Linköping University | Department of Computer and Information Science (IDA) Bachelor’s thesis, 18 ECTS | Cognitive Science Spring Term 2021 | LIU-IDA/KOGVET-G--21/023--SE
Investigating Acceptance
Among the Swedish
Population Towards
Energy-Saving
Behavioral Interventions
Andreas Perjons
Tutor: Daniel Västfjäll
Examiner: Rachel Ellis
Copyright
The publishers will keep this document online on the Internet – or its possible replacement – for a period of 25 years starting from the date of publication barring exceptional
circumstances. The online availability of the document implies permanent permission for anyone to read, to download, or to print out single copies for his/her own use and to use it unchanged for noncommercial research and educational purpose. Subsequent transfers of copyright cannot revoke this permission. All other uses of the document are conditional upon the consent of the copyright owner. The publisher has taken technical and administrative measures to assure authenticity, security and accessibility. According to intellectual property law the author has the right to be mentioned when his/her work is accessed as described above and to be protected against infringement. For additional information about the
Linköping University Electronic Press and its procedures for publication and for assurance of document integrity, please refer to its www home page: http://www.ep.liu.se/.
Abstract
Legislation, economic incentives and informational campaigns are traditional tools of government used to exert its influence on citizens. More recently, other behavioral interventions called nudges and boosts have also come into usage to influence behavior. Nudges exploit faults in human decision making, pushing the individual in a direction of a specific choice, hence the name nudge. Boosts instead try to foster existing competences in the individual, effectivizing decision making while still preserving the individual’s own agency. Both nudges and boosts have proved to be cost-effective ways of influencing behavior, making them attractive alternatives to traditional behavioral interventions. An a priori way to investigate the effectiveness of behavioral interventions without their implementation is by measuring their acceptance. This thesis investigates the acceptance for nudges and boosts compared to traditional behavioral interventions when used in the domain of energy saving practices. The results show that acceptance differs greatly depending on which behavioral intervention is used, which energy saving domain the behavioral intervention is applied to, and to an extent the demographic characteristics of the individuals exposed to the behavioral intervention.
Acknowledgements
I would like to express my gratitude towards my supervisor Daniel Västfjäll for giving me the opportunity to conduct the study covered in this thesis and for providing invaluable help along the way. I would also like to thank my examiner Rachel Ellis and my classmates for providing feedback on the contents of this thesis. Finally I would like to thank my roommates and members of WALL-E for being great friends, you know who you are.
Table of contents
1. Introduction ... 1
1.1 Traditional ways of behavioral intervention by government ... 1
1.2 Nudges ... 1
1.3 Boosts ... 2
1.4 Climate change and global warming ... 2
2. Purpose ... 3 2.1 Problem specification ... 3 2.2 Hypothesis... 4 2.3 Delimitations ... 4 3. Theoretical background ... 4 3.1 Decision making ... 4
3.2 Dual process theory ... 5
3.3 Nudges ... 6
3.4 Boosts ... 7
3.5 Nudges and boosts versus traditional tools of policy ... 7
4. Method ... 8
4.1 Participants and design ... 8
4.2 Material ... 9
4.3 Ethics ... 12
4.4 Data analysis ... 12
5. Results ... 14
5.1 Percentages for acceptance toward behavioral interventions ... 14
5.2 Percentages for perceived intrusiveness in behavioral interventions ... 15
5.3 Correlations for acceptance across descriptive variables ... 17
5.4 Correlations for perceived intrusiveness across descriptive variables ... 18
6. Discussion ... 19 6.1 Results ... 19 6.2 Method ... 21 6.3 Future research ... 21 7. Conclusions ... 22 8. References ... 23 9. Appendix ... 27
1
1. Introduction
Paternalism – “The interference of a state or an individual with another person, against their
will, and defended or motivated by a claim that the person interfered with will be better off or protected from harm” (Dworkin, 2020).
1.1 Traditional ways of behavioral intervention by government
This philosophical view is what has set the tone for legislation since the very formation of government itself, with the argument that rules need to be put in place to “protect citizens from themselves” for the sake of the greater good. This is ethically permissible at least according to the hard paternalistic branch: it is permissible to prevent an individual from performing some action, even if the individual acts willingly and knowledgeably, such as preventing voluntary suicide. To accomplish this the government needs to exert its power over citizens, which is mainly accomplished through mandates in the form of legislation. However in the governmental toolbox of behavioral influence there also exists some less coercive, but still effective, methods. Economic incentives are often put in place by governments to make specific choices (the ones preferred by the government) more enticing to the decision maker, while at the same time preserving the freedom of choice. For example, subsidies could be put in place for citizens who want to install solar panels in their homes, encouraging them to make this decision due to the now lower economic cost. Conversely, decisions not preferred by the government can also be affected by economic incentives, such as increasing the tax on cigarettes or alcohol to deter citizens from buying (and consuming) these products. Lastly, the government can influence behavior through information campaigns. This could for example be messages on billboards urging citizens to consume less meat, or commercials to remind citizens of paying their television license. These types of interventions are collectively referred to as
carrots (economic incentives), sticks (mandates) and sermons (information campaigns)
(Bemelmans-Videc, Rist & Vedung, 1998).
1.2 Nudges
However, in the last decade a new type of intervention used to influence behavior has grown in popularity among policymakers, namely the nudge. A nudge can be defined as “any aspect of the choice architecture that alters people's behavior in a predictable way without forbidding any options or significantly changing their economic incentives. To count as a mere nudge, the intervention must be easy and cheap to avoid” (Thaler & Sunstein, 2008). Nudges take advantage of flaws in human decision making to steer individuals into making the choice preferred by the choice architect – the author of the nudge. Variations of nudges are essentially endless and could for example be that certain wares in a grocery store are placed at eye level, being the first one to catch the buyer's attention, increasing its chances of being bought. Another example is default nudges, which means that a predetermined choice happens if the individual chooses to do nothing in a specific situation. This could be an extra tax applied to something, which needs to be actively opted out from if the individual wishes to avoid it. Advantages of
2
nudges include cost efficiency, and easy implementation and reversibility compared to traditional interventions, which is why they have garnered a lot of attention from policymakers. Governments all over the world are implementing their own behavioral science teams or “nudge units” at an ever increasing rate, examples being the Behavioural Insights Team in the United Kingdom (Halpern, 2015), and the Social and Behavioral Sciences Team in the United states (SBST, 2015) which was officially endorsed by at the time president Barack Obama (Obama, 2015). Behavioral science teams have also been established in Australia and Germany, with interest continuing to rise over the entire globe (Ly & Soman, 2013; Sunstein, 2016), not least in Europe (Whitehead et al., 2014).
