‘Nudging’ Sustainable Food Choices
The Role of Defaults, Frames,
Habits and Nature Relatedness
Sigrid Møyner Hohle
Master Thesis in Cultural and Societal Psychology
UNIVERSITY OF OSLO
May 2014
Copyright Sigrid Møyner Hohle
2014
‘Nudging’ Sustainable Food Choices.
The Role of Defaults, Frames, Habits and Nature Relatedness
Sigrid Møyner Hohle
http://www.duo.uio.no
Trykk: Reprosentralen, Universitetet i Oslo
ii
iii
Contents
Acknowledgements ...................................................................................................................vi
Abstract.....................................................................................................................................vii
Aim of Study .............................................................................................................................1
The Environmental and Health Impacts of Meat..................................................................1
Pro-Environmental Behavior ..................................................................................................2
Psychological Barriers to Pro-Environmental Behavior ......................................................3
Two Minds: System 1 and System 2 ...................................................................................4
Habits: The Same Procedure as Last Year, Miss Sophie?................................................5
Cognitive Dissonance: Thoughts and Behavior in Harmony ...........................................6
Denial: “Knowing and Not Knowing”................................................................................7
Facilitating Pro-Environmental Behavior through Nudging ...............................................8
Defaults: The Path of Least Resistance ..............................................................................9
Naming as Framing: Half-full or Half-empty..................................................................10
Nature Connectedness............................................................................................................11
Design of the Study and Hypotheses.....................................................................................13
Method and Materials............................................................................................................16
Design and Procedure ........................................................................................................16
Survey ..................................................................................................................................18
Analysis....................................................................................................................................21
Results......................................................................................................................................21
Hypothesis 1: Wrap Selection by Nudge ..........................................................................22
Hypothesis 2: Nature Relatedness as Moderator ............................................................23
Hypothesis 3: Meat Habits of the Customers...................................................................27
Discussion ................................................................................................................................29
The Effect of the Nudges....................................................................................................30
Interaction between Nature Relatedness and the Nudges ..............................................33
Breaking Meat Habits with a Nudge ................................................................................34
iv
Implications.........................................................................................................................36
Ethics of Nudging ...............................................................................................................40
Strengths and Limitations .................................................................................................41
Conclusion ...............................................................................................................................43
References ...............................................................................................................................44
Appendices ..............................................................................................................................50
Tables
Table 1.
Psychometric Properties of the Study Variables………………………………20
Table 2.
Nudge Condition and Wrap Choices...………………………………………...23
Table 3.
Binary Logistic Regression of Vegetarian Wrap Selection in Control versus
Default Frame Condition...……………………….…………………………...24
Table 4.
Binary Logistic Regression of Vegetarian Wrap Selection in Control versus
Framing Condition………….………..…………………..……………………26
Figures
Figure 1.
Control Poster…………………….……………………….…………..……..17
Figure 2.
Framing Poster.…………………….………………………………..……….17
Figure 3.
Default Poster.…………………….………………………………..…….…..18
Figure 4.
Default Frame Poster.…………….……………………………..…………...18
Figure 5.
Wrap Satisfaction……………………….……………………………………22
Figure 6.
Vegetarian Wrap Selection……………………….……………………..…...22
Figure 7.
Interaction of NR Perspective and Nudge on Wrap Selection………………25
Figure 8.
Interaction of NR Experience and Nudge on Wrap Selection……………….27
Figure 9.
Interaction of NR Self and Nudge on Wrap Selection.…………...…………27
Figure 10. Customers’ Meat Consumption Habit..………….…………..…..…...……...28
v
Acknowledgements
Writer: Sigrid Møyner Hohle
I developed the experimental design, the survey and the research material, set up the
experiment, instructed the café staff and collected the data. Nevertheless, special thanks go to
GreeNudge for supporting this research. My warmest thanks also go to Ebrahim and Marion.
Without you there would have been no experiment! I also thank all participants who
volunteered to participate in this study.
I wish to thank my main supervisor Jonas Rønningsdalen Kunst for all guidance, advice and
patience, and for your energy and belief in me and this project.
I also thank my second supervisor Per Espen Stoknes for introducing me to ecopsychology
and climate communication, for always inspiring me, and for valuable counseling.
Mamma, pappa, Anders and Knut: thank you for making me care about the environment and
nature in the first place, and for inspiring me and believing in me. And thanks to Bella, for
very valuable support.
Else Margrete, thank you the daily laughers and wonderful walking breaks.
Last, but not least, Vegard: Thank you! For all your help with this project, but most of all for
being such a fantastic person.
Oslo, April 15th 2014
vi
Abstract
Reduced consumption of meat could help mitigate climate change and health problems.
‘Nudging’ is one approach to individual behavior change. By nudging, the choice context is
designed to facilitate certain decisions. In the present field experiment, three different nudges
were tested, aiming at increasing vegetarian food selection among 353 customers in a café.
The menu was changed so that the vegetarian food options were either presented as the dish
of the day (DEFAULT) or given a more attractive name (FRAMING). Additionally, a
combination of the two (DEFAULT+FRAMING) was tested. To investigate whether the
effect of the nudges depended on individual factors, the customers filled out a survey in which
they reported their meat consumption habits and their sense of connection to the natural world
(Nature Relatedness). When looking at all customers, the nudges showed no effect on food
choice. However, to of the three nudges had an effect for customers with low Nature
Relatedness. Moreover, customers with high meat consumption in their daily lives were more
likely to select vegetarian food in DEFAULT+FRAMING. Hence, the findings indicate that
by giving vegetarian food a more attractive name and salient presentation, it can appeal to
customers otherwise unlikely to select vegetarian food.
Keywords: nudging, choice architecture, default rules, framing, nature relatedness,
habits, sustainable consumption, pro-environmental behavior, food choices.
vii
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Aim of Study
Reducing meat consumption is an effective way to reduce greenhouse gasses and other
forms of environmental degradation (Reisch, Eberle, & Lorek, 2013). Reducing meat
consumption, however, means that we need to change the behavior of people. One approach
to creating such behavioral change is nudging, a concept derived from decision theory (Thaler
& Sunstein, 2008). Nudging is the act of designing choice situations in such a way that they
facilitate certain preferred behaviors without restricting the individual’s freedom of choice
(Thaler & Sunstein, 2008). Some studies have explored the potential of increasing selection of
vegetarian dishes through nudging (Campbell-Arvai, Arvai, & Kalof, 2014), but none have
tested this in a real-life situation. I therefore took it upon me to set up a field experiment. I
wanted to explore whether I could increase the proportion of vegetarian food sold in a café by
creating new menus in which the vegetarian dishes were presented in a more salient and
attractive way. Personal factors such as habits and a person’s sense of connectedness to nature
are strongly related to environmental friendly behavior, including sustainable food choices
(Klöckner & Verplanken, 2013; Mayer & Frantz, 2004; Nisbet, Zelenski, & Murphy, 2009).
Therefore, I gave the customers a survey to test whether the new menus would work
differently for different customers, depending on their personal meat habits and connection to
nature. To measure connection to nature in the survey I used the Nature Relatedness scale
developed by Nisbet et al. (2009). Before presenting this study’s hypotheses and results, the
following sections will describe the climate effects of meat production, and present three
psychological barriers to environmental behavior. Nudging as a mean to reduce these barriers
and facilitate more sustainable food consumption will then be introduced. Finally, I will
discuss Nature Relatedness as a potential moderator of the nudges.
The Environmental and Health Impacts of Meat
There is common agreement that climate change is the world’s biggest challenge (N.
H. Stern, Britain, & Treasury, 2006; Utenriksdepartementet, 2011). Food systems contribute
between one fifth and one third of human-caused greenhouse gas emissions (Vermeulen,
Campbell, & Ingram, 2012), and livestock has been estimated to contribute 18 percent of
emissions (Steinfeld, Gerber, Wassenaar, Castel, & de Haan, 2006). Some claim that
livestock and their byproducts account for even larger amounts, up to 51 percent of annual
worldwide emissions (Goodland & Anhang, 2009). Regardless of the variation in these
estimates, meat consumption makes up one of the private consumption areas with the largest
impact on the environment (Reisch et al., 2013).
1
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
In most cases, meat production accounts for far more greenhouse gas emissions and
environmental degradation per calorie than vegetable farming (Carlsson-Kanyama & González, 2009). 1 kilo of beef accounts for approximately 80 times more emissions of greenhouse
gases than 1 kilo of Norwegian carrot, pumpkin or squash (Nymoen & Hille, 2012). When
Hedenus, Wirsenius, and Johansson (2014) compared different options for agriculture to
mitigate greenhouse gas emissions, they concluded that reducing emissions by production and
technology changes only will not be enough to limit the global average surface temperatures
from increasing by more than 2 °C above pre-industrial levels. Reduced ruminant meat and
dairy consumption is necessary to reach the 2 °C target with a high probability.
A diet consisting of less red meat and more vegetables is also a healthier diet. A high
intake of red meat is associated with increased risk of overall mortality, cardiovascular
disease mortality and cancer mortality, whereas a high intake of nuts, legumes and fruits is
associated with lower risk of various forms of cancer, diabetes, hypertension and arthritis
(Fraser, 1999; Pan et al., 2012; Sinha, Cross, Graubard, Leitzmann, & Schatzkin, 2009). In
spite of this, meat consumption in industrialized countries steadily increases. Whereas an
average Norwegian ate 44 kilos of meat in 1970, the annual consumption was 75 kilos in
2011 (Helsedirektoratet, 2012). How can we influence the behavior of people so that they get
a more sustainable and at the same time healthier diet, by eating less meat?
Pro-Environmental Behavior
An extensive amount of research within the field of environmental psychology has
investigated what factors are important to make individuals act environmental friendly, and
different terms have been coined for such behavior. This paper will use the term proenvironmental behavior (PEB), which can be defined as behavior that minimizes the negative
impact of one’s action on the environment, or even benefits it (Steg & Vlek, 2009). By this
definition, PEB is not necessarily motivated by environmental goals. Hence, people can act
pro-environmentally without intentions to do so.
When a person chooses to act or not act in a pro-environmental way, her behavior is a
product of both individual and external factors. Individual factors might be a person’s
knowledge, emotions, values, attitudes and norms. External predictors include physical
infrastructure, availability, product characteristics, and social, economic and cultural factors
(Kollmuss & Agyeman, 2002; Steg & Vlek, 2009; P. C. Stern, 2000). For instance, whether
an individual eats sustainable food such as vegetarian, local or organic food could depend on
her knowledge of the environmental impact of different diets, her personal norms and specific
2
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
attitudes towards these food types and toward environmental causes in general, as well as her
more overarching value system (De Groot & Steg, 2008; Kollmuss & Agyeman, 2002; P. C.
Stern, Dietz, Abel, Guagnano, & Kalof, 1999). However, contextual factors such as prices
and availability of these foods in her local supermarket, information about carbon footprints
or organic certification on the label of the product, the diets of her friends and family, and
what kind of food that is common and acceptable in her culture, could also play an important
role (Kollmuss & Agyeman, 2002; Steg & Vlek, 2009). In addition to being predicted by
individual and contextual factors, several theorists understand personal habits as a separate
barrier to PEB, as we shall see.
Psychological Barriers to Pro-Environmental Behavior
Early models of PEB assumed that better knowledge of environmental issues would
lead to pro-environmental attitudes, which would in turn result in PEB (Kollmuss &
Agyeman, 2002; Schultz, 2002b). The expectation was that educating people about
environmental issues and consequences of behaviors automatically would lead to more PEB.
This knowledge-deficit model seems to be the rationale behind many information and attitude
campaigns, which remain a widely used method to increase peoples’ PEB. This model
assumes that people are rational actors. In general, research on PEB is dominated by such
rational choice models, in which perceived consequences of behavior are seen as the
motivators of PEB (Schultz, 2002b; Steg & Vlek, 2009).
Over the past decades psychologists and economists have questioned the rationality of
human judgments and decisions (Simon, 1955; Thaler & Sunstein, 2003). The knowledgedeficit model of PEB was soon proven to be too simplistic (Kollmuss & Agyeman, 2002). It
is now accepted that only a small fraction of PEB can be explained by environmental
knowledge and awareness, and research indicates that information alone is not very effective
to foster PEB (Kollmuss & Agyeman, 2002; Schultz, 2002b). A substantial body of research
has revealed that decisions are not primarily rational, partly because they are influenced by
limited cognitive resources, incomplete information and heuristics that lead individuals to
make systematic blunders (Kahneman, Knetsch, & Thaler, 1991; Thaler, 1990; Thaler &
Sunstein, 2008; Tversky & Kahneman, 1974). Heuristics are mental shortcuts that individuals
are thought to rely on, “which reduce the complex tasks of assessing probabilities and
predicting values to simpler judgmental operations. In general, these heuristics are quite
useful, but sometimes they lead to severe and systematic errors.” (Tversky & Kahneman,
1974, p. 1124). Heuristic reasoning can lead to cognitive biases, which means that one
3
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
reasons in a way which is not in accordance with norms of logic and probability (Kahneman
& Tversky, 1979). For instance, the so-called optimistic bias makes people believe that they
are less vulnerable to threats related to both climate change and health problems than the
average person (Gifford, 2011; Gifford et al., 2009; Pahl, Harris, Todd, & Rutter, 2005;
Weinstein, 1980, 1982, 1987). The bias is associated with the belief that if something has not
yet happened; it is unlikely to happen also in the future (Weinstein, 1987).
Two Minds: System 1 and System 2
Departures from rationality can partly be understood by the two distinct cognitive
processes thought to underlie human reasoning, namely System 1 and System 2 (Kahneman,
2003). The operations of System 1 are automatic, heuristic-based, intuitive, fast, effortless,
associative, not available to introspection, and often emotionally charged. Because these
processes are often guided by habits, they can be hard to control or change. In contrast, it is
possible to consciously monitor and control the operations of System 2. These operations are
slower, effortful, analytic and flexible (Kahneman, 2003). Contrary to System 1, System 2
permits abstract reasoning and reasons according to logical standards. System 2 is thus known
as the rational system (Evans, 2003).
Rational theories of choice such as the knowledge-deficit model discussed above
imply that people always use the effortful and rational System 2 to process information and
make decisions. However, System 2 processing requires attention and cognitive effort, which
is often scarce in a complex and time-pressed world. Human decisions are therefore more
often guided by the fast System 1, and might for that reason be unconscious and influenced by
external and emotional factors (Kahneman, 2003; Knoll, 2010). As a result, individuals
sometimes make decisions that are not beneficial to them, defined by normative standards.
They do things that they would not have done if they possessed all information, had paid full
attention, had unlimited cognitive abilities and perfect self-control (Ariely, 2008; Milkman,
Rogers, & Bazerman, 2008; Thaler & Sunstein, 2008; Tversky & Kahneman, 1974).
One example of this could be when people pursuit an unsustainable lifestyle despite
their knowledge that our present consumption patterns will contribute to degrading the
environment for future generations. Another example is food choices. How come so many
people in the Western world are obese, and a majority of dieters fail to keep their weight low
(Hill, 2009)? Research suggests that eating and drinking are some of the most mindless
actions we do (Wansink & Sobal, 2007). Many of us simply eat what is put in front of us. For
instance, people tend to eat more than they want to eat, they do not know when they are full,
they eat more when the TV is on, and they eat more when portion sizes are large, even when
4
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
they dislike the food (!) (Kallbekken & Sælen, 2013; Skov et al., 2013; Wansink & Cheney,
2005; Wansink, Just, & Payne, 2009; Wansink & Kim, 2005; Wansink & Sobal, 2007).
Hence, food choices seem to be largely ruled by the “autopilot” processes of System 1 rather
than rational considerations (D. Cohen & Farley, 2008). Information about environmental or
health consequences of the food, which would require System 2 processing, might therefore
play a smaller role when selecting food and eating. On the other hand, intuitive, effortless
responses to situational cues and automatic habits might have a larger impact (D. Cohen &
Farley, 2008; Olstad, Goonewardene, McCargar, & Raine, 2014).
This leads us to the factor often understood as the main barrier to PEB, namely
personal habits (Kollmuss & Agyeman, 2002). In the following section I wish to explore how
habits, as well as the related mechanisms of cognitive dissonance and denial could be seen as
three facets of human cognitive processes that hinder PEB.
Habits: The Same Procedure as Last Year, Miss Sophie?
Recycling, taking the bus to work or buying the same lunch as every day are examples
of habits that are rarely considered at all. Verplanken and Aarts (1999) define habits as
“learned sequences of acts that have become automatic responses to specific cues, and are
functional in obtaining certain goals or end-states” (p. 104). Since habitual behavior is guided
by automatic cognitive processes rather than conscious reasoning, they do not require much
attention (Verplanken, Aarts, & Van Knippenberg, 1997), but they can also be irrational
behaviors, in the sense that they do not always lead to optimal decisions in terms of
maximized utility (Hastie & Dawes, 2010).
Habits work as barriers to PEB for several reasons. First of all because many behavior
changes require breaking old habits and establishing new ones (Dahlstrand & Biel, 1997), but
research suggests that the stronger the personal habit, the less likely is behavior change (Fujii,
Gärling, & Kitamura, 2001; Gifford, 2011; Klöckner & Verplanken, 2013).
