Content uploaded by Magda Osman
Author content
All content in this area was uploaded by Magda Osman on Jul 10, 2018
Content may be subject to copyright.
HBAS #1356304, VOL 0, ISS 0
Nudge: Concept, Effectiveness, and Ethics
Yiling Lina, Magda Osman, and Richard Ashcroft
QUERY SHEET
This page lists questions we have about your paper. The numbers displayed at left can be found in the text of the paper for reference. In
addition, please review your paper as a whole for correctness.
Q1: Au: Because DST stands for “dual-system theories” it does not need to be plural with a final “s”; thus the acronym is
always DST.
Q2: Au: The acronym PSC is used twice in this paragraph. Please write out PSC in both cases. Was this to stand for
“parallel constraint satisfaction”?
Q3: Au: Is “kind” correct in the sentence “Type 1 nudges kind minimally …”?
Q4: Au: Is “the evidence-based suggests that …” correct as phrased here?
Q5: Au: Arad and Rubinstein: Please indicate if this is an unpublished manuscript or other type of publication.
Q6: Au: [Hollands et al., 2015]: Please supply page range and publisher city.
TABLE OF CONTENTS LISTING
The table of contents for the journal will list your paper exactly as it appears below:
Nudge: Concept, Effectiveness, and Ethics
Yiling Lina, Magda Osman, and Richard Ashcroft
BASIC AND APPLIED SOCIAL PSYCHOLOGY
https://doi.org/10.1080/01973533.2017.1356304
5Nudge: Concept, Effectiveness, and Ethics
Yiling Lina, Magda Osman , and Richard Ashcroft
Queen Mary University of London
ABSTRACT
Nudges are psychologically informed tools designed to promote behavioral change in order to
improve health and well-being. In this review we focus on three areas of concern: theory, evidence
base, ethics. We begin by discussing the problems arising from the theoretical framework that
nudges are based on and propose an alternative framework that helps to classify nudges into two
types (Type 1 and Type 2). We then evaluate the evidence for nudges in the health domain,
drawing attention to critical empirical issues (internal and external reliability) that explain the
limited evidence base for their effectiveness. The review ends with an examination of the
implications of the theoretical and empirical issues we discussed with respect to current debates
regarding the ethics of nudge.
Noncommunicable diseases (NCDs)—principally, heart
diseases, stroke, cancer, diabetes, and chronic lung
25 diseases—are responsible for almost 70% of global
deaths (World Health Organization [WHO], 2017a).
However, most NCDs can be reduced by targeting
four main risk factors: tobacco use, physical
inactivity, harmful use of alcohol, and an unhealthy
30 diet. These factors are speculated, by some, to have a
common cause, which is poor health choices resulting
from our psychology (Thaler & Sunstein, 2008). That
is, we make choices in our day-to-day lives based on
heuristics (such as anchoring, availability, representa-
35 tiveness), and biases (optimism, overconfidence, status
quo) that drive poor lifestyle choices. Thus, based on
this speculation, a potential way of targeting NCDs is
to identify the psychological factors that contribute to
poor health choices and use behavioral interventions
40 to exploit social scientific research on human behavior:
“nudges.”
A nudge is “any aspect of the choice architecture that
alters people’s behavior in a predictable way without
forbidding any options or significantly changing their
45 economic incentives” (Thaler & Sunstein, 2008, p. 6).
Nudges influences choice behavior in a variety of ways
that include (a) provision of information (e.g., leaflets
about the benefits of climbing stairs), (b) correcting
misapprehensions about social norms (e.g., informing
50 individuals of peer group behavior such as statistics of
average alcohol intake), (c) altering the profiles of dif-
ferent choices (e.g., making healthy food appear more
prominent in the canteen), (d) implementing default
options (e.g., changing an organ donation legislative
55system from opt-in to opt-out; Bonell, McKee, Fletcher,
Wilkinson, & Haines, 2011). The guiding principle
behind these examples is to make the “better” option
more convenient or salient for the decision maker to
select; this option is better because it maximizes future
60health, wealth, and well-being.
The views on the use of nudges in the health domain
range from those praising their potential benefits (Mills,
2013; Saghai, 2013; Sunstein, 2015) to those that raise
doubts as to how they are used (Goodwin, 2012; Mols,
65Haslam, Jetten, & Steffens, 2015; Osman, 2016; Selinger
& Whyte, 2012) and whether they are effective (Bonell
et al., 2011; Marteau, Ogilvie, Ronald, Suhrcke, & Kelly,
2011; Rayner & Lang, 2011). If these doubts are
warranted, then one area that needs inspection is the
70theoretical framework on which nudges are built on,
given that this forms the rationale for how they are
supposed to operate (Baldwin, 2014). We use this as a
point of departure in our review by first examining
the proposed psychological mechanisms that underpin
75nudges and the problems associated with the theoretical
framework adopted. We use this opportunity to propose
an alternative theoretical account of nudges as a way to
rethink the evidence based of nudge interventions in the
health domain. We end the review by discussing how
80the issues we raise also have implications for ethical
debates (Saghai, 2013) and their impact on applied
social policy issues (Hansen & Jespersen, 2013).
none defined
CONTACT Magda Osman m.osman@qmul.ac.uk School of Biological and Chemical Science, Queen Mary University of London, Mile End Road, London
E1 4NS, United Kingdom.
© 2017 Taylor & Francis
3b2 Version Number : 11.0.3184/W Unicode (Apr 10 2014)
File path : P:/Santype(JATS)/Journals/TandF_Production/HBAS/v0n0/HBAS1356304/HBAS_A_1356304_J.3d
Date and Time : 27/10/17 and 18:16
Dual-system theoretical underpinnings
of nudges
85 The current theoretical framework used to support the
foundations of nudge is based on a dual system frame-
work (Kahneman & Frederick, 2002; Stanovich, 1999).
The idea is that our judgment, decision-making, and
reasoning processes are underpinned by two distinct
90 cognitive systems: System 1 and System 2. The received
view on dual-system theories (DST) is that generally
there are some family resemblances in the way the
two systems operate. In general System 1 processes are
heuristic-based, intuitive, biased, associative, automatic,
95 and System 2 processes are rule based, analytical,
flexible, and slow (Evans & Stanovich, 2013; Kahneman,
2011). Nudge theorists (Sunstein, 2015, 2016a; Thaler &
Sunstein, 2008) claim that the basis on which we make
poor lifestyle choices is commonly through the acti-
100 vation of System 1 type processes. Because of this, a
practical way of generating positive behavioral change
is to target System 1 process by reorienting the features
of the choice contexts on which heuristics and biases are
invoked; this is typically achieved covertly without the
105 decision maker’s awareness. This in turn relies on a dis-
tinction between implicit and explicit processes whereby
implicit processing occurs without awareness, whereas
explicit processing is deliberate and is accompanied by
awareness (Evans & Over, 1996; Stanovich & West,
110 2000). We focus on two core issues in the evaluation
of System 1/System 2 distinctions: (a) the nature of
the interaction between the two systems and (b) the lack
of precision around the details of key distinctions
between the two systems (Osman, 2016). Our evaluation
115 is designed to better understand how to target behavior
via nudges, because having a better idea of the actual
underlying mechanism that guides decision making
should reveal where and where not nudges are likely
to be effective (Grüne-Yanoff, 2016; Grüne-Yanoff &
120 Hertwig, 2015).
Critical issues with DST
Issue 1
The nature of the relationship between system 1 and 2.
DST
Q1 do not all make the same claims regarding the
125 relationship between systems (Evans & Stanovich,
2013). Some dual-system theorists claim that the
systems are interactive (Sloman, 1996), some claim that
they operate in parallel (Epstein, 1973, 1994; Evans &
Over, 1996), and some claim that they operate
130 serially (Gilbert, 1989); System 1 is the default system,
and only later does System 2 kick in to monitor the out-
puts of System 1 (Evans & Stanovich, 2013; Kahneman,
2003)—known as the default-interventionist approach.
If, as nudge proponents claim, Systems 1 and 2 are dis-
135sociated and they do not interact, then it makes better
sense to try to directly tap into System 1 processes to
generate behavioral change; this is the default system
that gives rise to many of the core decision-making
and reasoning processes that drive suboptimal lifestyle
140choices. Though, by doing so, nudge theorists and prac-
titioners need to identify which common suboptimal
behaviors are exclusively driven by System 1 processes,
which thus far the program of nudges has yet to do.
However, if it is the case that the two systems do
145interact, then does it still make sense to introduce an
intervention solely on System 1? If not, then nudges
may not operate in the way that they are intended. In
other words, if the grounds on which they are based is
theoretically problematic, this may also explain why
150they are not as effective as they are purported to be.
To explore Issue 1 further we consider the bat and
ball task (Kahneman & Frederick, 2002) that Thaler
and Sunstein (2008) used to illustrate the relationship
between Systems 1 and 2. The task involves presenting
155participants with this simple description and a question:
“A bat and ball cost $1.10 in total. The bat costs $1.00
more than the ball. How much does the ball cost?”
Typically the majority of participants will intuitively
answer 10 cents (Kahneman & Frederick, 2002) when
160the correct answer is 5 cents. Just based on this task
alone, DST vary significantly in their explanations for
this common error. One interpretation is that System 1
is invoked automatically and is the sole driver of the
error (Kahneman & Frederick, 2002, 2005), which is
165the favored interpretation of Thaler and Sunstein
(2008). An alternative explanation is that System 2 is
also in operation but fails to detect the error generated
by System 1 (Kahneman & Tversky, 1982), and a third
interpretation is that because System 2 is slower than
170System 1, System 2 detects the error but cannot inter-
vene quickly enough to prevent the error being made
(Gilbert, 1989; Stanovich & West, 2000). The role that
System 2 plays in this task in turn has implications for
how researchers develop methods to ameliorate the
175error effectively. Thus, the same problem extends to
nudges, in which the reasons behind a poor choice
being made, in turn will impact what appropriate inter-
vention is needed to reduce the chances of it continually
being made.
