationship between inputs
and outputs mirrors the extensive academic literature
on this topic. I you try to go beyond simple correlations,
the general result is nearly always the same. Chester
Finn, President o the Thomas Fordham Institute, an
education research organisation, and ormer United
States Assistant Secretary o Education, agrees. What
works, he says, takes place inside a black box that has
inputs coming in and outputs going out; but the inputs
do not predict the results and what goes on in the black
box is hard to quantiy.
The research does, though, at least point to some o
the diculties o seeing inside the black box. The first,
says Paul Cappon, ormer President o the Canadian
Council on Learning, is that in the study o education
we measure just a ew things, usually inputs more
than outputs because they are simpler and easier to
measure, not because they are more significant they
are not. Vibha Parthasarathi, a distinguished Indianeducationalist, adds that successul outcomes arise
rom the interplay o several actors, some tangible,
others intangible. What Ive seen in any number o
surveys is you measure what is measurable. The soter
inputs o education get let out. These inputs, however,
can be crucial, such as the cultural context in which
education occurs.
Th L C 2012 E p p: p...
7/30/2019 Curva dell'apprendimento
17/52
7/30/2019 Curva dell'apprendimento
18/52
7/30/2019 Curva dell'apprendimento
19/52
17
Second, straightorward correlations are dicult to
find because education involves complex, interrelated
processes rather than simple activities. Nahas Angula,
Prime Minister o Namibia who, as education minister,
oversaw the post-apartheid reconstruction o that
countrys education system, says that achieving good
outcomes is not really a question o spending money,
money, money. The question is how to get the most
out o the money you have spent. Dr Finn agrees:
Processes, more than inputs, are important. It is like
having a good cook versus a bad one: the ingredients
might be identical, but one produces something worth
eating. I education itsel is so complex, teasing out its
impact on broader societal phenomenon, like economic
growth, is harder still.This does not mean that education is a complete
mystery. Some key elements are apparent. Proessor
Schleicher explains that We have a good sense o what
makes a good education system. That doesnt answer
how you do it, but you can say these are the key actors.
The rest o this study will explore the most important o
those actors, bearing in mind that there is no single best
way to address them in every country. As with cuisine,
a variety o approaches may bring success. For example,
as we will discuss later, education in Finland and South
Korea two o the worlds top-perorming countries in
many benchmarks seem to have ew similarities other
than high academic achievement.
The main message o the lack o strong correlations,
though, should be humility. Brian Stecher, Associate
Director at RAND Education, says: We use jargon that
seems to explain student behaviour, but we really dont
understand the way students learn and the complex
mix o inputs amily, community and learning that
lead to skills and temperaments. I you compare
research in education to research in healthcare, you
see a dramatic diference in our knowledge o cause
and efect. Claudia Costin, Rio de Janeiros Municipal
Secretary o Education, adds that Reorming education
requires more than figures and analysis. You need to
avoid arrogance and the eeling o having a technocratic
approach.
Rather than being able to pronounce the last word,then, education research is still learning how to promote
better outcomes. The Data Bank itsel is only one step in
an efort that is hoped to last many years. The discussion
which ollows will look at several major issues relating
to successul educational outcomes, including national
income, culture, teaching quality and questions o choice
and accountability. In doing so, it seeks to be part o an
ongoing deepening o knowledge about education, and
to illuminate the key issues meriting urther investigation.
GDP per capita (US$PPP) Grade 8 PISA overall reading literacyGrade 8 PISA
PositivePositive
UN Human Development Index Upper secondary graduation rate Positive
UN Income Index Upper secondary graduation rate Positive
School lie expectancy(primary to tertiary)
Overall productivity o labour(GDP at US$PPP per worker)
Positive
Ratio o maximum teacher salary overaverage gross wage at primary level
Upper secondary graduation rate Negative
Ratio o maximum teacher salary overaverage gross wage at secondary level
Upper secondary graduation rate Negative
VARIABLE 1 VARIABLE 2 TYPE
Pearson plc
SELECTED STRONG RELATIONSHI PS FROM THE CORRELATION ANALYSIS
Note: Strong correlations, such as those shown here, are above a threshold o 0.65. Correlation te sts were conducted between two variab les over time
(on an annual basis). Each correlation reers to a minimum o 15 countries out o the sample.
Source: Economist Intelligence Unit.
7/30/2019 Curva dell'apprendimento
20/52
18
I : fp?
Two correlations rom the quantitative analysis indicate
a link between a countrys income and its educational
outcomes: higher GDP is associated with better
overall PISA scores, and the UNDPs Income Index is
a predictor o national secondary school graduation
rates. Similarly, PISA results correlate with national
GDP and Income Index scores in the years ollowing
the tests being administered. In both cases, however,the causation is not clear. In relation to the second
link, or example, those who were age 15 in 2009 and
2006 have had so little time in the labour orce that the
contribution o their skills is unlikely to have had much
efect yet on national income. That said, Proessor
Schleicher reports that PISAs extensive longitudinal data
on test-takers indicates that the tests predictive power
o ease o transition to work and initial income is high.
On the surace, this suggests a virtuous circle moneybuys good education, which instils higher earning
power. This seems to parallel an oten observed link
between socio-economic status and academic results
within countries. I anything, this association is growing
in the United States,2 but it is ar rom an American
phenomenon. It is present in European countries, such
as the United Kingdom and Italy, as well as, according to
a 1999 study by a World Bank researcher, in 43 largely
developing nations.3
Money, or both countries and individuals, does bringsobvious advantages. As Ms Parthasarathi notes or
amilies, wealth gives you access to schools where
you assume there are better teachers, etc, [and] ...
people who dont have the means miss out on a lot o
opportunities. The wider link to educational results,
however, is ar rom straightorward. Ms Parthasarathi
points out that, at the individual level, even something
as basic as student motivation can be greatly afected by
economic background.