Second, habits weaken the connection between intentions and behavior (Verplanken,
Aarts, van Knippenberg, & Moonen, 1998). Thus, intentions to act in an environmental way
might not suffice, if the non-environmental habit is strong. Likewise, food choice is better
predicted by habits than by attitudes and intentions (Köster, 2009), and habits are extra
important for food purchasing decisions under time pressure (Biel, Dahlstrand, & Grankvist,
2005).
Third, habits can lead to selective attention and misperceptions, because they make
people focus on information that confirms their choices, and neglect information that is not in
line with their habits. This “habitual mindset” contributes to the maintenance of habits (Steg
5
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
& Vlek, 2009; Verplanken & Aarts, 1999). For instance, the stronger habit people have for
driving, the less information they seek about other travel modes (Verplanken et al., 1997).
Besides, drivers who commute frequently by car overestimate the time it will take to
commute by public transport (Fujii et al., 2001). Thus, individuals with strong habits may
lack both accurate knowledge about other alternatives and motivation to look for it.
However, it is possible for habits to change when major contextual changes occur.
Fujii et al. (2001) surveyed car drivers before and during an eight-day freeway closure in
Japan, which made their habit of driving behavior impossible. The drivers who drove
frequently prior to the freeway closure were more likely to travel by public transport in the
future if they had largely overestimated the commuting time by public transport. That is; if
riding the bus turned out to be faster than they had anticipated, they were more likely to keep
riding by bus also in the future, than if there were small differences between expected and
experienced commuting time (Fujii et al., 2001). Thus, temporary structural changes might
catalyze durable behavioral change, especially if the new, non-habitual experience is
surprisingly positive. This is where nudging comes in, as we shall see later.
Cognitive Dissonance: Thoughts and Behavior in Harmony
Cognitive dissonance is related to habits and can be understood as a second barrier to
PEB. Cognitive dissonance was introduced as a theory of attitude change, and the basic idea
is that individuals seek consistency within themselves (Festinger, 1962). When cognitions
such as knowledge, opinions or beliefs about the environment, about ourselves, or about our
behavior––are inconsistent, the individual will experience psychological discomfort. This
distress will motivate the individual not only to reduce the dissonance, but also to avoid
information and situations that are likely to increase the dissonance (Festinger, 1962). This
theory can help explain the behavior of the car drivers previously mentioned who did not seek
out information about other travel modes (Verplanken et al., 1997), maybe because this could
have lead to discomfort if they realized that other alternatives were better than their habitual
behavior, driving.
Festinger (1962) points at different strategies used to increase the consistency. As we
shall see, these techniques can potentially hinder environmental consciousness and behavior.
For instance, a person who eats a lot of meat might experience cognitive dissonance when she
learns that meat production has negative impacts on the environment. As a response to this,
she might simply change her action, and reduce or stop meat intake. This will restore balance
between her knowledge and actions. There are strategies, however, for avoiding such
dissonance without having to change meat consumption. She might instead choose to believe
6
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
that meat production is not that bad for the environment, or focus on information saying that
meat production is a small problem compared to other issues, especially in Norway where
animals spend a lot of time outdoors on grass that could not be used for food cultivation
anyway, et cetera. She could also rationalize her behavior by thinking that she needs meat to
be healthy, by saying that other people eat more meat than her, or by thinking that she already
does her environmental duty by using public transport and recycling. The ultimate strategy
would simply be to avoid news and conversations about the issue.
Thus, cognitive dissonance represent a barrier to PEB because it can produce a filter
through which information is processed. Information threatening existing beliefs or behaviors
will be more difficult to accept. The theory can therefore help explain the gap often found
between environmental knowledge and behavior (Kollmuss & Agyeman, 2002; Schultz,
2002b)
Denial: “Knowing and Not Knowing”
The concept of denial originates from psychoanalysis, and was one of the defense
mechanisms proposed by Sigmund Freud (Hall, 1966). It refers to a refusal to accept reality
(Kollmuss & Agyeman, 2002). While both cognitive dissonance and denial can lead to refusal
to acknowledge climate problems, these concepts can be understood as different mechanisms.
Stoknes (2014) suggests that denial can be fuelled by dissonance, but represents a stronger
barrier because it involves refusing the issue completely. Feelings of powerlessness can lead
to denial, such as when people feel that they alone can do nothing to mitigate climate change
(Stoll-Kleemann, O’Riordan, & Jaeger, 2001). Pessimistic climate prospects can be too
frightening and bothersome for individuals to take in over time, and once people realize there
are no easy solutions, they might stop paying attention to climate issues (Stoknes, 2014).
S. Cohen (2001) proposes three varieties of denial, which have been used by among
other Norgaard (2011) to understand denial of climate change. Literal denial is the assertion
that something did not happen or is not true, à la “The climate isn’t changing.” In interpretive
denial, the facts are not denied, but interpreted in a different way. Euphemisms, technical
jargon and word changing can be reinterpreting tools, visible in justifications such as
“Climate change might happen, but it is not caused by humans”, or “The alternative to clean
Norwegian oil is dirty foreign coal.” The last form of denial is implicatory denial, in which
we do not reject that something is happening, but we fail to integrate the knowledge into
everyday life or transform it into action: “I know that climate change happens, and that it is
has human causes. But I have my life to live…” What is minimized here is not information,
but “the psychological, political or moral implications that conventionally follow” (Cohen,
7
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
2001, p. 8). Norgaard’s (2011) ethnographic study of Norwegian small town residents’ way of
coping with climate issues indicates that implicatory denial is widespread in Norway.
Norgaard (2011) observed that people she interviewed could talk about the weather in
general, but were unwilling or unable to talk about climate change in serious ways. She
describes their way of confronting climate issues as “knowing and not knowing, seeing but
not seeing” (p. 104). Thus, implicatory denial gives us the prospect of a situation in which
people have knowledge about the facts of climate change, but live their lives as if they did
not. Thus, as opposed to what rational choice theory predicts, in such states of denial,
information in itself will not suffice to encourage the pro-environmental behavior we are
seeking.
Facilitating Pro-Environmental Behavior through Nudging
How can individuals overcome the barriers made up by habits, cognitive dissonance
and denial, and make more pro-environmental food choices? In order to promote sustainable
food systems, Reisch et al. (2013) suggest that policy instruments include information-based,
market-based and regulatory instruments, as well as the latest addition to the political toolbox:
nudging. Giving people a small ‘nudge’ in the right direction can help individuals make better
decisions for themselves and the environment, as proposed by Thaler and Sunstein (2008).
We have already seen that the information-based approach might have limited potential,
partly due to the barriers made up by habits, cognitive dissonance and denial. Market-based
and regulatory instruments are beyond the scope of this paper. I will therefore discuss
nudging, which is focused on as tool to increase sustainable behavior in the present study.
Thaler and Sunstein (2008) propose that the context in which choices are made, the
choice architecture, largely affects decisions. We have already seen how food choices are
affected by situational cues, such as a TV that is on during food consumption or the size of
the food portions. But other decisions are also affected by choice architecture, such as which
alternative is presented as the default choice, order of presentation and framing of information
(Kallbekken, Sælen, & Hermansen, 2013; Thaler & Sunstein, 2003; Tversky & Kahneman,
1986).
This leads us to the next concept defined by Thaler and Sunstein (2008), namely
nudging. While rational theories of decision making focus on how the effortful System 2
makes decisions, and assume decisions to be coherent and constant over time (Tversky &
Kahneman, 1986), nudge techniques are based on the notion that System 1 often governs
decision-making, and that many decisions therefore are automatic and dependent on
8
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
situational cues (Thaler, Sunstein, & Balz, 2010). The verb nudge means “to push gently”
("nudge," n.d.). Nudging involves using choice architecture to alter “people’s behavior in a
predictable way without forbidding any options or significantly changing their economic
incentives” (Thaler & Sunstein, 2008, p. 6). Nudges still allow for people to make their own
choices, as they are easy interventions that are easy to avoid. Sunstein and Thaler (2003) call
this libertarian paternalism.
Sunstein and Reisch (2013) propose that choice architecture opens up new possibilities
for environmental protection, and that the potential of nudging in promoting PEB might be
even more effective than the standard tools used so far: economic incentives, mandates and
bans. Because food decisions often involve automatic reasoning, and less rational judgment
(Wansink, 2013), these could be especially influenced by a nudge. Therefore, in order to
promote sustainable food choices, it could be promising to design eating environments in
such a way that peoples’ instinctive actions are biased towards the sustainable food, which is
often also the healthy food. Sunstein (2013) mentions different types of nudges to promote
PEB. The next section presents two of these approaches, namely default choices and the
technique of increasing the salience of certain factors and variables, which I will call framing
in this paper.
Defaults: The Path of Least Resistance
Defaults are the option that will be selected or the settings that apply when an
individual make no active choice. Defaults have proven to have a large influence on behavior
(Johnson & Goldstein, 2003), especially when people have little knowledge about a product
(Sunstein & Thaler, 2003).
Pichert and Katsikopoulos (2008) found that most participants chose a green
(renewable) electricity utility when this was presented as the standard option, while the
majority chose a grey (fossil) utility when this was the default option. Food choices are
likewise affected by defaults. When Campbell-Arvai et al. (2014) made attractive, vegetarian
dishes the default choices, a majority (90%) of the participants chose a vegetarian dish,
compared to less than half (40%) when the vegetarian dishes were presented together with
meat dishes in the menu. The study also indicates that defaults are more powerful motivators
than information, as providing information about the environmental impact of meat did not
affect vegetarian food selection, neither when presented alone, nor in combination with the
defaults (Campbell-Arvai et al., 2014).
There are several reasons why defaults affect choices. Some of the effect can be
explained by inertia, or what Thaler and Sunstein (2008) name the “yeah, whatever” heuristic.
9
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
People might simply choose this option because it requires effort to gather information about
all the other possible choices, or because they do not pay attention when choosing (Johnson &
Goldstein, 2012). Considering the fact that an average person makes more than 200 choices a
day concerning food (Wansink & Sobal, 2007), it would simply not be possible to spend a lot
of cognitive effort on each decision. Food defaults could therefore be easily accepted. Further
time pressure could lead to particularly reduced attention, making default choices extra
influential. Another reason why defaults could be influential is implied endorsement (Johnson
& Goldstein, 2012). The decision maker will assume that this alternative is the default for a
reason. Because someone chose this to be first option, people will perceive it as the
recommended course of action (McKenzie, Liersch, & Finkelstein, 2006).
Naming as Framing: Half-full or Half-empty
Framing, or the wording of possible options, is also found to strongly influence
judgment and choices (Sunstein & Thaler, 2003). In decision research, the term frame refers
to “the decision maker’s conception of the acts, outcomes, and contingencies associated with
a particular choice” (Tversky & Kahneman, 1981, p. 453). An illustration of the framing
effect is the different “attractiveness” of a medical treatment depending on whether it is
framed in terms of mortality or survival. Both patients and physicians are more positive to
surgery as treatment if they are informed that the one-month survival rate is 90 percent, than
if told the more frightening-sounding fact that there is 10 percent mortality in the first month
after surgery (McNeil, Pauker, Sox Jr, & Tversky, 1982).
Kahneman (2003) defines framing effects as discrepancies between choice problems
that people after reflection will consider identical. Therefore, I understand the concept of
framing to include also the way in which a choice is presented, such as a product’s name.
Clearly, a name does not change a product, but it might make it appear different for System 1.
Why does framing affect choices? A central assumption in rational choice theory is
invariance in decision making, which means that people will prefer the same options
regardless of the manner in which they are presented (Tversky & Kahneman, 1986).
Accordingly, different frames should not affect choices. However, because decisions often
rely on the processes of System 1, and because these often are emotionally charged and
influenced by situational cues, individuals will process information differently depending on
how it is framed (Kahneman, 2003). The reflective System 2 does not do sufficient work to
find out whether reframing a question would provide a different answer (Thaler and Sunstein,
2008), or whether a product would still be the same with a different name.
10
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Brian Wansink has conducted a series of studies that illustrate how food menu labels
affect food selection and sensory perceptions. In one study, a number of dishes were given
either regular names like “Seafood Filet” and “Red Beans with Rice”, or descriptive names
like “Succulent Italian Seafood Filet” and “Traditional Cajun Red Beans with Rice”. The
descriptive labels increased sales, the customers’ intentions to repurchase the dish and the
amount they were willing to spend on the dish (Wansink, Painter, & Van Ittersum, 2001).
Furthermore, customers who ate dishes with descriptive names rated the food as more
appealing, tastier, more satisfying and even more caloric than those who ate the less
descriptively named counterparts (Wansink, Van Ittersum, & Painter, 2005).
Despite the encouraging results of increasing selection and sensory appeal by giving
food attractive menu labels, vegetarian food is still often simply named “Vegetarian” or
“Vegan” in canteens and restaurants. How will consumers perceive this name? If they are not
familiar to vegetarian food, this is a neutral name that says little about taste. Because it is so
general––it can mean all food that is not meat––it is not likely to produce any pleasant
memories or taste associations. The basic impression the name could invoke is that this dish
does not contain meat. It frames the food in terms of what it is missing. In addition, people
who eat meat hold negative beliefs towards vegetarian and vegan diets (Povey, Wellens, &
Conner, 2001). When meat eaters were asked to rate different diets, many described a
vegetarian diet as healthy, but also as nutritionally unbalanced, boring/bland and low in fat. A
vegan diet was in addition described as extreme, restrictive and unnatural (Povey et al., 2001).
It could therefore be expected that meat eaters will that assume that dishes named
“Vegetarian” or “Vegan” are boring and bland, and rather jump to the rest of the menu.
Moreover, food is an important part of identity (Belasco, 2008), and food habits and
preferences are shaped by personal food experience (Reisch et al., 2013). For people who are
not used to eating vegetarian foods, and do not identify as vegetarians, the name
“Vegetarian”/”Vegan” might give associations of something they “are not”, maybe something
that is “extreme”, as if these food items were meant only for vegetarian customers. Framing
the vegetarian dishes with a name that describes what they actually contain could possibly
make also non-vegetarian customers perceive them as more attractive and non-exclusive
choices.
Nature Connectedness
The last section described how nudges might increase PEB. As already mentioned,
PEB is determined not only by external factors, which nudging would be an example of, but
11
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
also by factors within the individual. Stern (2000) underscores that these factors play
together; so contextual factors can have different meanings to people with different attitudes
or beliefs. But to the best of my knowledge, little is known about whether nudges work
differently for different individuals.
One important individual factor to PEB seems to be the degree to which the individual
feels connected to the natural world. Ecopsychology hypothesizes that people who value and
feel concern for nature will be more willing to protect it (Gomes, Roszak, & Kanner, 1995),
and Schultz (2002a) suggests that the very source of environmental concern is peoples’ belief
of being a part of nature. When environmentalists were asked to describe the sources of their
environmental engagement, experiences in nature and of environmental destruction were
among the factors mentioned most often (Chawla, 1999). Thus, emotional connection to
nature seems to be important in sowing environmental awareness and concern.
The link between emotional involvement with nature and PEB has been examined
quantitatively in larger samples as well, and an individual’s sense of nature connection is
found to strongly predict sustainable attitudes and behaviors (Mayer & Frantz, 2004; Nisbet et
al., 2009). Nature Relatedness (NR) is one of the constructs used to describe such nature
connection, and can be defined as individual differences in cognitive, affective, and
experimental connection with nature. NR is understood to be similar to a personality trait that
is relatively stable over time and across situations (Nisbet et al., 2009; Zelenski & Nisbet,
2014). The trait is similar to an ecological identity or an ecological self, a term coined by
Arne Næss (1973), philosopher and founder of deep ecology. NR comprises both an
internalized identification with nature and an external perspective, which is a sense of agency
concerning human actions and their impact on other living things. It also reflects actual
experiences in and familiarity with nature (Nisbet et al., 2009).
NR is found to be positively related to environmentalism, vegetarianism, participation
in environmental organizations, fair trade and organic purchasing habits, while negatively
related to consumerist behavior (Mayer & Frantz, 2004; Nisbet et al., 2009). It seems like
people who feel that they are a part of nature want to protect it.
Thus, individuals’ sense of belonging to the natural world might play an important
role in shaping PEB, and especially when it comes to purchase behavior and meat
consumption. Therefore the present study asks the question: Does the individual factor Nature
Relatedness and the contextual factor nudging influence vegetarian food purchase in
interaction?
12
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Design of the Study and Hypotheses
Given the theories mentioned in the previous sections, it can be assumed that an
individual’s meat consumption will be affected by both external factors such as choice
architecture, individual factors such as nature relatedness, as well as cognitive barriers such as
personal habits, cognitive dissonance and denial. The aim of the present study is to explore
whether altering the situation in which food choices are made can increase the selection of
vegetarian food, even for people with a strong meat habit. I will also investigate whether these
changes in food presentation will work differently depending on the customers’ nature
relatedness. Three nudges will be tested: I will first make the vegetarian dish default choice,
second I will give it a more descriptive name, and third I will test the default presentation in
combination with the descriptive name.
To the best of my knowledge, only Campbell-Arvai et al. (2014) have tested the effect
of making vegetarian dishes the default choices. Their results indicated that a large majority
chose the vegetarian alternative when this was presented as the default. However, their study
was an experiment conducted in an artificial setting, with a main menu that consisted of only
vegetarian dishes. Thus, available research does not provide a profound understanding of how
food choices are altered by the presentation of the vegetarian options in a real-life setting. The
present study will therefore test the effect of different nudges in a more natural setting in
which meat dishes are easily available choices as well.
