NEW Supporting Statement Healthy Symbols - Part A 2022 Final

United States Food and Drug Administration
Quantitative Research on a Voluntary Symbol
Depicting the Nutrient Content Claim “Healthy” on Packaged Foods

OMB Control No. 0910-NEW
SUPPORTING STATEMENT
Part A: Justification
1. Circumstances Making the Collection of Information Necessary
This information collection supports Food and Drug Administration (FDA, us or we)
regulations and programs. Section 403(r)(1)(A) of the Federal Food, Drug, and Cosmetic
Act (FD&C Act) permits the use of label and labeling claims that characterize the level of
a nutrient in a food only if the claims are made in accordance with FDA’s regulations.
Such claims are referred to as “nutrient content claims.” We have issued regulations
under section 403(r)(1)(A) of the FD&C Act defining “implied nutrient content claims”
as those that imply that a food, because of its nutrient content, may be useful in achieving
a total diet that conforms to current dietary recommendations (“Food Labeling: Nutrient
Content Claims, General Principles, Petitions, Definition of Terms,” 58 FR 2302 at 2374,
January 6, 1993). We have found that a claim that a food, because of its nutrient content,
may be useful in maintaining healthy dietary practices is clearly a claim that characterizes
the level of nutrient in that food. The claim is essentially saying that the level of nutrients
in the food is such that the food will contribute to good health (58 FR 2302 at 2375). In
1994, we issued a definition of “healthy” as an implied nutrient content claim (59 FR
24232, May 10, 1994); the regulation is codified at 21 CFR 101.65(d)(2).
In 2018, FDA announced our Nutrition Innovation Strategy, outlining key priorities the
agency intended to pursue to reduce the burden of chronic disease through improved
nutrition and advance its public health mission. As one element of the Strategy we are
exploring the development of a graphic symbol to help consumers identify packaged food
products that meet FDA’s definition of “healthy.” The symbol would be a graphic
representation of the nutrient content claim “healthy” and, like the implied nutrient
content claim “healthy”, would be voluntary for packaged food companies; companies
could use the symbol on their food products if the products meet the FDA definition of
“healthy.”
In 2019 and 2020, FDA conducted a systematic review of the literature on front-ofpackage nutrition-related symbols. The global literature take-aways are:
•
•

A FOP rating system or symbol can help consumers identify and select
healthy foods;
Consumers generally prefer simple labels (such as the ones using a
summary system);
1

•
•
•

There is limited research on: (1) which type of summary system works
best, and; (2) whether consumers’ use of summary systems result in
healthier diets;
Some manufacturers have reformulated products following the
implementation of FOP nutrition symbols; there is some evidence of
increased sales of products bearing a FOP symbol; and
Institutional endorsement of logos may be related to greater confidence in
the label.

FDA also held three sets of eight focus groups to get consumer reaction, first to symbol
prototypes and then to possible FDA “healthy” symbols, with modifications made to the
symbols between each set of groups. The overall takeaways from the focus groups were
that participants were favorably disposed to the idea of FDA creating a “healthy” frontof-package symbol. Because the symbol would be coming from a trustworthy
organization like FDA, participants said shopping for healthy food items would be easier
and quicker. Participants agreed that “FDA” would need to be displayed on whatever
symbol is ultimately developed so that consumers can distinguish FDA’s designation
from a marketing effort by the manufacturer.
Certain symbol design elements resonated better with the focus group participant than did
others; e.g., rounded squares and checkmarks. Other elements engendered responses that
FDA did not desire. For example, symbols containing leaves or the color green invoked
environmental issues, “apples” invoked school.
Following the third set of focus groups, the contracting design firm and FDA continued
revising the symbols that had been tested in the focus groups, settling on the symbols
shown in Figure 1 as the set with which to begin the quantitative research.
We therefore request OMB approval of this collection of information as discussed in this
supporting statement.
Figure 1. Draft “Healthy” Symbols

2

2. Purpose and Use of the Information Collection
As part of our efforts to promote public health, we intend to conduct two consecutive
quantitative research studies—a survey (Study 1) and an experimental study (Study 2) to
explore consumer responses to the draft FOP symbols that companies could voluntarily
use on a food product as a graphic representation of the nutrient content claim “healthy.”
If research results suggest the need, the symbols will be fine-tuned following the survey
and again following the experimental study.
Study 1 will use non-probability survey methods, using a web-based panel to draw a
sample of 2,000 U.S. adults ages 18 and older who self-identify as primary food
shoppers. The sample will be balanced to the demographics of the U.S. population. The
survey instrument will focus on clarity, relevance, and appeal of the symbols in Figure 1.
Study 2 will be a controlled, randomized experiment that will use a 15-minute web-based
questionnaire to collect information from 5,000 U.S. adult members of an online
consumer panel. Conditions for Study 2 will be: (1) A set of draft FOP symbols,
including “no-symbol” controls; (2) three types of mock food products (i.e., a breakfast
cereal, a frozen meal, and a canned soup); (3) a “no-information” condition where no
explanation of the symbol is provided; and (4) a Uniform Resource Locator (URL)
condition, in which a URL is tested alongside the symbol. Each participant in Study 2
will be randomly assigned to a condition, which will include viewing a label image and
responding to various measures of the symbol’s effectiveness. Measures of response in
3

the experiment will include product perceptions (e.g., healthfulness and contribution to a
healthy diet), label perceptions (e.g., believability, trustworthiness, message effects), and
purchase/choice questions. The instrument will also collect information from participants
about their history of purchasing or consuming similar products; nutrition knowledge;
dietary interests; motivation regarding label use; health status; and demographic
characteristics.
The studies are part of FDA’s continuing effort to enable consumers to make informed
dietary choices and construct healthful diets. We intend to use the results to inform our
continued exploration of a symbol manufacturers could voluntarily use to represent the
nutrient content claim “healthy” on the food label. We will not use the results to develop
population estimates.
Experiment. Prior to administering the experimental study, the FDA will conduct nine
cognitive interviews with English-speaking adults to identify potential response error
caused by the questionnaire and study materials. A pre-test with 180 adults will also be
conducted prior to administering the experimental study. It is expected that there will be
minor adjustments to the study materials following the cognitive interviews and the
pretests.
The study will use a 15-minute Web-based questionnaire to collect information from
5,000 English-speaking adult members of an online consumer panel maintained by a
contractor. Researchers will endeavor to collect samples that reflect the U.S. Census on
gender, education, age, and ethnicity/race.
The research questions to be answered by the experimental study are:
1. Which symbol(s)…
(a) communicates that the product meets FDA’s definition of healthy?
(b) communicates that the food is a healthy choice?
(c) influences product purchase?
(d) reflects perceptions of greater trust and believability?
(e) gets consumer attention?
(f) elicits a positive affect?
(g) affects motivation to consume the foods?
2. How does exposure to the FDA Healthy definition (via an education piece)
affect outcome measures (from #1 above)?
3. How does inclusion of a URL affect outcome measures?
4. How does the checkmark versus a checkmark-alternative affect outcome
measures?
5. Does the product type mediate the effect of the symbol on selected outcome
measures (cereal vs. frozen meal vs. soup]?
6. How do demographics, health status, perceptions of product healthfulness,
motivation (and other covariates) mediate the symbol outcome measures?
Conditions for the experimental study will be: (1) a set of draft symbols, including “nosymbol” controls; (2) three types of mock food products (i.e., a breakfast cereal, a frozen
meal, and a canned soup); (3) a “no-information” condition where no explanation of the
4