180Issue 2
Lack of precision regarding the critical distinctions
between the two systems. There are essentially three
different proposals regarding the core qualitative dif-
ference between System 1 and System 2. One view is
2 Y. LIN ET AL.
185 that they can be differentiated based on their
demands on working memory (De Neys, 2006; Evans
& Stanovich, 2013; Oppenheimer, 2008); working
memory is a system for the temporary holding and
manipulation of information during the performance
190 of a range of cognitive tasks. A second claim is that
they differ depending on the extent to which meta-
cognitive processes are invoked (Thompson, 2009);
metacognition broadly refers to explicit knowledge
or beliefs of factors that affect the outcome of a cog-
195 nitive operation. The third is that they vary to the
extent to which representations are accessible (Kah-
neman, 2003); the ease (or effort) with which parti-
cular mental contents come explicitly to mind.
Essentially, the greater the dependency on working
200 memory, or metacognitive processes, or difficulty in
accessing representations suggests System 2 is in
operation, and the opposite applies for System 1.
There are three main problems with the apparent
qualitative distinctions between Systems 1 and 2. The
205 first is, and as has been highlighted by DST, that in
actual fact the three qualities essentially reduce to one
single factor, namely, dependency on working memory
(Evans & Stanovich, 2013). This in and of itself is not
necessarily a concern, as it suggests strong compatibility
210 between the theoretical claims. What is a concern is that
dependency on working memory is not all or nothing,
and modeling and empirical demonstrations of the
way in which high-order cognition relies on working
memory is based on quantitative differences, not
215 qualitative ones (e.g., Schmiedek, Oberauer, Wilhelm,
Süß, & Wittmann, 2007). Second, the purported core
qualitative distinction between Systems 1 and 2—be that
working memory, metacognition, or accessibility of
representations—is used to explain why one system is
220 implicit/automatic and the other is explicit. As with
working memory, measures identifying automatic and
explicit processes often rely on speed of response, which
is a continuous measure, meaning that there are a
variety of judgments, choices, and inferences that are
225 made, some of which are faster than others (Osman,
2004, 2007, 2013, 2014; Osman & Stavy, 2006). The rela-
tive nature by which automatic and explicit responses
are identified means that it makes better sense to claim
that some responses are faster than others, rather than
230 some are automatic and others are explicit. Thus far
no DST theorist has presented the necessary and suf-
ficient conditions by which to identify automatic
responses independently of explicit responses (Osman,
2004). Moreover, not all behaviors associated with
235 System 1 are fast, and not all behaviors associated with
System 2 are slow, and to accommodate this some
theorists have proposed a four-system framework
(Klaczynski, 2001; Sun, Slusarz, & Terry, 2005). Third,
the same initial observations that led to the formation
240of DST, and were used to theorize about the key
qualitative property that determines the difference
between the two systems, are redescribed to form
predictions of the same observations that identify
differences between the two systems. Put simply, it
245would be akin to detecting that sometimes people make
erroneous choices quickly, and correct choices slowly,
and then theorize that this is because of two underlying
systems— one of which is fast, the other which is
slow—each of which differs based on ease of access of
250information. From this, a prediction is formed that
outlines that when the slow system is being used, people
will make correct choices, but not when the fast system
is invoked. This is referred to as a particular type of
circular argument (self-dependent justification), which
255has been commonly found in the area of DST of
decision making, judgment, and reasoning (Hahn,
2011). Given the serious concerns discussed around
Issue 1 and 2, we propose an alternative.
A single system account of nudge
260Many dual-system theorists (Evans & Over, 1996; Evans
& Stanovich, 2013; Kahneman, 2003), as well as critics
of them (Osman, 2004, 2013, 2014), agree that a fully
dissociationist view of System 1 and System 2 is not
adequate for capturing the complexities in which
265decision-making processes operate. In light of this,
and other serious concerns with dual-system frame-
works, there are several single-system frameworks
(e.g., Kruglanski, 2013; Osman, 2004, 2007, 2013;
Osman & Stavy, 2006; Simon, Snow, & Read, 2004).
270Building on these unitary system frameworks, we extend
their proposals by suggesting that they reduce to a
parallel constraint satisfaction model (Simon et al.,
2004). This is essentially a connectionist approach in
which the spread of activation among nodes in the net-
275work is fully sufficient for the processing of an outcome
(e.g., a choice) and the basis on which a decision is
made (knowledge, evidence, beliefs) is coded in the
network through the pattern of weights among the
nodes (Read, Vanman, & Miller, 1997). PSC Q2processing
280is guided by the goal of maximizing consistency, which
means the need to reconfigure/reevaluate/update
knowledge/evidence/beliefs from multiple (potentially
conflicting) sources that bring about an outcome (i.e.,
judgment, choice, inference) so that both (representa-
285tions and outcome) are in alignment (i.e., coherent).
Where DSTs identify distinct types of processes that
can be classified as System 1 or System 2, in a PSC
model, variations in processing of information is
BASIC AND APPLIED SOCIAL PSYCHOLOGY 3
predicted and modeled according to the degree of
290 restructuring that needs to occur for coherence
(between knowledge and behavior) to be achieved
(Simon & Holyoak, 2002).
How does a single system account apply
to nudges?
295 We extend the parallel constraint satisfaction single-
system framework to help classify nudges into two types
that differ according to the degree to which processing
efforts are needed to maintain psychological coherence
(see Table 1 for examples). Type 1 nudges target
300 decision-making contexts that generate responses that
are not typically accompanied by critical inspection to
prompt reconsideration of the choices made. For
instance, familiar consumer-based contexts such as
supermarkets involve highly practiced patterns of
305 behavior leading to repetitive choices being made. So,
in response, Type 1 nudges involve simple interventions
such as rearranging the presentations of consumer items
in food isles to highlight options that would have
ordinarily been ignored. Type 1 nudges kind minimally
310 disrupt
Q3
the choice context to prompt some adjustment
in the way information within it is processed at the
point of decision, but not enough that the decision
maker detects any dissonance between the nudged
choice and their general value-system. Type 2 nudges
315 aim to promote a sustained reevaluation of the evidence
base on which people make their choices, and the
choices themselves, by disrupting the coherence
between the two. For example, long-term educational
campaigns promoting exercise present the benefits of
320 regular exercise, as well as the harmful effects of
continuing to be sedentary. Repeated exposure to infor-
mation of this kind is designed to create dissonance
resulting from the costs of maintaining poor habits
and the benefits of changing them. Given the cognitive
325 system’s need for coherence, a reevaluation and restruc-
turing of knowledge representations is needed to bring
in alignment the evidence base (knowledge/beliefs)
and choice behavior. It is worth reiterating here that
the two types of nudges differ according to the amount
330 of reevaluation of information on which people’s
choices are made and actions taken based on it, it is
not predicated on a difference between qualitative
differences in systems of thought. Now that we have
considered a simple way to conceptualizing nudges into
335types, we turn to the evidence base in order to critically
consider which type of nudge is shown to be effective.
Nudges: Evidence base in the health domain
In this section we examine a range of nudges (Table 1)
implemented in four health domains: (a) poor diet,
340(b) physical inactivity, (c) alcohol overconsumption
and (d) tobacco use. This is not a comprehensive review
of the evidence in the literature but instead is a focused
evaluation of core findings that provide a representative
impression of the pattern of evidence in this area (see
345the appendix for further information).
Poor diet
The size of packaging and the portions of food products
has dramatically increased over the past 30 years
(Young & Nestle, 2002), which in turn has affected
350our food consumption. A common Type 1 nudge
approach to address overconsumption has been to
change visual cues in a food environment which may
consist of the availability of certain foods; the variety
of food assortments; size of food packages and portions;
355or shape/size of plates, glasses, and bowls. These cues
are often used to imply a consumption norm that helps
regulate how much we eat or drink in a food environ-
ment (Wansink, 2004). In addition, many people adhere
to the norm “plate clean,” which means that in a food
360establishment if the plate is large, and the portions
match the size, people consume to the size of the por-
tion on the plate, and not to the point of being sated
(Schwarz, 1998). For example, depending on the plate
size, up to 45% more food is consumed in a Chinese
365buffet setting (Wansink & van Ittersum, 2013). A com-
parison of 58 studies (6,603 participants) in a recent
Cochrane review found that people consistently ate
more food when offered larger portions, packages, or
items of tableware than when offered smaller versions
370(Hollands et al., 2015). Introducing Type 1 nudges that
reduce plate size in food establishments was shown to
have a reasonable effect in reducing intake but was
Table 1. A summary of the evidence on the effectiveness of Type 1 and Type 2 nudges.
Type 1 nudges Type 2 nudges
Unhealthy diet Smaller plate sizes Limited effect Calorie labeling Traffic light labeling Ineffective effective
Physical inactivity Footprints Ineffective Motivational posters Mixed evidence
Harmful use of alcohol Adopting straight glassware Insufficient evidence Correct social norm misapprehension Mixed evidence
Tobacco use Shorter cigarettes Ineffective Health warnings on branded packs Mixed evidence effective
Health warnings on plain packs
4 Y. LIN ET AL.
dependent on participants being unaware of the
manipulation (Holden, Zlatevska, & Dubelaar, 2016).
375 However, it isn’t clear that the intervention successfully
generalizes to contexts beyond the one in which the
nudge was implemented. Moreover, there are many fac-
tors that significantly influence our eating habits. Eating
is often a social activity, and we take our cues as to how
380 much to consume from our dining partners, and dis-
tractions encourage us to overconsume (e.g., watching
television, watching movies at the cinema), which
suggests that there are multiple countervailing factors
that limit the scope of Type 1 nudges (Wansink, 2004).