More generally, a recent OECD report indicates that a
commitment to equity within an education system can
greatly diminish the correlation between amily income
and educational outcomes. It points to Finland, Canada
and South Korea, among others, as examples o success
in this area.4 This is consistent with research conducted
by the Canadian Council on Learning, says Mr Cappon.
Our composite learning index showed no direct
correlation between the wealth o a community and its
learning environment. It is not a given that you simply geta higher result with higher income levels.
Similarly, the tie between GDP per capita and PISA
results is ar rom linear. For countries with incomes
under $20,000 per person, economic growth appears
to bring rapidly improving educational results. Ater that
point, however, the gains become much less obvious.5
This type o result is common in economics, appearing
in areas such as the impact o national income on lie
expectancy: up to a certain point, the need is so greatthat almost any spending brings gains; thereater the way
that the money is spent becomes much more significant.
The tangible and intangible:income, culture andeducation outcomes
Th L C 2012 Th : ,
2 See Greg J. Duncan and Richard Murnane, eds.
Whither Opportunity? Rising Inequality, Schools, and
Childrens Life Chances, 2011.
3 Deon Filmer, Inequalities in Education:International Experience, in Ismail Sirageldin,
Human Development in the Twenty First Century.4 Equity and Quality in Education: Supporting
disadvantaged students and schools, 2012.
5 Does money buy strong perormance in PISA,
PISA in Focus, February 2012.
7/30/2019 Curva dell'apprendimento
21/52
7/30/2019 Curva dell'apprendimento
22/52
7/30/2019 Curva dell'apprendimento
23/52
7/30/2019 Curva dell'apprendimento
24/52
7/30/2019 Curva dell'apprendimento
25/52
7/30/2019 Curva dell'apprendimento
26/52
25Pearson plc
7/30/2019 Curva dell'apprendimento
27/52
25
Experts interviewed or this study repeatedly point
to several o these other actors which are essential in
promoting teacher quality:
Attracting the best people to the proession: Gettinggood teachers begins with recruiting talented individuals.
Finland and South Korea two perennially cited
examples o education success and the top countries in
our Index obtain their annual teacher intake rom the
top 10% and 5% o graduating students respectively.
The key to such success is the status in which teaching
is held culturally. Here money can have some efect,
not just as a simple inducement but as a signal o status.
The South Korean government uses high levels o
teacher pay in this way both to compensate or large
class sizes and to indicate the importance it accords to
the proession.
Providing the right training: The training o these new
recruits has to be appropriate to the conditions in which
they will work. This varies by country. The Finnish
system, or example, benefits rom teachers having
graduate degrees. On the other hand, Nahas Angula,
Prime Minister o Namibia, points out that his countrys
policy o requiring all teachers to have an undergraduatedegree may be driving up the cost o education when
other training would suce or primary grades.
Teacher training also needs to be ongoing. This has a
very practical reason that no teachers college course
will maintain complete relevance across decades o
work but also a demonstrative one. As Mr Cappon
notes, teachers need to be lielong learners themselves.
You cant inculcate a love o learning unless you live
it. Efective proessional development needs to
address not just upgrading the knowledge o teachers
providing, or example, a better understanding o new
technology and teaching strategies but also allow them
to advance along their career path into more senior
positions where relevant.
Treating teachers like proessionals: Consistent with the
need to promote the status o teaching is its treatment
as a proession. Mr Ratteree notes that things like
continual proessional development and proessionalautonomy can be powerul incentives or better learning
outcomes. Mr Cappon agrees: Teachers must be
seen as proessionals who exercise judgement, not just
technicians.
Implementing clear goals and efective oversight, and
then letting teachers get on with it: Proessors Hanushek
and Woessmann both point to this combination
o accountability and independence as consistently
correlated with improved outcomes. Says the latter:
Education economists emphasise the need to think
about incentives or people in the system to use
resources eciently. These are mostly ramed by the
surroundings o the education system, the accountability
system and whether schools can act autonomously.
There is clear evidence o strong relations between
these and improved outputs. Proessor Schleicher
agrees. High-perorming school systems, he says,
combine demanding standards, low tolerance o ailure,
and clear articulation o expectations with a lot oproessional responsibility within a collaborative work
organisation at the ront line, or both teachers and
schools.
None o these on their own is enough. Instead, they
orm an overlapping, and mutually supporting, set o
strategies to provide the high-quality teachers that are
so important or education and to use them in the most
efective ways.9
Pearson plc
9 For a similar discussion o the key success actors
in teacher development see Michael Barber and
Mona Mourshed, How the Worlds Best-Performing
School Systems Come Out on Top , McKinsey andCo., 2007, pp 15-23.
26 Th L C 2012 Sh h y: h
7/30/2019 Curva dell'apprendimento
28/52
Th h
In the English city o Guildord in 2011, every final-year
student in the Royal Grammar School earned at least
three A-levels, the highest secondary-school subject
qualification. The equivalent figure or the citys Kings
College or the Arts and Technology was just 69%.
Neither figure was a surprise, nor is such variation
exceptional. In most places, it is simply accepted that
specific schools like individual teachers have diferent
results which tend to persist over time. A naturalconclusion is that giving parents, and through them
students, the ability to choose better perorming
schools should lead to better outcomes.
Unortunately, this issue is ar more complex and not just
because o the range o systems through which choice
operates across the world including both publicly
and privately unded options. Whatever their specific
strong and weak points, all these arrangements need
accurate inormation. Getting it wrong can be harmul.
