Module 8 - Applying Behavioural Economics Tools to Improve Decision Making Flashcards
Describe the role of nudges.
A “nudge” is any aspect of the choice architecture that alters people’s behaviour in a predictable way without forbidding, penalizing or rewarding any options or significantly changing their economic incentives (i.e., it achieves behaviour change without actually changing the set of choices.). Perhaps the most frequently mentioned nudge is the setting of default rules, which are preset courses of action that take effect if nothing is specified by the decision maker. To count as a mere nudge, the intervention must be easy and cheap. Nudges are not mandates. Putting the fruit at eye level counts as a nudge. Banning junk food does not.
Explain the dimensions that can be used to classify nudge characteristics.
(1) Boosting self-control vs. activating a desired behaviour: The first dimension looks at whether a nudge is designed to boost self-control and help us follow through with a decision (such as contributing to a retirement plan). With certain behaviours, such as saving money or exercising, there is a discrepancy between what we would like to do and what we end up doing. Nudges that help boost self-control will correct this discrepancy.
(2) Externally imposed vs. self-imposed: The second dimension considers whether a nudge will be voluntarily adopted. Self-imposed nudges are voluntarily adopted by people who wish to enact a behavioural standard that they feel is important. Such nudges may include practices such as voluntarily asking for a reduction on one’s credit limit.
(3) Mindful vs. mindless: The third dimension considers whether a nudge will guide us to take a more cognitive, deliberate approach to decision making and remove some of the effects of the often unconscious behavioural influences present in the context or whether it will guide us toward a more automatic, implicit approach that utilizes well-established behavioural influences or heuristics.
(4) Encouraging vs. discouraging: The fourth dimension considers whether a nudge encourages or discourages behaviour. Encouraging nudges facilitate the implementation or continuation of a particular behaviour. Discouraging nudges, on the other hand, hinder or prevent behaviour that is believed to be undesirable.
Explain the significance of mapping the decision context in designing an effective nudging strategy.
Mapping or auditing the decision context (i.e., the decision-making process of the end user) is the first step in the process of designing an effective nudging strategy. This requires an analysis of the context and the task (i.e., how do we make decisions, what are the typical circumstances in which we do that, etc.) followed by identifying the key heuristics and influences that may affect the decision outcome. Auditing the decision-making process identifies factors that prevent us from following through with our intentions. These factors (“bottlenecks”) represent areas where a nudging strategy might yield quick dividends.
Outline aspects of the end user’s decision-making process to address in a context audit.
The four different aspects of the end user’s decision-making process to address are:
(1) The properties of the decision, including understanding the incentives and motivations associated with the decision, how much attention the decision receives and identifying the choices presented, especially the default option
(2) Information sources and how information related to the decision is gathered and presented
(3) Features of the individual’s mindset and whether emotions influence the outcome of the decision
(4) Environmental and social factors, such as peer pressure and lengthy application processes.
Outline factors to consider when prioritizing nudge selection to address bottlenecks in the decision-making process.
While it is always possible to combine nudges, when several nudges have been identified as being possible intervention devices, it is useful to prioritize. Factors to consider include:
(a) Operational costs associated with implementation
(b) What bottlenecks the nudges address. Prioritize based on where the bottlenecks lie in the decision-making process. Choose nudges that resolve bottlenecks that are further upstream in the decision-making process.
(c) Relative reach. Self-imposed nudges such as precommitment may not reach as many people compared to defaults or automatic enrollment. Although it may be in our best interest, we may not want to make an up-front commitment.
(d) Interventions like automatic enrollment have a high adoption rate but lead everyone to accepting the same terms and benefits. An automatic enrollment program, for example, may require an individual to contribute $200/month to a predetermined retirement savings plan. A significant portion of the target audience may not benefit from such a program, perhaps because the contribution amount is too high or because the investment fund does not match their risk appetite. What may be preferable is to allow us to determine our own contribution amount and select from a small assortment of investment funds. Determining whether segments of the target audience have different behavioural preferences will provide answers to this issue.
