Trends in Food Science & Technology 47 (2016) 82e92

Contents lists available at ScienceDirect

Trends in Food Science & Technology

journal homepage: http://www.journals.elsevier.com/trends-in-food-scienceand-technology

Commentary

Cool snacks: A cross-disciplinary approach to healthier snacks

for adolescents
Klaus G. Grunert a, *, Steen Brock b, Karen Bruns a, Tenna Christiansen c,
Merete Edelenbos d, Hanne Kastberg e, Stinne Gunder Strm Krogager f,
Line Holler Mielby d, Karen Klitgaard Povlsen g
Aarhus University, MAPP Centre, Bartholins All

e 10, DK-8000 Aarhus C, Denmark
Aarhus University, Department of Culture and Society, Jens Chr. Skous Vej 7, DK-8000 Aarhus C, Denmark
Langelandsgade 195, DK-8200 Aarhus N, Denmark
d
Aarhus University, Department of Food Science, Kirstinebjergvej 10, DK-5792 Aarslev, Denmark
e
Technological Institute, Gregersens Vej 1, DK-2630 Taastrup, Denmark
f
Aalborg University, Department of Communication, Nyhavnsgade 14, DK-9000 Aalborg, Denmark
g
Aarhus University, Department of Aesthetics and Communication, Helsingforsgade 14, DK-8200 Aarhus N, Denmark
a

b
c

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 16 June 2013
Received in revised form
6 October 2015
Accepted 11 October 2015
Available online 22 October 2015

Young people snack and their snacking habits are not always healthy. We address the questions whether
it is possible to develop a new snack product that adolescents will nd attractive, even though it is based
on ingredients as healthy as fruits and vegetables, and we argue that developing such a product requires
an interdisciplinary effort where researchers with backgrounds in psychology, anthropology, media
science, philosophy, sensory science and food science join forces. We present the COOL SNACKS project,
where such a blend of competences was used rst to obtain thorough insight into young people's
snacking behaviour and then to develop and test new, healthier snacking solutions. These new snacking
solutions were tested and found to be favourably accepted by young people. The paper therefore provides a proof of principle that the development of snacks that are both healthy and attractive to adolescents is possible if based on an interdisciplinary, concerted effort.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
Snacking, dened as food consumed between the three main
meals (de Graaf, 2006), is an important part of the daily life of
adolescents. Snacks are among the rst food products that adolescents buy with their own money for consumption outside a
family context (Brown, McIlveen, & Strugnell, 2000; Nicklaus,
Boggio, Chabanet, & Issanchou, 2004). Often snacks have a high
content of saturated fat, salt and rened sugar, and the public
debate suggests that adolescents tend to eat too much unhealthy
food resulting in negative consequences for public health such as
obesity and lifestyle-related diseases (Christensen, 2003; Fagt et al.,
2004). Public concern has given rise to numerous initiatives, mostly
using informational and educational tools, which aim at encouraging people to choose healthier alternatives. However, there is

* Corresponding author.
E-mail address: klg@mgmt.au.dk (K.G. Grunert).
http://dx.doi.org/10.1016/j.tifs.2015.10.009
0924-2244/ 2015 Elsevier Ltd. All rights reserved.

clearly also a role for new product development. If young people

prefer less healthy snack products, somehow the existing more
healthy alternatives seem to appear as less attractive to them. There
is no reason to believe that healthier alternatives are not chosen
just because they are healthier. It is thus worth asking whether it is
possible to develop a healthy snack product that adolescents will
nd attractive.
Developing such a product should be based on a thorough understanding of the target group, its snack preferences and the determinants of these preferences. The importance of consumer
insight in the new product development process is scientically
well-established (e.g., Im, Nakata, Park, & Ha, 2003), widely
acknowledged in the industry, and has led to a range of tools for
consumer-oriented new product development (Grunert et al.,
2008). Still, there have been no breakthrough successes in the
development of healthy snacks for adolescents. We believe this is
because of two major hurdles. First, the dominant approach of
analysing food preferences as a result of individual deliberation is
sters, 2009), but may be especially limited in
generally limited (Ko

