You are on page 1of 12

IIDM: Summary

Individual decision making:


• Decision under uncertainty

• Weights on criteria (multi-criteria analysis)

• Two types of thinking

• System 1: fast

• System 2: slow

Analytical/rational decision-making Model: PrOACT

Based on ve consecutive steps:

1. Identify the problem: frame

2. Clarify the objectives: values

3. Generate alternatives: options

4. Evaluate consequences: goals

5. Make tradeo s: interdependencies

Decisions under uncertainty (Tversky & Kahneman)

There are two types of thinking: fast and slow.

Fast thinking is automatic, e ortless, associative, intuitive, and generally generates accurate
short-term predictions.

Slow thinking is deliberate (done consciously), e ortful, ordinarily, and logical, it creates a feeling
of agency.

Fast thinking is prone to biases that slow thinking can help to detect and correct, but ego
depletion1 (tired mind) can hamper this process.

Biases: Cognitive ease

“Cognitive ease or uency is the measure of how easy it is for our brains to process information.”

Familiarity, certainty, coherent stories and good mood take us into a state of “ease”, which
translates into feeling good and true, the mind does not escape fast thinking (called system 1).

Associative machine: depending on what you feel/hear/see, you recognize & associate
di erently, this e ect is called priming2. (It causes con rmation bias: the brain looks for
con rming evidence (system 1) instead of the opposite (system 2))

The “Halo” e ect3: if you (dis-)like a thing, you tend to (dis-)like everything about it

What you see is all there is: fast thinking creates a story based on what is in view, without
considering what is missing.

1When the energy for mental activity is low, self-control is typically impaired, which would be
considered a state of ego depletion.
2Priming is a phenomenon whereby exposure to one stimulus in uences a response to a
subsequent stimulus, without conscious guidance or intention.
3 The halo e ect is a type of cognitive bias in which our overall impression of a person in uences
how we feel and think about their character. Essentially, your overall impression of a person ("He
is nice!") impacts your evaluations of that person's speci c traits ("He is also smart!"). Perceptions
of a single trait can carry over to how people perceive other aspects of that person.
1 of 12
ff
fi
fi
ff
ff
ff
ff
fl
ff
ff
fi
fi
fl
fl
Seeking causality: “system one” sees the world as cause-e ect scenarios, this might lead us to
see causalities where there are none.

Availability bias: We de ned the availability heuristic as the process of judging frequency by “the
ease with which instances come to mind.” The availability bias substitutes one question for
another: you wish to estimate the size of a category or the frequency of an event, but you report
an impression of the ease with which instances come to mind. Substitution of questions inevitably
produces systematic errors.

To counter the biases:

• Engage system 2 (slow thinking)

• Ask questions to probe what’s missing (con rmation bias, WYSIATI,…)

• Create a base rate (average of similar past events) and don’t anchor to what is told to you

• Evaluate alternatives in a table ( rst criteria for all alternatives, then next criteria,…)

This approach cannot be applied in a inter-/intra- organizational decision-making process


because problem formulation&knowledge might be contested! The aim of a negotiation is to nd
an acceptable solution and not necessarily the optimum! And the process matters as much as the
content of the decision!

Analytical/rational/project perspective: have the goals been realized at the lowest possible
cost?

Political/negotiation/process perspective: have goals been realized? Do parties perceive


results satisfactory? Was it a fair process? Have lasting relationships been developed (for future
cooperations)?

Wicked problem: a problem that is di cult or impossible to solve because of incomplete,


contradictory, and changing requirements that are often di cult to recognize. It refers to an idea
or problem that cannot be xed, where there is no single solution to the problem; and "wicked"
denotes resistance to resolution, rather than evil.

Dilemmas in General Theory of Planning [Rittel,


Webber]

There are two kinds of problems: tame4 and wicked ones.

The wicked problems have ten fundamental characteristics:

1. For the wicked problem, the formulation of the wicked problem is the problem! Because the
process of conceiving a problem and a solution is identical, therefore without
identifying a solution you cannot look for the problem, and the problem is not de ned.

2. Wicked problems have no criteria to de ne when a solution has been reached.

3. Solutions to wicked problems are not true-false but good-bad.

4. There are no immediate or ultimate tests of a solution to a wicked problem, since the
consequences of the solution go on forever and create a wave of consequences, it is not
possible to judge the solution until all the consequences reach an end.

5. Every solution is a “one-shot” operation since there is no opportunity to learn through trial-
and-error every attempt counts signi cantly (there is no coming back from a solution).

6. Wicked problems do not have an enumerable (or an exhaustively describable) set of potential
solutions, nor is there a well-described set of permissible operations that may be incorporated
into the plan.

7. Every wicked problem is essentially unique.

8. Every wicked problem can be considered as the symptom of another problem.

9. The discrepancy in representing a wicked problem can be explained in numerous ways, every
choice in the problem formulation determines the nature of the problem’s resolution.

4A problem is tame if an exhaustive formulation can be stated containing all the information the
problem-solver needs for understanding and solving the problem (provided he knows his “art”).
2 of 12
fi
fi
fi
fi
ffi
fi
fi
ffi
ff
fi
fi
10. The planner has no right to be wrong: planners are liable for the consequences of the actions
they generate.

