Nudge and the manipulation of choice: A framework for the responsible use of the nudge approach to behaviour change in public policy - Hansen & Jespersen - 2013 - Article

The book of Thaler and Sunstein (2008) suggests that policy-makers of public policies should arrange decision making in a way that behaviour change is promoted. Both society and individual interests should be considered.  Yes, discussions concerning the acceptability of nudge-based policy to change behaviour have emerged.

The book of Thaler and Sunstein (2008) suggests that policy-makers of public policies should arrange decision making in a way that behaviour change is promoted. Both society and individual interests should be considered.  Yes, discussions concerning the acceptability of nudge-based policy to change behaviour have emerged.

What are the objectives of this study?

According to Thaler and Sunstein (2008) people are always influenced by the context of decision making and nudging is just taking the liberty of using this for influencing these decisions. On the other hand, criticism of nudging techniques arises with the argument that nudging works by manipulation of citizens’ choices. The aim of the current article is to illustrate that arguments on both sides are flawed. The dual-process theory is used to show the epistemic transparency criterion described by Thaler and Sunstein (2008) to show that nudging is not about manipulation nor influencing choice. They create a framework in which four types of nudges are identified. They can provide a central component for nuanced, normative considerations in creating social policies.

How should nudging be approached in behavioural change policy?

Nudge means that human decision making is not rational but influenced by several social and environmental factors. Decision making is often influenced by subtle, insignificant-seeming contextual cues. Thaler and Sunstein (2008) suggest creating policy based on ways to influence people’s behaviour taking these contextual cues into consideration while avoiding the pitfalls and problems of traditional approaches. The definition by Thalor and Sunstein (2008) of nudge is that if subtle changes in the context of decision making can lead us astray from our best interest, the insights of how that happens can ‘nudge’ people in the right direction for health, happiness and wealth. This low-cost technique of changing political measures therefore is highly attractive.

What are the responsibilities of designing interventions based on nudge?

According to Thaler and Sunstein (2008) an architect of choice should take responsibility for the organization of context in which people make decisions. They suggest that these people should view themselves as choice architects. In this context, many people are considered choice architects. A policy-maker who adjusts the range of choice options is just as much a choice architect as someone who adjusts the choice structure. A choice architect should choose a way to design the context in which people make decisions and that he is responsible for. They also note there is nu such thing as a neutral design. It is impossible to be an agent in people’s decision making without influencing people’s behaviour and choices. The responsible thing to do is to incorporate and recognize the knowledge on nudge when designing the choice architecture, he is responsible for. According to the pro-nudge position it seems that nudging for his own benefit is always permitted. No one is forcing citizens or limiting their options to choose otherwise.

What are criticisms on the nudge approach?

Criticism of the nudge approach are that people are manipulating choices. It is said that nudging impairs people’s autonomy and ability to make our own moral choices. However, the characterization of nudging as such relies on the theory of agency that people assign to themselves. Policy recommendations that rely on this agency theory and the characterization as manipulation of choice is discussed in this article.

Why is the anti-nudge position not a literal non-starter?

Thaler and Sunstein (2008) describe nudging as the choice architecture that changes people’s behaviour in a way that no other options are forbidden or significantly changing their economic inventive. They argue the anti-nudge position is a literal non-starter because choice architecture may exist without the existence of a “nudger”. They argue that a neutral design of choice architecture does not exist even when the observed effects happened by accident. Even though architects do not intend to nudge towards particular ends, we are nudged to consequential and predictable behaviour even if a choice architect did not mean to. This is according to the position that nudging is a literal non-starter. However, according to the authors it is not. Nudging cannot be justified just because there is no neutral design for it. It is not a literal non-starter for the following reasons:

• The argument that neutral designs do not exist gives the pro-nudge approach an attractive defence. The argument says that we are always being nudged, regardless of anyone’s attempts. It’s argued that nudges are an inescapable effect of any type of decision making. It is supported by the observation that nudging does not promote behaviour that alters freedom of choice. Also, someone is always able to reject the behavioural change.

• The condition that is constructed for the literal non-starter is if a neutral design does not exist, this means we are always being nudged. Yet something important is being overlooked here. Someone can both unintentionally or intentionally influence behaviour. The distinction between these is important because of the notion of responsibility. Therefore, a nudge should be described as by Hausman and Welch, that it is the intentional attempt to influence one’s choices.

• The ends and means of libertarian paternalism means that certain ends and values should be promoted in influencing choices. Policy makers should nudge to promote things in interest of citizens. It seems preferable to any other alternative and serves in the interest of citizens.

• Finally, the principled freedom to choose differently is an important consideration in the debate. The fact that we have insights in the human fail ability in relation to decision making does not justify any choice intervention. The realistic definition of human agency should be taken into consideration, meaning that people do not always decide according to their reflected preferences. When people are nudged into a certain direction, they might lose the ability to choose otherwise.

Does nudging manipulate freedom of choice?

Nudging is described as the manipulation of choice, not behaviour. Because these concepts are not interchangeable, a further classification of the concept is needed. Therefore, the dual-process theory, actions and causes and types of nudges are discussed.

• The dual process theory states that the human brain distinct between two ways of human processing. One is intuitive, automatic thinking (fast thinking) and the other is rational, deliberate thinking -or reasoning (slow thinking). Kahneman refers to these two ways of thinking as system I and system II thinking. Automatic thinking is often uncontrolled, effortless, fast, associative, skilled and unconscious. Reflective -and rational thinking has the characteristics of effortful, slow, controlled, rule following and self-aware. The point of the theory is to describe that behaviour can result from either modes of thinking.