The first nudge is to present the vegetarian dish as the default choice. Sunstein and
Reisch (2013) advice choice architects who want to avoid a absolute default rule but
nonetheless want to promote selection of a green option, to “list it first, or use bold or a large
font, or adopt verbal descriptions that make it especially salient or appealing” (p. 22). In the
nudge DEFAULT, I will therefore present the vegetarian dish as the dish of the day, and give
it an extra salient presentation. Thus, the customers are still able to make other choices from
the rest of the menu. This can be considered a “soft” default.
In the second nudge FRAMING I will present the vegetarian option more attractively
by a descriptive name instead of the general name “Vegetarian”. As previously seen, Wansink
et al. (2001) increased the sale of different dishes by giving them more descriptive and
attractive names. It could therefore be expected that giving vegetarian dishes more descriptive
and attractive names would increase their selection. Wansink et al. (2001) suggest different
ways to generate descriptive menu labels, including geographical labels (“Italian”), affective
labels (“home-style”), sensory labels (“tender”) or a mix of these. Because the vegetarian
dishes in the present study are wraps with taco beans, tomato salsa and cilantro, the name
13
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
“Mexican Style” will be tested. The aim is to produce associations of the exotic, tasty and
spicy features of the dish, rather than framing it as a “non-meat” alternative. Olstad et al.
(2014) suggest that nudges might be more powerful if they are implemented in combination.
To explore the extent to which a combination of the default and framing is more potent than
the two nudges alone, the third intervention is a combined DEFAULT+FRAME nudge.
P. C. Stern (2000) argues that researchers should investigate how interventions aiming
at increasing PEB work in interaction with personal attitudes or beliefs. Little is known about
whether the effect of nudging differs with individual factors. As already mentioned, Nature
Relatedness is associated with pro-environmental attitudes and behaviors like vegetarianism
(Nisbet et al., 2009), and could therefore be assumed to predict vegetarian purchase in the
present study. But could there also be an interaction effect between nudging and NR, and
what could this effect look like?
ABC theory predicts personal attitudes to be strongly influential on PEB when context
is neutral – that is, when the situation neither hinders nor favors the PEB (P. C. Stern, 2000).
In contrast, the attitude-behavior relation becomes close to zero if the situation highly
encourages or discourages the PEB, for instance when the PEB is very difficult, timeconsuming or expensive to pursue, or in contrast; if the PEB is very easy, convenient and
lucrative. The nudges in the present study aim to make it easier to choose the vegetarian dish.
Hence, personal attitudes and worldview, such as NR, could be expected to play the largest
role for food selection in the “neutral” control condition. Customers with low NR – those with
the lowest sense of nature connection – should therefore have the lowest chance of selecting
vegetarian in the control condition. But in the nudge conditions, the context will encourage
vegetarian selection. According to ABC theory, personal variables should then become
unimportant. There is therefore a chance that customers with low NR will increase their
selection of vegetarian food in the nudge conditions. Customers with low NR should
consequently be the most affected by the nudge. This is also the most important group to
encourage to PEB, as they are the least likely to be environmentalists or vegetarians (Nisbet et
al., 2009). Customers with high NR, on the other hand, could be expected to have an equal
chance of selecting vegetarian food in all conditions, and hence be less affected by the nudge.
That is; in the control condition their personal attitudes towards nature should encourage them
to select vegetarian, and in the nudge conditions both their NR and the context will favor
vegetarian selection.
Last, could it be that the nudges could help break the barrier to PEB made up by
personal habits, and increase the likelihood that also customers who usually eat much meat
14
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
select vegetarian? We have seen that food decisions often rely on the automatic System 1 (D.
Cohen & Farley, 2008), and that System 1 is habit-driven (Kahneman, 2003). Habits are
especially strong predictors of food purchase under time pressure (Biel et al., 2005). Buying a
meal in a café usually involves time pressure. It can therefore be predicted that customers
with habits to consume much meat will be less likely to select a vegetarian wrap. But System
1 also relies on heuristics and is highly influenced by situational cues (Kahneman, 2003). The
nudges represent a change of situational cues, such that customers’ instinctive judgment and
heuristics should be biased towards the vegetarian food. Therefore, it could be expected that
customers with a strong meat consumption habit are more likely to select vegetarian when
“nudged”. If this is the case, we could expect to find that customers who select a vegetarian
wrap in the nudge condition will report higher average meat consumption than customers who
select a vegetarian wrap in the control condition.
Hypothesis 1a: The nudge DEFAULT will positively affect the selection of vegetarian wraps.
Hypothesis 1b: The nudge FRAMING will positively affect the selection of vegetarian wraps.
Hypothesis 1c: The nudge DEFAULT+FRAME will positively affect the selection of
vegetarian wraps.
Hypothesis 2: Nature Relatedness will moderate the effect of the nudges. Specifically,
nudging can be expected to be particularly effective among individuals with low Nature
Relatedness.
Hypothesis 3: The meat consumption habits of customers who selected a vegetarian wrap will
differ across the nudges. Specifically, customers who choose a vegetarian wrap in the nudge
conditions could be expected to have a stronger meat consumption habit than customers in the
control condition.
15
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Method and Materials
A field experiment was conducted in a café at a Norwegian university selling wraps,
pizzas and other dishes. I, the researcher, approached the customers that had bought a wrap
and asked them to fill out a survey. A total of 353 wrap buyers completed the survey, of
which approximately half were men (51.7%). The participants’ age spanned from 14 to 53
years (M = 25.04, SD = 5.59)1.
All respondents were informed about the purpose of the study and its confidentiality in
advance of answering. They were reminded about their right to withdraw from the study at
any time. The participants did not know that they participated in an experiment, as revealing
the purpose of the experiment in advance could possibly affect customers’ choices. The
changing of the food presentation was not considered a serious deception that could harm the
participants in any way. However, the participants were informed about the possibility to
contact the researcher in case of questions, or if they wanted to know the full purpose and the
results of the study. Moreover, as an incentive to answer the questionnaire, each participant
had the possibility to take part in the drawing of a 1000 Norwegian Kroner voucher.
Design and Procedure
A between-subjects design with four experimental groups (DEFAULT, FRAMING,
DEFAULT+FRAME and CONTROL) randomly rotating the menu presentation of the
vegetarian wraps from day to day was used. Each condition was tested on two different
weekdays2 within a period of three weeks in October and November 2013.
Four different posters were designed by the researcher and printed in two copies each.
They were created in the same design as the original posters in the café. Wrap prices were
equal for all wraps (46 NOK). In every condition, two identical posters were placed on the
long counter in the café: one by the entrance and one where the ordering and payment took
place. For every condition, the text was changed on the posters and at two blackboards placed
behind the counter, one at eye level (Blackboard 1), and one closer to the ceiling (Blackboard
2) (See Appendix 1).
In the CONTROL condition (see Figure 1), the poster presented nine different types of
wraps without highlighting any specific type. The meat-free choices were explicitly named
“Vegan” and “Vegetarian”. The ingredients of each wrap were listed below the name. These
1
5 individuals did not report their gender, 8 individuals did not report their age.
Of 10 experimental days, two days were omitted from analysis. The first day was omitted because it was a Friday. Fridays were excluded because this was the only day in which the competing neighbour café did not offer a
proper vegetarian meal, and because there were few customers on Fridays. The second day was excluded because the café was sold out of vegetarian wraps at 2.30 PM.
2
16
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
wrap names were also written on Blackboard 1. Blackboard 2 contained the text “Wraps. Try
our cold/warm wraps, chipotle style”. This condition was similar to how wraps were usually
presented in this café.
In the FRAMING condition (Figure 2), the name of the vegan and the vegetarian
wraps were changed to “Mexican Style”, on both the poster and the blackboards.
In DEFAULT (Figure 3), vegetarian wraps were again called “Vegan” and
“Vegetarian” but, importantly, made the wraps of the day, and therefore constituted the
default choice on that day. The other wrap types were presented in small font at the bottom of
the poster. Blackboard 1 said: “Wrap of the day: Vegetarian”, with a line circling the text. The
names of all other wrap types were written below. Blackboard 2 said: “Wraps. Try our
cold/warm wraps. Wrap of the day: Vegetarian”.
The DEFAULT+FRAME condition (Figure 4) was similar to DEFAULT, except that
the names “Vegetarian” and “Vegan” were changed to “Mexican Style” as in the FRAMING
condition.
Figure 1. CONTROL poster. Vegetarian wraps are
named “Vegetarian” and “Vegan”. No wrap
options are highlighted.
17
Figure 2. FRAMING poster. Vegetarian wraps are
named “Mexican Style”. No wrap options are highlighted. Dashed line in the figure highlights
difference to CONTROL condition but was not part
of the original poster.
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Figure 3. DEFAULT poster. Vegetarian wraps are
named “Vegetarian” (hot and cold) and made the
wrap of the day. Other wrap options are stated
below.
Figure 4. DEFAULT+FRAME poster. Vegetarian
wraps are named “Mexican Style” (hot and cold)
and made the wrap of the day. Other wrap options
are stated below. Dashed line in the figure highlights
difference to DEFAULT condition but was not part of
the original poster.
Survey
After having ordered their wrap, all wrap buyers received an identical survey. All
participants were asked if they had answered the survey before, to ensure that no customers
participated more than once. The survey took approximately 15 minutes to complete and had
been pilot-tested on a small group (N = 10). Because the instruments originally were in
English, they were translated to Norwegian by the researcher and then back-translated by a
native Norwegian speaker with a university degree in English language. The order of the
instruments was randomized to reduce potential interactions. The following questions and
instruments were included in the survey.3
Demographics. The respondents reported their gender and age.
3
The
questionnaire contained other instruments that will not be further discussed in this paper. See Appendix B
for a copy of the full questionnaire in Norwegian
18
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Meat consumption habit. Two items aimed at measuring the participants’ present
consumption of meat and vegetarian food. To be able to control for general food preferences,
participants were asked to indicate whether they considered themselves to be omnivore,
pescetarian, vegetarian or vegan. These answers were recoded to a binary variable
“vegetarian”, where 0 = eats meat, 1 = does not eat meat (vegetarians, pescetarians and
vegans combined).
Moreover, they were asked to indicate the number of their dinners that contain meat
over an average week. The question was phrased: “During an average week, how many of the
dinners that you eat contain meat? E.g.: If you normally eat meat for dinner every day, write
7. If you never eat meat for dinner, write 0.”
Nature Relatedness. Several scales and measures exist that assess individuals’ nature
connectedness. Tam (2013) did an empirical examination of nine different measures, and
found strong convergence among the concepts. They were strongly inter-correlated,
converged to a single factor, shared similar correlations with criterion variables, and showed
little unique predictive power. Tam (2013) concludes that the measures can be considered
markers of the same underlying construct. Of the various measures, the multidimensional
Nature Relatedness (NR) scale (Nisbet et al., 2009) proved to be among the best measures
(Tam, 2013). The scale has been found to have good internal reliability and test-retest
stability, and has been validated against similar measures (Nisbet et al., 2009; Tam, 2013).
Thus, a 20-item version of the NR scale was used to assess participants’ relationship to
nature. One item from the original scale was not included in the questionnaire because it has
been found to have very weak factor loadings to the rest of the scale (Nisbet et al., 2009). The
original scale has been found to have three subscales. NR-Self reflects the degree to which
people identify with the natural environment, representing an internal, ecological identity
(example item: “My relationship to nature is an important part of who I am.”). NRPerspective represents an extern, nature-related worldview, indicating how one’s relation with
nature is manifested through attitude and behavior (example item: “Animals, birds and plants
have fewer rights than humans”, reversed). The last subscale, NR-Experience, is indicative of
the familiarity people have with nature, reflecting both comfort with and desire to be out in
nature (example item: “I enjoy being outdoors, even in unpleasant weather”). (Nisbet et al.,
2009)
The items were rated using a 5-point Likert scale ranging from 1 (strongly disagree) to
5 (strongly agree), such that higher scores represent stronger connection to nature. About half
of the items are negatively worded, and were reverse scored before analyses. The
19
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
confirmatory maximum likelihood factor analysis supported the threefactor solution (see
Appendix C) found by Nisbet et al. (2009). Two items were excluded based on low loadings
in the factor analysis, and as this resulted in an increased reliability of the subscales. Hence,
18 items were computed into three subscales and a total scale. The reliability estimated by
Cronbach’s alpha was satisfactory for the full scale (α = .84) and for the subscales NRExperience, consisting of 6 items (α = .78) and NR-Self, consisting of 7 items (α = .81), but
not satisfactory for NR-Perspective, consisting of 5 items (α = .48). The unsatisfactory low
reliability was kept in mind during further analyses, and analyses were conducted both with
the three subscale and with the total scale, as Nisbet et al. (2009) suggests that the scale will
need revision to work optimally as a multifactorial scale. See Table 1 for psychometric
properties of the different NR-scales.
Wrap. At the end of the questionnaire, the participants were asked to respond to
questions about their wrap habits and satisfaction. They were asked to indicate which wrap
they had just bought, from a list containing all available wraps. For the vegetarian options, the
names on the list were “Vegan / Mexican style (hot)” and “Vegetarian / Mexican style
(cold)”.
Last, participants indicated to which extent they enjoyed their wrap, on a 6-point scale
ranging from 1 = very much, to 5 = not at all. In addition there was an option 6 = I have not
eaten it yet. Before analysis, responses 6 were excluded.
Table 1
Psychometric Properties of the Study Variables
Variable
M
SD
Range
Possible Observed
Skew
α
N
Meat Dinners per Week
4.55
2.19
0–7
0–7
-.826
-
342
Nature Relatedness Total
3.46
.583
1–5
1.78 – 4.72
-.269
.837
344
Nature Relatedness Experience
3.30
.824
1–5
1.00 – 4.83
-.356
.780
344
Nature Relatedness Self
3.35
.754
1–5
1.14 – 5.00
-.370
.808
344
Nature Relatedness Perspective
3.80
.680
1–5
2.00 – 5.00
-.167
.477
344
20
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Analysis
Before running the main analyses, a missing value analysis was performed for the NR
items. 9 subjects were excluded from this analysis because they only answered to the
questions regarding food habits and wrap choice, and thus had missing values for the rest of
the survey. All items had missing values, with the lowest percentage being 0.6% missing. The
item NR_Perspective_1 had the largest amount of missing values, with 10 subjects (2.9%) not
responding. The result of Little’s MCAR test was non-significant (p = .14), indicating that the
missing data was missing at random. Missing values were thus replaced using the
expectation-maximization (EM) method. EM estimates the parameters by maximizing the
complete data log likelihood function (Dong & Peng, 2013).
All analyses were conducted with SPSS Version 21. After a descriptive analysis of
characteristics of the customers, the wrap choices of the customers in each nudging condition
were explored descriptively in crosstabs. A chi-square test was used to determine whether
wrap choice differed by nudge condition. Next, binary logistic regressions with interaction
terms were used to test whether Nature Relatedness moderated the effect of nudge condition
on wrap choice. Last, as logistic regression could not be used to test whether habits moderated
the effect of nudge on wrap selection––because meat habits differed significantly across the
experimental groups (F (3, 342) = 7.77 p < .001)––analysis of variance (ANOVA) was used
to test for differences in meat habits between the nudge groups and between the customers
who bought a vegetarian wrap versus meat wrap.
Results
In terms of self-reported food consumption habits, most participants categorized
themselves as omnivores (89.5%). The remaining participants avoided animal products to
different extents, self-identifying either as vegetarian (5.7%), vegan (1.4%) or pescetarian
(3.4%). In further analysis, the three meat-avoidant categories are combined in one category
called “vegetarians” (10.5%).
Wrap satisfaction did not differ between customers who bought vegetarian wraps and
meat wraps (F (1, 247) = .05, p = .63); or between the nudge conditions F (3, 247) = 1.40, p =
.24. Furthermore, the interaction effect was non-significant (F (3, 247) = .58, p = .53),
indicating that the effect of wrap choice on wrap satisfaction did not differ by nudge condition
(see Figure 5). Vegetarians were excluded in this analysis. For more information about the
customers, see Appendix C.
21
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Figure 5. Wrap Satisfaction. The bars
display wrap satisfaction for customers
who ate meat wrap versus vegetarian
wrap in the different nudge conditions.
Vegetarian customers are excluded
(n = 247). Error bars indicate +/- 1 SD.
Hypothesis 1: Wrap Selection by Nudge
The first hypotheses were that the three
nudges would positively affect the
selection of vegetarian wraps. The
percentage of customers selecting a
vegetarian wrap in the different conditions
is displayed in Figure 6 and Table 2.
These analyses were run only for meat
eating participants, as no change in meat
consumption was plausible for vegetarians.
A chi-square test of independence was
performed to examine the relation between
Figure 6. Vegetarian Wrap Selection. Percentage of
omnivore customers who bought a vegetarian wrap
in the different nudge conditions. (n = 315).
nudge condition and wrap choice. The test was initially performed for nudge condition in
general, and then separately for each condition compared with control, to examine the effect
of each nudge to the baseline condition. There was a marginally significant difference
between the percentage of vegetarian wrap buyers by nudge χ2 (3, n = 315) = 7.36, p = .06.