K.G. Grunert et al. / Trends in Food Science & Technology 47 (2016) 82e92

the case of adolescents choosing snacks (Weijzen, de Graaf, &

Dijksterhuis, 2008). It is now widely accepted that food choice is
to a large extent governed by habits and routines (van't Riet,
Sijtsema, Dagevos, & De Bruijn, 2011) and that attempts to
change behaviours therefore need to be analysed not only based on
an analysis of the deliberate formation of preferences, but also on
an understanding of how the environment in which such choices
take place inuence decisions (Dolan et al., 2012). There is also
reason to believe that snack choices of adolescents are heavily
inuenced by social interaction in peer groups (Nrgaard, Hansen,
& Grunert, 2013), a phenomenon that, while acknowledged, has
received only limited attention in research on consumer food
choice (Kuenzel & Musters, 2007). Second, even the most careful
consumer insight study does not give clear directions as to how the
new product should look like. It needs to be translated into a
technologically feasible product specication, and earlier attempts
to formalize this process have had only limited success (Benner,
Linnemann, Jongen, & Folstar, 2003). The necessary interplay of
different disciplines in new product development has received
considerable attention in the innovation management literature
(Jacobsen et al., 2014), but there are few documented cases on how
these different disciplines indeed can play together in successful
product development.
In this paper we present the COOL SNACKS project, which was
an attempt to develop a healthful snack solution based on fresh
fruit and vegetables perceived as attractive by 10e16-year old adolescents. The project was based on the assumption that developing such a product is possible if it is based on an understanding of
adolescent snack choices that looks not only at individual preferences but also at daily routines and choice/consumption environments, and if the insight thus generated is turned into a physical

83

product in a transparent and systematic way where competencies

enabling customer insight and competencies enabling food production interact. The point of departure of the COOL SNACKS
project was therefore that such a development process is an
interdisciplinary task, i.e. a task combining insight from the social
sciences and humanities to obtain a deep understanding of adolescents' existing snacking behaviour, and then supplementing this
understanding with insight from food processing technology and
logistics to arrive at a solution that is both technically feasible and
attractive in the minds of adolescents. The project aims therefore to
demonstrate that the development of such a product is possible if
the right set of competencies is combined in the process. Fig. 1
summarizes the COOL SNACKS project work ow.
The rest of the paper follows the structure of Fig. 1. We rst
report a series of studies that had the aim of understanding adolescents' current snacking behaviour, leading to a set of requirements that a healthy, fresh fruit and vegetable-based snack
would need to full for adolescents to regard them as attractive. We
then describe how this set of requirements was transformed rst
into product concepts and then into physical prototypes, taking into
account a number of technological and logistical constraints. Then
we show how the COOL SNACK solutions were tested in a realworld school setting. We close with perspectives for user-driven
development of healthy food products.
2. Putting snacking behaviour in context
In order to successfully develop new healthy snacking solutions,
we formed an understanding of adolescents' current snacking
behaviour through a range of mutually complementary studies: an
analysis of adolescents' individual choices of existing snack

Fig. 1. Work ow in the COOL SNACKS project.

K.G. Grunert et al. / Trends in Food Science & Technology 47 (2016) 82e92

84

products, snacking habits and media use, and snacking behaviour

as part of general school routines.
2.1. Choice of existing snack products
A quantitative study of adolescents' (11-16 year-olds) individual
choices of existing snack products was conducted in a school
setting (in four urban and four rural schools in Denmark, for details
see Mielby, Edelenbos, & Thybo, 2012). Twenty-one unwrapped
snack products available on the Danish market were presented
unwrapped to participants, as we were not interested in the effects
of packaging or branding information. The study thus provides
information on snack preferences as revealed in snack choices.
The snacks covered seven categories:

Sweets (wine gum, liquorice, chocolate)

