The Inclusive Internet Index 2020 Methodology report - A report by The Economist Intelligence Unit
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
The Inclusive Internet Index 2020
Methodology report
A report by The Economist Intelligence Unit
Sponsored byThe world leader in global business intelligence
The Economist Intelligence Unit (The EIU) is the research and analysis division of The Economist Group, the sister company
to The Economist newspaper. Created in 1946, we have over 70 years’ experience in helping businesses, financial firms and
governments to understand how the world is changing and how that creates opportunities to be seized and risks to be managed.
Given that many of the issues facing the world have an international (if not global) dimension, The EIU is ideally positioned to be
commentator, interpreter and forecaster on the phenomenon of globalisation as it gathers pace and impact.
EIU subscription services
The world’s leading organisations rely on our subscription services for data, analysis and forecasts to keep them informed about
what is happening around the world. We specialise in:
•C ountry Analysis: Access to regular, detailed country-specific economic and political forecasts, as well as assessments of
the business and regulatory environments in different markets.
• Risk Analysis: Our risk services identify actual and potential threats around the world and help our clients understand the
implications for their organisations.
• Industry Analysis: Five year forecasts, analysis of key themes and news analysis for six key industries in 60 major
economies. These forecasts are based on the latest data and in-depth analysis of industry trends.
EIU Consulting
EIU Consulting is a bespoke service designed to provide solutions specific to our customers’ needs. We specialise in these key
sectors:
• EIU Consumer: We help consumer-facing companies to enter new markets as well as deliver greater success in current
markets. We work globally, supporting senior management with strategic initiatives, M&A due diligence, demand forecasting
and other issues of fundamental importance to their corporations. Find out more at eiu.com/consumer
• Healthcare: Together with our two specialised consultancies, Bazian and Clearstate, The EIU helps healthcare organisations
build and maintain successful and sustainable businesses across the healthcare ecosystem. Find out more at: eiu.com/
healthcare
• Public Policy: Trusted by the sector’s most influential stakeholders, our global public policy practice provides evidence-
based research for policy-makers and stakeholders seeking clear and measurable outcomes. Find out more at: eiu.com/
publicpolicy
The Economist Corporate Network
The Economist Corporate Network (ECN) is The Economist Group’s advisory service for organisational leaders seeking to better
understand the economic and business environments of global markets. Delivering independent, thought-provoking content,
ECN provides clients with the knowledge, insight, and interaction that support better-informed strategies and decisions.
The Network is part of The Economist Intelligence Unit and is led by experts with in-depth understanding of the geographies and
markets they oversee. The Network’s membership-based operations cover Asia-Pacific, the Middle East, and Africa. Through a
distinctive blend of interactive conferences, specially designed events, C-suite discussions, member briefings, and high-calibre
research, The Economist Corporate Network delivers a range of macro (global, regional, national, and territorial) as well as
industry-focused analysis on prevailing conditions and forecast trends.THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
Contents
The Inclusive Internet Index 2020: Methodology 2
Acknowledgements 2
Scoring criteria and categories 3
Country selection 3
Updates to the 2020 index 4
Sources and definitions 8
Data modeling 9
Estimating missing data points 11
Aggregation and weights 11
Appendix 1: Detailed indicator list 13
Appendix 2: Background indicator list 19
Appendix 3: List of indicators and weights 21
1 © The Economist Intelligence Unit Limited 2020THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
The Inclusive Internet Index 2020:
Methodology
C ommissioned by Facebook, the fourth iteration of the Inclusive Internet Index assesses and
compares countries according to their enabling environment for adoption and productive use of
the Internet. The index outlines the current state of Internet inclusion across 100 countries, and aims to
help policymakers and influencers gain a clearer understanding of the factors that contribute to wide
and sustainable inclusion.
This document contains The Economist Intelligence Unit’s (EIU) methodology for the index. The
research results are available at this website.1
This year’s index includes a total of 100 countries (80 core countries and 20 rotating countries).
This year’s 20 rotating countries replace 20 countries included in last year’s index; these two groups
are switched out on an annual basis while the core group of 80 countries remains unchanged. The
composition of countries in this year’s Index is provided in the Country selection section below.
The EIU made minor changes to the methodology this year. First, this iteration of the index
incorporates three new indicators and modifies five previously-included indicators. As the dynamics
of Internet inclusion change over time, the indicators in the index must adapt to measure new
phenomena, and to ensure that the index remains a leading benchmark. This is particularly true as new
technologies such as 5G networks are deployed. The details of these modifications are discussed in the
sections below.
Some features previously introduced into the second and third iterations of the index were
maintained in the latest version. This includes outreach to governments for data confirmation and the
“Value of the Internet” survey—which this year focuses on the relationship between the Internet and
financial inclusion.
Acknowledgements
The EIU drew on the expertise of highly respected experts in the field of connectivity and inclusion
to provide input on the methodology, data sources and modeling options for the index. This panel of
experts was convened initially in 2016 and has interacted annually as necessary.
Project director Sabu Mathai sabumathai@eiu.com
Project manager Robert Smith robertsmith@eiu.com
Project analyst Richard Pedersen richardpedersen@eiu.com
1
https://
theinclusiveinternet.eiu.
com
2 © The Economist Intelligence Unit Limited 2020THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
Scoring criteria and categories
Categories, indicators and weights used in the index were selected on the basis of EIU analysis, a
literature review, and consultation with industry experts and specialists from academia and NGOs.
The EIU gathered data and conducted the research and analysis for all qualitative and quantitative
indicators.
The index contains 56 indicators organized across four categories:
1) Availability: This category captures the quality and breadth of available infrastructure
required for access. At a basic level, connectivity is limited where the infrastructure needed
for connection is insufficient or unavailable. The category looks at Internet use, the quality of
the Internet connection, and the type and quality of infrastructure available for Internet and
electricity access in both urban and rural areas of the country.
