1. POLICY IMPACT EVALUATION THROUGH
PROSPERITY METRICS AND OPEN DATA SOURCES
Panagiotis Kokkinakos
Costas Koutras
Ourania Markaki
Sotirios Koussouris
John Psarras
National Technical University of Athens
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
2. Evaluating Policy Making via Prosperity Metrics
•Which metrics?
•What kind of indicators?
•For which kind of applications?
•How do we exactly interpret concepts like
• Quality of life
• Prosperity
• Performance measurement
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
2
?
3. The Social Indicators Zoo
• Aggregate (summary) indicators
• e.g. GDP
• Composite Indicators
• Combine individual indicators into a single composite index, possibly weighting
the indicators according to their relative importance – e.g. HDI
• Headline Indicators
• Sustainability Indicators
• International, National, Regional and Local indicators
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
3
4. Issues on the use of indicators
• Institutionalization: setup routine procedures and practices to ensure
the continuing existence of an indicator and to legitimize the method
and concept of the measure. (Drawbacks identified)
• Avoiding bias: indicators should be produced by professional
statistical agencies that have a strong awareness of policy issues,
without having responsibility for them.
• Maintaining the sensitive balance between (global) standardization
and local democracy (identifying the relevant measure to the
specific local circumstances)
4-5 June 2015
4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
4
5. A methodological framework of indicator
development [Coombes, Wong, 1994]
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
5
Step 1: Conceptual consolidation
Clarifying the concept to be represented by the analysis
Step 2: Analytical Structuring
Providing an analytical framework within which key factors will be analysed
Step 3: Identification of Indicators
Translation of key factors identified in step 2 into specific measurable indicators
Step 4: Synthesis of indicator values
Synthesizing the identified indicators into composite index/indices or into analytical summary
6. How can Policy Compass contribute?
• "Open data" movement and the "social computing" capabilities of
today open a fantastic observatory of experimentation with (and
for) Social Indicators
• Assess the limits of `causal analysis‘ in policy making, combining
indicators with causal models
• Assess the feasibility of assisting policy design and decision-
making with quantitative handles (Social Indicators) at the global &
local level
• Evaluate the ramifications and the consequences of mixing the
previous items with eGov techniques, eParticipation tools & Social
Media.
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
6
7. The Problem
• Internet has evolved into a rich source
for information but also to an
instrument of spreading misinformation
and propaganda
• Lack of consensus about a suitable
metric for measuring progress
• Difficulty of objectively assessing the
impacts of government policies
The Proposed Approach
Open Public Data
Prosperity Indicators
Fuzzy Cognitive Maps
Argumentation Technology
Deliberation Platforms and Social
Media
The Context
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
7
Improve the quality and
transparency of the policy
analysis and evaluation phases
of the policy cycle for a variety
of stakeholders, ranging from
citizens to policy makers
Policy
cycle
Analysis
Implementation
/Monitoring
8. The Proposed Approach: Overview
• A research prototype of an easy-to-use, highly visual and intuitive tool for:
• Constructing prosperity and other policy metrics with an easy-to-use visual
language for defining variables and functions over open data sources.
• Constructing graphs and charts visualizing metrics for selected
geographical regions and time periods.
• Annotating graphs and charts with political or policy events.
• Constructing causal models with an easy-to-use visual tool for Fuzzy
Cognitive Maps (FCM).
• Sharing and debating prosperity graphs and FCM across popular social
media platforms.
• Summarizing and visualizing the debates in argument maps and
conducting structured surveys about policy issues
• Aggregating opinions on policy issues, to formulate a common position in a
party or interest group.
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
8
9. Pillar I: Prosperity Indexes
• Prosperity metrics capture the level of welfare and quality of life in a
given region or society.
• Prosperity is a vague and subjective concept with essential
psychological, social and economic aspects.
• There is no consensus about how to objectively measure prosperity
• Indicators of economic growth:
• Gross Domestic Product (GDP)
• Genuine Progress Index (GPI)
• Index of Sustainable Economic Welfare (ISEW)
• GINI Index
• Alternatives:
• Human Development Index (HDI)
• Legatum Prosperity Index
• “Healthy life years statistics” by Eurostat
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
9
10. Pillar II: Open Public Data
• Open and unrestricted access to large scale data sets is essential for
political engagement and scientific research
• Available large scale data sets have nowadays their own self-
contained existence rules.
• Micro-data can be used to construct new indicators of multifaceted
nature.
• Sources of micro-data:
• Eurobarometer surveys
• European Union Statistics on Income and Living Conditions (EU-SILC) by Eurostat
• Urban Audit (the European cities Eurostat)
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
10
11. Pillar III: Fuzzy Cognitive Maps
• Fuzzy Cognitive Maps (FCMs) provide a well-founded, general-purpose and
intuitive method for modelling and simulating relationships between variables.
• FCMs have been introduced by B. Kosko (1986) as a fuzzified version of
Cognitive Maps, originally introduced by political scientist R. Axelrod (1976).
• An FCM is a fuzzy directed graph of nodes and edges, where nodes
represent fuzzy concepts, describing behavioral characteristics of a system
that occur to some degree, and directed edges represent the causal
relationships among these concepts.
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
11
• The graph edges are weighted by a real
value from the interval [-1, 1], which
expresses the strength of the relation
between two concepts.
• FCMs have been widely used to model
and simulate policies and their effects.
12. Pillar IV: Argumentation Technology
• Argumentation support systems are computer software for helping people
participate in various kinds of goal-directed dialogues in which arguments are
exchanged.
• The idea of using argumentation support systems for e-Participation can be
traced back at least to Horst Rittel’s pioneering work in the early 1970s who
used visual maps of arguments, to help people collaborate and find solutions
to what he called “wicked problems”.
• “Wicked problems” have no algorithmic, scientific or objectively optimal
solutions for a variety of reasons, including the lack of consensus among
stakeholders about utilities and values.
• Typically, e-Participation projects make use of generic groupware systems
(e.g. discussion fora, online surveys, etc.) not providing though specific
technical support for argumentation.
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
12
13. Pillar V: Deliberation Platforms and Social Media
• Deliberation platforms incarnate the efforts taken by government
agencies, to increase citizens’ engagement in their decision and
policy making processes.
• The first wave of deliberation platforms has witnessed extensive
information on government activities, decisions, plans and policies,
the proliferation of e-voting and e-consultation spaces, along with
various types of e-fora.
• Still, the first generation of deliberation platforms did not meet the
original expectations.
• The advent of Web 2.0 tools has created a more vivid environment
and the popularity of the social media has set a new battlefield for the
concept of e-Participation.
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
13
14. Overview of Main Concepts
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
14
15. Use Case Scenarios
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
15
16. Lessons Learnt
• There exists a growing concern on the limits of GDP-like measures for
measuring societal welfare. GDP is very successful, but it is impossible to
derive deep results on many aspects of everyday life with simple aggregate
indicators alone.
• It makes perfect sense to attempt constructing aggregate, headline or
composite indicators, to measure important social phenomena at the
regional /local level.
• There do not exist ‘objective’ or ‘neutral’ indicators of any kind. The
community does not pursue objective indicators, it rather attempts to
construct useful ones.
• A broad experimentation on the calculation and exploitation of indicators,
given the multitude of data available, is definitely interesting scientifically and
will provide useful "social" feedback
4-5 June 2015 4th International Symposium and 26th National Conference on Operational
Research 2015 – Chania, Greece
16