Give me the place to stand: Leverage analysis in systemic design
0 likes1,331 views
This document discusses leveraging social network analysis techniques like centrality measures and structural analysis to gain insights from systems models. It proposes that systems maps can be viewed as graphs that describe relationships between elements, enabling the application of metrics like closeness, betweenness, eigenvector centrality, and structural analysis methods. This could help identify key influential elements, bottlenecks, and isolated structures to find leverage points for change. Future research opportunities are discussed, such as developing guidelines for interpreting results and exploring additional metrics and algorithms.
![Three questions about systems models
• (1) How might we balance the trade-offs of “soft” and “hard” systems
thinking?
– Forrester (1994): “Systems thinking and soft OR […] rely on subjective use of unreliable
intuition for evaluating the complex structures that emerge from the initial description of the
real system.”
– Checkland (1984): “Systems engineering, based on defining goals or objectives, simply
did not work when applied to messy, ill-structured, real-world problems.”
• (2) How might we handle complexity?
– Jones (2014): Representative maps include input from more stakeholders
– Crowdsourcing (Lukyanenko & Parsons, 2012) and data science (Šćepanović, 2018)
offer tools to support large-scale data collection
• (3) How might we learn from these models?
– Models are excellent opportunities to find the most important
actors/phenomena/structures in a system: “leverage points” (Meadows, 1999)](https://image.slidesharecdn.com/6-murphy-jones-final-190321153940/85/Give-me-the-place-to-stand-Leverage-analysis-in-systemic-design-2-320.jpg)
![Applying centrality and structural
analysis to causal loop diagrams
Metric/Method Description Social Networks In Causal Loop Diagrams?
Reach The number of elements
within [x] steps of the given
element
Quick propagation of information
through the network; widely
accessible (Hanneman & Riddle,
2005)
The map is highly sensitive
to these elements
Reach efficiency The reach divided by the
degree of a given node
Efficient (non-redundant)
information spreading; high
exposure with limited influence
on the given element (Hanneman
& Riddle, 2005)
Quickly and efficiently
propagate change
throughout the rest of the
network; is not likely to be
highly influenced by the
rest of the system](https://image.slidesharecdn.com/6-murphy-jones-final-190321153940/85/Give-me-the-place-to-stand-Leverage-analysis-in-systemic-design-8-320.jpg)
![References
Borgatti, S. P. (2005). Centrality and network flow. Social Networks, 27(1), 55–71. https://doi.org/10.1016/j.socnet.2004.11.008
Checkland, P. (1985). From Optimizing to Learning: A Development of Systems Thinking for the 1990s. The Journal of the Operational Research Society, 36(9), 757–767. https://doi.org/10.2307/2582164
Forrester, J. W. (1994). System dynamics, systems thinking, and soft OR. System Dynamics Review, 10(2–3), 245–256. Retrieved from http://onlinelibrary.wiley.com/doi/10.1002/sdr.4260100211/abstract
Freeman, L. C. (1977). A Set of Measures of Centrality Based on Betweenness. Sociometry, 40(1), 35–41. https://doi.org/10.2307/3033543
Freeman, L. C. (1979). Centrality in social networks conceptual clarification. Social Networks, 1(3), 215–239.
Jones, P. H. (2014). Systemic Design Principles for Complex Social Systems. In G. S. Metcalf (Ed.), Social Systems and Design (pp. 91–128). Springer Japan. https://doi.org/10.1007/978-4-431-54478-4_4
Lukyanenko, R., & Parsons, J. (2012). Conceptual modeling principles for crowdsourcing (pp. 3–6). ACM. https://doi.org/10.1145/2390034.2390038
Meadows, D. (1997). Leverage Points: Places to Intervene in a System. Retrieved November 29, 2015, from http://www.donellameadows.org/wp-content/userfiles/Leverage_Points.pdf
Newman, M. (2010). Networks: An Introduction. Oxford, New York: Oxford University Press.
