Staging Model Abstraction – an example about political participation

Staging Model Abstraction, Bruce Edmonds, From Cases to General Principles, ABM + Theory, Hannover, July 2019. slide 1
Staging Model Abstraction
– an example about political
participation
Bruce Edmonds
Centre for Policy Modelling
Manchester Metropolitan University

SCID was a 5-year project funded by the EPSRC to apply
complexity methods to issues of social importance.
A collaboration between the:
• Centre for Policy Modelling at the MMU
• Cathy Marsh Centre at the UoM
• Dept. of Theoretical Physics at the UoM
Involved in the initial model development: Laurence Lessard-
Phillips, Ed Fieldhouse, Thomas Loughran and Bruce Edmonds
Involved in the abstraction stages: Luis Fernandez Lafuerza,
Louise Dyson, Bruce Edmonds and Alan McKane

Model ‘Theory’
• Just as machines extend our physical abilities,
models extend our mental abilities
• The model is a tool for understanding what we
observe (at least, ultimately)
• Some of the understanding is as a result of the
model but the rest is in the model
• With complex systems we often seem to be
forced to choose between rigour and relevance
• The trouble is science is needs both!
• It needs some level of generalisation but not
much

Intuitive understanding expressed in normal
language
Observations of the natural system of concern
Data obtained by measuring the
system
Model of the processes in the
system
ScientificComparisons
Common-SenseComparison
Model of the model

Our Basic Approach…
…is to stage abstraction with an intermediate, complex
model, that is then, itself, modelled (a ‘KIDS’ approach)
• The Data Integration Model (DIM) includes all that is
deemed relevant by social scientists – more like a
dynamic description than anything else
• The simpler models of the DIM are developed by
formal scientists but validated against the DIM
Data Evidence
Simple Model
Data Evidence
Simple Model
Complex Model
Representation
Simplification
DIM

Model ‘layers’ and the different kinds of
data used to support these
Underlying data about
population composition
Demographics of people in
households
Social network formation and
maintenance (homophily)
Influence via social networks
• Political discussions
Voting Behaviour

Discuss-politics-with person-23 blue expert=false
neighbour-network year=10 month=3
Lots-family-discussions year=10 month=2
Etc.
Memory
Level-of-Political-Interest
Age
Ethnicity
Class
Activities
AHousehold
An Agent’s Memory of Events
Etc.
Changing personal
networks over which
social influence occurs
Composed of households of
individuals initialised from
detailed survey data
Each agent has a rich variety of
individual (heterogeneous)
characteristics
Including a (fallible) memory of
events and influences

Evidence schema
Qualitative
Intuitive
understanding
Observations
of Phenomena
Quantitative
Models
Data
Text from interviewsTime Series Data etc.
A-B Simulation ‘Causal Stories’
Expert OpinionConclusions

Example Output: why do people vote (if
they do)
Intervention: voter
mobilisation
Effect: on civic
duty norms Effect: on habit-
based behaviour
Time
%ofvotersbyreason

Simulated Social Network at 1950
Established
immigrants: Irish,
WWII Polish etc.
Majority: longstanding
ethnicities
Newer
immigrants

Simulated Social Network at 2010

Outline of Model Reduction
• Iterative process of inspection of DIM, formulating simpler models, and
comparing them with output from the DIM
• Red and green processes were simplified, also parties and (initially)
social network

Comparison of Reduced and DIM
Models
Low and high turnout regimes
Broad agreement between models, but different
levels and different dynamics in transition region

With Different Kinds of Network
Blue = DIM Network, Green = With “clumped”
network, Red = no network (bistability in transition)

Synthetic Network
Kinds of network
A synthetic network that is composed of small
groups with some random inter-group connections
resulted in better fit of dynamics
Network in DIM

Fixed vs. Dynamic Networks
Reduced models with fixed ‘clumped’ network (blue)
and with a dynamic network where links between
households change (red)
Dynamism increases turnout due to diversity of links
over time allowing wider influence

