Nurtured Choice

What can you do to influence someone’s decision, if you cannot give them advice? In short, a possible approach is to take actions that satisfy two conditions:

1. Any communication that the action may result in, cannot be interpreted as advice by the decision maker. A simple case is when any information that the action leads to never reaches the decision maker. Another case is when some of the information that actions lead to is such that the decision maker sees no relationship with the decision they need to make.
2. The consequences of actions should change the environment in which the decision maker is defining the decision problem, whereby the changes should result in information that leads the decision maker to add or remove elements of the decision problem.

I call this the nurturing of choice. Instead of providing advice that is clearly directed at the elements of the decision problem (as I discussed in my book “The Analysis and Design of Advice“), the advisor takes actions that shape the environment that is assumed to directly influence the definition of the decision problem.

Which elements of the decision problem can be influenced by nurturing choice?

Suppose that you need to advise a decision maker who is rational in the sense of classical decision theory [1], and therefore makes decisions following the basic method of decision analysis [2]. Moreover, suppose that your advice cannot directly influence the elements of the decision problem, or equivalently, that anything you directly communicate to them (i.e., the verbal and written advice you may give) cannot influence the definition of the decision problem that that person or group came up with by applying decision analysis.

In classical decision analysis, a decision is the commitment to one of considered options, after comparing the anticipated consequences of these options over specific properties, acting as decision criteria. Decision analysis is concerned with proposing processes to collect information on, or define a decision problem using a small number of core concepts (i.e., which cannot be reduced to one another; none can be defined using the others).

When making a decision, the decision maker need information on:

• Goals: Situation considered desirable, or here, equivalently, partial description of a state of the world, considered desirable.
• Constraints: A Constraint is a statement which creates a distinction between two sets of Situations, namely those in which the Constraint is satisfied and those in which it is not. Contrary to a Goal, a Constraint is normally not associated with desirability – it is a given and it does not matter if it is desirable or not.
• Criteria: A Criterion is a property used to compare Consequences.
• Options (Alternatives): Mechanism by which Situations that do not satisfy Goals are changed into Situations which satisfy Goals; i.e., how to satisfy Goals.
• Consequence (of an Option on a Criterion): Situation brought about by an Option and its mapping to a value of each Criterion.
• Preference: Order over (all or a subset of) values of a Criterion, used to describe relative desirability.
• Importance: Order over (all or a subset of) Criteria, used to describe relative importance of criteria.
• Utility: Numerical representation of Preferences. 

The idea in the nurtured choice approach, is that the advisor needs to look for the factors that they believe determine any of the goals, constraints, criteria, and so on, that they expect the decision maker to come up with. Then, any action that the advisor takes should be aimed at influencing these factors, instead of directly being about these elements of the decision problem.

For example, if the advisor assumes that goals are a function of values that the decision maker wishes to be associated with, then working to emphasise specific values, and downplay others, is a possible way to nurture choice by influencing the formation of goals. Another approach is to work to influence the process by which goals are formed, especially if the decision maker forms goals through consultation of known stakeholders, in which case working with the stakeholders can influence the shaping of goals, by influencing the perceptions of feasibility, for example.

A model of choice, such as the one in classical decision theory (as above) is also an assumption. The advisor is the one making that assumption, that is, the advisor will have to make an assumption about how the decision maker is structuring the decision problem, in order to nurture choice. Even in a silly extreme case, of a decision maker who randomly picks an option, the advisor can still make assumptions about how the decision maker learns about which options are available; in such a case, nurturing choice involves influencing available options, and the fewer there are, the higher the probability that one of those will be picked relative to having more options. 

The point from all above, is that nurturing choice involves 

1. making and validating assumptions about how the decision maker thinks about the decision problem, 
2. making and validating assumptions about factors which influence the elements of the decision problem that will be available to the decision maker, and 
3. designing and taking actions which influence the factors mentioned above.

In many situations, however, this is the only feasible approach. 

References

1. Schoemaker, Paul JH. “The expected utility model: Its variants, purposes, evidence and limitations.” Journal of economic literature (1982): 529-563.
2. Keeney, Ralph L. “Decision analysis: an overview.” Operations research 30.5 (1982): 803-838.