Oliva, R. (2004). Model structure analysis through graph theory: partition heuristics and feedback structure decomposition. System Dynamics Review (Wiley), 20(4), 313–336. https://doi.org/10.1002/sdr.298
Ozbekhan, H. (1970). The predicament of mankind: A quest for structured responses to growing world-wide complexities and uncertainties (Original Proposal to the Club of Rome). Geneva, Switzerland: The Club of Rome.
Retrieved from http://quergeist.net/Christakis/predicament.pdf
Rittel, H. W., & Webber, M. M. (1973). Dilemmas in a general theory of planning. Policy Sciences, 4(2), 155–169. Retrieved from http://link.springer.com/article/10.1007/BF01405730
Šćepanović, S. (2018). Data science for sociotechnical systems - from computational sociolinguistics to the smart grid. Aalto University. Retrieved from https://aaltodoc.aalto.fi:443/handle/123456789/30187
Schoenenberger, L., Schmid, A., & Schwaninger, M. (2015). Towards the algorithmic detection of archetypal structures in system dynamics. System Dynamics Review (Wiley), 31(1/2), 66–85. https://doi.org/10.1002/sdr.1526
Stroh, D. P. (2015). Systems Thinking For Social Change: A Practical Guide to Solving Complex Problems, Avoiding Unintended Consequences, and Achieving Lasting Results. Chelsea Green Publishing.
Xie, J., Szymanski, B. K., & Liu, X. (2011). SLPA: Uncovering Overlapping Communities in Social Networks via A Speaker-listener Interaction Dynamic Process. ArXiv:1109.5720 [Physics]. Retrieved from
http://arxiv.org/abs/1109.5720](https://image.slidesharecdn.com/6-murphy-jones-final-190321153940/85/Give-me-the-place-to-stand-Leverage-analysis-in-systemic-design-21-320.jpg)
Give me the place to stand: Leverage analysis in systemic design
- 1. Give me the place to stand: Leverage analysis in systemic design RSD7: Models and processes of systemic design Ryan J. A. Murphy and Peter Jones October 24, 2018
- 2. Three questions about systems models • (1) How might we balance the trade-offs of “soft” and “hard” systems thinking? – Forrester (1994): “Systems thinking and soft OR […] rely on subjective use of unreliable intuition for evaluating the complex structures that emerge from the initial description of the real system.” – Checkland (1984): “Systems engineering, based on defining goals or objectives, simply did not work when applied to messy, ill-structured, real-world problems.” • (2) How might we handle complexity? – Jones (2014): Representative maps include input from more stakeholders – Crowdsourcing (Lukyanenko & Parsons, 2012) and data science (Šćepanović, 2018) offer tools to support large-scale data collection • (3) How might we learn from these models? – Models are excellent opportunities to find the most important actors/phenomena/structures in a system: “leverage points” (Meadows, 1999)
- 3. Ways forward: borrowing from social network analysis and systems dynamics • Many systems models (e.g., Causal Loop Diagrams) are graphs – Formal definition: a set of vertices (the elements of the system) and edges describing a relationship between the vertices (e.g., connections between elements) – Graph theory provides analytical methods for understanding graphs, such as: • Centrality analysis • Structural analysis – These methods have not been applied to soft systems models
- 4. Example from centrality analysis: Degree • The number of connections of a given element (Newman, 2010) – Indegree • The number of incoming connections. • An indicator of popularity – Outdegree • The number of outgoing connections. • An indicator of gregariousness
- 5. High-closeness SDGs: Inequality, Sustainable Consumption & Production, Peaceful & Inclusive Societies High-closeness targets exist as well (Model based on Le Blanc, 2015)
- 6. Applying centrality and structural analysis to causal loop diagrams Metric/Method Description In Social Networks In Causal Loop Diagrams? Degree The number of connections Higher connectivity to the rest of the network; influence, access, prestige (Newman, 2010) Immediate impact, sensitivity, resilience Indegree The number of incoming connections High inward connectivity to the rest of the network; sensitivity to information, influence (Newman, 2010) Receives change from many other elements; may be highly volatile or highly stable Outdegree The number of outgoing connections High outward connectivity to the rest of the network; rapid communication/high access to the rest of the network, highly infectious (Newman, 2010) Change in the given phenomena is felt by many other elements; impact, power
