2021Soci

al
Progr
essI
ndex
Met
hodol
ogyRepor
t

ScottSter
n
Jar
omi rHarmacek
MohamedHt i
ti
ch
Petr
aKr yl
ova
Acknowledgements
Constructing the Social Progress Index is a significant research effort which involves
months of desk research, data collection, cleaning, transformations and
calculations. This would not be possible without the leadership of Michael Green
and Luke Greeves, alongside the Board of Directors, and support from all our
colleagues. In particular, we would like to thank Sophie Sutherland, Brent Nagel
and the entire communications team for their editorial support and ensuring a very
smooth communications and design process.

Suggested Citation

Index
Social Progress Imperative: 2021 Social Progress Index. Social Progress Imperative.
Washington, DC. Available at: www.socialprogress.org

Methodology
Stern, S., Harmacek, J., Htitich, M., & Krylova, P.: 2021 Social Progress Index
Methodology Summary. Social Progress Imperative. Washington, DC. Available at:
www.socialprogress.org/global/methodology

Contact
For technical queries regarding the index, please contact Jaromir Harmacek
(jharmacek@socialprogress.org). For any other queries, please contact Brent
Nagel (bnagel@socialprogress.org) or visit our website www.socialprogress.org.
 2021 Social Progress Index Methodology Report

Introduction ................................................................................................................................................. 2
Social Progress Principles ............................................................................................................................. 3
Dimensions of Social Progress...................................................................................................................... 4
Components of Social Progress .................................................................................................................... 5
Indicator Selection ....................................................................................................................................... 5
Indicator Transformations............................................................................................................................ 8
A. Capped Indicators ........................................................................................................................................... 8
B. Log-transformed Indicators ............................................................................................................................ 8
C. Calculation of parity ..................................................................................................................................... 10
Determining the Country Sample................................................................................................................. 9
Index Calculation ........................................................................................................................................ 10
A. Missing Values .............................................................................................................................................. 10
B. Standardization ............................................................................................................................................ 11
C. Component Scores ........................................................................................................................................ 12
D. Dimension Scores .......................................................................................................................................... 13
E. Index Scores .................................................................................................................................................. 13
F. World Score Calculation................................................................................................................................ 13
Assessing Countries’ Relative Strengths and Weaknesses ......................................................................... 15
Structural Integrity of the Social Progress Index ........................................................................................ 16
Year-to-Year Results Comparison............................................................................................................... 16
Limitations.................................................................................................................................................. 19
Conclusion .................................................................................................................................................. 20
Appendix A: Indicator Definitions and Sources .......................................................................................... 21
Appendix B: Indicator Boundaries .............................................................................................................. 33
Appendix C: PCA-Derived Indicator Weights.............................................................................................. 35
Appendix D: Descriptive Statistics for 2019 Social Progress Index, Component, and Dimension Scores .. 36
Appendix E: Bibliography and Further Reading.......................................................................................... 37

Introduction
The Social Progress Index is a well-established measure, published since 2013, that is meant to catalyze
improvement and drive action by presenting social outcome data in a useful and reliable way. Composed
of multiple dimensions, the Social Progress Index can be used to benchmark success and provide a holistic,
transparent, outcome-based measure of a country’s wellbeing that is independent of economic indicators.
Policymakers, businesses, and countries’ citizens alike can use it to compare their country against others
on different facets of social progress, allowing the identification of specific areas of strength or weakness.

The 2021 Social Progress Index ranks 168 countries on social progress. We combine 53 social and
environmental outcome indicators to calculate an overall score for these countries, based on tiered levels
of scoring that include measures in health, safety, education, technology, rights, and more. We also
consider the data of 36 additional countries, calculating component and dimension scores when enough
data are available. In all, the Social Progress Index measures at least some aspects of social progress across
more than 99.97% of the world’s population.

2 | socialprogress.org
 This report describes the methodology used to calculate the Social Progress Index. We start by describing
the principles that establish the conceptual architecture of the index and provide an overview of the index
framework. We then detail the steps taken to select data and calculate the index. Finally, we discuss the
methodology behind assessing countries’ strengths and weaknesses, relative to their economic
prosperity. We conclude the report with limitations of year-to-year comparisons and information on
future directions.

Social Progress Principles

We define ‘social progress’ as the capacity of a society to meet the basic human needs of its citizens,
establish the building blocks that allow citizens and communities to enhance and sustain the quality of
their lives, and create the conditions for all individuals to reach their full potential. This definition,
established in consultation with a group of academic and policy experts, drives the framework of the Social
Progress Index. It alludes to three broad elements of social progress, which we refer to as dimensions:
Basic Human Needs, Foundations of Wellbeing, and Opportunity. Under each dimension are four
components whose underlying concepts relate and are guided by questions we seek to answer with
available data (see Figure 1.) Each component is further defined by a set of outcome indicators that
respond to the conceptual questions posed.

Figure 1 / Social Progress Index Component-Level Framework

Together, these interrelated elements combine to produce a given level of social progress. The Social
Progress Index methodology allows measurement of each component and each dimension, yielding an
overall score and ranking.

Our approach builds on a long line of work constructing country indexes to measure and assess various
facets of economic and social performance. However, the Social Progress Index is distinct in its core
methodological choices:

• A focus on non-economic dimensions of national performance

• A measurement approach based on outcome indicators, rather than input measures
• A holistic framework consisting of three broad dimensions of social progress, each of which is the
sum of four equally weighted components

3 | socialprogress.org
 • Calculation of each component as the weighted sum of a series of measures, with the weights
determined through principal component analysis

The Social Progress Index is explicitly focused on non-economic aspects of national performance. Unlike
most other national measurement efforts, we treat social progress as distinct though associated with
more traditional economic measures such as GDP per capita. In contrast, other indices such as the Human
Development Index or OECD Better Life Index combine economic and social indicators. Our objective is to
utilize a clear yet rigorous methodology that isolates the non-economic dimensions of social performance.

The Social Progress Index aims to be as outcome-based as possible. Both input and outcome-based
indexes can help countries benchmark their progress, but in very different ways. Input indexes measure
a country’s policy choices or investments believed (or known) to lead to an important outcome, while
outcome indexes directly measure the outcomes of these decisions or investments. Input indexes also
require a degree of consensus about how inputs lead to outcomes, as well as a process to calibrate the
relative importance of different input factors against outcome measures. In the field of social progress,
this would mean a clear consensus and understanding of which inputs lead to better social outcomes—a
field of research that is still growing and to which the Social Progress Index continues to contribute.

When there are multiple output measures or a lack of consensus on all the inputs that matter, or when
data related to inputs are highly incomplete, an outcome-oriented index may be more appropriate
(Fleurbaey and Blanchet, 2013). Following this logic, we designed the Social Progress Index as an outcome
index. The Social Progress Index has been designed to aggregate and synthesize multiple outcome
measures in a conceptually consistent and transparent way that will also be useful for decision-makers
benchmarking progress. The Social Progress Imperative continues to explore the role of input measures
and policies in determining a country’s performance.

Dimensions of Social Progress

At the topmost level of the framework, we synthesize three distinct though related questions that, taken
together offer insight into the level of social progress:

1) Does a country provide for its people’s most essential needs?

2) Are the building blocks in place for individuals and communities to enhance and sustain
wellbeing?
3) Is there opportunity for all individuals to reach their full potential?

Each of these questions describes a dimension of social progress, respectively: Basic Human Needs,
Foundations of Wellbeing and Opportunity. The first dimension, Basic Human Needs, assesses a
population’s capacity to survive with adequate nourishment and basic medical care, clean water,
sanitation, adequate shelter, and personal safety. These needs are still not met in many developing
countries and are often incomplete in some more prosperous countries.

Basic needs have been the predominant focus of research in development economics, but the second
dimension of social progress, Foundations of Wellbeing, deserves equal attention. It highlights the extent
to which a country’s residents can gain a basic education, obtain information and communicate freely,
benefit from a modern healthcare system, and live in a healthy environment conducive to a long life.
Nearly all countries struggle with at least one of these aspects.

4 | socialprogress.org
 Finally, any discussion of social progress must also include whether a country’s population have the
freedom and opportunity to make their own choices and pursue higher education. Personal rights,
personal freedom and choice, inclusiveness, and access to advanced education all contribute to the level
of opportunity within a given society. This dimension of the Social Progress Index is perhaps the most
controversial and most difficult to measure. Nonetheless, it is important to highlight that societies, high-
income or low-income, developed or developing, still struggle to meet the moral imperative to guarantee
the equality of opportunity for all citizens.

The multi-dimensional construction of the Social Progress Index should not be interpreted as a step-by-
step movement toward progress from one dimension to the next. Rather, the three dimensions are
interrelated and, in fact, statistically correlated. While we distinguish between these three aspects of
social progress, many issues they encompass interact with one another to drive more meaningful change.

Components of Social Progress

Under each dimension are four components. Components, like dimensions, are categories of outcomes,
rather than specific outcomes themselves. Each component highlights a separate aspect of the overall set
of outcomes that make up a dimension, building on both academic and policy literature. For example, the
Opportunity dimension includes the components Personal Rights, Personal Freedom and Choice,
Inclusiveness, and Access to Advanced Education. Each of these components describes a related, but
distinct aspect of what it means for a society to guarantee opportunity among its population. The Personal
Rights and Access to Advanced Education components describe the extent to which individuals can pursue
their own objectives to the best of their ability. Personal Freedom and Choice and Inclusiveness, on the
other hand, describe the extent of limits on individuals. Together, the four components offer a
conceptually coherent way of capturing how societies can empower (or limit) an individual’s autonomy,
freedom, and ability to progress.

The twelve components represent what we believe to be the most complete set of outcome categories
given our current understanding of social progress from diverse literature and given the current
availability of data. The Social Progress Imperative Advisory Board provided input into selecting the
dimensions and the elaboration of the components within each dimension, along with an iterative review
of relevant literature.

The framework was established in 2013, and we continue to ensure its relevance each year of publication.
We consult extensively with experts across disciplines on the twelve-component structure of the Social
Progress Index on an ongoing basis, ensuring it continues to capture the principal aspects of human
wellbeing and that the issues measured are comprehensive and apply to all societies, regardless of their
country’s level of economic development, political stature, or geography.

