ECDI2030 Technical Manual Sept 2023

THE EARLY CHILDHOOD DEVELOPMENT INDEX 2030
A NEW MEASURE OF EARLY CHILDHOOD DEVELOPMENT

© United Nations Children’s Fund (UNICEF), Division of Data,
Analytics, Planning and Monitoring, September 2023
Suggested citation: United Nations Children’s Fund, The Early

Childhood Development Index 2030: A new measure of early
childhood development, UNICEF, New York, 2023.
Acknowledgements
The preparation of this publication was led by Claudia Cappa
and Nicole Petrowski (Data and Analytics Section, UNICEF
Headquarters), along with Elga Filipa de Castro (formerly with Data
and Analytics Section, UNICEF Headquarters). Jean Marie Place
and Kathryn Werntz (independent consultants) assisted with the
preparation and revision of earlier drafts. Inputs were graciously
provided by Diego Armando Luna Bazaldua (World Bank), Marta
Rubio Codina (Inter-American Development Bank), Peter Halpin
(University of North Carolina), Adelle Pushparatnam (World Bank),
Ruth Argelia Vázquez Salas (National Institute of Public Health,
Mexico) and Nirmala Rao (The University of Hong Kong).
The report was edited by Tina Johnson, proofread by Lois Jensen

and designed by Era Porth (independent consultants).
Photo credit
Cover: © UNICEF/UNI333261/ Bhardwaj
© UNICEF/UN0312259/Sokol
3
Contents
Abbreviations 5
Executive summary 6
Introduction 9
Part One: Early childhood development measurement 10

ECD measurement tools: Past and present 12
MICS ECDI 13
Inclusion of ECD in the SDG monitoring framework 13
The ECDI2030 15
Part Two: Work process behind the ECDI2030 16
Part Three: Building the ECDI2030 20

Conceptual framework 21
Selecting the initial pool of items 24
Testing 25
Harmonizing a global dataset 32
Psychometric analyses 32
Standard-setting 36
Part Four: The final ECDI2030 42
Annexes 46
Endnotes 66
4 THE EARLY CHILDHOOD DEVELOPMENT INDEX 2030: A NEW MEASURE OF EARLY CHILDHOOD DEVELOPMENT
Abbreviations
BSID-III Bayley Scales of Infant and Toddler Development, Third Edition
CAPI computer-assisted personal interview
CREDI Caregiver-Reported Early Development Index
DHS Demographic and Health Surveys
EAP-ECDS East Asia-Pacific Early Child Development Scales
ECD early childhood development
ECDI Early Childhood Development Index
HKU The University of Hong Kong
IAEG-ECD Inter-agency Expert Group on Early Childhood Development Measurement
IAEG-SDGs Inter-agency and Expert Group on Sustainable Development Goal Indicators
IDELA International Development and Early Learning Assessment
IRT item response theory
MDGs Millennium Development Goals
MELQO Measuring Early Learning Quality and Outcomes
MICS Multiple Indicator Cluster Surveys
NLSCY National Longitudinal Study of Children and Youth
OECD Organisation for Economic Co-operation and Development
PRIDI Regional Project on Child Development Indicators
TAG Technical Advisory Group
SDGs Sustainable Development Goals
SMEs subject matter experts
SYC Survey of Young Canadians
UNESCO United Nations Educational, Scientific and Cultural Organization
UNICEF United Nations Children’s Fund
WHO World Health Organization
5
Executive summary
In 2015, UNICEF initiated a process of methodological work that
involved extensive consultations with experts, partner agencies
and national statistical authorities. Over the following five years,
a sequence of carefully planned technical steps were executed
incorporating both qualitative and quantitative methods to identify
the best items to measure SDG indicator 4.2.1: “Proportion of
children aged 24-59 months who are developmentally on track in
health, learning and psychosocial well-being, by sex”. This process
led to the development of the Early Childhood Development Index
2030 (ECDI2030).
The ECDI2030 addresses the need for nationally representative

and internationally comparable data on early childhood
development, collected in a standardized way. Data collected
with the ECDI2030 can be used by countries to monitor progress
against SDG Target 4.2.
The ECDI2030 can be integrated into existing national data

collection efforts. This is a public good and freely accessible to
all those countries interested in undertaking data collection on
ECD outcomes at the population level. The ECDI2030 has been
translated into a number of languages, including the six official UN
languages. It is accompanied by standard guidance and a set of
implementation tools that include interviewer guidelines, training
materials, syntaxes, tabulation plans and templates for reporting.
The ECDI2030 and accompanying implementation materials can
be found on the dedicated resource page.1
TARGET 4.2:
By 2030, ensure that all girls and boys
have access to quality early childhood
development, care and pre-primary education
so that they are ready for primary education
© UNICEF/UN067557/Helin
7
© UNICEF/UN0498266/Cus
Introduction
Early childhood has been recognized by the international community – and proven by decades of
research – to be a crucial time in a child’s life. During these early years, children develop the skills
and capacities they need for future learning and growth. Generally understood to apply to children
aged 8 years and younger, early childhood development (ECD) encompasses motor, cognitive,
language, socio-emotional and self-regulatory skills and capacities.
The importance of ECD is explicitly emphasized in the Convention on the Rights of the Child,
which has a special set of recommendations for implementing children’s rights during early
childhood. That it is a central component of global and national development is further evident from
its inclusion in the Sustainable Development Goals (SDGs). SDG Target 4.2 calls for countries to
“Ensure that, by 2030, all girls and boys have access to quality early childhood development, care
and pre-primary education so that they are ready for primary education”. One of the indicators for
tracking progress on Target 4.2 is indicator 4.2.1: “Proportion of children aged 24-59 months who
are developmentally on track in health, learning and psychosocial well-being, by sex”.
The SDGs, adopted by the international community in 2015, supersede the Millennium
Development Goals (MDGs) and significantly expand on them. They place more emphasis on
previously overlooked issues, including ECD, and make all governments, not just those of low- and
middle-income countries, accountable for achieving them.
The Inter-agency and Expert Group on SDG Indicators (IAEG-SDGs) initially classified indicator 4.2.1
as tier III (i.e., there was no internationally established way to measure it) and called for further
methodological work to develop a measure that can be used across countries for the purpose of
global monitoring and reporting. In order to assist governments in tracking progress on Target 4.2 –
and as custodian agency for indicator 4.2.1 – the United Nations Children’s Fund (UNICEF) led the
technical research and design of an improved ECD measurement tool that is universally relevant
for children aged 24 to 59 months, taking into account cultural and social differences in their
development. To oversee the work, UNICEF established the Inter-agency Expert Group on ECD
Measurement (IAEG-ECD), a taskforce that brought together representatives of national statistical
offices, experts from development organizations and academics.
This report describes how UNICEF, along with the IAEG-ECD, developed a measurement tool
that can be used to report against SDG indicator 4.2.1 and inform the development of evidence-
based ECD policies and programmes. The work on the ECDI2030 began in 2015 and was
completed in 2020.
The report is divided into four parts. Part One provides an overview of ECD, and its importance
in the SDGs and in informing evidence-based policymaking, and briefly looks at existing ECD
measures including an introduction to the ECDI2030. Part Two discusses the work process for
the development of the ECDI2030, while Part Three explains in detail the conceptual framework,
domains, subdomains and constructs of the new measure. The methodological work in
establishing its reliability and validity is also presented, along with an overview of the testing
conducted. Finally, Part Four introduces the final version of the ECDI2030 and its properties and
explains how to calculate the related indicator.
9
PART ONE
Early childhood
development
measurement
What is early childhood development (ECD)?
Early childhood development (ECD) is a maturational and iterative process involving an ordered
progression of motor, cognitive, language, socio-emotional and self-regulatory skills and capacities
across the first few years of life. Generally understood to apply to children up to the age of 8,
it is viewed as holistic and encompasses health, nutrition, social protection and stimulation, as
articulated in the Nurturing Care Framework developed jointly by UNICEF, the World Bank Group
and the World Health Organization (WHO).2
Why is this period of development so important?

Science has demonstrated that experiences in early childhood have lifelong implications for
learning and well-being. During these early years, from conception to infancy and early childhood,
a child’s newly developing brain is highly plastic and responsive to change. This is evidenced by the
billions of integrated neural circuits established through the interaction of genetics, environment
and experience. Therefore, ECD is a vital period for children that sets the stage for lifelong thriving
and is a critical window for human capital investment.
A child’s experiences in the early years are linked to educational achievement later in life as well
as the development of skills, capabilities and productivity in adulthood. This makes investing in
ECD one of the most critical and cost-effective investments a country can make. In fact, economic
analyses have found that investing in the early years of a child’s life yields some of the highest
rates of return to families, societies and countries.3
Do all children develop the same skills and at the same rate regardless of
where they live?
While the overall developmental process is similar across cultures, children develop at different
speeds and may reach developmental milestones at different times. What is considered ‘normal’
child development also varies across cultures and environments, since expectations and parenting
strategies may differ between countries as well as among cultural, ethnic or religious groups within
the same country. This makes measuring ECD in a way that allows for cross-cultural comparisons
especially challenging.
Why is the measurement of ECD so complicated?

Approaches to measuring ECD at scale have traditionally been based on screening and/or
diagnostic tests that depend on highly trained professionals and substantial administration time to
generate valid information, making them inadequate for large-scale population monitoring.
The multi-dimensional nature of ECD further complicates the task of generating accurate data at
population level through household surveys since specifically designed and validated instruments
that can be implemented in a standardized way are required to generate robust and comparable
data.
Additional constraints of population-level measurement in the context of multi-topic household