1.3 Boosts
Moreover, during the past few years a different intervention called boosts has also come into the spotlight. While nudges explicitly try to steer individuals into making a predetermined choice set by the choice architect, boosts instead tries to foster individuals’ competences, in turn effectivizing their decision making capabilities while at the same time preserving their own agency of decision (Hertwig & Grüne-Yanoff, 2017). This creates an even less paternalistic intervention that also aims to strengthen individuals’ competencies over time, whereas nudges only affect the behavior taking place at the time of the nudge. An example of a boost could be objective information, listing the pros and cons of a certain decision, helping the individual to make a decision based on the outcomes in relation to their own values. The question of whether to nudge or to boost is a current topic of discussion in the world of behavioral studies, with both interventions carrying their own pros, cons, and paradigms of usage (Grüne-Yanoff et al., 2018).
Carrots, sticks, sermons, nudges and boosts will henceforth collectively be referred to as
behavioral interventions or interventions in this thesis.
1.4 Climate change and global warming
Since the middle of the 20th century, humans have artificially affected the earth’s climate to an unprecedented degree, raising its average surface temperature by 1.2 ℃ (WMO, 2021). This causes severe effects on its ecosystems and is currently considered the largest threat to global health by the World Health Organization (IPCC, 2014). The majority of global warming is caused by emissions of greenhouse gasses, with carbon dioxide (CO2) and methane
contributing to 90% of this amount. These emissions are mostly produced by the burning of fossil fuel (coal, oil and natural gas) for energy consumption, be it in industry, homes or for transportation. As a measure to counteract global warming, the majority of the world's nations participate in the Paris Agreement put in place by the UNFCCC (2021). The goal of the Paris Agreement is to keep the average surface temperature from rising above 2 ℃, and pursuing the goal of limiting the increase to 1.5 ℃. While efforts to succeed with this goal are taking place on international levels, efforts need to be undertaken on every level of society to reach the goals of the Paris Agreement. A key factor here is strength in numbers, if individuals are making conscious effort to save energy and minimizing their carbon footprint the collective result will be immense.
3
This is where behavioral policy enters the picture. If governments can formulate effective behavioral interventions, the individual effort can and will increase drastically. Research on nudging for energy saving has yielded favorable results, with something as simple as informing households on their energy usage compared to neighboring households decreasing energy usage by 2% on average (Allcott, 2011). Keep in mind that this averaged 2% decrease applied to a treatment group of just over 300,000 households, which clearly shows the extremely high degree of return on investment for the clever usage of nudges. The results from further research point in the same direction: nudging for energy saving is an effective approach (e.g. Ayres, Raseman, & Shih 2013; Costa & Kahn 2013). However, knowledge of the full scope of the effectiveness of nudging is yet far from reached. Even if research on the acceptance for nudges has been conducted for some years by now (e.g. Reisch & Sunstein, 2016; Sunstein, Reisch & Kaiser, 2019; Sunstein, Reisch & Rauber, 2018), comparisons between nudges and traditional tools of policy have just started to take off (Almqvist, 2020; Hagman, 2019). Even further, the inclusion of boosts in this comparison has not yet been researched. Because of the untapped potential that the usage of nudges and boosts carries, it is in the interest of global health to further this research by investigating the effectiveness of nudges and boosts (not least in an energy saving context) in comparison to traditional tools of behavioral policy; carrots, sticks, sermons. A problem here is that the effectiveness of a behavioral intervention can only be known in hindsight, but an a priori way of probing for its effectiveness can be achieved by investigating the acceptance for the behavioral intervention in question. This is what has been done by Almqvist (2020) and the purpose of this thesis is to further this research by continuing to investigate acceptance for different types of behavioral interventions.
2. Purpose
The purpose of this thesis is to investigate acceptance among the Swedish population for the behavioral interventions carrot, stick, sermon, nudge, and boost when applied to energy saving practices. This serves as a foundation for further research and eventual implementation, since measuring acceptance serves as a partial predictor for the success of the intervention (Hagman, 2019). The primary aim of this thesis is to quantify the estimated acceptance for boosts and one type of nudge relative to the traditional behavioral interventions carrots, sticks and sermons in the Swedish population. The secondary aim of this thesis is to quantify the estimated acceptance for boosts and one type of nudge relative to the traditional behavioral interventions carrots, sticks and sermons in different relevant demographic subgroups of the Swedish population. The study covered in this thesis is conducted in collaboration with the research group JEDI-lab at Linköping university (2021) on behalf of the Swedish Research Council Formas (2021).
2.1 Problem specification
What is the internal level of acceptance for the behavioral interventions carrot, stick, sermon, nudge and boost among the Swedish population and relevant demographic subgroups of it, when used to foster energy saving practices?
4 2.2 Hypothesis
The hypothesis is that the level of acceptance for the behavioral interventions will be different from each other and rated in the following order, from highest to lowest: Boost, Sermon, Carrot, Nudge, Stick.
2.3 Delimitations
A survey was conducted in Swedish on a representative selection of 1016 Swedish citizens being the age of 18 or above. Save for descriptive data, collected data was quantitative and questions regarded policies and behaviors in relation to saving energy. The survey was performed online and respondents were gathered by the third-party company PFM Research (2021).
3. Theoretical background
3.1 Decision making
Until the 1950s, the paradigmatic model of human decision making was based on what is called the concept of homo economicus or economic man. This model portrays individuals as completely rational agents making optimal decisions based on whatever information necessary to make said decision. However, individuals often need to make decisions with incomplete information of the situation, while at the same time being given a small time frame to decide in. This is actually the case during the majority of the decision making that takes place in everyday life, making the model of homo economicus an ideal not applicable to current reality. Because of this, newer models of human decision making have been put forth, with the most famous being Herbert Simon's concept of bounded rationality (Simon, 1955). This concept proposes that rationality is constrained by factors such as problem complexity, cognitive capabilities of the individual and time available. Individuals are therefore unable to make completely rational decisions, hence rationality being bounded. To mitigate bounded rationality, individuals use cognitive heuristics, which can be thought of as mental shortcuts or rules of thumb, to simplify and effectivize decision making.
For example, “if one of two objects is recognized and the other is not, then infer that the recognized object has the higher value with respect to the criterion” (Goldstein & Gigerenzer, 2002). This is referred to as the recognition heuristic and might take effect when buying a product, you are more likely to buy a product from a brand that you recognize compared to one that you do not recognize due to its inferred higher value.
Reasoning through cognitive heuristics is an often cost-effective approach to decision making, saving time and cognitive resources. This is preferred by the mind, being a cognitive miser, which wants to take shortcuts and save resources wherever it can when processing information (Fiske & Taylor, 2013). The tendency to save time and resources is not without drawbacks though, reasoning through the use of cognitive heuristics might effectivize the process but decreases accuracy at the same time. This tradeoff in accuracy is often produced through
5
cognitive bias, a systematic error that occurs while reasoning through cognitive heuristics. For
example, humans show “a greater tendency to continue an endeavor once an investment in money, effort, or time has been made” (Arkes & Blumer, 1985). This is referred to as the sunk
cost fallacy, and means that it might be cheaper in the long run to purchase a new car than to
repair a currently owned one, but that this is refrained from due to the fact that money already has been invested into the old one.