Indicator Selection

At the most granular level of the Social Progress Index framework, we identify multiple independent
outcome measures – indicators – related to each component. Each set of indicators, grouped by
component, defines and measures the same aspect of social progress. Depending on data availability and
ongoing research into social outcomes, indicators may change with each edition of the Social Progress
Index. However, the concepts captured by each set of indicators (i.e. components) remains the same. The
2021 Social Progress Index includes 53 indicators, with 3-5 indicators per component (see Figure 2.)

5 | socialprogress.org
 Figure 2 / Social Progress Index Indicator-Level Framework

We only include indicators that are measured well, with consistent methodology, by the same
organization and across all (or essentially all) countries in our sample. We evaluate each indicator to
ensure that the procedures used to produce the measure are sound and that it captures what it purports
to capture. Data for each indicator must come from the same source to ensure consistency in
measurement across countries.

Data sources range from large international institutions like the United Nations to non-governmental
organizations such as Transparency International. We also include data collected via global surveys, such
as Gallup’s World Poll (sources are summarized in Appendix 1.) For each indicator, we evaluate the data
sources available and consider tradeoffs between the quality and precision of a social indicator and the
comprehensiveness of its country coverage. Figure 3 below depicts our decision tree for indicator
selection. Geographic coverage tends to exclude many high-quality indicators from consideration because
they only cover a subset of countries, such as OECD countries, or a particular region, such as Latin America.

6 | socialprogress.org
 Figure 3 / Indicator
Indicatorselection
Selection Treetree
decision

Included Indicators Eliminated Indicators

Does the indicator measure an economic,

social or environmental concept?

A social or environmental indicator An economic concept indicator

Does this indicator measure an input

or an outcome?

A concept that we are interested in because it Important mainly because it signals some -
is good or bad for its own sake thing else and is therefore an input indicator

What is the source of this indicator?

Widely reputable and the methods Unknown, uses biased methods,

it uses are sound or lacks rigorous data collection

How old are the data points?

Most data points are more than

Reasonably current 5-10 years old

How many geographic regions

does this indicator cover?

Fewer than 95% of the geographic

95-100% of geographic regions regions in the Index

SOCIAL PROGRESS INDEX

Additionally, we factor into our decision the age of the data, only considering the most recent available
data. Across the 168 countries we have a total of 8,700 raw (non-imputed) data points to calculate the
Social Progress Index for 2021. Most of the data are reflective of 2019 (49.9%) or 2020 (40.8%), 5.7% data
reflect year 2021. The least-recent data point comes from 2007 (dissatisfaction with affordability of
housing in Cuba).

A final important criterion for indicator data is that they are publicly available. We strive for transparency
both in terms of the data we use to inform the Social Progress Index, as well as our calculation

7 | socialprogress.org
 methodology. All the raw indicator data we use to calculate the Social Progress Index are published and
downloadable on our website at www.socialprogress.org.

Indicator Transformations
When comparing country-level data, we encounter issues that require us to transform the data for certain
indicators. In most cases, we transform data to meet clear upper or lower boundaries set by the indicator
definition. In others, we address extreme values that may skew results if left untreated. Our main two
techniques are to either cap an indicator, setting a clear upper or lower boundary cut-off value, or to log
an indicator. We also transform gender parity in secondary attainment to better reflect the parity
between boys and girls in a more gender-neutral fashion.

A. Capped Indicators

We impose a top and bottom boundary on six indicators, listed below in Figure 4. We set a floor at 0.03
for gender parity in secondary enrollment to allow for measurement error based on the recommendations
of UNESCO.1 The mobile telephone subscriptions indicator is capped at 100 subscriptions to reflect the
boundary set by its unit of measurement (number of subscriptions per 100 people). Child mortality rate
is capped at 197.35 deaths per 1,000 live births to limit the influence of a few significant outliers. The
political rights indicator is set to a floor of zero in line with the indicator’s definition. Similarly,
discrimination against minorities is set to a floor of one. Lastly, we cap years of tertiary schooling at five
years to avoid the influence of a few near-outliers on component-level performance.

Figure 4 / Capped Indicators

Indicators Cap
Child mortality rate 197.35
Gender parity in secondary enrollment 0.03 (bottom)
Mobile telephone subscriptions 100
Discrimination against minorities 1 (bottom)
Political rights 0 (bottom)
Years of tertiary schooling 5

B. Log-transformed Indicators

Five indicators––deaths attributable to unsafe water, sanitation and hygiene within Water and Sanitation,
deaths from interpersonal violence within Personal Safety, deaths from lead exposure within
Environmental Quality, citable documents, and quality weighted universities within Access to Advanced
Education––contain extreme values in relation to the rest of the indicator data distribution. Based on
external research, we determined that these extreme values are not erroneous and should be preserved
as a distinguishing characteristic of the countries they describe. As such, we transform these indicators
using natural log.2 Logging allows us to retain the unique differences between countries in performance
while creating a more sensible distribution that is less extreme.
1
UNESCO Institute for Statistics. “Global Education Digest 2010.” 2010, p. 17.
http://www.ungei.org/resources/files/GED_2010_EN.pdf
2
Prior to transformation, we add an alpha of 1 to deaths attributable to unsafe water, sanitation and hygiene,
to deaths from interpersonal violence and to deaths from lead exposure. We also add alpha of 0.01 to
quality weighted universities, and 0.0001 to citable documents. This ensures we can log all values within
the indicator, including zeros, while maintaining nearly the same relative differences between countries.

8 | socialprogress.org
 C. Calculation of parity

We transform gender parity in secondary attainment in Access to Basic Knowledge to reflect the absolute
distance from 1, where 1 represents an equal number of girls and boys enrolled. While in most countries,
more boys are enrolled in secondary education than girls, there are a select number of countries in which
the opposite is true. We therefore use the absolute distance from 1 to acknowledge the lack of parity for
both boys and girls across countries.

Determining the Country Sample

The 2021 Social Progress Index ranks 168 countries on social progress. We have selected these countries
by collecting all data available across all indicators and determining for which countries we can impute
data, and for which countries we will have incomplete information to calculate a Social Progress Index
score. Generally, a country cannot have more than one missing indicator per component to be included
in the final Social Progress Index score rankings. In some cases, we make exceptions to this rule,
particularly as it pertains to Access to Basic Knowledge and Access to Advanced Education, where data
are notoriously lacking. These exceptions are discussed in the next section.

Alongside the 168 ranked countries, we also include in our country sample 4 ‘partial’ countries. These
countries have enough data to calculate between nine to eleven of the twelve components, but not
enough data to calculate an overall Social Progress Index score. As with ranked countries, within those
nine to eleven components for which enough data are available there cannot be more than one indicator
missing per component.

Finally, we exclude from our original calculation sample an additional 59 countries or territories with
limited data, but we use the weights generated from PCA (described below) to calculate scores for these
countries when possible. These countries do not have enough data to calculate at least 9 components,
but 32 of them have enough data to calculate at least one component score. We do not impute missing
data for these countries during calculation, although we may do some imputations prior to calculation
(see below). Raw indicator data and scores for these 32 countries are included in the published dataset
on our website.

In addition, we exclude Venezuela and Afghanistan from all published documents and presentations due
to the uncertain situation in recent years or months which might not be fully reflected in all indicators and
thus might lead to misleading conclusions. Should anyone be interested in obtaining the countries’ results
and collected data, we would be happy to share these.

In this year’s edition, the 168 ranked countries include a full index score, ranks and relative performance
for the West Bank and Gaza. In order to do so, we implemented an approach different to other countries,
since some indicator sources provide data for the West Bank and Gaza while several others provide data
separately for the West Bank and for Gaza. In these cases, we calculated a population weighted average
to obtain one data point for the whole entity, which was then used in the overall index calculation.

9 | socialprogress.org
 Index Calculation

There are five core steps for calculating the Social Progress Index. We first address missing values, then
invert and standardize indicators so that they are comparable in scale. We then use Principal Component
Analysis (PCA) to aggregate indicators into a component score. Finally, we calculate dimension and overall
Social Progress Index scores by averaging components and dimensions, respectively. Each of these steps
is described in more detail below.

A. Missing Values

We ensure that all indicators included in the Social Progress Index are missing as few observations as
possible to avoid jeopardizing the statistical quality of the index. Missing values can stem from lack of
coverage by the data source, incomplete reporting by the country to international organizations, or
outdated data whose publication date is older than 2008. In cases where an indicator is missing a country
data point, we assess our imputation methodology both before and during index calculation. Imputations
used prior to calculation are included and marked in the published dataset on our website; imputations
generated during calculation are not.

Imputations prior to calculation:

We impute missing data prior to calculation under two scenarios: when a country lacks some, not all,
indicator data within the examined time period; and when there are gaps in the years of data for
indicators. These pre-calculation imputations are imperative to be able to include key countries in Social
Progress Index rankings. We mark and publish these values in our dataset available for download, as they
rely either on historical data from the same source or supplemental research.

In the first case, we carry back a future value for values used to calculate the Social Progress Indexes for
the years 2011-2020 in order to maintain a consistent sample. Similarly we carry forward a historical value
in those cases where historical data is available. In most cases we only carry forward or back a value for
the maximum of 5 consecutive years. In cases where more data points are missing, we rely on imputations
during calculations (see below).

Under the second scenario of pre-calculation imputations, we impute gaps between years by applying
linear interpolation. We do so to ensure smooth year-to-year estimates based on current and historical
data and by assuming linear change. In cases where there were data in the examined years, but not for
all years aligned with 2011 through 2021 Social Progress Indexes, we rely on data older than 2010 (if
available) to create linear estimations for the years in between. This is a necessary step in order to ensure
that our calculations of social progress over time do not exaggerate annual improvement or decline
merely due to gaps in the data points themselves.