surveys imply that instruments need to be short and that enumerators can effectively be trained in
their administration.
11
ECD measurement tools: Past and When UNICEF started the process of creating a
global, survey-based measurement tool for ECD
present
in 2006, there were only a few population-based
Interest in ECD surged at the turn of the twenty-first measures that could be used as reference. However,
century. The MDGs, established in 2000, called for the landscape has changed since that time, with a
governments of low- and middle-income countries to number of groups working to develop, test and
achieve and report on progress on eight goals related validate tools and measures of ECD for various
to human development and quality of life. While the purposes. Most notably and of greatest relevance,
MDGs did not have any specific goals or targets on these efforts include:
ECD, they did include aspects that contribute to early
child development, such as health (i.e., reducing child • Early Development Instrument (EDI), developed
by McMaster University
mortality) and investing in education.
• International Development and Early Learning
Despite consensus on the importance of inputs Assessment (IDELA), developed by Save the
affecting ECD in the MDGs, a population-based Children
measure was not readily available in 2000, particularly
for low- and middle-income countries. This was • The Regional Project on Child Development
probably due, in part, to the lack of a specific mention Indicators (PRIDI), developed by the Inter-American
of ECD in the goals and hence no globally coordinated Development Bank
drive, or investment, to develop such a measure.
• East Asia-Pacific Early Child Development Scales
(EAP-ECDS), developed by The University of Hong
The lack of an ECD measurement tool was problematic Kong (HKU) and UNICEF East Asia and the Pacific
since population-based data provide one of the most
effective ways to draw policy attention to the situation • The Measure of Development of Early Learning
of children, propel action and ensure progress in (MODEL), developed by the Measuring Early
creating equity in outcomes for all children. Learning Quality and Outcomes (MELQO) initiative
led by UNICEF, the United Nations Educational,
In recognition of this gap, UNICEF initiated a technical Scientific and Cultural Organization (UNESCO),
process to create a global measure of developmental World Bank and Brookings Institute
status in early childhood for use at population level,
• Caregiver-Reported Early Development Index
the Early Childhood Development Index (ECDI),
(CREDI), developed by Harvard University
within the context of the Multiple Indicator Cluster
Survey (MICS) programme. MICS surveys produce • Global Scales for Early Development (GSED),
nationally representative and comparable data on a developed by WHO
number of key indicators of the health and well-being
of children and their families. • International Early Learning and Child Well-being
Study from the Organisation for Economic Co-
operation and Development (OECD)
Global monitoring requires comparable and
standardized statistics. To achieve this, countries While all of the tools listed here are designed to
need a common set of data collection tools, including capture children’s development, they vary greatly
a core set of questions and a standard protocol to in terms of the domains covered, methodologies
minimize the variability of key aspects related to employed and applicable age groups.4 Most of these
implementation (such as measurement error due to are also limited in one or more of the following ways:
varying levels of capacity in implementing agencies/
partners). These tools need to be suitable for • relying on direct assessment of children and/or
population-level measurement of ECD outcomes and teachers’ reports
include questions that can be easily integrated into • not designed to produce population-level estimates
existing national data collection efforts.
• lacking sufficient testing and implementation at
scale across a large cross-section of countries
• designed for individual assessment or programme Inclusion of ECD in the SDG
evaluation rather than population-based measure-
monitoring framework
ment and global monitoring
In 2011, a series of published papers on early
childhood development7 and its influence on human
MICS ECDI development indicators later in life augmented the
The construction of the MICS ECDI in 2006 involved global discussion around the need to include specific
work with countries and partners using a multi- targets related to ECD in the MDGs’ successor, the
method approach. To establish the reliability and Sustainable Development Goals (SDGs).
validity of the survey items, psychometric properties
were examined at each stage in the development In anticipation of the adoption of the SDGs,
process, which also included an extensive literature UNICEF convened a technical consultation on the
review and pilot testing.5 measurement of ECD with some 35 key experts and
partners in January 2015. The main objectives of the
The MICS ECDI comprises four domains: literacy- meeting, among others, were to discuss the need
numeracy, physical development, social-emotional and possibilities for creating new measures of ECD,
development and learning. It contains 10 items to to reflect on opportunities for future methodological
measure whether a child is developmentally on track work and to discuss and generate consensus on a set
in each domain: of ECD indicators to be recommended for inclusion in
the post-2015 agenda.
Literacy-numeracy: The child can do at least two of
the following: identify/name at least 10 letters of the Following the creation of the Inter-agency and
alphabet; read at least 4 simple, popular words; and/ Expert Group on SDG Indicators (IAEG-SDGs) in
or know the name and recognize the symbols of all March 2015 during the 46th session of the UN
numbers from 1 to 10. Statistical Commission, work began on developing
and implementing a global indicator framework to
Physical: The child can pick up a small object with accompany the SDGs, which were formally adopted
two fingers, such as a stick or rock from the ground, by the UN General Assembly in September 2015.
and/or the mother/primary caregiver does not indicate As part of the global consultation process, UNICEF
that the child is sometimes too sick to play. submitted a proposal, with support and endorsement
from other UN agencies and partners in the global
Social-emotional: The child demonstrates at least ECD community, for inclusion of a dedicated indicator
two of the following: gets along well with other on early childhood development.
children; does not kick, bite or hit other children; does
not get distracted easily. These advocacy efforts resulted in the successful
inclusion of an ECD indicator (4.2.1) in the global
Learning: The child follows simple directions on indicator framework developed by the IAEG-SDGs,
how to do something correctly and/or when given agreed on at the 48th session of the UN Statistical
something to do is able to do it independently. Commission in March 2017 and subsequently adopted
by the UN General Assembly in July 2017. In its
original formulation, 4.2.1 referred to the “Proportion
The MICS ECDI was introduced in 2009 in the
of children under 5 who are developmentally on track
fourth round of MICS (MICS4) and has since also
in health, learning and psychosocial well-being, by
been collected in MICS5 and MICS6 as part of the
sex”.
questionnaire for children under 5, as well as in a
number of Demographic and Health Surveys (DHS).
Data on the MICS ECDI are now available for more To support implementation of the indicator frame-
than 80 (mostly low- and middle-income) countries. work, all SDG indicators were classified into three
tiers on the basis of their level of methodological
The process behind the development of the MICS
development and the availability of data at the global
ECDI is documented elsewhere.6
level. SDG indicator 4.2.1 was initially classified
as tier III, meaning the IAEG-SDGs deemed in
2017 that there was no internationally established
13
methodology available for its measurement. A new Beirut in March 2019, it reviewed workplans and tier
measurement tool therefore needed to be developed reclassification requests for tier III indicators. Indicator
and validated. As the custodian agency for indicator 4.2.1 was at risk of being dropped from the SDG global
4.2.1, UNICEF had the mandate to lead the technical monitoring framework because it is not possible
and methodological work on the development of an to have an indicator without an available measure.
improved measure of ECD outcomes. During this meeting, and in recognition of the efforts
and significant amount of work accomplished by
Although the MICS ECDI was significant in that it UNICEF and partners to develop such a measure, the
IAEG-SDGs approved the reclassification of indicator
was one of the first international population-based
4.2.1 as a multi-tier indicator,8 with the tier II portion
measures of early childhood development, it was
referring to children aged 24-59 months while the tier
not sufficient to measure the new dimensions and
III portion would continue to apply for children aged
breadth of ECD outlined in the SDG indicator. Beyond
0-23 months.
measuring a narrower scope of ECD, it was limited
in that it was designed only for children aged 3 to 4
This decision was subsequently re-evaluated at the
years. However, in the absence of a measurement IAEG-SDGs’ 11th meeting in November 2019. The
tool fully aligned with SDG indicator 4.2.1, the MICS Expert Group then recommended, as part of the
ECDI was recommended as an interim proxy measure 2020 comprehensive review process, to retain only
for global monitoring and reporting. the portion of indicator 4.2.1 referring to children
aged 24-59 months since methodological work being
As part of the process of developing and refining led by WHO to develop a population-level measure of
the global indicator framework, the IAEG-SDGs ECD outcomes for children aged 0-23 months had not
has met regularly since 2015. At its 9th meeting in been finalized.
© UNICEF/UNI211322/Rusanganwa
This recommendation by the IAEG-SDGs to delete Second, the SDGs call for all governments to
the portion of indicator 4.2.1 referring to children monitor and report on the indicators, meaning that
aged 0-23 months (which was still classified as tier the ECDI2030 needed to be universally relevant
III) was approved by the UN Statistical Commission for application and use in all countries, whereas
at its 51st session in March 2020. The Expert Group the MICS ECDI was developed primarily for use in
subsequently approved refinement of the indicator low- and middle-income countries. Therefore, the
name for 4.2.1 as: “Proportion of children aged 24- methodological work paid special attention to issues
59 months who are developmentally on track in of cultural and social relevancy.
health, learning and psychosocial well-being, by sex”.
Approval of this revised indicator name came from the The ECDI2030 was also designed, as was the
UN Statistical Commission during its 52nd session in MICS ECDI, to rely only on caregiver reports, not
March 2021. direct assessment. This decision responded to the
expressed need from countries for a relatively simple
The ECDI2030 and easy-to-administer tool that could be integrated
into existing data collection efforts and would not
While SDG monitoring and reporting was the create an added burden in terms of time, resources
motivation for developing a new measurement tool, and capacity.
this work was also undertaken in response to the
growth in the field of ECD measurement and recent
Further, a key part of the methodological work has
advances in research on child development. There
also been the establishment of formal collaborative
was also heightened interest among governments,
partnerships, including the creation of an Inter-agency
donors and the wider international community in
Expert Group on ECD Measurement (IAEG-ECD) as
collecting better data in order to inform ECD policies
well as a Technical Advisory Group (TAG). Annex A
and programmes. The ECDI2030 serves both
lists the membership of these two groups. See Part
purposes: to monitor and report on the SDGs; and
Two for more details on these partnerships.
to assist governments and other bodies in informing
policy and programmes.
The ECDI2030 is a population-level measure that can
be integrated into international household survey
The ECDI2030 measures the three domains
programmes (such as the MICS and DHS) or other
covered by SDG indicator 4.2.1: health, learning and
national data collection efforts. Therefore, similar to
psychosocial well-being. Some items are domain-
specific, while others tap into multiple constructs and the MICS ECDI, the ECDI2030 has a comparative
cross domains or subdomains. Rather than producing advantage in that it will not require the implementation
estimates of individual domains, the ECDI2030 of separate, dedicated survey efforts, which are often
generates a single overall prevalence estimate for all time and resource intensive.
three domains. This is what is needed by countries to
estimate the percentage of children developmentally Moreover, by embedding the ECDI2030 in existing
on track (i.e., to report on SDG indicator 4.2.1). data collection efforts, countries will be able to
produce disaggregated data on the ECDI2030 by key
The ECDI2030 was created to measure progress demographic and socio-economic variables. This is
against a specific SDG indicator (i.e., 4.2.1) and not always possible with other ECD measurement
therefore had to align with the indicator in a number tools and is key to addressing the ‘leave no one
of ways. behind’ agenda in the SDGs.
First, because the SDG indicator outlines three While the ECDI2030 provides governments with a
domains of ECD, UNICEF and its partners had tool to report on SDG indicator 4.2.1 and will produce
to develop a new conceptual framework built comparable and representative prevalence estimates
around these, with health being a domain that at the national level, it is not intended to evaluate
was insufficiently addressed in the MICS ECDI. programmes or interventions or to conduct clinical
Constructing a new conceptual framework also assessments of individual children.
required determining new constructs, subdomains
and items.
15
PART TWO
Work process
behind the
ECDI2030
The ECDI2030 is the result of a carefully planned methodological process that involved extensive
consultations with experts, partner agencies and national statistical authorities to generate the
shortest possible set of items to measure the core domains and subdomains of early childhood
development (ECD) in children aged 24 to 59 months.
In light of UNICEF’s mandate to undertake the methodological work of developing a new measure
of ECD outcomes, a two-day technical consultation was held in September 2016 with academic
and technical experts as well as key partners. Participants at the meeting formed a Technical
Advisory Group (TAG) composed of 11 academics and researchers from selected institutes,
agencies and universities. Members of the TAG have technical expertise in the measurement of
ECD and former experience with tool development and/or technical expertise in tool/instrument
testing and validation. An important outcome of the meeting was a well-defined plan with regards
to next steps in the process and a timeline for developing the ECDI2030.
To oversee the work, UNICEF formally established the Inter-agency Expert Group on ECD
Measurement (IAEG-ECD) in March 2017 at a side event to the UN Statistical Commission. The
IAEG-ECD was a global inter-agency advisory and coordination body whose overarching purpose
was to oversee the revision, testing and validation of the ECDI2030. The IAEG-ECD was chaired by
UNICEF and included partner agencies (OECD, UNESCO Institute for Statistics, WHO and World
Bank) as well as regional organizations (Inter-American Development Bank), international non-
governmental organizations (Save the Children), national research institutes (National Institute of
Public Health of Mexico) and national statistical offices (Palestinian Central Bureau of Statistics and
Statistics Canada). The IAEG-ECD was advised by and collaborated with the TAG.
Over the course of five years, UNICEF hosted a series of technical consultations with the IAEG-
ECD and TAG, as outlined in Figure 1. Between the in-person meetings, UNICEF also hosted virtual
webinars to regularly update the IAEG-ECD and TAG members on the status of the work and gain
their inputs at key moments of decision-making.
17
Figure 1
In-person technical consultations of the TAG and IAEG-ECD between 2016 and 2019
SEPTEMBER 2016 JANUARY 2018 NOVEMBER 2018 JUNE 2019
• Review the MICS ECDI • Review results of • Review results from • Review and discuss final
and discuss areas for cognitive testing in field test analyses and results
possible improvement four countries on item selection and
or revision • Gain consensus on come to a consensus
• Discuss implications next steps and timeline on the final draft set
• Explore the possible for modifying item set for constructing the
replacement of certain prior to field testing final measure • Review and discuss
domains or items in the results from the pilot
MICS ECDI, particularly of the standard-
with respect to the setting exercise
physical and social-
emotional domains • Discuss scoring of
the new measure
• Explore the possibility and development
of lowering the age of of the overall
eligible children and ‘performance profile’
related implications
on questionaire • Discuss plans for the
content/format global panel of the
standard-setting exercise
• Explore the possibility
of revising the analytical
plan for construction of
the total score, specifically
with respect to the weight
assigned to individual
items within each domain
as well as the contribution
of number of items
within each domain
• Identify next steps and
plans for validation,
including cognitive
and field testing
The methodological work involved a selection process and age coverage within each domain. In other
to identify the best items across 12 core subdomains words, all efforts were made to ensure that, after
within the three general domains of health, learning dropping any item, there were still items allowing
and psychosocial well-being (Figure 2). Conceptual measurement of all the conceptually identified
and psychometric criteria used for item selection subdomains for a given domain. In addition, age-
included the following: coverage was achieved by ensuring that selected
items resulted in a good distribution in terms of
• The initial pool of questions was selected from
item difficulty across all months of age.
previously validated instruments that produced
reliable results across different cultural and • Psychometric criteria included item difficulty,
socio-economic contexts and had relevance for discrimination, overall reliability and specificity in
policymaking. terms of maternal report versus direct assessment
• All questions were further tested, improved or of the child’s development.
discarded on the basis of results from cognitive
• Once the final set of 20 questions was identified,
testing in four countries, results from dedicated
a global panel of experts on ECD participated in
field testing in population-based household surveys
a standard-setting exercise to define the number
carried out in three countries and harmonized data
of milestones expected for each age group. This
points from an additional 30 countries.
led to the definition of the ECDI2030 cut scores,
• Successive rounds of item selection based on which identify children who are developmentally
psychometric testing were undertaken. Each on track in health, learning and psychosocial well-
round was carried out respecting content coverage being.
Figure 2
Work process behind development of the ECDI2030
LITERATURE REVIEW
Comprehensive review and expert rating on 500+ items from 20+ instruments (March 2017)
DRAFT SET OF QUESTIONS
Evaluating items in terms of comprehension, interpretation and cultural adequacy through cognitive testing in Bulgaria, Mexico, Uganda
and the United States (June-October 2017)
REVISED DRAFT SET OF QUESTIONS
Testing the administration of items in population-based surveys in Belize, Mexico and State of Palestine (June 2018-February 2019)
Harmonizing a global dataset with data from the ECDI2030 field tests in three countries and from an additional 30 countries
Identifying the best items using discrimination and difficulty properties
Undertaking further analyses to identify the minimum set of items fulfilling content and age coverage criteria
FINAL SET OF QUESTIONS
Defining cut scores through a standard-setting exercise (August 2019)
CONSTRUCTION OF THE INDEX
Translating the module into the six official UN languages
Developing manuals, training material, syntaxes for data analysis, tabulation plans and templates for reporting
COUNTRY IMPLEMENTATION
19
PART THREE
Building the
ECDI2030
Conceptual framework
Early childhood development (ECD) is considered a holistic process in that capabilities within
different domains are highly correlated and overlapping, particularly among very young children.
Therefore, the model selected for the conceptual framework assumes one underlying latent
construct of development.
The formation of the domains involved determining the most salient subdomains and constructs
within each domain based on a review of the research literature and existing evidence. The process
was also informed by substantial input from the Technical Advisory Group (TAG) through a desk
review and expert rating of many existing tools and measures of ECD to identify a pool of potential
items to capture the subdomains and their constructs. The initial conceptual framework is depicted
in Figure 3.
Figure 3
Initial conceptual framework of the ECDI2030
Learning
Early numeracy* Language, literacy and communication* Fine motor development*
• Counting (verbal and set production) • Letter/alphabet knowledge

• Number identification • Phonological awareness
• Number/size discrimination • Expressive language
• Receptive language
Executive function*
Mental flexibility
Working memory
Inhibition
Approaches to learning/play
Note: The executive function and approaches to learning/play subdomains are cross-cutting areas impacting each of the
other subdomains of learning.
Psychosocial well-being
Social development* Emotional development*
• Social competence • Empathy

• Relationship skills • Emotion knowledge
• Interpersonal conflict resolution
• Prosocial behaviour
• Social cognition
• Theory of mind
Self-regulation
Note: Self-regulation is a cross-cutting subdomain impacting the other subdomains of psychosocial well-being.
Health
Self-care* Gross motor development*
Note: Subdomains marked with an asterisk (*) were retained in the final ECDI2030.
21
A general description of each domain, subdomain expressive language/communication, pre-writing
and associated constructs follows, while additional and fine motor skills such as holding pens/chopsticks
details on the importance of the constructs to ECD and writing one’s name, among other things.16
and supporting research evidence in the learning and Communication is defined as “the child’s ability to
psychosocial domains are included in Annex B. verbally and non-verbally express needs, preferences,
as well as emotions, and to listen and respond to the
Learning communications of others”.17
Learning has been broadly conceptualized as
comprising a number of subdomains that reflect The language, literacy and communication subdomain
early pre-academic skills and competencies critical to is comprised of the following four constructs: letter/
the later acquisition of more complex skills as well alphabet knowledge, phonological awareness,
as academic success. Research literature and broad expressive language and receptive language.
agreement among technical experts has identified the
following five core subdomains of learning: (1) early It is important to measure language, literacy and
numeracy, (2) language, literacy and communication, communication because the early ability to read is
(3) fine motor development, (4) executive function directly linked with later reading ability and reading
and (5) approaches to learning/play. achievement.18 Vocabulary, knowledge of letters,
words and beginning and end sounds in particular
The executive function and approaches to learning/play have been found to predict later learning.19 These
subdomains are considered cross-cutting underlying skills are also associated with varied and rich verbal
areas impacting each of the other subdomains. interactions with parents, teachers and peers and with
the availability of books in the home environment.20
Early numeracy
The subdomain of early numeracy covers ‘numeric Further, it is important to note that this subdomain and
concept’, ‘number knowledge’ and ‘identification, its constructs can be categorized and differentiated
classification and sorting’9 as well as ‘comparison by age. For example, it is especially important to
and measurement’, ‘patterns’ and ‘geometry’.10 The measure vocabulary for children under 3 years of age,
constructs within this subdomain considered on the given it is a meaningful proxy for overall language
ECDI2030 include: counting (both verbal and set development.21 For older children aged 4 to 5 years,
production), number identification and number/size it has been shown in high-income countries that this
discrimination. age group’s language scores are linked with higher
school achievement between the ages of 6 and 15
Measuring early numeracy during early childhood years.22
is especially important because number sense in
infancy predicts the math abilities of preschoolers.11 Fine motor development
Early math skills are also the basis for readiness for The development of fine motor skills represents
more formal and complex math instruction in later mastery of and control over small, precise
primary schooling12 – an especially important point movements, primarily through the use of the hands
since early math skills are the strongest predictors and fingers.23 Typical skills developed in the early
of later school achievement, stronger even than early years include the ability to hold writing utensils (such
reading or early attention skills.13 Moreover, early as pencils or pens), using utensils for eating (such as
math skills are important to measure since they have spoons or chopsticks), writing one’s name, drawing
been shown to be a more powerful predictor of later and picking up/manipulating small objects.
reading skills than early reading in predicting later
math skills.14 The TAG initially decided to include this subdomain
as a component of learning instead of placing it
Language, literacy and communication under the health domain (which includes gross
The area of language on its own includes the motor development, as outlined below) because
“knowledge and use of words, both in print and in of its conceptual relevance to learning and the
oral form”,15 while language and literacy together can documented associations that it has with later
include letter identification, reading, receptive and academic and learning success.24 In addition to
demonstrating general physical development, fine Psychosocial well-being
motor development is also an indicator of executive While the literature offers diverse interpretations and
function and overall neurocognition.25 definitions, this domain has been conceptualized
as referring broadly to aspects of children’s social
Executive function and emotional development as well as the absence
The subdomain of executive function refers to the of mental health disorders. It captures skills and
ability to concentrate, focus, follow directions and competencies related to forming and maintaining
inhibit reactions.26 This definition has been elaborated healthy interpersonal relationships with adults and
on to also include goal formation, planning, carrying peers, as well as regulating and expressing emotions
out goal-directed plans and effective performance.27 in socially and culturally appropriate ways.
While the roots of rudimentary executive function
are discernable in the infant and toddler years,28 it Significant associations have been found between
predominately develops during the preschool years.29 psychosocial well-being and “outcomes in education,
Within the executive function subdomain, the employment, criminal activity, substance use, and
conceptual framework considers the following three mental health […], even after controlling for important
constructs: mental flexibility, working memory and child, family, and contextual characteristics”.36
inhibition.
Pursuant to the literature review and technical
Executive function is considered a cross-cutting consultations, three broad subdomains were
subdomain, posited by some to overlap with self- identified: (1) social development, (2) emotional
regulation.30 It is important to measure executive development and (3) self-regulation, which is
function because, together with self-regulation, it is proposed as a cross-cutting subdomain.
associated with many skills required for kindergarten,
including the ability to attend selectively, show Social development
appropriate social responses and stay engaged in The definition of social development refers to “the
academic tasks.31 Executive function also predicts ability of young children to interact and sustain
later school achievement, particularly in reading and relationships with others”.37 Six constructs were
math,32 and it may be especially important for keeping chosen to capture this subdomain: social competence,
children engaged in school over time.33 Finally, relationship skills, interpersonal conflict resolution,
executive function may be one of the most critical prosocial behaviour, social cognition and theory of
areas to measure due to its cross-cultural relevance. mind.
Approaches to learning/play The differences between the constructs of social

This subdomain is considered cross-cutting and cognition, theory of mind and empathy are nuanced.
includes skills and behaviour linked to the processes All three can refer to the ability to take the perspective
of learning; in other words, how children learn. The of other individuals, understand their emotional
literature on this subdomain indicates that it includes reactions and coordinate that into socially desirable
interest in learning, curiosity, creativity and initiative34 interactions.38 However, there are challenges in
as well as emotion regulation, attention and comparability due to differing frameworks and
persistence.35 terminologies. Therefore, these three constructs were
considered separately (with empathy categorized
Measuring approaches to learning in ECD is under the emotional development subdomain).
especially important because the development of
skills in its subcomponents is predictive of academic Emotional development
achievement in the later elementary grades and Emotional development refers to children’s
strengthens children’s ability to take advantage of the feelings about themselves and others and includes
learning opportunities that present themselves inside characteristics such as self-control, self-efficacy (i.e.,
and outside of school. the sense of being able to affect events) and the
ability to properly interpret the emotions of others.39
The two constructs that compose this subdomain on
the ECDI2030 are empathy and emotion knowledge.
23
Self-regulation citing reasons that stunting is (a) already a separate
Self-regulation is defined as “a child’s ability to SDG indicator and (b) considered a risk to ECD, not
recognize and control impulses, manage stress an outcome of it and that (c) psychometric analyses
and emotions, and exert self-control”.40 It also revealed it was not mapping onto the domain of
encompasses the ability to follow socially prescribed health. Immunization was also subsequently dropped
norms and rules, to be goal-directed and to be able as it is also a separate SDG indicator and was felt to be
to delay gratification.41 Therefore, it is considered as an input (rather than an outcome) that contributes to
a cross-cutting subdomain (i.e., present in both the healthy child development and also reflects coverage
social and emotional subdomains of psychosocial of services.
well-being).
The TAG also decided to include self-care as a
Measuring self-regulation in ECD is especially subdomain, which is linked to motor development but
important because it impacts children’s ability to was seen as an important developmental outcome
focus attention, control impulses or emotions and contributing to child health.
navigate relationships with others42 in order to cope
effectively with environmental demands.43 It has The initial conceptual framework was refined and
also been shown that self-regulation in childhood is reorganized in light of subsequent item testing and
associated with later mental health, life satisfaction analysis (for example, the fine motor development
and well-being, income and labour market outcomes, subdomain was finally placed under the health
measures of physical health, obesity, smoking, crime domain and some subdomains were dropped, such
and mortality.44 as approaches to learning/play).
Health
The development of the health domain has followed Selecting the initial pool of items
a slightly different approach than the other two After identifying the most relevant subdomains and
domains. constructs for all three domains based on expert
consultation, and taking into account the conceptual
Following a review of existing conceptual frame- framework, the next step was to create an inventory
works on child health, a literature review and an of items from existing tools that aim to measure child
assessment of data availability, there was an initial development at population level (as opposed to those
recommendation to create a composite indicator developed for other purposes such as individual child
of health comprising gross motor development, assessment). For tools with both caregiver-reported
immunization and stunting. The criteria used to select and direct assessment items, only those items based
these components of health were that they should: on the former were retained for consideration.
• directly relate to children rather than inputs to