A study o Beijing parental selection o primary schools
ound that excessive optimism about place availability
at better schools led parents to use up application
choices on schools that were already ull. Less optimistic
parents snapped up places at the next tier o schools,leaving only markedly worse ones or the children o
those making the initial mistake.10 On the other hand,
researchers in North Carolina ound that better, clearer
inormation on local schools increased the number o
low-income parents taking advantage o school choice,
and that the children so placed perormed better.11
As in any quasi-market, or choice to work, schools
have to reveal how well they are doing: choice and
accountability must go hand-in-hand.
Any accountability system, however, requires some
decision on what should be measured. Demographic
diferences between the children in the two Guildord
schools above might explain the gap in results ar
better than the education provided. Mr Cappon notes
o Canada: Social class and school choice tend to go
together. Indeed, much o the choice and accountability
debate continues because such other issues cloud
the picture.
School choiceand accountability:caveat scholacticus
y
10 Fang Lai, Elisabeth Sadoulet, Alain de Janvry,
The Adverse Efects o Pare nts School Selection
Errors on Academic Achievement: Evidence romthe Beijing Open Enrollment Program, Economics
of Education Review(2009) v28 n4: 485-496.
11 Justine S. Hastings and Jefrey M. Weinstein,
Inormation, School Choice, and AcademicAchievement: Evidence rom Two Experime nts,
The Quarterly Journal of Economics, (2008):
1373-1414.
27Pearson plc
7/30/2019 Curva dell'apprendimento
29/52
Recent research suggests that, at the system-wide
level, the potential or inormed choice helps raise
educational outcomes and reduces costs. In particular,
a cross-country comparison o the number o private,
oten aith-based, schools an indication o the degree
o choice with the 2003 PISA results ound that,
even ater controlling or other actors, the share o
schools that are privately operated has an economically
and statistically significant positive efect on student
achievement in mathematics, science, and reading.12
The benefits were greater than average or students
with a lower socio-economic status where such
private schools were publicly unded, as in Belgium and
the Netherlands. Proessor Woessmann, one o the
authors, explains: I there is more choice or parents,
and more non-governmental school operators so that
schools are not managed by one big state monopoly,
countries perorm much better.
Note: The score, which is on a scale o 0 to 1, is an ag gregate o the ollowing indicators: enrolment choices (reedom o enrolment choice at primary and lower
secondary education), the level o school choice (percentage o pupils living in an area with more than t wo schools), parental expectations, and fina ncial choiceand inormation (availability o school vouchers and governme nt responsibility or inorming parents on school choices (primary and lower secondary).
Source: Economist Intelligence Unit and OECD.
Sp
Nw Z
Th
Iy
Ch
A
B
U K
M
H K
Ch
Gy
Czh Rp
Nh
A
I
U S
Jp
Sw
Sh K
F
Swz
0.88
0.85
0.79
0.79
0.78
0.67
0.64
0.62
0.59
0.59
0.59
0.54
0.49
0.47
0.44
0.41
0.41
0.23
0.17
0.17
0.09
0.07
SCHOOL CHOICE SELECTED COUNTRIES, AGGREGATED SCORE, 2009
12 Ludger Woessmann and Martin West,
Competition rom private schools boostsperormance system-wide, Vox, http://www.
voxeu.org/article/competition-private-schools-
boosts-perormance-system-wide.
... the shareo schools thatare privatelyoperated hasan economicallyand statisticallysignificant positiveefect on studentachievementin mathematics,science, andreading.
Ludger Woessmann and
Martin West, Competitionrom private schools boosts
perormance system-wide
28 Th L C 2012 Sh h y: h
7/30/2019 Curva dell'apprendimento
30/52
How this choice drives the system to better results
in practice, however, is a matter o no little debate.
Indeed, any discussion involving market-like mechanisms
and education inevitably leads to contentious, oten
politicised, debate. Unortunately, the resultant heat has
shed little consistent light.
Vh h h
Some o the most investigated choice initiatives operate
in the US. Voucher programmes provide unding
generally assigned by lottery as the programmes are
almost invariably oversubscribed that pay or the
private education o underprivileged children. A 2008
review by Patrick Wol, Proessor o School Choiceat the University o Arkansas, looked at the ten best
studies o these programmes and ound widely varying
results.13 In general, all or some students who used
vouchers did better academically in certain fields,
especially maths. A more recent study by Mr Wol o
the long-standing Milwaukee voucher system brought
urther variability: voucher students there outdid peers
in reading but underperormed in maths.14
The impact o such programmes on abilities tends to be
unpredictable, but that may not be the point. Parents
almost invariably are satisfied with them, although
perhaps or reasons quite apart rom grades. Given
the public options available to some o these students,
physical saety is an issue: one study ound no academic
diferences or voucher users, but they did have lower
arrest rates.15
Another possible impact o choice is to create
competition so that all schools improve, especially
where they are made to give data on results. Debate on
the extent to which this has taken place and whether
competition was the driver o perceived change is also
ongoing.16 The one clear point is that vouchers, and
choice, do not seem to hurt existing school systems.17
A more widespread US experiment in using choice
and accountability to improve education has been
the growth o charter schools. These autonomous,
privately-run but publicly-unded schools open to all
students capacity permitting exist in 41 states.
In return or autonomy, these institutions are made
accountable. Charters are granted with binding
requirements to achieve certain levels o academic
success among students.
As with vouchers, the success o charter schools
as a whole is the ocus o intense debate. The largest
review to date o research presents a mixed picture.
The Center or Research on Education Outcomes
looked at research rom 15 American states and the
District o Columbia. It ound that, on average, students
in these schools tended to do slightly worse than thosein nearby public schools. But the broader message was
variety: 17% o charter schools do better, 46% are just as
good, and 37% do worse. Moreover, the success o the
schools depends on the way they are regulated. Roughly
even numbers o states had schools where students on
average did better than in traditional schools and schools
where students did worse.18
13 School Voucher Programs: What the Research
Says About Parental S chool Choice, Brigham
Young University Law Review, (2008): 415-446.