Imputations during calculation:

After constructing the dataset with pre-calculation imputations as noted above, we assess the number of
indicators each country is missing within a component. Using regression imputation, we generally impute
data only for those countries for which there is no more than one missing data point per component in
each of the twelve components (considered ‘ranked countries’) and for countries that have no more than
one missing indicator data point in nine to eleven components (considered ‘partial countries’). We use

10 | socialprogress.org
 our country sample data of ranked and partial countries (including both current and historical Social
Progress Index years, i.e. 2011-2021) to regress each indicator on the other indicators within a
component. By constraining the regression to within-component indicators, we can preserve the signal
that the indicator provides to PCA.

In the past, we have strictly adhered to only one missing indicator per component and continue to stress
the importance of this aspect of our methodology. However, we allowed for an exception to this rule
particularly within the Access to Basic Knowledge component where data availability poses a significant
limitation. Therefore, for two indicators within this component we applied a pre-imputation regression
methodology: we used indicators not directly included in the index which had a more complete global
coverage and were highly correlated with the indicators we needed to predict. We used the Institute for
Health Metrics and Evaluation indicators education in years per capita (total, males, females) and UNDP
indicator mean years of schooling (total) to predict total, males’, and females’ secondary attainment for
approximately 20 countries with missing data. The latter two variables were then used to calculate the
educational parity indicator. The two pre-imputed indicators (secondary attainment and gender parity in
secondary attainment) were then used again in the standard regression imputations described above. In
a similar fashion, we used two indicators from the Access to Advanced Education component (women
with advanced education and citable documents) with the highest correlations to pre-impute expected
years of tertiary schooling for Iraq.

We review each imputation to ensure accuracy. In some cases, we combine the regression trend with
observed data. For example, when the last observed value for a country is in 2012, we have nine missing
values that we impute by regression predictions. If the predicted data do not match the observed values,
we take the regression trend from the predictions and apply it on the observed data. If there are no
observed values for a country, we apply standard regression imputations as described above. In cases
where these imputations do not match expectations or qualitative research, we use regional cohort
estimates or carry values consistently across time to minimize bias. For example, for many Middle Eastern
countries where Gallup does not ask its survey question on gays and lesbians due to cultural sensitivities,
we consider assessments of countries set by the Human Dignity Trust based on LGBT criminalization laws.3
If a country is not assessed by the survey and criminalization includes the death penalty, we assign the
country zero value for the indicator.

The estimation of missing values is necessary prior to undertaking PCA, which requires a complete dataset
for the results to be sound. We do not impute values for countries that do not meet the criteria of ranked
or partial countries noted above; these countries are excluded from the main calculation process by which
PCA weights are determined.

B. Standardization

We convert indicators to the same scale in a three-step process. First, we set best- and worst- case
scenarios to provide concrete boundaries on both ends of the scale that are based on theoretical or
historical values. We then invert indicators when increasing values reflect lower social progress. Finally,
we standardize the indicators into z-scores prior to applying PCA.

The best- and worst-case scenarios are defined at the indicator level. For some indicators, the lower and
upper boundaries are straightforward, such as perceived criminality, which comes from the Global Peace
Index and is already measured on a scale from 1 (low) to 5 (high). For indicators that do not have a clear

3
Map of countries that criminalize LGBT people can be found here:
https://www.humandignitytrust.org/lgbt-the-law/map-of-criminalisation/

11 | socialprogress.org
 best or worst case or where the probability of reaching a boundary is extremely unlikely (e.g., child
mortality, for which the theoretical worst case would be that every child dies before the age of five), we
use a boundary based on the worst recorded performance five years prior to the first year of
measurement (i.e. 2011 Social Progress Index). Best- and worst-case data values are included with the
country dataset when PCA is applied. See Appendix B for the specific values used for each indicator’s
bounds.

Once we establish a full dataset with indicator values for 2011 through 2021 and the best- and worst-case
scenario, we invert indicators for which a higher value denotes lower social progress. There are 21
inverted indicators in the 2021 Social Progress Index. These include: undernourishment, maternal
mortality rate, child mortality rate, child stunting, deaths from infectious diseases, unsafe water-,
sanitation- and hygiene-attributable deaths, household air pollution-attributable deaths, dissatisfaction
with housing affordability, deaths from interpersonal violence, perceived criminality, transport related
fatalities, women with no schooling, gender parity in secondary attainment, premature deaths from non-
communicable diseases, outdoor air pollution attributable deaths, deaths from lead exposure, PM 2.5,
vulnerable employment, early marriage, young people not in education, employment or training, and
discrimination and violence against minorities.

As a final step prior to applying PCA, we standardize the indicators into z-scores. Doing so produces scores
with a mean of 0 and standard deviation of 1, ensuring the comparability of the indicators across the
dataset in measurement.

C. Component Scores

To calculate component scores, we aggregate the set of indicators within each component into a factor
using PCA and all ten years of data.4 PCA combines indicators in a way that captures the maximum amount
of variance in the data while reducing redundancy between indicators. It essentially assigns each indicator
a weight, a method we select over equal weighting to ensure that indicators are meaningfully contributing
to a component score, while accounting for similarities between them.

Within many of the twelve components, PCA generates similar weights for the indicators we include
because we ensure a fair level of correlation between them (e.g., not too high or low a correlation) prior
to finalizing our framework. However, for those cases in which indicators are less correlated with other
indicators within their component, such biome protection and greenhouse gas emissions in
Environmental Quality (discussed further below), we consider PCA a good statistical approach for
determining these indicators’ contribution to the component scores while remaining objective.

The formula below reflects indicator aggregation into a principal component, where c=Social Progress
Index component and i=indicator.

Formula 1 !"#$"%&%' )*+,& = /(1! ∗ 3%43-*'"5! )

-
!

Our choice of PCA as the basis for aggregation at the component level was also influenced by the quality
and quantity of data available on social progress. For PCA to be valid, each indicator must be relatively
free of measurement error (Dunteman, 1989). Thus, it should precisely measure what it was intended to

4
Each statistical program has several ways to calculate PCA, leading to slight differences in estimations
depending on both the command and program used. We use the following command in Stata: factor
[standardized indicator names], factor(1) pcf

12 | socialprogress.org
 measure and do so consistently across countries. Our design principles and the data we use fulfill this
requirement.

To convert each principal component into a component score on a scale of 0 to 100, we use a simple
min-max formula, where X=component value and j=country.

Formula 2 (8" − :"57' !*7&)

!"#$"%&%' 7-"5& = ∗ 100
- (;&7' !*7& − :"57' !*7&)

As noted in the prior section, only countries that are ranked or qualify as ‘partial’ are included in the
country sample that determines PCA-generated weights. For countries that do not have enough data to
calculate at least nine components, we use the weights generated by the original country sample to
calculate component scores when possible. If a country outside the ranked and partial country sample has
enough data to calculate all four components within a dimension, we proceed to calculate dimension
scores as well.

D. Dimension Scores

Each dimension is the arithmetic average of the four components that make up that dimension. Countries
that do not have scores in all four components of a given dimension do not have a dimension score. The
formula for calculating a dimension score is below, where d=dimension and c=component.

Formula 3 1
>3#&%73"% = / !"#$"%&%' 7-"5&
4 4 -
#

E. Index Scores

The overall Social Progress Index score is calculated as the arithmetic average of the three dimensions.
Countries that do not have scores in all three dimensions do not have a Social Progress Index score. The
formula for calculating a Social Progress Index score is below, where d=dimension.

Formula 4 1
@"-3*+ A5"B5&77 C%4&D 7-"5& = / >3#&%73"%
3 4
$

We provide the mean, standard deviation, minimum, and maximum values of the calculated component,
dimension, and Social Progress Index scores in Appendix D. In establishing country rankings for overall
performance, we divide country scores into six tiers based on hierarchical clustering.

F. World Score Calculation

In order to provide the most accurate assessment of world performance on social progress, we account
for countries’ populations as well as the statistical interaction between indicators. Therefore, to calculate
the world Social Progress Index score, we first aggregate indicators into population-weighted values using
data of all ranked and partial countries. We then apply the PCA weights generated by the original ranked
and partial country sample to derive component scores and proceed as noted above to calculate
dimension and the overall Social Progress Index scores. It is important to note that this method is different
than calculating population-weighted scores, and in essence treats the world as a country.

13 | socialprogress.org
 Tiers

Based on hierarchical cluster analysis we calculate six distinct tiers. Each tier groups countries with broadly
similar performance on Social Progress Index. Cluster analysis is calculated for each year separately and
therefore the cut-offs of the six tiers are different.

Assessing Countries’ Relative Strengths and Weaknesses

The component, dimension, and overall Social Progress Index scores are scaled from 0 to 100 to provide
an intuitive scale for the interpretation of absolute performance, benchmarking a country against the best
and worst-possible scenarios in terms of social progress performance. However, it is also useful to
consider relative performance, comparing the level of social progress among countries of similar levels of
economic development. For example, a lower-income country may have a low score on a certain
component, but could greatly exceed typical scores for countries with similar GDP per capita incomes.
Conversely, a high-income country may have a high absolute score on a component, but still fall short of
what is typical for comparably wealthy countries. For this reason, we have developed a methodology to
present a country’s strengths and weaknesses on a relative basis, comparing a country’s performance to
that of its economic peers. Results of this analysis are the basis of our country scorecards, which can be
found on our website.

We define the group of a country’s economic peers as the 15 countries closest in GDP PPP per capita.
Standard groupings of countries, such as the World Bank’s country income classifications, are not
appropriate for relative comparison of countries for two reasons. First, the groupings are too large,
representing excessively wide ranges of social performance and therefore few relative strengths and
weaknesses. Second, using these groups, countries at the top or bottom of a group may appear to have a
misleadingly large number of strengths or weaknesses simply because the group the country is being
compared to is at a much lower or higher level of economic development.

Each country’s GDP per capita is compared to every other country for which there is full Index data, and
the 15 countries with the smallest difference on an absolute value basis are selected for the comparator
group. We have found that groupings larger than 15 resulted in a wider range of typical scores and showed
too few relative strengths and weakness, while smaller groupings become too sensitive to outliers.
Additionally, to reduce the influence of year-to-year fluctuations in GDP data, we use a four-year average
(2017-2020).