The resulting inventory included more than 500 items
families/environment
drawn from over 20 instruments and tools. This initial
• measure current inputs/outcomes and not past bank of items included a number of direct assessment
ones tools such as the IDELA and EAP-ECDS. These were
eventually removed given the parameter of identifying
• be applicable over a wide age range within the only items that rely on caregiver or teacher reports.
group of children under 5 years of age The resulting set of items was grouped according to
domain, subdomain and construct.
• be equivalent and relevant across contexts for all
children Members of the TAG who are academics and
• be objective if possible technical experts in the field of ECD measurement
then participated in an exercise to rate these
Stunting was later removed from the domain following existing items against several predefined criteria that
an analysis and discussion of the results from the included whether the item has policy relevance, is
field test in Mexico (more details on this below), intervenable/actionable, has cross-cultural applic-
ability, is pragmatic/easy to administer, can accurately comprehend the underlying construct in survey
be reported on by caregivers and is known to have questions, recall needed information, judge their
strong empirical evidence/predictive validity. answer and, finally, map their answer onto one of the
available response categories. The data generated
Feedback received from the experts was consolidated from this process provide insights into whether a
into a single master database, and all items that question and its response options are capturing
were either not recommended by any experts or the original intent and meaning of the questions.
recommended by only one expert were removed. This Cognitive testing is a crucial step in designing new
resulted in a revised bank of items recommended by measures to inform item selection and ensure data
a minimum of two experts, which was then organized quality. It is especially important in the field of ECD
again by domain and within each domain by age (i.e., measurement given the complexity and culturally
those items relevant only for 2-year-olds, 3-year-olds, driven nature of early childhood development.
4-year-olds, 2–3-year-olds, 3–4-year-olds or 2–4-year-
olds). The items on the MICS ECDI had not previously
undergone cognitive testing. Therefore, cognitive
Observations from the experts indicated that the testing was conducted on the original set of 10 MICS
initial set of items was skewed towards the inclusion ECDI items in India and Jamaica in March and April
of items for younger children, i.e., those who are 2 2016. This involved a total of 45 parents of children
years old. Therefore, a further process of selection aged 2 to 4 years in Jamaica (n=20) and 2 to 5 years
was undertaken to identify additional items to in India (n=25).
include, ensuring that those selected received a
recommendation from at least two experts. All of The main findings from this round of testing included:
the MICS ECDI items were also included with the • Respondents found many questions in the MICS
exception of two: the item about kicking, biting and ECDI to be confusing.
hitting was excluded since a similar item was already
included from another measure; and the item about • Questions tended to be complex; respondents
being too sick to play was also removed given it has had difficulty mapping their experiences onto the
not been shown to be a good indication of physical simple, binary “yes/no” answer categories and,
development due to its vague nature. as a result, used varying patterns of interpretation
both across and within countries.
The resulting bank of 61 items (34 on learning, 20 on
• Many of these patterns appeared to be ‘out-of-
psychosocial well-being and 7 on health) was retained
scope’, indicating that some items were prone to
as the final set to undergo cognitive testing.
measurement errors.
As a result of these findings, two items were dropped,

Testing
and it was noted that the remaining question set
Cognitive testing would benefit from further revision and testing.
The purpose of cognitive testing is to evaluate
survey questions in order to assess respondent In 2017, UNICEF conducted cognitive testing on the
comprehension and interpretation, enhance cross- larger set of 61 draft ECDI2030 items compiled from
cultural appropriateness of items, reduce response a variety of existing measures and tools during the
bias and identify potential sources of measurement desk review and expert rating exercise (see previous
error.45 section for more details). The testing was conducted
in four field sites (in Bulgaria, Mexico, Uganda and
As a qualitative methodology, cognitive testing the United States) in collaboration with national
typically involves in-depth, semi-structured interviews implementing partners (Institute for Public Policies
with purposive (and hence non-representative) Studies in Bulgaria, National Institute of Public
samples. The primary benefit of cognitive interviewing Health of Mexico, Uganda Bureau of Statistics and
over non-qualitative evaluation methods is that it RTI International in the United States). Cognitive
provides rich, contextual data on how respondents testing across countries followed a standard protocol
25
for the overall research process, including interview aloud’ method and scripted or spontaneous verbal
methodology and data analysis. The main findings probes to collect data about the response process.
from the cognitive testing are included here, but Since cognitive testing took place in four rounds,
more detail can be found elsewhere.46 the interview script was adjusted to reflect the
changes to questions and to elicit information about
The research process was iterative, with the findings specific issues raised during the preceding rounds.
from the United States informing changes made to Mothers were asked to listen to each question as the
items and interview methodology in Bulgaria and interviewer read it aloud and then provide a response.
Mexico and findings from these two countries then Following this process, participants talked through
informing the testing in Uganda. what they thought about as they answered each
question. Follow-up probes were both general (e.g.,
Participant recruitment “In your own words, what is this question asking?”)
Within each country, the purposive sample of and question-specific, such as probes about terms or
respondents was recruited based on diverse concepts that might have been confusing (e.g. “What
characteristics, including child age (at least one does ‘identify’ mean as it is used in this question?”).
2–4-year-old child), child sex, household location, These scripted probes allowed interviewers to explore
level of caregiver education and child disability respondents’ understanding of the question and
status. In Mexico, mothers of indigenous ethnicity whether they comprehended constructs anticipated
were also specifically recruited. Local staff from to be problematic. Spontaneous ad-hoc probes
the implementing partner in each country used a explored inconsistencies in mothers’ responses.
variety of recruitment procedures to identify potential Response types included dichotomous (e.g., yes/no)
participants. In Bulgaria and Mexico, staff worked and frequency scale (e.g., never, rarely, sometimes,
through word-of-mouth invitations and snowball often, always) options depending on the item.
sampling to recruit mothers who were neighbours
or acquaintances. In Uganda, staff from the local Analysis
implementing agency worked with chairmen (local Information gathered from audio recordings of
leaders) from sampled areas prior to the data participants during interview probing was compared
collection to identify participants with the desired to item descriptions to determine matches and
attributes. mismatches between the intention of the item and
participants’ answers and interpretation of the item.
Interviewer training Information was also examined in terms of underlying
Training to prepare interviewers to conduct cognitive patterns related to the child’s age and other mother or
testing included two-day, in-country, interactive child demographics.
sessions that incorporated cognitive interviewing
techniques, practice and review of the interview guide Results
as well as review of project procedures and logistics. Key themes emerged from the four rounds of
Interviewers in all countries received standard training cognitive testing in Bulgaria, Mexico, Uganda and the
from a core team of trainers. Selected interviews United States that could potentially lead to systematic
were observed by trainers during fieldwork, either in- measurement error or response bias. Areas of concern
person or using video conferencing, with suggestions arose within every phase (i.e., comprehension,
and corrections provided to the interviewer after retrieval, judgement and response) of the question-
completion of the interview. response process, which is the underlying theory that
guides cognitive interviewing.47
Sample
Interviews were conducted in 2017 with 146 Issues with comprehension: Participants experienced
respondents across the four countries: Bulgaria (30 issues in understanding some questions. Four
interviews), Mexico (47 interviews), Uganda (39 main issues related to item interpretation and
interviews) and United States (30 interviews). comprehension are important to note:
Interview script • There was confusion over concepts and terms

Interviewers used a combination of the ‘think- used in some questions. For example, respondents
interpreted the item “Can (name) identify all • Respondents showed some confusion in
written numbers from 1 to 5?” in three ways, as understanding the difference between a child
illustrated in Figure 4. performing tasks versus performing them correctly
(e.g., counting 10 objects with mistakes vs. doing
• Respondents sometimes conflated their child’s so correctly).
willingness to perform a task with their actual
ability to perform it (e.g., ability to go to the • There was a lack of clarity when words have more
bathroom alone vs. willingness to do so). than one meaning (e.g., the word ‘distraction’ in
the question “Does (name) get distracted easily?”
has two meanings in Spanish).
Figure 4
Participants’ interpretation during cognitive testing of the item on written numbers
See (or point out) the symbol

and say the number
See and say numbers
See and distinguish

between numbers
Say (or recognize) and

write the numbers
Identify written
Say and write numbers
numbers from 1-5
Count and write the numbers
Count Know which number comes first
27
Issues with retrieval: In some cases, participants in “Does (name) get distracted easily?” had another
had never observed the behaviour being asked about possible meaning in Spanish and, in Mexico, some
(e.g., picking up a small object with two fingers) mothers answered this item as if they were asked
or were not familiar with objects referenced in the about whether the child is attracted to or likes a
question as examples, such as animals that are less variety of things. Likewise, some examples were
common in certain countries. not appropriate across countries or contexts. For
instance, the example provided in the item “Does
Issues with judgement: Some respondents, (name) frequently act impulsively or without thinking
especially with younger children, based their answers (e.g., running into the street without looking)?”
on perceptions or suppositions rather than actual may have different implications in rural and in urban
observation of the behaviour, such as when a child settings. In Uganda, a few mothers indicated that
has not yet started to speak. Other concerns included their child is always with them and others mentioned
participants varying in what they believe ‘counts’ as that they would not put the child in a position to be
an affirmative response to the question, such as able to run into the street. In a similar way, when
whether children needed to sing an entire song or asked “Does (name) know that a goat weighs more
draw an exactly straight line in order to affirm their than a mouse?”, most mothers in an urban setting in
child’s ability. Mexico mentioned that their children had never seen
a goat and did not know what it was.
Issues with response: Some questions were
problematic because the questions provided only Informed by the results of the cognitive testing,
yes or no response options but participants wanted recommendations for edits to item wording, response
to give more variability in their answer (e.g., by options, interviewer training instructions and other
responding ‘sometimes’), such as whether a child translation and implementation processes were
becomes extremely withdrawn or shy in new discussed among the TAG and IAEG-ECD members at
situations. Respondents mentioned that scaled a technical consultation in January 2018. This resulted
response options as opposed to a yes/no response in a modified set of 58 items to undergo field testing
option would more accurately allow representation of (see Annex C).
their children’s behaviour overall.
Field testing
Results from maternal and child subgroup analyses The primary goal of the field testing was to collect
suggested potential interpretative differences quantitative data to be able to assess the psychometric
based on maternal educational attainment and properties of the items in different cultural contexts. It
language/cultural background. For example, difficulty was also an opportunity to test administration aspects
understanding some questions or certain words was of the ECDI2030 and its implementation tools,
more frequent among mothers with lower education including training and instructions for interviewers,
levels. Subgroup analyses also revealed that the data analyses and tabulation plans.
comprehension of some items was impacted by
translation issues. In Mexico, for example, some Through the latter part of 2018 and early part of 2019,
questions were excessively wordy and therefore dedicated field testing was conducted on the draft set
difficult to understand, such as, “Can (name) easily of ECDI2030 items in Mexico (see Figure 5) and State
switch back and forth between activities such as of Palestine. There was also a field test in Belize as
going back to a game or playing with a toy after being part of the larger MICS6 pilot; as such, the sampling
interrupted?”, which was not easily understood by approach, instruments, eligible respondents and
almost all mothers and eventually dropped. Also, in fieldwork procedures are detailed further elsewhere.48
Uganda, there was evidence that respondents based
some of their answers on cultural expectations. The samples in each country were drawn using a
stratified, probabilistic three-stage cluster design
The different meanings that words can have in in order to ensure representativeness, and they
another language or their lack of appropriateness in included regional areas (West Bank and Gaza Strip in
the local context also interfered with the objective of the State of Palestine), metropolitan, urban and rural
some questions. For example, the word ‘distracted’ areas (both indigenous and non-indigenous), as well
Table 1
Field testing in Mexico, State of Palestine and Belize
Sample Households Response Instruments

Location size surveyed rate Stratification utilized Translation
Mexico • Puebla Women: 1,513 92.8% • Metropolitan • MICS household, Spanish
1,455 • Urban women’s and
• Rural- children under 5
Children:
indigenous questionnaires
1,641
• Rural-non- • ECDI2030 draft set
indigenous of questions
• Anthropometry
• ECD direct
assessment
State of • West Women: 897 99.5% • Region • MICS household, Arabic

Palestine Bank 888 • Urban women’s and
• Gaza • Rural children under 5
Children:
Strip • Refugee questionnaires
1,102
camps • ECDI2030 draft set
of questions
Belize • Toledo Women: 680 89.9% (for None • MICS household, Spanish
• Belize 583 children women’s and
City under age 5) children under 5
Children:
South questionnaires
214 under
Side • ECDI2030 draft set
age 5, of
of questions (only
which 120
subset of items
aged 24-59
from the learning
months
domain)
as refugee camps, depending on the country. The instruments were available via paper questionnaire or
final sample size across countries is shown in Table the CAPI system.
1. The surveys were implemented using computer-
assisted personal interviews (CAPI) loaded onto An adult household member (over the age of 18) was
handheld tablets. asked to respond to the household questionnaire,
which included modules to collect information on
In general, non-response due to absence of eligible household characteristics, water and sanitation,
household members was between 7.1 and 10.2 handwashing and education of household members.
per cent, depending on the country. Refusals to
participate in the survey accounted for about 1 per The women’s questionnaire was administered
cent for all countries. The response rate for the child only to mothers of children aged 2 to 4 years and
questionnaire was close to 100 per cent, and almost collected data on mothers’ background (education,
100 per cent of these questionnaires were answered etc.). In Mexico, modules on marriage/union and on
by the child’s mother. depressive symptomology were also included in this
questionnaire.
Field test instruments
Given the ECDI2030 objectives, standard MICS The questionnaire for children under 5 years of age
questionnaires were adapted so as to include only was administered to mothers of children aged 2 to
the relevant sections. The field test instruments were 4 years and included modules on child background
thus comprised of three questionnaires: a household characteristics (age, sex, etc.), birth registration,
questionnaire, a women’s questionnaire and a child functioning, child discipline and early childhood
questionnaire for children under 5 years of age. These development.
29
The ECD module included the ECDI2030 draft set of Table 2. The direct assessment items were adapted
58 items as well as questions about the availability of from validated ECD measurement tools including the
books and playthings in the home, early stimulation Bayley Scales of Infant and Toddler Development,
and responsive care from adult household members, Third Edition (BSID-III),49 Regional Project on Child
inadequate supervision and attendance in early Development Indicators (PRIDI),50 IDELA51 and the
childhood education. In Belize, only the learning draft Developmentally On Track (DOT) assessment
domain items were tested instead of the full set of tool52 developed by The University of Hong Kong
items. (HKU).
In Mexico, data were also collected on height and Fieldwork personnel training
weight for each child aged 2 to 4 years as well as on a Prior to fieldwork activities in each country, trainings
direct assessment of ECD, as explained next. took place that consisted of theoretical and practical
sessions. The contents included a general introduction
Direct assessment of early childhood development to the survey and application logistics (including
The items on the ECDI2030 do not rely on direct management of the CAPI), visiting households and
assessment. However, to compare mothers’ approaching eligible respondents (including how to
reports on their child’s abilities to the child’s actual address lack of response and related problems) and
performance related to those abilities, the field interview techniques, followed by simulated practice
test in Mexico also included direct assessment of of the questionnaires on paper and on tablet. Time
children. This module comprised 13 tasks related to was also devoted to discussion and clarification of
learning and gross motor development, as detailed in any concerns.
Table 2
Direct assessment items used in Mexico field test
Item Domain Adapted from
Child gives three blocks Learning IDELA
Child counts ten blocks Learning BSID-III
Child counts five blocks Learning BSID-III
Child identifies big and small objects Learning BSID-III
Child stacks three or more blocks Learning BSID-III
Child writes his/her name Learning DOT
Child describes position of a block (e.g., “in”, “on”, “above”, “under” another object) Learning DOT
Child uses verbs in past tense Learning PRIDI
Child names ten objects correctly Learning BSID-III
Child recognizes ten objects correctly Learning BSID-III
Child jumps up with both feet leaving the ground Gross motor BSID-III
development
Child throws a ball Gross motor BSID-III

development
Child catches a ball Gross motor PRIDI

development
In Mexico, training of nurses for administration of the CAPI questionnaires, additional information on the
ECD direct assessment was carried out by estimating interviewer’s and supervisor’s manuals, refinement
measurement variance in relation to results yielded of the ECD direct assessment application protocol
by gold-standard measurers. Reliability and accuracy and scoring guidelines and identification of final
of ECD direct assessment was measured by Cohen’s training topics that required additional content and/
Kappa and Bland-Altman graphs. Standardization was or clarification.
achieved by a Cohen´s Kappa value of at least 0.80
and agreement within 95 per cent limits in Bland- Final fieldwork activities
Altman53 graphs. All nurses with out-of-range values To ensure the overall quality of the information
for the direct assessment participated in an additional collected, there were multiple levels of operational
practical training and standardization session. oversight and supervision, including continuous
monitoring of data-quality indicators throughout the
Pre-test fieldwork.
Prior to the final fieldwork, a pre-test was conducted
in both Mexico and State of Palestine to test survey Interviewers visited sampled households, asking
instruments, fieldwork logistics and procedures, for respondents’ consent to participate in the
including the standardization of fieldworker interview survey. Within the household, mothers of children
techniques. The CAPI process was also tested for aged 2, 3 and 4 years were invited to participate in
system consistency and data transfer protocols. the interviews. Questionnaire application followed
In Mexico, standardization of the assessment standard MICS protocol, with fieldworkers returning
protocols for both the ECD direct measurement and to the household at least three times, if needed,
anthropometric measurements was also pre-tested. to complete all questionnaires. Failure to obtain
household consent to participate was recorded as
The pre-tests led to several modifications and household non-response, and failure to complete
improvements to the survey instruments and questionnaires after three visits to the household was
procedures, including adjustments to paper and recorded as individual non-response.
Figure 5
Overview of field test activities in Mexico in 2018
CARTOGRAPHIC
PRE-TEST LOGISTIC PRE-TEST
LISTING
(1-3 August) (4 September)
(July & August)
FINAL FIELDWORK
CARTOGRAPHERS’ (6-30 September)
TRAINING PRE-TEST TRAINING FINAL TRAINING
(18-20 July) (26-31 July) (27 August-3 September)
Institutional Supervision Supervision