• Actions and causes are often described in terms of state of the world meaning that an agent intentionally wants to bring about. There is an active deliberation that is determined by the available course of action depending on the situation. According to actions and causes a choice comes from the process of reflective thinking. A non-voluntary action is not described as a real action, but only a event that happens to you which was not, but could have been controlled.

• Two types of nudges, and only one aimed at reflective reasoning or choice. Even though both types of reasoning reflect automatic thinking processes, only one reflects on behaviour changing reflective reasoning. Type I nudges influence how behaviour is maintained by automatic processing without thinking reflectively about the consequences. Type II nudging is nudging to influence actions, behaviours. This is done deliberately by limiting the set of choices for example.

Based on these notions, a new definition of nudging is described: the behavioural influence in a predictable way, without prohibiting courses of action that were previously available, making alternatives costlier in terms of social sanctions and time.

What is the view of Thaler and Sunstein on transparency in nudging theory?

In democratic policy making, the transparency of measures is an important issue. The concepts deliberation, consent, respect and accountability should be considered. It should be clarified to what extend nudging works by manipulating choice or behaviour. Transparency is a measure of guiding principle according to Thaler and Sunstein. They derived it from the publicity principle, which means that the government should be able to protect its citizens. They adopt the principle for two reasons:

1. The government could be embarrassed if the policy and its grounds are disclosed.

2. The government should have respect for the people it governs.

The conclusion of Thaler and Sunstein is that this publicity principle is a good guideline. Even though Thaler and Sunstein advocate strong transparency, which means that the government should never be secretive about what they are doing in terms of nudging. Also, they ban subliminal priming as an alternative. Nudging should be manageable and disclosed to the people that are subject to it. However, the principles of subliminal priming are in accord with people’s moral intuitions. This means that it becomes restrictive and may even lead to an ethical paradox when applied to the nudging approach. Thaler and Sunstein also suggest an epistemic dimension of transparency, they argue that nudging is not necessarily about manipulation choices and behaviours but should be divided into transparent and non-transparent. Some choices might be objectionable, they are not visible and impossible to monitor.

• Transparent nudges are provided in a way that the intentionality and the pursued behavioural change is visible.

• Non-transparent nudges are nudges where the person is unable to reconstruct the intention of means by which the change in behaviour is constructed.

This notion of epistemic transparency can be used as a criterium for the evaluation whether nudging is a form of manipulation. The psychological sense of manipulation is the intention to change perception, behaviour or choices through deceptive or abusive tactics. The transparency of the nudge can determine its effectiveness. Nudges in the dark tend to be more effective than obvious ones.

What is the effect of nudging on the manipulation of choice?

Given the two types of nudging, transparent vs non-transparent, four different nudges can be distinguished. Some nudges will fall into a grey zone or will be difficult to qualify. However, the matrix is a valuable guide and basis for the typology for different types of nudges. It can also help with policy recommendations.

1. Transparent type 2 nudges are in the top left corner of the matrix. This intervention type is easy to reconstruct for the citizen. The intentions for the behavioural change are very clear, just as the way the reflective system is engaged. A popular example is the seat-belt alarm in a car.

2. Transparent type 1 nudges are situated in the bottom-left corner. Reflective thinking does not cause the behavioural change but happens as a side product. This makes it easier to reconstruct the way in which the nudge is used to influence behaviour. An example is playing relaxing music while people are boarding an airplane.  

3. Non-transparent type 2 nudges are in the matrix at the top-right corner. The success of the nudge depends on the unconscious activation of reflective thinking. People should be unable to reconstruct the incentive of the nudge. Subtle cues are used to activate reflective thinking for making specific choices. One example is the lottery where people start to overestimate the chance of winning.

4. Non-transparent type 1 nudges in the bottom-right corner resemble nudges that cause changes in behaviour without engaging the reflective thinking system. It is not likely the person will be aware of it and will neither be able to explain or reconstruct the nudge afterwards. One example is an experiment done by Wansink described in “mindless eating”. By reducing the size of plates in a cafeteria to a 10-inch dinner plate instead of a 12-inch one, people ate 22% less calories.

Thus, the facilitation of a consistent choice in a transparent manner is an example of system II thinking. Also, the non-transparent manipulation of choice. System I thinking is the transparent influence of behaviour and the non-transparent equivalent of behaviour.

What is the proposed framework for the responsible use of nudging?

Given the typology for nudging in the previous section, the authors developed a guideline for applying nudging in the creation of policy. It provides a framework based upon which you can decide whether a nudge is manipulative (non-transparent) or based on choice (reflective thinking).

1. Transparent type 2 nudges are not meant to influence people by psychological manipulation. They are used to promote decision making that can be seen by the citizen. They are in line with the preferences of people and provide feedback. They are facilitating the freedom of choice. The autonomy of people is protected. It can be explained as prompting of reflected choice.

2. Transparent type 1 nudges do not try to influence people’s choices, they influence automatic behaviour transparently. These nudges do work within the sense of manipulation and influence behaviour. Citizens are not free to ignore the response that the nudge causes. The principle of freedom is there, but most people are not able to ignore the consequences. This is because it works on automated behaviour.

3. Non-transparent type 1 nudges use psychological and mechanical manipulation. They cannot be regarded transparent or one of choice. They influence automatic behaviour without someone being aware of it. They operate outside of the radar. People can only avoid them by meaning of principle. Therefore, they are paternalistic and described as manipulating behaviour.

4. Non-transparent type 2 nudges are manipulating choice because it works with psychological manipulation to get people to unconsciously think or act a certain way. It does not influence autonomous decision making and are cases of straightforward paternalism. It is difficult to ascribe the responsibility to decision-makers because their actions are the result of manipulation by the architect of the choice.