Once the group of comparator countries is established, the country’s performance is compared to the
median performance of countries in the group. The median is used rather than the mean to minimize the
influence of outliers. If the country’s score is greater than (or less than) the average absolute deviation
from the median of the comparator group, it is considered a strength (or weakness). Scores that are within
one average absolute deviation are within the range of expected scores and are considered neither
strengths nor weaknesses. A floor is established so the thresholds are no less than those for poorer
countries and the minimum distance from median to strength or median to weakness is 1 point.

We define comparator groups for all countries, regardless of whether they have complete Social Progress
Index data or sufficient data for only some indicators, components, and dimensions. However, to maintain
stability in comparisons, only countries with full data across all components of the index are included in
comparator groups for other countries. Among ranked and partial countries, we do not calculate strengths
and weaknesses for Cuba, Eritrea, North Korea, South Sudan, Syria and Yemen due to missing GDP data.

14 | socialprogress.org
 Structural Integrity of the Social Progress Index

Throughout the indicator assessment and calculation process, we conduct statistical tests to ensure the
structural integrity of the Social Progress Index. Our goal is that no single indicator majorly affects a
country’s component, dimension, or overall score, and that the indicators within each component are
statistically related and compatible. To achieve this, we look at correlations between indicators and
between indicators and aggregated scores, Cronbach’s alpha, and the Kaiser-Meyer-Olkin measure of
sampling adequacy.

In understanding the correlations between indicators, we strive for indicators within components to show
correlations of between r=0.25 to r=0.92 (absolute values). Indicators with correlations below 0.25
generally show little conceptual and statistical relation to other indicators. Likewise, if two indicators are
too highly correlated (i.e., r>0.92), we find that the indicators overlap too much in concept and become
statistically redundant, which would place too much weight on the concepts they are capturing within the
component; we generally remove one of these indicators as well. In the 2021 SPI framework, correlation
coefficients range from 0.17 to 0.89. However, all correlations are statistically significant at the 1% level.

To evaluate the fit between indicators within each component, we calculate Cronbach’s alpha after we
transform the indicators and impute missing values. Cronbach’s alpha provides a measure of internal
consistency across indicators. An applied practitioner’s rule of thumb is that the alpha value should be
above 0.7 for any valid grouping of variables (Bland and Altman, 1997). As shown in Figure 5, all twelve
components meet the 0.7 threshold (although it is very close for the Personal Safety component).

Figure 5 / Cronbach’s Alpha for Each Component

Cronbach’s
Alpha
Nutrition and Basic Medical Care 0.94
Water and Sanitation 0.93
Basic Human Needs
Shelter 0.83
Personal Safety 0.70
Access to Basic Knowledge 0.90
Foundations of Access to Information and Communications 0.78
Wellbeing Health and Wellness 0.91
Environmental Quality 0.75
Personal Rights 0.92
Personal Freedom and Choice 0.82
Opportunity
Inclusiveness 0.85
Access to Advanced Education 0.86

Cronbach’s alpha is a good preliminary screen for conceptual fit; however, it does not provide a direct
measure of the goodness of fit of a factor analysis (Manly, 2004.) Rather, we assess goodness of fit using
the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy. Generally, KMO scores should be above
0.5. In our data, the mean KMO score is above 0.5 for all components, suggesting that the grouping of
indicators chosen for the components of the Social Progress Index provides a good measure of the
underlying construct.

15 | socialprogress.org
 Figure 6 / KMO for Each Component
Mean KMO
Nutrition and Basic Medical Care 0.86
Water and Sanitation 0.75
Basic Human Needs
Shelter 0.74
Personal Safety 0.60
Access to Basic Knowledge 0.84
Foundations of Access to Information and Communications 0.71
Wellbeing Health and Wellness 0.73
Environmental Quality 0.74
Personal Rights 0.88
Personal Freedom and Choice 0.74
Opportunity
Inclusiveness 0.82
Access to Advanced Education 0.82

Year-to-Year Results Comparison

Each year we conduct a comprehensive review of all indicators included in the Social Progress Index
framework to check data updates (which frequently include retroactive revisions) and whether new
indicators have been published that are well-suited to describing social progress concepts. Many data
sources that we use revise their data collection or estimation methods, which impacts not just newly
published data but also previously published data. The Social Progress Index undergoes the same process
for the sake of comparability. Using the 2021 Social Progress Index framework and methodology, we
provide comparable historical data for ten additional years of the Social Progress Index, from 2011 to
2020. Results for the years 2011 to 2020 are therefore different from results that we have previously
published.

It is important to note that while we establish an eleven-year time-series of social progress from 2011 to
2021, not all indicator data are updated on an annual basis. Therefore, change over time is best
interpreted over the entire span of these eleven years rather than focusing on annual change.

The underlying framework (components and dimensions) of the Social Progress Index has remained the
same as 2020. However, we added several new indicators and removed a few due to their discontinuation
or the lack of updated data. We also changed the sources and the measurement of a handful of indicators.
Additionally, of the 53 indicators, majority were retroactively revised by the data sources. We list indicator
changes by component below.

Nutrition and Basic Medical Care: Data sources for undernourishment and child mortality retroactively
revised prior years’ estimations. Deaths for the infectious diseases indicator was updated using a fresh
source of data (IHME Global Burden of disease 2019 instead of 2017). Data source for child stunting was
changed to IHME Global Burden of Disease 2019 due to lack of updates of the original source (IHME
Health-related SDGs).

Water and Sanitation: Due to the lack of updates to the original source (IHME Health-related SDGs) for
all indicators that were used last year in this component, we changed the sources of the data for all
indicators and slightly changed the measure for two indicators. Instead of measuring populations using
unsafe or unimproved water sources, and populations using unsafe or unimproved sanitation, we use
prevalence of populations using safe or improved water and prevalence of populations using safe or

16 | socialprogress.org
 improved water. Both indicators come from IHME Global Burden of Disease 2019 Covariates. The third
indicator (Deaths attributable to unsafe water, sanitation and hygiene) now comes from IHME Global
Burden of Disease 2019.

Shelter: Data sources retroactively revised prior years’ estimations of access to electricity. The data source
for household air pollution attributable deaths was changed to IHME Global Burden of Disease 2019 due
to lack of updates of the original source (IHME Health-related SDGs). We added one additional indicator
in this component – dissatisfaction with affordability of housing from the Gallup World Poll.5

Personal Safety: Data sources for political killings and torture retroactively revised prior years’
estimations. For perceived criminality, we used data from the 2020 edition of the Global Peace Index. Due
to the lack of updates of the original sources we changed the source of transportation related fatalities
(from IHME Health-related SDGs to IHME Global Burden of Disease 2019) and we changed the indicator
of homicide rate (UN-ODC) to deaths from interpersonal violence (IHME Global Burden of Disease 2019).

Access to Basic Knowledge: Data sources for primary school enrollment, and access to quality education
retroactively revised their prior years’ estimations. We changed the source for women with no schooling
(from IHME Global Educational Attainment Distributions 1970-2030 to IHME Global Burden of Disease
2019 Covariates) since it works with more broadly defined age-group (now age-standardized, last year
only 25-29 age category).

Access to Information and Communications:

Data sources for all four indicators retroactively revised prior years’ estimations.

Health and Wellness: The V-Dem indicator on access to quality healthcare retroactively revised prior
years’ estimations. The life expectancy at 60 indicator now comes from IHME Global Burden of Disease
2019 instead of IHME Global Burden of Disease 2017. The data source for access to essential health
services was changed to IHME Global Burden of Disease 2019 Covariates due to lack of updates of the
original source (IHME Health-related SDGs).

Environmental Quality: We decided to take out the greenhouse gas emissions (GHGs) indicator from the
framework to be able to carry out independent analyses between SPI and GHGs. The results of these
analyses are presented in our related findings papers. We replaced the indicator on biome protection with
species protection––both come from the Environmental Performance Index. We introduced a new
indicator of deaths from lead exposure (IHME Global Burden of Disease 2019) and used the same source
with updated data on outdoor air pollution attributable deaths. Data source for particulate matter 2.5
was changed to IHME Global Burden of Disease 2019 Covariates due to lack of updates of the original
source (IHME Health-related SDGs).

Personal Rights: V-Dem, which is the data source for freedom of expression, freedom of religion, access
to justice, and property rights for women retroactively revised prior years’ estimations.

5
As a result of experimenting with various alignment procedures in early stages of the index creation, this
indicator has remained differently aligned to our dataset when compared to the alignment of other
indicators: the first available year for each country has been taken as the most recent one. While this has
not impacted the values for SPI year 2021, historical values (2011-2020) for this indicator may be shifted
over years for a few countries. This shift has not affected any calculations since the correlation between
this alignment and the standard alignment of data for this indicator is almost 0.98.

17 | socialprogress.org
 Personal Freedom and Choice: Data sources for vulnerable employment and satisfied demand for
contraception retroactively revised prior years’ estimations. We added one additional indicator in this
component – young people not in education, employment or training (NEET) that comes from
International Labour Organization (ILO).

Inclusiveness: V-Dem, which is the data source for equality of political power by gender, equality of
political power by socioeconomic position, and equality of political power by social group retroactively
revised prior years’ estimations.

Access to Advanced Education: We changed the source for women with advanced education (from IHME
Global Educational Attainment Distributions 1970-2030 to IHME Global Burden of Disease 2019
Covariates) since it works with more broadly defined age-group (now age-standardized, last year only 25-
29 age category). We continued with the revised measurement of globally ranked universities to take into
account other universities (measured by V-dem) which is reflected in quality weighted universities.
UNESCO indicator on tertiary school life expectancy retroactively revised prior years’ estimations. We
added one additional indicator in this component – academic freedom index that comes from the V-Dem.

Limitations

The Social Progress Index measures how countries at the national level perform on a certain set of
indicators that meet the standards and concepts represented by the Social Progress Index framework. It
is an important tool that is used to compare countries and assess both absolute and relative levels of
performance on social progress to find best practices and to target areas which need improvement or
from which other countries can learn. While the Social Progress Index framework captures the multi-
dimensional concepts underlying social progress, we are limited in how we measure these concepts by
the data available from public sources. Country performance is dependent upon the data published by
other sources, and we defer to these sources to respond to country inquiries about the different aspects
of social progress (a full list of sources is included in Appendix A).