385 In this context, as an alternative to Type 1 nudges,
Type 2 nudges focus on improving the presentation of
information on which people make their food choices,
such as providing calorie counts on food menus to draw
attention to both healthy and unhealthy options
390 (Downs, Loewenstein, & Wisdom, 2009). However,
evidence suggests that this has had limited effects in
increasing healthy food choices (Loewenstein et al.,
2012). One reason for this is that calorie labels do not
provide an obvious reference point as to which specific
395 options are best for people. In and of themselves, they
do not reliably motivate people to systematically moni-
tor and translate calorie counts to shift their choices
over a sustained period. Indeed, a recent systematic
review of the impact of reading calorie labels at the
400 point of purchase or consumption has had little to no
effect on positively changing people’s choice behavior
(Kiszko, Martinez, Abrams, & Elbel, 2014; Sinclair,
Cooper, & Mansfield, 2014). The review found that
regardless of the length of the intervention, the Type 2
405 nudge was generally ineffective (4-week period: Elbel,
Gyamfi, & Kersh, 2011; Elbel, Kersh, Brescoll, & Dixon,
2009; K. L. Webb, Solomon, Sanders, Akiyama, &
Crawford, 2011; 2-month period: Holmes, Serrano,
Machin, Duetsch, & Davis, 2013); 13-month period:
410 Finkelstein, Strombotne, Chan, & Krieger, 2011). To
deal with these concerns, traffic light systems have been
used as a way to make calorie information more salient
and intuitively simpler to interpret (Sacks, Rayner, &
Swinburn, 2009; Sonnenberg et al., 2013). By using
415 simple visual cues (e.g., red ¼highly fat food, amber ¼
moderately healthy foods, green ¼very healthy foods),
the signals directly connect calorie counts to their
impact on health (House of Lords, 2011) and provide
a relevant reference point (Liu, Wisdom, Roberto, Liu,
420 & Ubel, 2014). An independent field study conducted
by Ipsos Mori showed that 35% of customers of a major
UK supermarket actively look at traffic light labels when
they shop, and 92% of those find these labels easy to
understand. Also, over the 12-week period, sales of food
425 items with mostly green traffic lights grew to 46.1%,
whereas those with mostly red traffic lights decreased
by 24% (House of Lords, 2011). Thus, in the United
Kingdom alone, several organizations (The National
Institute for Care and Health Excellence; The UK Food
430Standard Agency) have strongly encouraged food
manufacturers and food establishments to use text and
traffic lights on food labels/menus because they support
better understanding of which foods are healthy and
which are not.
435Physical inactivity
As with poor diet, another major global issue is the
significant decreases in regular exercise (Hallal et al.,
2012). A simple method of increasing physical activity
through a Type 1 nudge involves a point-of-decision
440prompt that uses visual cues in relevant contexts to
encourage people to take the more active of two options
(e.g., a choice between taking the stairs rather than the
escalator). For instance, painted footprints on stairwells
have been used to guide people to take the stairs over
445elevators. The evidence base for this is not encouraging.
Findings show that, perversely, the method increased
the selection of the less physical option (Åvitsland,
Solbraa, & Riiser, 2017). Although similar to the Type 1
nudge, a favored Type 2 nudge, is to also use a point-of-
450decision prompt but instead of covert visual cues,
educational information is presented that highlights
the benefits of regular exercise. This often involves
placing posters at start of stairwells or by elevators/
escalators that inform people about the calories they
455would burn or the net positive effects on their health
(i.e., increased heart rate; Andersen, Franckowiak,
Snyder, Bartlett, & Fontaine, 1998; Blamey, Mutrie, &
Aitchison, 1995; Brownell, Stunkard, & Albaum, 1980;
A. Lewis & Eves, 2012; Marshall, Bauman, Patch,
460Wilson, & Chen, 2002; Nomura, Yoshimoto, Akezaki,
& Sato, 2009; O. J. Webb & Eves, 2007). A recent review
of the evidence reported that across 11 studies the
success rate was at chance levels (Soler et al., 2010),
but a second review reported an overall positive effect
465ranging between 0.3% and 10.6% (Nocon, Müller-
Riemenschneider, Nitzschke, & Willich, 2010). A specu-
lation in the difference between these two reviews is that
there is variability in the length of the intervention. The
length of the intervention is often between 4 and
47012 weeks, but with some notable exceptions (24 weeks:
Kerr, Eves, & Carroll, 2001b; 9 months: Lee et al., 2012).
But further work has shown that there is no association
between effectiveness of the Type 2 nudge and length of
intervention. Another potential explanation for the
475mixed findings is that the locations in which the nudge
was implemented varied, and so it is hard to compare
BASIC AND APPLIED SOCIAL PSYCHOLOGY 5
like for like, for instance, comparing returning popula-
tions taking the stairs at train station versus those at a
shopping mall. To further uncover the precise reasons
480 for why Type 2 nudges vary in their effectiveness in this
domain, the types of informational prompts have been
evaluated. A specific message such as “7 minutes of stair
climbing protects your heart” was shown to be more
effective than a general message such as “Stay healthy,
485 use the stairs” (Puig-Ribera & Eves, 2010).
It is also worth noting that studies examining Types 1
and 2 nudges typically involve point-of-decision
prompts placed at stairwell with only one to two flights
of stairs. In order to experience any significant impact
490 on cardiorespiratory fitness women need to climb at
least six flights of stairs daily (Boreham, Wallace, &
Nevill, 2000), and men need to climb 25 flights to result
in any significant improvement in aerobic power (Fardy
& Ilmarinen, 1975). Even if some of the studies have
495 shown positive effects of nudges, they would fall short
of any meaningful impact on actual physical fitness
levels.
Alcohol overconsumption
The evidence base for nudges designed to reduce
500 alcohol overconsumption accounts for only 7.3% of all
behavioral intervention studies in the health domain
(Hollands et al., 2013), despite the severity of the prob-
lem (Magnusson, 2009). Akin to the Type 1 nudges used
to reduce food consumption via altering the size of food
505 containers, a similar rationale has been adopted in the
context of alcohol consumption. This typically involves
offering alcohol in tall, narrow glasses as opposed to
short, wide glasses in drinking establishments (bars
and public houses; Wansink & van Ittersum, 2005). This
510 is motivated by work showing that the rate of alcohol
consumption is related to the shape of glassware, which
is slower in a straight glass compared to a curved
glass (Attwood, Scott-Samuel, Stothart, Munafò, &
Campanella, 2012). A recent systematic review of stu-
515 dies examining the use of this nudge reported that there
was not enough evidence to estimate the effect on
reducing consumption (Hollands et al., 2015).
In line with other Type 2 nudges discussed so far, a
preferred method is to provide explicit information as
520 a means of generating behavioral change, for example,
providing a more accurate idea of the safe quantities
to consume through the use of social norm cues. As
social creatures, people are sensitive to majority influ-
ences, and this can be a strong pervasive influence on
525 behavior (Bullers, Cooper, & Russell, 2001; Ennett
et al., 2006; Pearson & West, 2003). That is, a wealth
of evidence from social psychology shows that people
behave in accordance with their peers (J. V. Wood,
1989), often as part of a group mentality; there is a
530strong drive to belong and be accepted by a group.
Given that consuming alcohol is typically a social
activity, the claim is that using social norm cues (i.e.,
the typical amount of alcohol a particular social group
consumes) is a more efficient way of helping people
535regulate their alcohol consumption by evaluating it rela-
tive to the consumption of their peers (Nishida, Akaoka,
& Nishizawa, 1975). For instance, heavy drinkers often
judge their alcohol consumption to be equal to or even
less than their peers, even though it is substantially
540greater (Perkins, Meilman, Leichliter, Cashin, & Presley,
1999; they feel as if they can reasonably justify their
behavior by rationalizing that it is no different than
their peers’). To correct misapprehensions of social
norms in a student population, several studies using
545self-reported survey responses have shown that Type 2
social norm interventions (through educational cam-
paigns) implemented over 1 year (Gomberg, Schneider,
& DeJong, 2001) and 5 years (Haines & Spear, 1996)
have successfully reduced alcohol consumption. How-
550ever, a different review of 66 studies analyzed alcohol
reduction at 4 months postintervention and found that
the effect sizes were small and were unlikely to be of
meaningful benefit in practice (Foxcraft, Moreira,
Santimano, & Smith, 2015). It is worth noting that when
555surveyed, students doubted the credibility of the edu-
cational campaign messages (Thombs, Dotterer, Olds,
Sharp, & Raub, 2004). In addition, some have suggested
that the effectiveness of the nudge needs to take into
account campus sizes in which norm misperception
560may be harder to correct if “everybody knows everyone
else,” and thus students are more confident in their esti-
mates of others’ drinking levels (Borsari & Carey, 2003).
Moreover, it is possible that the average or typical norm
used to compare drinking levels in these Type 2 nudges
565does not represent the ideal normative reference point
(M. A. Lewis & Neighbors, 2006). In other words,
feedback that involve best friends’ drinking rather than
typical student drinking level would be more specific
and may have a stronger influence, assuming the peers
570are actually consuming alcohol within healthy limits
(Baer, Stacy, & Larimer, 1991; Borsari & Carey, 2003;
M. A. Lewis & Neighbors, 2006).
Tobacco use
Another serious problem is tobacco consumption, which
575kills around 6 million people each year (WHO, 2016). To
target this, Type 1 nudges promoting smoking cessation
have focused on increasing the availability of shorter
cigarettes; however, a systematic review found that when
6 Y. LIN ET AL.
compared to standard-sized cigarettes, there was no over-
580 all reduction in tobacco consumption (Hollands et al.,
2015). Alternatively, a more common route is to adopt
educational campaigns that are typical of Type 2 nudges.
