There are times when we accept publically accept an approximation to move forward with life. Take, for instance, π (pi), better known for representing that irrational number of 3.14159 (and a lot of change) starring as the ratio of the circle’s circumference to its diameter. Yes, many of us recall our friend Archimedes and π from trigonometry class. Pi (π) is a mathematical constant derived from the area of a circle. Plug that π in to make sense of some spatial relationship. There is no riotous call for proofing unless it is on the exam or the SAT. Do students even proof anymore? The π is rationally accepted to serve a purpose to an end, to finish that darn homework on distributed systems. I will argue that the allegory of π can help us avoid clouded, myopic policy decisions.
A mental model is a cognitive tool. Success comes with altering those mental models. In other words, a mental model is not the dénouement. Formal simulations and mental models need each other. Mental models fall asleep on the job when complexity of the policy is dynamic (Sterman 2000). But it must be remember that policies are not spit out by modelling programs nor should they. So why am I calling upon modelling when policymakers already are bombarded with so much? I present the “policy π” in application of model testing to policymaking. “Policy π” focuses on making decisions that account for both systemic uncovered through the approximations garnered from formal modelling as well as practical issues posed in implementing the policy in the real world. Included in this “Policy π” is the common issue of policy ratification.
Regardless of the policy model used in a policy system,
when a new policy is written or committed in the system,
the administrator must consider [policy ratification as] how the new policy interacts with those
already existing in the system. (Agrawal et al 2005)

The model does NOT offer the “answer”. But the model is downright necessary to understand what is going on. When people brain storm, we battle with mental (qualitative) models pieced together as divergent points (inputs). We have all been there. Mentally, we approximate all the time. It is a natural part of the policy process. To the miasmic stench of permanent markers littering a flip chart, the decision makers and perhaps a silent minority are left to make sense of that complex data to get at the process that has no obvious end in sight. There needs to be “reflective conversation with the situation (policy )”of the simulated results (Schon 1992). The model does NOT forecast. The model may not be able to capture all of the connections to other policies that it is related to. It is a simulation. It is a model that is based on decision rules. It is a network with a priori boundaries.
Even with the best minds around the table, there still must be a mechanism to unleash the mental models around the table. Just as health policy would not move forward without epidemiology evidence, so should be the place of systems thinking in the policy process. Systems offers math to support or refute initial reactions to early conditions viewed under a policy. But the model will not be a predictive crystal ball. Much chatter has been circulated about the overreach of models into the world of prediction in policy. Agreed. But the meat of policy is not the model. In my opinion it should not be viewed in that way.
The policy π supported by the model, may give you sweet potato when the table (policymakers) ordered and expected the boysenberry that is now out of season. There may still be that flaky crust and reasonable price. But the input of a tuber over a berry is a notable change. Some of the ingredients may change. The taste will differ. Cooking time may change. You may not like boysenberry. The change to sweet potato could require a new produce distributor, who must then must recalibrate (with excitement) accommodating the needs of a new customer. Your fellow diners may try to convince you to hold out for boysenberry though there is a chance that maybe this sweet potato pie may remind you of Grandma’s. Sometimes a model changes that policy “pie” expectation. The model will not cover every possible scenario. Can you accept the presentation of the sweet potato? And perhaps next week, once again upending your policy gastronomy? Yes, a pie is a pie. Both a sweet potato and a boysenberry are dicotyledonous. Few would confuse the two. Each would present its own unique set of growth habits and susceptibilities. But food stuffs present different phenotypes and culinary experiences. Policy satiates, leaving some bellies full while others push away from the table wanting. Each policy π must be approached and respected as a necessity to prudent, systemic action to debate, not as a letdown of the expected π wished to grace the plate. How filling was that policy pi?

References
Agrawal, D et al (June 2005) Policy Ratification. Presented at Policies for Distributed Systems and Networks Sixth IEEE International Workshop
Schon D (1992).The theory of inquiry: Dewey’s legacy to education. Curriculum Inquiry. 22(2), 119-39
Sterman J (2002) All models are wrong: reflections on becoming a systems scientist. System Dynamics Review, 18, 501-531

Posted in Complexity theory, Computational models, Interdisciplinary, system dynamics, Systems thinking
Leave a comment

August 11, 2014

Well, my article in System Dynamics is finally in Early View!!! Reasoning dynamics of weight gain and loss

public and health professionals misconceptions System Dynamics Rev 2014

If you have the urge, here is the abstract, paper and citation. Pretty cool, HUH?

