Content uploaded by Lukas Paul Fesenfeld
Author content
All content in this area was uploaded by Lukas Paul Fesenfeld on Apr 26, 2023
Content may be subject to copyright.
Articles
https://doi.org/10.1038/s43016-022-00460-8
1Institute of Political Science and Oeschger Centre for Climate Change Research, University of Bern, Bern, Switzerland. 2Institute of Science,
Technology and Policy, ETH Zurich, Zürich, Switzerland. 3Geschwister Scholl Institute of Political Science, LMU Munich, Munich, Germany.
✉e-mail: lukas.rudolph@gsi.uni-muenchen.de
Food waste is a major hurdle in trying to reach the Sustainable
Development Goals (SDGs) and Paris climate targets1–4. Wasted
and lost food accounts for at least 8% of global anthropogenic
greenhouse gas emissions, 20% of freshwater consumption and 30% of
global agricultural land use5,6. Yet, according to the recent Food Waste
Index Report 2021 by the United Nations Environment Programme
(UNEP) ‘the opportunities provided by food waste reduction have
remained largely untapped and under-exploited’7. Although low-
to middle-income countries generate substantial amounts of food
waste, high-income countries contribute disproportionally to such
waste, especially of products with large environmental footprints
such as animal products5,7–10. Calls for a change in practices are mani-
fold, on the global (for example, the UN’s SDG 12) and local level (for
example, civil society initiatives). Yet, current efforts are not on track
to reach targets such as SDG 12.3, which calls for a halving of food
waste by 2030. Around one-third of all food products are wasted,
and waste is predicted to increase in the upcoming years as low- and
middle-income countries become richer9,10. According to the UNEP
and the Food and Agriculture Organization, ambitious public and
private food waste reduction measures are urgently needed7,11, but are
only implemented in rare cases (for example, the French food waste
law, Loi 2016-138)9,10. Potential measures include binding reduction
targets for the food industry, obligations for supermarkets not to
destroy unsold food products, and food waste fines10,12.
Ambitious regulation against food waste is widely presumed to
be politically challenging because such regulation may increase food
prices. In turn, price increases may result in consumer and citizen
opposition that would render stricter regulation politically unfeasi-
ble (for example, leading to protests and a loss of public support). In
democracies, citizens, who are both consumers and voters, are key
stakeholders, and public opinion is hence an important determinant
of the political feasibility of ambitious food waste regulation13. Yet,
there is a lack of robust empirical research on the mass public’s pref-
erences regarding food waste reduction regulation.
To address this research gap, we implemented a combination of
framing, conjoint and vignette experiments with a representative
random sample (N = 3,329) of voting-age Swiss citizens and con-
sumers (see Methods for details). Switzerland typifies a high-income
country responsible for large per-capita shares of global food waste
(around 300 kg per person per year are lost, compared to around
120–170 kg per person per year in low-income countries5,12). Our
study design and case selection also allowed us to mimic choice
situations that are realistic for our respondents: Switzerland has a
unique direct democratic decision-making setting in which Swiss
voters regularly and directly vote on public affairs in popular ini-
tiatives and referenda. Research has shown that survey-embedded
choice experiments indeed closely resemble real-world voting
behaviours of Swiss citizens14. Swiss voters also directly decide about
the direction of food waste regulation because they can occasionally
vote on food and agricultural initiatives. For example, in June 2021
Swiss citizens voted on two politically salient initiatives about pes-
ticide use in agriculture. In 2018, they voted on the so-called ‘Fair
Food initiative’ that would have obliged the federal government to
promote environmentally sound, animal-friendly and fairly pro-
duced food products. Although many voters supported the premises
of the ‘Fair Food initiative’, with only 38.7% votes in favour it did not
reach the necessary majority threshold, mainly because many voters
feared higher food prices15. While this example supports the wide-
spread assumption that voters are unlikely to accept ambitious food
waste regulation if this burdens them with direct costs, we argue
in line with a burgeoning environmental policy literature, however,
that appropriate policy framing, design and feedback could ensure
public support despite food price increases.
Policy framing, design and feedback
The environmental policy and public opinion literature suggests
three important factors that could affect public support for ambi-
tious food waste regulation. Building on this literature, our key
argument is that policy framing16–21 and design13,18,22–24 are essen-
tial factors through which policymakers and the private sector can
influence public support for ambitious but costly food waste regula-
tion. At the same time, we argue that potential feedback effects from
Policy framing, design and feedback can increase
public support for costly food waste regulation
Lukas Fesenfeld1,2, Lukas Rudolph 2,3 ✉ and Thomas Bernauer2
Stricter regulation of food waste reduction is widely presumed to increase food prices, which could render its implementation
politically unfeasible. Here we empirically tested whether specific policy framing, design and feedback could help ensure public
support despite potential food price increases. We used survey experiments with 3,329 citizens from a high-income country,
Switzerland. A combined framing and conjoint experiment shows that messages emphasizing national or international social
norms in favour of reducing food waste (policy framing) can increase public support for more ambitious reduction targets. Also,
most citizens support food waste regulation even if this leads to substantial increases in food prices, but only if such policies
set stringent reduction targets and are transparently monitored (policy design). Finally, a vignette experiment reveals that
voluntary industry initiatives do not crowd out individuals’ support for stricter governmental regulation, but potentially crowd
in support if industry initiatives are unambitious (policy feedback).
NATURE FOOD | www.nature.com/natfood
Articles NAture Food
voluntary food waste reduction initiatives (that is, private-sector
policies)25–34 do not crowd out but potentially crowd in public sup-
port for governmental food waste regulation.
These three factors have, so far, been studied mainly in isola-
tion and hence also in different survey and/or country populations.
Looking at all three factors simultaneously in one study speaks to
real-world policymaking, where policy framing, design and feed-
back usually occur together. We thus contribute to the literature
by examining in a unified research design whether policy fram-
ing, design, and feedback can help increase public support for food
waste reduction measures as public opinion research on related
environmental policy fields suggests—and whether and how these
three factors interrelate13,17–19,26,35.
First, based on the existing social norms and environmental
regulation literature36–40, we investigated the public support effects
of national and international social norm frames that emphasize
the importance of strong food waste reduction. We expected that
both would increase public support for ambitious but costly food
waste regulation. We also expected that signalling national norms
of reducing food waste might lead voters to perceive norms to be
more personally relevant and proximate, and hence trigger stron-
ger public support effects than the more distant international norm
frames. Existing literature36,37 on the effects of social norms on envi-
ronmental policy attitudes and behaviours tends to support this
argument, but we lack empirical evidence in the context of food
waste regulation.
Second, we expected that public support for food waste policies
is a function of policy design, especially a policy’s perceived binding
nature, stringency, scope, public control and consumer costs13,18,22–24.