Review Board Sample Final sample
approval design selection
Contact local authorities
Data processing
and analysis
Development of questionnaires, ECD direct assessment, CAPI programming, fieldwork materials
31
Field test data processing and initial analyses possessed data on one or more candidate items from
For each field test, a data-processing team performed the Measuring Early Learning Quality and Outcomes
standard data consistency checks and produced (MELQO), but country names were unknown as they
data quality tables. Sampling coordinators generated had been annonymized in the dataset.
household-, women- and child-specific sample
weights considering the sample design, selection In total, data were harmonized for the 58 candidate
probabilities and non-response. items across 33 countries, from a collective sample
of around 60,000 children aged 24 to 59 months
For each country, standard MICS indicators were (Table 3).
calculated using the data collected by the different
questionnaires. Using weighted data, a wealth index
was calculated using principal component analysis of
Psychometric analyses
selected household characteristics (e.g., crowding, Psychometric analyses were used to assess the
possession of durable goods in the house and having performance of individual items in all countries
a bank account). The resulting index was used to with available data. Even though each step in the
classify households in wealth quintiles. psychometric work involved different aims and
methods, two overarching principles guided all the
Initial data analyses focused on the Mexico and State steps and item selection. First, content coverage
of Palestine field tests, including estimating national-, was considered to ensure the selection of a
regional- and state-level percentages and 95 per sufficient number of items across all the conceptually
cent confidence intervals for all standard indicators, defined subdomains. Each subdomain needed to
as well as generating results by key disaggregation be represented by at least one item on the final
variables (e.g., child’s sex and age, maternal education ECDI2030, and the total number of subdomains
level, area of residence and wealth quintile). Basic should be maintained. Age coverage was also
descriptive statistics were also produced for the draft established as a core criterion for the learning and
ECDI2030 items to inspect overall endorsement of health domains such that an item would not be
each question by age group and sex, as well as the dropped if there were no other items with similar
proportion of ‘don’t know’ responses for each item probability of endorsement for the same age range.
(Annex D). The specific aims and results of each analytic step
are described in the following sections. The results
Data processing and statistical analyses were of the psychometric analyses are detailed in another
performed in SPSS v.20, Stata v.15 and R version paper.56
3.6.0.
Initial item screening
Aims: Pre-screening candidate items informed the
Harmonizing a global dataset analytic process to pare down and refine the draft set
In addition to the data from the three ECDI2030 field of items and remove any that had poor measurement
tests, data from a further 30 countries were used to properties.
pre-screen candidate items and inform the analytic
process to pare down and refine the draft set of items. Datasets used: Item selection of the 58 candidate
Countries shared access to raw data or statistical items was based on information from all countries in
summaries on candidate items (or their close the global dataset with available data (Table 3).
analogues) to form a global dataset. Four countries
had convenience sample data on 49 candidate items Methods: A decision-matrix was used to evaluate
from HKU,54 and 17 had convenience sample data the difficulty and discrimination properties of the
on 24 candidate items from the Caregiver Reported candidate items.
Early Development Instrument (CREDI).55 Data on 16
items from population-level data collected in Canada Item difficulty was assessed to ensure a selection of
in 2008-2009 – the National Longitudinal Study of items with adequate variability in terms of difficulty
Children and Youth (NLSCY) and Survey of Young across the ages of 24 to 59 months. To assess item
Canadians (SYC) – were also utilized. Ten countries difficulty, the percentage of children ‘passing’ each
item was estimated across each year of age. Two were used as an initial approach to inspect item
criteria were used in the decision-matrix related to discrimination on the basis of two criteria: (1) Item-
difficulty: (1) At least a 10-percentage-point difference rest correlations of the item with the domain score of
between children passing the item from ages 2 to at least 0.3 and (2) Item-rest correlations of the item
3 years old and (2) At least a 10-percentage-point with the full set of items score of at least 0.25.
difference between children passing the item from
ages 3 to 4 years old. These criteria were only applied Each item was assigned a ‘flag’ on the basis of
to items in the learning and health domains given the extent to which the above criteria were met.
that items within the psychosocial well-being domain Green-flagged items met all four criteria and red-
were not expected to show a clear age progression.57 flagged items did not meet any of the criteria. A
yellow flag was also used for items meeting only
Item discrimination reflects the item’s ability to some criteria or meeting all criteria but only in a
discriminate between children with better and worse few countries. Results are summarized in Annex E.
overall development. Point-biserial correlations
Table 3
Description of datasets used for initial item screening
Number of children Population-level

Country and agency/survey Age range in the sample probabilistic sampling
A. Mexico, National Institute of Public Health 2-4 years 1,641 Yes
B. State of Palestine, Palestinian Central Bureau of 2-4 years 1,099 Yes

Statistics
C. CREDI (17 countries1 aggregated) 2-3 years 16,317 No
D. The University of Hong Kong (Bangladesh, China, 3-4 years 633 No

India and Myanmar)
E. Belize, UNICEF and the Statistical Institute of Belize 2-4 years 120 Yes
F. Canada, NLSCY 2-4 years 6,865 Yes
G. Canada, SYC2010 2-4 years 4,782 Yes
H. Country 1, MELQO 2-4 years 5,669 No
I. Country 2, MELQO2 2-4 years 112 No
J. Country 3, MELQO2 2-4 years 355 No
K. Country 4, MELQO2 2-4 years 205 No
L. Country 5, MELQO 2-4 years 1,388 No
M. Country 6, MELQO 2-4 years 5,269 No
N. Country 7, MELQO 2-4 years 10,152 No
O. Country 8, MELQO2 2-4 years 3,680 No
P. Country 9, MELQO2 2-4 years 344 No
Q. Country 10, MELQO 2-4 years 6,779 No
1. Information from Bangladesh, Brazil, Cambodia, Chile, Colombia, Ghana, Guatemala, India, Jordan, Lao People’s Democratic Republic,
Lebanon, Nepal, Pakistan, Philippines, United Republic of Tanzania, United States and Zambia.
2. Due to the low proportion of children aged 2 years old, the decision-matrix only considered information from children aged 3 and 4
years old.
33
In addition to the classical test theory-based screening other items in the domain and/or the overall ECDI2030.
using the decision-matrix, a subset of 9 items on the Items that addressed problematic internalizing or
learning domain and 3 items on the health domain externalizing behaviours did not exhibit a strong
were compared to analogous direct-assessment correlation with other items in the pool or with
items administered in Mexico. If the caregiver- child age, but they were retained to ensure content
reported responses and the direct assessment coverage of the psychosocial subdomains. Finally, it
responses showed a large discrepancy (e.g., > 10% is important to note that the item “Recognizes three
of responses in the off-diagonal cells), the item was letters of the alphabet” was dropped despite being
flagged for follow-up and discussed with the TAG. flagged as green (i.e., meeting all four criteria) because
there was another item (“Recognizes five letters of
Main results: Items were dropped from the learning the alphabet”) that captures more advanced literacy
and health domains due to floor or ceiling effects in skills and was needed for age coverage criteria.
multiple countries. For example, “Pretends object is
something else” was removed because it was highly The results from application of the decision-matrix
endorsed in most countries regardless of child age are summarized in Figure 6 and Table 4. A total of 22
and, since this was the only item in the approaches items were dropped among candidate items due to
to learning subdomain, this subdomain was also poor measurement properties, reducing the 58-item
effectively removed. Items were omitted in the pool to 36 items that performed well across settings.
psychosocial domain due to low correlations with
Figure 6
Decision-matrix criteria
DIFFICULTY DISCRIMINATION
Percentage of children passing item Point-biserial correlations
Criteria 1 Criteria 2 Criteria 3 Criteria 4

At least 10-percentage- At least 10-percentage- Item-rest correlation Item-rest correlation
point difference from point difference from within domain of within full ECDI score
2 to 3 years old 3 to 4 years old at least 0.3 of at least 0.25
Red
Green Yellow
Does not meet
Meets all criteria Meets some criteria
any criteria
12 items 22 items
24 items
Content coverage and age coverage criteria
22 items dropped
Draft set 2 (36 items)
Only a single item was dropped based on the per cent of the normal distribution, because this
comparison with direct assessments (“Knows the range would be most relevant for determining on-
difference between big and small”). Caregivers track status, as defined by SDG 4.2.1. The TIF was
tended to underreport their children’s ability to computed using a unidimensional IRT model. In
distinguish sizes compared to the direct assessment. particular, the two-parameter logistic model (2PL)
was used for items with binary response items, and
As expected, the psychosocial items related to the graded response model (GRM) was used for
problematic internalizing and externalizing behaviours items with multiple categories.
were endorsed with similar frequency regardless of
age. It was also the case that these items had low ATA was used to maximize test reliability subject to
proportions of endorsement in all countries. For the following constraints:
example, the response ‘Always’ was endorsed for
the item “Seems to be sad or depressed” by about 2 1) The final test form would contain 20 items based
per cent of caregivers in Mexico and by 3 per cent of on logistical considerations.
caregivers in State of Palestine, and this was relatively
constant over age. This led to a dilemma in deciding 2) Items from each of the content-based subdomains
whether such items should be omitted because they must appear on the final test form. This condition
were not strongly related to other items, or be retained operationalized the content coverage requirements
because of their importance in defining the content of SDG indicator 4.2.1.58
coverage of the psychosocial domain. The consensus
of the IAEG-ECD and TAG was to retain one item Applying these constraints, ATA was implemented in
from each subdomain to ensure representation of all the R programming language using the mirt package
psychosocial subdomains, with the retained items to estimate IRT models and the eatATA package
having the better item-rest correlations overall. to run the test assembly. Resolving discrepancies
between the ATA test forms in the two samples was
A final review of the cognitive testing results also informed by input and consultation with the IAEG-
revealed that one item (“Says what others like or ECD and TAG, which focused on (a) the distribution
dislike”) performed poorly as it did not appear to work of subdomain coverage by difficulty and (b) the
as intended and the underlying concept did not seem face validity of the overall measure. This procedure
to translate well into a survey question. resulted in the current version of the ECDI2030.
IRT test assembly Main results: After screening, a total of 34 items

Aims: The objective of the automated test assembly remained in the item pool. These 34 items were then
(ATA) was to maximize the reliability of the ECDI2030 used for separate ATAs with the data from Mexico
over the range (-1.65, 1.65) SD units on the item and State of Palestine. As shown in Annex F, the ATAs
response theory (IRT) scale score. resulted in 17 out of 20 items in common across the
two countries. Of the 17 items, 16 were included
Datasets used: Mexico and State of Palestine. on the final ECDI2030. The remaining 4 items were
selected based on input from the TAG. The final
Methods: IRT-based test assembly was conducted ECDI2030 form is shown in the last column in the
separately in the two primary samples (Mexico and table in Annex F.
State of Palestine). All analyses utilized sampling
weights reflecting selection probabilities and survey Statistical properties of the ECDI2030
non-response. The IRT scale score of the ECDI2030 The methodological work resulted in a 20-item
was standardized to have a mean of zero and a measure with marginal IRT-based reliability
standard deviation of one in both samples. coefficients of .850 for the original ATA in Mexico
and .839 for the final ECDI2030 and .880 for the
The test information function (TIF) was used to original ATA in State of Palestine and .876 for the
operationalize reliability, and the objective function final ECDI2030. These results confirm that the ATA
was computed as TIF / (1 + TIF). The range of the procedure yielded a single test form with reasonably
IRT scale score was chosen to cover the middle 90 good approximation to the original ATA forms in both
35
Table 4
Stages of item pool screening and items dropped at each stage
Number of items Number of

Analysis Datasets used analysed items dropped
Classical test theory-based Mexico, State of Palestine, CREDI, HKU, 58 22
screening (decision-matrix) Belize, Canada and MELQO
Comparison with direct Mexico 9 learning domain items, 3 1

assessment health domain items
Psychosocial domain analyses Canada, Mexico, State of Palestine, 18 psychosocial domain 1

and review of cognitive testing CREDI, HKU and MELQO; CT results items
results from Bulgaria, Mexico, Uganda and
United States
countries. In addition, the final 20-item measure was Standard-setting

approximately 90 per cent as efficient as the 34-item
Items on the ECDI2030 capture specific
version (except at the higher end of the scale in State
developmental constructs nested within the three
of Palestine), suggesting that the ATA procedure
domains of health, learning and psychosocial well-
worked well at identifying a common minimum set of
being. Because the intention was to generate a single
items. A summary of the statistical properties of the
summary score reflecting the interlinkages among
final ECDI2030 is presented in Annex G.
these domains, and the holistic nature of ECD more
broadly, it was necessary to define some criteria in
The IRT scale score was highly correlated with the
order to transform the summative score obtained
unweighted total score in Mexico (0.96) and State
from the 20 items into a performance standard for
of Palestine (0.95), suggesting that a unidimensional
classifying children as ‘developmentally on track’.
model to report a single overall score based on
These specifically focused on defining expectations or
caregiver responses was appropriate for most
‘standards’ of what minimally on track children should
of the items on the ECDI2030. Some items in
be able to do by ages 24, 36 and 48 months, thereby
the psychosocial domain (i.e., those measuring
reflecting accumulated development up to that age.
internalizing and externalizing behaviours) were
The main methods and results of the standard-setting
problematic from a measurement perspective, given
are outlined below and documented in greater detail
their low correlation with age. Nevertheless, these
elsewhere.59
items were retained in the final measure due to
content coverage requirements.
To establish performance standards on the
ECDI2030, a criterion-referenced standard-setting
Correlations with concurrent validation measures
exercise using a modified Angoff approach was
revealed strong to moderate relationships in the
carried out. The Angoff method (and its variations)
anticipated direction in most cases (Annex G). For
is a common and widely employed test-centred
example, the correlation of the IRT scale score with
methodology for establishing criterion-referenced cut
child age was strong, positive and approximately
scores.60 It was selected over other methodologies
linear across the 24-59 month age range. The main
because having field test data from only two countries
exceptions were the relationship of the IRT scale
was not considered sufficient to generate norm-
score with exposure to early stimulation in both
referenced standards. That said, all the psychometric
countries (weak but positive correlations with the
information, including item parameters, was used
ECDI2030), and the relationship with wealth index
to inform the standard-setting exercise, and experts
quintiles in both countries (significant but small
calibrated their expectations considering the empirical
differences). However, comparing children from the
information (i.e., impact data).
wealthiest households with those from the poorest
showed anticipated differences in ECDI2030 scores.
In the Angoff method, subject matter experts (SMEs) During Round 1, SMEs individually reviewed the items
are tasked with reviewing each item, determining and submitted their ratings via an online survey form.
the knowledge or skills required to answer the item Following this, there was a facilitated discussion of
correctly and then making a judgement as to how a selected items where greater disparity in ratings was
subject, at the minimum threshold for a performance observed as well as of items where there was a high
level, will likely respond. degree of agreement. Additionally, the estimated
impact (i.e., the percentage of children who would
The first step of the standard-setting process was to be identified as ‘on track’ by applying the group’s
conduct a pilot in Mexico in which six SMEs met for recommended cut scores) using field test data from
1.5 days and practised this methodology using items Mexico and State of Palestine was also presented
previously discarded from the draft item set for the to the group. During Round 2, experts completed a
ECDI2030. The main recommendation generated by second round of ratings, informed by the feedback
the pilot was to utilize a modified Angoff approach and discussion from the first round.
in which SMEs indicated the likelihood that the
subject at each performance level would answer the Values were summed across all items for each expert
item correctly.61 For the ECDI2030 specifically, this to determine their individually recommended cut
translated into the following key questions: score for each age group. The group’s recommended
cut score (mean, median, range) and variability was
also calculated for each round and each age (Table 5).
• For items requiring a yes/no response: If you
For example, in Round 1, the panel’s recommended
asked 100 mothers/caregivers of children aged
mean cut score for a 24-month-old who is on track
24 months (for example) who were minimally on
was 7.0 items out of 36.
track in their development, how many of them
would endorse the item?
The standard-setting exercise was carried out
based on the 36 items that were retained after the
• For items that had more than two response
initial item screening. However, the psychometric
options: If you asked 100 mothers/caregivers of
analyses described above had reduced these to 20
children aged 24 months (for example) who were
items. Moreover, during subsequent analyses of the
minimally on track in their development, how many
psychosocial domain, items were dropped and two
out of 100 would mark each response option?
items (“Gets along with other children” and “Seems
to be sad or depressed”) from the candidate item
Following the pilot, a panel of 15 global SMEs convened pool were reinstated to form the final measure.
(virtually and in-person) in August 2019. The members
(Annex H) were identified based on nominations As a result, only 18 of the 20 items on the final
from the IAEG-ECD. They were all experts in ECD/ ECDI2030 were included in the standard-setting
developmental psychology and possessed significant exercise. A calibration method was therefore applied
experience in conducting standardized assessments to determine the difference in difficulty between
of children under age 5 in primary or tertiary health the set of 18 items and the final 20-item ECDI2030,
care and/or educational settings, worked in clinical, and then this relationship was used to obtain the
educational or research settings and collectively standard-setting recommendations for the final
represented 13 different countries spanning five ECDI2030. Table 6 shows the standard-setting results
continents. after calibrating the complete set of 20 items.
To begin, training was provided in order for SMEs to Identification of the final set of performance standards
develop a shared understanding of what it means The last step in identifying the performance standards
for a child to be considered developmentally on track for the ECDI2030 involved establishing the final cut
at each age (informed by their own experience and scores to identify children developmentally on track.
knowledge) and to understand how to translate their To inform this decision, field test data from Mexico
expectations into a standard-setting judgement. The and State of Palestine were once again used to
panel also participated in some practice sessions generate performance profiles according to different
prior to undertaking two rounds of standard-setting expectations of children’s performance based on the
for the 36 items from the draft set of the ECDI2030. average calibrated cut scores.
37
Table 5
Standard-setting results by round
Round 1 Mean Median Standard error Range
Round 1
24 months 7.0 4.7 1.4 4.3 - 9.7
36 months 17.1 16.4 1.3 14.5 - 19.6
48 months 26.7 26.6 1.0 24.7 - 28.8
Round 2
24 months 7.7 6.8 1.1 5.5 - 10
36 months 18.4 18.1 0.9 16.5 - 20.2
48 months 28.0 28.5 0.8 26.4 - 29.5
Table 6
Calibrated standard-setting results (20 items)
Standard
Round 1 Min Max Mean Median error Range
24 months 3.6 11.5 5.6 5.1 0.5 4.5 - 6.6
36 months 9.1 14.1 10.9 10.5 0.4 10 - 11.7
48 months 13.4 18.0 15.9 16.4 0.3 15.2 - 16.6
The proportion of developmentally on track children the different stages of designing the measure.
rendered by each performance profile was reviewed Moreover, further exploration of the field test
in each country, taking into account the underlying data suggested that there were very few children
expectation about children’s performance in general who were lacking proficiency in one domain but
and a number of key disaggregation variables. The were still classified as on track overall. Therefore,
identification of the final cut scores also took two it was determined that the application of a single
additional issues into consideration: cut score for each age range best supported the
holistic nature of ECD.
• A requirement on the minimum number of
items within each domain: The idea of requiring • The need to include intermediate cut scores for
a minimum number of items within each domain children between the ages of 24 and 35 months and
was to ensure that children could not be classified 36 and 47 months: Given how quickly development
as ‘on track’ if they were lacking proficiency in one occurs during these early ages, it was recognized
or more domains. It was decided, however, that that many children in these age groups might be
this was not necessary because content coverage identified as ‘on track’ if they were in the latter
is already a core attribute of the ECDI2030 and part of the year range (e.g., a 32-month-old judged
was addressed in several ways throughout against the expectations for a 24-month-old).
Therefore, the panel recommended the inclusion of Palestine. The dots in the chart represent the
of different cut scores by 6-month age groups for probabilities of obtaining different ECDI2030 scores
the younger children. for children between 24 and 59 months. Each dot
represents a month of age, and different colours are
These additional considerations informed calibration used to depict the probability for certain age groups
of the final set of standards from within the ranges (e.g., light blue dots for children aged 24 months
recommended by the global panel. For the 24-month- and yellow dots for children aged 59 months). As
old standard, the upper end of the recommended expected, the chart shows that the ECDI2030 score
range (6.6 rounded to 7) was selected on the basis is a function of age, with older children being more
of the estimated impact from the field test data. likely to score higher. Likewise, as ECDI2030 scores
Similarly, for the 48-month-old standard, the lower increase, the probabilities of obtaining such scores
end of the recommended range (15.2 rounded to decrease, also as a function of age. In addition to this,
15) was chosen based on the estimated impact. the chart also identifies how the final age-specific cut
Finally, the intermediate performance standards were scores are reflected in these probabilities. It indicates
identified as the median score within the range (i.e., that 84 per cent of children aged 24 to 59 months in
the difference between the performance standard
the country are developmentally on track in health,
at the start of an age level and the performance
learning and psychosocial well-being.
standard for the next age level). Thus defined, the final
age-specific cut scores adopted to identify children
The reasonableness of the final cut scores will be
developmentally on track were:
monitored and assessed in light of the availability of
• Children aged 24-29 months: 7 of the 20 items global data once more countries have implemented
• Children aged 30-35 months: 9 of the 20 items the ECDI2030. With the collection of data from a
sufficiently large enough number of countries, age-
• Children aged 36-41 months: 11 of the 20 items
specific norms could be established, thus producing
• Children aged 42-47 months: 13 of the 20 items a set of norm-referenced standards to compare to
• Children aged 48-59 months: 15 of the 20 items the criterion-referenced standards determined by the
standard-setting exercise.
Figure 7 illustrates the impact of the application of
these cut scores to the field test data from State
39
Figure 7
Predicted probabilities for ECDI2030 score by month of age and proportion of children
developmentally on track according to final cut scores, State of Palestine
0.9
0.8
0.7
0.6
Predicted probabilities
0.5 59 months
48 months
42 months
0.4
36 months
30 months
0.3 24 months
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
ECDI2030 score
Final cut scores recommended 24 30 36 42 48