2) Affordability: This category examines the cost of accessing the Internet and considers
initiatives, whether private or public, to lower costs or other ways to promote access. Cost of
access relative to income is a critical factor in Internet adoption. The category includes factors
that focus on price, such as the cost of a handset or fixed-line broadband, and the competitive
environment for wireless and broadband operators.
3) Relevance: This category considers the value of being connected, in terms of useful services
and content and the availability of local content. If people do not find value in being connected,
then Internet adoption is less likely. The category measures the availability of local content, such
as whether basic information or government services are available online in the local language.
It also measures whether content and services that stimulate economic activity, such as those
relating to health, finance, commerce or entertainment, are available online. The category
includes measures which determine the value of the Internet to consumers.
4) Readiness: This category measures the capacity among Internet users to take advantage
of being online. The category looks at measures such as the level of literacy and educational
attainment, the level of web accessibility, privacy regulations, the level of trust in different
sources of information found online, national female e-inclusion policies and spectrum policy.
Each category receives a score, calculated from a weighted average of the underlying indicator
scores (see “Weights”). Scores are then scaled from 0 to 100, where 100 represents the strongest
environment for the adoption and productive use of the Internet. The overall country score (adjusted)
is a weighted average of the category scores.
3 © The Economist Intelligence Unit Limited 2020THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
Country selection
The Inclusive Internet Index 2020 evaluates the state of Internet inclusion in 100 countries. The
countries selected reflect a mix of high-income, middle-income and low-income countries, with a
wide range of geographic and demographic representation. The 100 economies selected for the index
represent approximately 91% of the world population and 96% of global GDP.
The fourth year of the index continues the pattern of maintaining a core group of 80 countries and
including subsets of 40 non-core countries. In other words, the index swaps out 20 non-core countries
on an annual basis. This year, 20 new non-core countries were added to the study, replacing 20 non-
core countries that were included in 2019.2 The 100 countries included this year are:
Americas Asia Europe Middle East and Africa
Core Argentina, Brazil, Australia, Bangladesh, Austria, Belgium, Algeria, Botswana,
Canada, Chile, Cambodia, China, India, Bulgaria, Denmark, Burkina Faso,
Colombia, El Salvador, Indonesia, Iran, Japan, Estonia, France, Cameroon, Cote
Guatemala, Jamaica, Kazakhstan, Malaysia, Germany, Greece, d’Ivoire, Egypt, Ethiopia,
Mexico, Peru, United Mongolia, Myanmar, Hungary, Ireland, Italy, Ghana, Kenya, Kuwait,
States, Venezuela Pakistan, Philippines, Netherlands, Poland, Liberia, Madagascar,
Singapore, South Korea, Portugal, Romania, Malawi, Morocco,
Sri Lanka, Taiwan, Russia, Spain, Sweden, Mozambique, Namibia,
Thailand, Vietnam Turkey, United Kingdom Nigeria, Oman, Qatar,
Rwanda, Saudi Arabia,
Senegal, South Africa,
Sudan, Tanzania,
Uganda, United Arab
Emirates, Zambia
Rotating Cuba, Honduras, Azerbaijan, Hong Kong, Croatia, Latvia, Bahrain, Burundi,
Nicaragua, Paraguay, Laos, New Zealand, Lithuania, Slovakia Gabon, Lebanon,
Trinidad & Tobago Papua New Guinea, Zimbabwe
Uzbekistan
Updates to the 2020 Index
The EIU team made several improvements to the index as a result of feedback collected after the
launch of the 2019 index. These updates are discussed below.
2
The non-core countries
that appeared in the third Changes and additions to the index framework
edition of this index and A series of updates were made to the index framework to capture developments in the enabling
that were rotated out of
environment for Internet inclusivity. The table below summarizes the indicators that were modified or
this edition are: Angola,
Benin, Congo (DRC), Costa added to the framework.
Rica, Czech Republic,
Dominican Republic,
Ecuador, Finland, Guinea,
Israel, Jordan, Mali, Nepal,
Niger, Panama, Sierra
Leone, Switzerland,
Tunisia, Ukraine and
Uruguay.
4 © The Economist Intelligence Unit Limited 2020THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
Category Indicator Indicator name Modifications/ Additions Rationale for change
number
Availability 1.3.3 Network Adjusted the weighting With 3G networks being upgraded to
coverage (min. scheme for the Infrastructure 4G, 4G has increased in importance.
4G) sub-category. This involved Countries have increasingly made 4G
increasing the relative coverage their goal due to advancements
weighting of the 4G in network technology. While the
indicator to reflect the worldwide availability of 4G continues
increased importance of to expand, the speed of 4G around the
4G connectivity, and in world varies more greatly from country
turn reducing the relative to country.
weighting of the 3G indicator.
Availability 1.3.4 5G deployment NEW INDICATOR: This 5G deployment is a forward-looking
indicator measures indicator. Roll-out of 5G has begun in
whether 5G New Radio limited cities in advanced economies
(NR) technology has been in 2019. It is expected to have wide
deployed in any area of the implications for economies and societies
country, either as a trial or for at large. This indicator specifically
commercial or public use. measures deployment of 5G New
Radio (NR) technology, and acts as a
proxy for overall development of 5G
infrastructure.
Affordability 2.1.2 Mobile phone Adjusted the indicator to Available data regarding the cost of
cost (prepaid capture developments in data 500 MB mobile data plans are updated
tariff) plans. This entailed replacing less frequently and no longer the
the 500 GB data plan international standard for measuring
measure with a 1 GB data data plan pricing. With expanded
plan measure. 4G coverage and growing adoption,
statistics regarding data affordability
most commonly measure 1 GB
data plans. Also, Internet inclusivity
organizations such as the Alliance
for Affordable Internet (A4AI) have
increased their targets for Internet
affordability and now define affordable
access in terms of the price of a 1 GB
plan.