symbol is provided; and (4) a URL condition, in which a URL is tested alongside the
symbol. Participants will be randomly assigned to view a food label image and answer
questions about their perceptions and reactions to the label. Product perceptions (e.g.,
healthiness, and contribution to a healthy diet), label perceptions (e.g., believability and
trustworthiness), and purchase/choice questions will constitute the measures of response
in the experiment. To help understand the data, the instrument will also collect
information about participants’ background, such as purchase and consumption of similar
products; nutrition knowledge; dietary interests; motivation regarding label use; health
status and demographic characteristics.
3. Use of Improved Information Technology and Burden Reduction
FDA intends to use only electronic means to fulfill the agency’s request.
The research will use only web-based consumer panels and questionnaires. Web-based
questionnaires not only reduce the burden on participants, but also minimize possible
administration errors and expedite the timeliness of data processing. Compared to faceto-face interviews and mailed surveys, web-based data collections are less intrusive and
less costly.
4. Efforts to Identify Duplication and Use of Similar Information
No comparable data have been collected by any other entities because the exploration of
an FDA-sponsored symbol to depict the Nutrient Content Claim “Healthy” is specific to
the interests of the FDA. The experimental study and two surveys proposed here will
provide valuable information about the effectiveness of a “Healthy” symbol and
consumers’ reactions to them.
As discussed in Item 1, FDA began exploring developing a “Healthy” symbol by
conducting a systematic review of the scientific and grey literature on FOP nutrition
symbols. Results of the literature review revealed that FOP symbols have been
extensively studied and some large-scale literature reviews on FOP nutrition systems and
symbols had been conducted. In particular, the Institute of Medicine conducted a twophased literature review on FOP symbols (Refs. 1 and 2) and concluded that they are
beneficial to consumers and that simple systems or symbols are preferred. FDA began its
specific exploration using the findings from the literature to that date and is informed by
a continued monitoring of the literature throughout the research process.
5. Impact on Small Businesses or Other Small Entities
No small businesses will be involved in this information collection.
6. Consequences of Collecting the Information Less Frequently
This is a one-time data collection.

5

If this information is not collected, FDA will not know how consumers respond to
measures of the effectiveness of- nor evaluative feedback on- an FDA-developed symbol
for depicting the nutrient content claim “Healthy” which manufacturers can voluntarily
use on the food label if their product meets the FDA definition for using the nutrient
content claim.
The study is part of the agency’s continuing effort to enable consumers to make informed
dietary choices and construct healthful diets.
7. Special Circumstances Relating to the Guidelines of 5 CFR 1320
The collection fully complies with 5 CFR 1320.5(d) (2). There are no special
circumstances associated with this information collection. The study will not require
participants to: report the information more often than quarterly; provide a written
response in less than 30 days; submit more than one original plus two copies of the
information; or retain records for more than 3 years. The experimental study will
produce results that can be generalized to the response universe of study. The study will
not use statistical data that has not yet been reviewed or approved by OMB. The study
will not include a pledge of confidentiality that is: (1) not supported by authority
established in statute or regulation; (2) not supported by disclosure and data security
policies that are consistent with the pledge; or (3) which unnecessarily impedes sharing
of data with other agencies for compatible confidential use. Finally, the study does not
involve the submission of trade secrets, proprietary information or other confidential
information.
8. Comments in Response to the Federal Register Notice and Efforts to Consult Outside
the Agency
In accordance with 5 CFR 1320.8(d), in the Federal Register of May 7, 2021 (86 FR
24629), FDA published a 60-day notice requesting public comment on the proposed
information collection. We received 43 comments, 27 of which were PRA-related. The
remaining comments were non-responsive to the four PRA topics, and so we will not
address them in this document. A. Comments regarding the necessity and practical utility
of the information being collected and FDA responseSeveral comments addressed the
necessity and practical utility of collecting information on a voluntary symbol depicting
the nutrient content claim “healthy” on packaged foods.
(Comment 1) Some comments supported FDA’s proposed collection of information
through the three proposed quantitative consumer research studies. Some comments
expressly supported FDA’s end research goal of enabling consumers to make informed
dietary choices and construct healthful diets. Some supported FDA’s intention to
understand consumer responses to draft FOP symbols and gather data and other
information to inform our thinking on a “healthy” symbol. Many comments indicated the
importance of conducting this research before taking regulatory action on any symbol.
Some comments supported conducting the research in conjunction with development of a
proposed rule that would update the definition of “healthy” on food packages.