In general, individuals tend to prefer environmental policies that
they perceive as more ambitious and thus presumably more effec-
tive at solving the underlying problem41, while opposing policies
that they perceive as costly for themselves13. Citizens thus experi-
ence a potential trade-off between policy ambitiousness and policy
costs. Building on the environmental policy design literature13,18,23,42,
we argue that negative effects of policy-induced costs on public
support can be compensated for by the positive support effects of
those policy design attributes that increase the perceived policy
ambitiousness. Consequently, we expected that food waste regula-
tion leading to higher food prices (negative support effect) is most
likely to receive majority support if such regulation is also perceived
as ambitious (positive support effects), that is, is legally binding
(instead of voluntary), sets stringent reduction targets for the food
industry, applies to most firms in the sector and involves strong
transparency and public control mechanisms35. In addition, exist-
ing research has fallen short of investigating how policy framing
and policy design interact. We thus argue that social norm frames
can increase the salience of stricter food waste reduction targets and
increase support for more stringent targets and design elements
that enlarge the perceived ambitiousness of policies to achieve food
waste reductions.
Third, voluntary industry food waste reduction initiatives are
gaining ground43 and are more widespread than public forms
of food waste regulation in contexts such as the United States
or Switzerland. It is of particular interest, therefore, to explore
whether such voluntary industry initiatives crowd out or crowd in
public demand for ambitious state-led food waste reduction regu-
lation. While traditionally policy feedback research has focused on
Policy framing experiment (n = 1,231)
•Control group (n = 437)
no framing information
•National social norm frame (n = 389)
emphasizing national efforts and goal to
reduce food waste by 50% by 2030
•International social norm frame (n = 405)
emphasizing international efforts and goal
to reduce food waste by 50% by 2030
Policy design (conjoint) experiment (n = 1,231)
Design attribute Design value
1. Binding nature •Required by law
•Voluntary measures taken by food companies
2. Policy target
(by 2030)
•No increase of food waste
•10% decrease
•25% decrease
•50% decrease
3. Policy scope
(number of firms
covered)
•All firms
•Half of firms
•A few firms
4. Monitoring
mechanism
•Voluntary public report
•Obligatory public report without governmental
control
•Obligatory public report with governmental
control
5. Policy-induced
costs
•No food price increases
•2% increases
•5% increases
•10% increases
Support for
food waste
regulation
Fig. 1 | Combined policy framing and policy design (conjoint) experiment (n=1,231). From the full sample, we randomly selected 1,231 respondents to
participate in the combined framing and conjoint experiments (Methods). In the policy-framing experiment, the respondents were randomly confronted
with either a national social norm frame (n = 389), an international social norm frame (n = 405) or a control condition (n = 437). The national norm frame
provided information about Switzerland’s national efforts and goal to reduce food waste by 50% by 2030, while the international social norm frame
emphasized international efforts to reduce food waste by 50% by 2030. After that, respondents rated different policy proposals to reduce food waste on
a Likert scale from ‘strongly opposed’ (1) to ‘strongly in favour’ (7). They also had to choose between policy proposals. The policy proposals consisted of
five different design attributes: (1) the binding nature of the proposal; (2) the policy target (by 2030); (3) the policy scope (that is, the number of firms
covered by the policy); (4) the monitoring mechanism; and (5) an estimation of the policy-induced costs. For each of these five attributes, we randomly
varied the specific design values; for instance, the proposal could entail a more or less stringent target varying from no increase of today’s food waste to
a 50% decrease by 2030. Respondents chose and rated for four consecutive times two pairs of policy proposals. In general, we expected design features
that increase the perceived ambitiousness of reducing food waste (see attributes 1–4) to increase public support, while higher policy-induced costs would
decrease support (see attribute 5). We also expected that social norm frames, particularly the national norm frame, would increase public support and
positively interact with more stringent food waste reduction targets.
NATURE FOOD | www.nature.com/natfood
Articles
NAture Food
how earlier public policies influence latter public policies44–47, more
recently research has also started to investigate how earlier vol-
untary industry initiatives (that is, private-sector policies) impact
subsequent public policymaking26,27,34. An argument in this line of
environmental policy literature25–34,48 is that firms use voluntary
initiatives as a political strategy to prevent stricter governmental
regulations, and that even unambitious voluntary measures can
crowd out public demand for governmental interventions. In con-
trast to the existing literature, we propose that voluntary indus-
try food waste reduction initiatives do not crowd out but rather
crowd in public support for governmental food waste regulation.
In particular, we expected that broad (visible) voluntary initia-
tives may increase the salience of the food waste problem, lead
to more social norm pressure and thereby crowd in demand for
stricter governmental regulation of firms and personal food waste
reduction actions26,28,49. Moreover, shallow (non-stringent) indus-
try initiatives to reduce food waste are likely to be perceived as
non-ambitious and ineffective, and thus citizens would in turn
demand stiffer governmental regulation35.
Overall, we expected that policy framing, design and feedback
would not only individually affect public support for food waste
regulation but also interact with each other and jointly influence
public opinion. Our country context exemplifies a real-world case
where policy framing, design and feedback are simultaneously
present and likely to interact. For example, the Swiss food indus-
try engages in voluntary private initiatives to reduce food waste and
communicates this to the public43. At the same time, citizens experi-
ence social norm framing, for example, because the United Nations’
SDG to reduce food waste by 50% is publicly communicated and
discussed by media, politicians, scientists, industry and civil soci-
ety actors50. Moreover, public debates, in parliament and around
direct democratic initiatives, explicitly address how food waste
policies should be designed. As outlined above, we thus expected
that both social norm frames and voluntary industry initiatives alter
the salience of the food waste issue in general and focus attention
on specific design elements. In turn, the increased salience through
framing and feedback is likely to increase the support effects of spe-
cific aspects of the policy design, for example, the stringency of the
food waste regulation.
To test this expectation, we combined framing, conjoint and
vignette experiments in an integrated study design using the same
survey sample (that is, 40% of our sample was exposed to the con-
joint experiment, within which a third each was exposed to three
conditions of the policy-framing experiment; the other 60% of our
sample was exposed to the vignette experiment on policy feed-
back). This allowed us to empirically assess potential interactions
between policy framing and design, that is, whether social norm
frames increase support for more strictly designed food waste regu-
lations, by combining framing and conjoint experiments. Moreover,
by varying the design of voluntary food waste initiatives in the
vignette experiment, we could also test how differently designed
private-sector initiatives feed back into public opinion formation
and affect support for governmental food waste regulation.
Results and discussion
Policy design and framing can increase support for ambitious
food waste regulation. Turning to the experimental evidence, our
study supports the argument that policy design is a key factor in
trying to increase public support for costly food waste policies, and
that policy framing can increase public support for measures with
ambitious reduction targets. First, Fig. 1 outlines the set-up of the
combined policy framing and policy design (conjoint) experiment
(see Methods for further details).
Second, in Fig. 2, we present the average marginal component
effects of the different design attributes on citizens’ support for a
policy proposal to reduce food waste. In contrast to our expectations,
the results show that a proposal that entails purely voluntary mea-
sures by the food industry slightly increases support by around 0.1
scale points on a 7-point Likert scale compared to a proposal where
measures are required by law. This effect is, however, very weak and
compared to the other design elements does not substantially affect
overall public support. For example, proposals with a reduction goal
of 10% increase support by around 0.25 scale points compared with
the baseline option of no increase in food waste until 2030. Policy
proposals that include an ambitious reduction target of decreasing
food waste by 2030 by 25–50% even substantially increase support
by around 0.5 scale points on a 7-point Likert scale. Interestingly,
Voluntary measures
Required by law
No increase of food waste
10% decrease
25% decrease
50% decrease
All firms participate
Half of firms participate
Few firms participate
Voluntary report
Obligatory report
Obligatory with monitoring
No price increase
2% increase
5% increase
10% increase
−0.7
−0.6
−0.5
−0.4
−0.3
−0.2
−0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Average marginal component effects
Fig. 2 | Average marginal component effects of policy design attribute
values on citizens’ support for differently designed policy proposals
to reduce food waste. Support was measured on a 7-point Likert scale.