However, following up on this effect, no difference in the percentage of customers eating
vegetarian was observed in the DEFAULT, χ2 (1, n = 144) = 0.28, p = .87. FRAMING,
22
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
χ2 (1, n = 173) = 1.52, p = .22, or
DEFAULT+FRAME condition, χ2 (1, n =
Table 2
144) = 1.81, p = .18, compared to the
Nudge Condition and Wrap Choice. n = 315
Wrap choice
Meat wrap Vegetarian
Count (%)
wrap
Count (%)
Total wraps
Count (%)
Control
62 (84.9%)
11 (15.1%)
73 (100%)
Default
61(85.9%)
10 (14.1%)
71 (100%)
Framing 91 (91.0%)
9 (9.0%)
100 (100%)
Default
Frame
54 (76.1%)
17 (23.9%)
71 (100%)
suggests that the highest percentage of
Total
268 (100%) 47 (100%)
customers ate vegetarian in the
Note. Vegetarian customers are excluded.
control group. In contrast, a significant
difference was only observed between
DEFAULT+FRAME and FRAMING,
χ2 (1, n = 171) = 7.20, p = .01), with more
customers selecting vegetarian in
DEFAULT+FRAME.
Summing up, although Figure 6
315 (100%)
DEFAULT+FRAME condition, no
statistically significant differences were observed
comparing the nudge conditions with the control group.
Hypothesis 2: Nature Relatedness as Moderator
The second hypothesis was that customers’ NR score would moderate the effect of
nudge condition on wrap choice, in such a way that the nudge would have a positive effect on
vegetarian choice when NR was low. The moderation effect was tested with separate
moderated logistic regressions for each nudge with wrap choice as dependent variable. In
each regression, the experimental condition was coded as a binary categorical variable, so
each condition (FRAMING and DEFAULT+FRAME, coded 1) was tested against the
CONTROL condition (coded 0). The logistic regression model of the DEFAULT nudge is not
reported, because no direct effect or interaction effect of the nudge was found, and the model
had poor goodness of fit. It can, however, be found in Appendix D.
In the first step of the regression, NR Perspective, NR Experience and NR Self,4 and
nudge condition (FRAMING/DEFAULT+FRAME) were predictors of wrap choice. The
continuous NR variables were mean centered to facilitate interpretation. The variable
vegetarianism (0 = no, 1 = yes) was included as covariate to control for general food
preferences. In the second step, interaction terms of the nudge conditions and the respective
NR variables were added. There was no difference in NR between the different experimental
conditions, neither for NR-Perspective (F (3, 344) = .04, p = .97), NR-Experience
4
The model was also run with the total NR scale as predictor. See Appendix C for this logistic regression model.
23
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
(F (3, 344) = .55, p = .65) or NR-Self (F (3, 344) = 1.41, p = .24), and NR could therefore be
included in the model.
DEFAULT+FRAME. The main assumptions for logistic regression are absence of
multicollinearity and outliers, and a sufficiently large sample size (Pallant, 2010). As advised
by Field (2009) the model with the same variables was ran in linear regression to check for
multicollinearity, and yielded no problems of correlation between the predictor variables.
Two cases had high standardized residuals (larger than +/- 3), and were as advised by Pallant
(2010) checked. These were considered to look probable. The analysis was tried repeated
without the cases. However, because the new model yielded a weaker classification accuracy
rate than the model with all cases included, all cases were retained.
Including interaction terms in the model increased the explained variance from the
model, from between 27.5% (Cox & Snell R2) and 40.2% (Nagelkerke R2) to between 34.4
and 50.2% respectively. The accuracy rate rose from 84.3% correctly identified cases to
86.8%. A non-significant Hosmer and Lemeshow test (p = .07) indicated satisfactory fit of the
model.
Table 3
Binary Logistic Regression of Vegetarian Wrap Selection in Control versus Default Frame
Condition. n = 159
95% CI for Odds Ratio
Odds
B
SE (B)
Wald
df
p
Ratio
Lower
Upper
Constant
-2.45
.50
23.95
1
.000
.09
Default Frame
nudge
NR Perspective
1.03
.58
3.12
1
.077
2.81
.89
8.81
2.54
.80
10.17
1
.001
12.63
2.66
60.03
NR Experience
.24
.60
.16
1
.687
1.27
.39
4.12
NR Self
.37
.71
.28
1
.599
1.45
.36
5.81
Vegetarianism
4.81
1.24
15.01
1
.000
122.79
10.77
1400.24
-2.79
.90
9.68
1
.002
.06
.01
.36
-.15
.74
.04
1
.834
.86
.20
3.64
-.78
.83
.88
1
.347
.46
.09
2.33
NR Perspective x
Default Frame
NR Experience x
Default Frame
NR Self x
Default Frame
Note: R2 = .07 (Hosmer & Lemeshow), .34 (Cox and Snell), .50 (Nagelkerke). Model χ2 (8) = 67.00, p < .001.
The full model (Table 3) containing all predictors was statistically significant, χ2 (8, n
= 159) = 67.00, p < .001, indicating that the model was able to distinguish between
respondents who bought and did not buy a vegetarian wrap. Only two of the independent
variables by themselves made a unique statistically significant contribution to the model.
24
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Little surprisingly, vegetarianism was the strongest predictor of choosing a vegetarian wrap.
Moreover, NR Perspective was positively related to buying a vegetarian wrap. Here, for every
additional point at the NR Perspective score (ranging from 1 to 5), customers were almost 13
times more likely to select a vegetarian wrap. The DEFAULT+FRAME nudge by itself
marginally predicted vegetarian wrap selection. Customers who bought a wrap in this
condition were almost 3 times more likely to buy a vegetarian wrap, controlling for all other
factors in the model. This means that although cross-tabs were insignificant, participants in
this conditions did seem to have a higher likelihood to buy the vegetarian wrap when the
other variables were controlled for.
Supporting my predictions, there was a significant interaction effect between NR
Perspective and the DEFAULT+FRAME nudge. As illustrated in Figure 7, participants with
low NR Perspective had higher probability of selecting a vegetarian wrap in
DEFAULT+FRAME. Also participants with medium NR Perspective had a higher probability
of selecting a meat wrap in DEFAULT+FRAME, although this difference was only
marginally significant.
0,5
z = -1.42 p = .156
0
Odds Ratio
-0,5
-1
z = 1.75, p = .080
-1,5
Low NR Perspective
-2
Medium NR Perspective
-2,5
z = 2.92, p = .004
-3
High NR Perspective
-3,5
-4
Control
Default Frame
Nudge
Figure 7. Interaction of Nature Relatedness Experience and Nudge Condition on Wrap Selection.
Simple slopes of probability for buying a vegetarian wrap in DEFAULT+FRAME vs.
CONTROL: respondents with high (M + 1SD), medium (M) and low (M - 1SD) NR Perspective.
FRAMING. Assumptions for logistic regression were met. Multicollinearity was not
present. There were some highly influential cases, but these were checked and considered
probable. I tried to run the model excluding potential outliers that had high standardized
values and Cook’s distance above the advised values (Pallant, 2010). However, because this
25
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
yielded only trivial changes in the accuracy rate of the model, the original model was
estimated.
The full model (Table 4) containing all predictors was statistically significant, χ2 (8, n
= 192) = 95.39, p < .001. In Step 1, the model explained between 35.4% (Cox & Snell R2) and
56.2% (Nagelkerke R2) of the variance, and correctly classified 89.6% of cases. Including
interaction terms increased explained variance from the model, from between 39.2% (Cox &
Snell R2) and 62.1% (Nagelkerke R2). The accuracy rate rose to 92.7% correctly identified
cases. Hosmer and Lemeshow test was not significant (p = .57), indicating good fit of the full
model.
Table 4
Binary Logistic Regression of Vegetarian Wrap Selection in the Control Versus the Framing Condition.
n = 192
95% CI for Odds Ratio
Odds
B
SE (B)
Wald
df
p
Lower
Upper
Ratio
Constant
Framing nudge
NR Perspective
NR Experience
NR Self
Vegetarianism
NR Perspective x
Framing
NR Experience x
Framing
NR Self x
Framing
-2.43
-.38
2.50
.24
.38
4.65
.50
.72
.79
.59
.70
1.27
23.65
.28
10.02
.16
.30
13.34
1
1
1
1
1
1
.000
.597
.002
.687
.585
.000
.09
.68
12.22
1.27
1.47
104.86
.17
2.59
.40
.37
8.64
2.80
57.58
4.07
5.80
1272.81
-1.52
1.01
2.24
1
.135
.22
.03
1.61
-2.52
1.03
6.02
1
.014
.08
.01
.60
2.66
1.24
4.57
1
.033
14.29
1.25
163.71
Note: R2 = .57 (Hosmer & Lemeshow), .39 (Cox and Snell), .62 (Nagelkerke). Model χ2 (8) = 95.39, p < .001.
As shown in Table 4, two of the independent variables made a significant contribution
to the model by themselves (NR-Perspective and Vegetarianism). As in the first model,
vegetarians were more likely to buy a vegetarian wrap than meat eaters, and for every
additional point at the NR Perspective score, buying a vegetarian wrap was 12 times as likely.
The FRAMING nudge was not a significant predictor. There was an interaction effect
between both NR Experience and the FRAMING nudge (Figure 8), and between NR Self and
FRAMING (Figure 9). People with low NR Experience had higher probability of selecting a
vegetarian wrap in FRAMING, whereas customers with high NR Experience had lower
probability of selecting vegetarian in this condition. Both effects were marginally significant.
The interaction effect of NR Self was the other way around: Customers with high NR Self
were more likely to select vegetarian in FRAMING than in CONTROL, whereas customers
with low NR Self were less likely. Also these effects were only marginally significant.
26
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
0
-0,5
z = 1.95, p = .052
Odds Ratio
-1
-1,5
Low NR Experience
-2
Medium NR Experience
-2,5
z = -.51, p = .613
-3
High NR Experience
-3,5
-4
-4,5
z = -1.95, p = .051
-5
Control
Framing
Nudge
Figure 8. Interaction of Nature Relatedness Experience and Nudge Condition on Wrap Selection.
Simple slopes of probability for buying a vegetarian wrap in FRAMING vs. CONTROL: respondents
with high (M + 1SD), medium (M) and low (M - 1SD) NR Perspective.
1
z = 1.83, p = .068
Odds Ratio
0
-1
Low NR Self
z = .50, p = .618
-2
Medium NR Self
High NR Self
-3
-4
z = -1.71, p = .088
-5
Control
Nudge
Framing
Figure 9. Interaction of Nature Relatedness Self and Framing Nudge on Wrap Selection. Simple
slopes of probability for buying a vegetarian wrap in FRAMING vs. CONTROL: respondents with high
(M + 1SD), medium (M) and low (M - 1SD) NR Perspective.
Summing up, customers’ Nature Relatedness score seemed to affect wrap choice both
directly and in interaction with the FRAMING and DEFAULT+FRAME nudges, but these
interactions were only partly consistent with the hypothesis.
Hypothesis 3: Meat Habits of the Customers
The third hypothesis was that the meat consumption habits of participants who
selected a vegetarian wrap would differ across the nudge conditions. Customers who ate a
vegetarian wrap in the control condition were expected to have a habit to eat less meat than
customers who ate a vegetarian wrap in the nudge conditions. A 4 (Nudge: CONTROL,
FRAMING, DEFAULT, DEFAULT+FRAME) x 2 (Wrap choice: Meat, Vegetarian)
27
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
ANOVA was conducted with meat consumption habit per week as independent variable,
controlling for whether participants were vegetarian or not. The results showed that both
nudge (F (3, 342) = 7.77 p < .001, eta squared = .07), wrap choice (F (1, 342) = 38.97 p <
.001, eta squared = .11) and the interaction between nudge and wrap choice (F (3, 342) = 6.57
p < .001, eta squared = .06) significantly predicted how often participants ate meat per week.
Following up on the indirect effect, Figure 10 shows that only in the DEFAULT+FRAME
condition participants who bought the vegetarian wrap (M = 4.78, SE = .30) did not report a
generally lower tendency to eat meat in their daily lives than their peers buying a meat wrap
(M = 4.87, SE = .20). In all the other conditions, participants who bought a vegetarian wrap
(control: M = 2.74, SE = .33; framing: M = 3.22, SE = .40; default: M = 2.81, SE = .37) were
those who generally used to eat less meat in their daily lives than those buying a meat wrap
(control: M = 4.73, SE = .19; framing: M = 5.23, SE = .16; default: M = 4.70, SE = .19).
When results were considered separately for the four experimental groups, there was a
significant difference (p < .001) in meat consumption between customers who ate vegetarian
versus meat wrap in FRAMING, DEFAULT and the control group. However, there was no
significant difference between the wrap customers in DEFAULT+FRAME, F (1, 71) = 6.40,
p = .43, indicating that customers who ate vegetarian or meat wrap were equal in terms of
their usual eating habits.
Figure 10. Customers’ Meat Consumption Habit. The simple slopes illustrate the reported meat consumption of
customers who ate a vegetarian wrap compared to customers who ate a meat wrap, in the different nudge
conditions. Meat consumption is the reported number of dinners containing meat over an average week.
Vegetarianism is controlled for. N = 342.
28
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Discussion
The first objective of this study was to test three nudges in a real situation. In the first
nudge, vegetarian wraps were set up as “wrap of the day”, and thus made default choice. In
nudge number two the names of the vegetarian wraps were changed from “Vegan” and
“Vegetarian” to “Mexican Style”. In the third nudge a combination was tested, such that
“Mexican Style” was made the wrap of the day. All nudges were expected to lead to increased
sales of vegetarian wraps relative to meat wraps. Moreover, the study aimed to explore
whether the effect of the nudges was moderated by the customers’ sense of connection to
nature. The nudges were expected to be particularly effective among individuals with low
Nature Relatedness. Another aim was to investigate whether the nudges could encourage
vegetarian food selection also among customers with a strong meat consumption habit. It was
therefore hypothesized that customers who were nudged to select vegetarian would eat more
meat in their daily life, than customers who bought a vegetarian wrap in the control condition.
Was there an effect of these nudges? In short, no and yes. The nudges themselves had
no statistically significant effect on the overall frequency of vegetarian food selection. Thus,
the first hypotheses (1a, 1b and 1c) were rejected. However, when the results are broken down
by the answers given by customers in the questionnaire, it becomes evident that the nudges
worked differently for different people. That is to say, customers’ position on the Nature
Relatedness scale moderated the effects of two of the nudges, such that customers with low
connection to nature were more prone to be affected by the DEFAULT+FRAME and partly
by the FRAMING nudge. Thus, the second hypothesis was partly supported for two of the
nudges. Moreover, the DEFAULT+FRAME nudge seemed to attract a different type of
customers to select vegetarian than the other conditions. Customers who were nudged to
select a vegetarian wrap reported to eat the same amount of meat in their daily life as those
who bought a meat wrap, something that indicates that customers with strong meat
consumption habits were more likely to select vegetarian in this nudge. The third hypothesis
was therefore also supported, but only for the DEFAULT+FRAME nudge.
I will first discuss the effects of the nudge interventions separately, before I turn to the
moderation effect of the customers’ nature relatedness. Then follows a discussion on the
relation between nudging and habits, before a general discussion of nudging as a mean to
reduce psychological barriers to PEB. Last I will point to some implications and suggest a
path for future research, before I look at the strengths and limits of the present study.
29
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
The Effect of the Nudges
Default. Presenting the vegetarian wraps named “Vegetarian” as the default choice
did not increase the percentage of customers who selected this wrap. Thus, merely presenting
the vegetarian dish as the dish of the day does not seem to be enough to motivate more
individuals to choose vegetarian. This is contrary to the result of Campbell-Arvai et al.’s
(2014) study, in which a majority of those nudged chose a vegetarian dish when it was
presented as default.
Their study, however, differed from the present study in a number of ways. First of all,
it was a lab experiment with limited field validity. The participants were taken to an artificial
setting, a room where they were asked to select a dish as if they were ordering their actual
lunch or dinner (Campbell-Arvai et al., 2014). Their choices were likely influenced by the
fact that they neither were going to pay for the dish nor eat it. Moreover, reactivity bias is a
common research problem in which participants alter behavior due to the awareness that they
are observed. Some participants were given a menu consisting of only vegetarian dishes, but
had to consult a different menu placed 3,5 meters away from their table if they wanted a meat
dish. Whereas more customers would possibly want to seek out the meat menu in a real-life
situation, social desirability bias or expectation bias could encourage participants to choose an
item of the menu they were given first by the researcher, especially because they were not
going to get the dish anyway. What is more, the study did not give the participants in the
default condition a real choice, as the menu only displayed vegetarian dishes, and no meat.
This could be said to counter a central point in nudging, namely that a nudge should not
restrict freedom of choice (Thaler & Sunstein, 2008). Besides, if a café’s menu gives the
impression of offering only vegetarian dishes, meat-loving customers might not choose to go
to eat in this café in the first place.
The present study was a realistic field experiment where participants actually selected
and paid for the food, without knowledge of participating in a study, and meat dishes were
easily available choices. The results of the present study could indicate that it is more difficult
to enhance vegetarian selection through default nudges in a real-life situation than in an
experimentally controlled setting. It could also be, however, that the default menu in
Campbell-Arvai et al.’s (2014) study worked better because it presented five different
vegetarian dishes, whereas only two highly similar vegetarian options were available in the
present study. A wider range of vegetarian default dishes could have led more customers to
select a vegetarian dish.
30
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
We could also surmise that very attractive vegetarian dishes would be preferred by
most customers. In Campbell-Arvai et al.’s (2014) experiment, vegetarian dishes rated as very
attractive were selected far more often than their non-attractive counterparts. It could be that
the vegetarian bean wraps in the present study simply were found less attractive for other
reasons than the fact that they represented a vegetarian option.