Network decision-making: Multi-issue game


• Intraorganizational: hierarchy required for coordination

• Interorganizational: there is no hierarchy, there is a network of stakeholders

- Negotiated process approach

- Share solutions for unshared problems

Network decision-making can be managed by:

• Broadening the agenda

• Inviting stakeholders to mention issues

• Creating a perspective of gain for each of the stakeholders

• The multi-issue agenda will make the process attractive for all the stakeholders

The result of this process will be no longer one single problem but a multitude of issues
attractive to every stakeholder.

This process will increase the complexity but will result in incentives to:

• Participate in the decision-making process

• Take ownership of the nal decision

• Give and take (and share information)

• Show cooperative behavior

• Learning processes, about power and content

• Exercise peer pressure

A project approach will not work!

Project vs process mindset:

• Content vs power

• Problems vs interests

• Factors vs actors

Project Process

Problem Problem to solution (delineate) Solution to problem (broaden)

Goal Precise Broad, vague

Structure Reduce complexity Increase complexity

Information Facts and gures Negotiated knowledge

Decision making and Detailed Open


implementation

Evaluation E ectiveness Future relations

Management in Networks: Process not Project

In networks, an actor who de nes a problem must be aware that the problem doesn’t exist. There
is ‘only’ problem perception, and the question is whether other actors have the same perception
of the problem, therefore a project approach cannot be used, as there is no uniform problem
formulation!

So what the actors must do is to in uence the perception of the problem!

There are several matrixes in the articles useful to describe the di erences.

3 of 12
ff
fi
fi
fi
fl
ff
Data-driven decision making in a multi-actor
Law enforcement stages:

1. Norm setting

2. Gathering information

3. Judging information

4. Enforcement

Law enforcement is a big problem: too much enforcement can reduce the satisfaction of the
population and too little will cause tangible consequences.

Enforcing the law itself is a dilemma:

Just decision is “allow” Just decision is “sanction”

Decision is to allow OK Under-enforcement

Decision is to sanction Over-enforcement OK


The dilemma arises when we have no consensus on norms and no objective information available,
therefore the problem is de ned as “unstructured”.

A “tragedy” happens to people in a multi-actor context that face dilemmas and have con icts with
others because they do their job well.

Using data in law enforcement leads to two developments:

• Automatic law enforcement

• Automatic risk analysis

These two developments are based on the premise that AI leads to better risk analysis and better
risk analyses lead to better enforcement.

Data-driven decision-making is an example of a multi-actor context, in which generating data,


interpreting it, and deciding are done by di erent actors (multi-actors).

In each of these phases, there can be a hierarchy in the decision making and this can cause
dilemmas about the “just” decisions.

The value con ict matrix shows:

Coupled Decoupled

Explicit

Implicit
Coupling and decoupling are opposite responses to value con icts.

Coupled coping means dealing with values in relation to one another, so coupling always
manages multiple values in the same con ict.

Multiple values are “hybridized” in one task.

Decoupling means addressing values separately, so having a mono-value response to con icts.

Decoupling is usually done to assign the responsibilities of di erent values to di erent


departments or individuals.

Explicit values can be measured through KPI or similar, while implicit values cannot.

Moreover, in this context, there are stakeholders with di erent logic!

Data analysts have an information logic, while policy-makers have a decision logic.

Information logic Decision logic

Role Supporting decisions Control information resources

Quality mechanism Variety Selection

4 of 12
fl
fi
fl
ff
ff
ff
fl
ff
fl
fl
Information logic Decision logic

Main choices Methodological Ideological (politics)

Pragmatic (operational)

Logic of appropriateness vs logic of consequences

The logic of appropriateness de nes a basis for decision-making biased toward what social
norms deem right rather than what cost-bene t calculations consider best. Behavior in a speci c
situation is said to follow from the rules that govern the appropriate course of action for a given
role or identity. The rules that determine appropriateness are institutionalized in social practices
and sustained over time through learning. The logic of appropriateness can provide an
organization with institutional order, stability, and predictability. At the same time, it may run
counter to democratic principles by implying the substitution of tacit understanding for collective
deliberation.

The logic of consequences evokes self-interested rational actors with xed preferences and
identities whose behavior is determined by the calculation of expected returns from alternative
choices.

To manage this decision-making process what can be improved are:

• People involved

• Algorithms used

• Institutions that facilitate managing con icts:

• Between facilitators and decision makers

• Between those representing algorithms and work- oor professionals

• Between carriers of values

• Processes of analysis, for instance via participative modeling

Rationality and Politics of Algorithms

Humans have “bounded rationality”, and because of this decision-making processes tend to be
small incremental steps where sometimes means prevail over ends, so decision-makers should
somehow control and make sense of their information source and account for the information
they apply.

Decision-making can have two views:

Rational view Political view

Focus on what: activities Focus on who: actors

Focus on prede ned process steps Focus on real-life transactions

Focus on common goals Focus on individual goals


The rational view represents a clear process in which big data can enhance the various steps in
which information is required. The political view represents an erratic, dynamic process, in which
political or other goals partially determine when, where and how there is use for big data.