We also recognize that the indicators in many of the topics we measure are not perfect. We strive to
ensure each indicator meets our standards of quality; however, some issues are much more complex than
the numbers we use to communicate them. For example, equality of political power by gender (in
Inclusiveness) must consider laws that are in place that require female representation in government, as
well as account for places where women might not necessarily have the voice they are supposedly
provided under these laws. We view these indicators as a starting point for measurement and
conversation, and we continue to refine the index each year to accommodate more recent data with
greater geographic coverage that cover important aspects of social progress still not captured by the
current indicators available, including violence against women, national environmental degradation, fresh
water withdrawals, and more.

Furthermore, the Social Progress Index provides a view into how a country performs on average, which
helps inform the many policies and investments that affect social progress at the national level. However,
it is only a starting point: aggregate data can obscure substantial regional and state differences in
performance that are equally important to a country’s policy considerations, especially in geographically
large regions. For this reason, we have established several initiatives across Latin America, Europe, South
Asia, and North America to explore social progress at a disaggregated regional level. We apply the same
Social Progress Index framework to more localized geographic regions, contextualizing indicators and
concepts with the input of local stakeholders. These initiatives help further drive action from the broader
issues highlighted in the global Social Progress Index.

18 | socialprogress.org
 Conclusion

The Social Progress Index provides a benchmark by which countries can compare themselves to others,
and can identify specific areas of current strength or weakness. Additionally, scoring on a 0–100 scale
gives countries a realistic benchmark rather than an abstract measure. This scale allows us to track
absolute, not just relative, performance of countries over time on each component, dimension, and the
overall model.

The 2021 Social Progress Index results are a starting point for many different avenues of research into the
ways a country is successful or not and whether conclusions can be drawn about the overall effect of
social progress on economic growth. Furthermore, while disaggregated scores provide insight into the
behavior of the different components that contribute to a country’s performance, we believe
disaggregation within a country (e.g. regional or state) also provides important insight and actionable
information to those seeking to increase social progress. We continue to test our process and
methodology at the regional and city level, replicating the steps outlined in this report to produce
meaningful results in different areas of the world.

19 | socialprogress.org
Appendix A: Indicator Definitions and Sources
All data used to calculate the 2021 Social Progress Index and relevant analyses are the most recent available as of September 1, 2021.

Component Indicator name Definition Source Link

BASIC HUMAN NEEDS
Nutrition and Deaths from Age-standardized mortality rate from deaths caused by Institute for Health Metrics http://ghdx.healthdata.org/gbd-results-
Basic Medical infectious diseases HIV/AIDS, tuberculosis, diarrhea, intestinal infections, and Evaluation tool
Care (deaths/100,000) respiratory infections, otitis media, meningitis,
encephalitis, diptheria, whooping cough, tetanus, measles,
varicella, herpes zoster, malaria, Chagas disease,
leishmaniasis, typanosomiasis, schistosomiasis,
cysticercosis, cycstic echinococcosis, lymphatic filariasis,
onchocerciasis, trachoma, dengue, yellow feber, rabies,
intestinal nematode infections, food-borne trematodiases,
leprosy, ebola, zika virus, guinea worm disease, sexually
transmitted diseases excluding HIV, hepatitis, and other
infectious diseases per 100,000 people.
Child mortality Probability of dying between birth and exactly 5 years of UN Inter-agency Group for http://www.childmortality.org
rate (deaths/1000 age, expressed per 1,000 live births. In the SPI model, Child Mortality Estimation
live births) values are capped at 197.35 deaths per 1,000 live births.

Child stunting Risk-weighted prevalence of stunting in children under 5 as Institute for Health Metrics http://ghdx.healthdata.org/gbd-results-
(0=low risk; measured by the summary exposure value (SEV) for child and Evaluation tool
100=high risk) stunting. Measured on a scale from 0% to 100%.
Maternal mortality Maternal deaths per 100,000 livebirths in women aged 10- Institute for Health Metrics http://ghdx.healthdata.org/record/ihm
rate 54 years. and Evaluation e-data/gbd-2017-health-related-sdgs-
(deaths/100,000 1990-2030
live births)
Undernourishment The prevalence of undernourishment expresses the Food and Agriculture http://www.fao.org/economic/ess/ess-
(% of population) probability that a randomly selected individual from the Organization of the United fs/ess-fadata/en/
population consumes an amount of calories that is Nations
insufficient to cover her/his energy requirement for an
active and healthy life. The indicator is computed by
comparing a probability distribution of habitual daily
dietary energy consumption with a threshold level called
the minimum dietary energy requirement. Both are based
on the notion of an average individual in the reference
population.

20 | socialprogress.org
Component Indicator name Definition Source Link
Water and Access to Proportion of population with access to improved toilet Institute for Health Metrics http://ghdx.healthdata.org/record/glob
Sanitation improved types as defined by the Joint Monitoring Program (JMP). and Evaluation al-burden-disease-study-2019-gbd-
sanitation 2019-covariates-1980-2019
(proportion of
population)
Access to Proportion of population with access to improved water Institute for Health Metrics http://ghdx.healthdata.org/record/glob
improved water sources as defined by the Joint Monitoring Program (JMP). and Evaluation al-burden-disease-study-2019-gbd-
source (proportion 2019-covariates-1980-2019
of population)
Unsafe water, Age-standardized death rate attributable to unsafe water, Institute for Health Metrics http://ghdx.healthdata.org/gbd-results-
sanitation and sanitation and hygiene (per 100,000 people). In the SPI and Evaluation tool
hygiene model, this indicator is logarithmically transformed.
attributable deaths
(deaths/100,000)
Shelter Household air Age-standardized deaths from household air pollution from Institute for Health Metrics http://ghdx.healthdata.org/gbd-results-
pollution solid fuels per 100,000 people. Household air pollution and Evaluation tool
attributable deaths includes exposure to particulate matter less than 2.5
(deaths/100,000) microns in diameter (PM2.5) due to the use of solid fuels
for cooking, including coal, charcoal, wood, agricultural
residue, and animal dung.
Dissatisfaction The percentage of respondents answering no to the Gallup World Poll https://ga.gallup.com/
with housing question, “In the city or area where you live, are you
affordability satisfied or dissatisfied with the availability of good,
(0=low; 1=high) affordable housing?”
Access to The percentage of the population with access to electricity. SE4ALL Global Tracking https://data.worldbank.org/indicator/E
electricity (% of Framework (World Bank, G.ELC.ACCS.ZS
population) International Energy
Agency, and the Energy
Sector Management
Assistance Program)
Usage of clean The proportion of population primarily using clean cooking World Health Organization https://apps.who.int/gho/data/node.m
fuels and fuels and technologies for cooking ain.SDGFUELS712?lang=en
technology for
cooking (% of
population)
Personal Deaths from Age-standardized deaths rate (per 100,000 people) from Institute for Health Metrics http://ghdx.healthdata.org/gbd-results-
Safety interpersonal interpersonal violence. Interpersonal violence is defined as and Evaluation tool
violence as death or disability from intentional use of physical force
(deaths/100,000) or power, threatened or actual, from another person or
group not including military or police forces. In the SPI
model, this indicator is logarithmically transformed.

21 | socialprogress.org
Component Indicator name Definition Source Link
Transportation Age-standardized rate of deaths per 100,000 people due to Institute for Health Metrics http://ghdx.healthdata.org/gbd-results-
related fatalities injuries related to transportation. These injuries include and Evaluation tool
(deaths/100,000) road injuries (death or disability due to unintentional
interaction with an automobile, motorcycle, pedal cycle, or
other vehicles) as well as other transport injuries.

Perceived An assessment of the level of domestic security and the Institute for Economics and http://visionofhumanity.org/indexes/gl
criminality (1=low; degree to which other citizens can be trusted. Measured on Peace Global Peace Index obal-peace-index/
5=high) a scale of 1 (majority of other citizens can be trusted; very
low levels of domestic security) to 5 (very high level of
distrust; people are extremely cautious in their dealings
with others; large number of gated communities, high
prevalence of security guards). Data comes from the 2020
edition of the Global Peace Index.
Political killings Physical violence index scaled 0 to 1 that is based on Varieties of Democracy (V- https://www.v-
and torture (0=low indicators that reflect violence committed by government Dem) Project dem.net/en/data/data/v-dem-dataset-
freedom; 1=high agents and that are not directly referring to elections. v111/
freedom) Variable name: v2x_clphy

FOUNDATIONS OF WELLBEING
Access to Women with no Proportion of females (age-standardized) with no Institute for Health Metrics http://ghdx.healthdata.org/record/glob
Basic schooling schooling. and Evaluation al-burden-disease-study-2019-gbd-
Knowledge (proportion of 2019-covariates-1980-2019
females)
Equal access to Country experts' aggregated evaluation of the question, Varieties of Democracy (V- https://www.v-
quality education "To what extent is high quality basic education guaranteed Dem) Project dem.net/en/data/data/v-dem-dataset-
(0=unequal; to all, sufficient to enable them to exercise their basic v111/
4=equal) rights as adult citizens?" measured on a scale of 0 to 4.
0: Extreme. Provision of high quality basic education is
extremely unequal and at least 75 percent (%) of children
receive such low-quality education that undermines their
ability to exercise their basic rights as adult citizens.
1: Unequal. Provision of high quality basic education is
extremely unequal and at least 25 percent (%) of children
receive such low-quality education that undermines their
ability to exercise their basic rights as adult citizens.
2: Somewhat equal. Basic education is relatively equal in
quality but ten to 25 percent (%) of children receive such
low-quality education that undermines their ability to
exercise their basic rights as adult citizens.
3: Relatively equal. Basic education is overall equal in
quality but five to ten percent (%) of children receive such