- 7. Applying centrality and structural analysis to causal loop diagrams Metric/Method Description In Social Networks In Causal Loop Diagrams? Betweenness Frequency of participation in the shortest path between two other elements Member has a high degree of control; the network is dependent on the member; bottlenecking, control, influence (Freeman, 1979) Phenomena is a gateway or bottleneck for change; change strategies must consider how to prevent blocking Closeness Average length of the shortest paths between the given vertex and every other vertex in the graph High visibility to the rest of the network and information spreads easily from this member; independence from the rest of the graph (Freeman, 1979) Phenomena is highly powerful; likely to be resistant to change, and therefore a key indicator of success or failure Eigenvector Connectedness to other well- connected elements Influence of highly influential elements; influence (Newman, 2010) High-impact phenomena; likely key phenomena to change in pursuit of a given strategy
- 8. Applying centrality and structural analysis to causal loop diagrams Metric/Method Description Social Networks In Causal Loop Diagrams? Reach The number of elements within [x] steps of the given element Quick propagation of information through the network; widely accessible (Hanneman & Riddle, 2005) The map is highly sensitive to these elements Reach efficiency The reach divided by the degree of a given node Efficient (non-redundant) information spreading; high exposure with limited influence on the given element (Hanneman & Riddle, 2005) Quickly and efficiently propagate change throughout the rest of the network; is not likely to be highly influenced by the rest of the system
- 9. Example from structural analysis: Level partitions (Oliva, 2004)
- 10. Example: Education systems change • Level partition only results in two levels • Cycle partition results in a single cycle set – Not surprising • Loop inclusion graph: 3 (4) 2 (7) 4 (8) 1 (4)5 (2) 6 (8) 7 (8) 8 (8) 9 (8) 10 (8) 11 (7) 12 (7) 13 (7) 14 (7) 15 (7) 16 (7) 17 (9) 18 (9) Zero One Two Three (Model based on Murphy, 2016)
- 11. Example: Education systems change • Level partition only results in two levels • Loop inclusion graph:
- 12. Example: Education systems change • Level partition only results in two levels • Cycle partition results in a single cycle set – Not surprising • Loop inclusion graph: 3 (4) 2 (7) 4 (8) 1 (4)5 (2) 6 (8) 7 (8) 8 (8) 9 (8) 10 (8) 11 (7) 12 (7) 13 (7) 14 (7) 15 (7) 16 (7) 17 (9) 18 (9) Zero One Two Three
- 16. Applying centrality and structural analysis to causal loop diagrams Metric/Method Detail Dynamics models In Causal Loop Diagrams? Level partition Which variables are dependent on which? Hierarchy of causal structure (Oliva, 2004) Elements at the “bottom” of the hierarchy are uncontrollable within the system; elements at the top are highly dependent on the rest of the system Cycle partition Which other variables share the same predecessors or successors? Illustrates cycle set “dominance” → sub-cycles sets must be understood before their “parents” (but not that useful as most elements in models sit in the same cycle set; Oliva, 2004) Sub-cycle set elements dictate the behaviour of supercycles Shortest Independent Loop Set A decomposition of the cycle partition showing which loops are included in which - Illustrates a loop hierarchy - With level partitioning, gives an ordering from simple loops to complex loops - Shows isolated loop structures (Oliva, 2004) - Simple loops are easier to experiment with than more complex loops - Inner loops will influence the behaviour of their containing loops - Isolated structures are more easily manipulated
- 17. Discussion • Important centrality measures: – Closeness might be used to find key indicators of success (recall rule 4 of Rittel & Webber, 1973), especially in combination with structural analysis – High betweenness elements are bottlenecks – Reach efficiency indicates elements that are minimally influenced themselves but are potentially powerful sources of impact elsewhere – Eigenvector centrality indicates high-influence elements in general • (are these the leverage points of the system?) • Structural analysis is potentially powerful – Especially in combination with centrality measures
- 18. Limitations • Does this go beyond the ease-of-use of systems thinking techniques? • What is the “unit” of change? – SNA metrics were developed to model the flow of information... What flows in a systems map? • Need for normalization – What is the role of delay? Same/opposite connections? • Interpretation is (still) important
- 19. Future research • Ontological guidelines for mapping and normalization • Guidelines for interpretation and use • Explore additional metrics – Compare with different types of network flows (e.g., Borgatti, 2005) – Community detection (e.g., Xie, Szymanski, & Liu, 2011) – Automated identification of archetype patterns (e.g., Schoenenberger, Schmid, & Schwaninger, 2015) • Weighted metrics + algorithms to implement them – E.g., reach efficiency weighted by eigenvector value • Further testing of validity/utility • The need for clear case studies with which to experiment • Systems dynamics vs. systems thinking: from dichotomy to spectrum?