(e) The long-term effectiveness of the nudge and whether the intervention could lead to the development of new, more beneficial habits.
Explain the role of process and evaluation outcomes in testing the effectiveness of nudges.
An outcome evaluation confirms that the nudge has produced the desired outcome. For instance, an outcome evaluation of the “planning aid to increase savings participation” would simply demonstrate that people who were randomly assigned to have access to the nudge participated at a greater rate than people who did not. A process evaluation seeks support for the underlying mechanism. For instance, people who were nudged should report a greater ease in comprehending materials and a shorter time in completing the necessary forms. Ideally, the testing of nudges incorporates both a process evaluation and an outcome evaluation.
Explain the concepts of “choice architecture” and “choice architect.”
The way a choice is presented influences what the decision maker does. Much like the architect of a building who affects the behaviours of the building’s inhabitants through the placement of doors, hallways, staircases and bathrooms, “choice architects” (anyone who presents people with choices) can influence choice in many ways, including varying the presentation order of choice alternatives, framing, the order of attributes and their ease of use, and the selection of defaults. Choice architecture influences decision making by simplifying the presentation of options, automatically evoking particular associations or making one option more salient or easier to choose than the alternatives. Everyone, from a parent presenting a child with bedtime options, to a government providing its citizens with pension options is a choice architect. While it is tempting to think that choices can be presented in a neutral way, the reality is that there is no neutral architecture—Any way a choice is presented influences how the decision maker chooses. Consider a well-known example: All choice presentations have a (usually implicit) default, even if the default is that no choice is made, preserving the status quo. The default option will be chosen more often than if another option is designated as the default.
2 Describe the two categories of tools available to choice architects.
The two categories of tools are:
(1) Tools used in structuring or setting up the choice task (i.e., determining what to present to decision makers), such as using the optimal number of options or using technology to assist in the choice task
(2) Tools available for describing choice options (i.e., how to present to decision makers), such as partitioning options into groups or categories.
Outline factors that impact the optimal number of alternatives to present to the decision maker.
When determining the optimal number of alternatives to present, the choice architect needs to balance two considerations:
(1) More options increase the chances of offering a preference match to the decision maker.
(2) More options place a greater cognitive burden on the decision maker because of the additional need to evaluate options.
Balance is affected by the willingness of the decision maker to engage in the choice process, the decision maker’s satisfaction with the decision process, and more generally the nature of the processes that will be used to make the decision. Balance is also contingent upon characteristics of the individual decision maker (e.g., level of processing capacity). This makes it hard to identify a simple recommendation for the optimal number of alternatives to present. Some general guidelines apply: One wants the fewest number of options that will encourage a reasoned consideration of tradeoffs among conflicting values and yet not seem too overwhelming to the decision maker and an adequate number of options to avoid generating context-specific preferences (a well-known phenomenon in choice, where the presence or absence of one option influences what is chosen). One recommendation that balances these considerations is having four or five nondominating options. One could start with this limited choice set but also provide the decision maker with the option of considering more options, if desired.
Explain how using technology-based decision aids while structuring the choice task can be both beneficial and detrimental to the decision maker.
Information technology can assist the decision maker with the choice task— choosing what to buy, what activities to participate in or what to attend. Technology-based tools such as search engines or product recommendation systems can help to identify attractive choice alternatives that the decision maker is not aware of and to filter out ones that are not of interest. Interactive decision aids can be used to compare choice alternatives in terms of their attractiveness on various feature dimensions. Automatic personalization of user interfaces can be used to reflect the decision maker’s preferences. Technology-based decision aids can be designed to steer decision makers toward choosing products, services or activities that are individually and/or socially desirable—i.e., healthy, environmentally friendly, etc.— without restricting their freedom to choose. Such decision aids can be highly beneficial. However, these tools can also predictably influence choices through very subtle architectural features, such as the set of other products that are presented alongside a recommended alternative or product attributes that are made more salient by the system. Decision aids can also steer decision makers toward choosing products, services or activities.