by global panel months months months months months
Children aged 24 to 59 months developmentally on track

84%
© UNICEF/UN0465328/Pancic
41
PART FOUR
The final
ECDI2030
The final ECDI2030 is comprised of 20 items covering 3 domains and 12 subdomains
(Figure 8). The learning domain is made up of 5 subdomains as measured by 11 items, the
psychosocial well-being domain is made up of 4 subdomains as measured by 5 items and the
health domain has 3 subdomains measured by 4 items.
Figure 8
Content coverage of the ECDI2030
3 DOMAINS
12 SUBDOMAINS
Learning
EXPRESSIVE LANGUAGE
NUMERACY LITERACY PRE-WRITING
EXECUTIVE FUNCTION
Psychosocial well-being Health

EMOTIONAL DEVELOPMENT
SELF-CARE
SOCIAL DEVELOPMENT
EXTERNALIZING GROSS
FINE MOTOR
MOTOR
DEVELOPMENT
DEVELOPMENT
INTERNALIZING
20 items
43
Figure 9 shows the age coverage of the ECDI2030 50 per cent of children have passed the item and
and the variation of item difficulty across age groups. thus corresponds to the most difficult items in the
The main objective of the chart is to exemplify how the ECDI2030; teal shows the age band by when 50 per
final set of items allows for capturing the acquisition cent to 75 per cent of children have passed the item;
of milestones across domains and ages. and dark blue shows the age band by when 75 per
cent to 90 per cent of children have passed the item
For each item, the chart shows the percentage of and thus corresponds to the less difficult items in the
children for whom the item was endorsed or, in other ECDI2030.
words, the percentage of children who ‘pass’ the
item. The different colours quantify the proportion Even though a pronounced variation by age for the
of children passing the item by month of age: items in the psychosocial well-being domain is not
purple shows the age band by when 25 per cent to expected, the probabilities of endorsement are also
Figure 9
Probabilities of endorsement of the final ECDI2030 items by month of age, Mexico field
test data
25%-50% 50%-75% 75%-90% all > 90%
Says 10 or more words

Names an object consistently
Does an activity without
asking for help or giving up
Says sentences of 3 or more words
Uses correctly ‘I’, ‘you’, ‘she’, ‘he’
Says sentences of 5 or more words
Gives correct amount (3)
Knows numbers 1 to 5
Counts to 10
Recognizes 5 letters of the alphabet
Writes his/her name
Kicks, bites, hits

Seems to be sad or depressed
Gets along well with other children
Offers to help
Asks about familiar people
Jumps with both feet

Walks on an uneven surface
Fastens/unfastens buttons
Dresses him/herself
24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59
Figure 10
Calculation of the ECDI2030 indicator
The 20 items in the ECDI2030 are used to generate

one single indicator
15
Number of children aged 24 to 59 months

13 who have achieved the minimum number of
milestones expected for their age group
ECDI2030 =
11
Total number of children aged 24 to 59 months
SDG 4.2.1
Proportion of children aged 24 to 59 months
who are developmentally on track in health,
SDG indicator learning and psychosocial well-being
4.2.1
81%
Total indicator Total on track 75%

24 to 29 30 to 35 36 to 41 42 to 47 48 to 59 value and 71%
months months months months months disaggregated
values
MINIMUM NUMBER OF
MILESTONES EXPECTED
Poorest Richest
BY AGE GROUP children children
shown for these items. In this domain, a blue stripped ECDI2030 in order for that child to be classified as
bar is used for two items (“Kicks, bites, hits” and ‘developmentally on track’. The resulting data reflect
“Seems to be sad or depressed”), which measure the proportion of children aged 24 to 59 months who
manifest behaviours rather than developmental are developmentally on track in health, learning and
milestones. psychosocial well-being (i.e., SDG indicator 4.2.1).
And because the data can be disaggregated by key
The 20 items are used to calculate a single score by demographics (such as wealth quintiles, as illustrated
applying the final set of performance standards that in Figure 10) and subnational areas, the use of the
were adopted. By way of example, the mother of a ECDI2030 to measure indicator 4.2.1 can also help
child between the ages of 24 and 29 months would advance the SDG commitment to leave no one (child)
need to endorse at least 7 of the 20 items on the behind.
45
Annexes
Annex A. Membership of the Inter-agency Expert Group on ECD Measurement (IAEG-ECD) and
Technical Advisory Group (TAG)
IAEG-ECD members
Claudia Cappa, Senior Adviser, Statistics and Monitoring, Division of Data, Analytics, Planning and Monitoring,
UNICEF (Chair)
Marta Rubio Codina, Senior Economist and Child Development Specialist, Inter-American Development Bank
Rami Al Dibs, Director of Health Statistical Department, Palestinian Central Bureau of Statistics
Tarun Dua, Medical Officer, Department of Mental Health and Substance Abuse, World Health Organization
Alaka Holla, Senior Economist and Program Manager, Strategic Impact Evaluation Fund, World Bank Group
Dafna Kohen, Assistant Director, Health Analysis Division, Statistics Canada
Silvia Montoya, Director, UNESCO Institute for Statistics
Rowena Phair, Project Leader, Education and Social Progress, OECD
Lauren Pisani, (former) Advisor in Learning Research, Save the Children
Adelle Pushparatnam, Education Specialist, World Bank Group
Ruth Argelia Vázquez Salas, Researcher, National Institute of Public Health, Mexico
TAG members
Santiago Cueto, Principal Investigator, Grupo de Analisis para el Desarrollo (GRADE), Peru
Edward Frongillo, Professor and Director, Global Health Initiatives, University of South Carolina, United States
Melissa Gladstone, Senior Lecturer in Neurodevelopmental Paediatrics and International Child Health, University
of Liverpool, United Kingdom
Peter Halpin, Associate Professor, Quantitative Methods, University of North Carolina, United States
Magdalena Janus, Professor, Department of Psychiatry and Behavioural Neurosciences, Offord Centre,
McMaster University, Canada
Gillian Lancaster, Professor of Medical Statistics, School of Medicine, Keele University, United Kingdom
Dana Charles McCoy, Associate Professor, Graduate School of Education, Harvard University, United States
Abigail Raikes, Assistant Professor and Director, Global Early Childhood Development Research, University of
Nebraska Medical Center, United States
Nirmala Rao, Professor, Faculty of Education, The University of Hong Kong, Hong Kong, China
Ghassan Shakhshir, Chair of the Advisory Expert Group, Palestine National Child Institute, An Najah University,
State of Palestine
Hollie Hix-Small, Associate Professor of Special Education and Coordinator of the Early Intervention/Early
Childhood Special Education (EI/ECSE) Program, Portland State University, United States
47
Annex B. Subdomains and constructs in the learning and psychosocial well-being domains
Domain Subdomain Construct Definition and importance to ECD
Learning Early numeracy Counting Rote counting refers to children having knowledge of the names and order of
(verbal and set counting words. However, this does not indicate an understanding of quantity
production) (i.e., just because a child can count to five does not mean he/she can count five
objects). Being able to accurately count requires a number of skills that include
knowledge of the counting word sequence; an understanding of one-to-one
correspondence; knowledge that the last number word in the count is the
cardinal value of the set; knowledge that objects can be counted in any order;
and an understanding that any collection of objects can be counted.62
Along with other skills related to number sense, both verbal counting and set
production are predictive of later achievement in mathematics.63
Number Number identification, which represents the ability to identify numerals and
identification understand their corresponding numericities, is important to measure because
this skill is at the core of many other numerical competencies.64 At its most
basic level, it is the ability to form an accurate association between the written
word of a number and its symbol equivalent (e.g., knowing that ‘two’ and
‘2’ have the same meaning).65 Numerical identification has been found to be
predictive of later math abilities.66
Number/size Number/size discrimination refers to a child’s ability to compare things of

discrimination different numbers or size, e.g., “that spoon is bigger” or “she is taller than
he is”. This ability to discriminate is important because measurement-related
procedures have the potential to later serve as cognitive tools that children can
use to organize the way they reason in math problem-solving.
Language, Letter/alphabet This construct involves skills such as the ability to recognize and name
literacy and knowledge letters and knowledge of the sounds associated with them. Development of
communication these skills during early childhood is critical since this knowledge forms the
foundation of later literacy skills such as reading and spelling. In fact, among
other reading-readiness skills, letter identification has been recognized as the
strongest predictor of later literacy achievement.67
Phonological The concept of phonological awareness has been defined as “a skill that
awareness allows kids to recognize and work with the sounds of spoken language”. 68
In the preschool years, this can be demonstrated by children’s abilities to
identify words that rhyme or to clap out the number of syllables in a name.
Development of phonological awareness is crucial given its strong and
consistent links with early reading and spelling success in every language in
which it has been researched.69
Expressive This construct is defined as “the ability to communicate verbally with others”. 70
language There is great variety in the development of expressive language capacities,
resulting in a range of what is defined as ‘normal’ development.71 Measuring
expressive language is especially important because children who demonstrate
difficulty in expressing themselves have been found to be at greater risk for
language, social and academic problems in later years.72
Receptive This construct refers to the ability to hear, listen and comprehend oral
language information. Children who have difficulties understanding others may find it
challenging to follow instructions and/or appropriately respond to questions or
requests. Having foundational oral language skills are also critical for general
reading achievement at older ages.73
Executive Mental This set of skills helps children to both sustain attention when it is required as
function flexibility well as shift attention in response to different demands and settings. Sustained
attention has been found to modestly, but consistently, predict academic
achievement outcomes, both in preschool and in the early grades, even after
controlling for the effects of other cognitive and language abilities.74 Children’s
inability to pay attention or ‘do what is expected of them’ has been associated
with anti-social behaviours, peer rejection and lower academic achievement.75
Learning Executive Working Working memory refers to the ability to hold information in the mind and
function memory manipulate that information in order to perform tasks.76 It is an important
component of ECD since, along with attention, it has been found to be
associated with academic outcomes and years of schooling, with stronger
relationships documented in the early childhood period.77 Working memory
was also identified as the best predictor of reading, spelling and math skills six
years after assessment.78
Inhibition Inhibition, also known as inhibitory control, has been identified as one of
the first executive function skills to emerge, developing quickly during the
preschool years.79 Measuring inhibition during ECD is important because
control skills help children suppress inappropriate behaviours and thoughts
that may distract them from instruction.80 Further, when measured prior to
kindergarten age, inhibitory control skills predict math skills in kindergarten.
When measured in kindergarten, inhibitory control predicts math and early
literacy skills in primary school.81
Psycho- Social Social Social competence is a construct broadly defined as “children’s ability to be
social development competence effective in their social interactions with respect to achieving their goals”. 82 It
well-being is distinct from emotional competence or regulatory competence in that it is
the enactment, or behavioural manifestation, of other competencies.83 Social
competence is a composite term, representing a multitude of skills, which
is advantageous in that it leads to many broad, describable and long-term
outcomes but is disadvantageous in that many of the associated skills are
interlinked and therefore difficult to differentiate from one another.
A substantial body of evidence links social competence to a range of
later life outcomes and demonstrates how it affects children across other
developmental domains.84 It has also been found that children with early
social competence skills are more likely to have higher IQs, positive self-worth
and better mental health.85 Lastly, social competence (along with emotional
competence) provides the foundation for mastery of a wide range of skills
crucial for successful academic achievement.86
Relationship Relationship skills have been defined as the “ability to establish and maintain
skills healthy and rewarding relationships with diverse individuals and groups.
The ability to communicate clearly, listen well, cooperate with others, resist
inappropriate social pressure, negotiate conflict constructively, as well as
seek and offer help when needed”. 87 Relevant group relations in the preschool
years include interaction beyond the dyad, friendships, gender segregation,
dominance hierarchies and peer acceptance and rejection.88
Measuring relationship skills is especially important since a bi-directional
influence is said to exist between peer relationships and psychosocial and
emotional disorders in children. That is, peer problems may contribute to the
onset of disorders, while children with disorders may find themselves “at odds
with their peers from the very first years of life”. 89
Interpersonal Interpersonal conflict resolution refers to the strategies and methods used
conflict by individuals and groups to peacefully negotiate interpersonal disputes. It
resolution includes the ability to express emotions relating to an interpersonal conflict,
e.g., sharing a toy. Interpersonal conflict resolution skills in the early years
develop when a child learns to inhibit aggressive behaviours. While aggression
manifests early and declines over the first five years, it is considered abnormal
when it is pervasive, frequent and severe.90
Capturing interpersonal conflict resolution skills is necessary because of its
salience on later aggression and its impact on a child’s later life outcomes.
Further, persistent and highly aggressive behaviour is associated with co-
occurring language problems, impulsivity, hyperactivity, poorly regulated
negative emotions and defiance. The directionality of these associations,
however, is not clear.91
49
Psycho- Social Prosocial Prosocial behaviour, also termed as helpful or kind behaviour, has been defined
social development behaviour as “voluntary actions specifically intended to benefit or improve the well-being
well-being of another individual or group of individuals”. 92 Examples of prosocial behaviour
include being respectful, helping, sharing, consoling, comforting, cooperating and
protecting others from real or potential harm.93 Prosocial behaviour is considered
to be a construct that lies within the moral and cultural dimensions of development
and exists across different subnational, national, cultural and faith contexts.
Measuring prosocial behaviour in early childhood is especially important because
its development and manifestation during these years is associated with peer
acceptance, empathy, self-confidence and emotion regulation skills.94 Prosocial
behaviour in early childhood has also been shown to be an antecedent to later
positive relationship trajectories.95
Social Social cognition refers specifically to “the way in which people process, remember,
cognition and use information in social contexts to explain and predict their own behaviour
and that of others”. 96 Skills such as the ability to understand, describe and predict
people’s mental states all constitute a pathway for children to develop strong social
cognition.97
Social cognition in the early years is critical because its related skills may impact
the quality of relationships and school success. Children with better social cognition
tend to be better communicators, more socially competent, more popular with
peers, happier at school and academically more advanced.98 Further, the literature
shows that children who are not able to discern the thoughts and feelings of others
have a higher probability of acting aggressively and experiencing peer rejection.99
Theory of Theory of mind refers to “the understanding of epistemic mental states such as
mind knowledge and belief, as well as motivational mental states such as desire and
emotion, and their consequences on people’s behaviours”. 100 In other words, it is
the ability to attribute mental states to both oneself and others and to recognize
others’ perspectives even if they differ from one’s own. It is the cognitive aspect
of interpreting other people’s intentions, desires and beliefs, while empathy is the
emotional aspect.
Theory of mind has important links with children’s success in social interactions and
in academics. Children with strong theory of mind are better communicators, better
at resolving conflicts with friends and engage in more complex pretend play. They
are rated more socially competent by teachers, are happier in school, more popular
with peers and produce school work that is more advanced in some ways.101
Emotional Empathy Empathy is defined as “the ability to perceive, understand, and react to other
development people’s emotions appropriately”. 102 It refers specifically to being able to feel what
another person is feeling or might be expected to feel.
Measuring empathy is especially important because it is a precursor to prosocial
behaviours. Empathy is distinguished from theory of mind, which describes the
mind of the child comprehending the mental states of others, while empathy refers
to the ‘heart’ of how that child comprehends the mental states of others.103
Emotion This construct is closely linked to the concept of emotion recognition, which refers
knowledge to the “awareness of feelings”. 104 Emotion knowledge is the ability to recognize one
is experiencing an emotion, comprehending one’s emotional experience within the
constraints of emotion scripts and the social context, and realizing that one’s own
inner and outer emotional states may differ from that of others.
Emotion knowledge has been recognized as being key for children when moving
into the world of peers and getting along with peers and adults.105 Therefore, a
child’s emotional knowledge is said to be both a precursor to social competence and
a skill that expands with social competence.106
Further, it has been posited that emotion knowledge is a construct that has
predicted academic success.107 Children who are weak on this construct are
unable to discern the thoughts and feelings of others and therefore have a higher
probability of behaving aggressively and experiencing peer rejection.108
Annex C. Draft set of ECDI2030 items for field testing
Summary Domain Subdomain
Knows numbers 1 to 5 Learning Numeracy
Gives correct amount (3) Learning Numeracy
Counts to 5 Learning Numeracy
Counts to 10 Learning Numeracy
Knows difference between big/small Learning Numeracy
Recognizes 3 letters of the alphabet Learning Literacy
Recognizes 4 simple words Learning Literacy
Stacks small objects (3) Learning Pre-writing
Fastens/unfastens buttons Health Fine motor skills
Writes his/her name Learning Pre-writing
Names an object consistently Learning Expressive language
Says 10 or more words Learning Expressive language
Asks using ‘what,’ ‘which,’ ‘where,’ ‘who’ Learning Expressive language
Says sentences of 3 or more words Learning Expressive language
Uses correctly ‘I,’ ‘you,’ ‘she,’ ‘he’ Learning Expressive language
Uses correctly ‘on,’ ‘in,’ ‘under’ Learning Expressive language
Asks using ‘why’ Learning Expressive language
Uses past tense Learning Expressive language
Identifies 7 objects Learning Expressive language
Sings short song or repeats from memory Learning Expressive language
Talks about what he/she sees, hears or does Learning Expressive language
Does an activity without asking for help or giving up Learning Executive function
Stops when told ”no” or “stop doing that” Learning Executive function
Acts impulsively (reversed) Learning Executive function
Follows instructions of more than one step Learning Executive function
Switches from one activity to another Learning Executive function
Concentrates on activity Learning Executive function
51
Summary Domain Subdomain
Pretends object is something else Learning Approaches to learning
Kicks, bites, hits Psychosocial Externalizing behaviour
Becomes withdrawn, shy Psychosocial Internalizing behaviour
Shows respect Psychosocial Social skills
Likes to meet new adults Psychosocial Social skills
Takes turns Psychosocial Social skills
Seeks help Psychosocial Social skills
Gets along with familiar children Psychosocial Social skills
Shares things Psychosocial Social skills
Asks about familiar people Psychosocial Social skills
Offers to help Psychosocial Emotional skills
Gets interested in a person who is sick Psychosocial Emotional skills
Helps with household chores Psychosocial Social skills
Says what others like or dislike Psychosocial Emotional skills
Calms down after periods of exciting activity Psychosocial Emotional skills
Shows when he/she needs to use the bathroom Psychosocial Emotional skills
Seems to be sad or depressed Psychosocial Internalizing behaviour
Says when he/she is unhappy, sad, angry Psychosocial Emotional skills
Destroys things (reversed) Psychosocial Externalizing behaviour
Dresses him/herself Health Self-care
Jumps with both feet Health Gross motor skills
Throws ball or stone Health Gross motor skills
Catches ball Health Gross motor skills
Walks on an uneven surface Health Gross motor skills
Runs without falling or bumping into objects Health Gross motor skills
Annex D. Results from field testing of draft set of ECDI2030 items in Mexico and State of Palestine
Percentage of mothers endorsing the items, by child age, Mexico
Domain- Don’t know Age