- 20. Conclusion • A novel use of centrality measures and structural analysis is found by importing them into systems thinking – These measures are easy to implement in many mapping and diagramming applications • We may be able to make systems thinking approaches more rigorous without the intractability of systems dynamics
- 21. References Borgatti, S. P. (2005). Centrality and network flow. Social Networks, 27(1), 55–71. https://doi.org/10.1016/j.socnet.2004.11.008 Checkland, P. (1985). From Optimizing to Learning: A Development of Systems Thinking for the 1990s. The Journal of the Operational Research Society, 36(9), 757–767. https://doi.org/10.2307/2582164 Forrester, J. W. (1994). System dynamics, systems thinking, and soft OR. System Dynamics Review, 10(2–3), 245–256. Retrieved from http://onlinelibrary.wiley.com/doi/10.1002/sdr.4260100211/abstract Freeman, L. C. (1977). A Set of Measures of Centrality Based on Betweenness. Sociometry, 40(1), 35–41. https://doi.org/10.2307/3033543 Freeman, L. C. (1979). Centrality in social networks conceptual clarification. Social Networks, 1(3), 215–239. Jones, P. H. (2014). Systemic Design Principles for Complex Social Systems. In G. S. Metcalf (Ed.), Social Systems and Design (pp. 91–128). Springer Japan. https://doi.org/10.1007/978-4-431-54478-4_4 Lukyanenko, R., & Parsons, J. (2012). Conceptual modeling principles for crowdsourcing (pp. 3–6). ACM. https://doi.org/10.1145/2390034.2390038 Meadows, D. (1997). Leverage Points: Places to Intervene in a System. Retrieved November 29, 2015, from http://www.donellameadows.org/wp-content/userfiles/Leverage_Points.pdf Newman, M. (2010). Networks: An Introduction. Oxford, New York: Oxford University Press. Oliva, R. (2004). Model structure analysis through graph theory: partition heuristics and feedback structure decomposition. System Dynamics Review (Wiley), 20(4), 313–336. https://doi.org/10.1002/sdr.298 Ozbekhan, H. (1970). The predicament of mankind: A quest for structured responses to growing world-wide complexities and uncertainties (Original Proposal to the Club of Rome). Geneva, Switzerland: The Club of Rome. Retrieved from http://quergeist.net/Christakis/predicament.pdf Rittel, H. W., & Webber, M. M. (1973). Dilemmas in a general theory of planning. Policy Sciences, 4(2), 155–169. Retrieved from http://link.springer.com/article/10.1007/BF01405730 Šćepanović, S. (2018). Data science for sociotechnical systems - from computational sociolinguistics to the smart grid. Aalto University. Retrieved from https://aaltodoc.aalto.fi:443/handle/123456789/30187 Schoenenberger, L., Schmid, A., & Schwaninger, M. (2015). Towards the algorithmic detection of archetypal structures in system dynamics. System Dynamics Review (Wiley), 31(1/2), 66–85. https://doi.org/10.1002/sdr.1526 Stroh, D. P. (2015). Systems Thinking For Social Change: A Practical Guide to Solving Complex Problems, Avoiding Unintended Consequences, and Achieving Lasting Results. Chelsea Green Publishing. Xie, J., Szymanski, B. K., & Liu, X. (2011). SLPA: Uncovering Overlapping Communities in Social Networks via A Speaker-listener Interaction Dynamic Process. ArXiv:1109.5720 [Physics]. Retrieved from http://arxiv.org/abs/1109.5720