The parallel drawn by Simon is that decision making through bounded rationality works as a pair of scissors, with one blade being “cognitive limitations” (heuristics) and the other being “structures of the environment” (regularities). These scissors are then able to cut (extract information from) the problem space so that it becomes comprehensible and a decision can be made (Gigerenzer & Selten, 2002). Decisions made this way are not optimal. However, according to the concept of bounded rationality, this is not the goal of human decision making either. Finding the optimal solution to every problem that an individual is faced with on a daily basis is simply too complex to be feasible. Instead, individuals rely on satisficing, which entails searching through the available problem space (by using the “bounded rationality scissors”) until a good enough solution given time and resources needed to invest is found. Therefore, individuals never consider the entire range of available options, but instead evaluate possible courses of action on a case-by-case basis until a sufficiently good solution is found. This shifts the view of human decision making from the concept of homo economicus to the concept of
homo heuristicus, or heuristic man (Brighton & Gigerenzer, 2008).
3.2 Dual process theory
A common model for information processing in the mind of the heuristic man is the dual
process theory or two systems theory, which is the theory that thought arises in two separate
ways. One variant of this theory stems from Daniel Kahneman’s research on heuristics and
biases (2011). This model asserts that the mind is split into two specific systems, System 1 and System 2, which process information in separate ways. System 1 is fast and requires a low
amount of cognitive resources to use, with decision making being performed mainly through intuitive reasoning. It is automatic and stimulus-response driven, therefore tightly coupled to emotion and perception. It can be thought of as an “auto-pilot” for the mind and is responsible for things such as reflexes and decision making based on intuitive reasoning. The problem with System 1 is that it has no way to modulate itself, making it prone to cognitive biases. System 2 is slower and requires more cognitive resources to use, although decision making is instead based on more analytical reasoning, making it suitable for more complex decisions. It needs to be used consciously, but has the abilities to “handle abstract information, falsify, and use deductive reasoning, and reason about the past and future events” (Hagman, 2019). System 1 and System 2 are not completely separated, partial calculations relating to System 2-reasoning are often off-loaded to System 1. Relating to the theory of the cognitive miser, System 1 is highly preferred by the mind and therefore always active. The usage of System 2 needs to be “switched on” consciously, so when it is not, decision making is done through System 1. Human decision making is therefore considered imperfect and error-prone according to these theories, which is what nudges take advantage of.
6 3.3 Nudges
The fact that human decision making is not completely rational by itself creates the possibility to influence it through external means, which is what nudges do. Any aspect in choice architecture with the goal to steer the individual being nudged in a predictable direction intended by the choice architect, i.e. make the choice preferred by the author of the nudge, without forbidding any choices or altering their economic incentive significantly can be considered a nudge. The intent of this is to help individuals make better decisions, since the fundamental principle in nudge theory is that humans are “fallible, inconsistent, ill-informed, unrealistically optimistic, and myopic, and they suffer from inertia and self-control problems” (Sunstein, 2014; Thaler & Sunstein, 2008). Therefore, through the usage of nudges, decision making is effectivized and steered in some direction. The range of what is considered to be a nudge is extremely broad, with many ways to separate and characterize them. One framework for this is Hagman’s nudge acceptance model (2019). Here nudges are separated on factors such as if the nudge is pro-self or pro-social, i.e. being of primary benefit to the individual or society. Another factor nudges are separated on is the intrusiveness to freedom of choice the nudge elicits. Even if nudges are to be considered less paternalistic than for example legislation, nudges still have the power to greatly influence decisions. Depending on if the nudge targets System 1 or System 2, the nudge is to be considered differently intrusive to freedom of choice. A purely informational nudge aimed at System 2 is much easier to guard against compared to a nudge aimed at eliciting an emotional response from System 1, due to the lack of self-modulation in System 1. Other factors nudges are separated on are the alternatives to the nudge, i.e. could the goal of the nudge be reached in another way, and choice architect of the nudge. Depending on these characteristics of the nudges their effectiveness and intrusiveness vary, with their separate levels of opposition following suit. Ethical concerns have been raised with nudges being considered too intrusive and low in transparency, making them hard (or impossible) to guard against. Opponents of nudges are often citing nudges as being unethical (Rebonato, 2012), manipulative (Wilkinson, 2013), threatening to personal integrity (Helbing et al., 2017), and being of benefit to the choice architects themselves (Berg, 2014).
This whole debate about the ethics of nudges poses some real questions and is an important one, but is not going to be elaborated on further here. However, acceptability of nudges is in and of itself something that might make or break a nudge and is therefore detrimental to the effectiveness of a nudge. As described by Jack Brehm (1989), consider the following:
“When you put your quarters in a softdrink machine, you would not like the machine to start flashing a large Coca Cola sign at you, complete with a flashing arrow that keeps moving to the button for Coke (Classical, of course). Or at least I don't think you would like that, even if you intended to obtain a Coke when you started to put your quarters in the machine”.
The phenomenon that would occur here is called psychological reactance and is characterized by strong feelings of aversion and unpleasant motivational arousal to stimuli that intrude on the freedom of choice. The consequences of psychological reactance is often obstinacy and unwillingness to cooperate, which if caused by a nudge would defeat the purpose of it. It is
7
therefore paramount that the individuals being nudged are accepting of the nudge to ensure its success. However, depending on the type of nudge, acceptability differs greatly. For example, in the separation of nudges into pro-self and pro-social, acceptability is higher for nudges considered pro-self (Hagman, 2019). Depending on the effectiveness of the nudge, acceptance also differs. Cadario and Chandon (2019) found an inverse relationship between acceptance for nudges and their effectiveness. A non-transparent nudge with the intent to increase bias in System 1-reasoning would prove more effective than a more transparent nudge with the intent to decrease bias in System 2-reasoning (provided that psychological reactance does not occur), but would at the same time be more intrusive. As intrusiveness increases, the chances that the nudge elicits psychological reactance increases as well. This creates a situation where effectiveness of nudges can be increased by increasing intrusiveness, but at some point the rate of intrusiveness becomes too high for the nudge to provide effective results. The problem here is that this drop-off-point can only be observed in hindsight, the furthest possible knowledge that can be gained beforehand is by investigating perceived acceptance for the specific nudges. On the positive side however, this type of research on acceptability of nudges have yielded cautious-supportive to favorable opinions to the use of nudges depending on population nationality (e.g. Almqvist, 2020; Hagman, 2019; Reisch & Sunstein, 2016; Sunstein, Reisch & Kaiser, 2019; Sunstein, Reisch & Rauber, 2018).
3.4 Boosts
As a result of the ongoing ethical debate concerning the use of nudges, the boost has grown into popularity over the past few years. It relies on the same psychological concepts as nudging, but instead of viewing the human as fallible, inconsistent, ill-informed etcetera, and trying to influence behavior, the principle of boosting instead acknowledges the bounds of human decision making, and tries to foster its competences instead (Hertwig & Grüne-Yanoff, 2017). This brings the focus from influencing behavior to influencing competence, which strengthens the individual over time while at the same time preserving the freedom of choice. Instead of steering the individual to a predetermined choice, boosts try to effectivize decision making based on the individual’s own agency. This makes boosts less intrusive, less paternalistic, and more ethical due to the preservation of freedom of choice, however a theoretical gap exists on the effectiveness and public acceptance of boosting.