subdomain Construct Item or missing 2 years 3 years 4 years Total
Learning - Single-word verbal Says 10 or more words 0.16 84.52 93.70 96.31 91.56
Expressive communication
language Says 15 or more words 0.22 56.72 82.20 85.12 74.74
Says 20 or more words 0.22 38.17 66.57 70.60 58.53
Early sentence Says sentences of 3 or more words 0.02 55.10 89.98 95.75 80.40
construction
Says sentences of 4 or more words 0.00 38.54 83.07 91.02 71.04
Object recognition Names an object consistently 0.00 79.84 92.03 96.99 89.71
Identifies 7 objects 0.00 92.44 95.73 98.16 95.49
Person recognition Uses correctly ‘I,’ ‘you,’ ‘she,’ ‘he’ 0.05 48.27 72.46 84.78 68.73
Pragmatics Uses correctly ‘on,’ ‘in,’ ‘under’ 0.01 41.96 72.19 77.77 64.09
Temporal Uses past tense 0.00 20.46 56.83 76.62 51.67

understanding
Asking questions Asks using ‘what,’ ‘which,’ ‘where,’ 0.20 53.40 77.04 87.58 72.87
‘who’
Asks using ‘why’ 0.00 45.29 82.36 95.02 74.46
Other Sings short song or repeats from 0.01 65.47 84.27 95.27 81.87
memory
Talks about what he/she sees, hears or 0.00 59.02 87.21 96.99 81.26
does
Learning - Letter identification Recognizes 3 letters of the alphabet 0.13 15.35 29.72 57.34 34.63
Literacy
Recognizes 5 letters of the alphabet 0.12 4.79 16.62 35.26 19.22
Reading words Recognizes 4 simple words 0.11 7.13 8.93 16.39 10.95
Learning - Counting Gives correct amount (3) 0.27 24.06 46.70 72.27 48.14
Numeracy
Counts to 5 0.19 15.48 42.64 74.65 44.83
Counts to 10 0.11 3.46 19.97 38.20 20.87
Number identification Knows numbers 1 to 5 0.21 13.93 30.35 57.57 34.44
Number/size Knows difference between big/small 0.26 60.17 77.96 93.84 77.61
discrimination
Learning - Writing Writes his/her name 0.02 2.43 6.91 29.61 13.38
Pre-writing
Object manipulation Stacks small objects (3) 0.27 85.76 93.35 94.91 91.37
Learning - Following directions Follows instructions of more than one 0.24 73.14 82.72 83.37 79.76
Executive step
function
Mental flexibility Does an activity without asking for 0.00 58.62 73.20 82.71 71.69
help or giving up
Switches from one activity to another 0.01 72.88 78.57 84.80 78.86
Concentrates on activity 0.00 84.19 90.82 91.28 88.77
53
Learning - Inhibition/control Stops when told ”no” or “stop doing 0.00 69.85 71.38 71.53 70.93
Executive that”
function
Acts impulsively (reversed) 0.11 44.77 53.69 56.70 51.78
Learning - Approaches to learning play Pretends object is something else 0.04 69.82 84.07 74.36 75.92
Psychcosocial Sharing Shares things:

- Social
development Always 0.00 39.59 36.99 45.54 40.85
Sometimes 46.08 54.70 44.32 48.19
Never 14.33 8.31 10.14 10.96
Respect Shows respect:
Always 0.23 44.62 47.05 52.29 48.08
Sometimes 52.28 49.19 45.38 48.88
Never 2.89 3.34 2.27 2.81
Getting along with Gets along with other children 0.34 83.04 90.17 93.34 88.91
others
Helpfulness Offers to help 0.05 69.90 81.68 88.17 80.04
Helps with household chores:
Always 0.04 43.37 46.81 52.10 47.52
Sometimes 40.68 47.27 43.56 43.77
Never 15.89 5.92 4.28 8.66
Attachment/trust Becomes withdrawn, shy:
Never 0.03 25.64 31.71 28.93 28.72
A few times a year 19.49 15.70 22.48 19.34
Monthly 23.86 23.60 19.25 22.16
Weekly 19.89 15.21 18.76 18.01
Daily 11.12 13.75 10.52 11.74
Seeks help:
Always 0.19 57.80 53.04 53.47 54.77
Sometimes 37.96 41.56 41.38 40.30
Never 4.17 5.40 4.68 4.74
Other Likes to meet new adults 1.15 64.66 65.42 69.76 66.69
Psychosocial Empathy Gets interested in a person who is

- Emotional sick:
development
Always 0.60 35.00 48.65 51.42 45.08
Sometimes 42.99 43.09 44.11 43.41
Never 20.85 8.11 3.99 10.90
Psychosocial Aggressiveness Kicks, bites, hits:
- Emotional
development Not at all 0.02 41.83 51.50 62.63 52.19
The same or less 38.73 39.19 29.45 35.62
More 13.07 5.69 5.63 8.13
A lot more 6.33 3.62 2.29 4.05
Destroys things (reversed) 0.00 68.51 68.43 71.10 69.39
Self-control/ Shows when he/she needs to use the 0.00 74.85 92.81 93.51 87.08
regulation bathroom
Says when he/she is unhappy, sad, 0.05 41.42 66.04 80.13 62.79
angry
Emotion regulation Calms down after periods of exciting 0.03 75.99 76.87 83.15 78.78
activity
Patience Takes turns 0.61 42.78 54.35 64.33 54.00
Affect Seems to be sad or depressed:
Never 0.00 57.93 56.27 47.78 53.84
Sometimes 41.65 41.12 49.90 44.38
Always 0.42 2.61 2.33 1.78
Emotion recognition Asks about familiar people 0.00 78.61 92.30 94.89 88.65
Says what others like or dislike 0.24 29.68 55.37 81.91 56.13
Health - Dresses him/herself 0.05 24.94 54.79 74.50 51.77

Self-care
Fastens/unfastens buttons 0.04 36.68 54.59 77.88 56.80
Jumps with both feet 0.40 73.79 90.81 97.25 87.41
Health - Throws ball or stone 0.00 96.74 97.57 99.58 98.00

Gross motor
Catches ball 0.39 56.98 67.94 72.78 65.99
Walks on an uneven surface 0.32 76.88 85.77 88.87 83.90
Runs without falling or bumping into 0.36 81.65 84.21 94.46 86.95
objects
Health - No versus yes 2.32* 79.08 83.43 84.18 81.92

Stunting
* The child was not measured as he/she was not present or it was not allowed.
Note: Items in bold are in the final ECDI2030.
55
Percentage of mothers endorsing the items, by child age, State of Palestine

Learning - Single-word verbal Says 10 or more words 0.00 86.71 92.55 97.13 91.85
Expressive communication
language Says 15 or more words 0.32 79.41 86.61 93.59 86.18
Says 20 or more words 0.12 71.37 85.55 91.84 82.27
Early sentence Says sentences of 3 or more words 0.12 75.30 94.80 98.50 88.68
construction
Object recognition Names an object consistently 0.31 80.11 95.42 96.71 90.09
Identifies 7 objects 0.00 87.15 94.99 98.56 93.21
Person recognition Uses correctly ‘I,’ ‘you,’ ‘she,’ ‘he’ 0.13 63.00 85.73 96.14 80.58
Pragmatics Uses correctly ‘on,’ ‘in,’ ‘under’ 0.25 76.42 93.04 95.58 87.62
Temporal understanding Uses past tense 0.16 35.01 62.14 73.69 55.71
Asking questions Asks using ‘what,’ ‘which,’ ‘where,’ 0.14 69.79 91.25 94.62 84.28
‘who’
Asks using ‘why’ 0.14 39.42 75.13 87.34 65.69
Other Sings short song or repeats from 0.14 75.55 87.56 94.61 85.34
memory
Talks about what he/she sees, hears 0.00 67.71 89.78 94.62 83.06
or does
Learning - Letter identification Recognizes 3 letters of the alphabet 1.07 4.67 15.52 55.95 24.63
Literacy
letters of the alphabet 0.02 0.42 2.84 36.14 12.80
Reading words Recognizes 4 simple words 0.83 3.30 11.13 28.28 13.78
Learning - Counting Gives correct amount (3) 3.74 30.28 64.85 87.85 59.36
Numeracy
Counts to 5 0.95 31.54 61.50 86.10 58.26
Counts to 10 0.65 15.34 37.01 68.34 39.09
Number identification Knows numbers 1 to 5 0.36 21.44 41.19 67.47 42.34
Number/size Knows difference between big/small 2.23 56.62 81.55 94.56 76.42
discrimination
Learning - Writing Writes his/her name 0.00 3.54 8.46 33.02 14.62
Pre-writing
Object manipulation Stacks small objects (3) 0.20 83.91 92.42 96.67 90.61
Learning - Following directions Follows instructions of more than 0.60 79.94 90.47 89.94 86.34
Executive one step
function
Mental flexibility Does an activity without asking for 1.35 53.61 72.08 85.81 69.61
help or giving up
Switches from one activity to another 0.62 76.13 86.64 93.17 84.82
Concentrates on activity 0.95 82.05 91.23 94.61 88.88
Learning - Inhibition/control Stops when told ”no” or “stop 0.01 76.70 76.97 77.50 77.04
Executive doing that”
function
Acts impulsively (reversed) 2.12 58.05 65.06 64.71 62.31
Learning - Approaches to learning play Pretends object is something else 0.71 76.89 77.68 83.13 79.16
Psychcosocial Sharing Shares things:

- Social
development Always 1.29 37.85 48.27 50.90 45.21
Sometimes 47.59 43.59 42.20 44.64
Never 12.84 7.57 5.43 8.85
Respect Shows respect:
Always 1.48 42.40 64.88 61.07 55.18
Sometimes 44.74 28.83 33.84 36.45
Never 9.24 5.90 5.09 6.89
Getting along with others Gets along with other children 1.13 85.84 89.50 90.74 88.53
Helpfulness Offers to help 3.30 65.74 76.29 89.59 76.66
Helps with household chores:
Always 0.37 45.93 47.57 45.72 46.35
Sometimes 39.70 47.12 50.14 45.32
Never 13.38 5.32 4.14 7.97
Attachment/trust Becomes withdrawn, shy:
Never 0.99 50.19 50.91 43.81 48.32
A few times a year 25.94 24.08 32.70 27.59
Monthly 10.44 11.30 11.59 11.07
Weekly 7.77 9.89 7.70 8.38
Daily 4.05 3.31 3.48 3.65
Seeks help:
Always 0.34 52.18 44.87 43.75 47.25
Sometimes 42.84 47.18 52.14 47.17
Never 4.08 7.95 4.11 5.24
Other Likes to meet new adults 1.49 67.29 79.29 81.05 75.35
Psychosocial Empathy Gets interested in a person who

- Emotional is sick:
development
Always 1.43 46.92 61.79 60.25 55.70
Sometimes 37.48 30.65 33.87 34.27
Never 12.72 7.44 4.93 8.60
57
Psychosocial Aggressiveness Kicks, bites, hits:
- Emotional
development Not at all 0.27 24.01 32.46 30.39 28.60
The same or less 58.94 52.06 57.03 56.27
More 12.51 10.65 9.84 11.08
A lot more 3.85 4.80 2.75 3.77
Destroys things (reversed) 0.14 71.77 73.98 79.27 74.88
Self-control/regulation Shows when he/she needs to use 0.09 60.81 92.70 96.54 81.96
the bathroom
Says when he/she is unhappy, 0.83 66.48 79.94 83.51 76.05

sad, angry
Emotion regulation Calms down after periods of 0.27 72.27 81.22 81.05 77.80
exciting activity
Patience Takes turns 1.24 49.76 53.82 65.07 55.97
Affect Seems to be sad or depressed:
Never 1.21 71.01 74.13 67.41 7076
Sometimes 27.02 21.52 31.01 26.68
Always 1.42 1.77 0.87 1.34
Emotion recognition Asks about familiar people 0.27 88.14 95.44 94.55 92.40
Says what others like or dislike 3.09 26.91 56.00 67.38 48.77
Health - Dresses him/herself 0.28 42.17 77.12 90.09 68.21

Self-care
Fastens/unfastens buttons 1.11 31.27 66.32 78.70 57.18
Jumps with both feet 0.67 76.42 95.09 97.82 88.96
Health - Throws ball or stone 0.02 94.47 96.93 98.41 96.49

Gross motor
Catches ball 1.01 80.67 94.21 93.22 88.79
Walks on an uneven surface 0.86 73.55 91.88 94.54 85.86
Runs without falling or bumping 0.06 72.05 84.64 90.54 81.83

into objects
Note: Items in bold are in the final ECDI2030.
Annex E. Results of the decision-matrix
Flw Domain Result according to decision-matrix
Recognizes 3 letters of alphabet Green Learning Item passed all criteria but was dropped as it was
considered that another item made it redundant
Counts to 5 Yellow Learning Item was too easy in CREDI and HKU and not
discriminant in NLSCY 2008
Likes to meet new adults Yellow Psychosocial Item was not discriminant in Mexico and HKU
Gets along with other children Yellow Psychosocial Item was not discriminant in Mexico
Shows when he/she needs to use the Yellow Psychosocial Item was not discriminant in Mexico and HKU
bathroom
Catches ball Yellow Health Item was too easy for children aged 3 to 4 years
Recognizes 10 letters of alphabet Red Learning Item was too difficult and with little variation across
ages in Mexico, State of Palestine and some MELQO
countries
Recognizes 4 simple words Red Learning Item was too difficult in all countries except HKU
Stacks small objects (3) Red Learning Item was too easy in all countries with almost no
variation across ages
Says 15 or more words Red Learning Item was too easy and less discriminant than another
similar item
Says sentences of 4 or more words Red Learning Item was too easy in Belize and Mexico
Identifies 7 objects Red Learning Item was too easy and with little variation with age in
Mexico, State of Palestine and HKU
Stops when told “no” or “stop doing that” Red Learning Item was not discriminant in most countries
Acts impulsively (reversed) Red Learning Item was not discriminant in most countries
Switches from one activity to another Red Learning Item was not discriminant in Mexico
Concentrates on activity Red Learning Item was not discriminant in Mexico and CREDI
Pretends object is something else Red Learning Item was not discriminant in most countries
Becomes withdrawn, shy Red Learning Item was not discriminant in any country
Seeks help Red Psychosocial Item was not discriminant in any country
Seems to be sad or depressed Red Psychosocial Item was not discriminant in most countries
Destroys things (reversed) Red Psychosocial Item was not discriminant in most countries
Throws ball or stone Red Psychosocial Item was too easy and with low discrimination in many
countries
59
Annex F. Results of ATA in Mexico and State of Palestine and the final version of the ECDI2030
Item ATA Mexico ATA State of Palestine Final ECDI2030