Individual vs. Household
Immigration
With immigrants of either a lower or higher political
interest than natives, individual immigration resulted
in higher level of turnout than household
immigration, due to asymmetry of influence process
Low Interest Immigrants High Interest Immigrants
Individual
Immigration
Family
Immigration

Final Comparison
Reduced model + dynamic ‘clumped’ network +
household immigration has good fit to DIM turnout
dynamics but amenable to complete checking and
much simpler and easier to experiment with

What was Learnt?
The comparisons suggested that:
• distinctive ‘high’ and ‘low’ turnout regimes exist,
depending on rate of discussions
• when in extreme turnout regimes, the system is
stable, whereas in the transition regime turnout
can 'flip' between low and high
• the lack of bistability in the transition is due to the
‘locality’ and sparseness of the network
• the dynamism of the network resulted in higher
turnout levels
• household immigration reduced turnout compared
to individual immigration

Final Picture – indirect but staged
knowledge!
Data-Integration Simulation Model
Micro-Evidence Macro-Data
Reduced Simulation Model
Analytic Model
Even Simpler Simulation Model
Reported in
this
presentation
Further
iteration of
this
approach

Conclusions
• Relevance and rigour combined via a ‘chain’ of highly
related models (not in a single model)
• Abstraction achieved in stages, not in one ‘jump’
• Each model acts as a check on the other
• Further iterations of model simplification in progress
towards completely analytic models
• Whether generalisation is possible is a contingent fact, it
may be that what you are modelling is more like zoology
than physics regardless for a wish for general laws
• A model may be too complex to understand, but they are
still useful – we might understand a model of the model, or
a model of the mode of the model etc.
• Limiting ourselves to models we can directly understand is
a mistake, and will limit what we can study (just as
keeping to analytic mathematics did)

References
Description of base model (the DIM): Fieldhouse, E; Lessard-Phillips, L;
and Edmonds, B. (2016) Cascade or echo chamber? A complex agent-
based simulation of voter turnout. Party Politics. 22(2):241-256.
DOI:10.1177/1354068815605671
Detail of evidential base of the DIM: T. Loughran, L. Lessard-Phillips, E.
Fieldhouse, B. Edmonds, The voter model – a long description, SCID
project document, http://scidproject.files.wordpress.com/2015/05/the-
voter-modeldescription-final-draft.docx (2015)
Base model available at: https://www.comses.net/codebases/4368
First stage of abstraction (described here): Lafuerza LF, Dyson L,
Edmonds B, McKane AJ (2016) Staged Models for Interdisciplinary
Research. PLoS ONE, 11(6): e0157261.
DOI:10.1371/journal.pone.0157261
Description of simplified model: L.F. Lafuerza, L. Dyson, B. Edmonds,
A.J. McKane, arXiv:1604.00903 (2016)
Further abstraction stage (not discussed here): Lafuerza, LF, Dyson, L,
Edmonds, B & McKane, AJ (2016) Simplification and analysis of a model
of social interaction in voting, European Physical Journal B, 89:159.
DOI:10.1140/epjb/e2016-70062-2

The End
Bruce Edmonds: bruce@edmonds.name
Centre for Policy Modelling: http://cfpm.org
Manchester Metropolitan University: http://mmu.ac.uk
More about the SCID project: http://cfpm.org/scid
These slides at: http://slideshare.com/BruceEdmonds

Staging Model Abstraction – an example about political participation

Empfohlen

Empfohlen

Weitere ähnliche Inhalte

Ähnlich wie Staging Model Abstraction – an example about political participation

Ähnlich wie Staging Model Abstraction – an example about political participation (20)

Mehr von Bruce Edmonds

Mehr von Bruce Edmonds (19)

Kürzlich hochgeladen

Kürzlich hochgeladen (20)

Staging Model Abstraction – an example about political participation