6

Other comments opposed FDA research on a “healthy” symbol. Some of these
comments suggested the research is unnecessary, claiming that a single food is not
“healthy” or “unhealthy,” that overall diet matters more than individual foods, or that
symbols are industry marketing. A few comments suggested a “healthy” symbol could
be particularly misleading to, or misinterpreted by, people who are experiencing eating
disorders. Some comments also questioned whether a “healthy” symbol would: (1) have
a positive and meaningful impact on improving health; or (2) lead consumers to
overconsume foods bearing the symbol.
(Response 1) We intend to conduct this research now, in conjunction with further work
on updating our definition of the claim “healthy” and before taking regulatory action on
any symbol. Our intended research will help us better understand how consumers might
respond to and use a graphic symbol to identify packaged food products that meet our
definition of “healthy.” This research will help address many points raised in the
comments, such as how consumers might react to and understand a “healthy” symbol and
misinterpretations they may have.
While we agree that there are some symbols that may be used exclusively for industry
marketing, companies could use any FDA “healthy” symbol we develop and finalize only
when the product displaying the symbol meets FDA’s regulatory definition of “healthy.”
This could help consumers make more informed dietary choices and construct healthful
diets. The comments claiming that a single food is not “healthy” or “unhealthy” and that
overall diet matters more than individual foods are commenting on the “healthy” claim
itself, which we do not intend to test in this research. Rather, we intend to test consumer
reactions to symbols that could be a graphic representation of the claim. Nonetheless, we
note that a “healthy” symbol, such as the ones FDA is exploring in our research, could
help consumers choose food products, as part of their overall diet, that meet FDA’s
regulatory definition of “healthy.” The research is not designed to study long-term health
effects or consumer consumption patterns. We reiterate that this research is about
graphical representations of the nutrient content claim “healthy” – in other words, we
intend to study only the symbol, not the claim itself. Depending on the results of this
data collection, we may decide to test additional symbols or revise our current symbols.
(Comment 2) Many comments expressed a preference for conducting the research after
we revise our regulatory definition of “healthy,” as they wondered whether the definition
of the claim could influence both the design and consumer understanding of the symbol.
Some expressed concern that testing a symbol without clearly communicating what the
symbol means could lead to ambiguous results. One comment expressed concern that, by
conducting testing only on the symbols in the notice, we would not consider testing any
other symbols in the future. A few comments contended that FDA’s testing would be
invalid if the mock products used in testing do not meet FDA’s updated “healthy”
definition.
(Response 2) FDA has an existing definition for the claim “healthy,” and in the Federal
Register of September 28, 2016, we announced our intent to exercise enforcement

7

discretion around some criteria for the claim (see “Use of the Term ‘Healthy’ in the
Labeling of Human Food Products: Guidance for Industry,” 81 FR 66487 at 66527).
However, as part of this data collection, we have included experimental conditions in
which participants will read general information outlining the use of the claim “healthy”
only for purposes of this study. This will help us better understand how consumers might
respond to the symbols we are proposing to test if participants understand a “healthy”
definition, even if not necessarily an updated definition. While the symbol is intended to
represent the nutrient content claim “healthy,” our research on the symbol is not
dependent on specific criteria for “healthy.” We are researching general consumer
perceptions and impressions of the symbols themselves, not the definition that may
underly those symbols, and as such, we do not need to wait until we have a final, updated
regulatory definition of “healthy” before conducting this research. Moreover, the
symbols being tested would not need to be modified with a changing definition of
“healthy;” the symbol would remain a simple graphic representation of the “healthy”
claim.
Regarding the claim that our testing would be invalid if the mock products used in testing
do not meet FDA’s updated “healthy” definition, our mock products represent broad and
basic food categories. They include foods such as vegetables and whole grains with
limited nutrients of concern (e.g., sodium or saturated fat) that would meet our current
definition of “healthy” and would help consumers build a diet consistent with the Dietary
Guidelines for Americans, 2020-2025.
B. Comments regarding the accuracy of our burden estimates, including the
validity of the methodology and assumptions used, and FDA response
Many comments discussed the accuracy of FDA’s estimate of the burden for this
information collection, including the validity of FDA’s methodology and the assumptions
used.
(Comment 3) Several comments alleged that we provided limited details about our
proposed research studies and encouraged us to publish additional information on the
proposed scope and methodology of our consumer research to allow for more
comprehensive input from experts in the field of consumer research. One comment
suggested we “pre-register” details of the proposed studies on AsPredicted.org or
ClinicalTrials.gov so that stakeholders could better understand the primary outcome of
the research, hypotheses, analytic plan, and power analysis used.
(Response 3) We described the research in the 60-day notice, providing information on
research design, measures, sampling, and sample size. Many comments substantively
addressed these issues, and so we believe there was enough information about the studies
in the 60-day notice for the public – including consumer researchers – to comment on the
research.
We specified in the 60-day notice that we intend to use the results to inform our
continued exploration of a symbol manufacturers could voluntarily use to represent the

8

nutrient content claim “healthy” on the food label (86 FR 24629 at 24631). The comment
did not provide sufficient information regarding additional details that it believed
necessary for stakeholders to better understand this primary outcome, hypotheses,
analytic plan, and power analysis used, and it did not explain what additional details
might be available via pre-registration that would not be available in our Federal Register
notices. Therefore, we are unable to provide those details here, and we also decline to
pre-register our studies.
(Comment 4) One comment questioned the ordering of the quantitative research, asking
why the experimental studies come before the surveys. Other comments suggested we
use the two surveys to test draft symbols first to narrow down options and test the “final”
symbols in the experiment, or to conduct preliminary research to narrow the options for
the experiment.
(Response 4) We conducted several phases of qualitative research to solicit input from
consumers, allowing us to evaluate symbol prototypes and design elements to learn what
resonated with consumers. Through that process we narrowed our draft symbol options.
After considering public comments, we have reconsidered the order of the research, and
plan to conduct one survey with a larger sample size (instead of two surveys with smaller
sample sizes each) before the experimental study. In other words, we will reorder the
studies and combine the two surveys into one, which will allow us to test all symbols in a
single survey. While our proposed information collection is intended to help us better
understand how consumers might respond to and use a graphic symbol that indicates
packaged food products meet FDA’s definition of “healthy,” and all the draft symbols we
proposed to test would allow us to do that, we expect conducting a single survey first will
help us further revise and narrow down the set of symbols.
(Comment 5) One comment suggested we use a more naturalistic study environment,
such as an online store setting, instead of using images.
(Response 5) Online store settings and other naturalistic study environments have been
successfully employed in some studies on food labeling effects. One advantage of
employing such naturalistic study environments is that they more closely reflect
participants’ actual shopping experience. However, there are substantial additional costs
associated with using such research settings, and results in these settings generally do not
differ appreciably from results garnered through the simple random-assignment-tocondition design that we proposed. Therefore, we decline to change our study
environment.
(Comment 6) One comment suggested that FDA separate different aspects of the symbols
to isolate consumer perceptions of the word “healthy,” the graphic itself, and the graphic
accompanied by the word “FDA.” One comment suggested that FDA should test each
symbol with and without “FDA.”
(Response 6) Separating each aspect of the symbols for our testing would increase the
number of conditions exponentially, making the design impractical. We instead elected