Knows numbers 1 to 5 + + +
Gives correct amount (3) + + +
Counts to 10 + + +
Recognizes 5 letters of the alphabet + + +
Fastens/unfastens buttons + + +
Writes his/her name + + +
Names an object consistently – + +
Says 10 or more words + – +
Says 20 or more words – – –
Asks using ‘what,’ ‘which,’ ‘where,’ ‘who’ – – –
Says sentences of 3 or more words + + +
Says sentences of 5 or more words + + +
Uses correctly ‘I,’ ‘you,’ ‘she,’ ‘he’ – + +
Uses correctly ‘on,’ ‘in,’ ‘under’ – + –
Asks using ‘why’ + – –
Uses past tense – – –
Sings short song or repeats from memory – – –
Talks about what he/she sees, hears or does + – –
Does an activity without asking for help or giving up + + +
Follows instructions of more than one step – – –
Kicks, bites, hits + + +
Shows respect – – –
Takes turns – – –
Gets along with other children – – +
Asks about familiar people + + +
Offers to help + + +
Gets interested in a person who is sick + + –
Helps with household chores – – –
Calms down after periods of exciting activity – – –
Seems to be sad or depressed + + +
Dresses him/herself + + +
Jumps with both feet + + +
Walks on an uneven surface + + +
Runs without falling or bumping into objects – – –
Note: ‘+’ indicates the item was included on the test form, ‘-’ indicates that the item was not included on the test form.
Annex G. Statistical properties of the final ECDI2030 and associations with child background
characteristics, Mexico and State of Palestine
Statistical properties of the final ECDI2030, Mexico
Percentage of children Point-biserial

passing the item correlations Item-level analyses
2 3 4 Discrimi-
Domain Item years years years Domain Age nation Difficulty
Learning Can (name) say at least ten or more words like 84% 94% 96% 0.79 0.33 1.90 -1.88
“Mama” or “ball”?
Can (name) speak using sentences of three or more 54% 90% 96% 0.87 0.63 2.95 -0.98
words that go together, for example, “I want water”
or “The house is big”?
Can (name) speak using sentences of five or more 20% 66% 82% 0.80 0.61 2.22 -0.21
words that go together?
If you show (name) an object he/she knows well, 79% 92% 97% 0.72 0.45 1.85 -1.74
such as a cup or animal, can he/she consistently
name it?
Can (name) correctly use any of the words ‘I,’ ‘you,’ 47% 72% 85% 0.76 0.44 1.72 -0.68
‘she,’ or ‘he’, for example, “I go to the store,” or “He
eats rice”?
Can (name) recognize at least five letters of the 4% 17% 35% 0.83 0.53 2.01 1.14
alphabet?
If you ask (name) to give you three objects, such as 23% 46% 72% 0.77 0.49 1.61 0.07
three stones or three beans, does (he/she) give you
the correct amount?
Can (name) count 10 objects, for example, 10 3% 19% 39% 0.83 0.51 1.95 1.08
fingers or blocks, without mistakes?
Does (name) know all numbers from 1 to 5? 14% 29% 58% 0.76 0.51 1.59 0.58
Can (name) write his/her own name? 2% 7% 29% 0.80 0.60 1.84 1.52
Can (name) do an activity such as colouring without 59% 73% 82% 0.60 0.30 0.90 -1.20
repeatedly asking for help or giving up too quickly?
Psychosocial Does (name) get along well with other children? 83% 90% 93% 0.85 0.22 0.77 -3.17
well-being
Does (name) ask about familiar people other than 78% 92% 95% 0.81 0.37 1.48 -1.86
parents when they are not there, for example,
“Where is Grandma?”?
Does (name) offer to help someone who seems to 69% 82% 88% 0.90 0.27 0.98 -1.68
need help?
How often does (name) seem to be very sad or 98% 97% 98% 0.49 -0.11 0.03 -146.01
depressed?
Compared with children of the same age, how 81% 90% 92% 0.83 0.23 0.62 -3.41
much does (name) kick, bite or hit other children or
adults?
Health Can (name) dress him/herself, that is, put on pants 25% 55% 74% 0.87 0.51 1.26 -0.07
and shirt without help?
Can the child fasten and unfasten buttons without 36% 54% 78% 0.94 0.44 1.19 -0.29
help?
Can (name) jump up with both feet leaving the 74% 90% 97% 0.82 0.50 1.10 -2.12
ground?
Can (name) walk on an uneven surface, for 77% 86% 88% 0.72 0.22 0.75 -2.44
example, a bumpy or steep road, without falling?
61
Associations between the final ECDI2030 sum score and child background characteristics, Mexico
Mean Lower bound Upper bound
Mexico 13.2 12.9 13.5
24 to 35 months 10.1 9.6 10.6
36 to 41 months 13.4 13.1 13.7
42 to 59 months 15.8 15.4 16.1
Girls 13.6 13.0 14.1
Boys 12.5 12.4 13.2
Mother's education level primary or less 13.1 12.7 13.7
Mother's education level upper secondary or higher 13.4 12.4 14.2
Children not attending early childhood education1 12.9 12.4 13.3
Children attending early childhood education1 15.2 14.9 15.5
Children without early stimulation 11.7 11.0 12.4
Children with early stimulation 13.6 13.3 14
Children without books 12.4 12.0 12.8
Children with books 14.7 14.3 15.2
Children who are stunted 12.6 11.8 13.4
Children who are not stunted 13.3 12.9 12.7
Children living in poorest 20% of households 13.1 12.6 13.3
Children living in richest 20% of households 13.9 13.2 14.4
1
These results refer to children aged 36 to 49 months only.
Statistical properties of the final ECDI2030, State of Palestine
Percentage of children Point-biserial