14 The Comprehensive Longitudinal Evaluation of the
Milwaukee Parental Choice P rogram: Summary ofFinal Reports, February 2012, http://www.uark.
edu/ua/der/SCDP/Research.html.15 Julie Berry Cullen, Brian A. Jacob, and Steven
Levitt, The Efect o S chool Choice onParticipants: Evidence From Randomized
Lotteries,Econometrica, (2006), 74: 11911230.
16 See: Caroline Hoxby, School Choice and School
Productivity: (Or Could School Choice Be A Tide That
Lifts All Boats?), 2002, NBER Working Paper 8873,
an influential, article advocating this argument,
http://www.nber.org/papers/w8873; GregForster,A Win-Win Solution: The Empirical Evidence
on School Vouchers, 2011; David N. Figlio and
Cecilia Elena Rouse, Do Accountability and VoucherThreats Improve Low-performing Schools?, 2005,
NBER Working Paper 11597.
17 Research on voucher programmes in Chile have
produced similarly contrasting results to those
in America (Francisco Gallego, School Choice,
Incentives, and Academic Outcomes: Evidence or
Chile, paper 39, Econometric Society 2004 L atinAmerican Meetings; Chang-Tai Hsieh and Miguel
Urquiola, The efects o generalized school
choice on achievement and stratification: Evidencerom Chiles voucher program,Journal of Public
Economics (2006) 90: 14771503).
18 Center or Research on Education Outcomes,
Multiple Choice: Charter School Performance in
16 States, June 2009.
7/30/2019 Curva dell'apprendimento
31/52
30 Th L C 2012 Sh h y: h
7/30/2019 Curva dell'apprendimento
32/52
Sh h p Where such provision is poor, however, choice and
accountability can be essential. James Tooley, Proessor
o Education Policy at Newcastle University, has done
extensive research into the huge number o unocial
private schools used by economically underprivileged
students in developing countries. In many cases, rather
than trusting state provision, amilies are willing to
spend oten a substantial part o their income to send
children to these unregistered schools. The reason is
simple: parents know that education is important but
public provision is sub-standard or illusory. Proessor
Tooley ascribes parents decisions in this area to their
mistrust o state-school teachers, who are accused o
absenteeism, poor teaching habits and poor attitudes
toward students themselves.
As with any unocial activity, it is hard to assess its
ull scope. Proessor Tooley notes that the best data
rom India shows around a quarter attending private
schools in rural areas, and other research indicatesaround 6570% do so in urban areas. He thereore
estimates the overall total at around 40% or more
a figure consistent with his own, less detailed research
in communities in Ghana, Kenya, and Nigeria.20
These schools exist because they provide results:
Proessor Tooleys research in a variety o locations
has ound significantly better reading, mathematics,
and English skills. Similarly, World Bank-supported
researchers rom the Learning and EducationalAttainment in Punjab Schools (LEAPS) project ound
that in that Pakistani state, students in such private
schools were on average 1.5 to 2.5 years ahead ocounterparts in government schools, even though the
latter spent three times as much per pupil.21
What makes these private schools so much more
efective is not immediately clear, says Proessor Tooley.
They typically have ewer resources, class sizes vary
widely and oten the teachers are not as well trained or
do not have as much teaching experience. He concludes
that there is a missing ingredient [rom public schools
that exists] in private schools. It must be accountability.
The teachers have to teach, otherwise they get
removed; the schools need to please parents.
The extreme situation aced by these parents gives
the same message as the correlation between PISA
outcomes and private-school numbers: choice and
accountability can have an important impact on results.
On the other hand, the experience o school choice
in the US shows that the way these mechanisms
work are complex, require parents to have as much
inormation as possible and can penalise wrong choicesas much as reward right ones. Rio de Janeiros Ms
Costin points out, however, that the efort needed to
bring in parents is worth it even in the poorest areas:
They are not second-class citizens. Their opinion is
important. Parents know which school is a good school.
Social pressure or quality can be exerted even by
illiterate parents.
20 It should be noted or disclosure purposes thatPearson, who commissioned this report rom the
Economist Intelligence Unit, is a minority investor
in a chain o schools in Ghana co-ounded by
James Tooley.
21 James Tooley, Yong Bao, Pauline Dixon,John Merrifield, School Choice and Academic
Perormance: Some Evidence From Developing
Countries,Journal of School Choice, 2011,
5: 139; Baladevan Rangaraju, James Tooley,
Pauline Dixon, The Private School Revolution in
Bihar: Findings from a survey in Patna Urban, 2012;World Bank, Learning and Educational Achievement
in Punjab Schools Report Summary, 2008.
7/30/2019 Curva dell'apprendimento
33/52
7/30/2019 Curva dell'apprendimento
34/52
7/30/2019 Curva dell'apprendimento
35/52
34 Th L C 2012 R h: ,
7/30/2019 Curva dell'apprendimento
36/52
In considering country-level benefits, the more common
area o study has also been economic. On a basic level,
education helps. Our correlation analysis shows a strong
link between average years in school or school lie
expectancy and labour productivity. This does not
surprise Namibias Mr Angula: A well-educated nation
is likely to be innovative. I dont think that you have to
go to the statistical evidence to find that. People are
able to use knowledge or economic development. It is
not simply that better educated people themselves are
more productive. Extensive research has ound a spill-
over efect rom education, with benefits arising both
rom how the educated share their knowledge with
others and how they are better able to pick up new skills
themselves by building on their existing education.
