A recipe for systems thinking and modelling

Problem-oriented thinking and analysis can be summarized by the acronym P’HAPI denoting the five steps: Problem, Hypothesis, Analysis, Policy and Implementation. To make it easier to remember, pronounce it like “Be Happy”. Having such a recipe, helps you be less confused, more goal oriented, more productive, and hence more happy. Discuss the steps with your fellow students. Keep P’HAPI in mind whenever you get confused about the process of systems thinking and modeling or when writing an abstract or a report.

P’HAPI is consistent with other formal methods of analysis:

P’HA: represents the scientific method

P: represents policy analysis and operations research

I: represents implementation: management or public administration

P’-Problem

A problem is something of great importance to stakeholders and represents a state of affairs that is different from what is to be desired. The problem reflects a discrepancy (gap) between the system state and the goal. A dynamic problem is a problem that develops over time. Typically, we are all concerned about how things are changing and how they may change. We refer to an observed or a hypothetical future problem behavior as a reference mode. As a first step, the reference mode can be expressed in words. For instance, an animal population has been observed to deplete faster and faster, and we are afraid that it will continue to decrease in the future. This reference mode indicates the time horizon for the problem development, whether the problem develops quickly or slowly. The reference mode also indicates a boundary for what system parts need to be considered to explain and understand the problem development. The reference mode helps focus your thinking on the problem at hand – and nothing else.

H-Hypothesis

The hypothesis is your ideas about the cause-and-effect relationships that give rise to the problem behavior (the reference mode). In order to test the hypothesis more thoroughly, it must be expressed in terms of structure diagrams or a formal simulation model. It is recommended to start with a simple, understandable hypothesis that captures the essence of the problem. For instance, the population size has exceeded its carrying capacity and the system has passed a tipping point such that a reinforcing feedback loop dominates the system behavior.

A-Analysis

The main purpose of the analysis is to test your hypothesis: is it likely to explain the reference mode? Are the hypothesized feedback loops likely to explain depletion over time? Does graphical integration help you check your ideas about what explains the problem behavior? If such systems thinking leave you in doubt, model simulations should be considered. Independent of what method you use, there are important questions to ask yourself. Are the hypothesized cause and effect relationships consistent with available knowledge? Does your hypothesis really explain the problem behavior? Can you rule out alternative hypotheses? Should the hypothesis be reformulated and improved? Should the hypothesis be rejected? This process of repeated formulation and testing is called learning. If you reach a reliable and useful explanation for the problem, your hypothesis becomes your theory that you can use when exploring policies to deal with the problem.

P-Policy

A policy is a set of rules that transform information into actions. Policy proposals can come from stakeholders or from analysts including yourself. The policy analysis is similar to the hypothesis testing. Now the focus is on new policies to see if they perform better than the historical policies.

There is one important and subtle difference between hypothesis testing and policy testing. While stakeholders tend to agree that the original problem is a problem, they are much less likely to agree about the benefits and costs of alternative policies. Hence, the policy analysis typically requires that you to find out more about what stakeholders care about. If you learn that they care about a particular “side-effect” of some policy proposals, you may have to revise your theory to make sure that it captures this “side-effect” as a real and expected effect. When this effect is captured by the model, stakeholders are likely to be more comfortable with your analysis.

I-Implementation

The policy analysis can be seen as a theoretical exercise where we search for first best policies. Diffusion analysis have found that decision-makers are more convinced by observable, positive experiences than by complex theoretical knowledge. It also matters if policies are compatible with existing systems and can be tested locally. Effective communication is essential for implementation. Analysts need to learn about issues that they have not considered. Decision-makers need to rethink their current ways of doing things and become aware of their own biases regarding problem definitions and understanding. Being aware of the ideas and assumptions that are not usually communicated, can help the implementation process. If “world views” change, anything can happen. However, in practice one will typically settle for a minimum of mutual understanding.

Closing remark

Now that you have seen all the steps in P’HAPI, go back and notice how each point naturally leads to the next. The problem requires an explanation (the hypothesis), the hypothesis must be thoroughly tested (analyzed), then the acquired knowledge (theory) is used to test alternative policies to avoid problems, and finally these policies are considered for implementation. This said, be aware that systems thinking and the modeling process is iterative: what you learn at one stage may have implications also for earlier stages. Typically, the analysis in part A leads to changes in the hypothesis H, and may even lead to changes in the problem definition P’. Contact with stakeholders for the purpose of policy implementation may call for extensions of the model (theory) to acknowledge “side-effects”.