passing the item correlations Item-level analyses
2 3 4 Discrimi-
Domain Item years years years Domain Age nation Difficulty
Learning Can (name) say at least ten or more words like 86% 93% 97% 0.79 0.44 1.48 -2.17
“Mama” or “ball”?
Can (name) speak using sentences of three or more 75% 95% 99% 0.87 0.65 2.98 -1.40
words that go together, for example, “I want water”
or “The house is big”?
Can (name) speak using sentences of five or more 39% 77% 93% 0.80 0.69 3.45 -0.51
words that go together?
If you show (name) an object he/she knows well, 80% 95% 97% 0.72 0.57 2.46 -1.58
such as a cup or animal, can he/she consistently
name it?
Can (name) correctly use any of the words ‘I,’ ‘you,’ 63% 86% 96% 0.76 0.65 3.86 -0.92
‘she,’ or ‘he’, for example, “I go to the store,” or “He
eats rice”?
Can (name) recognize at least five letters of the 2% 11% 50% 0.83 0.49 4.97 0.87
alphabet?
If you ask (name) to give you three objects, such as 30% 65% 88% 0.77 0.61 2.26 -0.29
three stones or three beans, does (he/she) give you
the correct amount?
Can (name) count 10 objects, for example, 10 15% 37% 68% 0.83 0.57 2.96 0.31
fingers or blocks, without mistakes?
Does (name) know all numbers from 1 to 5? 21% 41% 67% 0.76 0.54 2.29 0.24
Can (name) write his/her own name? 3% 9% 33% 0.80 0.38 2.07 1.36
Can (name) do an activity such as colouring without 54% 72% 86% 0.60 0.56 1.53 -0.76
repeatedly asking for help or giving up too quickly?
Psychosocial Does (name) get along well with other children? 86% 89% 91% 0.85 0.35 0.70 -3.16
well-being
Does (name) ask about familiar people other than 88% 95% 95% 0.81 0.44 1.36 -2.34
parents when they are not there, for example,
“Where is Grandma?”?
Does (name) offer to help someone who seems to 66% 76% 90% 0.90 0.44 1.13 -1.30
need help?
How often does (name) seem to be very sad or 99% 98% 99% 0.52 0.07 0.21 -20.67
depressed?
Compared with children of the same age, how 84% 84% 88% 0.78 0.05 0.32 -5.52
much does (name) kick, bite or hit other children or
adults?
Health Can (name) dress him/herself, that is, put on pants 42% 77% 91% 0.86 0.50 1.17 -0.82
and shirt without help?
Can the child fasten and unfasten buttons without 31% 66% 79% 0.84 0.52 1.23 -0.30
help?
Can (name) jump up with both feet leaving the 76% 95% 98% 0.78 0.43 1.50 -1.87
ground?
Can (name) walk on an uneven surface, for 73% 92% 95% 0.87 0.43 1.57 -1.60
example, a bumpy or steep road, without falling?
63
Associations between the final ECDI2030 sum score and child background characteristics, State of Palestine
Mean Lower bound Upper bound
State of Palestine 14.1 13.7 14.4
24 to 35 months 11.1 10.6 11.6
36 to 41 months 14.3 13.7 14.8
42 to 59 months 16.4 16.0 16.8
Girls 14.5 14.0 14.9
Boys 13.6 13.2 14.1
Children not attending early childhood education1 14.8 14.4 15.2
Children attending early childhood education1 17.7 17.3 18.1
Children without early stimulation 13.0 12.2 13.9
Children with early stimulation 14.2 13.8 14.5
Children without books 13.5 13.1 13.8
Children with books 16.2 15.5 16.8
Children living in poorest 20% of households 13.4 12.8 14.0
Children living in richest 20% of households 15.0 14.2 15.7
1
These results refer to children aged 36 to 49 months only.
Annex H. Membership of the global panel on ECD standard-setting
Claudia Regina Lindgren Alves, Professor, Department of Pediatrics, School of Medicine, Universidade Federal de
Minas Gerais, Minas Gerais, Brazil
Yvonne Becher, Chief Executive/Director, Programme Development and Learning, The Child Development
Centre, Hong Kong
Maureen Black, Professor, University of Maryland School of Medicine & RTI International, Baltimore, United
States
Gauri Divan, Director, Child Development Group, Sangath, Delhi, India
Kirsten Donald, Senior Specialist, Division of Developmental Paediatrics, Red Cross War Memorial Children’s
Hospital and Deputy Director, Neuroscience Institute, University of Cape Town, Cape Town, South Africa
Melissa Gladstone, Senior Lecturer in Neurodevelopmental Paediatrics and International Child Health, University
of Liverpool, Liverpool, United Kingdom
Frances Page Glascoe, Professor of Pediatrics, Vanderbilt University, Nashville, United States
Jennifer Grisham-Brown, Professor and Faculty Director of Early Childhood Laboratory, University of Kentucky,
Lexington, United States
Gwendoline Kandawasvika, Senior Consultant, Primary Health Sciences Department, Faculty of Medicine and
Health Sciences, University of Zimbabwe, Harare, Zimbabwe
Shazia Maqbool, Professor & Chair Developmental-Behavioural Paediatrics Department, The Children’s Hospital
and Institute of Child Health, Lahore, Pakistan
Fahmida Tofail, Scientist & Senior Consultant Physician, International Centre for Diarrhoeal Disease Research,
Dhakka, Bangladesh
Erika Marcela Osorio Valencia, Head of the Department of Developmental Neurobiology, National Institute of
Perinatology, Mexico City, Mexico
Meta Van den Heuvel, MD, PhD, Hospital for Sick Children, Toronto, Canada
Tao Xin, Deputy Director of the National Assessment Center for Education Quality, Ministry of Education, Beijing,
China
Pia Zeinoun, (former) Assistant Professor, American University of Beirut, Beirut, Lebanon
65
Endnotes
1. United Nations Children’s Fund, ‘Early Childhood 13. Duncan, Greg J., et al., ‘School Readiness and Later
Development Index 2030: A new tool to measure Achievement’, Developmental Psychology, vol. 43, no. 6,
SDG indicator 4.2.1’, undated, <https://data.unicef.org/ 2007, pp. 1428–1446.
resources/early-childhood-development-index-2030-
ecdi2030>, accessed 3 January 2022. 14. Ibid.
2. World Health Organization, United Nations Children’s Fund 15. Devercelli, Amanda Epstein et al., ‘Measuring
and World Bank Group, Nurturing Care for Early Childhood Child Development and Early Learning’, Early
Development: A framework for helping children survive Learning Partnership Guidance Note, World Bank,
and thrive to transform health and human potential, WHO, Washington, D.C., July 2016, <https://documents.
Geneva, 2018. worldbank.org/en/publication/documents-reports/
documentdetail/659701473955877219/measuring-child-
3. The Lancet, ‘Advancing Early Childhood Development: development-and-early-learning>, accessed 3 January
From science to scale’, Series Papers, 4 October 2016, 2022.
<www.thelancet.com/series/ECD2016>, accessed 3
January 2022. 16. ‘Meeting Report: Technical Consultation on SDGs indicator
4.2.1’.
4. For an overview of these tools, see: Fernald, Lia C. H., et
al., A Toolkit for Measuring Early Childhood Development 17. Moore, Kristin Anderson, et al., ’Flourishing from the Start:
in Low- and Middle-Income Countries, World Bank, What is it and how can it be measured?’, Child Trends
Washington, D.C., 2017, and the accompanying ‘ECD Research Brief, March 2017, <https://www.childtrends.org/
Measurement Inventory’, <https://documents.worldbank. publications/flourishing-start-can-measured>, accessed 21
org/en/publication/documents-reports/documentdeta June 2023.
il/384681513101293811/a-toolkit-for-measuring-early- 18. Early Child Development in Social Context.
childhood-development-in-low-and-middle-income-
countries>, accessed 3 January 2022. 19. Duncan et al., ‘School Readiness and Later Achievement’.
5. For a detailed overview of the development of the MICS 20. Early Child Development in Social Context.
ECDI, see: Loizillon, Anaïs, et al., ‘Development of the
21. Fernald et al., A Toolkit for Measuring Early Childhood
Early Childhood Development Index in MICS Surveys’,
Development in Low- and Middle-Income Countries.
MICS Methodological Papers, No. 6, Data and Analytics
Section, Division of Data, Research and Policy, United 22. Duncan et al., ‘School Readiness and Later Achievement’.
Nations Children’s Fund, New York, 2017.
23. Early Child Development in Social Context.
6. Ibid.
24. Halle, Tamara G., et al., In the Running for Successful
7. The Lancet, ‘Early Child Development in Developing Outcomes: Exploring the evidence for thresholds of school
Countries 2011’, Series Papers, 24 September 2011, readiness, Office of the Assistant Secretary for Planning
<www.thelancet.com/series/child-development-in- and Evaluation, US Department of Health and Human
developing-countries-2>, accessed 3 January 2022. Services, Washington, D.C., 2012.
8. United Nations Department of Economic and Social Affairs, 25. Anderson, Kate, and Abbie Raikes, ‘Key Measurement
Statistics Division, ’Ninth Meeting of the Inter-Agency Questions for SDG 4.2.1’, Discussion paper for Global
and Expert Group on the Sustainable Development Goal Alliance to Monitor Learning (GAML) Task Force 4.2,
Indicators, 26-28 March 2019: Report’, STA/441/2/162A/3, unpublished, 2017.
22 May 2019, <https://unstats.un.org/sdgs/meetings/iaeg-
sdgs-meeting-09/>, accessed 3 January 2022. 26. ‘Meeting Report: Technical Consultation on SDGs indicator
4.2.1’.
9. United Nations Children’s Fund, ‘Meeting Report: Technical
Consultation on SDGs indicator 4.2.1 – Revising the Early 27. Jurado, Maria-Beatriz, and Monica Rosselli, ‘The Elusive
Childhood Development Index (ECDI), 14-15 September Nature of Executive Functions: A review of our current
2016’, UNICEF, New York, 2017, unpublished. understanding’, Neuropsychology Review, vol. 17, no. 3,
2007, pp. 213–233.
10. Borisova, Ivelina, and Nell O’Donnell, Early Literacy and
Math Toolkit: Teacher training component – Facilitator’s 28. Administration for Children and Families, Head Start Early
guide, Save the Children USA, New York, 2013. Learning Outcomes Framework: Ages birth to five, US
Department of Health and Human Services, Washington,
11. Starr, Ariel, Melissa E. Libertus and Elizabeth M. Brannon, D.C., 2015; Fernald et al., A Toolkit for Measuring Early
‘Number Sense in Infancy Predicts Mathematical Abilities Childhood Development in Low- and Middle-Income
in Childhood’, Proceedings of the National Academy of Countries.
Sciences, vol. 110, no. 45, 2013, pp. 18116–18120.
29. Ibid.
12. Child Trends and Center for Child Health Research, Early
Child Development in Social Context: A chartbook, The 30. Jones, Stephanie M., et al., ‘Assessing Early Childhood
Commonwealth Fund, Fairfield, CT, 2004. Social and Emotional Development: Key conceptual and
measurement issues’, Journal of Applied Developmental
Psychology, vol. 45, 2016, pp. 42–48; Fernald et al., A Toolkit for 46. Cappa, Claudia, et al., ‘Identifying and Minimizing Errors
Measuring Early Childhood Development in Low- and Middle- in the Measurement of Early Childhood Development:
Income Countries. Lessons learned from the cognitive testing of the ECDI2030’,
International Journal of Environmental Research and Public
31. Rimm-Kaufman, Sara, and Lia Sandilos, ‘School Transition and Health, vol. 18, no. 22, 2021, pp. 1-13.
School Readiness: An outcome of early childhood development’,
in Encyclopedia on Early Child Development, edited by Richard 47. Tourangeau, Roger, ‘Cognitive Science and Survey Methods: A
E. Tremblay, Michel Boivin and Ray DeV. Peters, Centre of cognitive perspective’, in Cognitive Aspects of Survey Design:
Excellence for Early Childhood Development, Montreal, 2017, Building a bridge between disciplines, edited by Thomas B.
<www.child-encyclopedia.com/school-readiness/according- Jabine et al., National Academy Press, Washington, D.C., 1984.
experts/school-transition-and-school-readiness-outcome-early-
childhood>, accessed 3 January 2022. 48. Khan, Shane, M., et al., ‘Results of a Multi-Topic Field Test in
Belize 2019’, MICS Methodological Papers, No. 10, Data and
32. Best, John R., Patricia H. Miller and Jack A. Naglieri, ‘Relations Analytics Section, Division of Data, Analytics, Planning and
between Executive Function and Academic Achievement from Monitoring, United Nations Children’s Fund, New York, 2019.
Ages 5 to 17 in a Large, Representative National Sample’,
Learning and Individual Differences, vol. 21, no. 4, 2011, pp. 49. Bayley, Nancy, Bayley Scales of Infant and Toddler Development,
327–336. 3rd ed., Harcourt Assessment, San Antonio, TX, 2006.
33. Glewwe, Paul, Sofya Krutikova and Caine Rolleston, ’Do Schools 50. Inter-American Development Bank, ‘Regional Project on Child
Reinforce or Reduce Learning Gaps Between Advantaged and Development Indicators (PRIDI)’, <https://www.iadb.org/
Disadvantaged Students? Evidence from Vietnam and Peru’, en/sector/education/pridi/home#:~:text=WHAT%20IS%20
Economic Development and Cultural Change, vol. 65, no. 4, PRIDI%3F,aged%2024%20to%2059%20months>, accessed 3
2017, pp. 699-739. January 2022.
34. Head Start Early Learning Outcomes Framework. 51. Save the Children, ‘International Development and Early
Learning Assessment (IDELA)’, <https://idela-network.org>,
35. Rimm-Kaufman and Sandilos, ‘School Transition and School accessed 3 January 2020.
Readiness’.
52. Rao, Nirmala, et al., Developmentally on Track Assessment Scale
36. Halle, Tamara G., and Kristen E. Darling-Churchill, ‘Review of and Caregiver Survey, Faculty of Education, The University of
Measures of Social and Emotional Development’, Journal of Hong Kong, 2020.
Applied Developmental Psychology, vol. 45, 2016, pp. 8–18.
53. Bland, J. Martin, and Douglas G. Altman, ‘Measuring Agreement
37. Rubin, Kenneth H., and Linda Rose-Krasnor, ‘Interpersonal in Method Comparison Studies’, Statistical Methods in Medical
Problem Solving and Social Competence in Children,’ in Research, vol. 8, no. 2, 1999, pp. 135–160.
Handbook of Social Development: A lifespan perspective, edited
by Vincent B. Van Hasselt and Michel Herson, Plenum Press, 54. Information from Bangladesh, China, India and Myanmar
New York, 1992, pp. 283–323. (children aged 3 to 4 years old).
38. Save the Children, Social Emotional Learning Brief, Save the 55. Information from Bangladesh, Brazil, Cambodia, Chile, Colombia,
Children, undated, <https://resourcecentre.savethechildren.net/ Ghana, Guatemala, India, Jordan, Lao People’s Democratic
pdf/226._sc_social_emotional_learning_-_brief.pdf/>, accessed 3 Republic, Lebanon, Nepal, Pakistan, Philippines, United Republic
January 2022. of Tanzania, United States and Zambia (children aged 2 to 3
years old).
39. Early Child Development in Social Context.
56. Halpin, Peter, et al., ‘Monitoring Early Childhood Development at
40. Anderson Moore et al., Flourishing from the Start. the Population Level: The ECDI2030’, preprint, 2023, <https://
psyarxiv.com/6qcjb/>, accessed 24 February 2023.
41. Goodman, Alissa, et al., ‘Social and Emotional Skills in Childhood
and Their Long-term Effects on Adult Life: A review for the Early 57. Factor analyses were initially performed on all items, but
Intervention Foundation’, Institute for Education Cabinet Office following technical consultation with the TAG and IAEG-ECD, it
Child Poverty and Social Mobility Commission, London, 2015. was agreed that the results from factor loading would lead to
similar, but less informative, results as those derived from the
42. Halle et al., In the Running for Successful Outcomes. discrimination functions.
43. Campbell, Susan B., et al., ‘Commentary on the 58. With regards to this constraint, a simple application would
Review of Measures of Early Childhood Social and require that at least one item from each subdomain be
Emotional Development: Conceptualization, critique, and included on the final test form. However, this would result in
recommendations’, Journal of Applied Developmental the final measure being comprised almost entirely of items
Psychology, vol. 45, 2016, pp. 19–41. from the expressive language subdomain. This was deemed
44. Goodman et al., ‘Social and Emotional Skills in Childhood and an inadequate distribution of items by content domains.
Their Long-term Effects on Adult Life’. Therefore, it was required instead that a minimum of 4 items
from the health domain and 5 items from the psychosocial
45. Miller, Kristen, et al., eds., Cognitive Interviewing Methodology, domain be included on the final test form. This ensured that
John Wiley and Sons, Hoboken, NJ, 2014. the representation of these two domains on the final version
67
of the ECDI2030 was approximately proportional to their 69. Bradley, Lynette, and Peter Bryant. ‘Categorizing Sounds
representation in the screened item pool. Additionally, the and Learning to Read: A causal connection’, Nature, vol.
maximum number of items from the expressive language 301, no. 5899, 1983, pp. 419–421; Roth, Froma P., Deborah
subdomain was restricted to 5. L. Speece and David H. Cooper, ‘A Longitudinal Analysis
of the Connection Between Oral Language and Early
59. Petrowski, Nicole, et al., ‘Establishing Performance Reading’, Journal of Educational Research, vol. 95, no.
Standards for Child Development: Learnings from the 5, 2002, pp. 259– 272; Wagner, Richard K., and Joseph
ECDI2030’, submitted for publication, 2023. K. Torgesen, ‘The Nature of Phonological Processing
60. Angoff, William H., ‘Scales, Norms, and Equivalent Scores’, and Its Causal Role in the Acquisition of Reading Skills’,
in Educational Measurement, 2nd ed., edited by Robert L. Psychological Bulletin, vol. 101, no. 2, 1987, pp. 192–212.
Thorndike, American Council on Education, Washington, 70. Early Child Development in Social Context.
D.C., 1971, pp. 508–600.
71. Ibid.
61. Plake, Barbara S., and Gregory J. Cizek, ‘Variations on a
Theme: The modified Angoff, extended Angoff, and yes/ 72. Ibid.
no standard setting methods’, in Setting Performance
Standards: Foundations, methods and innovations, 2nd ed., 73. Duncan et al., ‘School Readiness and Later Achievement’;
edited by Gregory J. Cizek, Routledge, New York, 2012. Stanovich, Keith E., Anne E. Cunningham and Dorothy J.
Feeman, ‘Intelligence, Cognitive Skills, and Early Reading
62. Gelman, Rochel, and Charles R. Gallistel, The Child’s Progress’, Reading Research Quarterly, vol. 19, no. 3, 1984,
Understanding of Number, Harvard University Press, pp. 278–303.
Cambridge, MA, 1978. As cited in United Nations
Educational, Scientific and Cultural Organization, et al., 74. Duncan et al., ‘School Readiness and Later Achievement’.
Overview: MELQO Measuring Early Learning Quality and 75. Early Child Development in Social Context.
Outcomes, UNESCO, Paris, United Nations Children’s
Fund, New York, and Brookings Institution and the World 76. Head Start Early Learning Outcomes Framework.
Bank, Washington, D.C., 2017.
77. Stipek, Deborah, and Rachel A. Valentino, ‘Early Childhood
63. Jordan, Nancy C., et al., ‘Early Math Matters: Kindergarten Memory and Attention as Predictors of Academic Growth
number competence and later mathematics outcomes’, Trajectories’, Journal of Educational Psychology, vol. 107,
Developmental Psychology, vol. 45, no. 3, 2009, pp. 850– no. 3, 2015, pp. 771–788.
867; Clements, Douglas H. and Julie Sarama, Learning and
78. Alloway, Tracy P., ‘Working Memory Is a Better Predictor
Teaching Early Math: The Learning Trajectories Approach,
of Academic Success than IQ’, Psychology Today, 21
2nd ed., Routledge, New York, 2014.
December 2010.
64. Geary, David C., ‘From Infancy to Adulthood: The
79. Diamond, Adele, and Bertha Doar, ‘The Performance
development of numerical abilities’, European Child and
of Human Infants on a Measure of Frontal Cortex
Adolescent Psychiatry, vol. 9, no. 2, 2000, pp. S11–S16.
Function, the Delayed Response Task’, Developmental
65. Mix, Kelly S., Catherine M. Sandhofer and Arthur J. Psychobiology, vol. 22, no. 3, 1989, pp. 271–294; Diamond,
Baroody, ‘Number Words and Number Concepts: Adele, Natasha Kirkham and Dima Amso, ‘Conditions
The interplay of verbal and nonverbal quantification in under which Young Children Can Hold Two Rules in
early childhood’, in Advances in Child Development Mind and Inhibit a Prepotent Response’, Developmental
and Behavior, vol. 33, edited by Robert V. Kail, Elsevier Psychology, vol. 38, no. 3, 2002, pp. 352–362.
Academic Press, New York, 2005, pp. 305–346.
80. Alexander, Karl L., Doris R. Entwisle and Susan L.
66. Chard, David J., et al., ‘Using Measures of Number Sense Dauber, ‘First Grade Classroom Behavior: Its short- and
to Screen for Difficulties in Mathematics: Preliminary long-term consequences for school performance’, Child
findings’, Assessment for Effective Intervention, vol. 30, Development, vol. 64, no. 3, 1993, pp. 801–814; Blair,
no. 2, 2005, pp. 3–14. Clancy, ‘School Readiness: Integrating cognition and
emotion in a neurobiological conceptualization of children’s
67. McBride-Chang, Catherine, ‘The ABCs of the ABCs: The functioning at school entry’, American Psychologist, vol.
development of letter-name and letter-sound knowledge’, 57, no. 2, 2002, pp. 111–127; Sáez, Leilani, et al., ‘Relations
Merrill-Palmer Quarterly, vol. 45, no. 2, 1999, pp. 285–308; among Student Attention Behaviors, Teacher Practices,
Schatschneider, Christopher, et al., ‘Kindergarten Prediction and Beginning Word Reading Skill’, Journal of Learning
of Reading Skills: A longitudinal comparative analysis’, Disabilities, vol. 45, no. 5, 2012, pp. 418–432.
Journal of Educational Psychology, vol. 96, no. 2, 2004,
pp. 265–282; Wagner, Richard K., Joseph K. Torgesen 81. Blair, Clancy, and Rachel Peters Razza, ‘Relating
and Carol A. Rashotte, ‘Development of Reading-Related Effortful Control, Executive Function, and False Belief
Phonological Processing Abilities: New evidence of Understanding to Emerging Math and Literacy Ability in
bidirectional causality from a latent variable longitudinal Kindergarten’, Child Development, vol. 78, no. 2, 2007, pp.
study’, Developmental Psychology, vol. 30, no. 1, 1994, pp. 647–663.
73–87.
82. Blandon, Alysia Y., Susan D. Calkins and Susan P. Keane,
68. Johnson, Kelli, ‘What Is Phonological Awareness?’, ‘Predicting Emotional and Social Competence during
undated, <www.understood.org/en/articles/phonological- Early Childhood from Toddler Risk and Maternal Behavior’,
awareness-what-it-is-and-how-it-works>, accessed 3 Development and Psychopathology, vol. 22, no. 1, 2010,
January 2022. pp. 119–132.
83. Campbell et al., ‘Commentary on the Review of Measures of 101. Astington and Edwards, ‘The Development of Theory of Mind
Early Childhood Social and Emotional Development’. in Early Childhood’.
84. Ibid. 102. Wang and Wang, ‘The Mind and Heart of the Social Child’.
85. Early Child Development in Social Context. 103. Ibid.
86. Campbell et al., ‘Commentary on the Review of Measures of 104. ‘Measuring Child Development and Early Learning’.
Early Childhood Social and Emotional Development’; Fernald
et al., A Toolkit for Measuring Early Childhood Development in 105. Denham, Susanne A., ‘Dealing with Feelings: How children
Low- and Middle-Income Countries. negotiate the worlds of emotions and social relationships’,
Cognition, Brain, Behavior, vol. 11, no. 1, 2007, pp. 1–48.
87. Collaborative for Academic, Social, and Emotional Learning,
‘CASEL’s SEL Framework: What are the core competence 106. Campbell et al., ‘Commentary on the Review of Measures of
areas and where are they promoted’, 1 October 2020, <https:// Early Childhood Social and Emotional Development’.
sel4ma.org/resource/casels-sel-framework-what-are-the- 107. Denham, ‘Dealing with Feelings’.
core-competence-areas-and-where-are-they-promoted-2/>,
accessed 21 June 2023. 108. Denham et al., ‘Preschool Emotional Competence: Pathway
to social competence?’; Fernald et al., A Toolkit for Measuring
88. Hay, Dale F., Alexandra Payne and Andrea Chadwick, ‘Peer Early Childhood Development in Low- and Middle-Income
Relations in Childhood’, Journal of Child Psychology and Countries.
Psychiatry, vol. 45, no. 1, 2004, pp. 84–108.
89. Ibid.
90. Keenan, Kate, ‘Development of Physical Aggression from Early
Childhood to Adulthood’, in Encyclopedia on Early Childhood
Development [online], edited by Richard Tremblay, 2012.
91. Ibid.
92. Knafo-Noam, Ariel, ed., ‘Prosocial Behaviour: Synthesis’, in
Encyclopedia on Early Childhood Development [online], edited
by Richard Tremblay, Michel Boivin and Ray DeV Peters, 2016.
93. Dunfield, Kristen A., ‘A Construct Divided: Prosocial behavior
as helping, sharing, and comforting subtypes’, Frontiers in
Psychology, vol. 5, 2014, pp. 1-13.
94. Knafo-Noam, ‘Prosocial Behaviour: Synthesis’.
95. Ladd, Gary W., ‘School Transitions/School Readiness: An
outcome of early childhood development ~ Perspective:
Children’s social and scholastic development – Findings from
the Pathways Project,’ in Encyclopedia on Early Childhood
Development [online], edited by Richard Tremblay, Michel
Boivin and Ray DeV Peters, 2009.
96. Bulgarelli, Daniela, and Paola Molina, ‘Social Cognition in
Preschoolers: Effects of early experience and individual
differences’, Frontiers in Psychology, vol. 7, 2016, pp. 1-11.
97. Astington, Janet W., and Margaret J. Edward, ‘The
Development of Theory of Mind in Early Childhood’, in
Encyclopedia on Early Childhood Development [online], edited
by Richard Tremblay, Michel Boivin and Ray DeV Peters, 2010.
98. Ibid.
99. Denham, Susanne A., et al., ‘Preschool Emotional
Competence: Pathway to social competence?’, Child
Development, vol. 74, no. 1, 2003, pp. 238–256.
100. Wang, Zhenlin, and Lamei Wang, ‘The Mind and Heart of the
Social Child: Developing the empathy and theory of mind
scale’, Child Development Research, vol. 1, 2015, pp. 1–8.
69
United Nations Children’s Fund
Data and Analytics Section
Division of Data, Analytics, Planning and Monitoring
3 United Nations Plaza
New York, NY 10017, USA
Email: [email protected]
Website: data.unicef.org

ECDI2030 Technical Manual Sept 2023

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

ECDI2030 Technical Manual Sept 2023

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ECDI2030 Technical Manual Sept 2023

Uploaded by

Copyright:

Available Formats

THE EARLY CHILDHOOD DEVELOPMENT INDEX 2030

A NEW MEASURE OF EARLY CHILDHOOD DEVELOPMENT

Suggested citation: United Nations Children’s Fund, The Early

The report was edited by Tina Johnson, proofread by Lois Jensen

Part One: Early childhood development measurement 10

Part Two: Work process behind the ECDI2030 16

Part Three: Building the ECDI2030 20

Part Four: The final ECDI2030 42

The ECDI2030 addresses the need for nationally representative

The ECDI2030 can be integrated into existing national data

Why is this period of development so important?

Why is the measurement of ECD so complicated?

Additional constraints of population-level measurement in the context of multi-topic household

SEPTEMBER 2016 JANUARY 2018 NOVEMBER 2018 JUNE 2019

DRAFT SET OF QUESTIONS

REVISED DRAFT SET OF QUESTIONS

Identifying the best items using discrimination and difficulty properties

FINAL SET OF QUESTIONS

Defining cut scores through a standard-setting exercise (August 2019)

CONSTRUCTION OF THE INDEX

Translating the module into the six official UN languages

Early numeracy* Language, literacy and communication* Fine motor development*

• Counting (verbal and set production) • Letter/alphabet knowledge

Social development* Emotional development*

• Social competence • Empathy

Self-care* Gross motor development*

Approaches to learning/play The differences between the constructs of social

• directly relate to children rather than inputs to

As a result of these findings, two items were dropped,

Interview script • There was confusion over concepts and terms

See (or point out) the symbol

See and say numbers

See and distinguish

Say (or recognize) and

Count and write the numbers

Count Know which number comes first

Sample Households Response Instruments

State of • West Women: 897 99.5% • Region • MICS household, Arabic

Item Domain Adapted from

Child gives three blocks Learning IDELA

Child counts ten blocks Learning BSID-III

Child counts five blocks Learning BSID-III

Child identifies big and small objects Learning BSID-III

Child stacks three or more blocks Learning BSID-III

Child writes his/her name Learning DOT

Child uses verbs in past tense Learning PRIDI

Child names ten objects correctly Learning BSID-III

Child recognizes ten objects correctly Learning BSID-III

Child throws a ball Gross motor BSID-III

Child catches a ball Gross motor PRIDI

Institutional Supervision Supervision

Contact local authorities

Development of questionnaires, ECD direct assessment, CAPI programming, fieldwork materials

Number of children Population-level

B. State of Palestine, Palestinian Central Bureau of 2-4 years 1,099 Yes

C. CREDI (17 countries1 aggregated) 2-3 years 16,317 No

D. The University of Hong Kong (Bangladesh, China, 3-4 years 633 No