Affordability 2.1.3 Mobile phone Adjusted the indicator to Available data regarding the cost of
cost (postpaid capture developments in data 500 MB mobile data plans is updated
tariff) plans. This entailed replacing less frequently and no longer the
the 500 GB data plan international standard for measuring
measure with a 1 GB data data plan pricing. With expanded
plan measure. 4G coverage and growing adoption,
statistics regarding data affordability
most commonly measure 1 GB
data plans. Also, Internet inclusivity
organizations such as the Alliance
for Affordable Internet (A4AI) have
increased their targets for Internet
affordability and now define affordable
access in terms of the price of a 1 GB
plan.
Relevance 3.1.1 Availability This indicator measures All but one country received a perfect
of basic whether the country has score on the existing indicator, so the
information domestic news websites that indicator has been revised to include
in the local provide information online in a higher bar that assesses whether or
language both official and non-official not local news sites are available in local
language(s). Previously, it languages. If domestic news websites
had measured whether are available in local languages, adoption
domestic news websites becomes more likely.
provide information in official
languages only.
5 © The Economist Intelligence Unit Limited 2020THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
Category Indicator Indicator name Modifications/ Additions Rationale for change
number
Relevance 3.2.3 e-Health content Added a higher scoring Score distribution has narrowed
option for countries that in each iteration of the index. The
provide e-Health services proposed modification aims to capture
via the Ministry of Health (or differentiation by tracking the provision
equivalent) website. of government e-Health services.
Relevance 3.2.8 Open data NEW INDICATOR: This Added because open data policies
policies measures the existence of promote transparency that can increase
government “open data” availability, affordability, efficiency,
policies that promote the competition, and innovation in the
dissemination of public- provision of essential services, and
sector (public and publicly enable oversight that makes government
funded) data. more accountable.
Readiness 4.3.3 Existence Added a higher scoring The existing indicator provided only
of national option for countries that limited differentiation in scores. The
broadband have developed or revised proposed modification also accounts
strategy a national broadband for technology neutrality with regard to
strategy within the past fixed and mobile broadband.
two years. Also, the
question has been revised
to capture technologically
neutral strategies (i.e., they
emphasize either fixed or
mobile broadband, or both).
Readiness 4.3.7 Government NEW INDICATOR: Roll-out of 5G has already begun in
efforts to This indicator assesses limited cities in advanced economies
promote 5G government efforts to in 2019. It is expected to have wide
promote 5G roll-out across implications for economies and societies
the use case applications writ large, and governments should be
(e.g., FWA, eMBB, mMTC, IoT, working to help accelerate the adoption
URLLC). of such a significant technological leap.
Backscoring for selected indicators
The EIU has backscored a number of indicators in this year’s index to account for revisions to previously
reported data.
Data publishing institutions may revise previously published data in light of new information.
Backscoring index data incorporates these revisions, making the most accurate data available in
the index. Having identified datasets that include revisions to prior year data, the EIU team has
incorporated those historical data revisions into the data collection and analysis process for this index
edition. Indicators in the index which have been backscored are based on ITU data, Gallup World Poll
gender gap data, UNESCO literacy data, and UNDP education data.
Users of the index conducting historical data analysis should utilize only the data published in this
edition of the index as reference data for year-on-year analysis.
Note: Due to the introduction of new metrics, revisions to existing indicators, and backscoring for
selected indicators, as well as the use of new sources and updates to the weighting system, the scores and
rankings for the 80 core countries may not be directly comparable between the 2020 and 2019 editions of
the Inclusive Internet Index. For additional information on comparability issues, please refer to the YOY
CHANGE tab in the accompanying Excel workbook.
6 © The Economist Intelligence Unit Limited 2020THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
Value of the Internet survey
The 2018 index introduced a novel survey, evaluating the value of the Internet to people in all index
countries.3 Repeated for the 2019 and 2020 editions of the index, the “Value of the Internet” survey
explores the ways in which the Internet brings value to people’s lives, from employment and shopping
to entertainment and self-expression. The survey was conducted among 4,953 people across 99
countries, who were interviewed using local languages. Both CATI (computer-assisted telephone
interviewing) and online methodologies were employed, depending on the market. Quotas were set at
the country level using standard census criteria to ensure consistent sampling and representation, and
to allow for reliable cross-country comparisons.
The following quotas were required within each country:
l Minimum sample size: 51 complete respondents
l Age: Minimum of 20% for each category: Millennials4 (born 1981–2001), Gen X (born 1965–1980),
Baby Boomers (born 1946–1964)
l Gender: Minimum of 30% male and 30% female respondents
l Household income: 50/50 split above and below the country median
l Community type: Mix of urban (major cities) and non-urban (suburban and rural) in each country;
Minimum of 10% rural respondents in each macro region (Asia, North America, Latin America,
Europe, Middle East & North Africa, Sub-Saharan Africa)
The main findings of the survey are captured in the Executive Summary paper, available here:
https://theinclusiveinternet.eiu.com/summary
Survey data in the index
The research team used survey data to build several indicators related to Relevance and Readiness,
focusing on user perceptions of factors such as trust and privacy. Such data were not easily available
through desk research and an existing survey with comparable data could not be found for all countries
included in the index. Below is a list of indicators that used survey data:
Category Indicator Indicator name Survey question
3
Azerbaijan, Cuba, Iran, number
Papua New Guinea and Relevance 3.2.2 Value of e-finance Which of the following forms of useful information have
Venezuela were excluded you accessed via the Internet at least once in the past year?
The indicator is ranked by responses indicating ‘Information
from the survey sample about personal finance’.
this year. At the time of Relevance 3.2.4 Value of e-health Which of the following forms of useful information have
fieldwork, no reliable you accessed via the Internet at least once in the past year?
online sample source was The indicator is ranked by responses indicating ‘Information
about health and fitness’.
available that could be
confidently verified. Given Relevance 3.2.5 e-Entertainment usage How often do you use the Internet for entertainment
purposes? The indicator is ranked by responses indicating
the target audience, phone ‘Several times a day’, ‘Every day’ and ‘Several times a week’.
access would also have
Relevance 3.2.7 Value of the Internet for How often do you purchase goods via the Internet? The
been extremely limited. e-commerce indicator is ranked by responses indicating ‘About once a
month’, ‘About once a week’ and ‘More than once a week’.