9

to use a full factorial design with simple random assignment to condition, to give us
results on the performance of the various symbol designs. Using random assignment to
condition, we may be able to eliminate some symbols without needing to test particular
attributes in any one symbol. We may consider alternate study designs when we have a
narrower set of symbols.
One finding of our literature review was that institutional endorsement may be related to
greater confidence in the symbol. Our focus group research affirmed that participants
regarded symbols with “FDA” as more trustworthy than symbols without “FDA.”
Therefore, for the intended research, we are testing draft symbols with “FDA.” We may
consider additional research on this point depending on the results.
(Comment 7) One comment recommended using images of real food products in the
experimental studies instead of using mock product images.
(Response 7) FDA does not agree with the recommendation to use images of real
products in the experimental studies. Mock images remove the potential for brand biases,
a source of response error that has been demonstrated to affect the way individuals
answer survey questions. Mock food product labels psychologically remove the salience
of branded product informational cues present in the retrieval stage of the response
process (Refs. 3 and 4). Additionally, the mock product labels we designed are visually
similar to labels consumers could expect to see in stores for each given product category.
We confirmed this assertion in our qualitative testing by noting that participants
perceived the mock product labels as ones with which they were unfamiliar, but which
were plausible for the food product depicted.
(Comment 8) One comment suggested that we should assess “multi-tier symbols” in
addition to the symbols we intend to test. The comment suggested that multi-tier
symbols are those that use, for example, an increasing number of stars to indicate to the
consumer that a choice is “good,” “better,” or “best.” The comment argued that a multitiered approach could encourage consumers to make incremental improvements in their
diets, enable manufacturers to reformulate products to meet the initial tier of the system,
and increase the number of foods with at least some healthful benefits that could carry a
symbol.
Another comment suggested that FDA consider a symbol that warns consumers about
high levels of unhealthy nutrients. Another comment asserted that we should also test
what it suggested were more neutral FOP labels, such as traffic lights, nutrition scoring
symbols, and warning symbols, to better assist consumers in making healthy choices and
motivate manufacturers to make healthier foods.
(Response 8) For the purpose of this study, we are testing only symbols that would be a
graphic representation of the nutrient content claim “healthy” – a food that could bear
that claim could also bear the symbol. FDA’s systematic literature review suggested that
a summary indicator – the type we are proposing to test – would have the greatest utility
to depict the “healthy” claim to a broad array of consumers, especially those with lower

10

education or lower health literacy. As such, we disagree that we should test other kinds
of symbols to depict the nutrient content claim “healthy.” We are testing different draft
symbol designs based on our literature review and the feedback we collected through our
focus group research. Our current study plans are limited to testing summary symbols
depicting the nutrient content claim “healthy” to get reactions to design elements and to
reduce the current number of symbols under consideration. Because there are no
“healthy” tiers in the nutrient content claim, we decline to test a tiered symbol.
(Comment 9) One comment encouraged us to consider testing the “healthy” symbol
alongside other current voluntary FOP labels – rather than as the only symbol on a
package – to determine the effect of other FOP labels on the efficacy of the “healthy”
symbol.
(Response 9) Our studies are designed to test general consumer responses to the symbols
presented. Testing additional variables, such as the effect of other packaging elements on
the symbols, is outside the scope of this research. We may decide to test “healthy”
symbols alongside other FOP symbols in later research depending on the results from this
data collection.
(Comment 10) One comment recommended randomizing participants to see subgroups of
symbols, claiming it would be an efficient use of resources.
(Response 10) FDA agrees with the recommendation that participants be randomly
assigned; however, we disagree with the recommendation to have participants view
subgroups of symbols. We plan to randomly assign participants to see a single symbol
condition, including product type, information on the definition of healthy, URL/noURL, and set of symbols. Viewing a single symbol condition precludes the effects of
biases that may result from having viewed and responded to questions about one symbol
affecting responses about another symbol (Ref. 4). Therefore, even if we might use
fewer resources by assigning participants to see subgroups of symbols, the practice would
introduce biases and confounds that could make interpreting the results very difficult.
(Comment 11) One comment recommended incorporating time limits for a choice task to
better mimic real-life scenarios where consumers have only limited time to shop.
(Response 11) The current experimental study design is random assignment to condition
with no “choice task.” While time-constraint studies can be useful to test certain
variables, our research goal is for the participants to provide thoughtful responses,
unaffected by the stress that a time limit could impose.
(Comment 12) A few comments recommended that the studies be adequately powered to
enable FDA to do appropriate statistical analysis.
(Response 12) Our studies are designed to have the appropriate statistical power to
conduct all necessary statistical analysis. We will test hypotheses related to betweenlabel differences. We will impose no a priori direction of differences, if any (i.e., we

11

assume all tests are two-tailed). The target sample size (5,000 for the experimental study
and 2,000 for the survey) will yield enough observations to provide adequate power to
identify 4-way interactions of a medium size (Ref. 5).
(Comment 13) One comment recommended that we test the draft “healthy” symbols in
the context of restaurants.
(Response 13) Our research plans include testing symbols solely on packaged foods.
Testing “healthy” on a packaged food label versus in a restaurant minimizes the many
confounding factors inherent in studying claims in a restaurant environment, such as the
enormous variance in size and content of materials used to sell restaurant food. Keeping
the studies limited to packaged food labels allows FDA to better isolate various effects of
the symbols to strictly test consumer perceptions about the symbols. Additionally, as we
noted in response to another comment, we have no reason to believe that adding
additional test product categories would change the study outcomes given our goal of
testing consumer responses to the symbols.
(Comment 14) Some comments recommended that FDA include more than three mock
product types in the experimental studies because of the potential that consumer
perceptions of a “healthy” symbol might be different on different products. One
comment suggested including a variety of food categories in the studies, while a few
other comments recommended including specific product categories, including beverages
and fresh produce, so FDA could assess consumer reactions to, or preferences for
placement of, a symbol on those products.
(Response 14) FDA disagrees with the recommendation to add more product types to the
studies. We are proposing to test “healthy” symbols on a set of mock products that
belong to large food categories, with many product types within each category. The
broad product categories for those mock products are likely to contain multiple products
that currently meet FDA’s regulatory definition of “healthy.”
For our research, we chose three packaged foods that are commonly consumed and that
are clearly distinct food types. The selected products will give us sufficient information
on general consumer responses to the symbols to continue development of a proposed
“healthy” symbol. We also note that adding any products would increase the scope and
cost of the studies while providing limited new information and that the comments
provided no evidence that additional test products from other food categories would
impact our study outcome.
We decline to include a beverage as one of the mock products. While beverages that
meet FDA’s definition of “healthy” could bear any “healthy” symbol we finalize, the
same is true of any packaged food, and as we explained above, we have no reason to
believe that adding additional test product categories would change the study outcomes.
We decline to add fresh produce to the studies for the same reasons.