22 | socialprogress.org
Component Indicator name Definition Source Link
low-quality education that probably undermines their
ability to exercise their basic rights as adult citizens.
4: Equal. Basic education is equal in quality and less than
five percent (%) of children receive such low-quality
education that probably undermines their ability to
exercise their basic rights as adult citizens.
Variable name: v2peedueq_osp
Primary school Total number of students of official primary school age who UN Educational, Scientific, http://data.uis.unesco.org/
enrollment (% of are enrolled in any level of education, expressed as a and Cultural Organization
children) percentage of the total population of official primary Institute for Statistics
school age. Statistic is termed 'total net primary enrollment
rate.'
Secondary school Population with at least some secondary education (% ages United Nations http://hdr.undp.org/en/data
attainment (% of 25 and older) Development Programme
population aged (UNDP) Human
25+) Development Data
Gender parity in The absolute deviation from parity (=1) in secondary United Nations http://hdr.undp.org/en/data
secondary education attainment of women and men. In the SPI Development Programme
attainment model, an absolute deviation from parity lower than 0.03 is (UNDP) Human
(distance from treated as parity. Development Data
parity)
Access to Access to online The availability of e-participation tools on national UN Department of https://publicadministration.un.org/ego
Information and governance government portal for of the following uses: e-information Economic and Social Affairs vkb/en-us/Data-Center
Communications
(0=low; 1=high) – provision of information on the Internet; e-consultation – E-Government Survey
organizing public consultations online; and e-decision-
making – involving citizens directly in decision processes. E-
participation is defined as the process of engaging citizens
through ICTs in policy, decision-making, and service design
and delivery in order to make it participatory, inclusive, and
deliberative.
Internet users (% The estimated number of Internet users out of the total International http://www.itu.int/en/ITU-
of population) population, using the Internet from any device (including Telecommunications Union D/Statistics/Pages/stat/default.aspx
mobile phones) in the last 12 months.
Media censorship Country experts' aggregated evaluation of the question, Varieties of Democracy (V- https://www.v-
(0=frequent; "Does the government directly or indirectly attempt to Dem) Project dem.net/en/data/data/v-dem-dataset-
4=rare) censor the print or broadcast v111/
media?" measured on a scale of 0 to 4.
0: Attempts to censor are direct and routine.
1: Attempts to censor are indirect but nevertheless routine.
2: Attempts to censor are direct but limited to especially
sensitive issues.

23 | socialprogress.org
Component Indicator name Definition Source Link
3: Attempts to censor are indirect and limited to especially
sensitive issues.
4: The government rarely attempts to censor major media
in any way, and when such exceptional
attempts are discovered, the responsible officials are
usually punished.
Variable name: v2mecenefm_osp
Mobile telephone Subscriptions to a public mobile telephone service using International http://www.itu.int/en/ITU-
subscriptions cellular technology, including the number of pre-paid SIM Telecommunications Union D/Statistics/Pages/stat/default.aspx
(subscriptions/100 cards active during the past three months, expressed as the
people) number of mobile telephone subscriptions per 100
inhabitants. In the SPI model, values are capped at 100
mobile telephones per 100 people.
Health and Life expectancy at The average number of years that a person of 60 to 64 Institute for Health Metrics http://ghdx.healthdata.org/gbd-results-
Wellness 60 (years) years old could expect to live, if he or she were to pass and Evaluation tool
through life exposed to the sex- and age-specific death
rates prevailing at the time of his or her 60 years, for a
specific year, in a given country, territory, or geographic
area.
Premature deaths Mortality rate due to cardiovascular diseases, cancers, Institute for Health Metrics http://ghdx.healthdata.org/record/ihm
from non- diabetes, and chronic respiratory diseases among and Evaluation e-data/gbd-2017-health-related-sdgs-
communicable populations aged 30–70 years. 1990-2030
diseases
(deaths/100,000)
Equal access to Country experts' aggregated evaluation of the question, Varieties of Democracy (V- https://www.v-
quality healthcare "To what extent is high quality basic healthcare guaranteed Dem) Project dem.net/en/data/data/v-dem-dataset-
(0=unequal; to all, sufficient to enable them to exercise their basic v111/
4=equal) political rights as adult citizens?" measured on a scale of 0
to 4.
0: Extreme. Because of poor-quality healthcare, at least 75
percent (%) of
citizens’ ability to exercise their political rights as adult
citizens is undermined.
1: Unequal. Because of poor-quality healthcare, at least 25
percent (%) of citizens’ ability to exercise their political
rights as adult citizens is undermined.
2: Somewhat equal. Because of poor-quality healthcare,
ten to 25 percent (%) of citizens’ ability to exercise their
political rights as adult citizens is undermined.
3: Relatively equal. Basic health care is overall equal in
quality but because of poor-quality healthcare, five to ten

24 | socialprogress.org
Component Indicator name Definition Source Link
percent (%) of citizens’ ability to exercise their political
rights as adult citizens is undermined.
4: Equal. Basic health care is equal in quality and less than
five percent (%) of citizens cannot exercise their basic
political rights as adult citizens.
Variable name: v2pehealth_osp
Access to essential The universal health coverage (UHC) measures the Institute for Health Metrics http://ghdx.healthdata.org/record/glob
health services coverage of 9 tracer interventions and risk-standardized and Evaluation al-burden-disease-study-2019-gbd-
(0=none; 100=full death rates from 32 causes amenable to personal 2019-covariates-1980-2019
coverage) healthcare, including vaccine-preventable diseases (e.g.,
diphteria, tetanus, measles), respitory infections, cancer
(breast, cervical, uterine, testicular), heart diseases,
diabetes, kidney disease, and the adverse effects of
medical treatment.
Environmental Outdoor air The number of deaths resulting from ambient particulate Institute for Health Metrics http://ghdx.healthdata.org/gbd-results-
Quality pollution matter pollution, including emissions from industrial and Evaluation tool
attributable deaths activity, households, cars and trucks, expressed as the rate
(deaths/100,000) per 100,000 people, age adjusted.
Deaths from lead Age-standardized death rate attributable lead exposure Institute for Health Metrics http://ghdx.healthdata.org/gbd-results-
exposure (per 100,000 pop.). Lead exposure is defined as acute and Evaluation tool
(deaths/100,000) exposure, measured by micrograms of lead per decilitre of
blood, and chronic exposure, measured by micrograms of
lead per gram of bone. In the SPI model, this indicator is
logarithmically transformed.
Particulate matter Population-weighted mean levels of annual exposure to Institute for Health Metrics http://ghdx.healthdata.org/record/glob
pollution (mean suspended particles smaller than 2.5 microns in and Evaluation al-burden-disease-study-2019-gbd-
annual exposure, aerodynamic diameter (PM2.5), which are capable of 2019-covariates-1980-2019
µg/m3) penetrating deep into the respiratory tract and causing
severe health damage.
Species protection An index of how well a country's terrestrial protected areas Environmental https://epi.yale.edu/epi-
(0=low;100=high) overlap with the ranges of its vertebrate, invertebrate, and Performance Index results/2020/component/epi
plant species. The Species Protection Index is calculated
using remote sensing data, global biodiversity informatics, Map of Life https://mol.org/indicators/
and integrative models to map suitable habitat for over
30,000 terrestrial species at high resolutions. A score of
100 indicates full coverage of all species' ranges by a
country's protected areas, and a score of 0 indicates no
overlap.

25 | socialprogress.org
Component Indicator name Definition Source Link
OPPORTUNITY
Personal Access to justice Country experts' aggregated evaluation of the question, Varieties of Democracy (V- https://www.v-
Rights (0=non-existent; "Do citizens enjoy secure and effective access to justice?" Dem) Project dem.net/en/data/data/v-dem-dataset-
1=observed) Responses are collected on an ordinal scale, and then v111/
converted to a 0-1 scale. 0 signifies secure and effective
access to justice is non-existent, and 1 signifies secure and
effective access to justice is almost always observed.
Variable name: v2xcl_acjst
Freedom of Country experts' aggregated evaluation of the question, Varieties of Democracy (V- https://www.v-
expression (0=no "To what extent does government respect press & media Dem) Project dem.net/en/data/data/v-dem-dataset-
freedom; 1=full freedom, the freedom of ordinary people to discuss v111/
freedom) political matters at home and in the public sphere, as well
as the freedom of academic and cultural expression?"
Variable name: v2x_freexp
Freedom of Country experts' aggregated evaluation of the question, "Is Varieties of Democracy (V- https://www.v-
religion (0=no there freedom of religion?" measured on a scale of 0 to 4. Dem) Project dem.net/en/data/data/v-dem-dataset-
freedom; 4=full 0: Not respected by public authorities. Hardly any freedom v111/
freedom) of religion exists. Any kind of religious practice is outlawed
or at least controlled by the government to the extent that
religious leaders are appointed by and subjected to public
authorities, who control the activities of religious
communities in some detail.
1: Weakly respected by public authorities. Some elements
of autonomous organized religious practices exist and are
officially recognized. But significant religious communities
are repressed, prohibited, or systematically disabled,
voluntary conversions are restricted, and instances of
discrimination or intimidation of individuals or groups due
to their religion are common.
2: Somewhat respected by public authorities. Autonomous
organized religious practices exist and are officially
recognized. Yet, minor religious communities are
repressed, prohibited, or systematically disabled, and/or
instances of discrimination or intimidation of individuals or
groups due to their religion occur occasionally.
3: Mostly respected by public authorities. There are minor
restrictions on the freedom of religion, predominantly
limited to a few isolated cases. Minority religions face
denial of registration, hindrance of foreign missionaries
from entering the country, restrictions against
proselytizing, or hindrance to access to or construction of
places of worship.
26 | socialprogress.org
Component Indicator name Definition Source Link
4: Fully respected by public authorities. The population
enjoys the right to practice any religious belief they choose.
Religious groups may organize, select, and train personnel;
solicit and receive contributions; publish; and engage in
consultations without undue interference. If religious
communities have to register, public authorities do not
abuse the process to discriminate against a religion and do
not constrain the right to worship before registration.
Variable name: v2clrelig_osp
Political rights (0 An evaluation of three subcategories of political rights: Freedom House https://freedomhouse.org/report-
and lower=no electoral process, political pluralism and participation, and types/freedom-world
rights; 40=full functioning of government on a scale from 0 (no political
rights) rights) to 40 (full political rights). Some countries and
territories score below zero on the questions used to
compose the indicator. In the SPI model, data below zero
are treated as zero.
Property rights for Country experts' aggregated evaluation of the question, Varieties of Democracy (V- https://www.v-
women (0=no "Do women enjoy the right to private property?" measured Dem) Project dem.net/en/data/data/v-dem-dataset-
rights; 5= full on a scale of 0 to 5. v111/
rights) 0: Virtually no women enjoy private property rights of any
kind.
1: Some women enjoy some private property rights, but
most have none.
2: Many women enjoy many private property rights, but a
small proportion enjoys few or none.
3: More than half of women enjoy most private property
rights, yet a smaller share of women have much more
restricted rights.
4: Most women enjoy most private property rights but a
small minority does not.
5: Virtually all women enjoy all, or almost all, property
rights.
Variable name: v2clprptyw_osp
Personal Satisfied demand The percentage of total demand for family planning among United Nations Population http://www.un.org/en/development/de
Freedom and for contraception married or in-union women aged 15 to 49 that is satisfied Division sa/population/theme/family-
Choice (% of satisfied with modern methods. planning/cp_model.shtml
demand)
Perception of The perceived level of public sector corruption based on Transparency International www.transparency.org/cpi
corruption (0=high expert opinion, measured on a scale from 0 (highly corrupt)
corruption; to 100 (very clean).