4
Adult Gen Z (1996–2000)
are included in this group.
7 © The Economist Intelligence Unit Limited 2020THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
Category Indicator Indicator name Survey question
number
Readiness 4.2.2 Trust in online privacy How confident are you that your activity online is private?
The indicator is ranked by responses indicating ‘Somewhat
confident’ and ‘Very confident’.
Readiness 4.2.3 Trust in government websites To what extent do you trust the information you receive
and apps from the following sources online?—‘Government websites/
apps.’ The indicator is ranked by responses indicating ‘Mostly’
and ‘Completely’.
Readiness 4.2.4 Trust in non-government To what extent do you trust the information you receive
website and apps from the following sources online?—‘Non-government
websites/apps that are based in my country.’ The indicator is
ranked by responses indicating ‘Mostly’ and ‘Completely’.
Readiness 4.2.5 Trust in information from To what extent do you trust the information you receive
social media from the following sources online?—‘Other people using
social media.’ The indicator is ranked by responses indicating
‘Mostly’ and ‘Completely’.
Readiness 4.2.6 e-Commerce safety To what extent do you agree with the following
statements?—‘Making purchases online is safe and secure.’
The indicator is ranked by responses indicating ‘Somewhat
agree’ and ‘Strongly agree’.
Sources and definitions
The EIU project team collected and analyzed all of the quantitative and qualitative data in the Inclusive
Internet Index 2020. We gathered the data from reputable international, national and industry sources,
including the EIU’s own internal databases. The data collection process lasted from September 2019 to
December 2019. Any changes to source data after December 2019 are not accounted for in this version
of the index.
In creating the Inclusive Internet Index, the EIU relied heavily on publicly available sources. This
research approach has the benefit of creating a fully transparent and repeatable methodology.
The main sources used in the Inclusive Internet Index are the EIU’s internal databases, Alexa
Internet, Cisco, Gallup, Google, GSMA, International Energy Association (IEA), International
Telecommunications Union (ITU), Ookla, OpenSignal, TeleGeography, United Nations Conference on
Trade and Development (UNCTAD), United Nations Development Program (UNDP), United Nations
Educational, Scientific and Cultural Organization (UNESCO), World Bank, country ministries, national
statistics offices, city administrative offices, national telecommunications authorities, domestic news
websites and industry associations.
Data availability is a critical issue in this index and in this field of study. The latest publicly available
data are not always up to date or recent, leading to pronounced effects in assessing performance in
such a fast-changing field. In addition, several international sources rely on data reported by countries.
Country governments may each adopt different methodologies to gather or analyze data, or lack the
means to report on the most recent data, which causes variations in data quality and timeliness.
To mitigate these risks, the EIU validated 10 selected indicators through a data confirmation process
in collaboration with the telecommunications ministry (or its equivalent) in each of the 100 countries in
this edition of the index. The following indicators were vetted with ministries:
8 © The Economist Intelligence Unit Limited 2020THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
Indicator name Unit Main source(s)
Internet users % of households ITU
Fixed-line broadband subscribers Per 100 inhabitants ITU
Mobile subscribers Per 100 inhabitants ITU
Network coverage (min. 2G) % of population ITU
Network coverage (min. 3G) % of population ITU
Network coverage (min. 4G) % of population ITU
Mobile phone cost (prepaid tariff) % of monthly GNI per capita A4AI, ITU, World Bank
Mobile phone cost (postpaid tariff) % of monthly GNI per capita ITU, World Bank
Level of literacy % of population UNESCO
Schools with Internet access % of schools UNESCO
In cases where the ministries provided data that were comparable (in terms of definitions and
methodologies used in calculation) to other data in the series, they were included in the study.
The research team hopes to improve the accuracy and type of data provided by both public agencies
and private companies to aid the understanding of Internet inclusion and welcomes comments and
suggestions to this end.
As discussed above, the index also uses data from a survey conducted by the EIU called “Value of the
Internet”. In total, nine indicators were based on survey results.
The index uses the latest year of available data as the reference year for all data series. Where data
existed for 2019, that value was used. If not, we used the following sequence:
1. Where available, we used data from the most recent year prior to 2019.
2. If the data were older than five years, we used other reliable data sources if the definitions and
technical notes in the series that were used to fill the data gaps aligned with those of the same
series.
3. If neither historical same-source data nor alternative source data were available, we used a data
imputation approach (see “Estimating missing data points”).
Data modeling
Indicator scores are transformed and then aggregated across categories to enable a comparison of
broader concepts across countries. The process of transforming involves rebasing the raw indicator
data to a common unit so that it can be aggregated. All indicators in this model are transformed into
a 0 to 100 scale, where 100 refers to the strongest enabling environment for Internet inclusion and 0
indicates the weakest environment for Internet inclusion.
Most indicators are transformed on the basis of a min/max normalization, where the minimum
and maximum raw data values across the 100 countries are used to bookend the indicator scores. The
indicators for which a higher value indicates a more favorable environment for Internet inclusion, such
as access to mobile phones, have been transformed on the basis of:
x = (x - Min(x)) / (Max(x) - Min(x))
where Min(x) and Max(x) are, respectively, the lowest and highest values in the 100 countries for
any given indicator. The value is then changed from a 0–1 value to a 0–100 score to make it directly
9 © The Economist Intelligence Unit Limited 2020THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
comparable with other indicators. This in effect means that the country with the highest raw data value
will score 100, while the lowest will score 0 for all indicators in the index.