In many Western countries, health warnings on cigarette
packages are among the most common means of
585 increasing smokers’ awareness of the risks of smoking
(Hammond et al., 2006). It is now mandatory that consu-
mers of tobacco products have a “fundamental right to
health information, including accurate information about
the harms of tobacco use” (WHO, 2015). A comprehen-
590 sive review by Hammond et al. (2006) found that
smokers’ knowledge of toxic constituents in tobacco
smoke was very low even among smokers in affluent
and educated countries in the world. Although this is a
common practice for many countries across the world,
595 the style of presentation of health information differs
between countries, making it difficult to evaluate the
effectiveness of these messages on reducing consumption.
More recently, Type 2 nudges such as plain cigarette
packaging standardize the shape, color, and method of
600 opening the package, as well as the health warnings them-
selves. The aim is to fulfill several objectives that include
reducing the attractiveness of consuming tobacco
(Hammond, Daniel, & White, 2013; Hammond &
Parkinson, 2009; Moodie & Mackintosh, 2013; Moodie,
605 Mackintosh, Hastings, & Ford, 2011) and restricting use
of the pack as a form of advertising and promotion while
increasing the size of the health warnings (Maynard,
Munafò, & Leonards, 2013; Moodie et al., 2012; Munafò,
Roberts, Bauld, & Leonards, 2011). A large review of 37
610 studies concluded that plain packaging was rated as less
attractive and contained poorer quality products than
branded packaging (WHO, 2017b), although this did
not indicate the impact on tobacco consumption.
Empirical work looking at the impact of health
615 warnings on tobacco consumption is still in its infancy,
but the findings are promising and suggest that they
indeed reduce acute craving and are often associated
with more negative perceptions of smoking (Brose,
Chong, Aspinall, Michie, & McEwen, 2014). In the short
620 term, plain packaging has been shown to encourage
cessation for up to 2 weeks (Moodie & Mackintosh,
2013). In the medium and long term, there is evidence
to suggest that plain packaging decreased tobacco
consumption 6 months postintervention (Dunlop,
625 Dobbins, Young, Perez, & Currow, 2014) as well as
12 months postintervention (Wakefield et al., 2015).
Meanwhile Australia, as the first WHO member to
implement standardize packaging, has also seen a stat-
istically significant decline in smoking prevalence as a
630 result of this Type 2 nudge (Australian Government,
Department of Health, 2016).
Empirical issues concerning nudge
Having examined the available evidence of both Type 1
and Type 2 nudges, we evaluate the methodological
635issues concerning the implementation of nudges
designed to promote health behaviors. As with any
intervention designed to improve behavior, the most
reliable way to confidently make casual inferences about
a manipulation and its possible effect is to compare it
640against a control condition (randomized control trials).
However, randomized control trials are hard to intro-
duce in field work, and so this, along with other factors,
limits the ability to draw firm conclusions as to the
effectiveness of nudges. Beyond this, we next raise two
645key points that we consider need addressing in future
empirical research on nudge.
Internal reliability of experiments examining nudges
Internal reliability refers to the extent to which a
measure is consistent within itself, namely, it generates
650the same behavior each time it is used within the same
context. The Type 1 nudges just reviewed suggest that
overall the evidence base is mixed and that the replic-
ability of positive nudge interventions is hard to
establish. This raises questions about the reasons for
655when Type 1 nudges do work, and why the effects are
hard to replicate. Loewenstein, Bryce, Hagmann, and
Rajpal (2015) speculated that the limited effectiveness
of Type 1 nudges results from peoples lack of deep
insight into how the nudge is designed to influence their
660behavior. Ashcroft (2013) proposed that the effective-
ness of both Type 1 and 2 nudges in general may
depend on the various heterogeneous motivations/value
systems people have with regards to changing their
behavior in line with a healthier lifestyle. In addition,
665the fact that nudges are highly context dependent
(Kosters & Van der Heijden, 2015) means that some
Type 1 nudges are less likely to work in some contexts
rather than others, and a clearer understanding of the
context in which they are implemented is needed. Thus,
670a critical step in devising research programs around
nudging in health domains is to establish the internal
reliability of nudges over time at an epidemiological level
but also at the individual level. However, the research
practices so far have yet to adopt methodological techni-
675ques that tackle any of these issues in depth (i.e., asses-
sing motivation needs, levels of awareness of nudges,
characterizing the contexts in which they are implemen-
ted) in order to better establish internal reliability.
External reliability of experiments examining nudge
680External reliability refers to the extent to which a
measure varies from one use to another. With most
BASIC AND APPLIED SOCIAL PSYCHOLOGY 7
nudge field experiments, the difficulty is in reproducing
the same conditions in different contexts under which
the original intervention was assessed. For instance,
685 consider nudges designed to increase physical activity.
The informational prompts have been used in various
environmental settings such as libraries (Russell,
Dzewaltowski, & Ryan, 1999), underground stations
(Blamey et al., 1995), and office buildings (Coleman &
690 Gonzalez, 2001). In meta-analytic reviews of these stu-
dies (Andersen et al., 1998; Blamey et al., 1995; Brownell
et al., 1980; Kerr, Eves, & Carroll, 2001a; Kerr et al.,
2001b; Marshall et al., 2002; Nomura et al., 2009), none
suggested that there was a consistent pattern of evidence
695 across the different contents that were studied. The
positive impact on behavior as indicated in these studies
increased stair use over elevators/escalators varied from
around 2% to 12% but not controlling for length of time
in which the measure was implemented, that is, 1 month
700 or 3 months. Similar inconsistencies have also been
noted for nudges that extend beyond the health domain.
For instance, nudges used to increase civic behaviors,
such as recycling, volunteering, voting, petitioning,
donating, and debating, have shown that the variation
705 in how long and where they are implemented may
explain why overall effect size is as low as 9% (John
et al., 2013). Thus, with findings such as this there need
to be more efforts in standardizing the ways in which
nudges are examined in the wild in order to establish
710 external reliability.
Thus, from an empirical perspective the picture
appears to be somewhat bleak with respect to establish-
ing good evidence for the effectiveness of nudges in the
health domain. The main problem is that it is hard to
715 draw any firm conclusions as to their effectiveness in
the long term (i.e., positive change over 1 year or more),
which should be the ultimate goal of assessing their
effectiveness. Often because the studies are conducted
in the field, as with many field-based studies, the prob-
720 lems are that it is hard to run studies on a large sample
with proper controls, and it is rare to find field studies
that also carry out follow-ups to examine the effects of
the nudges in the long term. This does not undermine
the program of nudge per se but simply that the
725 evidence to date does not allow researchers to draw
strong conclusions about its general effectiveness in
generating meaningful positive behavioral change.
Moreover, as noted, the limitations in drawing firm
conclusions is restricted not only to their effectiveness
730 and reliability over time but also in establishing the gen-
eralizability of positive behavioral change beyond the
context in which nudges are implemented. Further-
more, the small effect sizes reported in empirical studies
means that translating their positive results at a
735population level may render them less effective than
typical social policy methods (i.e., mandates, bans,
taxes). Given that there is a growing list of international
governments wanting to apply nudge to public policy
on important issues such as health and well-being, there
740is clearly a need to establish further empirical rigor in
order to better establish the effectiveness of these beha-
vior interventions (Osman, 2016).
General theoretical reconsiderations of the
nudge evidence base
745As discussed earlier in this review, a concern for the
nudge program is that the theoretical foundations on
which it is built are problematic. We have proposed that
the types of nudges that have been developed fall into
two broad categories, which differ depending on the
750extent to which they promote a reevaluation of infor-
mation that informs better decisions (i.e., maximizing
long-term gains), so as to bring the new information
and choice behavior into greater alignment (greater
coherence). This is in contrast to a position of Thaler
755and Sunstein (2008) and Sunstein (2014, 2016c) that
nudges differ according to the underlying differences
between System 1 and System 2 thought operations.
Building on our proposals and the evidence we
reviewed, one reason why Type 1 nudges seem to be
760ineffective and tend to be short-lived is because they
do not engage the decision maker on any substantial
level to reexamine the basis on which the decisions
are made so as to meaningfully shift their choice beha-
vior. This is consistent with Loewenstein et al.’s (2015)
765claim regarding the level of insight that people have as
to the underlying basis on which Type 1 nudges are
designed to influence their behavior. Indeed, without
prompting people to think about and acknowledge that
they might be eating/drinking less as a result of smaller
770dinnerware/glassware, any behavioral change is not
likely to become sticky (i.e., habitual), or reliably gener-
alize to other contexts outside of where the nudge is
present. It has long since been known that habits
require sustained and explicit association between situa-
775tional cues and learned behavioral responses (Hull,
1943), often through repetition of a behaviors in the
same context (W. Wood, Quinn, & Kashy, 2002) for
the behaviors to generalize beyond them.