Explain the role of defaults in structuring the choice task.
“Defaults” are the options that are preselected if an individual does not make an active choice. They are one of the most powerful and popular tools available to the choice architect—an effective tool in choice architecture when there is inertia or uncertainty in decision making. Collections of default settings or default configurations determine the way decision makers initially encounter products, services or policies. “Reuse defaults” come into play with subsequent uses of a product. The range of default policies includes:
Simple defaults: Choosing one default for all
Random defaults: Assigning a configuration at random, for instance, as an experiment
Forced choice: Withholding the product or service by default and releasing it to the recipient only after an active choice is made
Sensory defaults: Change according to what can be inferred about the user, for example, websites change language dependent on country of origin of the visitor.
For products and services that are frequently chosen, the range of defaults includes:
Persistent defaults: Past choices are remembered
Reverting defaults: The last changes made to the default configuration are forgotten
Predictive defaults: Intelligently alter reuse defaults based on observation of the user.
Outline ethical issues associated with the use of defaults and policy to address these issues.
The ethical acceptability of using a default to guide choice has much to do with the reason the default is having an effect. When decision makers are aware that defaults may be set as recommendations in some cases or manipulation attempts in other cases, they exhibit a level of marketplace “metacognition” (i.e., awareness and understanding of one’s own thought processes) that suggests they successfully retain autonomy and freedom of choice. However, if defaults have an effect because decision makers are not aware that they have choices or because the transaction costs of changing from the default are too high, defaults impinge upon liberty. An often-prudent policy, though not a cure-all, is to set the default to the alternative that most people prefer when making an active choice, without time pressure and in the absence of any default.
Many choices that we face involve outcomes that unfold over long periods of time.
Intertemporal choice is concerned with the relative value we assign to payoffs at different points in time. It generally finds that we are biased toward the present and tend to discount the future. The fact that many choices involve outcomes that unfold over long periods of time impacts decision-maker behaviour in three specific ways:
(1) We tend to be myopic (i.e., shortsighted) and prefer to receive positive outcomes early, leading us to yield to immediate temptations and heavily discount future outcomes when making decisions. Time (temporal) discounting shows the differences in the relative valuation placed on rewards, usually money or goods, at different points in time, by comparing its valuation at an earlier date with one for a later date. It shows that present rewards are weighted more heavily than future ones.
(2) Uncertainty about the future can cause our preferences for future outcomes to be unclear, such that certain types of outcomes are systematically over- or underweighted. For example, uncertainties in life expectancy can affect decisions about financial products with future payouts such as savings, annuities and reverse mortgages. Similarly, uncertainties about the likelihood and extent of global climate change can reduce the political will for mitigation. In dealing with this uncertainty, the decision maker can become overly focused on certain highly salient or desirable future outcomes and fail to consider satisfactory second-best alternatives.
(3) We are often overly optimistic about the future and assume that we can accomplish more than we actually will. We expect to have both more time and more money in the future than we do today and overestimate the probability that desired outcomes will occur as planned.
The choice architect considers the intertemporal biases that can result from the fact that many choices involve outcomes that unfold over long periods of time.
Outline tools available to the choice architect to address intertemporal bias.
Tools available to address intertemporal bias include:
(a) Order of consideration. Drawing attention to the delayed options can refocus the decision maker, generating more patient choices. Refocusing the decision maker toward “satisficing” by considering second-best outcomes can lead to less choice deferral and higher choice satisfaction. When we make decisions by satisficing rather than optimizing, decisions are often simply good enough in light of the costs and constraints involved. Satisficing individuals choose options that meet their most basic decision criteria.
(b) Placing limited windows on opportunities. This can overcome the tendency to think that the future holds more resources.