It must be noted that the study focuses on comparing data across countries and between the 2020
and 2019 versions of the study. To that end, year-on-year comparisons of data took into account the
core 80 countries that were included in both the 2019 and 2020 iterations. These comparisons also took
into account updated data for previous years as part of our backscoring process, discussed above.
Several adjustments were made to quantitative indicators to manage the data structure or to account
for outliers. These adjustments are summarized in the table below:
Indicator Adjustment
Mobile subscribers Mobile cellular telephone subscriptions are subscriptions to a public mobile telephone
service that provide access to the PSTN (public switched telephone network) using cellular
technology. This includes, and is split into, the number of postpaid subscriptions and the
number of active prepaid accounts (i.e., that have been used during the past three months)
and applies to all mobile cellular subscriptions that offer voice communications. It excludes
subscriptions via data cards or USB modems, subscriptions to public mobile data services,
trunked private mobile radio, telepoint, radio paging and telemetry services. There is a cap
on mobile subscriptions at 130. All countries that exceed this value will receive 130 as the
maximum possible value. This cap accounts for differences in SIM user behavior, including
influxes in tourism, migrant workers, and other factors that can lead to the over-estimation of
the number of subscribers.
Wireless operators’ market The wireless operators’ market share was calculated using a commonly accepted method
share called the “Hirschman-Herfindahl Index”. Market concentration, however, does not follow a
strictly linear pattern. To account for that, the EIU used three scoring bands as follows: HHI
< 3,000, “unconcentrated”; HHI 3,000–4,000, “moderately concentrated”; and HHI > 4,000,
“highly concentrated”. This reflects the nature of the telecom industry, which tends to have
fewer players than most other industries. It is important to note these thresholds when using
the “Simulator” function in the Excel workbook. Changes to scores and ranks in the “Simulator”
function will be recorded only if a value is changed so that it moves to a different scoring band.
Broadband operators’ market The broadband operators’ market share was calculated using a commonly accepted method
share called the “Hirschman-Herfindahl Index”. Market concentration, however, does not follow a
strictly linear pattern. To account for that, the EIU used three scoring bands as follows: HHI
< 3,000, “unconcentrated”; HHI 3,000–4,000, “moderately concentrated”; and HHI > 4,000,
“highly concentrated”. This reflects the nature of the telecom industry, which tends to have
fewer players than most other industries. It is important to note these thresholds when using
the “Simulator” function in the Excel workbook. Changes to scores and ranks in the “Simulator”
function will be recorded only if a value is changed so that it moves to a different scoring band.
10 © The Economist Intelligence Unit Limited 2020THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
Estimating missing data points
5
For i) ad hoc weighting
schemes, the analyst
In cases where data were incomplete or missing, EIU analysts developed customized estimation
simply chooses the models to estimate data points, where appropriate. The concern at this stage of the data treatment
contribution of each
variable to the final process was assigning missing data using statistical methods. This was done for 7 indicators, which
composite indicator. While presented missing values that could not be obtained through comparable series, historical data or
an attempt to base the
weights is possible using regional estimates. Missing data were therefore populated using a modeling approach for the following
a theoretical framework indicators:
that assigns different
priorities to different
sub-dimensions, the Indicator Number of missing data points
final weight is—to some e-Commerce content 3
extent—always ad hoc.
A variant of the ad hoc Average mobile download speed 8
approach is to use a Average mobile latency 8
structured methodology
to determine the weights, Average mobile upload speed 8
although not one that Mobile phone cost (prepaid tariff) 5
is based upon statistical
optimization. For example, Mobile phone cost (postpaid tariff) 2
it may be possible to use Smartphone cost (handset) 2
survey data to weight
indicators by importance,
or a “traffic-light” system, For those indicators where data availability presented incomplete datasets, missing data were
where indicators can be
put into one of a number assigned using a regression-based approach. In order to calculate index values for countries with
of categories of high, missing data, we assigned missing values by predicting data using the Ordinary Least Squares (OLS)
medium or low weight.
method.
For ii) statistical
(optimization) methods,
the most common Aggregation and weights
approach is to use
principal component Methods to aggregate the transformed variables into a final composite indicator can be broadly
analysis (PCA). Intuitively, separated into two types: i) ad hoc weighting schemes; and ii) statistical (optimization) methods.5
the idea of PCA is to
reduce high-dimensional Given the difficulties in assigning weights, many composite indices resort to an equal weighting
data (several variables
scheme, allowing all variables to enter uniformly. The advantage of an equal weighting approach is
and sub-components
of an index) into lower- transparency, while the clear disadvantage is the lack of an underlying theoretical justification for equal
dimensional data by
grouping highly correlated treatment of all variables and sub-dimensions. Other approaches that allow users to adjust weights or
sub-components and that present a number of weighting scenarios pose the risk of obfuscating the index outcome.
variables into a linear
combination. In most The weighting process for any index is a qualitative determination that may reflect the biases or
cases, the factor loadings suppositions of the researchers. One method for gauging the soundness of a weighting scheme is to
for the first component
are used as the weights think of the weights as implicit trade-offs among the sub-dimensions of an indicator. As such, a short
for the final index. The
advantage of this approach
survey and consultation with individual experts was used to reflect on-the-ground priorities and the
is that the weights are practical shortcomings of existing data around Internet inclusion.
statistically determined
and, as a result, free from The four indicator categories used in the index are weighted according to a “lifecycle” approach,
value judgements. The under which the most important category is Availability, followed by Affordability, then Relevance, and,
disadvantage lies in the
lack of transparency: finally, Readiness. At the foundation of this approach is the conceptual sequencing of these categories
multivariate statistical in the lifecycle of activities that underlie an inclusive Internet: (1) if access to the Internet were not
methods are relatively
complex and the available due to limited infrastructure in a country, then affordability would matter less; (2) once access
methodology of such
indices is difficult to
becomes available and affordable, then relevant content will likely be a major driver of adoption; and
convey to a wider public. (3) once there is relevant adoption, the ability to take advantage of Internet access, as measured by
11 © The Economist Intelligence Unit Limited 2020THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
readiness, will in all probability become a factor. Although not always the case, the research team found
that this sequence applied to the vast majority of countries.