12

(Comment 15) A few comments recommend comparing foods with a “healthy” symbol
and foods that may have healthful attributes but do not meet the regulatory definition of
“healthy,” to evaluate whether use of the symbol might result in discouraging purchase of
foods that have important nutrients but that do not meet the definition of “healthy.”
Another comment suggested testing a variety of products (“healthier” and “less healthy”)
for each food product category included in the studies.
(Response 15) The studies are not designed to test purchasing behavior, and so we
decline to add mock products for that purpose. Rather, this research is designed to test
general consumer responses to the symbols themselves. Additionally, a product could
only bear the symbol if it qualified to bear the “healthy” claim itself – the symbol is a
graphic representation of the claim – and we are not testing the claim definition or its
effects here.
One of the study assumptions is that, like the nutrient content claim “healthy,” any food
bearing a “healthy” symbol on its label must meet the regulatory definition of “healthy.”
Therefore, to test consumer responses to the symbols, we do not need to test the ancillary
effects of a “healthy” symbol on foods that do not bear the claim. Moreover, FDA
intends to test “healthy” symbols on a set of mock products whose product categories are
likely to contain multiple products that currently meet FDA’s regulatory definition of
“healthy” – we are making no claims about the relative healthfulness of any product.
Using these mock products in our research will provide us with sufficient information to
understand how consumers might respond to a “healthy” symbol on food packaging, and
that information is our goal with these studies. Testing the selected products will give us
sufficient information to continue development of a proposed “healthy” symbol.
The comment did not provide an explanation for its recommendation to test “healthier”
and “less healthy” products for each food product category. We are not testing
“healthier” and “less healthy” versions of a given product in this research effort, as the
goal of the research is to gauge participants’ reactions to a symbol. Additionally, we are
working on updating our definition of “healthy” and will describe our proposed updated
definition in any related rulemaking. It would be inappropriate to assign relative
“healthfulness” to comparator products. Products bearing the “healthy” symbol, which
would be a graphic version of the nutrient content claim “healthy,” would have to meet
the criteria for using the claim.
(Comment 16) One comment noted a symbol with the term “FDA” may cause confusion
if that symbol is used on any products regulated by the U.S. Department of Agriculture
(USDA) and urged us to consider the potential for such confusion in our research.
Another suggested that we engage with the Food Safety and Inspection Service (FSIS)
during our proposed consumer research on the “healthy” symbol to develop a symbol that
could apply to all products that meet the “healthy” claim criteria.
(Response 16) We cannot comment on whether or how USDA might allow an FDA
“healthy” symbol on the products it regulates that meet FDA’s definition of “healthy.”

13

However, we intend to coordinate with our federal partners, including USDA, as
appropriate, as we continue our work on a “healthy” symbol.
(Comment 17) A few comments asserted it was important to consider the symbols’
placement on packaging and noted that food packaging size, type, and appearance vary,
suggesting that FDA should study how consumers may respond to a “healthy” symbol on
a wider variety of packaging formats than are currently proposed for the studies. One
comment suggested testing the symbol on different locations on the package and with
varying prominence. Another encouraged flexibility in specified requirements (e.g.,
placement, type size, color format, scannable images) surrounding the FOP symbol.
(Response 17) We anticipate that any “healthy” symbol we propose would be
standardized in certain ways. However, the purpose of this research is to gauge consumer
responses to the symbols we are testing, not to decide on a single symbol, its potential
placement on packaging, or what aspects we would require, should a company choose to
use the symbol.
(Comment 18) A few comments suggested we include questions or conduct additional
research to assess the potential for consumer misunderstanding of the symbols. Some
comments suggested that we investigate whether consumer perceptions of some symbols
might imply messages other than “healthy” (such as “organic,” “natural,” “plant-based,”
and “minimally processed”) or whether the symbols we are testing may appear similar to
other existing or abandoned symbols (such as the USDA Organic Seal, Smart Choices, or
any of USDA’s bioengineered food symbols). One comment claimed that a lack of
legibility of the text in the symbol could cause consumer confusion. Another comment
recommended that FDA avoid leaf or nature imagery in the symbol because it could
imply that the product is plant-based, “natural,” or unprocessed. One comment
encouraged us to examine how appealing the symbols are to consumers, and another
comment described the proposed symbols as too simplistic.
(Response 18) We have selected study designs and draft symbols that we expect, when
used together, will reveal how consumers will react when they see such symbols on a
food label. We included questions in our studies on what the symbols lead participants to
believe about the products bearing them. We also expect to hear from participants
whether the symbols we are testing are perceived as too complex, too simplistic, or
invoke concepts other than “healthy.”
We agree that any symbol we propose should be legible and minimize imagery that our
research indicates could widely lead consumers to think the symbol means something
unintended. As such, we will add an open-ended question to the experimental study
asking what the symbol brings to mind to help determine if any symbols should be
removed from consideration or revised on this basis. Moreover, we agree that the FDA
symbol design for “healthy” should not be easily confused with other existing symbols
and should be viewed as professional and credible by consumers. We expect to get some
data on these points through this round of testing and may undertake further research
before we make any formal regulatory decision on a symbol.