3.5 Nudges and boosts versus traditional tools of policy
Nudges and boosts provide an enticing alternative to traditional behavioral interventions (carrots, sticks and sermons) due to their cost-effectiveness, and ease of implementation and reversibility. Even if general support for nudges have been observed, the public is still in preference of traditional tools of policy (Almqvist, 2020). This does not rule out the fact that nudging and boosting can be used to great effectiveness, but emphasizes the need for slow and closely monitored implementation, hence more research is needed. Since nudging for energy saving has shown to be an effective approach in earlier research, the question that now stands to answer is whether influencing energy saving can be optimized further by choosing the correct intervention.
8
4. Method
4.1 Participants and design
A study was conducted on a representative selection of the Swedish population to investigate acceptance for the behavioral interventions carrot, stick, sermon, nudge and boost when used to foster energy saving practices. With the independent variable being the type of intervention used, the study used a survey with within-subjects design. 1016 Swedish citizens participated in the study, see Table 1 for descriptive statistics. Participants were recruited through the third-party company PMF Research (2021) and received a small monetary compensation for completing the study. Data collection ran in April 2021.
Table 1. Descriptive statistics for respondents
Age Minimum age 18 Maximum age 81 Mean age 50.2 Median age 44 Standard deviation 18.44 Gender Men 499 (49.1%) Women 515 (50.8%)
Prefer not to say 2 (0.1%)
Level of education
No finished education 5 (0.5%)
Elementary school or equivalent 97 (9.5%)
High school 424 (41.7%) University 490 (48.3%) Political leaning Left 353 (34.7%) Right 439 (43.2%) Don’t know 221 (21.8%) No answer 3 (0.3%)
9
Party voted for if election today
Social Democrats (S) 207 (20.3%) Moderates (M) 206 (20.2%) Don’t know 168 (16.4%) Sweden Democrats (SD) 115 (11.3%) Left Party (V) 86 (8.5%) Center Party (C) 77 (7.6%) Liberals (L) 50 (4.9%) Christian Democrats (KD) 39 (3.8%) Green Party (MP) 38 (3.7%) Other 25 (2.5%)
Feminist Initiative (FI) 4 (0.4%)
No answer 1 (0.3%)
4.2 Material
The survey used was constructed through the online survey tool Qualtrics and divided into nine blocks. Block 1 contained a survey description and inquiry of consent as well as questions about age and gender. Blocks 2 - 8 each corresponded to one of the domains energy
conservation, mobility and transportation, waste avoidance, recycling, consumerism, and vicarious behaviors toward conservation, and smoking. Each block contained 6 items, one
example of each of the behavioral interventions positive carrot (positive economic incentive),
negative carrot (negative economic incentive), stick, sermon, nudge, and boost (see Box 1 for
extract and Appendix for full version). The study was confined to using one type of nudge, being a default type nudge due to its general effectiveness (Hummel & Maedche, 2019). The domains used were based on the General Ecological Behavior (GEB) scale (Kaiser, 1998), with the domain smoking being added as a control domain for independent comparison. For each intervention the participants were asked whether the intervention should be considered acceptable and/or intrusive on a binary yes/no-scale. The order of presentation for blocks 2 - 8 as well as the order of items in each block were randomized. Between block 8 and 9 an attention check was placed, prompting respondents to answer a specific alternative out of four possible alternatives. Block 9 contained 12 items, two for each domain on the GEB-scale, that surveyed respondents about their own ecological behaviors. Items contained a statement about ecological behavior, with responses being on a 4-point Likert-scale ranging from “never” to “often” (see Box 2 for extract and Appendix for full version). The last part of the survey also included descriptive questions about completed education, whether the participants considered
10
themselves to be on the left or right side of the political spectrum and which party in the Swedish Riksdag they would vote for if there was an election today. These questions were put last to prevent bias affecting earlier responses.
Box 1. Survey extract of items measuring acceptance and intrusiveness in domain energy conservation, translated from Swedish to English (see Appendix for full version)
Electricity, water and heating
Imagine that the government wants to decrease the usage of energy in society, therefore a couple of methods have been put forth to facilitate this. Human behavior can be influenced in many ways. How appropriate do you find it that authorities use the following methods with respect to acceptability and intrusiveness? Acceptability means that you would consent to its usage in reality even if you did not like it. Intrusiveness means that the intervention is intrusive to the individual’s possibility of choice.
[Positive carrot]
Government subsidies for individuals who wants to install solar panels in their home The intervention is acceptable
◻ Yes ◻ No
The intervention is intrusive ◻ Yes
◻ No
[Negative carrot]
When households exceeds a certain amount of electricity used each day the prize they pay for continued usage increases
The intervention is acceptable ◻ Yes
◻ No
The intervention is intrusive ◻ Yes
◻ No [Stick]
Households are only allowed to consume a certain amount of water each day The intervention is acceptable
◻ Yes ◻ No
11 The intervention is intrusive
◻ Yes ◻ No [Sermon]
Information that the amount of electricity that is used in society needs to decrease and how this can be accomplished are sent to households
The intervention is acceptable ◻ Yes
◻ No
The intervention is intrusive ◻ Yes
◻ No [Nudge]
An additional environmental tax is added to electricity, but this can be opted out from if you actively contact your energy distributor and state this
The intervention is acceptable ◻ Yes
◻ No
The intervention is intrusive ◻ Yes
◻ No [Boost]
Educational material meant to foster more informed decisions about electricity consumption are sent to households
The intervention is acceptable ◻ Yes
◻ No
The intervention is intrusive ◻ Yes
12
Box 2. Survey extract of items measuring individual habits related to energy consumption, translated from Swedish to English (see Appendix for full version)
Individual habits regarding energy consumption
Here a few statements will be presented that are related to energy consumption. Choose the alternative that best describes your own habits. If no correct alternative exists, choose the alternative that least bad describes your own habits.
[Domain Energy conservation]
I wait until I can fill up one or more whole machines before I do my laundry ◻ Often
◻ Sometimes ◻ Rarely ◻ Never
[Domain Energy conservation]
If I am the last person to leave a room I make sure to turn off the lights ◻ Often
◻ Sometimes ◻ Rarely ◻ Never
4.3 Ethics
All participants were informed at the start of the survey that their participation was voluntary, that their responses were anonymous and would only be used for independent research. To proceed further the participants had to give explicit consent.