The dashed horizontal lines separate policy attributes; the dots on the
solid vertical line at 0 on the x axis denote the baseline comparison of
each policy attribute. Error bars represent 95% confidence intervals.
See Supplementary Table 1 for the detailed regression output and
Supplementary Fig. 4 for the respective marginal mean plot.
Voluntary measures
Required by law
No increase of food waste
10% decrease
25% decrease
50% decrease
All firms participate
Half of firms participate
Few firms participate
Voluntary report
Obligatory report
Obligatory with monitoring
No price increase
2% increase
5% increase
10% increase
−0.7
−0.6
−0.5
−0.4
−0.3
−0.2
−0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
No information International/national social norm
Average marginal component effects
Fig. 3 | Average marginal component effects of policy design attribute
values on citizens’ support for differently designed policy proposals to
reduce food waste by combined social norm frames. The control group
effects are marked in blue while the combined norm treatment effects are
marked in red. Support was measured on a 7-point Likert scale. The dashed
horizontal lines separate policy attributes; the dots on the solid vertical line
at 0 on the x axis denote the baseline comparison of each policy attribute.
Error bars represent 95% confidence intervals. See Supplementary Table 2
for detailed regression outputs and Supplementary Fig. 5 for the respective
marginal mean plot. See Supplementary Table 4 and Supplementary Fig. 6
for average marginal component effects differentiated by the national and
international norm frame.
NATURE FOOD | www.nature.com/natfood
Articles NAture Food
and in contrast to our expectation, citizens do not prefer the most
stringent 50% reduction target significantly more compared with
the less stringent 25% reduction target (P value from a Wald test
of equality of coefficients is 0.82 for the rating and 0.06 for the
choice task).
The insignificant difference between the two targets suggests a
potential ceiling effect of higher policy stringency on policy sup-
port. In other words, citizens value more stringent policy targets
but a stringent target itself is insufficient to ensure majority support
and other design elements are also needed. Other attributes, such
as the scope of the policy (that is, the number of firms covered) and
the monitoring mechanism, also significantly affect public support.
In line with our expectations, citizens prefer policy proposals that
cover all firms in the food sector to proposals that have smaller cov-
erage. For example, a proposal that covers only half of firms receives
around 0.22 scale points less support than a proposal that covers
all firms in the sector. A proposal that covers only a few firms in
the sector receives even 0.51 scale point less support than a pro-
posal with the widest coverage. Similarly, in line with our expecta-
tions, a mandatory clause to publish reports (via the internet) about
firms’ progress to reduce food waste increases support for a policy
proposal by around 0.14 scale points, and an additional clause that
public authorities control such reports increases support by around
0.21 scale points compared with voluntary reports.
Information about the policy-induced costs, however, may
reduce support for food waste reduction policies. Yet, our results do
not fully confirm our expectations. In contrast, the results indicate
that support for policy proposals that only lead to a slight food price
increase of 2–5% do not significantly reduce policy support. Only
a strong increase of food prices by 10% significantly reduces policy
support by about 0.35 scale points. This finding suggests that there
is no linear negative effect of higher prices on policy support but
that support for food waste regulation only decreases from a certain
threshold of price increases (that is, public support is price inelas-
tic for smaller food price increases but becomes elastic if prices
increase further51).
To put these effects into perspective, our largest effects of half
a scale point move respondents 7% along the scale from strongly
opposed to strongly in favour of food waste regulation, or 30% of
the standard deviation of 1.67 of the rating scale that 68% of all
observed variation falls into. This is a substantial effect, which also
shows up in the relatively similar results for the conjoint choice
tasks presented in Supplementary Fig. 3 (where a 50% reduction
target is 25 percentage points more likely to be chosen compared
with the baseline of no food waste increase). Notably, as can be
seen from Supplementary Fig. 4 displaying the respective marginal
means, favourable attribute levels for reduction target and participa-
tion push preferences to around 4.5–4.6 on the 7-point rating scale,
which shifts an average respondent from indifference to approval of
food waste regulation in absolute terms.
Next, we investigated to what extent social norm frames increase
public support and interact with the different policy design ele-
ments. Figure 3 shows the average marginal component effects of
the different design attributes on respondents’ support for a policy
Initiative design attribute Initiative design value
1. Voluntary initiative target
(by 2030)
•No increase of food waste
•10% decrease
•25% decrease
•50% decrease
2. Voluntary initiative scope
(number of firms covered)
•All firms
•Half of firms
•A few firms
3. Monitoring mechanism •Voluntary public report
•Obligatory public report with
governmental control
4. Initiative-induced costs •No food price increases
•2% increases
•5% increases
•10% increases
Policy feedback (vignette) experiment (n = 1,897)
•Control group (n = 614)
no information about voluntary industry initiatives
• Vignette treatment groups (n = 1,283)
randomly varying the design options of voluntary initiatives
• Perceived
ambition of
voluntary
initiative
• Support for
food waste
regulation
Fig. 4 | Policy feedback (vignette) experiment (n=1,897). From the full sample, we randomly selected a subsample of 1,897 respondents (see Methods
for details). While 614 respondents were randomly allocated to a placebo control group without any information about industry initiatives to reduce food
waste, 1,283 respondents were confronted with randomly varying information about a voluntary food waste reduction initiative. Each of the respondents
was assigned to a single information treatment and afterwards answered standard survey questions on perceived policy ambition and support for
government food waste regulation. The vignette treatments randomly varied information about the scope of the voluntary industry initiative (that is,
the number of firms participating), the stringency of the 2030 food waste reduction target, the monitoring mechanism and the expected impact of the
voluntary industry initiative on food prices. As primary outcome variables of interest, we measured how the randomly varied information about the design
of voluntary initiatives would affect citizens’ perception of the ambition of voluntary industry initiatives (manipulation check) and citizens’ preferences for
government regulations to reduce food waste. In addition, we also assessed effects on citizens’ intentions and perceived self-efficacy to reduce their own
food waste (see Methods and Supplementary Tables 7 and 8 for details).
NATURE FOOD | www.nature.com/natfood
Articles
NAture Food
proposal to reduce food waste for the control group (in blue, no
social norm information) compared with the combined national
and international social norm frame group (in red, Swiss and global
efforts to reduce food waste by 50% by 2030). The results of inter-
acting policy frames and design elements indicate that both inter-
national and national social norms can increase the positive support
effects of more ambitious reduction targets for the food industry.
While a target of reducing food waste by 10% by 2030 does not
increase support in the control group without social norm empha-
sis, in the two social norm framing groups support increases sig-
nificantly and substantially by 0.29–0.35 scale points. We observe a
similar picture for the more stringent reduction targets. While in the
control group support for policy proposals with a reduction target
of 25–50% increases by around 0.35 scale points, in the two social
norm framing groups support increases by around 0.60 scale points.