The evidence we have reviewed regarding Type 2
780nudges typically involve interventions that involve the
provision of explicit information that is directly connec-
ted to the pursuit of a clearly identified goal, which in
turn is associated with a specific choice behavior (e.g.,
reduce unhealthy eating, alcohol consumption, tobacco
785consumption); this has been in the form of providing
8 Y. LIN ET AL.
calorie information, a peer group’s alcohol consump-
tion, or health warnings on cigarette packages. More
to the point, Type 2 nudges seem to be effective in
reducing poor health behaviors such as alcohol con-
790 sumption (Haines, Barker, & Rice, 2003; Haines &
Spear, 1996) and cigarette smoking (Hancock, Abhold,
Gascoigne, & Altekruse, 2002; Hancock & Henry,
2003; Linkenbach, Perkins, & DeJong, 2003) for a
period equal to or greater than 12 months. The evidence
795 shows that through repeated intervention over long
periods, some Type 2 nudges (particularly those
correcting misapprehensions of social norms) can lead
to sustainable behavioral change over longer periods
(i.e., over 1 year postintervention). Thus, we suggest
800 here that in order to establish reliable methods that pro-
mote critical reexamination of one’s values, attitudes,
and motivations, we advocate that Type 2 nudges
should be more frequently used, and over sustained
periods (i.e., at least 6- to 12-month educational
805 campaigns). The rationale for this is that, unlike Type 1
nudges, Type 2 nudges typically encourage a form of
reevaluation of behavior through explicit means, and
this helps to maintain greater coherence between the
information on which new choice behaviors are made
810 coherently. We speculate that it is for this reason that
the evidence-based suggests that
Q4 they are relatively more
effective in leading to meaningful sustained behavioral
change than Type 1 nudges (see Table 1).
Ethical implications
815 Our review of the theoretical foundations of nudge, as
well as the evidence base examining the efficacy of
nudges in the health domain, suggests that there is good
reason to focus on implementing Type 2 nudges over
Type 1 nudges. That is to say, there are theoretical
820 grounds on which Type 2 nudges can be argued to have
a more sustainable and deeper impact on generating
positive behavioral change, and in line with this, the
current evidence base indicates, to some degree, that
they are more effective than Type 1 nudges. Not only
825 are there theoretical and empirical grounds for promot-
ing Type 2 nudges, but here we briefly discuss the
ethical reasons that corroborate this conclusion.
By definition, nudges are designed to influence
choice behavior, but not at the expense of forbidding
830 any options. That is, they preserve people’s rights to
freely choose whatever option they like, but that the
nudge is designed to highlight the option deemed better
for them in the long run. This is why the nudge pro-
gram is liberal paternalistic (Sunstein, 2016c; Thaler &
835 Sunstein, 2008). This approach is designed to spare pol-
icymakers any ethical concerns that they might likely
face, because a certain choice behavior is being encour-
aged by the state that promotes a certain value-based
lifestyle approach but that preserves the right an indi-
840vidual has to do otherwise (for an in depth discussion,
see Osman, 2016). What has come into question is
whether nudges easily allow people to do otherwise.
Debates have arisen because some have argued that
the way in which nudges operate, specifically Type 1
845nudges, does not preserve freedom of choice, because
choice behavior is predominately influenced without
the awareness of the individual (Ashcroft, 2013;
Blumenthal-Barby & Burroughs, 2012; Bovens, 2009;
Dworkin, 2012; Osman, 2016; Saghai, 2013). That is, it
850would be hard to do otherwise, if one is choosing an
option without any awareness of how and why it is
being chosen. From an empirical standpoint this is a
moot point because so far there simply isn’t enough evi-
dence to suggest that Type 1 nudges work, and if they
855do, there is no good evidence to suggest that they are
influencing behavior on a nonconscious level (Osman,
2014). Rather, the ethical problems that researchers
have raised consider the rationale behind the nudge
program itself, and the extent to which an endeavor that
860has global appeal should be a cause for concern. Nudge
defenders (Sunstein, 2016b; Thaler & Sunstein, 2008)
appeal to comparisons with typical social policy tools
such as mandates, taxes, and bans as a way to explicitly
and more heavy-handedly steer people to behave in a
865manner that maximizes their own and society’s good
(i.e., seat belt laws, fat tax, smoking bans). Although a
critical point to highlight here is that typical policy mea-
sures are explicit, and often accompanied by educational
campaigns, so the public is well aware of the basis on
870which their choices are being modified, even if they
aren’t necessarily happy about it (Osman, 2014; Weber,
2017). In this regard, it is worth also considering the fact
that public surveys of nudges also suggest that the
public show much higher approval ratings for Type 2
875over Type 1 nudges (Arad & Rubinstein, 2015;
Felsen, Castelo, & Reiner, 2013; Hagman, Andersson,
Västfjäll, & Tinghög, 2015; Hedlin & Sunstein, 2016;
Jung & Mellers, 2016; Mazzocchi et al., 2015; Reisch &
Sunstein, 2016; Reisch, Sunstein, & Gwozdz, 2016;
880Sunstein, 2016c; Sunstein, Reisch, & Rauber, 2017).
Again, this goes to show that, whether or not people
are going to modify their behavior in light of nudges,
or more typical policy methods of behavioral change,
they are supportive of explicit methods that signal what
885methods are being used and how they change behavior
over those that seek to do this covertly, especially with-
out their consent (Osman, 2016). On a macro level, in a
democratic society, nudges, like other governance
interventions, would be subject to evaluation by the
BASIC AND APPLIED SOCIAL PSYCHOLOGY 9
890 government official that represents the citizens’ interest
(Sunstein, 2016a). If there are indeed strong objections
against nudge, then public officials would be and
should be attentive to public views of their consent to
implement them, as well as acquiring evidence to
895 examine their effectiveness. Similarly, the duty of public
officials is to promote the welfare of the citizen in the
long term, but this cannot be done in a liberal society
without discussion and debate.
In conclusion, the motivation behind this review was
900 a fairly simple: to better understand how nudges work.
With that in mind, once this is achieved, then the better
armed we (social scientists, policymakers, practitioner)
are in designing ways of intervening on behaviors to
achieve the best outcome for individuals that need
905 and want it. In this review we argue that any meaningful
change in behavior arises from developing a consistently
coherent basis on which people understand the reasons
for their decisions and how they enact them. If, through
nudges, we want to encourage people to help them-
910 selves, particularly in targeting serious problems around
NCDs, we need to make the goal of helping oneself
making better lifestyle choices a coherent and sustained
approach. We argue that for theoretical, empirical, and
ethical reasons, this is best achieved through Type 2
915 rather than Type 1 nudges.
Funding
This work was supported by the Life Sciences Initiative
Studentship (LSIPGRS) [Grant Number LSIPGRS].
ORCID
920
Magda Osman http://orcid.org/0000-0003-1480-6657
References
Andersen, R. E., Franckowiak, S. C., Snyder, J., Bartlett, S. J., &
Fontaine, K. R. (1998). Can inexpensive signs encourage
925 the use of stairs? Results from a community intervention.
Annals of Internal Medicine, 129, 363–369.
Arad, A., & Rubinstein, A. (2015). The people’s perspective on
libertarian-paternalistic policies. Tel Aviv University, Tel
Aviv, Israel.
Q5
930 Ashcroft, R. E. (2013). Doing good by stealth: Comments on
“Salvaging the concept of nudge”. Journal of Medical Ethics,
39(8), 494. doi:10.1136/medethics-2012-101109
Attwood, A. S., Scott-Samuel, N. E., Stothart, G., Munafò,
M. R., & Campanella, S. (2012). Glass shape influences
935 consumption rate for alcoholic beverages. PLoS ONE, 7(8),
e43007. doi:10.1371/journal.pone.0043007
Australian Government, Department of Health. (2016). Post-
Implementation Review Tobacco Plain Packaging 2016.
Retrieved from http://ris.pmc.gov.au/sites/default/files/
940posts/2016/02/Tobacco-Plain-Packaging-PIR.pdf
Åvitsland, A., Solbraa, A. K., & Riiser, A. (2017). Promoting
workplace stair climbing: sometimes, not interfering is
the best. Archives of Public Health, 75, 2. doi:10.1186/
s13690-016-0170-8
945Baer, J. S., Stacy, A., & Larimer, M. (1991). Biases in the
perception of drinking norms among college students.
Journal of Studies on Alcohol, 52(6), 580–586.
Baldwin, R. (2014). From regulation to behaviour change:
Giving nudge the third degree. The Modern Law Review,
95077(6), 831–857. doi:10.1111/1468-2230.12094
Blamey, A., Mutrie, N., & Aitchison, T. (1995). Health
promotion by encouraged use of stairs. BMJ, 311(7000),
289–290. doi:10.1136/bmj.311.7000.289
Blumenthal-Barby, J. S., & Burroughs, H. (2012). Seeking
955better health care outcomes: The ethics of using the
“nudge.” The American Journal of Bioethics, 12(2), 1–10.
doi:10.1080/15265161.2011.634481
Bonell, C., McKee, M., Fletcher, A., Wilkinson, P., &
Haines, A. (2011). One nudge forward, two steps back.
960BMJ, 342, d401. doi:10.1136/bmj.d401
Boreham, C. A. G., Wallace, W. F. M., & Nevill, A.
(2000). Training effects of accumulated daily stair-
climbing exercise in previously sedentary young women.
Preventive Medicine, 30(4), 277–281. doi:10.1006/pmed.
9652000.0634
Borsari, B., & Carey, K. B. (2003). Descriptive and injunctive
norms in college drinking: A meta-analytic integration.
Journal of Studies on Alcohol, 64(3), 331–341.
Bovens, L. (2009). The ethics of nudge. In T. Grüne-Yanoff &
970S. O. Hansson (Eds.), Preference change: Approaches from
philosophy, economics and psychology (pp. 207–219).
Dordrecht: Springer Netherlands.
Brose, L. S., Chong, C. B., Aspinall, E., Michie, S., & McEwen,
A. (2014). Effects of standardised cigarette packaging on
975craving, motivation to stop and perceptions of cigarettes
and packs. Psychology & Health, 29(7), 849–860.
doi:10.1080/08870446.2014.896915
Brownell, K. D., Stunkard, A. J., & Albaum, J. M. (1980).
Evaluation and modification of exercise patterns in the
980natural environment. The American Journal of Psychiatry,
137(12), 1540–1545.