4.4 Data analysis
For each behavioral intervention in each domain, percentages of “Yes”-answers to the questions of whether the behavioral intervention were considered acceptable and/or intrusive were calculated. A total average percentage for acceptance and perceived intrusiveness for all domains together as well as all behavioral interventions together was also calculated. An
environmental literacy-score was calculated by combining answers from items concerning
individual ecological behaviors, giving each answer a set of 1 to 4 points depending on if the answer indicated ecological behavior. Two items were reversed due to being negatively correlated with environmentally conscious behavior. Cronbach’s alpha for answers from items concerning individual ecological behaviors was calculated to measure consistency of items.
13
Environmental literacy-score served as a measurement of environmental consciousness in
individuals.
Chi2-tests were performed on each set of percentages for acceptance and perceived intrusiveness corresponding to the six behavioral interventions in each domain to determine if a significant difference in acceptance and perceived intrusiveness existed. Chi2-tests were also performed on each separate combination of percentage pairs corresponding to behavioral interventions in each domain to determine if a significant difference in acceptance and perceived intrusiveness existed. Chi2-tests were also performed on the total average acceptance and perceived intrusiveness percentages for both domains as well as behavioral interventions to determine if a significant difference in acceptance and perceived intrusiveness existed. To investigate if acceptance and perceived intrusiveness differed between GEB-scale domains and the independent domain smoking, Chi2-tests between acceptance and perceived intrusiveness percentages for GEB-scale total averages and acceptance and perceived intrusiveness percentages for independent domain smoking was performed.
To investigate acceptance and perceived intrusiveness for demographic subgroups, a set of Pearson correlation tests were performed. Difference in acceptance and perceived intrusiveness depending on gender was investigated by comparing percentages for acceptance and perceived intrusiveness to respective gender (binary scale of 1 and 2 corresponding to women and men). Difference in acceptance and perceived intrusiveness depending on political leaning was investigated by comparing percentages for acceptance and perceived intrusiveness to a 3-point scale being 1 – left leaning, 2 – no leaning, 3 – right leaning. Difference in acceptance and perceived intrusiveness depending on environmental literacy was investigated by comparing percentages for acceptance and perceived intrusiveness to a point scale of 12 – 48 being the score calculated from items about individual ecological behavior. Difference in acceptance and perceived intrusiveness depending on education was investigated by comparing percentages for acceptance and perceived intrusiveness to a 4-point scale being 1 – no completed education, 2 – elementary school or equivalent, 3 – high school, 4 – university.
All statistical testing was performed in Python 3, 41 respondents were excluded from the analysis for failing to complete the attention check.
14
5. Results
This section is divided into four subsections. The first section contains the percentages of acceptance for each combination of behavioral intervention – domain, total average acceptance percentages for each behavioral intervention as well as domain, Chi2 test statistics for the difference between behavioral interventions in each domain as well as the total average acceptance. Non-significant Chi2 test statistics between intervention pairs in each domain are also presented through subscripts next to percentages, see Table 2. Acceptance percentages for a non-GEB-scale domain smoking is also presented for comparison, along with its respective
Chi2-values. The second section contains the same information as the first but for perceived intrusiveness instead of acceptance. The third section contains the results from Pearson r correlation testing between total average acceptance percentages for each behavioral intervention and the descriptive variables gender, political leaning, environmental literacy, and
level of education. The fourth section contains the same information as the third but for
perceived intrusiveness instead of acceptance.
5.1 Percentages for acceptance toward behavioral interventions
Looking at the total average acceptance across all domains, respondents accepted the behavioral interventions boost, sermon, positive carrot, and negative carrot in a majority of the cases, see Table 1. Behavioral interventions stick and nudge were only accepted in a minority of the cases. Acceptance for individual interventions differs greatly depending on domain. The intervention positive carrot was only accepted in 37% of cases for the domain
Vicarious behaviors toward conservation while being accepted in a high majority for the rest
of the domains, having a standard deviation of 21.36 in the total average percentages across all domains. The intervention stick was accepted by a majority of respondents in domains Waste
avoidance and Recycling while being accepted by only a minority in other domains, also having
a standard deviation of 21.36 in the total average percentages across all domains. Conversely, standard deviation for behavioral intervention boost measured at 2.21, accepted in around 90% of cases for all domains. Chi2-test over total average acceptance between domains proved no significant difference between domains nor paired domains. Chi2-test across each separate domain proved significant difference for acceptability towards behavioral interventions across all domains. Results from Chi2-testing for separate combinations of behavioral intervention pairs where difference was not significant are subscripted in Table 1. For non-GEB-scale reference domain smoking, similar results for acceptance as in GEB-scale domains were obtained. Chi2-test between total average acceptance percentages and acceptance percentages for independent comparison variable smoking did not prove significant.
15
Table 2. Percentages for accepted behavioral interventions across domains and intervention types. Same letter next to percentages denotes interventions where Chi2-difference were not significant across domain
5.2 Percentages for perceived intrusiveness in behavioral interventions
Looking at the total average perceived intrusiveness across all domains, respondents rated the behavioral intervention stick as intrusive in a majority of the cases, see Table 3. The behavioral intervention nudge was rated as intrusive in 50% of cases and the behavioral interventions
boost, sermon, positive carrot, and negative carrot was rated as intrusive in a minority of the
cases. Perceived intrusiveness for individual interventions differs greatly depending on domain. The intervention stick was only considered intrusive in 31% and 43% of cases for domains Waste avoidance and Recycling respectively, while being rated as intrusive in a high majority for the rest of the domains. Standard deviation for behavioral intervention stick measured at 20.59. Conversely, standard deviation for behavioral intervention boost measured at 1.46 and was considered intrusive in only a small minority of cases. Chi2-test over total
16
average acceptance between domains proved no significant difference between domains, but proved significant for paired domains Energy conservation - Recycling, Recycling -
Consumption, and Recycling - Vicarious behaviors towards energy conservation. Chi2-test across each separate domain proved significant difference for acceptability towards behavioral interventions across all domains. Results from Chi2-testing for separate combinations of behavioral intervention pairs where difference was not significant are subscripted in Table 2. For non-GEB-scale reference domain smoking, similar results for perceived intrusiveness as in GEB-scale domains were obtained. Chi2-test between total average perceived intrusive percentages and perceived intrusive percentages for independent comparison variable smoking did not prove significant.
Table 3. Percentages for behavioral interventions considered intrusive across domains and intervention types. Same letter next to percentages denotes interventions where Chi2 -difference were not significant across domain
17
5.3 Correlations for acceptance across descriptive variables
Pearson correlation testing was performed on total average acceptance percentages over every domain for each behavioral intervention and descriptive variables gender, political leaning,
environmental literacy, and level of education, see Table 5. Descriptive variable gender
correlated significantly with total average acceptance percentage for each separate behavioral intervention, women having a positive correlation with high acceptance compared to men. Descriptive variable political leaning correlated significantly with total average acceptance percentage for each separate behavioral intervention, left-leaning respondents having a positive correlation with high acceptance compared to right-leaning respondents. Descriptive variable
environmental literacy correlated significantly with total average acceptance percentage for
each separate behavioral intervention, respondents with high environmental literacy score having a positive correlation with high acceptance compared to respondents with low environmental literacy score. Cronbach’s alpha for items measuring environmental literacy was measured at α = 0.639. Descriptive variable level of education correlated significantly with total average acceptance percentage for separate behavioral interventions negative carrot,
sermon and boost, respondents high in level of education having a positive correlation with
high acceptance compared to respondents low in level of education.