In general, tools that translate aspects of the choice into immediate salient outcomes are more successful than those that attempt to manipulate heavily discounted future costs and benefits.
Explain how structuring the choice task as a single choice vs. a series of configuration decisions affects the decision maker’s search process
Structuring the choice task as a single choice vs. a series of configuration decisions has implications for the way in which information is searched, how choices are made and how they are justified. In a typical choice context, a decision maker needs to choose between a relatively small set of alternatives and typically is asked to choose only one product (e.g., buying grocery products, buying clothing, choosing a certain service provider, etc.). When decision makers are confronted with configurators (i.e., software systems for selecting the options for highly customizable products, such as for cars) they may use different strategies to deal with the complexity of multiple decisions that need to be made than those they use in the simpler, classical choice context. These strategies can affect their choice outcomes differently and therefore suggest that different tools of choice architecture might be needed. For example, decision makers are more likely to choose default levels of attributes when they begin with attributes that offer a greater number of configuration options than when they begin with attributes that offer a smaller number of options.
Decision makers commonly first screen alternatives on the basis of a subset of attributes and then make alternative-based comparisons for the remaining set of alternatives after screening. In this context, formatting the screening stage by facilitating comparisons on one attribute but not others will lead to a stronger preference for options favored by the focal attribute. The problems that come with long or complex search processes can be reduced with decision staging. Decision staging involves deciding what information to focus on, as well as choices between attributes and alternatives. For example, when we narrow down our options, we often tend to screen alternatives on the basis of a subset of attributes, and then we compare alternatives. Choice architects may not only break down complex decisions into multiple stages to make the process easier, but they can also work with an understanding of sequential decision making by facilitating certain comparisons at different stages of the choice process.
Explain how partitioning options and using attributes in choice option descriptions can impact choice behaviour. Suggest tools that consider this behaviour.
When we allocate a limited resource (e.g., money, attention, probabilistic belief, importance weights) over a fixed set of possibilities (e.g., investment opportunities, consumption options or time periods, events or attributes), we are typically biased toward even allocation over each group or category that has been identified. For example:
(a) In personal investment, we tend toward allocating 1/n of our savings to each of the options that are singled out in retirement plans.
(b) In consumer choice, we tend to seek variety when choosing multiple goods for future consumption, and we tend to favour spreading out consumption over different time periods.
(c) In decision analysis, we are biased toward assigning equal probabilities to each event that could occur and equal importance weights to each attribute that is explicitly identified.
This pervasive tendency toward even allocation provides a powerful tool to choice architects. Partitioning, i.e., physically partitioning resources into smaller units, can be used to alter this behaviour. When a resource is divided into smaller units, we encounter additional decision points—a psychological hurdle that encourages us to stop and think. Judgments and choices can be strongly influenced by the particular groups or categories into which the set of possibilities is partitioned, including:
(a) By assigning favoured investment options to separate superordinate categories (e.g., domestic and international stock index funds to choice sets A and B) and disfavoured investment options to a single superordinate category (e.g., several high-load exotic mutual funds to choice set C), we can nudge greater investment into the favoured options.
(b) By segregating healthy food menu options into separate menu categories (e.g., fruits, vegetables) and integrating unhealthy options into a single menu category (e.g., cookies, candies), we can nudge participants to choose a greater number of healthy options and a smaller number of unhealthy options.
(c) By segregating later time periods into separate categories and integrating earlier time periods into a single category, we can induce greater patience in consumption. By splitting more important attributes (e.g., practicality of an automobile) into a greater number of subcategories (e.g., safety, gas mileage, warranty) and combining less important attributes into a single category (e.g., stylishness—design, stereo, horsepower), we can increase the importance given to the more important attributes when we choose among product offerings.
Explain how an individual’s intrinsic preferences or beliefs impact partitioning.