Importantly, people who eat meat can hold negative attitudes towards a vegetarian
diet, and believe it to be boring and low in fat (Povey et al., 2001). As reasoned earlier, the
name “Vegetarian” might only attract customers already familiar with and positive to
vegetarian food, and thus not making a difference in selection for other customers, even when
it is made the wrap of the day.
In addition to attractiveness, the salience of the default option, for instance the
background rate of how many chooses the dish of the day compared to all other menu options,
can be important. In the present study, the menu consisted of seven non-vegetarian dishes in
addition to the vegetarian wraps of the day. “Wrap of the day” might therefore not be a very
“strong” default, given that customers could find many other wrap options in the menu. More
vegetarian dishes relative to meat dishes would maybe represent a stronger default.
Framing. Would more customers select a vegetarian wrap when it was given the
name Mexican Style instead of the less attractive names Vegetarian/Vegan? This nudge also
failed to produce any significant change in selection. The result deviates from previous
studies that have found attractive and descriptive food names to increase selection (Wansink,
Just, Payne, & Klinger, 2012; Wansink et al., 2001). On the other hand, there are other studies
that have found no impact of framing through descriptive menu labels on healthy food
selection (Olstad et al., 2014). The effectiveness of food labels likely depends on several
factors in the choice situation, as well as the name that is tested. “Mexican Style” was not a
very well thought out name, nor tested. To increase the attractiveness food labels should give
customers associations of something they know is tasteful (Wansink et al., 2001), it is
uncertain to what degree Mexican Style managed to do this. Optimally, the study should have
used pilot-tested names.
Before the experiment the vegetarian wraps in the café were named “Vegetarian” and
“Vegan”. One additional possibility is therefore that those who normally chose these wraps
did not understand that “Mexican Style” was the same wrap, and thus resorted to order
something else instead. If this happened, it could have helped nullify the effect of those who
actually were affected by the nudge. I will return to this point later.
31
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Default Frame. Would more customers select the vegetarian wrap when it was both
named Mexican Style and presented as the wrap of the day? The percentage of customers who
chose a vegetarian wrap in DEFAULT+FRAME did not differ significantly from the control
condition. But looking at the percentages there seemed to be at least a tendency towards more
sales of vegetarian wraps with this nudge. Indeed, when controlling for Nature Relatedness in
the logistic regression, DEFAULT+FRAME had a direct, marginally significant effect on
vegetarian wrap selection. It is also notable that significantly more customers chose a
vegetarian wrap in DEFAULT+FRAME compared to the FRAMING, indicating that it was
more effective than the FRAMING-only nudge.
Why did the combination seem slightly more potent than the two other nudges
separately? It is possible that the name Vegetarian in the DEFAULT nudge was so
unappealing that this option was not considered despite its eye-catching presentation. What is
more, simply presenting the wrap as Mexican Style in the FRAMING nudge might have
attracted little attention from customers when other dishes well known to the regular
customers were given equal attention on the menu. The combination, however, could have
succeeded in both making the dish sound more appetizing and of interest for all customers, as
well as making it more visible. The result could be that customers paid more attention to the
dish, and maybe selected it because of convenience or laziness – the “yeah whatever”
heuristic (Thaler & Sunstein, 2008).
A high number of regular customers could have weakened the effect of the nudges.
Defaults are more effective when customers have little knowledge about a product (Sunstein
& Thaler, 2003). Since habits are strongly predictive of food purchase (Biel et al., 2005), it is
possible that the effect of the nudges was reduced because many customers were familiar with
the café’s menu. Almost half of the participants reported to be visiting the café at least 2-3
times a month, and more than half reported that they bought the same wrap “sometimes”,
“often” or “every time” they visited. 14% bought the same wrap every time. Many customers
probably ordered without even looking at the menu, thus bypassing the nudge.
It is also possible that “Mexican Style” attracted the customers because it was a new
name, and therefore appeared to be a new item in the menu. Ideally, the experiment should
have been held for a longer time period to investigate whether the results reflected a novelty
effect that eventually would decrease.
To sum up, although none of the nudges significantly increased vegetarian food
selection, DEFAULT+FRAME seemed to be the most powerful nudge among the three. This
indicates that in order to enhance vegetarian food choice, it may not be enough to simply
32
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
change one aspect of the choice situation. As suggested by Olstad et al. (2014), nudges could
be more efficient in combination, or along with other incentives.
Interaction between Nature Relatedness and the Nudges
While the nudges alone seemed to have limited effect on customers’ food choice,
results from the logistic regression analysis, partly supporting Hypothesis 2, were more
encouraging. The results reveal that a certain type of customers––the customers who scored
low on one of the three NR scales, namely NR Perspective––were more prone to be affected
by the DEFAULT+FRAME nudge. That is, customers who reported the least nature-related
attitudes and behaviors, such as care about other species, had a higher probability of selecting
a vegetarian wrap when it was called Mexican Style and presented as the wrap of the day than
in the control condition. Customers with medium NR-Perspective were also more likely to
choose vegetarian when nudged, although this effect was only marginally significant. In
contrast, individuals with high NR-Perspective––those who possess a highly nature-related
worldview––were not affected by the nudge. These customers were equally likely to select a
vegetarian wrap whether they were nudged or not.
This is a noteworthy finding regarding choice architecture, as it indicates that nudging
might be most effective for those who are otherwise the least likely to act in a proenvironmental way. NR-Perspective was, as anticipated, a strong predictor of vegetarian
purchase by itself in all conditions in the present study. This was no surprise, as NRPerspective “reflects an external, nature-related worldview, a sense of agency concerning
individual human actions and their impact on all living things” (Nisbet et al. “2009), p. 723),
and is related to love of animals, vegetarianism and environmentalism (Nisbet et al., 2009).
The fact that participants with low NR-Perspective were more affected by the nudge therefore
implies that the nudge worked on people who are otherwise unlikely to eat vegetarian or care
about the environment. This finding represents, to my knowledge, new and important
understanding of how nudges can be expected to work.
Though not tested in this study, an other way to understand the results could be that
the nudge downplays the effect of NR, such that NR no longer predicts food choice when the
vegetarian food is presented with the DEFAULT+FRAME nudge. The nudge increases the
influences of contextual factors on choice at the expense of personal factors. This is in line
with the ABC theory, which assumes that personal variables such as attitudes will predict
PEB when the context is neutral, while attitudes will be less important if the behavior is
highly supported by the context (P. C. Stern, 2000).
33
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Considering all results together, the DEFAULT+FRAME seemed to be the most
promising nudge, especially for groups otherwise less likely to select a vegetarian meal. But
the logistic regression found support for two more interaction effects, which indicate that also
the FRAMING nudge increased vegetarian wrap selection among certain customers. These
effects were partly contrary to the theoretical expectations, as customers with high NR-Self
actually were less likely to select vegetarian when nudged. We can only speculate why this
happened, but it could be because the customers were not aware that “Mexican Style” was
vegetarian, as I will discuss in a while. Because FRAMING did not affect overall wrap
selection, and because the effects of the NR scales were unclear and only marginally
significant, I will not discuss these results in further detail. What all the interaction effects
point to, however, is that nudges work differently for different people. The results underscore
P. C. Stern’s (2000) concern that contextual factors aiming at promoting PEB should be
investigated in relation to individual factors to reveal their dependence on personal attitudes
or beliefs.
It should also be noted that the results contribute to the understanding of NR, as they
reveal that personal relationship to nature is important for actual PEB, and not only for selfreported PEB, as previous studies have shown (Nisbet et al., 2009). Interestingly, only NRPerspective was predictive of vegetarian purchase, but that could be specific for this particular
behavior, as NR-Perspective specifically reflects attitudes towards animals. The other NR
scales could be equally important for other forms of PEB, such as identifying as an
environmentalist, as other studies has found (Nisbet et al., 2009).
Individuals with high NR-Perspective had a stable preference to buy a vegetarian dish
regardless of the way in which it was presented, while people with low NR-Perspective
bought it more often when it was presented more attractively and as the default choice.
Therefore, while nudging can give people a small push towards the sustainable choice in the
decision moment, increasing peoples’ overall connection to nature may provide a deeper,
more intrinsic motivational force to protect nature and the environment in the long run and
across different contexts, as suggested in ecopsychology and by several theorists (Crompton
& Kasser, 2009; Gomes et al., 1995; Schultz, 2002a).
Breaking Meat Habits with a Nudge
The third hypothesis was that customers who were nudged to buy a vegetarian wrap
would have generally stronger meat consumption habits than customers who bought a
vegetarian wrap in the control condition. This hypothesis was supported, but only for the
DEFAULT+FRAME nudge. That is; in all conditions except DEFAULT+FRAME, there was
34
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
a significant difference in average meat consumption between customers who ate a vegetarian
wrap and customers who ate a meat wrap. Customers who selected a vegetarian wrap
generally ate small amounts of meat in their daily lives, while those who selected a meat wrap
ate larger amounts of meat in their daily lives. This is not surprising, since habits are strong
determinants for food behavior (Biel et al., 2005; Köster, 2009).
But in the DEFAULT+FRAME condition, customers who ate vegetarian wraps
reported to eat in average the same amount of meat in their daily lives as customers who
chose a meat wrap. Meat habit no longer seemed to be related to wrap choice, and it therefore
seems like DEFAULT+FRAME made “high-meat eaters” consider the vegetarian choice
more often. This indicates that the “Wrap of the day: Mexican Style” appealed to a different
group of customers than the other menus that were tested.
This finding might have several explanations. The food-judging processes of System 1
rely on both habits and situational cues (D. Cohen & Farley, 2008; Kahneman, 2003). The
nudges represent situational cues. It is possible that the DEFAULT+FRAME nudge
outweighed the strong effect of personal habits on food purchase, and therefore made even
customers with a strong meat habit consider the vegetarian wrap out of convenience.
As discussed previously, the menu item “Vegetarian” could push away meat-loving
customers, because they might associate “Vegetarian” with a boring and non-satisfying dish,
thinking that is meant exclusively for vegetarian customers. The name “Mexican Style”, in
combination with a more visible presentation, could have helped the vegetarian food appear
more interesting especially for meat-lovers.
Nevertheless, it needs to be considered that the customers who ordered Mexican Style
might not have been aware that it was vegetarian, as this was not explicitly stated in the menu.
A larger percentage of customers reported to be vegetarian in the control condition (21.9%)
than in the nudge conditions (FRAMING: 6.0%, DEFAULT: 11.2%, DEFAULT FRAME:
9.8%). These percentages indicate that less vegetarians bought a wrap overall in the nudge
conditions than in the control condition. Even though there were few vegetarians also in
DEFAULT, when the vegetarian wrap was explicitly named “Vegetarian”, these percentages
could imply that some vegetarians did not understand that “Mexican Style” was vegetarian,
and instead bought something else in the café, such as the explicitly named “Vegetarian
pizza”. If vegetarians misunderstood the posters, non-vegetarians most likely could have done
so too. Even though the wrap ingredients were stated clearly, we know that food selection is a
quick and almost mindless process (D. Cohen & Farley, 2008; Wansink & Sobal, 2007).
Many customers probably made the decision of which wrap to buy after a short glimpse at the
35
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
menu, and some might have missed the fact that this dish did not contain meat. Though not
empirically supported, there seems to be a dominant norm in our society that food should
contain meat to be a proper meal. Especially men tend to view meat as an essential part of a
meal (Kubberød, Ueland, Tronstad, & Risvik, 2002). It is therefore possible that customers
assumed that the “wrap of the day” had to contain meat, because someone had made it the
default choice. In spite of this, customers were equally satisfied whether they ate a vegetarian
wrap or meat wrap, and whether the wrap was named “Vegetarian” or “Mexican style”.
Therefore, even if someone was not aware that they bought a vegetarian wrap, this does not
seem to have affected their satisfaction with the wrap.
Implications
Implications for society. The so-called “libertarian paternalistic” method of nudging
could in some situations be seen a more acceptable approach to behavioral change than
coercion, and possibly represent a compromise between those who want to decrease high
environmental exposures and libertarians who advocate free choice (Campbell-Arvai et al.,
2014). Nudging could also be a more effective tool for promoting PEB than traditional tools
like information provision. Information about the environmental impacts of food has been
found ineffective in reducing selection of meat dishes (Campbell-Arvai et al., 2014).
According to theory of cognitive dissonance (Festinger, 1962), such information is likely to
be especially ineffective for customers with strong meat habits and low NR-Perspective, as it
would conflict their existing beliefs and behavior. People with low NR-Perspective agree to
that humans should rule over other species, and they also believe that nothing they do will
change problems on other places on the planet (Nisbet et al., 2009). In light of theory of
cognitive dissonance, individuals with low NR-Perspective are therefore not likely to be
affected by information questioning the ethical aspects of eating animals, or about
environmental problems related to meat production, since it would contradict their way of
understanding the world. Likewise, for high-meat eaters’, meat constitutes an important part
of the daily diet. Negative information about meat will come into conflict with their strong
habits, and would for that reason be downplayed (Festinger, 1962).
Choice architecture could therefore, as the present results suggest, be more efficient in
promoting vegetarian food choice particularly for people with non-vegetarian habits and low
NR worldviews. A nudge could represent a shortcut to sustainable food choice, bypassing
considerations of the reflective System 2 that would be negative towards behavioral change,
and instead making it attractive to System 1. As pointed out by Campbell-Arvai et al. (2014)
nudges could counter individuals’ tendency to choose the option accompanied with short-term
36
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
benefits, such as taste or to satisfy a habit, at the expense of the choices that give more longtime benefits, such as better health or environment. This could be helpful for both individuals
and environment, since the optimistic bias can lead people to underestimate their risk of being
harmed by a health or environmental problem (Pahl, Harris, Todd, & Rutter, 2005; Weinstein,
1987).
As previously mentioned, habits and cognitive dissonance are related to denial
(Stoknes, 2014). Could nudging help break all these three barriers, as proposed by Stoknes
(2014), and represent a way towards more PEB also in the long run? Not only are people with
strong meat consumption habits prone to downplay negative information about meat. Since
people with strong habits are less likely to seek information about alternative courses of
action (Verplanken et al., 1997), high-meat eaters could also be unlikely to seek information
about vegetarian food. But as this study shows, nudges could make high-meat eaters more
willing to test vegetarian food. Whether they will continue to buy vegetarian food in the
future will likely depend on their satisfaction with the food, given that hedonic appreciation
and past experience are important determinants for food choice (Campbell-Arvai et al., 2012;
Köster, 2009). Fujii et al. (2001) found that individuals who were forced to break a habit for a
short period of time were more likely to keep doing the new, non-habitual behavior also in the
future, if their experience of the new behavior was more pleasant than they had expected.
Maybe high-meat eaters can be more positive to eat vegetarian food in the future if they are
nudged to try it and experience it to be more tasteful and fulfilling than anticipated.
Furthermore, if people are nudged to act more environmental-friendly, it can be
plausibly argued that cognitive dissonance and denial concerning climate issues could be
mitigated (Stoknes, 2014). Theory of cognitive dissonance predicts that individuals who are
nudged to eat less meat will experience less dissonance when confronted with information
about environmental problems related to meat consumption, because what they do will now
be better harmonized with the new information. They could therefore become more open to
information saying that a more plant-based diet is more sustainable and healthful, and be
motivated to continue to eat less meat. If their behavior changes, their attitudes could also
change.
Nudging could also be a way out of implicatory denial. Denial of climate issues can
stem from feelings of hopelessness and personal powerlessness (Stoll-Kleemann et al., 2001).
If nudging succeeds in making sustainable options easy to choose, people could become more
aware of possible sustainable courses of actions, and thereby feel more empowered
37
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
confronted with climate issues. This could contribute to reduce powerlessness, climate denial
and reinforce hope.
I therefore suggest that nudging might have the potential to change pro-environmental
attitudes and behavior also in the long run, but this remains a hypothesis that needs to be
further tested. Habits in general, and especially eating habits, are extremely resistant to
permanent change (Gifford, 2011; Klöckner & Verplanken, 2013). The present study only
indicates that nudging might alter choices and break habits in the moment of decision. Little
is known about how nudges affect future behavior, but one study suggests that the effect
decays quickly when the nudge is removed (Allcott & Rogers, 2012). However, if repeated
regularly, participants’ PEB was gradually kept more stable over time, meaning that nudges
could change habits when repeated (Allcott & Rogers, 2012).
Nevertheless, Raihani (2013) underlines the need for choice architects to be aware that
nudges might lead to behavior change that is limited to a specific context, or even produce
negative spillovers, and thereby negate the original positive effect. Especially for decisions
that are performed more than once, such as food selection, there is a risk that habits might go
“back to start” in the future (Raihani, 2013). Hence, even though nudges might help reduce
barriers to PEB, more research is needed to understand how they work over time.
In spite of these uncertainties, the present study has some implications for “food
choice architects”––that is, those who design menus and decide how food is presented. The
explicit menu label “Vegetarian” seems to appeal mostly to customers who already eat little
meat in their daily life. On the other hand, vegetarian dishes that are given more attractive and
descriptive names, and are presented as the default dish, seem to become attractive also for
customers who usually eat much meat and have low internal motivation to select vegetarian
food. A more appealing food name could describe the actual ingredients and reflect the taste
of the dish, and should give pleasant associations (Wansink et al., 2001). At the same time, it
seems important to make it clear that it is a vegetarian dish by stating the ingredients clearly,
such that vegetarian as well as meat-hungry customers can easily orientate.