Table 4. Pearson correlation coefficients for total average acceptance percentages for each separate behavioral intervention measured against descriptive variables
18
5.4 Correlations for perceived intrusiveness across descriptive variables
Pearson correlation testing was performed on total average perceived intrusiveness percentages over every domain for each behavioral intervention and descriptive variables gender, political
leaning, environmental literacy, and level of education, see Table 4. Descriptive variable gender correlated significantly with total average perceived intrusiveness percentage for each
separate behavioral intervention, women having a negative correlation with high perceived intrusiveness compared to men. Descriptive variable political leaning correlated significantly with total average perceived intrusiveness percentage for each separate behavioral intervention, left-leaning respondents having a negative correlation with high perceived intrusiveness compared to right-leaning respondents. Descriptive variable environmental literacy correlated significantly with total average perceived intrusiveness percentage for each separate behavioral intervention, respondents with high environmental literacy score having a negative correlation with high perceived intrusiveness compared to respondents with low environmental literacy score. Descriptive variable level of education correlated significantly with total average perceived intrusiveness percentage for separate behavioral interventions stick, sermon and
boost. Respondents high in level of education had a positive correlation with high perceived intrusiveness for behavioral intervention stick and a negative correlation with high perceived intrusiveness for behavioral interventions sermon and boost, compared to respondents low in level of education.
Table 5. Pearson correlation coefficients for average perceived intrusiveness over domains in each behavioral intervention measured against descriptive variables
19
6. Discussion
6.1 Results
The aim of this study was to investigate how the behavioral interventions positive carrot,
negative carrot, stick, sermon, nudge, and boost were rated on levels of acceptance and
perceived intrusiveness in the domains energy conservation, mobility and transportation, waste
avoidance, recycling, consumerism, and vicarious behaviors towards conservation. Along with
this the study also aimed at performing explorative comparisons between demographic variables gender, political leaning, environmental literacy, and level of education. The hypothesis for this study was that respondents would rate the behavioral interventions on an average over all domains from highest to lowest in acceptance boost, sermon, positive carrot,
negative carrot, nudge, stick, and intrusiveness rated inversely. The results obtained from this
study shows three main points of interest.
The first point of interest in the results are the significant differences in acceptance and perceived intrusiveness between the different interventions. These differences can be grouped the following: sermon and boost – high acceptance and low intrusiveness, positive carrot and
negative carrot – medium acceptance and medium intrusiveness, stick and nudge – low
acceptance and high intrusiveness. This rank ordering is consistent with the hypothesis of the study, however the differences between behavioral interventions in each respective grouping are too small to be significant. The rank ordering of the different behavioral interventions found here is also consistent with earlier research on acceptance for behavioral interventions among the same demographic (Almqvist, 2020), but with lower rates of acceptance overall being obtained in this study. This is most likely due to this study being different from Almqvist’s study in the sense that it was confined to the evaluation of behavioral interventions used for energy saving purposes, with Almqvist investigating differences in broader domains. Another differing factor from previous research is that this study included the behavioral intervention
boost which has not been included in previous research. This factor might have created
different reference points in the survey (items being compared to each other), which has been shown in earlier research to influence decision making (e.g. Fredrick et al., 2009; Hsee and Zhang, 2010; Kahneman, 2011; Kahneman & Tversky, 2000). Results might also have been affected by the current COVID-19 pandemic (WHO, 2021).
The second point of interest is the difference in acceptance and perceived intrusiveness between the same type of behavioral intervention but in different domains. Overall, levels of acceptance and perceived intrusiveness follows a general trend in each domain consistent with the average percentages. However, outliers still exist such as behavioral intervention stick being rated at an acceptance percentage of 75% and 68% in domains Waste avoidance and Recycling respectively. Total average acceptance percentage for behavioral intervention stick was measured at 42% and domains except for Waste avoidance and Recycling averaging around 30%. A similar but inverse relationship can be observed in the ratings of perceived intrusiveness over domains for behavioral intervention stick. A possible explanation for this is that some form of subconscious framing effect occurs here (Druckman, 2001), where waste
20
avoidance and recycling is something that is generally considered acceptable to encourage whereas consumerism for example is not. This interpretation is also consistent with the results regarding acceptance for behavioral intervention stick in domain Consumerism, percentage being measured the lowest out of all domains at only 20%. Because of this view, more paternalistic (and effective) behavioral interventions might therefore be preferred due to the domain itself being subconsciously framed as something that is to be encouraged and vice versa. Another possible explanation for this is the difference in test items between domains, see Appendix and method limitation below.
The third point of interest are the differences in the correlation testing between the total average acceptance percentages and the demographic variables. Correlation for demographic variable
gender showed that women were significantly more likely to rate behavioral interventions as
acceptable and less likely to rate them as intrusive compared to men. Gender difference is consistent with results from Almqvist (2020), Diepeveen et al. (2013) and Sunstein, Reisch & Kaiser (2019), even if earlier research only covered nudges the difference is expected to apply to behavioral interventions in general. However even if correlation is significant Pearson r was only measured at an average of 0.11 and ˗0.13 for acceptance and perceived intrusiveness respectively which indicates a not particularly strong, albeit significant, correlation. Correlation for demographic variable political leaning showed that respondents identifying themselves as politically left were significantly more likely to rate behavioral interventions as acceptable and less likely to rate them as intrusive, compared to respondents identifying themselves as politically right or that did not know. These results are consistent with results found in Almqvist (2020). If we also infer that left-leaning individuals would be more collectivistic, this difference would also be consistent with results found in Hagman et al. (2015) and Jung and Mellers (2016). Correlation for political leaning was stronger than demographic variable gender, Pearson r being measured at an average of 0.19 and ˗0.17 for acceptance and perceived intrusiveness respectively. Correlation for demographic variable environmental literacy showed that respondents high in environmental literacy were significantly more likely to rate behavioral interventions as acceptable and less likely to rate them as intrusive, compared to respondents low in environmental literacy. Correlation for environmental literacy was stronger than demographic variable political leaning, Pearson r being measured at an average of 0.32 and ˗0.22 for acceptance and perceived intrusiveness respectively. This relationship can be explained by the fact that individuals high in environmental literacy most likely have a greater concern for the environment as well, which in turn gives rise to the nudge partisan bias described by Hagman (2019). Correlation for demographic variable level of education showed some significant differences in acceptance and perceived intrusiveness, however no systematic correlation was found for neither acceptance nor perceived intrusiveness across intervention type. The differences are however broadly pointing towards individuals high in level of education being more likely to rate behavioral interventions as acceptable and less likely to rate them as intrusive, compared to respondents low in level of education. This is roughly in line with findings by Almqvist (2020).