The diculty or policymakers, though, is deciding what
sort o education works best when so many actors
afect the economy. Predictably, quality appears to
be more important than duration. In one analysis,
Proessors Hanushek and Woessmann ound that
when cognitive skills, as measured by PISA scores, are
correlated with GDP, then the impact o total years
o schooling becomes irrelevant. In other words,
how long it took to learn was less important than that
learning had occurred.26 This may seem obvious, but it is
directly applicable to decisions such as starting primary
education a year earlier or using the same resources or
teacher training.
Overallproductivityof
labour(GDPatUS$PPPperworker)
RELATIONSHIP BETWEEN SCHOOL LIFE EXPECTANCY AND LABOUR PRODUCTIVITY 19902011
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
1991 0.78 0.80 0.84 0.85 0.87 0.83 0.81 0.79 0.82 0.81 0.80 0.78 0.75 0.74 0.73 0.74 0.74 0.72 0.74 0.70 0.69
1992 0.79 0.80 0.83 0.85 0.86 0.83 0.81 0.80 0.83 0.81 0.80 0.78 0.75 0.73 0.72 0.73 0.74 0.73 0.74 0.70
1993 0.79 0.79 0.83 0.84 0.86 0.83 0.82 0.80 0.83 0.81 0.80 0.78 0.74 0.72 0.72 0.73 0.75 0.72 0.73
1994 0.78 0.79 0.83 0.84 0.86 0.84 0.82 0.80 0.82 0.81 0.80 0.78 0.74 0.72 0.71 0.74 0.74 0.72
1995 0.77 0.78 0.82 0.84 0.86 0.83 0.82 0.80 0.82 0.81 0.79 0.77 0.74 0.71 0.72 0.73 0.73
1996 0.76 0.77 0.82 0.85 0.86 0.84 0.81 0.79 0.82 0.80 0.79 0.77 0.74 0.72 0.71 0.72
1997 0.75 0.77 0.83 0.85 0.86 0.83 0.81 0.79 0.82 0.80 0.78 0.77 0.74 0.71 0.70
1998 0.75 0.78 0.83 0.85 0.86 0.83 0.80 0.78 0.81 0.79 0.78 0.77 0.74 0.70
1999 0.76 0.79 0.84 0.84 0.86 0.82 0.80 0.78 0.81 0.79 0.79 0.78 0.74
2000 0.77 0.79 0.82 0.84 0.85 0.82 0.79 0.78 0.81 0.79 0.79 0.77
2001 0.78 0.77 0.82 0.83 0.85 0.81 0.79 0.78 0.81 0.80 0.79
2002 0.78 0.77 0.81 0.83 0.84 0.81 0.79 0.78 0.82 0.80
2003 0.77 0.77 0.81 0.82 0.84 0.81 0.80 0.79 0.82
2004 0.77 0.76 0.81 0.82 0.84 0.81 0.80 0.79
2005 0.77 0.76 0.80 0.82 0.84 0.82 0.80
2006 0.77 0.76 0.81 0.83 0.85 0.82
2007 0.76 0.77 0.81 0.83 0.85
2008 0.77 0.77 0.81 0.83
2009 0.77 0.77 0.82
2010 0.77 0.77
2011 0.78
Note: The scatter matrix shows the correlation o school lie expectancy or all years against a ll possible uture years or overall productivity o labour.The correlation in each set o years is well above our threshold or strong correlations o 0.65.
Source: Economist Intelligence Unit and UNESCO.
26 Eric A. Hanushek and Ludger Woessmann,
Education and Economic Growth,in Dominic J. Brewer and Patrick J . McEwan, eds.
Economics of Education (2010).
School lie expectancy (primary to tertiary)
35Pearson plc
7/30/2019 Curva dell'apprendimento
37/52
More complicated than quality is the question o
what sort o content in an education system will yield
the best labour market and economic outcomes.
For example, some countries prize strong vocational
school programmes while others preer more unified
systems. One advocate o vocational education is
Proessor Schwartz, who says o the US that having a
system ocused entirely on preparing students or our-
year colleges and universities is a major problem. Only
30% o young Americans actually get a our-year degree
by their mid-twenties, and many o those wind up in
jobs that didnt require a degree. The consequence o
not having a strong post-secondary vocational system
is that most young Americans reach their mid-twenties
without the skills and credentials needed or success in a
technology-driven economy.
Mr Angula, whose country is looking to bolster its
vocational education system, adds that systems need to
create linkages between the school and the community,
and the school and the economy, so that education
should have a meaning in the context that it is practised.
Sometimes it is hard or students to apply their
knowledge or skills. Without seeing any relevance, they
might simply leave education.
INDONESIA
THAILAND
COLOMBIA
ROMANIA
RUSSIA
MEXICO
TURKEY BULGARIA ARGENTINA
SLOVAKIA
CZECH REPUBLIC
HUNGARYPORTUGAL
KOREA
NEW ZEALAND
AUSTRALIA
BELGIUM
NETHERLANDS
UNITED STATES
NORWAY
AUSTRIAIRELAND
GREECE
JAPAN
CANADA
FINLAND
SWEDEN
SPAIN
GERMANYFRANCE
ITALY
UK
DENMARK
ISRAEL
CHILE
CHINA
School lie expectancy (primary to tertiary)
0 2 4 6 8 10 12 14 16 18 20
Overallproductivityoflabour(GDPatUS$PPPperworker)
100,000
90,000
80,000
70,000
60,000
50,000
40,000
30,000
20,000
10,000
0
RELATIONSHIP BETWEEN SCHOOL LIFE EXPECTANCY (IN 1995) AND LABOUR PRODUCTIVITY (IN 2010)
SWITZERLAND
Note: The scatter chart shows the correlation o school lie expectancy in 1995 against overall productivity o labour in 2010 or 37 countries.The correlation is 0.817, well above our threshold or strong correlations o 0.65.