Crisps (potato crisps, popcorn, salted peanuts)
Nuts and dried fruit (mixed, unsalted nuts, mixed dried fruit)
Baked sweet snacks (mufn, chocolate chip cookie, rum ball)
Baked savoury snacks (mini pizza, sausage roll)
Fresh fruit (grapes, banana, apple, pear)
Fresh vegetables (mini carrot, mini cucumber, cherry tomato,
sugar pea)

These snacks were picked based on interviews with adolescents

about their perception of snack products (what is a snack?) and
from screening for available snack products in Danish supermarkets. Snack choice data was collected for 387 adolescents. Background data on the adolescents such as age, gender, level of hunger
was also collected.
The results of the study showed that grapes, rum balls, mini
pizzas and apples were chosen most frequently (7e18.6% of all
choices made), whereas salted peanuts, mixed nuts and popcorn
were chosen less frequently (.8e1.8% of all choices made) in this
school setting (Fig. 2). The fact that adolescents chose a combination of relatively healthy (grapes and apples) and unhealthy snacks
(rum ball and mini pizza) is encouraging considering adolescents'
general snack consumption (Piernas & Popkin, 2010; Savige,
MacFarlane, Ball, Worsley, & Crawford, 2007). The results also

indicate that no single product dominates the choices made, and

that any new product with wide appeal therefore will be difcult to
construct around one main ingredient.
We also investigated whether different segments of adolescents
choose different types of snacks, as consumers are not all alike
ster, 2003). Fig. 3 depicts a Principal Component
(Grunert, 2002; Ko
Analysis (PCA) plot of snack choices and their relation with background variables.
As can be seen, there was a gender difference in the choice of
snack types. Boys chose the more lling, baked sweet and savoury
snacks, whereas girls chose fresh fruit snacks. This is in agreement
with existing literature on adolescents' preference for food
including fresh fruit and vegetables (Cooke & Wardle, 2005; Le
Bigot Macaux, 2001; Nu, MacLeod, & Barthelemy, 2007; Reynolds,
Hinton, Shewchuk, & Hickey, 1999). Less clear tendencies were
seen with regard to urbanity, while no major relationship was
found for the adolescents' age. The full details of the study are reported in Mielby, Edelenbos, and Thybo (2012).
In summary, adolescents chose both unhealthy and healthy
snacks, and choices are gender specic. Boys chose more lling
while girls chose less lling snacks. Gender specic choices and
preferences were also found in other studies exploring adolescents'
preferences for fruit and vegetable snacks (Mielby, Jensen,
Edelenbos, & Thybo, 2013; Mielby, Kildegaard, Gabrielsen,
Edelenbos, & Thybo, 2012; Mielby, Nrgaard, Edelenbos, & Thybo,
2012).
2.2. Snacking, food habits and media use
The study of snacking, food habits and media use consisted of
twelve focus groups and ten individual in-depth interviews in the
4th and 9th grades (10e11 and 15-16-year-olds) at various schools
in mainland Denmark (for details, see Krogager, 2012). Only three
focus groups included adolescents of both genders, the other ones
were either all boys or girls. The main purpose of the interviews
was to get the adolescents to talk about their daily routines as
regards food consumption and media use, to discuss food and
media preferences, and to relate this to snacking behaviour. The
focus group interviews were structured around creative and

Fig. 2. Adolescents' snack choice frequency (%) based on 21 snack products available on the Danish market. Modied from Mielby, Edelenbos, et al., 2012.

K.G. Grunert et al. / Trends in Food Science & Technology 47 (2016) 82e92

85

Fig. 3. PCA plot of adolescents' snack choice frequency among 21 snack products available on the Danish market and their relation with background variables. Modied from
Mielby, Edelenbos, et al., 2012.