The impact of partitioning will tend to be strongest among decision makers with weaker intrinsic preferences or beliefs and to diminish or disappear among those with stronger intrinsic preferences or beliefs. For example, wine novices asked to choose among several different varieties of white wine were more likely to diversify over grape if wines were grouped by grape type and more likely to diversify over country of origin when wines were grouped in that manner. This effect was greatly weakened among wine experts. Partitioning will tend to exert the strongest paternalistic influence on those who need the greatest guidance and will have the weakest effect on those who require the least guidance.
Explain how designing attributes can impact choice behaviour.
We predict our satisfaction with different alternatives by weighing the pros and cons of the different attributes of the choice options. For example, car buyers will consider attributes such as style, cost, safety, reliability, capacity and fuel economy. These attributes may be important in their own right or because they help us achieve more fundamental objectives. An ideal decision incorporates all of the relevant attributes and weighs them to the degree that it allows us to achieve our objectives.
Choice architects influence decision-making behaviour by making particular attributes more or less salient (i.e., prominent or important). Accuracy in choices can be improved by adhering to the principles of parsimony, linearity, comparability and evaluability.
(a) Parsimony: Decision makers, especially less numerate decision makers, understand more information and weigh important information better in choices that require less cognitive effort. If overwhelmed by too many attributes, we may simplify our decision by focusing on only one. Cognitive effort can be reduced by using smaller sets of attributes and highlighting the meaning of only the most important attributes. However, this approach must be balanced by the need to include all attributes.
(b) Linearity: A decision attribute may have a nonlinear relationship to a more fundamental objective. For example, many people expect monthly credit payments to have a roughly linear relationship to the payback period, when in reality the payback period increases sharply when monthly payments barely cover interest. This misperception can be corrected by credit card statements that list the monthly payment needed to eliminate a balance in three years. Nonlinear attributes (e.g., monthly payments in relation to credit card payback periods) can be made more understandable by rescaling or translating them (e.g., monthly payment needed to eliminate the balance in a certain number of years).
(c) Comparability: This is achieved by expressing the same attribute (e.g., cost) of different activities or products (e.g., newspaper subscriptions, cable costs, per use music download) on a standardized scale (e.g., monthly dollar cost) and allows us to compare the relative value more accurately.
(d) Evaluability: For highly quantitative information that can be difficult to process because the numbers are challenging or the domain is unfamiliar, breaking numbers into categories, such as grades, or assigning endpoints clearly labeled as good or bad facilitates evaluation. When numbers are supplemented with evaluative labels, we can integrate more information into judgments and access affective reactions more quickly. For example, newly proposed U.S. Environmental Protection Agency (EPA) labels contain information about carbon dioxide (CO2) emissions, but no one is familiar with what is a “good” or “bad” level of CO2. The labels try to remedy this problem by rating a given car on a one to ten scale that is linear in CO2 reduction.
Describe the tools attribute translation and attribute expansion.
Choice architects may make some attributes, such as those with externalities (i.e., side effects or consequences), that might otherwise be neglected more salient by using the following tools.
(a) Attribute translation: Mapping an attribute to its consequences for other objectives. Research suggests that we bring to mind only half the objectives we care about in a decision. There may benefits in explicitly mapping an attribute to its consequences for other objectives. For example, a car’s gas consumption is directly related to the cost of driving the car and to the CO2 emissions from the car. However, we may fail to translate gas consumption to either scale because the math is challenging or we don’t recognize the consequence on that objective without a reminder. The newly proposed EPA labels translate gas consumption to both driving costs and CO2 emissions to draw more attention to these objectives.
(b) Attribute expansion: Increasing the use of an attribute by changing the scale on which it is expressed. For example, a car’s gas consumption can be expressed over short distances (100 kilometers) or long distances (10,000 kilometers). Expanding the denominator makes the numerators larger and makes the differences between alternatives appear larger, leading these expanded attributes to receive more weight in choice.
Explain how characteristics of the decision maker can influence the effectiveness of nudges.