27 | socialprogress.org
Component Indicator name Definition Source Link
100=low
corruption)
Early marriage (% The percentage of women aged 15-19 years who are United Nations Population https://www.un.org/en/development/d
of married women married or in-union. Division esa/population/theme/marriage-
aged 15-19) unions/marriage_estimates.asp
Young people not The proportion of youth who are not in employment and International Labor https://ilostat.ilo.org/data/
in education, not in education or training. Youth are defined as persons Organization
employment or between the ages of 15 and 24 years. The series is part of
training (% of the ILO modelled estimates.
youth)
Vulnerable Contributing family workers and own-account workers as a International Labor https://data.worldbank.org/indicator/S
employment (% of percentage of total employment. Organization/World Bank L.EMP.VULN.ZS
total employment)
Inclusiveness Equality of political Country experts' aggregated evaluation of the question, "Is Varieties of Democracy (V- https://www.v-
power by gender political power distributed according to gender?" measured Dem) Project dem.net/en/data/data/v-dem-dataset-
(0=unequal power; on a scale of 0 to 4. v111/
4=equal power) 0: Men have a near-monopoly on political power.
1: Men have a dominant hold on political power. Women
have only marginal influence.
2: Men have much more political power but women have
some areas of influence.
3: Men have somewhat more political power than women.
4: Men and women have roughly equal political power.
Variable name: v2pepwrgen_osp
Equality of political Country experts' aggregated evaluation of the question, "Is Varieties of Democracy (V- https://www.v-
power by social political power distributed according to social groups Dem) Project dem.net/en/data/data/v-dem-dataset-
group (0=unequal (defined by caste, ethnicity, language, race, religion or v111/
power; 4=equal some combination thereof)?" measured on a scale of 0 to
power) 4.
0: Political power is monopolized by one social group
comprising a minority of the population. This monopoly is
institutionalized, i.e., not subject to frequent change.
1: Political power is monopolized by several social groups
comprising a minority of the population. This monopoly is
institutionalized, i.e., not subject to frequent change.
2: Political power is monopolized by several social groups
comprising a majority of the population. This monopoly is
institutionalized, i.e., not subject to frequent change.
3: Either all social groups possess some political power,
with some groups having more power than others; or
different social groups alternate in power, with one group

28 | socialprogress.org
Component Indicator name Definition Source Link
controlling much of the political power for a period of time,
followed by another – but all significant groups have a turn
at the seat of power.
4: All social groups have roughly equal political power or
there are no strong ethnic, caste, linguistic, racial, religious,
or regional differences to speak of. Social group
characteristics are not relevant to politics.
Variable name: v2pepwrsoc_osp
Equality of political Country experts' aggregated evaluation of the question, "Is Varieties of Democracy (V- https://www.v-
power by political power distributed according to socioeconomic Dem) Project dem.net/en/data/data/v-dem-dataset-
socioeconomic position?" measured on a scale of 0 to 4. v111/
position 0: Wealthy people enjoy a virtual monopoly on political
(0=unequal power; power. Average and poorer people have almost no
4=equal power) influence.
1: Wealthy people enjoy a dominant hold on political
power. People of average income have little say. Poorer
people have essentially no influence.
2: Wealthy people have a very strong hold on political
power. People of average or poorer income have some
degree of influence but only on issues that matter less for
wealthy people.
3: Wealthy people have more political power than others.
But people of average income have almost as much
influence and poor people also have a significant degree of
political power.
4: Wealthy people have no more political power than those
whose economic status is average or poor. Political power
is more or less equally distributed across economic groups.
Variable name: v2pepwrses_osp
Discrimination and Group Grievance indicator. Discrimination, powerlessness, Fund for Peace Fragile https://fragilestatesindex.org/
violence against ethnic violence, communal violence, sectarian violence, States Index
minorities (0=low; and religious violence, measured on a scale on 0 (low
10=high) pressures) to 10 (very high pressures). In the SPI model,
data below one are treated as one.
Acceptance of gays The percentage of respondents answering yes to the Gallup World Poll https://ga.gallup.com/
and lesbians question, “Is the city or area where you live a good place or
(0=low; 1=high) not a good place to live for gay or lesbian people?”
Access to Citable documents Citable documents - articles, reviews and conference Scimago Journal & Country https://www.scimagojr.com/countryran
Advanced (documents/1000 papers - per 1,000 population. In the SPI model, this Rank k.php
Education people) indicator is logarithmically transformed.

29 | socialprogress.org
Component Indicator name Definition Source Link
Academic freedom Aggregated evaluation of the question, "To what extent is Varieties of Democracy (V- https://www.v-
(0=low; 1=high) academic freedom respected?", measured on a scale of 0 Dem) Project dem.net/en/data/data/v-dem-
to 1. The index is an aggregated measure capturing dataset-v111/
freedom to research and teach, freedom of academic
exchange and dissemination, institutional autonomy,
campus integrity, and freedom of academic and cultural
expression.
Women with Proportion of females (age-standardized) with 12–18 years Institute for Health Metrics http://ghdx.healthdata.org/record/glob
advanced of education. and Evaluation al-burden-disease-study-2019-gbd-
education 2019-covariates-1980-2019
(proportion of
females)
Expected years of Number of years a person of tertiary school entrance age UN Educational, Scientific, https://data.unodc.org/
tertiary schooling can expect to spend within tertiary education. For a child of and Cultural Organization
(years) a certain age a, the school life expectancy is calculated as Institute for Statistics
the sum of the age specific enrolment rates for the levels of
education specified. The part of the enrolment that is not
distributed by age is divided by the school-age population
for the level of education they are enrolled in, and
multiplied by the duration of that level of education. The
result is then added to the sum of the age-specific
enrolment rates. The indicator seeks to show the overall
level of development of an educational system in terms of
the average number of years of schooling that the
education system offers to the eligible population,
including those who never enter school. In the SPI model,
values are capped at 5 years of expected tertiary schooling.
Quality weighted The number of universities in a country weighted by the Times Higher Education https://www.timeshighereducation.co
universities quality of universities, measured by university rankings on World University Rankings, m/world-university-
(points) any of the three most widely used international QS World University rankings/2020/world-ranking
assessments. Universities in the top 400 on any list are Rankings, and Academic https://www.topuniversities.com/unive
given double weight. Not ranked universities are given 5% Ranking of World rsity-rankings/world-university-
weight of the top ranked universities. In the SPI model, this Universities; Varieties of rankings/2021
indicator is logarithmically transformed. Variable name Democracy (V-Dem) Project http://www.shanghairanking.com/ARW
from the V-Dem dataset: v2canuni. ; SPI calculations U2019.html
https://www.v-
dem.net/en/data/data/v-dem-dataset-
v111/

30 | socialprogress.org
Component Indicator name Definition Source Link

GDP per capita, GDP per capita based on purchasing power parity (PPP). World Bank http://data.worldbank.org/indicator/NY
PPP (constant PPP GDP is gross domestic product converted to .GDP.PCAP.PP.KD
2017 international dollars using purchasing power parity rates.
international $) An international dollar has the same purchasing power
over GDP as the U.S. dollar has in the United States. GDP at
purchaser's prices is the sum of gross value added by all
resident producers in the economy plus any product taxes
and minus any subsidies not included in the value of the
products. It is calculated without making deductions for
depreciation of fabricated assets or for depletion and
degradation of natural resources. Data are in constant
2017 international dollars.