practical elements and competences. First participants were asked

to report on their food intake yesterday and on their media use
yesterday and then to talk about their food and media preferences
in general. In this process, adolescents were asked to make media
co-productions (Thomson, 2008): Collaborative visual works using
media representations of food. The younger participants (10-11year-olds) made collages of food pictures from a sample of magazines, yers and catalogues. Collages are well-suited to supplement
verbal data collection techniques with media-related visual elements (Buckingham, 2009) and are regarded as well suited to
investigate the more hedonic aspects of consumption (Belk, Ger, &
Askegaard, 2003). The older group (15-16-year-olds) made media
campaigns in support of and against single food items such as
chicken or hamburgers. Each focus group was divided into two
smaller groups that had to agree on a food item and then produce
slogans for a) a healthier eating campaign and b) commercial advertisements. This part of the study showed that adolescents'
knowledge on healthy and unhealthy food items is indeed broad
and that they are well-informed media users that are able to mimic
media campaigns and even make humorous and creative
embellishments.
As a last component of this study, adolescents were allowed to
snack while talking and doing their creative work. On the table
there were four bowls with fresh vegetables, fresh berries, cinnamon buns and sweets, representing more (vegetables and berries)
and less healthy (buns and sweets) snacks. The bowls were not
introduced to the adolescents, but when boys or if girls asked if
they were allowed to try some, they were invited to do so. Some of
the girls never asked if they could try some and we had to introduce
the snacks.
The design strived to prioritize saying as well as doing in an
attempt to capture both adolescents' attitudes to media and food, as
expressed in their verbal expressions, and their actual practices in
connection with media and food, as expressed by the action-based
parts of the design, i.e., collages and media campaigns (Reckwitz,
2002; Schatzki, 1996; Schatzki, Knorr-Cetina, & von Savigny,
2001; Warde, 2005).

The results of the study suggested that a collective conclusion

about adolescents as food consumers and media users is impossible. Adolescents are as diverse as adults in terms of food and
media use and preferences. Gender turned out to be a major
discriminator with regard to daily media routines and preferences
and everyday food consumption and preferences. It seemed that
there are only a few common denominators in boys' and girls'
media use as regards preferences for digital media, TV and print
media. Not surprisingly, digital media played a central role in adolescents' everyday routines. However, girls and boys seemed to
use these media for different purposes: Girls focused on social
networks, Facebook, mobile phones, Skype and other similar services and boys used digital media for game playing.
When it comes to food preferences, boys and girls crave
different things. Boys love meat of every kind, and meat sometimes
worked as a social stimulus in the conversation. The collages produced by 4th grade boys focused on big roasts and sausages. In
contrast, girls talked about fresh fruit and berries and cakes; in our
study, girls often made collages with aesthetic representations of
fresh berries, sweets and decorated cakes. Most boys and all girls
demonstrated diverse competences related to food, taste, quality
and health as they were cutting and pasting their collages. Boys
related especially to pictures of recognizable and real food, whereas
girls also appreciated visual aesthetics and were often competent
producers of aesthetic collages.
The campaigns produced by 9th graders negotiated a wide range
of mediated discourses about food. The discussions that unfolded
when the adolescents had to agree on a food item showed that they
were competent users and negotiators of normative public discourses and media debates about food quality, nutrition awareness,
and animal welfare. They would mimic well-known slogans with a
twist of humour and irony. Adolescents had explicit and very
detailed knowledge of nutrition and healthy food, but were mainly
interested in experiments with alcohol and traditional snacks. They
reproduced normative views in the group, but talked about
breaking the rules outside school, and they had intimate knowledge of media discourses and advertisements of alcoholic