Presenting vegetarian dishes more saliently and attractively is from a commercial
perspective a low-cost––or even net profitable––intervention. All it requires is making new
posters or menus. In addition, all dishes were equally priced in this experiment. These small
changes will therefore not lead to decreased profit of the café if people select more vegetarian
food and less meat. There is rather a potential of increased profit, as beans are cheaper than
meat. Thus, a small, well-designed non-intrusive nudge towards the vegetarian dishes could
give triple benefits to the company, the customers’ health and the environment.
38
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Last it needs to be stated that in order to bring about the kind of behavioral changes
required to solve climate problems and achieve sustainable food systems, nudging is clearly
not enough. However, such end-user interventions could be part of the systemic solution,
together with information-based, market-based and regulatory instruments (Reisch et al.,
2013).
Implications for research. The present results indicate that future research on
nudging would benefit from changing the question from if nudging is efficient; to for whom it
is efficient. Future research should aim at including individual factors when testing nudges, to
get a better understanding of what nudges are most effective for what groups. Different
individual moderators of nudges could be tested, such as demographic and socioeconomic
variables, as well as personal environmental measures such as attitudes, habits and values.
This study also points to new nudges that could be tested. A focus group could have
been set up to discuss possible names for vegetarian dishes, and participants could have rated
the attractiveness of different names, in order to tailor names that sound appetizing to both
vegetarians and omnivores. One way to ensure that customers understand that a dish is
vegetarian, but at the same time evaluate it as more attractive, could be to let the name reflect
specifically what the dish contains, such as “Mexican Bean Style” or “The Real Mexican
Style, with Beans”. A small green leaf at the side of the vegetarian dishes could be a way to
facilitate the choice for vegetarian customers. Other frames should be tested as well, such as
framing the vegetarian options as the healthiest, most popular or trendy dish.
Vegetarian dishes could be made default choices in a number of ways. An intervention
that could be tested could be that all dishes were vegetarian per default, so customers who
wanted meat needed to ask for this in addition, and maybe pay an extra charge. This would
represent a stronger default than the one tested in this study. In the present study the “Dish of
the day” was equally priced with the other dishes. A reduced price could have made it more
attractive, as it is suggested that nudges may be more potent when implemented along with
more powerful economical motives (Olstad et al., 2014), and research shows that price
reductions on healthy foods can increase both their purchase and consumption (An, 2013).
Future research should also include and test a various range of vegetarian options.
Because choice architecture is a relatively new concept and technique, more research
is needed to be better understand when different techniques are effective. The efficacy of
nudging seems to be complex and context-specific (Olstad et al., 2014). Working out nudges
is a trial-and-error process, and it is unlikely to find nudges that work across all contexts.
Therefore, research should test diverse nudges in various contexts and for different PEB.
39
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Last, little is known about how nudges affect future behavior and attitudes (Raihani,
2013). Using longitudinal designs including self-reporting apps, researchers should further
explore the link between nudging and habits, and explore whether behavior drops back to the
habitual when the nudge is taken away. Can the nudge work in such a way that people open
up their eyes to vegetarian food, and choose it more often also in the future? Possible
spillover effects – both positive and negative – should also be tested. If some nudges can
change habits over time, then how many and how frequent nudges are needed to change
behavior? And more, can nudges through changing habits decrease cognitive dissonance and
denial in the long run?
Ethics of Nudging
The nudges implemented in this study are techniques that impact consumers’ choices
without their awareness. Can this be considered ethical? Nudge tactics have been criticized
for being a subtle form of manipulation, or even coercion, and Goodwin (2012) argues that
nudging in policymaking needs to be altogether rejected. However, it can be argued that there
will always be a social background that affects our decisions, and some paternalism, in the
form of effects on individual decisions, is almost unavoidable (Sunstein, 2014). Further, the
huge marketing industry is already employing a long range of these tactics to increase product
sales. Airports, for instance, are nudging people by having all passengers to walk through
liquor stores on their way to the departure gates. Thaler and Sunstein (2003) state that
”[w]hen paternalism seems absent, it is usually because the starting point appears so natural
and obvious that its preference-shaping effects are invisible to most observers. But those
effects are nonetheless there” (Thaler & Sunstein, 2003, p. 177).
Consider the present study. In the “neutral” control condition, the vegetarian dishes
were named “Vegetarian” and “Vegan”. This cannot really be considered a neutral situation,
as these particular names are likely to form customers’ expectations and desires to select these
products. A menu has to call the dishes something, and if there is a dish of the day, one dish
needs to be selected for this role. And all these features are likely to affect customers’ choices.
So: what criterions should be decisive of which foods we select as defaults, and how we
frame them? Sunstein and Thaler (2003) propose that there exist different criterions for
choosing. Should we do it in such a way that we maximize the profit of the café? Or should
the alternatives be presented completely at random? Or maybe the choice architecture should
make customers choose what they would have chosen if there was no choice architecture? But
if it is true that some choice architecture will always be present, it is impossible to imagine
such a true, unbiased preference. Not only will the present choice architecture shape
40
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
decisions, but also our habits were most likely once formed by some choice situation. A
different alternative will be to design the choice situation in such a way that the customers do
what is in their best interest in the long term (Thaler & Sunstein, 2003). As previously stated,
most of us would benefit from a more plant-based diet, both from a health perspective (Fraser,
1999; Pan et al., 2012; Sinha et al., 2009) and environmental perspective (Hedenus et al.,
2014; Reisch et al., 2013). Thus, making it easier to choose sustainable food should be in
everybody’s interest.
Furthermore, many people might know that their behavior is harming the environment
and want to act in a sustainable way, but habits might hinder them (Verplanken et al., 1998).
In this situation, information or instructions will not help, since people already perform the
behavior despite the knowledge that it is harmful (Bolderdijk, Lehman, & Geller, 2013).
People may also be hindered to act sustainably because of lack of cognitive capacities. Duffy
and Verges (2009) nudged people to recycle, and suggest that many people intend to recycle,
but might fail from doing so because of limited cognitive capacity. Equally, people might
intend to eat more sustainable food, but find their cognitive capacity occupied by hunger, time
pressure or a conversation in the decision moment, so environmental considerations regarding
the food are hindered. They might instead choose a habitual meat dish, even if they in
retrospect would have been satisfied with the vegetarian choice. The right nudge could make
it easier to choose according to environmental intentions.
However, it is essential not to restrict choices, and in general it is important to let the
nudges be transparent, such that they are easy to detect and easy to avoid (Hansen &
Jespersen, 2013; Sunstein, 2014). Manipulating people to eat vegetarian food against their
will would not only be unethical, but could also cause negative spillovers in the future
(Raihani, 2013), and a nudge that appears to be coercion might lead to the opposite behavior
(Bolderdijk et al., 2013). These are examples of what may be termed “bad” choice
architecture.
Strengths and Limitations
The aim of this field experiment was to understand how nudging works in real-life
situations. Because it was conducted in a natural setting, the field validity is thought to be
good. But since it was a field experiment, not all variables could be controlled, and there are
several limitations that warrant consideration. It is possible that the effects found in this study
are specific to the particular sample. The sample consisted of mostly university students, and
many of them were returning customers in the café. Moreover, this study was limited to
41
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
vegetarian food choice, and cannot necessarily be generalized to other forms of environmental
behavior or food choices.
Furthermore, the results of the study, and especially the main effect of the nudges,
could have been clearer with a larger sample. However, this was a field experiment. The
customers did not know that the wrap posters changed from day to day and was part of the
experiment. I was concerned that they might begin to notice if the experiment lasted for much
longer, as some customers came back several times. If rumor spread that there was an
experiment in the café, this could have affected customers’ choices. For that reason, I decided
to end data collection after 10 days. Future research will benefit from a larger sample.
A problem of the data collection was that the participants were given the survey after
selecting wrap. Thus, the experimental variable (nudge condition) was presented and the
outcome (wrap selection) had taken place before the measurement of the moderator variable
(NR). Nudge condition and wrap selection could therefore have affected responses to the NR
measure. Although no such effects were observed statistically, in future studies, NR should
also have been measured first to see if there are any interaction effects. However, it would be
difficult to do it otherwise. The very goal was to see what wrap the participants selected when
they did not know that they participated in an experiment, so their choices would not be
affected by experiment bias.
In spite of Nisbet et al.’s (2009) suggestion that the NR measure will need
modifications to work optimally as a multi-factorial scale, it was used as a multi-factorial
solution in this study, as the factor analysis clearly yielded a three-factor solution, and the
logistic regression with the subscales gave the most interpretable model. Still, the reliability
of NR-Perspective was not satisfactory, meaning that the validity of these results is unclear.
Considering the low reliability of the subscale, the present results support the need of a
revised scale. They also suggest that a multi-factorial scale could be the most promising
solution, as the NR scales seemed to predict different outcomes and therefore be different
constructs.
When participants were asked to indicate the number of their dinners that contain meat
over an average week, the survey did not specify whether “meat” included fish. Participants
could therefore interpret this question differently. Some added “fish included” to their
response, whereas those who asked the researcher about the question, were told that fish was
not included. This number can therefore not be interpreted as a reliable number of meat
consumption in the sample. However, because the answers are thought to be equally biased
across conditions, it was considered acceptable to include the measure.
42
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Conclusion
This experiment tried to nudge more café customers into choosing vegetarian wraps
instead of meat wraps. In order to do so, three different nudge interventions were applied. The
first of these was to rename and frame the wraps called "Vegan" and "Vegetarian" to the more
attractive "Mexican Style". In the second nudge the Vegetarian wrap was made the dish of the
day, or default choice. The third intervention applied both of these nudges by making the
"Mexican style" the dish of the day. Neither of the three nudges helped increase the sale of
vegetarian wraps overall, but interestingly they did seem to have a different effect on different
customers, depending on the customer’s self-reported degree of Nature Relatedness. The
customers who reported to have a less nature-related worldview were unlikely to select the
wrap called "Vegetarian", but more likely to select the renamed version "Mexican Style"
when it was presented as the dish of the day. This way of presenting the vegetarian dish also
made customers who usually ate much meat select it more often. The study thus illustrates
that food choices are affected by both internal factors such as personal nature connectedness
and meat consumption habits, as well as external factors such as the way in which dishes are
presented. The right nudge might be a powerful tool to increase vegetarian food selection
among people who are otherwise unlikely to select vegetarian food. By building on new
knowledge and in combination with focus groups and pilot testing, it is likely that more effective nudges can be designed in future research.
43
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
References
Allcott, H., & Rogers, T. (2012). The short-run and long-run effects of behavioral
interventions: Experimental evidence from energy conservation. Cambridge: Working
paper, NBER
An, R. (2013). Effectiveness of subsidies in promoting healthy food purchases and
consumption: a review of field experiments. Public Health Nutr, 16(7), 1215-1228.
Ariely, D. (2008). Predictably irrational. New York: HarperCollins
Belasco, W. (2008). Food. Oxford: Berg.
Biel, A., Dahlstrand, U., & Grankvist, G. (2005). Habitual and value-guided purchase
behavior. Ambio: a journal of the human environment, 34(4), 360-365. doi:
10.1579/0044-7447-34.4.360
Bolderdijk, J. W., Lehman, P. K., & Geller, E. S. (2013). Encouraging pro-environmental
behaviour with rewards and penalties. In L. Steg, A. E. Van den Berg & J. I. M. De
Groot (Eds.), Environmental psychology: an introduction (pp. 233 - 242). Chichester:
BPS Blackwell.
Campbell-Arvai, V., Arvai, J., & Kalof, L. (2014). Motivating sustainable food choices: The
role of nudges, value orientation, and information provision. Environment and
Behavior, 46 (4),453-475. doi: 10.1177/0013916512469099
Carlsson-Kanyama, A., & González, A. D. (2009). Potential contributions of food
consumption patterns to climate change. The American journal of clinical nutrition,
89(5), 1704S-1709S.
Chawla, L. (1999). Life paths into effective environmental action. The Journal of
Environmental Education, 31(1), 15-26. doi: 10.1080/ 00958969909598628
Cohen, D., & Farley, T. A. (2008). Eating as an automatic behavior. Preventing chronic
disease, 5(1).
Cohen, S. (2001). States of denial: knowing about atrocities and suffering. Cambridge: Polity.
Crompton, T., & Kasser, T. (2009). Meeting environmental challenges: The role of human
identity: WWF-UK Godalming, UK.
Dahlstrand, U., & Biel, A. (1997). Pro-‐environmental habits: Propensity levels in behavioral
change. Journal of Applied Social Psychology, 27(7), 588-601. doi: 10.1111/j.15591816.1997.tb00650.x
De Groot, J. I. M., & Steg, L. (2008). Value orientations to explain beliefs related to
environmental significant behavior. How to measure egoistic, altruistic, and
biospheric value orientations. Environment and Behavior, 40(3), 330-354.
Dong, Y., & Peng, C.-Y. J. (2013). Principled missing data methods for researchers.
SpringerPlus, 2(1), 1-17. doi: 10.1186/2193-1801-2-222
Duffy, S., & Verges, M. (2009). It matters a hole lot. Perceptual affordances of waste
containers influence recycling compliance. Environment and Behavior, 41(5), 741749. doi: 10.1177/0013916508323737
Evans, J. S. B. (2003). In two minds: dual-process accounts of reasoning. Trends in cognitive
sciences, 7(10), 454-459.
Festinger, L. (1962). A theory of cognitive dissonance (2nd ed.). Stanford: Stanford university
press.
Field, A. (2009). Discovering statistics using SPSS: (and sex and drugs and rock 'n' roll). Los
Angeles: SAGE.
Fraser, G. E. (1999). Associations between diet and cancer, ischemic heart disease, and allcause mortality in non-hispanic white california seventh-day adventists. The American
journal of clinical nutrition, 70(3), 532-538.
Fujii, S., Gärling, T., & Kitamura, R. (2001). Changes in drivers’ perceptions and use of
public transport during a freeway closure: Effects of temporary structural change on
44
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
cooperation in a real-life social ilemma. Environment and Behavior, 33(6), 796-808.
doi: 10.1177/00139160121973241
Gifford, R. (2011). The dragons of inaction: Psychological barriers that limit climate change
mitigation and adaptation. American psychologist, 66(4), 290-302. doi:
10.1037/a0023566
Gifford, R., Scannell, L., Kormos, C., Smolova, L., Biel, A., Boncu, S., . . . Hine, D. (2009).
Temporal pessimism and spatial optimism in environmental assessments: An 18nation study. Journal of environmental psychology, 29(1), 1-12. doi:
10.1016/j.jenvp.2008.06.001
Gomes, M. E., Roszak, T., & Kanner, A. D. (1995). Ecopsychology: restoring the earth,
healing the mind. San Francisco: Sierra Club Books.
Goodland, R., & Anhang, J. (2009). Livestock and climate change: what if the key actors in
climate change are... cows, pigs, and chickens?. Worldwatch Institute. Retrieved from
www.worldwatch.org/files/pdf/Livestock%20and%20Climate%20Change.pdf
Goodwin, T. (2012). Why we should reject ‘nudge’. Politics, 32(2), 85-92. doi:
10.1111/j.1467-9256.2012.01430.x
Guagnano, G. A., Stern, P. C., & Dietz, T. (1995). Influences on attitude-behavior
relationships a natural experiment with curbside recycling. Environment and
Behavior, 27(5), 699-718.
Hall, C. S. (1966). Freuds psykologi. En grundbok. København: Hans Reitzels Forlag.
Hansen, P. G., & Jespersen, A. M. (2013). Nudge and the manipulation of choice. Eur. J. Risk
Reg, 1, 3-28.
Hastie, R., & Dawes, R. M. (2010). Rational choice in an uncertain world: The psychology of
judgment and decision making. Thousand Oaks: Sage.
Hedenus, F., Wirsenius, S., & Johansson, D. J. A. (2014). The importance of reduced meat
and dairy consumption for meeting stringent climate change targets. Climatic Change,
1-13. doi: 10.1007/s10584-014-1104-5
Helsedirektoratet (2012). Utviklingen i norsk kosthold (IS: 2036). Oslo: Helsedirektoratet.
Hill, J. O. (2009). Can a small-changes approach help address the obesity epidemic? A report
of the Joint Task Force of the American Society for Nutrition, Institute of Food
Technologists, and International Food Information Council. The American journal of
clinical nutrition, 89(2), 477-484.
Hille, J. (2013). Livsløpsgrillen. Arbeidsnotat 02/2013. Framtiden i våre hender.
Nymoen, L. L., & Hille, J. (2012). Klimavennlig mat i sykehjem. Bioforsk rapport, 7(1). Ås:
Bioforsk Økologisk.
Johnson, E. J., & Goldstein, D. (2003). Do defaults save lives?. Science. 302 (5649), 13381339. doi: 10.1126/science.1091721
Kahneman, D. (2003). Maps of bounded rationality: Psychology for behavioral
economics. The American economist review, 93(5), 1449-1475.
Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1991). Anomalies: The endowment effect,
loss aversion, and status quo bias. The journal of economic perspectives, 5(1), 193206.
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk.
Econometrica: Journal of the Econometric Society, 263-291.
Kallbekken, S., & Sælen, H. (2013). ‘Nudging’hotel guests to reduce food waste as a win-win
environmental measure. Economics Letters, 119(3), 325-327.