31 | socialprogress.org
 Appendix B: Indicator Boundaries
Indicator Best case Worst case
Child mortality rate (deaths/1000 live births) 0 197.35
Undernourishment (% of population) 2.5 55.7
Child stunting (0=low risk; 100=high risk) 0 40.36
Maternal mortality rate (deaths/100,000 live births) 0 742.34
Deaths from infectious diseases (deaths/100,000) 0 1732.55
Unsafe water, sanitation and hygiene attributable deaths (deaths/100,000) 0 295.62
Access to improved water source (proportion of population) 1 0.346
Access to improved sanitation (proportion of population) 1 0
Household air pollution attributable deaths (deaths/100,000) 0 438.34
Usage of clean fuels and technology for cooking (% of population) 100 0
Access to electricity (% of population) 100 0
Dissatisfaction with housing affordability (0=low; 1=high) 0 1
Perceived criminality (1=low; 5=high) 1 5
Transportation related fatalities (deaths/100,000) 0 98.38
Deaths from interpersonal violence (deaths/100,000) 0 103.606
Political killings and torture (0=low freedom; 1=high freedom) 1 0
Women with no schooling (proportion of females) 0 0.914
Primary school enrollment (% of children) 100 36.865
Equal access to quality education (0=unequal; 4=equal) 4 0
Secondary school attainment (% of population aged 25+) 100 3.2
Gender parity in secondary attainment (distance from parity) 0.03 0.838
Access to online governance (0=low; 1=high) 1 0
Internet users (% of population) 100 0
Media censorship (0=frequent; 4=rare) 4 0
Mobile telephone subscriptions (subscriptions/100 people) 100 0
Access to essential health services (0=none; 100=full coverage) 100 16.334
Life expectancy at 60 (years) 28.36 11.409
Premature deaths from non-communicable diseases (deaths/100,000) 0 1503.93
Equal access to quality healthcare (0=unequal; 4=equal) 4 0
Particulate matter pollution (mean annual exposure, µg/m3) 0 98.223
Outdoor air pollution attributable deaths (deaths/100,000) 0 200.294
Species protection (0=low;100=high) 100 0
Deaths from lead exposure (deaths/100,000) 0 98.128
Access to justice (0=non-existent; 1=observed) 1 0
Freedom of expression (0=no freedom; 1=full freedom) 1 0
Freedom of religion (0=no freedom; 4=full freedom) 4 0
Property rights for women (0=no rights; 5= full rights) 5 0
Political rights (0 and lower=no rights; 40=full rights) 40 0
Vulnerable employment (% of total employment) 0 94.4

32 | socialprogress.org
 Young people not in education, employment or training (% of youth) 0 48.993
Early marriage (% of married women aged 15-19) 0 62.707
Satisfied demand for contraception (% of satisfied demand) 100 4.7
Perception of corruption (0=high corruption; 100=low corruption) 100 0
Equality of political power by gender (0=unequal power; 4=equal power) 4 0
Acceptance of gays and lesbians (0=low; 1=high) 1 0
Equality of political power by social group (0=unequal power; 4=equal power) 4 0
Equality of political power by socioeconomic position (0=unequal power;
4=equal power) 4 0
Discrimination and violence against minorities (0=low; 10=high) 1 10
Citable documents (documents/1000 people) 5.954 0
Women with advanced education (proportion of females) 0.95 0
Expected years of tertiary schooling (years) 5 0
Quality weighted universities (points) 1043.4 0
Academic freedom (0=low; 1=high) 1 0

33 | socialprogress.org
 Appendix C: PCA-Derived Indicator Weights
Indicator Unscaled Scaled
Deaths from infectious diseases (deaths/100,000) 0.22 0.20
Child mortality rate (deaths/1000 live births) 0.23 0.21
Child stunting (0=low risk; 100=high risk) 0.22 0.20
Maternal mortality rate (deaths/100,000 live births) 0.23 0.21
Undernourishment (% of population) 0.21 0.19
Access to improved sanitation (proportion of population) 0.36 0.34
Access to improved water source (proportion of population) 0.35 0.33
Unsafe water, sanitation and hygiene attributable deaths (deaths/100,000) 0.36 0.33
Household air pollution attributable deaths (deaths/100,000) 0.33 0.29
Dissatisfaction with housing affordability (0=low; 1=high) 0.13 0.11
Access to electricity (% of population) 0.34 0.30
Usage of clean fuels and technology for cooking (% of population) 0.33 0.29
Deaths from interpersonal violence (deaths/100,000) 0.32 0.24
Transportation related fatalities (deaths/100,000) 0.31 0.23
Perceived criminality (1=low; 5=high) 0.38 0.28
Political killings and torture (0=low freedom; 1=high freedom) 0.35 0.26
Women with no schooling (proportion of females) 0.25 0.22
Equal access to quality education (0=unequal; 4=equal) 0.21 0.18
Primary school enrollment (% of children) 0.21 0.18
Secondary school attainment (% of population aged 25+) 0.25 0.21
Gender parity in secondary attainment (distance from parity) 0.24 0.21
Access to online governance (0=low; 1=high) 0.35 0.28
Internet users (% of population) 0.37 0.30
Media censorship (0=frequent; 4=rare) 0.20 0.16
Mobile telephone subscriptions (subscriptions/100 people) 0.33 0.26
Life expectancy at 60 (years) 0.30 0.26
Premature deaths from non-communicable diseases (deaths/100,000) 0.27 0.24
Equal access to quality healthcare (0=unequal; 4=equal) 0.27 0.24
Access to essential health services (0=none; 100=full coverage) 0.30 0.27
Outdoor air pollution attributable deaths (deaths/100,000) 0.35 0.27
Deaths from lead exposure (deaths/100,000) 0.34 0.26
Particulate matter pollution (mean annual exposure, µg/m3) 0.35 0.27
Species protection (0=low;100=high) 0.25 0.19
Access to justice (0=non-existent; 1=observed) 0.23 0.20
Freedom of expression (0=no freedom; 1=full freedom) 0.24 0.21
Freedom of religion (0=no freedom; 4=full freedom) 0.21 0.18
Political rights (0 and lower=no rights; 40=full rights) 0.24 0.21
Property rights for women (0=no rights; 5= full rights) 0.22 0.19

34 | socialprogress.org
 Satisfied demand for contraception (% of satisfied demand) 0.26 0.20
Perception of corruption (0=high corruption; 100=low corruption) 0.29 0.22
Early marriage (% of married women aged 15-19) 0.27 0.21
Young people not in education, employment or training (% of youth) 0.18 0.14
Vulnerable employment (% of total employment) 0.29 0.22
Equality of political power by gender (0=unequal power; 4=equal power) 0.27 0.22
Equality of political power by social group (0=unequal power; 4=equal power) 0.26 0.20
Equality of political power by socioeconomic position (0=unequal power; 4=equal
power) 0.26 0.21
Discrimination and violence against minorities (0=low; 10=high) 0.22 0.18
Acceptance of gays and lesbians (0=low; 1=high) 0.25 0.20
Citable documents (documents/1000 people) 0.28 0.23
Academic freedom (0=low; 1=high) 0.15 0.13
Women with advanced education (proportion of females) 0.27 0.23
Expected years of tertiary schooling (years) 0.28 0.23
Quality weighted universities (points) 0.22 0.19

Appendix D: Descriptive Statistics for 2021 Social Progress Index,

Component, and Dimension Scores
The following descriptive statistics are based on the sample of 174 countries for which we can
calculate at least 9 components for the 2021 Social Progress Index.

Obs. Mean Standard Deviation Minimum Maximum

Social Progress Index 170 67.25 15.10 32.50 92.63
Basic Human Needs 173 75.94 16.21 29.91 96.85
Foundations of Wellbeing 173 67.62 15.48 34.17 93.80
Opportunity 171 58.04 15.96 23.37 89.30
Nutrition and Basic Medical Care 174 84.57 14.37 36.42 98.99
Water and Sanitation 174 76.17 23.27 14.80 99.27
Shelter 174 76.89 18.86 26.79 96.87
Personal Safety 173 66.35 14.62 29.46 96.18
Access to Basic Knowledge 173 74.68 19.57 23.14 99.51
Access to Information and
Communications 174 66.18 20.85 5.67 98.18
Health and Wellness 174 61.97 16.13 21.03 92.10
Environmental Quality 174 67.19 14.37 23.95 95.15
Personal Rights 173 69.18 22.05 3.19 97.91
Personal Freedom and Choice 173 62.74 15.10 26.67 91.16
Inclusiveness 172 46.64 16.98 4.26 83.77
Access to Advanced Education 173 53.84 18.53 19.70 89.60

35 | socialprogress.org
 Appendix E: Bibliography and Further Reading
Bland, J. M., and D. G. Altman. “Cronbach’s Alpha.” BMJ (Clinical Research Ed.) 314, no. 7080
(1997): 572.
Dunteman, George H. Principal Components Analysis. SAGE, 1989.
Fleurbaey, M. and D. Blanchet. Beyond GDP: Measuring Welfare and Assessing Sustainability,
Oxford University Press, 2013.
Manly, Bryan F. J. Multivariate Statistical Methods: A Primer. CRC Press, 1994.

For further reading on social progress, development, and composite indices, we recommend:

Bishop, Matthew, and Michael Green. The Road from Ruin: How to Revive Capitalism and Put
America Back on Top. New York: Crown Business, 2011.
Delgado, Mercedes, Christian Ketels, Michael E. Porter, and Scott Stern. The Determinants of
National Competitiveness. Working Paper. National Bureau of Economic Research, July
2012. http://www.nber.org/papers/w18249.
Fehder, Daniel, Michael E. Porter, and Scott Stern. “The Empirics of Social Progress: The
Interplay between Subjective Well-Being and Societal Performance.” AEA Papers and
Proceedings, 108 (2018): 477-482.
Furman, Jeffrey L., Michael E. Porter, and Scott Stern. “The Determinants of National Innovative
Capacity.” Research Policy 31, no. 6 (2002): 899–933.
Gehl, Katherine M., and Michael E. Porter. “Why Competition in the Politics Industry is Failing
America.” Harvard Business School, 2017.
https://www.hbs.edu/competitiveness/Documents/why-competition-in-the-politics-industry-
is-failing-america.pdf
Joint Research Centre-European Commission. Handbook on constructing composite indicators:
methodology and user guide. OECD Publishing, 2008.
Kuznets, Simon. "Economic Growth and Income Inequality." The American Economic
Review 45, no. 1 (1955): 1-28.
Porter, Michael E. Competitive Advantage of Nations: Creating and Sustaining Superior
Performance. Simon and Schuster, 2011.
Porter, Michael E. Competition in Global Industries. Harvard Business Press, 1986.
Sen, Amartya Kumar. Commodities and Capabilities. North-Holland Publ., 1985.
Sen, Amartya Kumar. Development as Freedom. Oxford University Press, 1999.
Stevenson, Betsey, and Justin Wolfers. Economic growth and subjective well-being:
Reassessing the Easterlin paradox. No. w14282. National Bureau of Economic Research,
2008.
Stiglitz, Joseph, Amartya Sen, and Jean-Paul Fitoussi. "The measurement of economic
performance and social progress revisited." Reflections and overview. Commission on the
Measurement of Economic Performance and Social Progress, Paris (2009).

36 | socialprogress.org