Source: Economist Intelligence Unit and UNESCO.
36 Th L C 2012 R h: ,
7/30/2019 Curva dell'apprendimento
38/52
S The questions o the appropriate education content to
best ensure uture economic growth and how best to
equip students to ace an uncertain uture are also at the
core o reorms in some o the more successul school
systems, particularly in Asia. Singapores Proessor Lee
explains that o todays job titles compared to those
o 1995, many are very new; the skills are very new.
We anticipate that evolution will be ast into the uture.
For over a decade, his countrys Ministry o Education
has engaged in uture scanning to identiy the likely skills
needed in the coming years, and adjusted its oferings
to students accordingly. More important, since 1997,
says Proessor Lee, Singapore has shited away rom
teaching rote knowledge to a firm oundation in the
basics o maths, science, and literacy combined with an
inculcation o how to understand and apply inormation.
We eel it contributes toward the students acquiring
knowledge and skills o cognition and creativity attributes
which are very important in the 21st century landscape.
Both o these developments reflect an attitude thateducation systems need to be prepared or ongoing
change rather than seek a single, best end state.
No education system can remain static, writes
Singapores Prime Minister, Lee Hsien Loong, in
the oreword to a recent report on education and
geopolitics in the 21st century. The world is changing
rapidly. Technology is transorming our l ives. The skills
needed in the uture will be very diferent rom those
needed today.27
Singapore is not alone. Shanghai students finished first
in the latest PISA tests, but China is also shiting toward
a much greater emphasis on creativity. Proessor Zhao
explains that the countrys leadership believes the
economy is moving quickly rom a labour-intensive
one to a knowledge economy. It needs creative talent.
Indeed, he finds it ironic that China is moving more in
the direction o Western models even while politicians
in those countries sometimes praise that o traditional
Asian education. South Korean schools, meanwhile, arenow being encouraged to develop creativity, character
and collaboration.
Teaching people how to work together is indeed o
growing relevance to the economy. According to
Ms Parthasarathi, A lot o education in the second
hal o the 20th century has made children fiercely
27 Foreword to Michael Barber, Katelyn Donnelly
and Saad Rizvi, Oceans of innovation: The Atlantic,
the Pacific, global leadership and the future ofeducation, 2012.
No education system can remain static. The worldis changing rapidly. Technology is transorming ourlives. The skills needed in the uture will be verydiferent rom those needed today.
Lee Hsien Loong, Prime Minister o Singapore
7/30/2019 Curva dell'apprendimento
39/52
7/30/2019 Curva dell'apprendimento
40/52
7/30/2019 Curva dell'apprendimento
41/52
7/30/2019 Curva dell'apprendimento
42/52
41Pearson plc
7/30/2019 Curva dell'apprendimento
43/52
The methodology appendix describes in more detail
the Indexs construction and relevant data issues.
The broader message o this lengthy disclaimer is that
the Index is very much a first step. We hope that, as
understanding o the outcomes o education grows,the Index will become more complex and nuanced as
well as be populated with more robust and varied data.
For now, however, it is better to light a candle than curse
the statistical darkness.
Wh h h h
Given the attention paid to the results o international
education tests, the leading countries in the cognitive
skills category o the Index come as no surprise.
The top five Finland, Singapore, Hong Kong,
South Korea and Japan all score more than one
standard deviation above the norm in this part o the
Index. The educational attainment category, based on
literacy and graduation rates, tells a slightly diferent
story. Here South Korea leads, ollowed by the UK,
Finland, Poland and Ireland, with Japan, Hong Kong
and Singapore urther down the table. Because o their
strength in both measures, then, Finland and South
Korea are the clear overall leaders o the Index.These results mirror the conventional wisdom: already
in 2007, the BBC reerred to the two countries as
among the superpowers o education.28 But what do
these have in common that might help to identiy the
keys to educational success? On the ace o it, there is
remarkably little.
In many ways, it is hard to find two education systems
more diferent. South Koreas schools are requently
described as test-driven, with a rigid curriculum and anemphasis on rote learning. Most striking is the amount o
time spent in study. Once the ormal school day is over,
the majority o students go to private crammer schools,
orhagwons. According to OECD data, o 15-year-old
students or whom data was available in 2009, 68%
engaged in private study o the Korean language, 77% in
mathematics, 57% in science and 67% in other subjects.
In later years, students typically do ar more privately.
The government has become so worried about the
extent o these studies that it has banned hagwons
rom being open ater 10pm, but still needs to send
out patrols to shut down those which mask illegal,
ater-hour teaching by posing as sel-study libraries.
On the other hand Finland, in the words o Proessor
Schwartz, is a wonderul case study. Kids start school
later; school hours are shorter than most others; they
dont assign homework; their teachers are in ront o
kids less. By one estimate, Italians go to school three
years longer. The PISA data shows that very ew Finns
take out-o-school lessons either, and those who do
typically do worse on standardised tests, suggesting
that this is largely remedial help. Finally, the systemhas a reputation or being ocused on helping children
understand and apply knowledge, not merely repeat it.
The existing data also paint a picture o two distinct
approaches. In some cases, the systems are widely
diferent: average teacher salaries in South Korea are
over twice the national average, while those in Finland
are almost exactly average; pupil-teacher ratios, on the
other hand, are much higher in South Korea. Where
the two systems are similar, they are usually near theaverage or all countries in the Index. The only diference
is school choice, where both are highly restrictive.
That said, the vast amount o ater-school private
education in South Korea brings into question the
relevance o that metric.