Kallbekken, S., Sælen, H., & Hermansen, E A. (2013). Bridging the Energy Efficiency Gap:
A Field Experiment on Lifetime Energy Costs and Household Appliances. Journal of
Consumer Policy, 36(1), 1-16.
45
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Klöckner, C. A., & Verplanken, B. (2013). Yesterday's habits preventing change for
tomorrow? In L. Steg, A. E. Van den Berg & J. I. M. De Groot (Eds.), Environmental
psychology: An introduction (pp. 197-209). Chichester: BPS Blackwell.
Knoll, M. A. Z. (2010). The role of behavioral economics and behavioral decision making in
Americans' retirement savings decisions. Social Security Bulletin, 70(4), 1-23.
Kollmuss, A., & Agyeman, J. (2002). Mind the gap: why do people act environmentally and
what are the barriers to pro-environmental behavior? Environmental education
research, 8(3), 239-260.
Kubberød, E., Ueland, Ø., Tronstad, Å., & Risvik, E. (2002). Attitudes towards meat and
meat-eating among adolescents in Norway: a qualitative study. Appetite, 38(1), 53-62.
Köster, E. P. (2009). Diversity in the determinants of food choice: A psychological
perspective. Food Quality and Preference, 20(2), 70-82.
Mayer, F. S., & Frantz, C. M. (2004). The connectedness to nature scale: A measure of
individuals’ feeling in community with nature. Journal of environmental psychology,
24(4), 503-515.
McKenzie, C. R. M., Liersch, M. J., & Finkelstein, S. R. (2006). Recommendations implicit
in policy defaults. Psychological science, 17(5), 414-420.
McNeil, B. J., Pauker, S. G., Sox Jr, H. C., & Tversky, A. (1982). On the elicitation of
preferences for alternative therapies. The New England journal of medicine, 306(21),
1259-1262.
Milkman, K. L., Rogers, T., & Bazerman, M. H. (2008). Harnessing our inner angels and
demons: What we have learned about want/should conflicts and how that knowledge
can help us reduce short-sighted decision making. Perspectives on Psychological
Science, 3(4), 324-338.
Nisbet, E. K., Zelenski, J. M., & Murphy, S. A. (2009). The Nature Relatedness scale linking
individuals' connection with nature to environmental concern and behavior.
Environment and Behavior, 41(5), 715-740. doi: 10.1177/0013916508318748
Norgaard, K. M. (2011). Living in denial: climate change, emotions and everyday life.
Cambridge: MIT press.
nudge. (n.d.). In Merriam-Webster.com. Retrieved from http://www.merriamwebster.com/dictionary/nudge.
Nymoen, L. L., & Hille, J. (2012). Klimavennlig mat i sykehjem. Bioforsk rapport, 7(1). Ås:
Bioforsk Økologisk.
Næss, A. (1973). The shallow and the deep, long-‐range ecology movement. A summary.
Inquiry, 16(1-4), 95-100.
Olstad, D. L., Goonewardene, L. A., McCargar, L. J., & Raine, K. D. (2014). Choosing
healthier foods in recreational sports settings: a mixed methods investigation of the
impact of nudging and an economic incentive. International Journal of Behavioral
Nutrition and Physical Activity, 11(1), 6.
Pahl, S., Harris, P. R., Todd, H. A., & Rutter, D. R. (2005). Comparative optimism for
environmental risks. Journal of Environmental Psychology, 25(1), 1-11.
Pallant, J. (2010). SPSS survival manual: A step by step guide to data analysis using SPSS.
Crows Nest: Allen & Unwin.
Pan, A., Sun, Q., Bernstein, A. M., Schulze, M. B., Manson, J. E., Stampfer, M. J., . . . Hu, F.
B. (2012). Red meat consumption and mortality. Results from 2 prospective cohort
studies. Archives of internal medicine, 172(7), 555-563. doi:
10.1001/archinternmed.2011.2287
Pichert, D., & Katsikopoulos, K. V. (2008). Green defaults: Information presentation and proenvironmental behaviour. Journal of environmental psychology, 28(1), 63-73.
46
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Povey, R., Wellens, B., & Conner, M. (2001). Attitudes towards following meat, vegetarian
and vegan diets: an examination of the role of ambivalence. Appetite, 37(1), 15-26.
doi: 10.1006/appe.2001.0406
Raihani, N. J. (2013). Nudge politics: efficacy and ethics. Frontiers in psychology, 4.
Reisch, L., Eberle, U., & Lorek, S. (2013). Sustainable food consumption: an overview of
contemporary issues and policies. Sustainability: Science, Practice, & Policy, 9(2), 725.
Schultz, P. W. (2002a). Empathizing with nature: The effects of perspective taking on
concern for environmental issues. Journal of social issues, 56(3), 391-406.
Schultz, P. W. (2002b). Knowledge, information, and household recycling: Examining the
knowledge-deficit model of behavior change. New tools for environmental protection:
Education, information, and voluntary measures, 67-82.
Simon, H. A. (1955). A behavioral model of rational choice. The quarterly journal of
economics, 69(1), 99-118.
Sinha, R., Cross, A. J., Graubard, B. I., Leitzmann, M. F., & Schatzkin, A. (2009). Meat
intake and mortality: A prospective study of over half a million people. Archives of
internal medicine, 169(6), 562-571. doi: 10.1001/archinternmed.2009.6.
Skov, L. R., Schmidt, K., Hansen, P. G., Skov, K. L., E., Mikkelsen B., & A., Pérez-Cueto F.
J. (2013). The smaller the piece the healthier consumption - a choice architectural
experiment in behavioural nutrition. Paper presented at the The 20th International
Congress of Nutrition, Grenada, Spain. Retrieved from:
www.inudgeyou.com/inudgeyou-does-health-nudge-experiment-on-buffetarrangement/
Steg, L., & Vlek, C. (2009). Encouraging pro-environmental behaviour: An integrative review
and research agenda. Journal of environmental psychology, 29(3), 309-317.
Steinfeld, H., Gerber, P., Wassenaar, T. D., Castel, V., & de Haan, C. (2006). Livestock’s
long shadow: environmental issues and options. Rome: United Nations Food and
Agriculture Organization.
Stern, N. H., Britain, G., & Treasury, H. M. (2006). Stern review: The economics of climate
change. London: HM Treas./Cabinet Off.
Stern, P. C. (2000). New environmental theories: toward a coherent theory of environmentally
significant behavior. Journal of social issues, 56(3), 407-424.
Stern, P. C., Dietz, T., Abel, T., Guagnano, G. A., & Kalof, L. (1999). A value-belief-norm
theory of support for social movements: The case of environmentalism. Human
ecology review, 6(2), 81-98.
Stoknes, P. E. (2014). Rethinking climate communications and the psychological climate
paradox. Energy Research & Social Science. doi: 10.1016/j.erss.2014.03.007
Stoll-Kleemann, S., O’Riordan, T., & Jaeger, C. C. (2001). The psychology of denial
concerning climate mitigation measures: evidence from Swiss focus groups. Global
environmental change, 11(2), 107-117.
Sunstein, C. R. (2013). Behavioral economics, consumption, and environmental protection. In
L. Reisch & T. J. (Eds.), Forthcoming in Handbook on Research in Sustainable
Consumption.
Sunstein, C. R. (2014). Why nudge?: The politics of libertarian paternalism: Yale University
Press.
Sunstein, C. R., & Reisch, L. (2013). Automatically green: Behavioral economics and
environmental protection. Available at SSRN 2245657.
Sunstein, C. R., & Thaler, R. H. (2003). Libertarian paternalism is not an oxymoron. The
University of Chicago Law Review, 1159-1202.
47
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Tam, K.-P. (2013). Concepts and measures related to connection to nature: Similarities and
differences. Journal of environmental psychology.
Thaler, R. H. (1990). Anomalies: Saving, fungibility, and mental accounts. The journal of
economic perspectives, 193-205.
Thaler, R. H., & Sunstein, C. R. (2003). Libertarian paternalism. The American economic
review, 93(2), 175-179. doi: 10.2307/3132220
Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving decisions about health, wealth,
and happiness: Yale University Press.
Thaler, R. H., Sunstein, C. R., & Balz, J. P. (2010). Choice Architecture.
Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases.
Science, 185(4157), 1124-1131.
Tversky, A., & Kahneman, D. (1981). The framing of decisions and the psychology of choice.
Science, 211(4481), 453-458.
Tversky, A., & Kahneman, D. (1986). Rational choice and the framing of decisions. The
Journal of Business, 59(4), 251-278. doi: 10.2307/2352759
Utenriksdepartementet. (2011). Klima og bærekraftig utvikling. Oslo: Retrieved from
http://www.regjeringen.no/nb/dep/ud/dok/regpubl/prop/2011-2012/prop-1-s20112012/8/3.html?id=657710.
Vermeulen, S. J., Campbell, B. M., & Ingram, J. S. I. (2012). Climate change and food
systems. Annual Review of Environment and Resources, 37(1), 195.
Verplanken, B., & Aarts, H. (1999). Habit, attitude, and planned behaviour: Is habit an empty
construct or an interesting case of goal-directed automaticity? European Review of
Social Psychology, 10(1), 101-134. doi: 10.1080/14792779943000035
Verplanken, B., Aarts, H., & Van Knippenberg, A. (1997). Habit, information acquisition,
and the process of making travel mode choices. European Journal of Social
Psychology, 27(5), 539-560. doi: 10.1002/(SICI)1099-0992
Verplanken, B., Aarts, H., van Knippenberg, A., & Moonen, A. (1998). Habit versus planned
behaviour: A field experiment. British Journal of Social Psychology, 37(1), 111-128.
doi: 10.1111/j.2044-8309.1998.tb01160.x
Wansink, B. (2013). Convenient, attractive, and normative: The CAN approach to making
children slim by design. Childhood Obesity, 9(4), 277-278. doi:
10.1089/chi.2013.9405
Wansink, B., & Cheney, M. M. (2005). Super bowls: serving bowl size and food
consumption. Jama, 293(14), 1723-1728. doi: 10.1038/ijo.2013.169
Wansink, B., Just, D. R., & Payne, C. R. (2009). Mindless eating and healthy heuristics for
the irrational. The American economic review, 165-169. doi: 10.1257/aer.99.2.165
Wansink, B., Just, D. R., Payne, Collin R., & Klinger, M. Z. (2012). Attractive names sustain
increased vegetable intake in schools. Preventive medicine, 55(4), 330-332. doi:
10.1016/j.ypmed.2012.07.012
Wansink, B., & Kim, J. (2005). Bad popcorn in big buckets: portion size can influence intake
as much as taste. Journal of nutrition education and behavior, 37(5), 242-245. doi:
10.1016/S1499-4046(06)60278-9
Wansink, B., Painter, J. E., & Van Ittersum, K. (2001). Descriptive menu labels' effect on
sales. The Cornell Hotel and Restaurant Administration Quarterly, 42(6), 68-72. doi:
10.1016/S0010-8804(01)81011-9
Wansink, B., & Sobal, J. (2007). Mindless eating: The 200 daily food decisions we overlook.
Environment and Behavior, 39(1), 106-123. doi: 10.1177/ 0013916506295573
Wansink, B., Van Ittersum, K., & Painter, J. E. (2005). How descriptive food names bias
sensory perceptions in restaurants. Food Quality and Preference, 16(5), 393-400. doi:
10.1016/j.foodqual.2004.06.005
48
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Weinstein, N. D. (1980). Unrealistic optimism about future life events. Journal of personality
and social psychology, 39(5), 806.
Weinstein, N. D. (1982). Unrealistic optimism about susceptibility to health problems.
Journal of behavioral medicine, 5(4), 441-460.
Weinstein, N. D. (1987). Unrealistic optimism about susceptibility to health problems:
Conclusions from a community-wide sample. Journal of behavioral medicine, 10(5),
481-500.
Zelenski, J. M., & Nisbet, E. K. (2014). Happiness and feeling connected. The distinct role of
Nature Relatedness. Environment and Behavior, 46(1), 3-23. doi:
10.1177/0013916512451901
49
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Appendix A
Wrap Presentation in the Field Experiment
Figure A.1. The Wrap Presentation in the DEFAULT+FRAME Condition. Blackboard 1 to the right and
Blackboard 2 to the upper left. Poster on the counter.
50
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Appendix B
Survey
SPØRRESKJEMA OM MATVANER
Dette spørreskjemaet har som formål å undersøke matvaner og folks holdninger til
samfunnsmessige spørsmål. Det vil ta ca. 10-15 minutter å fylle ut. Alle som svarer, er med i
trekningen av et gavekort på 1000 kroner.
Deltakelse er frivillig, og du kan trekke deg når som helst uten å oppgi noen forklaring. Alle
dine svar vil være fullstendig anonyme. Opplysninger om deg vil holdes konfidensielle, og
ingen enkeltpersoner vil kunne gjenkjennes i den ferdige oppgaven.
Noen spørsmål kan være ganske like. Det er likevel viktig at du svarer på hvert enkelt av dem.
Forsøk å svare så spontant som mulig. Unngå å tenke lenge gjennom hvert spørsmål, men
svar det som er din umiddelbare følelse. Og husk: Vi er interessert i dine synspunkter. Det
finnes ingen riktige eller gale svar. Svar derfor slik som du virkelig føler, ikke tenk på hva du
tror ”folk flest” føler. Svar bare så ærlig og oppriktig som mulig på alle spørsmålene. Les nøye
gjennom påstandene og teksten før spørsmålene, så du er sikker på at du har forstått hva det
spørres etter.
Dersom du har spørsmål angående studien, eller ønsker å vite de endelige resultatene, er du
velkommen til å kontakte meg på e-post: sigrimho@student.sv.uio.no.
Med vennlig hilsen
Sigrid Møyner Hohle, masterstudent ved Universitetet i Oslo
51
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
I løpet av en gjennomsnittlig uke, hvor mange av middagene du spiser inneholder kjøtt?
For eksempel: Hvis du vanligvis spiser kjøtt til middag hver dag, skriv 7. Hvis du aldri spiser
kjøtt til middag, skriv 0.
□
Antall middager
Angi i hvilken grad du er enig i de neste påstandene, ved å sette ett kryss ved siden av
hver påstand.
Sterkt
uenig
1. Styresmaktene bør kontrollere i hvilket
tempo råvarer brukes, for å sikre at de
varer så lenge som mulig.
2. Jeg er i mot styresmakter som prøver å
få råvarer til å vare lenger ved å
kontrollere og regulere hvordan de
brukes.
3. En av de viktigste grunnene til å holde
innsjøer og elver rene, er at mennesker
da har et sted å kjøre båt, bade, padle,
dykke etc.
4. Vi må holde elver og innsjøer rene for å
beskytte miljøet, IKKE for å sikre
steder hvor mennesker kan kjøre båt,
bade, padle, dykke etc.
5. Jeg er IKKE en person som anstrenger
meg for å spare naturressurser.
6. Når det er mulig, prøver jeg å spare
naturressurser.
7. Det er viktigere å bevare
arbeidsplasser enn å beskytte miljøet.
8. Å beskytte miljøet er viktigere enn å
sikre fortsatt økonomisk vekst.
Litt
uenig
□
2
□
2
□
2
□
2
1
1
1
1
□
□
□
□
Verken
enig
eller
uenig
□
3
□
3
□
3
□
3
□
□
□
□
Litt
enig
□
4
□
4
□
4
□
4
□
□
□
□
Sterkt
enig
□
5
□
5
□
5
□
5
□
□
□
□
□
□
□
□
□
□
□
□
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
52
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
De neste påstandene handler om dine matvaner. Angi i hvilken grad du er enig, ved å
sette et kryss ved siden av hver påstand.
Sterkt
uenig
1. Jeg har tenkt å spise mye vegetarisk mat i
framtiden.
2. I framtiden tror jeg de fleste middagene jeg
spiser, vil inneholde kjøtt.
3. Jeg vil prøve å lage mange vegetarmåltider i
tiden framover.
4. I tiden framover tror jeg at jeg kommer til å spise
mye kjøtt.
□
□
□
□
Litt
uenig
□
□
□
□
Verken
enig
eller
uenig
□
□
□
□
Litt
enig
Sterk
t enig
□
□
□
□
□
□
□
□
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
I løpet av den neste uken, hvor mange av middagene du spiser tror du vil inneholde
kjøtt?
For eksempel: Hvis du tror du kommer til å spise kjøtt til middag hver dag, skriv 7. Hvis du
tror du ikke kommer til å spise kjøtt til middag noen dag, skriv 0.
□
53
Antall middager
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Det fins mange arter i verden, blant annet mennesker og andre pattedyr. Hvor enig er du i
følgende påstander om arter generelt? Angi i hvilken grad du er enig, ved å sette ett
kryss ved siden av hver påstand.
Husk at mennesket også er en art.
Sterkt
uenig
1. Visse arter er rett og slett underlegne andre.
2. For å oppnå det man vil, er det noen ganger
nødvendig å bruke makt mot andre arter.
3. Det er greit at noen arter er bedre stilt i livet
enn andre.
4. Ingen arter burde få dominere på jorda.
5. Alle arter bør ha like muligheter i livet.
6. Det er nok bra at visse arter er overlegne og
andre underlegne.
7. Det hadde vært bra om arter var likeverdige.
8. Noen ganger må andre arter holdes på
plassen sin.
9. Vi bør gjøre det vi kan for å skape like vilkår
for ulike arter.