Table 3 in the paper is interesting but extensive to explain.

Equality, diversity and inclusion in an organization


Kurt Lewin’s unfreeze-change-refreeze model is used as a framework for organizational change.

This involves rstly unfreezing – where organizations stop and recognize that there are problems
that can be addressed and should be addressed.

They then move onto changing behavior in the organization. This involves drawing on research
about what works, giving people the tools that they need to make better decisions, and
experimenting with a range of approaches to improve outcomes. Finally, they refreeze that new

5 of 12
fi
fi
fi
fl
fi
fl
fi
fi
behavior and undergo a continuous process of improvement, improving how data is collected and
drawing on evidence to understand what works best for them.

Interorganizational decision making: Value multiplicity:


Re ections on stakeholder engagement
Intraorganizational decision-making (how organizations make decisions):

Interorganizational decision-making has speci c stakeholders that interact together without any
hierarchy (value multiplicity), there are many issues and one solution.

Intraorganizational decision-making has a hierarchy (value singularity), this greatly in uences the
decision-making process.

There are stakeholders trying to coordinate other stakeholders on di erent layers (e.g. higher
manager to lower management).

Since there is a hierarchy, the values cascade from the top to the down, if things go wrong there
is an escalating e ect from the bottom to the top of the organization.

There is one problem and one solution.

When there are problems in the organization (e.g. in an engineering environment speci c
engineers have an idea of what is needed, while other engineers have an opposite vision, the less
knowledgeable management might not have the capabilities to understand the best option so the
negotiation moment cannot happen in the top of the hierarchy) the hierarchical organizational
structure fails and turns into something similar to the network negotiation (more structured
but similar).

Organization structure decision-making like:

• Phase model:
• Feasibility

• Cascading from the management to give input

• Concept

• Escalating from the engineering to the management

• Design

• Realize

The cascading and escalating processes happen between every phase, and each phase ends
with a “decide” moment.

Among all these phases alternatives, consequences, objectives, and risks are discussed,
analyzed, and managed.

Risk management:
Risks are “opinions” on uncertainty, they can be managed through:

• Prioritization (Matrix)

• By making risks actionable (Bow tie)

• The problem here is that impacts are subjective! It’s di cult to judge impacts especially if the
project is new (very risky)

Stakeholders represent a special risk:


Their risk lies in the unknown of stakeholders (How much will they impact your organization? Will
they go against your project?)

Two traditional ways to manage stakeholders:

• Manage stakeholders outside (can predict what they want and defend your decisions)
• Decide and defend

• Manage stakeholders inside (real engagement of stakeholders, cannot be done for all projects)
• Deliberate and deliver

Managers prefer the outside model.

Stakeholders: a person with an interest or concern in something, especially a business.

6 of 12
fl
ff
fi
ffi
ff
fl
fi
Examples are: government, NGOs, employees,…

We engage stakeholders because:

• Legal requirements

• Increase awareness

• Develop capacity building

• Promote social learnings

• Increase network, transparency and trust

• …

The rst thing to do when interacting with stakeholders: identify them with a stakeholder map.

After, the engagement process starts:

1. Stakeholder identi cation (who are they?)

2. Stakeholder characterization (what are their values?)

3. Stakeholder structuration: degree of involvement de nition (how much should they be


involved?)

4. Choice of participatory techniques (how do I engage with them?)

5. Implementation of participatory techniques (engage with them)

6. Evaluation (were the goals met?)

Stakeholders are classi ed in typologies based on


di erent criteria:

• Classi cation based on power, legitimacy and


urgency:

7 of 12
ff
fi
fi
fi
fi
fi
• Classi cation based on in uence:

• Classi cation based on power vs. interest grid:

• Classi cation based on power and support:

8 of 12
fi
fi
fi
fl
• Classi cation based on power, interest and attitude:

These frameworks can help you identify who is really in uential, but mapping the relationships
among stakeholders and the context is also important.

• Representation of the stakeholders and their relationships:

9 of 12
fi
fl
• Representation of the relationship among stakeholders and the strength of their ties:

For each stakeholder, a di erent engaging method is to be used based on how impactful the
stakeholder is! (e.g. survey, personal meeting, agreements,…)

Careful: the power of your decision depends on the quality of the stakeholders decision!

The stakeholder involvement can be seen as a process:

10 of 12
ff
It’s very important to build consensus among stakeholders, but it is not always possible
(sometimes even the law is a barrier).

It’s useful to go ahead and behind on questions to create consensus.

11 of 12
Knowledge co-production processes can be set up to increase cooperation and consensus:

“Iterative and collaborative processes involving diverse types of expertise, knowledge and actors
to produce context-speci c knowledge and pathways toward a sustainable future”.

To use both scienti c and societal approaches a mixed approach can be used:

This focus on stakeholders is done in the intraorganizational decision making because also in this
case there is a lot of negotiation in networks since the coordination sometimes lacks.

12 of 12
fi
fi

You might also like