21 6.2 Method
The biggest concern about this study is the irregularity in items across domains designed to measure acceptance and intrusiveness. The methodological decision was made to try to diversify test items instead of keeping them the same across all GEB-domains. This was done to increase external validity, the generalizability of the results would otherwise be narrowed down. For example, if items for behavioral intervention negative carrot only consisted of explicitly stated taxes then the results would only be generalizable to taxes. Negative carrot refers to economic incentives which encompasses a broader range of interventions than only taxes. Further, asking participants to respond to the same intervention repeatedly may have led to lower engagement with the survey. Even if this choice was made explicitly, test items were still quite similar due to constraints in possibilities of item formulation, see Appendix for full survey. The downside of this choice is that it might compromise internal validity as a trade-off. For example, legislation as represented in the behavioral intervention stick can be applied in a broad range of areas. However, what is being legislated against can differ greatly in impact on respondents (cf. legislation determining the amount of trips one can do with commercial airliners each year vs. legislation forbidding the sale of non-renewable plastic bags in grocery stores), which in turn might impact the acceptance and perceived intrusiveness for the behavioral intervention in question.
Another discrepancy that might have affected internal reliability is the way behavioral interventions were framed. Research conducted by Hagman (2019) shows conflicting evidence whether nudges are affected by framing, however earlier research shows that the public support for taxes are affected by framing (McCaffery & Baron, 2004) and that this is thought to apply to nudges as well (Almqvist, 2020). The study could also have controlled for more demographic variables such as income to ensure greater external validity.
6.3 Future research
While acceptance for nudges is a field of research that has been established to some degree over the past few years, research concerning the comparison of nudges to other behavioral interventions is still in its infancy. This study along with ones conducted by Almqvist (2020) and Hagman (2019) are among the first ones to explore this realm, this study being the first one to also compare the behavioral intervention boost to nudges and traditional interventions. Because of this, an extremely large range of possibilities for further research exists that can build on what exists as of now. Concerning this study, the first discerning factor of where to continue research lies within the choice of area of application. Since this study was confined to investigate the acceptance for behavioral interventions exclusively in the area of energy saving practices, a multitude of domains where the application of behavioral interventions is possible along with their corresponding acceptance still exists to be investigated. If the same domain of application is chosen, the variations that are possible also exist in multitude. Since the type of nudge used in this study was confined to a default opt-out-nudge, other types of nudges can be investigated. In an even finer sense, the actual examples for behavioral interventions used can also be changed. As long as the number of respondents remain adequate, more in-depth analyses of acceptance based on demographic variables are also possible.
22
Because the cost-effectiveness of nudges and boosts is so high, and also because of the especially high level of acceptance for boosts observed in this study, further research on and application of these newer types of behavioral interventions are likely to prove lucrative.
7. Conclusions
This study has shown that acceptance and perceived intrusiveness for different behavioral interventions used to foster energy saving practices differ greatly and that this depends on several factors. They include the type of intervention used, the domain in which the intervention is applied, and to an extent the demographic characteristics of individuals exposed to the intervention. Since acceptance and perceived intrusiveness of behavioral interventions is a good predictor for their effectiveness in practical use, these factors are important for choice architects to take into consideration when formulating them. The amount of studies in this area of science is still very limited, so the hope is that this work can further the expansion of research and knowledge on behavioral interventions and their use.
23
8. References
Allcott, H. (2011). Social norms and energy conservation. Journal of Public Economics, 95
(9-10), 1082- 95. doi:10.1016/j.jpubeco.2011.03.003.
Almqvist, G. (2020). The Good Place: Essays on Nudging. (Doctoral thesis, Stockholm School of Economics). Retrieved from:
https://www.hhs.se/contentassets/a6f8c612f38b47de8c49ac8a5a8b9614/sse-phd-diss_2020_kappa_gustav-almqvist_the-good-place.pdf
Arkes, H. R, & Blumer, C. (1985). "The psychology of sunk cost". Organizational Behavior
and Human Decision Processes, 35(1), 124–140. doi:10.1016/0749-5978(85)90049-4
Ayres, I., Raseman, S., & Shih, A. (2013). Evidence from two large field experiments that peer comparison feedback can reduce residential energy usage. Journal of Law, Economics,
& Organization, 29(5), 992-1022. doi:10.1093/jleo/ews020
Bemelmans-Videc, M. L., Rist, R. C., & Vedung, E. O. (Eds.). (2011). Carrots, sticks, and
sermons: Policy instruments and their evaluation. New Jersey: Transaction Publishers.
Berg, L. (2014). Who benefits from behavioural economics?. Economic Analysis and Policy,
44(2), 221-232
Brehm, J. W. (1989). Psychological Reactance: Theory and Applications. Advances in
Consumer Research 16. Reds. Thomas K. Srull, Provo, UT: Association for Consumer
Research, pp. 72–75
Brighton, H., & Gigerenzer, G. (2008). Bayesian brains and cognitive mechanisms: Harmony or dissonance. The probabilistic mind: Prospects for Bayesian cognitive science, ed. N.
Chater & M. Oaksford, 189-208.
Cadario, R., & Chandon, P. (2019). Viewpoint: Effectiveness or consumer acceptance? Tradeoffs in selecting healthy eating nudges. Food Policy. 85, 1-6.
Costa, D. L., & Kahn. M. E. (2013). Energy conservation “nudges” and environmentalist ideology: Evidence from a randomized residential electricity field experiment. Journal of the
European Economic Association, 11(3), 680-702. doi:10.1111/jeea.12011
Diepeveen, S., Ling, T., Suhrcke, M., Roland, M., & Marteau, T. M. (2013). Public
acceptability of government intervention to change health-related behaviours: a systematic review and narrative synthesis. BMC Public Health, 13(756).
Druckman, J. N. (2001). The Implications of Framing Effects for Citizen Competence.
24
Dworkin, G. (2020). Paternalism. In E. N. Zalta (Ed.), The Stanford Encyclopedia of
Philosophy (Fall 2020 Edition). Palo Alto, CA: Metaphysics Research Lab, Stanford
University.
Fiske, S. T., & Taylor, S. E. (2013). Social Cognition: From Brains to Culture. London: SAGE Publications Ltd. ISBN: 978-1-4462-5814-9
Formas. (2021). Formas. Retrieved from: https://www.formas.se/en/start-page.html
Frederick, S., Novemsky, N., Wang, J., Dhar, R., & Nowlis, S. (2009). Opportunity cost neglect. Journal of Consumer Research, 36(4), 553-561.
Gigerenzer, G., & Selten, R. (2002). Bounded Rationality: The Adaptive Toolbox. Cambridge, MA: The MIT Press.
Goldstein, D. G., & Gigerenzer, G. (2002). Models of ecological rationality: The recognition heuristic. Psychological Review. 109(1), 75-90.
Grüne-Yanoff, T., Marchionni, C., & Feufel, M. A. (2018). Toward a framework for selecting behavioural policies: How to choose between boosts and nudges. Economics and
Philosophy, 34, 243-266.