Bullers, S., Cooper, M. L., & Russell, M. (2001). Social network
drinking and adult alcohol involvement. Addictive Beha-
viors, 26(2), 181–199. doi:10.1016/S0306-4603(00)00099-X
985Coleman, K. J., & Gonzalez, E. C. (2001). Promoting stair use
in a US–Mexico border community. American Journal of
Public Health, 91(12), 2007–2009. doi:10.2105/AJPH.91.
12.2007
De Neys, W. D. (2006). Dual processing in reasoning: Two
990systems but one reasoner. Psychological Science, 17,
428–433. doi:10.1111/j.1467-9280.2006.01723.x
Downs, J. S., Loewenstein, G., & Wisdom, J. (2009). Strategies
for promoting healthier food choices. American Economic
Review, 99, 159–164. doi:10.1257/aer.99.2.159
995Dunlop, S. M., Dobbins, T., Young, J. M., Perez, D., &
Currow, D. C. (2014). Impact of Australia’s introduction
of tobacco plain packs on adult smokers’ pack-related per-
ceptions and responses: results from a continuous tracking
survey. BMJ Open, 4(12), e005836. doi:10.1136/bmjopen-
10002014-005836
10 Y. LIN ET AL.
Dworkin, G. (2012). Lying and nudging. Journal of Medical
Ethics, 39(8), 496–497.
Elbel, B., Gyamfi, J., & Kersh, R. (2011). Child and adolescent
fast-food choice and the influence of calorie labeling: a
1005 natural experiment. International Journal of Obesity
(London), 35, 493–500. doi:10.1038/ijo.2011.4
Elbel, B., Kersh, R., Brescoll, V. L., & Dixon, L. B. (2009).
Calorie labeling and food choices: a first look at the effects
on low-income people in New York City. Health Affairs,
1010 28(6), w1110–w1121. doi:10.1377/hlthaff.28.6.w1110
Ennett, S. T., Bauman, K. E., Hussong, A., Faris, R., Foshee,
V. A., Cai, L., & DuRant, R. H. (2006). The peer context
of adolescent substance use: Findings from social network
analysis. Journal of Research on Adolescence, 16, 159–186.
1015 doi:10.1111/j.1532-7795.2006.00127.x
Epstein, S. (1973). The self-concept revisited: Or a theory of a
theory. American Psychologist, 28, 404–416. doi:10.1037/
h0034679
Epstein, S. (1994). Integration of the cognitive and the
1020 psychodynamic unconscious. The American Psychologist,
49, 729–724.
Evans, J. S. B. T., & Over, D. E. (1996). Rationality and
reasoning. East Sussex, UK: Psychology Press.
Evans, J. S. B. T., & Stanovich, K. E. (2013). Dual-process
1025 theories of higher cognition: Advancing the debate.
Perspectives on Psychological Science, 8(3), 223–241.
doi:10.1177/1745691612460685
Fardy, P. S., & Ilmarinen, J. (1975). Evaluating the effects and
feasibility of an at work stairclimbing intervention program
1030 for men. Medicine and Science in Sports, 7(2), 91–93.
Felsen, G., Castelo, N., & Reiner, P. B. (2013). Decisional
enhancement and autonomy: Public attitudes towards overt
and covert nudges. Judgement and Decision Making, 8(3),
202–213. doi:10.1177/2332858416674007
1035 Finkelstein, E. A., Strombotne, K. L., Chan, N. L., & Krieger, J.
(2011). Mandatory menu labeling in one fast-food chain in
King County, Washington. American Journal of Preventive
Medicine, 40(2), 122–127. doi:10.1016/j.amepre.2010.
10.019
1040 Foxcraft, D. R., Moreira, M. T., Santimano, N. M. L., & Smith,
L. A. (2015). Social norms information for alcohol misuse
in university and college students (Review). Cochrane
Database of Systematic Reviews, 12. doi:10.1002/14651858.
CD006748.pub4
1045 Gilbert, D. T. (1989). Thinking lightly about others:
Automatic components of the social inference process. In
J. S. Uleman & J. A. Bargh (Eds.), Unintended thought
(pp. 189–211). New York, NY: Guilford Press.
Gomberg, L., Schneider, S. K., & DeJong, W. (2001). Evalu-
1050 ation of a social norms marketing campaign to reduce
high-risk drinking at the University of Mississippi*. The
American Journal of Drug and Alcohol Abuse: Encompass-
ing All Addictive Disorders, 27(2), 375–389.
Goodwin, T. (2012). Why we should reject “nudge.” Politics,
1055 32(2), 85–92.
Grüne-Yanoff, T. (2016). Why behavioural policy needs
mechanistic evidence. Economics and Philosophy, 32(3),
463–483. doi:10.1017/S0266267115000425
Grüne-Yanoff, T., & Hertwig, R. (2015). Nudge versus boost:
1060 How coherent are policy and theory? Minds and Machines,
26(1–2), 1–52. doi:10.1007/s11023-015-9367-9
Hagman, W., Andersson, D., Västfjäll, D., & Tinghög, G.
(2015). Public views on policies involving nudges. Review
of Philosophy and Psychology, 6(3), 439–453. doi:10.1007/
1065s13164-015-0263-2
Hahn, U. (2011). The problem of circularity in evidence,
argument, and explanation. Perspectives on Psychological
Science, 6(2), 172–182.
Haines, M. P., Barker, G., & Rice, R. (2003). The Northern Illi-
1070nois University experiment: A longitudinal case study of
the social norms approach. In H. W. Perkins (Ed.), The
social norms approach to preventing school and college age
substance abuse: A handbook for educators, counselors,
and clinicians (pp. 21–34). San Francisco, CA: Jossey-Bass.
1075Haines, M. P., & Spear, S. F. (1996). Changing the perception
of the norm: A strategy to decrease binge drinking among
college students. Journal of American College Health : J of
ACH, 45(3), 134–40. doi:10.1080/07448481.1996.9936873
Hallal, P. C., Andersen, L. B., Bull, F. C., Guthold, R., Haskell,
1080W. & Ekelund, U., for the Lancet Physical Activity Series
Working Group. (2012). Global physical activity levels:
Surveillance progress, pitfalls, and prospects. The Lancet,
380(9838), 247–257.
Hammond, D., Daniel, S., & White, C. M. (2013). The effect
1085of cigarette branding and plain packaging on female youth
in the United Kingdom. The Journal of Adolescent Health:
Official Publication of the Society for Adolescent Medicine,
52, 151–157. doi:10.1016/j.jadohealth.2012.06.003
Hammond, D., Fong, G. T., McNeill, A., Borland, R.,
1090Cummings, K. M., & Hammond, D. (2006). Effectiveness
of cigarette warning labels in informing smokers about
the risks of smoking: Findings from the International
Tobacco Control (ITC) Four Country Survey. Tobacco
Control, 15, 19–25. doi:10.1136/tc.2005.012294
1095Hammond, D., & Parkinson, C. (2009). The impact of
cigarette package design on perceptions of risk. Journal of
Public Health, 31(3), 345–353. doi:10.1093/pubmed/fdp066
Hancock, L., Abhold, J., Gascoigne, J., & Altekruse, M. (2002,
August). Applying social norms marketing to tobacco
1100cessation and prevention: Lessons learned from three cam-
paigns (Working Paper No. 6). The Report on Social Norms.
Available from http://socialnorms.org/wp-content/uploads/
2015/05/WP6.pdf
Hancock, L., & Henry, N. (2003). Perceptions, norms and
1105tobacco use in college residence hall freshmen: Evaluation
of a social norms marketing intervention. San Francisco,
CA: Jossey-Bass.
Hansen, P. G., & Jespersen, A. M. (2013). Nudge and the
manipulation of choice a framework for the responsible
1110use of the nudge approach to behaviour change in public
policy. European Journal of Risk Regulation, 1, 3–28.
Hedlin, S., & Sunstein, C. R. (2016). Does active choosing
promote green energy use: Experimental evidence. Ecology,
43, 107.
1115Holden, S. S., Zlatevska, N., & Dubelaar, C. (2016). Whether
smaller plates reduce consumption depends on who’s
serving and who’s looking: A meta-analysis. Journal of the
Association for Consumer Research, 1(1), 134–146.
Hollands, G. J., Shemilt, I., Marteau, T. M., Jebb, S. A., Kelly,
1120M. P., Nakamura, R., … Ogilvie, D. (2013). Altering micro-
environments to change population health behaviour:
Towards an evidence base for choice architecture
BASIC AND APPLIED SOCIAL PSYCHOLOGY 11
interventions. BMC Public Health, 13(1), 1218. doi:10.1186/
1471-2458-13-1218
1125 Hollands, G. J., Shemilt, I., Marteau, T. M., Jebb, S. A., Lewis,
H. B., Wei, Y., … Ogilvie, D. (2015). Portion, package or
tableware size for changing selection and consumption of
food, alcohol and tobacco. In G. J. Hollands (Ed.),
Cochrane database of systematic reviews. UK: Wiley & Sons.
Q6
1130 Holmes, A. S., Serrano, E. L., Machin, J. E., Duetsch, T., &
Davis, G. C. (2013). Effect of different children’s menu
labeling designs on family purchases. Appetite, 62, 198–202.
House of Lords. (2011). Behaviour change. London, UK:
Author. Retrieved from http://www.parliament.uk/
1135 business/lords
Hull, C. L. (1943). Principles of behaviour: An introduction to
behaviour theory. New York, NY: Appleton-Century Crofts.
John, P., Moseley, A., Wales, C., Smith, G., Richardson, L.,
Stoker, G., & Cotterill, S. (2013). Nudge, nudge, think, think:
1140 Experimenting with ways to change civic behaviour.
London, UK: Bloomsbury Academic.