Abdel-Hamid T., Ankel F., Battle-Fisher M., Gibson B., Gonzalez-Parra G., Jalali M., Kaipainen K., Kalupahana N., Karanfil O., Marathe A., Martinson B., McKelvey K., Sarbadhikari S. N., Pintauro S., Poucheret P., Pronk N., Qian Y., Sazonov E., Oorschot K. V., Venkitasubramanian A. and Murphy P. (2014), Public and health professionals’ misconceptions about the dynamics of body weight gain/loss, Syst. Dyn. Rev., doi: 10.1002/sdr.1517

Abstract: Human body energy storage operates as a stock-and-flow system with inflow (food intake) and outflow (energy expenditure). In spite of the ubiquity of stock-and-flow structures, evidence suggests that human beings fail to understand stock accumulation and rates of change, a difficulty called the stock–flow failure. This study examines the influence of health care training and cultural background in overcoming stock–flow failure. A standardized protocol assessed lay people’s and health care professionals’ ability to apply stock-and-flow reasoning to infer the dynamics of weight gain/loss during the holiday season (621 subjects from seven countries). Our results indicate that both types of subjects exhibited systematic errors indicative of use of erroneous heuristics. Indeed 76% of lay subjects and 71% of health care professionals failed to understand the simple dynamic impact of energy intake and energy expenditure on body weight. Stock–flow failure was found across cultures and was not improved by professional health training. The problem of stock–flow failure as a transcultural global issue with education and policy implications is discussed.

Posted in Computational models, system dynamics, Systems thinking
Leave a comment

July 15, 2014

Is being liquid enough to innovate toward policy success?

Comparison of requirements of the liquid network to the mathematical requirements of Social Networks

Steven Johnson (2010), the popular author of “Where Good Ideas Come From- The Natural History of Innovation”, wrote that innovation is structure based. A cornerstone of Johnson’s (2010) book is this idea of an informal, liquid network in which ideas collide. Johnson (2010) visualized the liquid network as the Hogarth’s painting “Humours of an Election”. Ideas generate from informal meetings not that company annual retreat in Florida ac-cording to Johnson (2010). The closer the people’s ideas are in space and time, the better the chance for the ideas to collide with a few new (idea) bonds formed along the way. Johnson’s use of network would better be labeled as idea management rather than being complexity driven. But for policy to be able to account for the long-term implications of idea management, ethical deliberation drives the complexity of organization. The included table compares the offerings of the liquid network of Johnson (2010) to the social networks for idea management. Collision unfortunately lacks the formality and accountability afforded by the structural relationships of social networks. Broadening idea generation beyond a “bump” illustrates the immensity of the effect of the social and political reach of policy.
People connect within a social space. According to Alba and Kadushin (1976), the measure of proximity is concerned with pairs of individuals and how “distant” they are. Traditionally, proximity is all what diffusion of social commodities, often defined by information sharing, as a flow (Alba and Kadushin 1976). In a liquid network of Johnson’s conceptualization, ideas bump into each other due to the act of discourse as an event. First, there must be connections to other people and the environment supports the development of ties. Johnson concurs. There are the things colliding in a network which in Johnson’s (2010) case is the idea. Discourse turns that “private solid state to a (public) liquid network” (Johnson 2010). In what Johnson (2010) called the state of adapting exaptation, people can informally engage due to the “bump factor” of these chance meetings (Johnson 2010). To Johnson (2010), innovations linger in a “slow hunch” countering the eureka moment. The power of leverage points values the ability to find a small change that can make a huge impact. This is not the same as the epiphany. Systems that undulate slowly can be most frustrating to the urgency placed on having policies work.
The liquid network is mildly bound enough to be a network but not enough to strangle innovation with rigidness. Ideas are innovative because of the people that offer them with their skills and prestige. Some ideas are drowned out or flatly ignored. Connecting the ideas as a structure is often descriptive. Which idea (or person with the idea) was most connected? Which idea structurally demonstrated power in the network of ideas? Then this whole idea of playing with private ideas within a public space by basing it on the structure of the irrational agents is the more interesting, and fruitful question to pose.
Johnson (2010) proposed that platforms, or orchestrated spaces, may foster innovation. Within these platforms, liquid networks which do not restrict movement within an organization and their innovation ideas may be better supported to thrive. The ideas take on a fluid state without outside restriction. Does the event become more consequential than the interaction supported by the event? Networks are systems in which the nodes (the things that are connected) are connected by ties. When one is interested in reasons for why people affiliate over an idea, two mode networks could be employed (Borgatti and Everett 1997). The people would be agents (set 1 mode) while the IDEAS serve as a “series of events” (set 2 mode) that they are sharing. In a two-mode network, both the idea and agent indicates the elements of the network. The idea becomes an element to be weighted with the agents and chance meetings. So there is a tie be-tween the agent and the event. The tie formation could be due to “bump factor”:
1) to accept the chance meeting and
2) engage in the innovating chance meeting.
Both nodes engage in it since Johnson (2010) is basing adapting exaptation on engagement not its possibility. But a concern with approaching such two-mode networks, which mode takes the primary position? For instance, Johnson (2010) does not divorce the person from the idea and but for his paradigm the idea is element of focus. The idea is the primary node. The idea comes from the person (secondary node). The argument would also be made that if innovation is viewed as a dynamic, social experiment, the person as the primary mode may say more about why that idea even came to be in the first place. While the platform may be nature be useful in revisiting existing resources such as ideas in a liquid network, the problem often uncovered in policy is the lack of new ideas and continued reliance on existing ideas that may lead to attribution errors. To Johnson (2010), the ideas are used as bonding agents. Two-mode networks deal with the nature of affiliation. The idea is no longer just assumed but is now used as a means to investigate the structure in which ideas are embedded within the ties connecting the author of the idea (see Smith and Christakis 2008). The approach of Johnson (2010) had ideas brought “together” by its messengers as one frame. I do not believe that Johnson (2010) is marginalizing the person behind the idea of the liquid network. By stressing the network, the structure at two mode levels appears to capture the reality of how innovation in a liquid network more likely occurs.
To Johnson (2010) removing boundaries to the liquid network leads to chaos. This may be wise as any chaotic system does not return back to original state before the spiraling began. This chaos could hamper innovation by changing the new starting point and disrupting the momentum started in creating the innovative (see Feigenbaum 1983). As Johnson (2010) was intending avoiding chaos for a malleable liquid state, there is no history of events connected with a state of chaos. When things are chaotic, the past matters far less, only the energies running amok are important. What significance might this hold for developing innovation? Unlike chaotic systems, complexity is tied to its history of events upon which the system was built (see Buchanan 2000).
What is applicable to policymakers under liquid networks, according to Johnson, is less promising. In offering the example of Howard Dean’s Presidential Campaign “supernova”, Johnson (2010) admitted that “political leadership involves some elements (decision making and oratory) that aren’t best outsourced to a liquid network.” So how do policymakers innovate and think creatively under political stressors? Using Wasserman and Faust (1994) taxonomy, two-mode networks have three possible purposes under Johnson’s (2010) liquid networks:

(1) The affiliation is the tie of the person/ideas (node) to the informal chance meetings (event).
(2) The “bump factor” (if it supports interaction) may make interaction, therefore innovation, more likely.
(3) In order to framing the adapting exaptation realistically, the measurement of homophily is required.
The two mode network is difficult to interpret, let alone analyze. But how ever said that policymaking was easy? People tend to clique around policy issues, perhaps based on the process of idea generation or even bunch out due to personal access due to co-location. Two modes capture this bunching up as cliques that overlap, perhaps by party affiliation or side of a legislative issue. But the point of this methods tale is not the complicated analysis but that that coming together to work policy out that on the surface can be deceivingly simplistic. This is a great example of how the mental model of a liquid network does not apply neatly to policymaking but still has a place in the discussion. Unlike the informal liquid network, the innovation required by policy must translate into proper strictures under governance. The chaos of a liquid network must be operationally linked to the structural nature of complexity. But this fact does not mean that networks, whether liquid or mathematically understood, are not useful in discerning policy. While the informal ideas that are eventually brought to the policy table may utilize adapting exaptation, the policymaking process strips away the lacquer of casualness.

References

Alba R, Kadushin C (1976) The intersection of social circles- a new measure of social prox-imity in networks. Socio Meth Res 5(1): 77-101
Borgatti S, Everett M (1997) Network analysis of 2-mode data. Soc Network 19: 243-269
Buchanan M (2000) Ubiquity- The Science of History…Or Why the World is Simplier Than We Think. Weidenfeld & Nichols, London
Feigenbaum M J (1983) Universal behavior in nonlinear systems. Physica 7: 16-39
Johnson S (2010) Where Good Ideas Come from. Riverhead-Penguin, New York
Smith K, Christakis N (2008). Social Networks and Health. Annual Review of Sociology 34: 405-429