The strong norm framing effect on the policy stringency attribute
can be explained by the fact that both the national and international
norm frame explicitly highlight the goal of reducing food waste by
50% by 2030. In essence, this might have primed respondents to
pay special attention to the proposed stringency of the policy target
(compared with other policy design features).
We decided to present the two norm frames jointly because,
against our expectations, we find few significant differences
between national and international social norm frames that only
concern two policy design attributes. As shown in Supplementary
Table 4 and Supplementary Fig. 6, for the national norms treatment
the effect of breadth is relatively stronger (compared with baseline)
than the effect of the international norm treatment. For the obliga-
tory reporting with monitoring, however, the effect of the inter-
national norms treatment (compared with baseline) is relatively
stronger than of the national norms treatment.
Taken together, the combined framing and conjoint experiment
clearly shows how important policy design is to ensure public sup-
port for costly food waste reduction policies. A majority of Swiss
citizens appears to accept even substantial policy-induced food
price increases of up to 10% if the proposal entails ambitious food
waste reduction targets, wide firm coverage and public monitoring
mechanisms. Social norm frames emphasizing national or interna-
tional efforts to reduce food waste can further enhance support for
policy proposals with more stringent reduction targets.
Voluntary industry initiatives can crowd in support for stricter
governmental regulation. To test for the potential effect that volun-
tary industry initiatives against food waste have on public support
for government interventions to reduce food waste, we conducted
a second vignette survey experiment with a separate subsample
(n = 1,897; see details in Fig. 4 and Methods). We varied similar
attributes as in the conjoint experiment (besides the voluntary and
not legally binding nature of the initiative, which was fixed to this
attribute level). Rather than assessing individual support for these
differently designed voluntary food waste reduction initiatives,
here we studied how such information would affect citizens’ per-
ception of the ambition of voluntary industry initiatives (manipula-
tion check) and citizens’ preferences for government regulations to
reduce food waste.
Our manipulation check (Supplementary Table 5) indicates that
the different vignette attributes did, as expected, affect citizens’ per-
ceived ambition of industries’ voluntary initiatives to reduce food
waste. Figure 5 shows the effect of randomly varied information on
industry initiatives on citizens’ preferences for government regu-
lation to reduce food waste. These results suggest that broad and
unambitious voluntary initiatives do not reduce but rather increase
demand for stricter government food waste regulation. A decrease
in depth (that is, the stringency of the voluntary food waste reduc-
tion targets) reduces the perceived level of ambitiousness of vol-
untary food waste reduction initiatives (Supplementary Table 5)
and leads to more demand for government food waste regulation
(by around 0.14 points on a 5-point Likert scale, all other things
being equal). Also, in line with our expectation that broad volun-
tary initiatives can increase issue salience (visibility) and thus sup-
port for regulation, we find that an increase in breadth (that is, the
number of firms participating in the voluntary initiative) leads to
higher demand for government regulation, in our case by around
0.13 points on a 5-point Likert scale (compared to few participating
companies). Information on monitoring and food price increases
does not affect demand for regulation.
In sum, our experiment suggests more positive than negative
feedback effects of voluntary firm actions. Note that the degree of
such positive feedback effects should be interpreted with some cau-
tion because our robustness tests using the Benjamini–Hochberg52
procedure support the robustness of our findings at a false discov-
ery rate of 20% but not of 10% (see Methods and Supplementary
Table 9 for details). Overall, however, we provide robust evidence
that voluntary industry initiatives do not crowd out public support
for governmental food waste regulation.
Implications for policymaking and future research on food
waste. Our study shows that policymakers can use appropriate
policy framing and design to achieve sufficient public support
and make ambitious but costly food waste regulation politically
feasible. Our evidence suggests that policy frames emphasizing
national or international social norms in favour of reducing food
waste can increase public support for more ambitious reduction
targets. Furthermore, a majority of Swiss citizens support food
waste regulation even if this leads to substantial increases in food
prices, but only if such policies set stringent reduction targets and
are transparently monitored. We also show that the risk of shal-
low voluntary industry action undermining support for ambitious
governmental food waste reduction and crowding out consum-
ers’ intentions to reduce their food waste is low. On the contrary,
we find that broad and unambitious voluntary industry initiatives
potentially induce more support for ambitious government food
waste regulation.
Control
Any treatment
All companies
Half of companies
Few companies
No increase
10% reduction
25% reduction
50% reduction
No reporting/monitoring information
Reporting and monitoring
No price information
0% price increase
10% price increase
3.2 3.3 3.4 3.5 3.6
Marginal means for preferences for regulation (index)
Fig. 5 | Marginal means for citizens’ preference for government food
waste regulation based on average marginal component effects of
vignette treatments. Marginal means for the total control and the average
of all treatment groups are marked in blue (above the solid horizontal line),
while marginal means for the specific vignette treatments are marked in
red (below the solid horizontal line). Dashed horizontal lines separate the
attributes of the specific vignette treatment. Preferences for government
food waste regulation were measured with an additive index of two items
and a 5-point Likert scale with higher values indicating higher support
ratings (Methods). Error bars represent 95% confidence intervals. See
Supplementary Table 6 for the detailed regression outputs.
NATURE FOOD | www.nature.com/natfood
Articles NAture Food
The main policy implication of our research is that Swiss citizens
support ambitious but costly measures for reducing food waste and
that majority backlash against such measures is not very likely. It
also shows, however, that appropriate policy design and framing are
key to nurturing public support. Finally, policymakers do not have
to be overly concerned that voluntary industry initiatives under-
mine support for ambitious government policies, as long as such
public policies set stringent reduction targets, are transparently
monitored and policymakers explain to voters the importance of
adopting these measures.
Our study offers a unified research design to simultaneously
assess the effects of policy framing, design and feedback on public
opinion. Building on this analytical template, we see several fruitful
avenues for future research. First, similar studies could be carried
out comparing responses in high-, middle- and low-income coun-
tries to explore to what extent demand for food waste reduction is
income-elastic and/or dependent on institutional contexts. Doing
so would help in assessing the external validity of our result (see also
Supplementary Section 1). Second, further studies could employ
natural experiments (for example, during public referenda) or field
experiments (for example, in public cafeterias or supermarkets) to
validate the results in different real-world settings. In particular,
social norms and policy feedback effects evolve dynamically over
time. Real-world exposure to voluntary industry food waste reduc-
tion campaigns over time and active deliberation about the impor-
tance of reducing food waste among friends and family are likely
to trigger stronger social norm and feedback effects than treat-
ments embedded into singular, cross-sectional survey experiments.
Our experimental design does not allow us to draw causal infer-
ences about the exact mechanisms through which the social norm
frames affect public support. Future studies could investigate these
mechanisms in more detail. Moreover, institutional signals and the
adoption of binding government policies against food waste might
lead to stronger social norm and feedback effects than the adoption
of voluntary private regulation measures39. Third, future research
should also assess the risks of potential negative feedback effects.
For instance, previous research has shown that norm messages
communicating the prevalence of non-sustainable behaviours can
decrease policy support and increase unsustainable behaviours53,54.