14

(Comment 19) One comment suggested that FDA test other terms besides “healthy,” such
as “nourishing” or “nutrient-dense.”
(Response 19) We are not testing the nutrient content claim “healthy;” rather, we are
testing consumer responses to graphic representations of the claim. We similarly do not
intend to test other terms.
(Comment 20) Several comments supported conducting research on the use of an
accompanying URL with the “healthy” symbol; however, others stated the purpose for
including a URL was unclear, and one comment expressed concern that a URL would not
work as well in a brick-and-mortar retail setting. Another comment stated that future
labeling could include the use of other technologies, such as a QR code or digital
watermark, to provide consumers access to all the labeling information included on the
package and suggested that we incorporate digital disclosure flexibility into our labeling
regulations because technology continues to evolve. Other comments suggested that
consumers may not have internet access in stores or may not know how to use QR codes,
while another comment suggested that researchers could develop unique QR codes for
each condition and track participant use.
(Response 20) Our research efforts on the “healthy” symbol are intended to collect
sufficient data for the development and finalization of a “healthy” symbol. We are
studying several dimensions of a proposed symbol, including the inclusion of a URL as
part of the symbol. This research will help us better understand study participants’
reactions to and understanding of those different elements.
Our preliminary research indicated that participants are interested in learning more about
the symbols, and a URL can serve as a representation to participants that more
information is available. The current study design proposes to test a URL alongside
some of the symbols to gauge the ability of the URL to indicate that information about
the symbol is available and to assess the degree to which a URL improves confidence and
trust in the symbol. We are not studying participants’ actual ability to access the URL in
stores or elsewhere. We are also not considering inclusion of other technologies, such as
a QR code or digital watermark, in this information collection because a URL will help
us gauge whether participants want a way to access additional information about the
symbol. Further, a QR code or digital watermark would not indicate government
involvement in the way a URL ending “.gov” may, and we are interested in how
participants will respond to a “.gov” URL.
While we agree that technology changes over time, we are only studying consumer
responses to the symbols in this research. It would be premature to comment on any
requirements surrounding any symbol we might propose. However, we could consider
other digital elements, such as QR codes or digital watermarks, in future research
depending on our future research goals.

15

(Comment 21) Some comments raised concerns that the use of FDA’s name as part of the
“healthy” symbol. The comments said that the use of FDA’s name could create the
appearance of an FDA endorsement of a given food.
(Response 21) We are testing draft symbols with “FDA.” We note that we are studying
several dimensions of a proposed symbol to help us better understand study participants’
reactions to and understanding of those different elements, including any impression of
an FDA endorsement.
(Comment 22) A few comments expressed uncertainty as to whether FDA research
participants would come from a nationally representative sample and recommended
paying particular attention to or using quota samples similar to the demographic
breakdown of the U.S. population regarding sex, age, race/ethnicity, income, and
education. Some comments also stated that FDA should consider oversampling from
certain groups at highest risk for dietary-related disparities, asserting that it is important
to ensure that any proposed healthy symbol works well among all populations. One
comment noted this is especially important for lower-education groups who, the comment
asserted, may be less likely to use or understand the package’s nutrition label. Some
comments also requested that FDA screen participants to ensure a sample large enough to
collect responses from food-allergic individuals, caregivers to food-allergic individuals,
and parents.
(Response 22) We designed our studies to test consumer responses to draft symbols in a
randomized controlled setting, with participants drawn from a general population. Our
research collection is not intended to produce population estimates. However, we intend
to select the samples in each study to be reflective of the general U.S. population (e.g.,
sex, race/ethnicity). We believe our approach is reasonable because any “healthy”
symbol we finalize will be available to the general U.S. population.
(Comment 23) One comment suggested that we test a Spanish language version of the
symbol and consider whether including “FDA” as part of the symbol would resonate with
consumers of products sold in other countries.
(Response 23) Our regulations, at 21 CFR 101.15(c), generally require that the labeling
of all food offered for sale in the United States be in English, and outline requirements
for manufacturers that also choose to label their products in additional
languages. Because we generally require only English labeling, and manufacturers may
choose whether to use or include foreign-language labeling, we are testing only an
English-language version of the symbol in this set of studies.
As for products sold in other countries, the nutrient content claim “healthy,” and any
related symbol we finalize, are specifically for products marketed and sold in the United
States. We decline to comment on marketing and sales in, or the food labeling
requirements of, other countries.

16

(Comment 24) One comment argued that we should not generalize the results from this
study to all FOP label systems.
(Response 24) We agree that findings from this research should not be generalized to all
FOP label systems.
C. Comments regarding ways to enhance the quality, utility, and clarity of the
information to be collected, and FDA response
Several comments suggested ways to enhance the quality, utility, and clarity of the
information about “healthy” symbols to be collected.
(Comment 25) Some comments stated FDA should conduct thorough research regarding
the development and finalization of a symbol for “healthy,” and should collect
comprehensive data so that FDA’s final decision promotes health.
(Response 25) FDA agrees with the comments. Our research goal is to explore consumer
responses to draft symbols that could represent the nutrient content claim “healthy.” The
goal of the symbol is to help consumers make more informed dietary choices.
(Comment 26) Several comments recommended we change certain aspects of the
questions we include in the experimental study. Some comments suggested that we
select specific outcome measures, such as purchase intent, sales data, ability of the
symbols to attract consumer attention, long-term behavior change, consumer perceptions
of the taste and cost of products bearing the “healthy” symbol, the healthfulness of the
products consumers purchased, the number of “healthier” products purchased in a
shopping setting, and any unintended consequences of the symbol.
One comment recommended adding covariates, such as health status, particularly for
conditions that are related to nutrition, such as diabetes, weight status, and hypertension,
to help us understand responses. Regarding the interpretation of measurements, one
comment suggested we avoid “believability” or “trustworthiness” as indicators of which
symbol can help people make more informed dietary choices, claiming that these are not
strong predictors of behavior. The comment cited a study on cigarette pack warning
messages that found that measures on the effects of the warning message resulted more in
intended behavior change than did measures on attitude perceptions (Ref. 6).
Another comment recommended FDA provide an option for open-ended responses to
gauge consumers’ perceptions of “healthy.”
(Response 26) The intended studies cover the key measurements and covariates that will
help us understand consumer perceptions of the symbols. The comments did not provide,
and we are not aware of, evidence that adding covariates or measurements would enhance
the quality, utility, or clarity of the information we intend to collect. We will evaluate
our draft symbols based on our analysis of all – not just a subset – of these measurements.
We acknowledge that there are measurements we are not including in this research effort