The two systems, though, do share some important
aspects when examined closely. When you look
at both, you find nothing in common at first, says
Proessor Schleicher, but then find they are very similar
in outlook. One element o this is the importance
assigned to teaching and the eforts put into teacher
recruitment and training. As discussed above, the
practices o the two countries difer markedly, but the
status which teaching achieves and the resultant high
quality o instruction are similar. Proessor Schleicher
adds that both systems also have a high level o ambition
or students and a strong sense o accountability, but
again these are articulated diferently. In South Korea,
accountability is exam driven; in Finland, it is peeraccountability, but the impact is very similar.
28 Finland stays top o global class, 4 December
2007, http://news.bbc.co.uk/1/hi/7126562.stm
7/30/2019 Curva dell'apprendimento
44/52
43Pearson plc
7/30/2019 Curva dell'apprendimento
45/52
Finally, there are cultural parallels. The two societies
are highly supportive o both the school system itsel
and o education in general. O course, other countries
are also highly supportive o education, but what may
set Finland and South Korea apart is that in both, ideasabout education have also been shaped by a significant
underlying moral purpose.
Although discussions o Korean attitudes to education
requently reerence Conucian ideals, under a quarter
o South Koreans were even literate by the end o the
Korean War. In the decades that ollowed, education
was not just about sel-improvement: it was a way to
build the country, especially as the Japanese colonial
power had restricted the access o ethnic Koreansto schooling. The immediate cause o this drive has
disappeared, but it has helped inculcate a lasting ethic
o education which only strengthened the more
widespread attitude in Asia that learning is a moral duty
to the amily and society as well as a necessary means o
individual advancement.
In Finland, the ethos is diferent but no less powerul.
As Mr Mackay explains, that country has made
a commitment as a nation to invest in learning as a
way o liting its commitment to equity. They wish
to lit the learning o all people: it is about a moralpurpose that comes rom both a deeper cultural
level and a commitment at a political-social level.
In other words, education is seen as an act o
social justice.
Both o these moral purposes can cause diculties in
diferent ways. The high expectations and pressure
mean that studies regularly find South Korean teenagers
to be the least happy in the OECD. In Finland, the
egalitarian system seems less efective at helping highlytalented students to perorm to the best o their
ability than at making sure average results are high.
Nevertheless, the power o these attitudes in shaping
cultural norms and political decisions in ways that help
education attainment overall are undeniable. Mr Angula,
ater many years as a teacher, Minister o Education,
and Prime Minister, believes that the key ingredient
[in creating a successul education system] is or
everybody to be committed and to understand that
they are doing a public good.
The two societies [that score highest] are highly
supportive o both the school system itsel ando education in general.
7/30/2019 Curva dell'apprendimento
46/52
7/30/2019 Curva dell'apprendimento
47/52
7/30/2019 Curva dell'apprendimento
48/52
47Pearson plc
7/30/2019 Curva dell'apprendimento
49/52
national ministry or statistics bureau
data were sanity-checked and then used
i deemed internationally comparable.
C wh
In order to make indicators directlycomparable across all countries in the
Index, all values were normalised into
z-scores. This process enables the
comparison and aggregation o diferent
data sets (on diferent scales), and also
the scoring o countries on the basis
o their comparative perormance.
A z-score indicates how many standard
deviations an observation is above
or below the mean. To compute the
z-score, the EIU first calculated each
indicators mean and standard deviation
using the data or the countries in the
Index, and then the distance o the
observation rom the mean in terms o
standard deviations.
The overall Index score is the weighted
sum o the underlying two category
scores. Likewise, the category scores
are the weighted sum o the underlying
indicator scores. As recommended by
the Advisory Panel, the deault weightor the Index is two-thirds to cognitive
skills and one-third to educational
attainment. Within the cognitive skills
category, the Grade 8 tests score
accounts or 60% while the Grade 4
tests score accounts or 40% (Reading,
Maths and Science all account or
equal weights). Within the educational
attainment category, the literacy rate
and graduation rates account or equal
weights. The user can, however, change
the weightings and recalculate scores
according to personal preerence via the
website that accompanies this report.
A
Because indexes aggregate diferent data
sets on diferent scales rom diferent
sources, building them invariably
requires making a number o subjective
decisions. This index is no diferent. Eacharea or caution is descr ibed below.
Z- PISA, TIMSS PIRLS
It is important to note that, strictly
speaking, the z-scores or PISA, TIMSS
and PIRLS are not directly comparable.
The methodology applied both by
the OECD and the International
Association or the Evaluation o
Educational Achievement (IEA) to
calculate the perormance o theparticipating countries consists o
comparing the perormance o the
participating countries to the respective
mean perormance. (The countries
raw test scores beore normalisation
are not published; just their scores in
comparison to the other participants.)
Thus, which countries participate in
each test and how well they perorm in
comparison to the other participants
has a direct impact on the resultingfinal scores. Given that the sample o
countries that take the PISA, TIMSS
and PIRLS tests are not exactly the
same, there are limitations to the
comparability o their scores.
The EIU has chosen not to change these
scores to account or this lack o direct
comparability; however, it did consider
other options along the way. The main
alternative suggestion rom the Advisory
Panel was to use a pivot country inorder to transorm the z-scores o other
countries in comparison to that pivot
countrys z-score. Although this method
is used in some studies, ater substantial
consideration, the EIU decided not to
employ this method or the purpose o
an index. The resulting z-scores ater
transormation depend heavily on thechoice o pivot country; choosing one
country as a pivot over another afects
countries z-scores quite substantially.
The EIU did not eel it was in a position
to make such a choice. Despite these
limitations to test scores direct
comparability, the EIU believes that the
applied methodology is the least invasive
and most appropriate to aggregate
these scores.
G
Some members o the Advisory Panel
questioned the use o graduation
rates in the Index in that it is not clear
whether they add value as a comparative
indicator o education perormance.