As the nature of connectivity and inclusion changes, it will be necessary to revisit the weighting
system applied to this index to see whether the logic still holds.
The weights assigned to each category are as follows:
Category Weight
Availability 40%
Affordability 30%
Relevance 20%
Readiness 10%
Total 100%
As part of the lifecycle approach, further adjustments were made to the weights of individual
indicators. These are intended to lessen any bias in factors relating to income and geography, and to
balance the influence of indicators across a sub-category or category. A summary of these adjustments
can be found below:
Indicator Sub-category weight Rationale
INFRASTRUCTURE
Network coverage (min. 3G) / 9.1% / 18.2% 4G coverage is now considered the standard to which
Network coverage (min. 4G) countries should strive when developing network
infrastructure. Therefore, the weightings for the 3G and
4G network coverage indicators were adjusted, with
greater weight given to the latter. As 5G developments
progress, the weightings for both of these indicators may
be adjusted again in the future.
POLICY
National digital identification system 15.4% Digital identification systems were traditionally more
relevant on account of how they were applied to
e-government services, and were generally not of broader
relevance to the entire Internet. However, their use is
becoming more prevalent around the globe, and they are
a potentially useful tool for promoting trust in the Internet.
Therefore, the weighting for this indicator was increased.
Examining the weighting scheme by comparing the relative importance of different dimensions is an
important tool for conducting robustness checks. To this end, the EIU has provided a way to compare
the effects of different weighting schemes on country rankings in the dashboard tool.
Despite the care that has been taken in selecting the indicators, categories and weights, no index
of this kind can ever be perfect. The EIU recognizes there are many different methods for weighting
an index. The weighting assigned to each category and indicator can be changed by users on the
‘Custom Weights’ tab of the dashboard tool to reflect different assumptions about their relative levels
of importance. This functionality enables users to create customized weightings that allow them to test
their own assumptions about the relative importance of each category and indicator. Users can also set
a weighting to zero to completely remove the influence of any category, indicator or sub-indicator on
the index results and country rankings. In addition to the pre-set weighting offered in the dashboard
tool, users can save one other bespoke weight setting and compare the effect of different settings on
country rankings and scores.
12 © The Economist Intelligence Unit Limited 2020APPENDIX THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
Appendix 1: Detailed indicator list
The categories, sub-categories and indicators are:
No. Indicator Unit Description Source
OVERALL 0-100 The overall score is the weighted sum of the following category scores: 1 to 4.
1 AVAILABILITY 0–100 This category captures the quality and breadth of available infrastructure
required for access. Connectivity is limited if the infrastructure to connect
is insufficient or unavailable. The score for the availability category is the
weighted sum of the following indicator scores: 1.1 to 1.4.
1.1 USAGE 0–100 Increased usage usually indicates greater connectivity, even if this may be
concentrated among certain groups. The usage score is the weighted sum of
the following indicators: 1.1.1 to 1.1.5.
1.1.1 Internet users % of households This measures the number of people who have used the Internet in the past ITU
12 months. A higher number of people using the Internet indicates greater
connectivity.
1.1.2 Fixed-line broadband Per 100 inhabitants This measures fixed-line broadband subscriptions per 100 inhabitants. ITU
subscribers The higher the number of subscriptions, the greater the level of Internet
connectivity.
1.1.3 Mobile subscribers Per 100 inhabitants This measures mobile-cellular telephone subscriptions per 100 inhabitants. ITU
A higher number of smartphones increases the propensity to use the
Internet and related services, especially advanced mobile services, though
this may be concentrated among certain groups.
1.1.4 Gender gap in Internet % difference, male This measures the gap between male and female access to the Internet (% EIU, Gallup, ITU
access to female access male access - % female access / % male access). Positive values indicate
that male access exceeds female access. A smaller or negative gap indicates
greater female connectivity.
1.1.5 Gender gap in mobile % difference, male This measures the gap between male and female access to mobile phones EIU, Gallup
phone access to female access (% male access - % female access / % male access). Positive values indicate
that male access exceeds female access. A smaller or negative gap indicates
greater female connectivity.
1.2 QUALITY 0–100 The higher the quality of the available infrastructure for access, the easier it is
to use a broader range of Internet sites and related services. The quality score
is the weighted sum of the following indicators: 1.2.1 to 1.2.7.
1.2.1 Average fixed broadband Kbps This measures average fixed broadband upload speed. Averages are based Ookla
upload speed on Ookla’s analysis of Speedtest data. A faster speed is a positive indicator for
better performance.
1.2.2 Average fixed broadband Kbps This measures average fixed broadband download speed. Averages are Ookla
download speed based on Ookla’s analysis of Speedtest data. A faster speed is a positive
indicator for better performance.
1.2.3 Average fixed broadband ms This measures average fixed broadband latency (or how long it takes for data Ookla
latency to travel between its source and destination). Averages are based on Ookla’s
analysis of Speedtest data. A faster speed is a positive indicator for better
performance.
1.2.4 Average mobile upload Kbps This measures average mobile upload speed. Averages are based on Ookla’s Ookla
speed analysis of Speedtest data. A faster speed is a positive indicator for better
performance.
1.2.5 Average mobile Kbps This measures average mobile download speed. Averages are based on Ookla
download speed Ookla’s analysis of Speedtest data. A faster speed is a positive indicator for
better performance.
1.2.6 Average mobile latency ms This measures average mobile latency (or how long it takes for data to Ookla
travel between its source and destination). Averages are based on Ookla’s
analysis of Speedtest data. A faster speed is a positive indicator for better
performance.