Fourth, the effects of voluntary industry initiatives on public opin-
ion might not only depend on the scope of covered firms but also on
the type of firms (for example, size of firms, role within the supply
chain, public visibility). Further research could assess this in more
detail. Finally, consumer misperceptions of their food waste might
explain why citizens support supply-side policies that set stringent
reduction targets for the food industry but oppose demand-side
policies that directly target consumers13. Future studies should
investigate how to change such misperceptions and accelerate food
waste reduction across different socioeconomic contexts.
Methods
The research design for our study relies on the combination of a conjoint
experiment55 with framing and vignette experiments56. These were embedded in
a population-representative survey implemented in Switzerland from 4 June 2019
to 21 October 2019 (80% of fieldwork completed by 7 July 2019) within wave 3 of
the Swiss Environmental Panel57. The data and survey instruments for this study
can be accessed via https://doi.org/10.23662/FORS-DS-1220-1. Replication code
to reproduce the analyses presented in this study are available in the Harvard
Dataverse via https://doi.org/10.7910/DVN/MPFAJW.
Data and sample representativeness. The ongoing, multiwave, dual-mode
Swiss Environmental Panel focuses on the attitudes and preferences of the Swiss
population on environmental issues. Wave 3 of this panel focuses on beliefs,
behaviour and policy preferences concerning food waste. The Swiss Environmental
Panel is conceived by ETH Zurich in cooperation with the Swiss Federal Office for
the Environment (FOEN). The fieldwork for the Swiss Environmental Panel was
completed by ETH Zurich and the Decision Science Laboratory of the Department
of Humanities, Social and Political Sciences, ETH Zurich. Pen-and-paper
questionnaires and corresponding online questionnaires (administered with
Qualtrics) were prepared by the authors. In total 3,229 participants responded to
the survey, of whom 601 (18.62%) replied via pen-and-paper, and 2,627 (81.38%)
via the online questionnaire. It should be noted that we invited the survey
population by post, followed by two postal reminders to participate in the survey.
All respondents received a personalized online token in the invitation letter and
contact information if they preferred a mail-in pen-and-paper questionnaire. With
the second reminder, all respondents were sent a pen-and-paper questionnaire by
default. Because we had no interviewers present in households, interviewer effects
cannot bias our results. Median interview length was 13.79 min (online sample).
Due to Switzerland’s multilingualism, we fielded the survey in the country’s
three major languages: German, French and Italian. The survey was approved by
the Ethics Commission of ETH Zurich (decision EK 2019-N-43). The response
rate to this survey was 70.36% of wave 1 participants. Wave 1 participants were
recruited based on a random sample of the Swiss population aged over 15 yr from
the population register of the Federal Statistical Office (BFS/SRPH). It therefore
comprises Swiss residents. Sampling was conducted as a simple random sample on
the level of NUTS-2 regions; the response rate was 32.17%. The BFS sample mirrors,
excluding random errors and uneven response rates, the Swiss resident population.
As discussed in Rudolph et al.58, there is no evidence that take-up of the survey is
particularly biased on sociodemographic respondent and population characteristics.
Because the sample is based on a random draw of the resident population, we do not
have to worry about the frequently discussed match of online-access quota-samples
to the general population59,60. We hence conduct our experiment in a high-quality,
address-based representative sample of the Swiss population, and are confident that
our survey speaks to the preference formation of Swiss citizens at large.
As presented in Rudolph et al.40, assessing to what extent results from this
Swiss sample generalize to other countries will require additional research, and
our study hopefully provides a useful template for this. It is noteworthy, however,
that Swiss citizens’ hold very similar attitudes towards regulator y policy of business
activities as citizens in other high-income countries. While regulation in the area
of food waste is not addressed specifically (to date, there is unfortunately a paucity
of public and cross-country comparative research on this topic), it still indicates
general preferences of Western countries towards government regulation. We
follow Rudolph et al.40, who show these similarities based on data from the 2016
International Social Survey61. This suggests that using a similar or even identical
study designs in other high-income democratic countries is likely to produce
somewhat similar findings. Detailed results are reported in Supplementary Fig. 1.
Research design. The research design for our study relies on standard survey
items, and on conjoint, framing and vignette survey experiments. All respondents
saw the standard survey items first. Subsequently, 40% of our sample were exposed
to the conjoint experiment (n = 1,231 excluding missing observations), and within
the conjoint experiment a third each was exposed to three conditions of the policy
framing experiment. In total, 60% of our sample (n = 1,897 excluding missing
observations) were exposed to the vignette experiment on policy feedback, with
one-third serving as placebo group therein, and two-thirds receiving the vignette
attributes in a full factorial design.
Conjoint experiment. In the conjoint experiment, we confronted respondents
with four tasks requiring them to make a choice. For each task, they had to choose
their preferred policy out of two policies, A and B, for dealing with food waste in
Switzerland and rate the policies individually (7-point scale). Policies A and B were
displayed in a fully randomized way from the full set of levels. Supplementary Fig.
2 illustrates how such a choice task was presented to respondents.
We chose the dimensions of the conjoint experiment such that they mirror
our theoretical arguments (see Fig. 1 for an overview). The first attribute, ‘binding
nature’, depended on whether food waste policies were enacted via private or public
regulation. The second attribute, ‘policy target’, varied the policy stringency, that
is, how ambitious policies are (from 0% to 50% reduction). The third attribute,
‘policy scope’, measured industry coverage and varied from ‘a few’ to ‘all firms’
active in the sector. The fourth attribute, ‘monitoring mechanism’, covered
whether corporations have to report and/or are being monitored in their food
waste reduction efforts. We propose that respondents’ perceived ambitiousness of
the policy increases for the second, third and fourth attributes. Finally, the fifth
and last attribute, ‘policy-induced costs’, ranging from 0% to 10%, captured the
implications of the policy proposal for food prices.
Overall, these dimensions capture: (1) whether respondent attitudes are shaped
by private industry versus public regulation, that is, who implements a policy; (2)
whether respondent attitudes are shaped by policy ambitiousness, as covered with
the three dimensions of policy target, policy scope and monitoring mechanism;
and (3) cost implications, and hence the trade-off involved for citizens (policy
coming at a price).
Framing experiment. In a framing experiment, we additionally assessed whether
a social norm framing of the food waste reduction policy affected respondents’
preferences on regulation. For this, we presented experimentally varied emphasis
frames to respondents. These were crossed over with the conjoint experiment as
stated above. Before respondents entered the conjoint task, they were presented
information as follows (translation from the German original):
NATURE FOOD | www.nature.com/natfood
Articles
NAture Food
A placebo group saw a general statement that highlighted: “There are different
opinions on how to deal with food waste in Switzerland.”
A first framing treatment group then read an international (descriptive) social
norm frame: “All countries in the world have set themselves the goal of reducing
food waste by companies and households by half (50%) by 2030 as part of the
United Nations’ Sustainable Development Agenda. This is supposed to better
protect the environment and the climate. The United Nations has called on all
countries to develop an action plan against food waste. This shows that the issue
enjoys broad international support.”