Unlike test results and literacy rates,
standards to gaining an upper secondary
and tertiary degree do difer across
countries. Notwithstanding, the EIU
believes that graduation rates do add
value in evaluating a national educationalsystems perormance, as there is
common acceptance that national
education systems should aim or their
citizens to gain educational qualifications,
especially at the secondary level.
Including graduation rate data in the
Index thereore awards countries that
have put this aim into practice, albeit at
varying levels o quality.
Because o the variation in how countries
measure graduation rates, the EIU
ollowed the Panels suggestion in using
OECD graduation rate data, which use
one main definition. When OECD
data were not available, national ministry
or statistics bureau data were sanity-
checked and then used i deemed
comparable. In some cases, no dataon graduation rates were available.
In this case, the EIU awarded the
country the mean score or this
indicator. One disadvantage o giving
a country the mean score is that i in
reality it perorms worse than the
average in this indicator, the Index
boosts its score, and vice versa.
The EIU used the most recent data
available. Because graduation rates
are based on the pattern o graduationexisting at the time, they are sensitive to
changes in the educational system, such
as the addition o new programmes or
a change in programme duration. As an
extreme example, Portugals upper
secondary graduation rate increased
rom a range between 50% and 65%
in the early 2000s to 2008, to 104% in
2010, as a result o the governments
New Opportunities programme,
launched to provide a second chance orthose individuals who let school early
without a secondary diploma. In order
to treat countries consistently, the Index
takes the 2010 figure. Although this
inflates Portugals score in this indicator,
this inflation should eventually all out
o the Index should it be updated on
an annual or bi-annual basis. Given the
limitations o graduation rate data, the
EIU ollowed the Panels suggestion o
giving a smaller weighting (one-third) to
educational attainment.
48 Th L C 2012 App 1: Mhy h Q Cp Th L C
7/30/2019 Curva dell'apprendimento
50/52
It is also important to note that the
tertiary graduation rate indicator covers
only tertiary-type A programmes.
Tertiary-type B programmes are not
included. This methodology was chosen
largely because not all countries collectdata and organise their education
systems along the lines o A and B.
As per the OECD, tertiary-type
A programmes are largely theory-
based and are designed to provide
qualifications or entry into advanced
research programmes and proessions
with high requirements in knowledge
and skills. These programmes are
typically delivered by universities, and
their duration ranges rom three to five
years, or more at times. Tertiary-type
B programmes are classified at the
same academic level as those o type
A, but are oten shorter in duration
(usually two to three years). They are
generally not intended to lead to urther
university-level degrees, but rather to
lead directly to the labour market.
Although excluding tertiary-type B
programmes makes or a more relevant
comparison among countries, it alsoslightly disadvantages a number o
countries that have particularly high
type B graduation rates (as these rates
are not included). These countries
are Canada, Ireland, Japan and New
Zealand. Nonetheless, this exclusion
has a limited impact on these countries
ranking in the Index.
Oh
The EIU had wanted to include other
education perormance indicators in
the Index, such as how well national
education systems prepare students
or the labour market and theperormance o vocational studies.
However, data availability was a limiting
actor. The EIU ound that sucient
data were not available that isolates
educational attainment within labour
market outcomes; and internationally
comparable data on vocational studies
covering all countries in the Index were
not readily available either.
C
With the comprehensive-data
countries data rom the Data Bank, a
correlations exercise was undertaken
in order to test relationships across
countries between education inputs,
outputs and wider society. The EIU
tested or correlations between the
inputs to and outputs o education, the
inputs to education and socio-economic
environment indicators (as a proxy
or wider society), and the outputso education and socio-economic
environment indicators.
Dfi
hh
The correlation coecient is a
measure o the degree o linear
relationship between two variables.
While in regression the emphasis is on
predicting one variable rom the other,
in correlation the emphasis is on thedegree to which a linear model may
describe the relationship between two
variables. Importantly, the presence o a
correlation does not imply causality.
In order to ensure that relationships
being ound were indeed strong,
the EIU looked or at least a 0.65
level o correlation (the higher it is,the stronger the relationship). It is
important to acknowledge that some
social science research uses a lower
level o correlation, but the EIU wished
to maintain a high level to avoid finding
relationships between indicators that
might not be significant.
C
Correlation tests were conducted on
an indicator-by-indicator basis, betweentwo variables over time (on an annual
basis) and at three-year growth rates
(or example, the three-year growth
rate o 1999 (199699) against the three-
year growth rate o 2007 (200407)).
For the latter tests, adjustments were
made to include TIMSS and PIRLS tests
even though these are not taken every
three years (they are taken every our
and five years respectively). The EIU
used the same time lags across countrieson the same indicator, as per the Panels
suggestions.
When looking or evidence o a strong
correlation, the EIU sought a strong
relationship over time. For example,
although there may have been evidence
o a strong correlation between one
input variable in 1990 and an output
variable in 2005; a strong level o
correlation would also need to be ound
or 1991 and 2006, 1992 and 2007, and
so on, or at least a number o years. In
addition, correlation tests were only run
i there were at least 15 countries with
relevant data or both o the indicators
being assessed.F f h
The EIU did not find a great number
o strong relationships. Given the
complexity o education, this was not
totally surprising. However, other
actors may also account or the lack o
correlations. For one, not all indicators
were available going back 1520 years
in time. There was also a lack o data
availability or some countries (some
o this due to the Data Banks ocus
on ensuring that data being used were
internationally comparable). Finally,
other qualitative actors that are dicult
to measure, such as culture and the
quality o teaching, were not included in
the Data Bank. These actors may have a
significant impact on education outputs,
but the EIU was not able to take these
into account within the correlations
exercise.
7/30/2019 Curva dell'apprendimento
51/52
7/30/2019 Curva dell'apprendimento
52/52