10. Vi ville hatt færre problemer dersom vi
behandlet alle arter mer likt.
□
□
□
□
□
□
□
□
□
□
Litt
uenig
□
□
□
□
□
□
□
□
□
□
Verken
enig
eller
uenig
□
□
□
□
□
□
□
□
□
□
Litt
enig
□
□
□
□
□
□
□
□
□
□
Sterkt
enig
□
□
□
□
□
□
□
□
□
□
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
54
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Det fins mange typer grupper i verden: menn og kvinner, etniske og religiøse grupper,
nasjonaliteter og politiske grupper. Hvor enig er du i følgende påstander om grupper
generelt? Angi i hvilken grad du er enig, ved å sette ett kryss ved siden av hver påstand.
Sterkt
uenig
1. Visse menneskegrupper er rett og slett
underlegne andre.
2. For å oppnå det man vil, er det noen
ganger nødvendig å bruke makt mot
andre grupper.
3. Det er greit at noen grupper er bedre
stilt i livet enn andre.
4. Ingen grupper burde få dominere i
samfunnet.
5. Alle grupper bør gis like muligheter i
livet.
6. Det er nok bra at visse grupper er
overlegne og andre underlegne.
7. Det hadde vært bra om grupper var
likeverdige.
8. Noen ganger må andre grupper holdes
på plassen sin.
9. Vi bør gjøre det vi kan for å skape like
vilkår for forskjellige grupper.
10. Vi ville hatt færre problemer dersom
vi behandlet alle mennesker mer likt.
11. Alle grupper bør stå sammen for det
felles beste.
12. Alle typer grupper bør stå sammen
samlet.
13. Sammen bør alle grupper være som én
stor familie.
14. Forholdet mellom forskjellige grupper
bør være én for alle og alle for én.
55
□
□
Litt
uenig
□
□
Verken
enig
eller
uenig
□
□
Litt
enig
□
□
Sterkt
enig
□
□
1
2
3
4
5
1
2
3
4
5
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Angi i hvilken grad du er enig i de neste påstandene, ved å sette ett kryss ved siden av
hver påstand.
Sterkt
uenig
1. Det er meningen at menneskene skal
herske over resten av naturen.
2. Menneskene er blitt skapt eller har
utviklet seg for å dominere over resten
av naturen.
3. Planter og dyr har like mye rett til å
eksistere som mennesker.
4. Planter og dyr eksisterer først og fremst
for å bli brukt av mennesker.
5. Mennesker er en del av økosystemet i
like stor grad som andre dyr.
6. Mennesker er ikke noe viktigere i
naturen enn andre levende vesener.
7. Naturen eksisterer først og fremst for at
mennesker skal kunne bruke den.
8. Naturen i alle sine former bør
kontrolleres av mennesker.
9. Jeg tror IKKE at mennesker ble skapt
eller utviklet seg for å dominere over
resten av naturen.
10. Mennesker er ikke viktigere enn andre
arter.
□
□
Litt
uenig
□
□
Verken
enig
eller
uenig
□
□
Litt
enig
□
□
Sterkt
enig
□
□
1
2
3
4
5
1
2
3
4
5
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
□
1
□
2
□
3
□
4
□
5
Angi i hvilken grad du er enig i de neste påstandene, ved å sette ett kryss ved siden av
hver påstand.
Sterkt
uenig
1. Jeg liker å være utendørs, selv i dårlig
vær.
2. Noen arter er det rett og slett
meningen at skal dø ut eller bli
utryddet.
3. Vi mennesker har rett til å bruke
naturressurser som vi vil.
□
□
Litt
uenig
Verken
enig
eller
uenig
□ □
□ □
Litt
enig
□
□
Sterkt
enig
□
□
1
2
3
4
5
1
2
3
4
5
□
1
□ □
2
3
□
4
□
5
56
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Sterkt
uenig
4. Mitt drømmeferiested er et øde sted i
villmarka.
5. Jeg tenker alltid på hvordan
handlingene mine påvirker miljøet.
6. Jeg liker å grave i jorda og få jord på
hendene.
7. Min tilknytning til naturen og miljøet
er en del av min åndelighet.
8. Jeg er veldig bevisst på
miljøproblemer.
9. Jeg legger merke til dyre- og planteliv
uansett hvor jeg er.
□
□
Litt
uenig
Verken
enig
eller
uenig
□ □
□ □
Litt
enig
□
□
Sterkt
enig
□
□
1
2
3
4
5
1
2
3
4
5
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
□
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
14. Mine følelser for naturen påvirker ikke
1
hvordan jeg lever livet mitt.
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
1
2
3
4
5
10. Jeg drar ikke ofte ut i naturen.
11. Ikke noe jeg gjør, kan forandre
problemer andre steder på kloden.
12. Jeg er ikke atskilt fra naturen, men en
del av naturen.
13. Tanken på å være dypt inne i skogen,
unna sivilisasjonen, er skremmende.
15. Dyr, fugler og planter burde ha færre
rettigheter enn mennesker.
16. Jeg legger merke til naturen rundt meg
til og med når jeg er midt i byen.
17. Mitt forhold til naturen er en viktig del
av hvem jeg er.
18. Det er unødvendig å verne natur og
miljø, for naturen klarer å komme seg
etter en hvilken som helst menneskelig
påvirkning.
□
□
□ □
□ □
□
□
□
□
19. Jeg tenker mye på dyrs lidelser.
1
2
3
4
5
20. Jeg føler meg veldig forbundet med
jorda og med alt levende.
1
2
3
4
5
Kjønn:
□ Kvinne □ Mann
57
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Alder:
□
år
Velg kategorien som passer best for dine matvaner. Kryss av for kun én kategori.
1.
2.
3.
4.
□ Jeg spiser både kjøtt og fisk.
□ Jeg er vegetarianer (unngår å spise kjøtt og fisk).
□ Jeg er veganer (unngår å spise alle produkter fra dyr).
□ Jeg er pescetarianer (unngår å spise kjøtt, men spiser fisk og sjømat).
Har du svart på dette spørreskjemaet før?
□ Ja
□ Nei
Kjøpte du wrap på Break i dag?
□ Ja
□ Nei
Hvis du svarte nei på spørsmålet over, kan du se bort fra de neste spørsmålene.
Hvis du svarte ja, vennligst svar på de neste spørsmålene.
Hvilken wrap spiste du på Break i dag?
□ Taco m/kjøtt (kald)
2. □ BLT (kald)
3. □ Kylling pesto (kald)
4. □ Kylling curry (kald)
5. □ Vegetar/mexican style (kald)
1.
□ Taco m/kjøtt (varm)
7. □ Svinekjøtt (varm)
8. □ Kylling (varm)
9. □ Vegan/mexican style (varm)
10. □ Tunfisk
6.
58
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Hvor ofte spiser du wrap på Break?
1.
2.
3.
4.
5.
□ Aldri (dette er første gang jeg spiser wrap her)
□ Sjelden (1 gang i måneden eller sjeldnere)
□ Av og til (2-3 ganger i måneden)
□ Ofte (1-4 ganger i uka)
□ Alltid (5 ganger i uka)
Hvis du har spist wrap på Break før: Hvor ofte pleier du å spise den typen du spiste i
dag?
1.
2.
3.
4.
5.
□ Aldri (det er første gang jeg spiser denne typen)
□ Sjelden
□ Noen ganger
□ Ofte
□ Alltid (jeg spiser denne typen hver gang jeg er her)
Hvor godt likte du wrapen du spiste i dag?
1.
2.
3.
4.
5.
□ Svært godt
□ Godt
□ Helt OK
□ Dårlig
□ Svært dårlig
Dersom du har kommentarer eller presiseringer til noen av spørsmålene, skriv dem gjerne på
baksiden av dette arket.
59
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Appendix C
Factor Analysis of the Nature Relatedness Scale
All wrap buyers were included in the factor analysis (N = 344). The scale items were
normally distributed. Bartlett’s test of sphericity was significant: χ2 (190) = 1876.00, p = .000
and the KMO value was satisfactory (KMO = .86). However, the correlation matrices
displayed several correlations below .3. This was especially the case for the items within the
subscale Perspective.
An initial exploratory maximum likelihood factor analysis supported a fivefactor solution
(eigenvalues 5.45, 1.91, 1.38, 1.21, and 1.08), accounting for 55.11% of total variance. This
solution gave no interpretable result. Furthermore, Nisbet et al. (2009) suggest three factors in
the original model, and Catell’s scree plot (see Figure C.1) method suggested a three-factor
model. Thus, to test the assumed underlying factors NR-Self, NR-Perspective, and NRExperience, a confirmatory maximum likelihood factor analysis was conducted, requesting
three factors. An oblique rotation (direct oblimin) was chosen because the factors were
expected to correlate. The three factors (eigenvalues 5.45, 1.91, and 1.38), accounted for
43.69% of the total variance.
An interpretable three-factor structure was obtained (see Table C.1), where a majority of
the items were without cross loadings and loaded on one of the three factors. The correlation
matrix shows that the items loading on the first factor correspond to the NR-Self concept,
representing how people identify with nature. The items loading to the second factor represent
people’s familiarity with and desire to spend time in nature, the NR-Experience, whereas the
items assessing the external, nature-related worldview, the NR-Perspective items, loaded in
the third factor.
Two items did not fit the structure proposed by Nisbet et al. (2009). Item NR_Self_4 (“I
am not separate from nature, but a part of nature”) and NR_Self_6 (“Even in the middle of the
city, I notice nature around me”) did not have strong loadings to any factors, but loaded the
most strongly to the Experience factor. Item NR_Self_8 (“I think a lot about the suffering of
animals”) had weak loadings, and loaded more strongly to the Perspective scale. Cronbach’s
alpha analysis was performed to explore the reliability of the scale with or without these
items. Cronbach’s alpha of the subscale NR_Self was increased without the items NR_Self_4
and NR_Self_8, but would decrease without NR_Self_6. Taking into consideration the factor
analysis and the Cronbach’s alpha analysis, items NR_Self_4 and NR_Self_8 were excluded,
but NR_Self_6 was retained. Though some items in NR_Perspective had weak loadings, they
60
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
all loaded most strongly to the third factor, and the Cronbach’s alpha of the subscale would
decrease if any of the items were removed. Thus, these items were all retained.
The structure proposed by Nisbet et al. (2009) was therefore supported, with two
modifications. 18 items were computed into three sums and one total sum, based on the scales
developed by Nisbet et al. (2009). The reliability estimated by Cronbach’s alpha was
satisfactory for the full scale (α = .837), NR-Experience (α = .780) and NR-Self (α = .808),
but not satisfactory for NR-Perspective (α = .477). Nisbet et al. (2009) likewise found the NRPerspective factor to have the lowest reliability in the scale (α = .66). Even though NR is a
multidimensional concept, Nisbet et al. (2009) considered a one-factor structure more
promising, because several items are loading on multiple factors, and the subscales are
strongly inter-correlated. The authors suggest that the items require changes to achieve an
optimal multi-factorial solution. Nevertheless, because there seemed to be a clear threefactorial structure, I chose to proceed with both the three subscales and a total scale, keeping
in mind that the subscale NR-Perspective had poor internal reliability.
Figure C.1. Scree Plot of the 20 Nature Relatedness Items.
61
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Table C.1
Factor Loadings for Confirmatory Factor Analysis with Oblique
Rotation of Nature Relatedness Scale.
Factor 1
Factor 2
Factor 3
(Self)
(Experience)
(Perspective)
E1
.70
-.03
-.18
E2
.67
-.04
.09
E3
.62
.08
-.00
E4
.35
.29
.09
E5
.52
.10
.02
E6
.60
-.07
.06
S1
-.09
.70
.04
S2
.30
.48
-.08
S3
-.07
.67
.10
S4
.19
.17
.12
S5
.32
.46
.03
S6
.27
.24
.02
S7
.45
.49
-.04
S8
.01
.28
.36
S9
.29
.52
.06
P1
.06
-.11
.41
P2
-.03
.10
.46
P3
.17
-.04
.20
P4
-.13
.07
.62
P5
-.02
.13
.18
Items
NR-Experience
NR-Self
NR-Perspective
Note. Factor loadings > .32 are in boldface.
62
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Appendix D
Descriptive Statistics: Participants’ Café and Wrap Habits
Customers’ frequency of visiting the café and of purchasing the same wrap as they bought
on the experiment day can be found in Figure D.1 and D.2. A majority of the participants
(85%) had been to the café at least once before, and almost half (44%) reported to visit the
café at least 2-3 times à month (“Sometimes” in the questionnaire). 14% of those who
answered reported to buy the same wrap as they bought today, every time, and more than half
(56%) reported to buy the same wrap sometimes, often or every time they visit the café. For 3
Percentage of customers
out of 10, it was the first time they bought the wrap they had chosen today.
45
40
35
30
25
20
15
10
5
0
40.9 %
31.5 %
15.0 %
12.1 %
0.6 %
Never
Rarely
Sometimes
Often
Every time
Frequency of visiting the café
Figure D.1. Frequency of Visiting the Café. The diagram shows how
often the customers visit the café, in percent. N = 340.
Percentage of customers
35
31.6 %
30
25
20.5 %
21.4 %
20
13.9 %
12.7 %
15
10
5
0
Never
Rarely
Sometimes
Often
Every time
Frequency of selecting today's wrap
Figure D.2. Frequency of Purchasing this Wrap. The figure shows
how often the customers choose the same wrap as today. N = 332.
63
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Appendix E
Logistic Regression of Default Frame versus Control with full NR Scale
Binary logistic regression with DEFAULT+FRAME condition versus CONTROL was run
with total Nature Relatedness score as predictor (see table E.1). The full model was
significant, χ2 (4, n = 159) = 55.47, p < .001. However, the model containing the three NR
subscales was reported, for several reasons. First of all, the factor analysis yielded a
multidimensional structure of NR. Secondly, the logistic regression model with the subscales
revealed that one subscale explained variance, whereas the two others did not. The difference
between the subscales was considered relevant to present. Thirdly, the model with three scales
had stronger classification accuracy rate (86.8%) than the model with one scale (84.3%), and
explained more variance (Cox and Snell R2: 33.4%, Nagelkerke R2: 50.2%) than the model
where the total NR scale was predictor (between 26.6 and 38.9%).
Three cases were tried excluded because of high standardized residuals, which lead to
stronger classification accuracy rate (87.7%), but inflated odd ratio values (Vegetarianism,
OR = 4.333E+10). Taken together, reporting the model with three subscales was considered
more promising. However, the logistic regression model for NR total can be found below.
only the models with the three subscales were reported for FRAMING and DEFAULT.
Table E.1
Binary Logistic Regression of Vegetarian Wrap Selection in Control versus Default Frame Condition.
n = 159
95% CI for
Odds
Odds
Ratio
B
SE (B)
Wald
df
p
Ratio
Lower
Upper
Constant
-2.05
.40
26.23
1
.000
.13
-2.05
.40
Default Frame nudge
.64
.50
1.63
1
.202
1.90
.71
5.08
NR
1.60
.74
4.67
1
.031
4.96
1.16
21.16
Vegetarianism
4.40
1.08
16.72
1
.000
81.43
9.89
670.60
NR x Default Frame
-2.17
.92
5.58
1
.018
.11
-2.17
.92
Note: R2 = .60 (Hosmer & Lemeshow), .27 (Cox and Snell), .39 (Nagelkerke). Model χ2 (3) = 49.25, p < .001.
64
EFFECTS OF NUDGING ON SUSTAINABLE FOOD CONSUMPTION
Appendix F
Logistic Regression of Default versus Control
Binary logistic regression was conducted with DEFAULT nudge, NR subscales and
vegetarianism as predictors in Block 1, interaction terms between subscales and nudge in
Block 2 (see table F.1). The full model was significant, χ2 (3, n = 165) = 81.615, p < .001, but
had poor goodness of fit according to the Hosmer-Lemeshow test (p = .022). When three
outliers were excluded based on high standardized residuals, the Hosmer-Lemeshow test was
no longer significant, indicating good fit (p = .568), and classification accuracy rate increased
from 89.7% to 92.9%. However, the revised model gave very large odds ratios (NR
Perspective, OR = 1228.12 and Vegetarianism, OR = 1.435E+12), which could be due to too
few participants in some cells. Thus, the original model with all cases is reported, but its poor
goodness of fit should not be considered.
Table F.1
Binary Logistic Regression of Vegetarian Wrap Selection in Control versus Default Condition. n = 165
95% CI for
Odds
Odds Ratio
B
SE (B)
Wald
df
p
Ratio
Lower
Upper
Default nudge
.55
.62
.78
1
.378
1.73
.51
5.84
NR Perspective
2.59
.80
10.37
1
.001
13.33
2.76
64.46
NR Experience
.24
.61
.16
1
.689
1.28
.39
4.21
NR Self
.36
.72
.25
1
.620
1.43
.35
5.83
Vegetarianism
5.07
1.21
17.58
1
.000
159.38
14.89
1706.37
NR Perspective
x Default
NR Experience
x Default
NR Self
x Default
Constant
-1.73
.97
3.17
1
.075
.18
.03
1.19
-.36
.86
.18
1
.672
.67
.13
3.75
.07
.98
.01
1
.940
1.08
.16
7.32
-2.49
.51
24.25
1
.000
.08
Note: R2 = .02 (Hosmer & Lemeshow), .39 (Cox and Snell), .58 (Nagelkerke). Model χ2 (8) =82.63, p < .001.
65