A second framing treatment group then read a national (descriptive)
social norm frame: “Switzerland has set itself the goal of reducing food waste
by companies and households by half (50%) by 2030 as part of the sustainable
development agenda. This is supposed to better protect the environment and the
climate. A majority of the Swiss Parliament, with votes from all parliamentary
groups, has mandated the Federal Council to develop an action plan against food
waste. This shows that the issue enjoys broad support in Switzerland.”
Before each consecutive conjoint task, the treatment vignettes were restated
as “In view of the broad international consensus on reducing food waste, various
packages of measures are now being discussed”, for the international social norm
frame, and as “In view of the broad consensus in Switzerland on reducing food
waste, various packages of measures are now being discussed”, for the national
social norm frame.
With these framing treatments, we can test whether respondent preferences
can be moved into support for more ambitious policy packages, while being more
accepting of price increases by a strong statement on international or domestic
norm setting. Following March and Olsen62, a norm that effectively guides
behaviour needs: (1) to be made aware to citizens; (2) to be set by a dominant
institution; and (3) to provide clear prescriptions and adequate resources (that is,
doable action in an unambiguous way). Our framing experiment was designed to
fulfil these three criteria because we assume that prior knowledge of citizens on an
existing norm on the issue is low, dominant institutions were shown to have set the
norm (UN and international consensus for the international norm frame, and the
Swiss parliament in a non-partisan way in the domestic norm frame), including a
clear prescription (50% policy goal for food waste reduction).
Vignette experiment on policy feedback. To test the effects of policy feedback,
we build on current literature that works with experimental vignettes that provide
varying descriptions on private industry initiatives on the policy matter before
eliciting preferences on government regulation, attitudes towards the policy goal
and consumers’ intentions to reduce their own food waste26,27.
To this end, we randomly exposed participants in this study arm to varying
information about voluntary measures by Swiss firms on food waste. In particular,
participants were randomly assigned to either a placebo group or a treatment
group that saw information on company behaviour that varied along four
dimensions: policy scope, policy target, monitoring mechanism and price increases
(similar to the conjoint dimensions). Figure 4 provides an overview.
The placebo text that was administered to respondents read as follows
(translation from the German original): “There are different opinions on how to
deal with food waste in Switzerland. We would like to hear your personal opinion
on this.”
The exact wordings of the experimental vignettes were composed of the levels
of the attributes presented in Fig. 4. As an example, we present here the most
stringent and the most lenient version of these vignettes.
The most stringent vignette reads:
• Almost all companies in the Swiss food industry will take voluntary measures
as part of a new campaign starting in autumn 2019.
• ese companies voluntarily commit to reducing their food waste by one-half
(50%) by 2030.
• ey also commit to accurately record the amount of food waste they produce,
publish this information, and have it monitored by the state.
• ese measures will make food products from these companies no more
expensive (0%).
The most lenient vignette reads:
• A few companies in the Swiss food industry will take voluntary measures as
part of a new campaign starting in fall 2019.
• ese companies voluntarily commit to maintaining their food waste at cur-
rent levels until 2030 and thus not increase it any further.
After respondents read these vignettes, we elicited attitudes towards the
perceived ambition of corporate efforts and governmental food waste regulation
with 5- and 7-point Likert scales.
A first item serves as a manipulation check to investigate whether the
experimental vignettes affected the perceptions of respondents on the ambition
of companies to reduce food waste. The item read: “How strong or weak do you
consider the commitment of the Swiss food industry to voluntarily reduce food
waste? Please indicate your answer on a scale from 1 (very weak) to 7 (very strong).”
Subsequently, we fielded several items on preferences towards government
regulation of the food sector. Because we use several outcome measures for the
same underlying concept of demand for regulation, we follow Mutz56 who proposes
to combine items into a more robust measurement of our dependent variable. We
use the following two items to build an additive index of individuals’ preferences
for government food waste regulation by adding the respective scores of the
two items together and dividing the sum by 2. Higher scores indicate a higher
preference for government food waste regulation. The two items used to build
the additive index of preferences for government food waste regulation read as
follows: “To what extent do you agree or disagree with the statements below?
Please indicate your answer on a scale from 1 (completely disagree) to 5
(completely agree).”
• Voluntary measures by companies and households are sucient to reduce
food waste in Switzerland to an acceptable level. (We reverse-coded this item
to build the additive index.)
• Government measures are needed to reduce food waste.
We additionally investigated treatment effects on individuals’ perceived
self-efficacy and intentions to reduce their personal food waste (compared with
supporting governmental regulations for stricter industry food waste reductions).
We measured respondents’ perceived self-efficacy and intentions to reduce their
own food waste via the following two survey items: “To what extent do you agree
or disagree with the statements below? Please indicate your answer on a scale from
1 (completely disagree) to 5 (completely agree).”
• Individual consumers like me can do very little about the problem of food
waste. (In Supplementary Table 7, we use a reversed scale for this item to make
interpretation easier.)
• In the future, I will do more myself to prevent food waste.
Our survey indicates that most Swiss citizens substantially underestimate
their own food waste contributions (for example, compared with retailers, see
Supplementary Figs. 7 and 8). However, we do not see that voluntary industry
initiatives crowd out individuals’ intentions to reduce their own food waste
(Supplementary Table 8). Instead, Supplementary Table 7 shows that such
initiatives rather crowd in consumers’ perceived self-efficacy to reduce their food
waste (by around 0.1 points on a 5-point Likert scale, all other things being equal,
although significant only at the 10% level, and without substantial heterogeneity if
we differentiate different types of voluntary company initiatives).
Taken together, we expect that compared with the control group the display of
an experimental vignette, and even more so of stringent experimental vignettes,
leads to a policy feedback process that in turn affects preferences for government
regulation, consumers’ perceived self-efficacy and intentions to reduce their own
food waste. Importantly, this assumes that our vignettes actually moved respondent
attitudes towards the stringency of voluntary industry initiatives. As outlined in
our manipulation check (Supplementary Table 5), we conclude that our vignette
experiment successfully and strongly manipulated respondent beliefs on the
ambition of the food industry to tackle food waste.
Estimation strategy. We analysed the data collected in the conjoint experiment
based on average marginal component effects55 for both the rating and binary
choice outcome and additionally display marginal means in Supplementary Fig.
4 based on the suggestions by Leeper et al.63. We have opted to present the rating
outcome in the main paper, and only present the (substantially similar) results for
the choice outcome in the Supplementary Information because ratings allow us a
finer-grade evaluation of respondent preferences, where absolute levels of support
or opposition to regulation can also be assessed. Moreover, the rating scale allows a
better comparison of treatment effects on public support for food waste regulation
across the different experimental parts of the study.
Similarly, we analysed the data on the vignette experiment on policy feedback
data by comparing: (1) conditional means for the placebo group versus any
vignette treatment exposure, and (2) average marginal component effects based on
the full factorial vignette design within the experimental treatment group. Average
marginal component effects are estimated based on a linear regression model, from
which we predict marginal means.
Robustness checks. One issue that needs to be discussed with respect to robustness
concerns multiple hypothesis testing64,65. First, we would like to highlight that
multiple hypothesis testing is a less problematic issue in the context of conjoint
compared with vignette experiments. Conjoint experiments have been specifically
developed to allow researchers to test multiple hypotheses within the same study.