13 © The Economist Intelligence Unit Limited 2020APPENDIX THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
No. Indicator Unit Description Source
1.2.7 Bandwidth capacity Bit/s per Internet This measures the total used capacity of international Internet bandwidth, ITU
user in bits per second per Internet user. Used international Internet bandwidth
refers to the average traffic load (expressed in bits per second) of
international fiber optic cables and radio links carrying Internet traffic. More
Bits/s indicates better quality.
1.3 INFRASTRUCTURE 0–100 The wider the coverage of infrastructure for Internet access, the easier it is
for people to be connected. The infrastructure score is the weighted sum of
the following indicators: 1.3.1 to 1.3.7.
1.3.1 Network coverage (min. % of population This measures the percentage of people covered by 2G networks (number of ITU
2G) people as a percentage of the total population). The higher the percentage,
the greater the number of people connected.
1.3.2 Network coverage (min. % of population This measures the percentage of people covered by 3G networks (number of ITU
3G) people as a percentage of the total population). The higher the percentage,
the greater the number of people connected.
1.3.3 Network coverage (min. % of population This measures the percentage of people covered by 4G networks (number of ITU
4G) people as a percentage of the total population). The higher the percentage,
the greater the number of people connected.
1.3.4 5G Deployment Qualitative rating This indicator measures whether 5G New Radio (NR) technology has been EIU, Ookla
0–2; 2 = best deployed in any area of the country, either as a trial or for commercial or
public use.
1.3.5 Government initiatives to Qualitative rating This indicator looks at whether the government provides public Wi-Fi access EIU country
make Wi-Fi available 0–2; 2 = best in the largest city in the country and whether the Wi-Fi is free to connect to. research
Initiatives that come at no cost to the consumer are likely to promote usage.
1.3.6 Private-sector initiatives Qualitative rating This indicator looks at whether the largest privately owned ISP provides EIU country
to make Wi-Fi available 0–2; 2 = best public Wi-Fi access to its customers in the largest city in the country and research
whether the Wi-Fi is free to connect to. Initiatives that come at no cost to the
consumer are likely to promote usage.
1.3.7 Internet exchange points Number of IXPs This indicator measures the number of Internet exchange points (IXPs) in EIU, PCH,
per 10 million each country. The higher the number of IXPs, the wider the infrastructure PeeringDB,
inhabitants coverage. TeleGeography
1.4 ELECTRICITY 0–100 Electricity is needed to power the infrastructure and hardware required for
Internet access. More extensive electricity access increases the number of
people who are connected. The electricity score is the weighted sum of the
following indicators: 1.4.1 to 1.4.2.
1.4.1 Urban electricity access % of population This indicator measures the urban electrification rate (%). The higher the IEA, World Bank
percentage of population with access to electricity, the easier it is for people
to gain access to the Internet.
1.4.2 Rural electricity access % of population This indicator measures the rural electrification rate (%). The higher the IEA, World Bank
percentage of population with access to electricity, the easier it is for people
to gain access to the Internet.
2 AFFORDABILITY 0–100 This category looks at the cost of access to the Internet. Cost of access
relative to income is a critical factor in Internet adoption. The score for the
affordability category is the weighted sum of the following indicator scores:
2.1 to 2.2.
2.1 PRICE 0–100 The cost of access relative to income is an important factor for Internet
adoption. Generally, the lower the cost of access, the higher the adoption
rates. The price score is the weighted sum of the following indicators: 2.1.1
to 2.1.4.
2.1.1 Smartphone cost Score of 0–100; 100 This measures the indexed scores of the price of an entry-level handset to GSMA
(handset) = most affordable the consumer, as a percentage of GNI per capita. Generally, the lower the
cost of a smartphone handset, the higher the adoption rates.
2.1.2 Mobile phone cost % of monthly GNI This measures the price of a prepaid 1 GB mobile data plan, as a percentage A4AI, ITU, World
(prepaid tariff) per capita of monthly income. Generally, the lower the mobile phone data cost, the Bank
higher the adoption rates.
14 © The Economist Intelligence Unit Limited 2020APPENDIX THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
No. Indicator Unit Description Source
2.1.3 Mobile phone cost % of monthly GNI This measures the price of a postpaid 1 GB mobile data plan, as a percentage ITU, World Bank
(postpaid tariff) per capita of monthly income. Generally, the lower the mobile phone data cost, the
higher the adoption rates.
2.1.4 Fixed-line monthly % of monthly GNI This measures the price of fixed-line monthly broadband to the consumer as ITU, World Bank
broadband cost per capita a percentage of monthly income. Generally, the lower the broadband cost,
the higher the adoption rates.
2.2 COMPETITIVE 0–100 A healthy, competitive environment usually leads to lower prices for
ENVIRONMENT consumers. The competitive environment score is the weighted sum of the
following indicators: 2.2.1 to 2.2.4.
2.2.1 Average revenue per user USD This measures the average revenue per user (ARPU) for wireless operators. Axiata,
(ARPU, annualized) Generally, the higher the ARPU, the higher the adoption rates. GSMA, ITU,
TeleGeography
2.2.2 Wireless operators' HHI score This measures the market concentration among all wireless operators. The EIU,
market share (0–10,000) Hirschman-Herfindahl Index measures the concentration of markets as TeleGeography
follows: HHI < 3,000, “unconcentrated”; 3,000 ≤ HHI < 4,000, “moderately
concentrated”; and HHI ≥ 4,000, “highly concentrated”. A lower HHI score
indicates a more competitive environment.
2.2.3 Broadband operators' HHI score This measures the market concentration among all broadband operators. EIU,
market share (0-10,000) The Hirschman-Herfindahl Index measures the concentration of markets as TeleGeography
follows: HHI < 3,000, “unconcentrated”; 3,000 ≤ HHI < 4,000, “moderately
concentrated”; and HHI ≥ 4,000, “highly concentrated”. A lower HHI score
indicates a more competitive environment.