Hagman, W. (2019). When are nudges acceptable?: Influences of beneficiaries, techniques,
alternatives and choice architects. (Doctoral thesis, Linköping Studies in Arts and Sciences,
759). Linköping: LiU-Tryck. Retrieved from:
http://www.divaportal.org/smash/get/diva2:1265064/FULLTEXT01.pdf
Hagman, W., Andersson, D., Västfjäll, D., & Tinghög, G. (2015). Public views on policies involving nudges. Review of Philosophy and Psychology, 6(3), 439-453.
Halpern, D. (2015). Inside the nudge unit: How small changes can make a big difference. London, UK: WH Allen.
Helbing, D., Frey, B. S., Gigerenzer, G., Hafen, E., Hagner, M., Hofstetter, Y., ... & Zwitter, A. (2017, February 25). Will Democracy Survive Big Data and Artificial Intelligence? [Article]. Retrieved from
https://www.scientificamerican.com/article/will-democracy-survive-big-data-andartificial-intelligence/
Hertwig, R., & Grüne-Yanoff, T. (2017). Nudging and Boosting: Steering or Empowering Good Decisions. Perspectives on Psychological Science, 12(6), 973–986
Hsee, C. K., & Zhang, J. (2010). General evaluability theory. Perspectives on Psychological
25
Hummel, D., & Maedche, A. (2019). How effective is nudging? A quantitative review on the effect sizes and limits of empirical nudging studies. Journal of Behavioral and Experimental
Economics, 80, 47-58.
Intergovernmental Panel on Climate Change. (2014). AR5 Synthesis Report: Climate Change
2014. Geneva, Switzerland: IPCC
JEDI-lab. (2021). JEDI-lab. Retrieved from: https://liu.se/forskning/jedi-lab
Jung, J. Y., & Mellers, B. A. (2016). American attitudes toward nudges. Judgment and
Decision Making, 11(1), 62-74.
Kahneman, D. (2011). Thinking, fast and slow. New York: Farrar, Straus and Giroux.
Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1991). Anomalies: The Endowment Effect, Loss Aversion, and Status Quo Bias. Journal of Economic Perspectives., 5(1), 193–206. doi:10.1257/jep.5.1.193
Kahneman, D., & Tversky, A. (Eds.). (2000). Choices, values and frames. New York: Cambridge University Press and the Russell Sage Foundation.
Kaiser, F. G. (1998). A General Measure of Ecological Behavior. Journal of Applied Social
Psychology. 28(5), 395-422.
Ly, K., & Soman, D. (2013). Nudging around the world (Research Report Series). Retrieved from the Rotman School of Management, University of Toronto:
https://inside.rotman.utoronto.ca/behaviouraleconomicsinaction/files/2013/12/Nudging-Around-The-World_Sep2013.pdf
McCaffery, E. J., & Baron, J. (2004). Framing and taxation: Evaluation of tax policies involving household composition. Journal of Economic Psychology, 25(6), 679-705.
Obama, B. (2015). Executive order — Using behavioral science insights to better serve the
American people (Executive Order 13707). Washington, DC: The White House. Retrieved
from the White House:
https://obamawhitehouse.archives.gov/the-press-office/2015/09/15/executive-order-using-behavioral-science-insights-better-serve-american
PFM Research. (2021). PFM Research. Retrieved from: https://www.pfmresearch.se/
Rebonato, R. (2012). Taking liberties: A critical examination of libertarian paternalism. Basingstoke: Palgrave Macmillan.
26
Reisch, L. A., & Sunstein, C. R. (2016). Do Europeans like nudges?. Judgment and Decision
Making 11(4), 310–325
Simon, H. A. (1955). A behavioral model of rational choice. The quarterly journal of
economics, 69(1), 99-118.
Social and Behavioral Sciences Team (2015). Social and behavioral sciences team 2015
annual report. Washington, DC: National Science and Technology Council, Executive Office
of the President. Retrieved from the Social and Behavioral Sciences Team: https://sbst.gov/assets/files/ 2015-annual-report.pdf.
Sunstein, C. R. (2014). Why nudge? The politics of libertarian paternalism. New Haven, CT: Yale University Press.
Sunstein, C. R. (2016). The council of psychological advisers. Annual Review of Psychology,
67, 713–737.
Sunstein, C. R., Reisch, L. A., & Kaiser, M. (2019). Trusting nudges? Lessons from an international survey. Journal of European Public Policy, 26(10), 1417-1443.
Sunstein, C. R., Reisch, L. A., & Rauber, J. (2018). A worldwide consensus on nudging? Not quite, but almost. Regulation & Governance, 12(1), 3-22.
Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving decisions about health, wealth
and happiness. London: Penguin Books.
United Nations Framework Convention on Climate Change. (2021). The Paris Agreement. Retrieved 2021-04-23 from: https://unfccc.int/process-and-meetings/the-paris-agreement/the-paris-agreement
Wilkinson, T. M. (2013). Nudging and Manipulation. Political Studies, 61, 341-355.
World Health Organization. (2021). WHO Coronavirus (COVID-19) Dashboard. Retrieved 2021-05-21 from: https://covid19.who.int/
World Meteorological Organization. (2021). The State of the Global Climate 2020. Retrieved 2021-04-18 from: https://public.wmo.int/en/our-mandate/climate/wmo-statement-state-of-global-climate
27
9. Appendix
Survey has been translated from Swedish to English. Items measuring acceptance and perceived intrusiveness have been annotated with their respective intervention type. Items measuring individual habits relating to energy-saving practices have been annotated with their respective GEB-domain.
Acceptance for behavioral interventions with the aim of reducing energy consumption In this survey we would like your opinion on a number of scenarios describing different attempts at influencing the behavior of individuals (referred to here as behavioral interventions) in situations relating to energy consumption, as well as your opinion on statements concerning your own habits regarding energy consumption. Your participation is voluntary and the answers that you give will be anonymous so that they can not be connected with you. The survey is conducted by Linköping university (contact: Andreas Perjons, andpe813@student.liu.se) and the results will only be used for independent research. In this survey we are interested in the societal opinion on different types of interventions used to encourage different types of behaviors in situations related to energy consumption with the goal of decreasing it.
Press below and then “Next” if you agree to these terms.
◻ I have understood and accept that the information that I provide will be saved and used according to the description above
Individual information Age [Textbox] Gender ◻ Male ◻ Female ◻ Other
◻ Prefer not to say
A number of examples of behavioral interventions will be shown to you whose aim is to decrease energy consumption in different domains of society. The behavioral interventions can be rated based on if you consider them to be acceptable and intrusive. Acceptable means that you would consent to their usage in the real world, even if you did not like it. Intrusive means that the behavioral intervention is intrusive to the individual’s ability to make independent choices.
Electricity, water and heating
Imagine that the government wants to decrease the usage of energy in society, therefore a couple of methods have been put forth to facilitate this. Human behavior can be influenced in many ways. How appropriate do you find it that authorities use the following methods