Hainmueller et al.55, for example, state: “Practically, this implies that empirical
researchers can apply the proposed methods to test multiple causal hypotheses
about multidimensional preferences and choices.” Concerning multiple outcomes
in the vignette experiment, we relate to best practices by creating index variables
for related outcomes instead of separately testing them. This not only reduces
measurement error56 but also reduces potential problems associated with multiple
hypothesis testing64,65. Where we conduct tests on multiple experimental treatments
simultaneously, we report adjusted P values based on the procedure pioneered
by Benjamini and Hochberg52 to safeguard against reporting false discoveries.
For an acceptable false discovery rate (FDR) of 20%, all our main conclusions
uphold. For an acceptable FDR of 10%, the results of the policy design experiment
uphold, while some results of the policy framing experiment become insignificant.
Concerning the policy feedback experiment, our results on how the breadth
NATURE FOOD | www.nature.com/natfood
Articles NAture Food
and depth of voluntary business initiatives affect regulatory demand become
insignificant at an FDR of 10% (see Supplementary Section 5 for further details)
Data availability
The data sets generated during and/or analysed during the current study and
relevant study documentation are available for scientific use after registration at
https://doi.org/10.23662/FORS-DS-1220-1.
Code availability
Statistical code used to analyse the data sets during the current study are available
in the Harvard Dataverse public repository at https://doi.org/10.7910/DVN/
MPFAJW.
Received: 11 May 2021; Accepted: 13 January 2022;
Published: xx xx xxxx
References
1. Springmann, M. et al. Options for keeping the food system within
environmental limits. Nature 562, 519–525 (2018).
2. Tilman, D. & Clark, M. Global diets link environmental sustainability and
human health. Nature 515, 518–522 (2014).
3. Poore, J. & Nemecek, T. Reducing food’s environmental impacts through
producers and consumers. Science 992, 987–992 (2018).
4. Bajželj, B. et al. Importance of food-demand management for climate
mitigation. Nat. Clim. Change 4, 924–929 (2014).
5. Waste and the Right to Adequate Food: Making the Connection (FAO, 2018).
6. Makov, T., Shepon, A., Krones, J., Gupta, C. & Chertow, M. Social and
environmental analysis of food waste abatement via the peer-to-peer sharing
economy. Nat. Commun. 11, 1156 (2020).
7. Forbes, H., Quested, T. & O’Connor, C. UNEP Food Waste Index Report 2021
(UNEP, 2021).
8. Reynolds, C. et al. Consumption-stage food waste reduction interventions—
what works and how to design better interventions. Food Policy 83, 7–27 (2019).
9. Gil, J. Going to waste. Nat. Food 1, 192 (2020).
10. Barrera, E. L. & Hertel, T. Global food waste across the income spectrum:
implications for food prices, production and resource use. Food Policy 98,
101874 (2021).
11. Global Call for Zero Tolerance on Food Loss and Waste (FAO, 2017); https://
www.fao.org/news/story/en/item/1038090/icode/
12. Beretta, C. & Hellweg, S. Lebensmittelverluste in der Schweiz: Umweltbelastung
und Verminderungspotenzial (ETH Zürich, 2019).
13. Fesenfeld, L., Wicki, M., Sun, Y. & Bernauer, T. Policy packaging can make
food system transformation feasible. Nat. Food 1, 173–182 (2020).
14. Hainmueller, J., Hangartner, D. & Yamamoto, T. Do survey experiments
capture real-world behavior? External validation of conjoint and vignette
analyses with a natural experiment. Proc. Natl Acad. Sci. USA 112,
2395–2400 (2014).
15. Tresch, A., Bernhard, L., Lauener, L. & Scaperrotta, L.
VOTO-Studiezureidgenössischen Volksabstimmung vom 23. September 2018
(FORS, ZDA, Link, 2018); https://www.voto.swiss/wp-content/uploads/
2018/11/VOTO_Bericht_23.09.2018_DE.pdf
16. Bain, P. G. et al. Co-benets of addressing climate change can motivate action
around the world. Nat. Clim. Change 6, 154–157 (2016).
17. Stokes, L. C. & Warshaw, C. Renewable energy policy design and framing
inuence public support in the United States. Nat. Energy 2, 17107 (2017).
18. Fesenfeld, L. e Political Feasibility of Transformative Climate Policy—Public
Opinion about Transforming Food and Transport Systems (ETH Zurich, 2020);
https://doi.org/10.3929/ethz-b-000425564
19. Druckman, J. & McGrath, M. e evidence for motivated reasoning in
climate change preference formation. Nat. Clim. Change 9, 111–119 (2019).
20. Fesenfeld, L. P. & Rinscheid, A. Emphasizing urgency of climate change is
insucient to increase policy support. One Earth 4, 411–424 (2021).
21. Fesenfeld, L., Sun, Y., Wicki, M. & Bernauer, T. e role and limits of
strategic framing for promoting sustainable consumption and policy. Glob.
Environ. Change 68, 102266 (2021).
22. Ingold, K., Stadelmann-Steen, I. & Kammermann, L. e acceptance of
instruments in instrument mix situations: citizens’ perspective on Swiss
energy transition. Res. Policy 48, 103694 (2019).
23. Wicki, M., Fesenfeld, L. P. & Bernauer, T. In search of politically feasible
policy-packages for sustainable passenger transport: insights from choice
experiments in China, Germany, and the USA. Environ. Res. Lett. 14,
084048 (2019).
24. Bergquist, P., Mildenberger, M. & Stokes, L. Combining climate, economic,
and social policy builds public support for climate action in the US. Environ.
Res. Lett. 15, 54019 (2020).
25. Druckman, J. N. & Valdes, J. How private politics alters legislative
responsiveness. Quart. J. Polit. Sci. 14, 115–130 (2019).
26. Malhotra, N., Monin, B. & Tomz, M. Does private regulation preempt public
regulation? Am. Polit. Sci. Rev 113, 19–37 (2019).
27. Kolcava, D., Rudolph, L. & Bernauer, T. Voluntary business initiatives can
reduce public pressure for regulating rm behaviour abroad. J. Eur. Public
Policy 28, 591–614 (2021).
28. Maxwell, J. W., Lyon, T. P. & Hackett, S. C. Self-regulation and social welfare:
the political economy of corporate environmentalism. J. Law Econ. 43,
583–618 (2000).
29. Urpelainen, J. Frontrunners and laggards: the strategy of environmental
regulation under uncertainty. Environ. Resour. Econ. 50, 325–346 (2011).
30. Bendor, J., Diermeier, D., Siegel, D. A. & Ting, M. A Behavioral eory of
Elections (Princeton Univ. Press, 2011).
31. Denicolò, V. A signaling model of environmental overcompliance. J. Econ.
Behav. Organ. 68, 293–303 (2008).
32. Bernauer, T. & Cadu, L. In whose interest? Pressure group politics,
economic competition and environmental regulation. J. Public Policy 24,
99–126 (2004).
33. Potoski, M. & Prakash, A. Green clubs: collective action and voluntary
environmental programs. Annu. Rev. Polit. Sci. 16, 399–419 (2013).