Jung, J. Y., & Mellers, B. A. M. (2016). American attitudes
toward nudges. Judgement and Decision Making, 11(1),
62–74.
1145 Kahneman, D. (2003). Maps of bounded rationality: A
perspective on intuitive judgement and choice. In T.
Frangsmyr (Ed.), Lex prix Nobel: The Nobel prizes 2002
(pp. 449–489). Stockholm, Sweden: Almquist & Wiksell
International.
1150 Kahneman, D. (2011). Thinking fast and slow. London, UK:
Allen Lane.
Kahneman, D., & Frederick, S. (2002). Representativeness
revisited: Attribute substitution in intuitive judgment.
In T. Gilovich, D. Griffin, & D. Kahneman (Eds.), Heuris-
1155 tics and biases: The psychology of intuitive judgment
(pp. 49–81). New York, NY: Cambridge University Press.
Kahneman, D., & Frederick, S. (2005). A model of heuristic
judgment. In K. J. Holyoack & R. G. Morrison (Eds.),
The Cambridge handbook of thinking and reasoning
1160 (pp. 267–293). New York, NY: Cambridge University Press.
Kahneman, D., & Tversky, A. (1982). On the study of statisti-
cal intuitions. In D. Kahneman, P. Slovic, & A. Tversky
(Eds.), Judgement under uncertainty: Heuristics and biases
(pp. 493–508). New York, NY: Cambridge University Press.
1165 doi:10.1016/0010-0277(82)90022-1
Kerr, J., Eves, F. F., & Carroll, D. (2001a). Getting more
people on the stairs: the impact of a new message format.
Journal of Health Psychology, 6(5), 495–500. doi:10.1177/
135910530100600503
1170 Kerr, J., Eves, F. F., & Carroll, D. (2001b). Six-month
observational study of prompted stair climbing. Preventive
Medicine, 33(5), 422–427. doi:10.1006/pmed.2001.0908
Kiszko, K. M., Martinez, O. D., Abrams, C., & Elbel, B. (2014).
The influence of calorie labeling on food orders and con-
1175 sumption: a review of the literature. Journal of Community
Health, 39(6), 1248–1269. doi:10.1007/s10900-014-9876-0
Klaczynski, P. A. (2001). Analytic and heuristic processing
influences on adolescent reasoning and decision making.
Child Development, 72(3), 844–861.
1180 Kosters, M., & Van der Heijden, J. (2015). From mechanism
to virtue: Evaluating Nudge theory. Evaluation, 21,
276–291. doi:10.1177/1356389015590218
Kruglanski, A. W. (2013). Only one? The default
interventionist perspective as a unimodel—Commentary
1185on Evans & Stanovich (2013). Perspectives on Psychological
Science, 8, 242–247. doi:10.1177/1745691613483477
Lee, K. K., Perry, A. S., Wolf, S. A., Agarwal, R., Rosenblum,
R., Fischer, S., … Silver, L. D. (2012). Promoting routine
stair use evaluating the impact of a stair prompt across
1190buildings. American Journal of Preventive Medicine, 42,
136–141. doi:10.1016/j.amepre.2011.10.005
Lewis, A., & Eves, F. (2012). Prompt before the choice is
made: Effects of a stair-climbing intervention in university
buildings. British Journal of Health Psychology, 17, 631–643.
1195doi:10.1111/j.2044-8287.2011.02060.x
Lewis, M. A., & Neighbors, C. (2006). Social norms
approaches using descriptive drinking norms education:
A review of the research on personalized normative feed-
back. Journal of American College Health, 54, 213–218.
1200doi:10.3200/JACH.54.4.213-218
Linkenbach, J., Perkins, H. W., & DeJong, W. (2003). Parent’s
perceptions of parenting norms: using the social norms
approach to reinforce effective parenting. San Francisco,
CA: Jossey-Bass.
1205Liu, P. J., Wisdom, J., Roberto, C. A., Liu, L. J., & Ubel, P. A.
(2014). Using behavioral economics to design more
effective food policies to address obesity. Applied
Economic Perspectives and Policy, 36, 6–24. doi:10.1093/
aepp/ppt027
1210Loewenstein, G., Asch, D. A., Friedman, J. Y., Melichar, L. A.,
Volpp, K. G., Lowenstein, G., … Volpp, K. G. (2012). Can
behavioural economics make us healthier? BMJ, 344, e3482.
doi:10.1136/bmj.e3482
Loewenstein, G., Bryce, C., Hagmann, D., & Rajpal, S. (2015).
1215Warning: You are about to be nudged. Behavioral Science
& Policy, 1, 35–42.
Magnusson, R. S. (2009). Rethinking global health challenges:
Towards a “global compact” for reducing the burden of
chronic disease. Public Health, 123, 265–274.
1220Marshall, A. L., Bauman, A. E., Patch, C., Wilson, J., & Chen,
J. (2002). Can motivational signs prompt increases in inci-
dental physical activity in an Australian health-care facility?
Health Education Research, 17, 743–749. doi:10.1093/her/
17.6.743
1225Marteau, T. M., Ogilvie, D., Ronald, M., Suhrcke, M., & Kelly,
M. P. (2011). Judging nudging: Can nudge improve
population health? BMJ, 342, d228.
Maynard, O. M., Munafò, M. R., & Leonards, U. (2013).
Visual attention to health warnings on plain tobacco
1230packaging in adolescent smokers and non-smokers. Addic-
tion, 108, 413–419. doi:10.1111/j.1360-0443.2012.04028.x
Mazzocchi, M., Cagnone, S., Bech-Larsen, T., Niedźwiedzka,
B., Saba, A., Shankar, B., … Traill, W. B. (2015). What is
the public appetite for healthy eating policies? Evidence
1235from a cross-European survey. Health Economics, Policy
and Law, 10, 267–292. doi:10.1017/S1744133114000346
Mills, C. (2013). Why nudges matter: A reply to Goodwin.
Politics, 33(1), 28–36. doi:10.1111/j.1467-9256.2012.01450.x
Mols, F., Haslam, S. A., Jetten, J., & Steffens, N. K. (2015).
1240Why a nudge is not enough: A social identity critique of
governance by stealth. European Journal of Political
Research, 54(1), 195. doi:10.1111/1475-6765.12073
Moodie, C., & Mackintosh, A. M. (2013). Young adult women
smokers’ response to using plain cigarette packaging: a
1245naturalistic approach. BMJ Open, 3(3), e002402.
doi:10.1136/bmjopen-2012-002402
12 Y. LIN ET AL.
Moodie, C., Mackintosh, A. M., Hastings, G., & Ford, A.
(2011). Young adult smokers’ perceptions of plain packag-
ing: A pilot naturalistic study. Tobacco Control, 20(5),
1250 367–373. doi:10.1136/tc.2011.042911
Moodie, C., Stead, M., Bauld, L., McNeill, A., Angus, K.,
Hinds, K., … O’Mara-Eves, A. (2012). Plain tobacco
packaging: A systematic review. London. Retrieved from
http://dspace.stir.ac.uk/bitstream/1893/24420/1/Moodie_
1255 et_al_2012_PHRC_PlainPacks.pdf
Munafò, M. R., Roberts, N., Bauld, L., & Leonards, U. (2011).
Plain packaging increases visual attention to health warn-
ings on cigarette packs in non-smokers and weekly smokers
but not daily smokers. Addiction, 106, 1505–1510.
1260 doi:10.1111/j.1360-0443.2011.03430.x
Nishida, Y., Akaoka, I., & Nishizawa, T. (1975). Effect of sex
hormones on uric acid metabolism in rats. Experientia,
31, 1134–1135. Retrieved from http://www.ncbi.nlm.nih.
gov/pubmed/1204720
1265 Nocon, M., Müller-Riemenschneider, F., Nitzschke, K., &
Willich, S. N. (2010). Review Article: Increasing physical
activity with point-of-choice prompts–a systematic review.
Scandinavian Journal of Public Health, 38, 633–638.
doi:10.1177/1403494810375865
1270 Nomura, T., Yoshimoto, Y., Akezaki, Y., & Sato, A. (2009).
Changing behavioral patterns to promote physical activity
with motivational signs. Environmental Health and Preven-
tive Medicine, 14, 20–25. doi:10.1007/s12199-008-0053-x
Oppenheimer, D. M. (2008). The secret life of fluency. Trends
1275 in Cognitive Sciences, 12(6), 237–241. doi:10.1016/j.tics.
2008.02.014
Osman, M. (2004). An evaluation of dual-process theories of
reasoning. Psychonomic Bulletin & Review, 11, 988–1010.
doi:10.3758/BF03196730
1280 Osman, M. (2007). Can tutoring improve performance on a
reasoning task under deadline conditions? Memory &
Cognition, 35, 342–351.
Osman, M. (2013). A case study: Dual-process theories of
higher cognition-commentary on Evans & Stanovich
1285 (2013). Prospect Psychological Science, 8, 248–252.
doi:10.1177/1745691613483475
Osman, M. (2014). Future-minded: The psychology of agency
and control. New York, NY: Palgrave Macmillan.
Osman, M. (2016). Nudge: How far have we come?
1290 Œconomia. History, Methodology, Philosophy, 6–4, 557–
570. Retrieved from https://oeconomia.revues.org/2490
Osman, M., & Stavy, R. (2006). Development of intuitive
rules: Evaluating the application of the dual-system frame-
work to understanding children’s intuitive reasoning.
1295 Psychonomic Bulletin & Review, 13, 935–953. doi:10.3758/
bf03213907
Pearson, M., & West, P. (2003). Drifting smoke rings: Social
network analysis and Markov processes in a longitudinal
study of friendship groups and risk-taking. Connections,
1300 25(2), 59–76.