Choice architecture at its best promises better decisions, healthier lives and improved finances. But some “nudges” have gone wrong simply because a nudge can have multiple effects that may depend on characteristics of the decision maker and how he or she processes and draws meaning from information or on what his or her goals are.
The effectiveness of nudges can be increased if there is sensitivity to variations between individuals. In some cases, the right choice architecture may differ by these individual characteristics. Differences can be dealt with by customizing information. For example, prevention-focused people, who tend to focus on avoiding negative outcomes in their pursuit of goals, are more susceptible to loss frames, while individuals with a stronger promotion focus, who tend to pursue positive outcomes eagerly, are more easily influenced by gain frames.
Explain some of the challenges inherent in using the decision maker’s experience of the selected choice in evaluating the effectiveness of a choice architecture intervention. Provide example of strategies for addressing these challenges.
Most theories of choice (implicitly) assume that the utility of an outcome estimated ex ante (i.e., before the event) equals its utility when it is experienced ex post (after the event). A growing body of research has documented numerous ways in which we fail to accurately predict how we will feel about the outcomes of our choices. We often overestimate the impact of differences in income on well-being and underestimate the impact of an empty stomach on our grocery shopping decisions. A related problem is the underprediction of adaptation to enduring changes. For example, research demonstrates that we greatly overestimate the duration of our emotional responses to the denial of academic tenure and the breakup of a relationship. But there is evidence that those who have experience with a situation make more accurate predictions about adaptation.
Some features of existing policy and choice environments reflect at least a tacit knowledge of these phenomena. For example, many consumer protection laws provide cooling off periods, during which a consumer can cancel a choice without penalty. One function that experienced agents and advisors often serve is to encourage a decision maker to consider not only the features of an option that are salient at the time of choice but also those that will be more important when the outcomes are experienced. These and other interventions that bear on the decision maker’s knowledge about future outcomes are part of the choice architect’s tool kit.
Outline elements identified in the MINDSPACE framework that impact behaviour and their related effects
Messenger: We are heavily influenced by who communicates information.
Incentives: Our responses to incentives are shaped by predictable mental shortcuts, such as strongly avoiding losses.
Norms: We are strongly influenced by what others do.
Defaults: We “go with the flow” of preset options.
Salience: Our attention is drawn to what is novel and seems relevant to us.
Priming: Our acts are often influenced by subconscious cues.
Affect: Our emotional associations can powerfully shape our actions.
Commitments: We seek to be consistent with our public promises and to reciprocate acts.
Ego: We act in ways that make us feel better about ourselves
Some of the elements explain largely automatic effects on behaviour (e.g., Norms, Defaults, Salience, Priming, Affect), while other effects relate to elements that draw more on reflective processing (e.g., Messenger, Incentives, Commitments, Ego). Traditional theories of behaviour change that relate to the reflective system inform and provide value; however, to understand how influences on the reflective system work, we need to account for the influence of the automatic system. For example, we better understand how incentives work when we account for the automatic effects of loss aversion alongside the more considered weighing of costs and benefits.
Describe factors that influence the weight we give to information.
Factors that influence the weight we give to information include:
(a) Reactions we have to the source of that information. We are affected by the perceived authority of the messenger, whether formal or informal. For example, there is evidence that we are more likely to act on information if experts deliver it.
(b) Demographic and behavioural similarities between the expert and the recipient can improve the effectiveness of the intervention.
(c) Peer effects
(d) Feelings we have for the messenger. For example, we may irrationally discard advice given by someone we dislike.
(e) More rational and cognitive means. For example, whether there is a consensus across society (Do lots of different people say the same thing?) and the consistency across occasions (Does the communicator say the same thing in different situations?).
Describe the five main, related insights from behavioural economics that impact the effectiveness of incentive design.
(1) Losses loom larger than gains
(2) Reference points matter
(3) We overweight small probabilities
(4) We mentally allocate money to discrete bundles.
(5) We live for today at the expense of tomorrow