34. Schmid, N., Haelg, L., Sewerin, S., Schmidt, T. S. & Simmen, I. Governing
complex societal problems: the impact of private on public regulation
through technological change. Regul. Gov. 15, 840–855 (2021).
35. Kolcava, D., Rudolph, L. & Bernauer, T. What mode of environmental
regulation do post-industrial societies prefer and why? Glob. Environ. Change
68, 102226 (2021).
36. Higgs, S. Social norms and their inuence on eating behaviours. Appetite 86,
38–44 (2015).
37. Robinson, E., omas, J., Aveyard, P. & Higgs, S. What everyone else is
eating: a systematic review and meta-analysis of the eect of informational
eating norms on eating behavior. J. Acad. Nutr. Diet. 114, 414–429 (2014).
38. Nyborg, K. et al. Social norms as solutions. Science 354, 42–43 (2016).
39. Huber, R. A., Anderson, B. & Bernauer, T. Can social norm interventions
promote voluntary pro environmental action? Environ. Sci. Policy 89,
231–246 (2018).
40. Rudolph, L., Kolcava, D. & Bernauer, T. Public demand for extraterritorial
environmental and social public goods provision. Preprint at https://doi.
org/10.31219/osf.io/y3cbp (2021).
41. Huber, R., Wicki, M. & Bernauer, T. Public support for environmental policy
depends on beliefs concerning eectiveness, intrusiveness, and fairness.
Env. Polit. 29, 649–673 (2020).
42. Beiser-McGrath, L. F. & Bernauer, T. Could revenue recycling make eective
carbon taxation politically feasible? Sci. Adv. 5, eaax3323 (2019).
43. Save Food Campaign (PUSCH Praktischer Umweltschutz, 2021); https://
savefood.ch/de/
44. Pierson, P. When eect becomes cause: policy feedback and political change.
World Polit. 45, 595–628 (1993).
45. Jordan, A. & Matt, E. Designing policies that intentionally stick: policy
feedback in a changing climate. Policy Sci. 47, 227–247 (2014).
46. Soss, J. & Schram, S. F. A public transformed? Welfare reform as policy
feedback. Am. Polit. Sci. Rev. 101, 111–127 (2007).
47. Campbell, A. L. Policy makes mass politics. Annu. Rev. Polit. Sci. 15,
333–351 (2012).
48. Werfel, S. H. Household behaviour crowds out support for climate change
policy when sucient progress is perceived. Nat. Clim. Change 7,
512–515 (2017).
49. Werner, T. Gaining access by doing good: the eect of sociopolitical
reputation on rm participation in public policy making. Manage. Sci. 61,
1989–2011 (2015).
50. Food Waste in der Schweiz (SRF, 2019); https://www.srf.ch/news/schweiz/
food-waste-in-der-schweiz-schweizer-haushalte-werfen-eine-million-
tonnen-lebensmittel-weg
51. Green, D. P. e price elasticity of mass preferences. Am. Polit. Sci. Rev. 86,
128–148 (1992).
52. Benjamini, Y. & Hochberg, Y. Controlling the false discovery rate: a practical
and powerful approach to multiple testing. J. R. Stat. Soc. B 57, 289–300 (1995).
53. Rinscheid, A., Pianta, S. & Weber, E. U. What shapes public support for
climate change mitigation policies? e role of descriptive social norms and
elite cues. Behav. Public Policy 5, 503–527 (2021).
54. Sparkman, G., Howe, L. & Walton, G. How social norms are oen a barrier
to addressing climate change but can be part of the solution. Behav. Public
Policy 5, 528–555 (2021).
55. Hainmueller, J., Hopkins, D. & Yamamoto, T. Causal inference in conjoint
analysis: understanding multidimensional choices via stated preference
experiments. Polit. Anal. 22, 1–30 (2014).
56. Mutz, D. C. Population-Based Survey Experiments (Princeton Univ. Press,
2011).
57. Rudolph, L. et al. Schweizer Umwelt-Panel: Dritte Erhebungswelle:
Lebensmittelabfälle (ETH Zürich, 2020); https://doi.org/10.3929/ethz-b-
000440843
NATURE FOOD | www.nature.com/natfood
Articles
NAture Food
58. Rudolph, L., Quoß, F. & Bernauer, T. Swiss Environmental Panel Study
2018–2019, Wave 1–3, Cumulative Data (distributed by FORS, Lausanne,
2021; ETH Zurich, 2020); https://doi.org/10.23662/FORS-DS-1220-1
59. Ansolabehere, S. & Schaner, B. F. Does survey mode still matter? Findings
from a 2010 multi-mode comparison. Polit. Anal. 22, 285–303 (2014).
60. Baker, R. et al. Summary report of the AAPOR task force on non-probability
sampling. J. Surv. Stat. Methodol. 1, 90–143 (2013).
61. Edlund, J. & Lindh, A. e ISSP 2016 role of government module: content,
coverage, and history. Int. J. Sociol. 49, 99–109 (2019).
62. March, J. G. & Olsen, J. P. in e Oxford Handbook of Political Science (ed.
Goodin, R. E.) 478–497 (Oxford Univ. Press, 2009).
63. Leeper, T. J., Hobolt, S. & Tilley, J. Measuring subgroup preferences in
conjoint experiments. Polit. Anal. 28, 207–221 (2020).
64. Viviano, D., Wuthrich, K. & Niehaus, P. (When) should you adjust inferences
for multiple hypothesis testing? Preprint at https://arxiv.org/abs/2104.13367
(2021).
65. Greenland, S. Analysis goals, error–cost sensitivity, and analysis hacking:
essential considerations in hypothesis testing and multiple comparisons.
Paediatr. Perinat. Epidemiol. 35, 8–23 (2021).
Acknowledgements
We thank F. Quoss for her valuable support. We thank I. Stadelmann-Steffen,
G. Brückmann, D. Kolcava, participants at the Swiss Political Science Association Annual
Conference 2021, and participants in research seminars at ETH Zurich and LMU
Munich who provided valuable feedback on earlier drafts of this paper. We thank
R. Buchs, J. Bruker, C. Waldner and N. Radowsky for excellent research assistance.
The Swiss Federal Ministry of the Environment provided financial support for fielding
the survey.
Author contributions
L.R and L.F. contributed equally to this study. T.B. acquired the grant funding for data
collection. L.F., L.R. and T.B. contributed equally to developing the study concept and
designing the survey embedded experiment. L.R. had the lead in gathering the data. L.F.
had the lead in developing the theoretical argument. L.R. had the lead in data analysis.
L.R and L.F. had the joint lead in writing the paper.
Competing interests
The authors declare no competing interests.
Additional information
Supplementary information The online version contains supplementary material
available at https://doi.org/10.1038/s43016-022-00460-8.
Correspondence and requests for materials should be addressed to Lukas Rudolph.
Peer review information Nature Food thanks Manuel Fischer and the other, anonymous,
reviewer(s) for their contribution to the peer review of this work.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
© The Author(s), under exclusive licence to Springer Nature Limited 2022
NATURE FOOD | www.nature.com/natfood
A preview of this full-text is provided by Springer Nature.
Content available from Nature Food
This content is subject to copyright. Terms and conditions apply.