17

(e.g., long-term behavior changes). These studies are designed to explore consumer
responses to the draft symbols, and inclusion of variables such as long-term behavior
changes would be premature.
We plan to use a variety of measures to help understand the potential impact of a
voluntary FOP symbol for “healthy,” and intend to use “believability” and
“trustworthiness” as outcome measures because well-established scientific literature has
shown that consumers’ attitudes and perceptions affect their behavior (Refs. 7 to 9).
Additionally, we note that the cigarette-pack study one comment cited qualified its
findings as unsure if the same would be found in other message or product scenarios
(Ref. 6). Because the published literature does not indicate that “effects perception
measures” have been tested in the food label domain, we will add some questions to the
experimental study to evaluate their use as outcome measures compared to “message
effects measures.”
We disagree with the suggestion to query consumers on their perception of “healthy.”
Our research is designed to test consumer responses to the draft symbols, not determine
consumer perceptions of “healthy.”
D. Comment regarding ways to minimize the burden of the collection of
information on respondents, including through the use of automated collection
techniques, when appropriate, and other forms of information technology, and
FDA response
One comment discussed minimizing the information collection burden on respondents to
our proposed “healthy” symbol research.
(Comment 27) One comment supported the proposed research and noted that the use of
online surveys will help alleviate participant and administrative burden while ensuring
that the research reaches sufficient participants.
(Response 27) We agree with the comment for the purposes of this research.
E. Nonresponsive comments to the PRA
Some comments addressed aspects of “healthy” symbols that are outside the scope of this
information collection or addressed issues other than the “healthy” symbol research.
These discussed, for example, the definition of “healthy,” potential impacts of the
“healthy” nutrient content claim generally, whether the symbols should be voluntary or
mandatory, and whether we should develop an accompanying consumer education
campaign. These are outside the scope of this information collection, and we will not
address them here. Interested parties will have an opportunity to comment on any
“healthy” symbol we propose and any proposed updated definition of the nutrient content
claim “healthy” in response to their respective Federal Register notices.

18

9. Explanation of Any Payment or Gift to Participants
Experiment and Survey. Participants in the cognitive interviews will be recruited from
an online consumer panel, and all cognitive interviews will be held online. Each
participant will receive a cash incentive of $75 to participate in a one-hour interview.
Survey. The overall sampling frame for the surveys is the Ipsos Online Panel. The Panel
is an actively managed research access panel that uses multi-source recruitment to
maintain a base of respondents’ representative of national demographic distributions. The
Panel includes individuals who have volunteered to take part in market research.
Survey participants receive an incentive for participating in the Panel. They receive
points for completion of surveys, and the points can be redeemed for cash or prizes.
Experiment. Experimental study participants will be drawn from a panel maintained by
Prodege. Prodege incentivizes its Internet panel members with a digital currency for
participation in surveys and other online activities. This digital currency can be used to
purchase gift cards with well over a hundred options to choose from and/or can direct it
to 501C3 charitable organizations. The appropriate incentive that panel members receive
for participation is based on multiple factors including an approximate length of the
survey and audience (i.e., business professionals vs. general consumers). Some examples
of incentive partners include Amazon, Pizza Hut, Best Buy, Macy’s, American Airlines,
Hertz, Target, iTunes, etc. Charitable groups include groups like The American Red
Cross and Wounded Warriors. There is no additional payment or gift associated with
participation in the study proposed here.
Consumers are invited to join the Panel directly through Prodege's network of portal sites
and complete a double opt-in registration process with multiple verification steps
including CAPTCHA, IP Address verification, and mobile device reputation
check. Currently, Prodege's Panel has over 100 million participants worldwide.

10. Assurance of Confidentiality Provided to Participants
In preparing this Supporting Statement, we consulted our Privacy Office to ensure
appropriate identification and handling of information collected.
This ICR will collect personally identifiable information (PII). The PII collected
typically consists of name and contact information. PII is collected on behalf of the FDA
by a contractor or vendor who conducts surveys. PII is collected in the context of studies
on consumer self-reported responses to survey questions which include reacting to a
variety of symbols that FDA is exploring for representing the nutrient content claim
“healthy.” Information collected by the vendor or contractor will be summarized into
aggregate form, sent in aggregate to FDA (no PII will be included), and destroyed after

19

the study or interview has been completed. Collected PII is used to notify potential
respondents of their selection and includes name and contact information. All individual
information collected will be kept secure by the vendor or contractor. FDA and any
vendor or contractor will disclose identifiable information only to the extent authorized
by the individual or required by law. Contractors or vendors maintaining information
will destroy it in accordance with applicable records retention and other requirements per
contract terms after the aggregate information has been provided to FDA and the survey
has been completed. In keeping with IRB/Human Subjects Research protocols, the
clearance process ensures that study data is appropriately secured (e.g., housed on the
Contractor’s servers, password protected, separate storage areas for each study, access
controlled).
FDA determined that although PII is collected it is not subject to the Privacy Act of 1974
and the particular notice and other requirements of the Act do not apply. Specifically, the
contractor does not use name or any other personal identifier to routinely retrieve records
from the information collected.
All data will be collected with an assurance that the participants' answers will be kept
secure to the extent provided by law, and the study instruments will contain a statement
to that effect. Information is protected from disclosure under the Freedom of Information
Act (FOIA) under sections 552(a) and (b) (5 U.S.C. 552(a) and (b)), and by part 20 of the
agency’s regulations (21 CFR part 20). As noted above, identifying information will not
be included in the data files delivered by contractors to the agency.
The contractors will not share personal information regarding panel members with any
third party without the participant’s permission unless it is required by law to protect their
rights or to comply with judicial proceedings, court order, or other legal process.
Identifying information will not be included in the data files delivered to the agency.
FDA will receive data for analysis in aggregate form. Although Prodege and Ispos retain
contact information on participants for honoraria purposes, individually identifiable
information is not shared with anyone, including FDA and its contractors; it is stored
separately from the survey data file and is not linked in any way to participant responses.
The contractors maintain restricted access to all data preparation areas (i.e., receipt and
coding). All data files on multi-user systems will be under the control of a database
manager, with access limited to project staff on a “need-to-know” basis only. The
contractors will take the following security measures to ensure separation between
participants’ identity and their survey data. First, questionnaires and survey instruments
have no personally identifying information (PII) on them. No participant name, address,
email address, phone number or any other kind of PII appears on the instruments. The
only way a survey is identified is with a digital identification number. Second, while the
invitation method, whether email, mail or direct mail will inherently have PII information
included, this will not be combined with survey responses, so the responses from the
20