Perkins, H. W., Meilman, P. W., Leichliter, J. S., Cashin, J. R.,
& Presley, C. A. (1999). Misperceptions of the norms for
the frequency of alcohol and other drug use on college
campuses. Journal of American College Health, 47(6),
1305 253–258. doi:10.1080/07448489909595656
Puig-Ribera, A., & Eves, F. F. (2010). Promoting stair
climbing in Barcelona: Similarities and differences with
interventions in English-speaking populations. European
Journal of Public Health, 20(1), 100–102. doi:10.1093/
1310eurpub/ckp059
Rayner, G., & Lang, T. (2011). Is nudge an effective public
health strategy to tackle obesity? No. BMJ, 342, d2177–
d2177. doi:10.1136/bmj.d2177
Read, S. J., Vanman, E. J., & Miller, L. C. (1997). Connection-
1315ism, parallel constraint satisfaction processes, and gestalt
principles: (Re)introducing cognitive dynamics to social
psychology. Personality and Social Psychology Review, 1,
26–53.
Reisch, L. A., & Sunstein, C. R. (2016). Do Europeans like
1320nudges? Judgement and Decision Making, 11(4), 310–325.
Reisch, L. A., Sunstein, C. R., & Gwozdz, W. (2016). Better
than a whip? European attitudes toward health nudges.
Food Policy, 69, 1–10. doi:10.1016/j.foodpol.2017.01.007
Russell, W. D., Dzewaltowski, D. A., & Ryan, G. J. (1999). the
1325effectiveness of a point-of-decision prompt in deterring
sedentary behavior. American Journal of Health Promotion,
13(5), 257–259. doi:10.4278/0890-1171-13.5.257
Sacks, G., Rayner, M., & Swinburn, B. (2009). Impact of front-
of-pack “traffic-light” nutrition labelling on consumer food
1330purchases in the UK. Health Promotion International, 24,
344–352. doi:10.1093/heapro/dap032
Saghai, Y. (2013). Salvaging the concept of nudge. Journal of
Medical Ethics, 39(8), 487–493. doi:10.1136/medethics-
2012-100727
1335Schmiedek, F., Oberauer, K., Wilhelm, O., Süß, H. M., &
Wittmann, W. W. (2007). Individual differences in compo-
nents of reaction time distributions and their relations to
working memory and intelligence. Journal of Experimental
Psychology: General, 136, 414.
1340Schwarz, N. (1998). Warmer and more social: Recent develop-
ments in cognitive social psychology. Annual Review of
Sociology, 24, 239–264. doi:10.1146/annurev.soc.24.1.239
Selinger, E., & Whyte, K. P. (2012). Nudging cannot solve
complex policy problems. European Journal of Risk
1345Regulation, 3(1), 26–31.
Simon, D., & Holyoak, K. J. (2002). Structural dynamics of
cognition: From consistency theories to constraint satisfac-
tion. Personality and Social Psychology Review, 6, 283–294.
Simon, D., Snow, C. J., & Read, S. J. (2004). The redux of
1350cognitive consistency theories: evidence judgments by
constraint satisfaction. Journal of Personality and Social
Psychology, 86, 814–837.
Sinclair, S. E., Cooper, M., & Mansfield, E. D. (2014). the
influence of menu labeling on calories selected or
1355consumed: A systematic review and meta-analysis. Journal
of the Academy of Nutrition and Dietetics, 114, 1375–1388.
doi:10.1016/j.jand.2014.05.014
Sloman, S. A. (1996). The empirical case for two systems of
reasoning. Psychological Bulletin, 119, 3–22. doi:10.1037/
13600033-2909.119.1.3
Soler, R. E., Leeks, K. D., Buchanan, L. R., Brownson, R. C.,
Heath, G. W., Hopkins, D. H. & the Task Force on
Community Preventive Services. (2010). Point-of-decision
prompts to increase stair use a systematic review update.
1365American Journal of Preventive Medicine, 38(2, Suppl),
S292–S300. doi:10.1016/j.amepre.2009.10.028
Sonnenberg, L., Gelsomin, E., Levy, D. E., Riis, J.,
Barraclough, S., & Thorndike, A. N. (2013). A traffic light
food labeling intervention increases consumer awareness
1370of health and healthy choices at the point-of-purchase.
BASIC AND APPLIED SOCIAL PSYCHOLOGY 13
Preventive Medicine, 57, 253–257. doi:10.1016/j.ypmed.
2013.07.001
Stanovich, K. E. (1999). Who is rational?: Studies of individual
differences in reasoning. Mahwah, NJ: Erlbaum.
1375 Stanovich, K. E., & West, R. F. (2000). Advancing the ration-
ality debate. Behavioral and Brain Sciences, 23(5), 701–717.
doi:10.1017/S0140525x00623439
Sun, R., Slusarz, P., & Terry, C. (2005). The interaction of the
explicit and the implicit in skill learning: a dual-process
1380 approach. Psychological Review, 112(1), 159.
Sunstein, C. R. (2014). Why nudge? The politics of libertarian
paternalism. New Haven, CT: Yale University Press.
Sunstein, C. R. (2015). Nudges, agency, and abstraction: A
reply to critics. Review of Philosophy and Psychology, 6(3),
1385 511–529. doi:10.1007/s13164-015-0266-z
Sunstein, C. R. (2016a). Do people like nudges?
Administrative Law Review, Forthcoming. doi:10.2139/
ssrn.2604084
Sunstein, C. R. (Ed.). (2016b). The ethics of influence:
1390 Government in the age of behavioral science. Cambridge,
UK: Cambridge University Press.
Sunstein, C. R. (2016c). People prefer System 2 nudges (kind
of). Duke Law Journal, 66, 121–168.
Sunstein, C. R., Reisch, L. A., & Rauber, J. (2017). A
1395 worldwide consensus on nudging? Not quite, but almost.
Regulation and Governance. Advance online publication.
doi:10.2139/ssrn.2955693
Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving
decisions about health, wealth, and happiness. New Haven,
1400 CT: Yale University Press.
Thombs, D. L., Dotterer, S., Olds, R. S., Sharp, K. E., & Raub,
C. G. (2004). A close look at why one social norms
campaign did not reduce student drinking. Journal of
American College Health, 53(2), 61–68. doi:10.3200/
1405 JACH.53.2.61-70
Thompson, V. A. (2009). Dual process theories: A metacogni-
tive perspective. In J. S. B. T. Evans & K. Frankish (Eds.), In
two minds: Dual processes and beyond (pp. 171–196).
Oxford, UK: Oxford University Press.
1410 Wakefield, M., Coomber, K., Zacher, M., Durkin, S., Brennan,
E., & Scollo, M. (2015). Australian adult smokers’ responses
to plain packaging with larger graphic health warnings
1 year after implementation: results from a national
cross-sectional tracking survey. Tobacco Control,
1415 24(Suppl. 2), ii17–ii25. doi:10.1136/tobaccocontrol-2014-
052050
Wansink, B. (2004). Environmental factors that increase the
food intake and consumption volume of unknowing
consumers. Annual Review of Nutrition, 24, 455–479.
1420doi:10.1146/annurev.nutr.24.012003.132140
Wansink, B., & van Ittersum, K. (2005). Shape of glass and
amount of alcohol poured: comparative study of effect of
practice and concentration. BMJ, 331, 1512–1514.
doi:10.1136/bmj.331.7531.1512
1425Wansink, B., & van Ittersum, K. (2013). Portion size me:
Plate-size induced consumption norms and win–win
solutions for reducing food intake and waste. Journal of
Experimental Psychology: Applied, 19(4), 320–332.
doi:10.1037/a0035053
1430Webb, K. L., Solomon, L. S., Sanders, J., Akiyama, C., &
Crawford, P. B. (2011). Menu labeling responsive to con-
sumer concerns and shows promise for changing patron
purchases. Journal of Hunger & Environmental Nutrition,
6(2), 166–178. doi:10.1080/19320248.2011.576210
1435Webb, O. J., & Eves, F. F. (2007). Promoting stair climbing:
Intervention effects generalize to a subsequent stair ascent.
American Journal of Health Promotion, 22(2), 114–119.
doi:10.4278/0890-1171-22.2.114
Weber, E. U. (2017). Breaking cognitive barriers to a sustain-
1440able future. Nature Human Behaviour, 1, 13. doi:10.1038/
s41562-016-0013
World Health Organization. (2015). Health warnings.
Retrieved from http://www.euro.who.int/en/health-topics/
disease-prevention/tobacco/policy/who-framework-conven
1445tion-on-tobacco-control-who-fctc/key-areas-of-tobacco-
control-policy/health-warnings.
World Health Organization. (2016). WHO | Tobacco.
Retrieved from http://www.who.int/mediacentre/fact-
sheets/fs339/en/
1450World Health Organization. (2017a). NCDs | Noncommunic-
able diseases and their risk factors. Retrieved from http://
www.who.int/ncds/introduction/en/
World Health Organization. (2017b). Plain packaging of
tobacco products: evidence, design and implementation.
1455Retrieved from http://apps.who.int/iris/bitstream/10665/
207478/1/9789241565226_eng.pdf?ua=1
Wood, J. V. (1989). Theory and research concerning social
comparisons of personal attributes. Psychological Bulletin,
106(2), 231–248.
1460Wood, W., Quinn, J. M., & Kashy, D. A. (2002). Habits in
everyday life: Thought, emotion, and action. Journal of
Personality and Social Psychology, 83(6), 1281–1297.
doi:10.1037//0022-3514.83.6.1281
Young, L. R., & Nestle, M. (2002). The contribution of
1465expanding portion sizes to the US obesity epidemic.
American Journal of Public Health, 92, 246–249.
14 Y. LIN ET AL.