3 RELEVANCE 0–100 This category describes the value of being connected, in terms of useful
services and content and the availability of local content. If people do not
find value in being connected, then Internet adoption is less likely. The score
for the relevance category is the weighted sum of the following indicator
scores: 3.1 to 3.2.
3.1 LOCAL CONTENT 0–100 A key barrier for adoption is when local content does not meet local needs.
The higher the amount of local content, the higher likelihood of Internet
adoption. The local content score is the weighted sum of the following
indicators: 3.1.1 to 3.1.3.
3.1.1 Availability of basic Qualitative rating This indicator measures whether the country has domestic news websites EIU country
information in the local 0–3; 3 = best that provide information online in local language(s). It gives additional research
language credit for countries with news websites in local languages other than the
official language. If domestic news websites are available in local languages,
adoption becomes more likely.
3.1.2 Concentration of Qualitative rating This measures the proportion of websites in the top 25 most-visited websites Alexa Internet
websites using country- 0–3; 3 = best that use a country code top-level domain (ccTLD). The higher the proportion,
level domains the more likely there are popular websites catering to local content needs.
3.1.3 Availability of Qualitative rating This measures whether the government of the largest city in the country has EIU country
e-government services in 0–2; 2 = best a website that offers transactional services, including applying for a business research
the local language license or permit. The availability of government services online is likely to
increase adoption.
3.2 RELEVANT CONTENT 0–100 This measures whether there are content and services online that stimulate
economic or social activity. The relevant content score is the weighted sum of
the following indicators: 3.2.1 to 3.2.8.
3.2.1 e-Finance content Qualitative rating This measures whether the largest retail banking institution offers online EIU country
0–2; 2 = best banking services. Online banking services are likely to stimulate economic research
activity.
3.2.2 Value of e-finance % This is an indicator taken from the EIU “Value of the Internet” survey. The EIU survey
indicator looks at country-level responses to questions about personal
finance. A higher proportion of respondents that value e-finance in their
country suggests that more relevant content is available.
15 © The Economist Intelligence Unit Limited 2020APPENDIX THE INCLUSIVE INTERNET INDEX 2020
METHODOLOGY REPORT
No. Indicator Unit Description Source
3.2.3 e-Health content Qualitative rating This measures whether the Ministry of Health in the country has a website EIU country
0–3; 3 = best that provides information regarding healthcare, and also provides e-Health research
services functionalities. Easily available health information is likely to inform
both social and economic activity, and increase adoption.
3.2.4 Value of e-health % This is an indicator taken from the EIU “Value of the Internet” survey. The EIU survey
indicator looks at country-level responses to questions about health and
fitness. A higher proportion of respondents that value e-health in their
country suggests that more relevant content is available.
3.2.5 e-Entertainment usage % This is an indicator taken from the EIU “Value of the Internet” survey. EIU survey
The indicator looks at country-level responses to questions about how
often respondents use the Internet for entertainment purposes. A higher
proportion of respondents that use the Internet for entertainment in their
country suggests that more relevant content is available.
3.2.6 e-Commerce content Score of 0–100; 100 This indicator seeks to measure the availability—and extent—of electronic UNCTAD
= best commerce (e-commerce) in the country, which can serve both as a way to
buy products and to sell them. E-content refers to both electronic (online)
and mobile commerce. Greater availability of online services/e-commerce is
generally thought to increase Internet adoption.
3.2.7 Value of e-Commerce % This is an indicator taken from the EIU “Value of the Internet” survey. The EIU survey
indicator looks at country-level responses to questions about how often
respondents purchase goods via the Internet. A higher proportion of
respondents that use the Internet for purchasing goods in their country
suggests that more relevant content is available.
3.2.8 Open data policies Qualitative rating This indicator measures the existence of government “open data” policies EIU country
0–2; 2 = best that promote the dissemination of public-sector (public and publicly funded) research
data, as well as the existence of government open data platforms.
4 READINESS 0–100 Readiness is a measure of the capacity among Internet users to take
advantage of being online. The score for the readiness category is the
weighted sum of the following indicator scores: 4.1 to 4.3.
4.1 LITERACY 0–100 In order to find and use Internet content, users must have basic and digital
literacy. The literacy score is the weighted sum of the following indicators:
4.1.1 to 4.1.4.
4.1.1 Level of literacy % of population This indicator assesses the extent of literacy within countries. In order to use UNESCO
the Internet for useful purposes, such as to read news and access health or
educational information, people must be able to read. The higher the level of
literacy, the higher the capacity to take advantage of being online.
4.1.2 Educational attainment Years of schooling This indicator measures educational attainment through average years of Pew Research
schooling (ISCED 1 or higher). Internet adoption tends to be higher among Center, UNDP
highly educated groups. The greater the number of years of schooling, the
higher the capacity to take advantage of being online.
4.1.3 Support for digital Qualitative rating This measures whether the government has a plan or strategy that addresses EIU country
literacy 0–3; 3 = best digital literacy for students and training for teachers. Higher digital literacy research
increases the capacity of users to take advantage of being online.
4.1.4 Level of web accessibility Qualitative rating This measures whether the national government website passes W3C EIU country
0–4; 4 = best guidelines on web accessibility. If websites are not accessible to people with research
disabilities, there are fewer opportunities to use them.
4.2 TRUST & SAFETY 0–100 A secure and safe connection and higher cultural acceptance generally
increase the capacity to take advantage of being online. The trust and safety
score is the weighted sum of the following indicators: 4.2.1 to 4.2.6.
4.2.1 Privacy regulations Qualitative rating This measures whether the country has data protection law(s) and whether EIU country
0–2; 2 = best there are legal or financial penalties in place for firms that do not follow the research
law. Clear and transparent laws and financial penalties mean users can tell
what is legally acceptable within the country, which increases their capacity
to take advantage of being online.
16 © The Economist Intelligence Unit Limited 2020You can also read