survey are not linked to the PII. Third, screener data shall be considered part of the
survey data. The contractors will provide the results of the screener questions for all
panelists, regardless of whether they qualify for the study. However, the contractors will
not retain responses to screening questions for those who are deemed ineligible for any
other purpose outside the scope of this project. Fourth, the contractors will retain study
records for the duration of the study. Upon final delivery of data files to the contractors
and completion of the project, the contractors will destroy all study records including data
files upon request. The contractors will not be able to supply or access this information
for any reason, even at the request of FDA or the contractors, once destroyed. Finally,
data coming directly from the survey engine are stored in a proprietary database. While
this data is not encrypted, once inside the firewall, they are stored in a relational database
protected by several layers of intrusion detection and access control. Data files delivered
to the contractors will be sent via encrypted files.
All electronic data will be maintained in a manner that is consistent with the Department
of Health and Human Services ADP Systems Security Policy as described in DHHS ADP
Systems Manual, Part 6, chapters 6-30 and 6-35. All data will also be maintained in
consistency with the FDA Privacy Act System of Records #09-10-0009 (Special Studies
and Surveys on FDA Regulated Products).
11. Justification for Sensitive Questions
The survey and experimental study instruments do not include any questions that are of a
sensitive nature, such as sexual behavior and attitudes, religious beliefs, and other matters
that are commonly considered private.
12. Estimates of Annualized Burden Hours and Costs
12a. Annualized Hour Burden Estimate
FDA estimates the burden of this collection of information as follows:

Activity

Table 1. Estimated Annual Reporting Burden1
Number of
responses
Total annual Average burden
Number of
per
responses
per response
respondents
respondent

Study 1 (Survey)
Cognitive interview
screener

75

Study 2 (Experiment)
Cognitive interview
screener2

75

1

75

0.083

Total
hours
6

(5 minutes)
1

75

0.083
(5 minutes)

21

6

Study 1 (Survey)
Cognitive interview

5

Study 2 (Experiment)
Cognitive interview

9

Study 1 (Survey) Pretest

1

5

1

5

(60 minutes)
1

9

1

9

(60 minutes)
60

1

60

0.17

10

(10 minutes)
Study 2 (Experiment)
Pretest
Study 1 (Survey)

180

1

180

0.25

45

(15 minutes)
2,000

1

2,000

0.17

340

(10 minutes)
Study 2 (Experiment)

5,000

1

5,000

0.25

1,250

(15 minutes)
Total

1,671

There are no capital costs or operating and maintenance costs associated with this collection of
information.

1

12b. Annualized Cost Burden Estimate
The annualized cost to all participants for the hour burden for the collection of
information is $36,762. (1,671 x $22). The rate of $22 per hour is the 2021 median
wage rate in the U.S., rounded to the nearest dollar. 1
13. Estimates of Other Total Annual Costs to Participants and/or Recordkeepers/Capital
Costs
There are no capital, operating, or maintenance costs associated with this data collection.
14. Annualized Cost to the Federal Government
The estimated total cost to the Federal Government for this information collection is
$300,000. This includes the value of the task orders to develop and conduct the
collection of information and the value of a Full-Time-Employee to develop and monitor
the data collections and then analyze the results.
15. Explanation for Program Changes or Adjustments
This is a new data collection.
1

http://www.bls.gov/oes/current/oes_nat.htm, accessed December 2021.
22

16. Plans for Tabulation and Publication and Project Time Schedule
The Agency will use the study results to help inform the continued exploration of a
symbol for voluntary use on the food label to depict the nutrient content claim “Healthy.”
The Agency anticipates disseminating the results of the research after the final analyses
of the data are completed, reviewed, and cleared. Results of the research may be
summarized for publication in a peer-reviewed scientific journal. The planned schedule
for project activities is shown in Table 2.
Table 2. Project Schedule
Date
Within 3 days after
receipt of OMB
approval of collection
of information
Within 135 days after
notification to
contractors
Within 180 days after
notification to
contractors
Within 6 months after
receipt of final data
files
Within 18 months after
receipt of final data
files
Within 18 months after
receipt of final data
files
Within 24 months after
receipt of final data
files

Activity
Notification to the contractors to
proceed with data collection
activities

Audience
Not
applicable

Completion the experimental study
and both surveys

Not
applicable

Delivery by the contractors of
experimental study and survey final
data files
Delivery of oral and written
preliminary summaries

Not
applicable

Delivery of a written final report of
summaries and analytical findings

FDA

Response to information requests

FDA and
public

Submission of manuscript(s) of
journal article(s) to disseminate
information and analytical findings

Public

FDA

Activities associated with the outcomes of this research will primarily consist of written
and oral presentations as well as a written final report. In addition, journal manuscripts
and oral and/or poster presentations will be planned for disseminating information to the
public, including professionals, academics, and industry and consumer organizations.
The dialogues will help improve the effectiveness of the agency’s regulatory and
education initiatives in promoting and protecting the public health.

23

17. Reason(s) Display of OMB Expiration Date is Inappropriate
The OMB approval and expiration date will be displayed on all materials associated with
the study.
18. Exceptions to Certification for Paperwork Reduction Act Submissions
No exceptions are requested.
References
1. IOMa (Institute of Medicine). 2010 Examination of Front-of-Package Nutrition Rating
Systems and Symbols: Phase I Report. Washington, DC: The National Academies
Press.
2. IOMb (Institute of Medicine). 2012. Front-of-Package Nutrition Ratings Systems and
Symbols: Promoting Healthier Choices. Washington, DC: The National Academies
Press.
3. Podsakoff, P.M., MacKenzie, S.B., Podsakoff, N.P. 2012. Sources of method biases
in social science research and recommendations on how to control it. Annual Review
of Psychology, 63, pp. 539-569.
4. Sheff, J.N. 2011. Biasing brands. Cardozo Law Review, 32(4), pp. 1245-1314.
5. Cohen, J. 1992. A power primer. Psychol. Bull., 112 (1), pp. 155-159.
6. Baig, S.A., Noar, S.M., Gottfredson, N.C., Lazard, A.J., Ribisl, K.M., Brewer, N.T
2021. Incremental Criterion Validity of Message Perceptions and Effects Perceptions
in the Context of Anti-smoking Messages. Journal of Behavioral Medicine, 44, pp.
74-83.
7. Azjen, I. 2011. The theory of planned behaviour: Reactions and reflections.
Psychology and Health, 26(9), pp. 1113-1127.
8. Azjen, I. 2014. The theory of planned behaviour is alive and well, and not ready to
retire: a comentary on Sniehotta, Presseau, and Araujo-Soares. Healthy Psychology
Review, 9(2), pp. 131-137.
9. Ajzen, I. 2016. Consumer attitudes and behavior: The theory of planned behavior
applied to food consumption decisions. Italian Review of Agricultural Economics,
70(2), pp. 121–138.

24