86

K.G. Grunert et al. / Trends in Food Science & Technology 47 (2016) 82e92

beverages and fast food products. They seemed to prefer unhealthy

food and media products relating to them, but with a twist indicating their awareness that they are out of bounds. The actual
snacking practices were also structured around routinely accepted
normative presuppositions; all groups ate the fresh vegetables rst,
then the berries, thirdly cinnamon buns and lastly the sweets. This
practice applied to both boys and girls across ages, and it was articulated as the snacking played out. All groups talked about the
candy as the most desirable snack e but all adolescents ate the
healthier fresh fruit and vegetable snacks rst and the sweets at the
end.
The study showed that the daily media routines and food consumption patterns differed according to gender and age. Across age,
girls and boys had different preferences for food and they
demonstrated a large set of competences in relation to articulating
their desires. At the same time, they were disciplined and able to
postpone and negotiate own desires socially e at least in this school
context.
2.3. Snacking behaviour and daily school routines
The relationship between snacking behaviour and general
school routines was investigated at four different public schools
and one private school, as well as in 38 stores and food outlets in
the vicinity of these schools. The schools were selected in order to
get a good coverage of regional differences in Denmark. Adolescents were interviewed in fteen small focus groups and ten large
groups in class. A questionnaire, which mapped both adolescents'
actual consumption during the day and their preferences if they
could choose freely (i.e., unconstrained by availability and price),
was answered by 180 participants. We also observed eating practices in the class room (where much of the eating during school
breaks takes place in Danish schools), in the school canteens, at
food outlets and stores, and nally we conducted interviews with
staff at the schools and in the stores and food outlets.
The resulting data was used to draw a broad picture of adolescents' daily routines and social competences in relation to their
food consumption. It soon became clear that there still is what
might be called a lunch box culture in Danish schools, which
means that during the school day most adolescents have access to a
lunch box, which is prepared at home, usually by the parents. Such
a lunch box usually contains a variety of open sandwiches, pieces of
fresh fruit, biscuits, snacks, and sometimes fresh salads. As a
consequence of the Danish lunch culture, consumption of food and
snacks in schools from the moment adolescents leave home and
until they return home, has a characteristic temporal structure.
Adolescents may supplement what is in their lunch box in different
ways throughout the day in terms of what they purchase, from
which outlet and who they shop with.
After the rst series of observations and interviews we began
looking for adolescent behaviour in terms of the temptations they
are offered at school, by stores and by outlets in the vicinity of the
schools and how they give expression to these temptations. Thus
we observed and recorded the snacking behaviour of adolescents
both as a supplement to their lunch boxes and as socially competent engagement with certain groups, spaces, and places.
The ve schools visited were very different in terms of geography, social and ethnic factors. However, the general pattern we
found was that at any school, if given the choice and option, adolescents would spend between DKK 30e50 daily on snacking. The
type of snacks would be wide-ranging. No particular kinds of
sweets, cakes, or soft drinks dominated. If anything, it was the
actual invitations and offers from the various stores, outlets,
and canteens that inuence choices. Choice of snack was clearly
part of the way in which adolescents positioned themselves in a

, 1991). First, there was the capability of

number of ways (Harre
appropriating physical places, such as places to sit and eat, thus
realizing a kind of ownership towards these places. Second, there
was the capability of demonstrating and manifesting social relations to staff, friends, and e surprisingly e a variety of family
bonds (for more details see Brock & Kayser-Nielsen, 2009). Third,
there was the capability of expressing personal interests and
preferences facing both peer groups and family. Fourth, there was
the capability of standing as a unique individual, where the individual adolescent's self-concerns were displayed bravely or
shamefully without challenging the social dimensions of a given
snacking situations.
In relation to all four kinds of capabilities, a quality assessment
of the food and snack products in terms of convenience, taste,
texture, wrapping, size, and ability to t certain forms of mobility
was crucial for the adolescents. Particularly we recorded a marked
interest in products characterized by varying layers or elements
each possessing different recognizable tastes and textures. Also,
products with a sprinkled surface were popular.
In terms of perspectives for new and healthy snack solutions,
the results of this investigation thus point at a plastic snacking
product in the sense that it would allow adolescents to display their
social competences. Accordingly, purchase and consumption of a
snack should involve some kind of personal touch and nish, and
the product should be shareable. Also it should suit the mobility
within and across social places and spaces. Finally, the product
should meet the requirements of adolescents in relation to texture
and assessment of quality.
2.4. Key requirements for COOL SNACKS
Taken together, the three studies led to the formulation of the
following key requirements concerning the development of ideas
and concepts for new snacking solutions targeted at adolescents in
a school context. The key requirements were divided into three
preferences, three routines and one competence:
Preferences
P1. Both healthy and unhealthy
P2. Gender-specic preferences as regards taste and energy
density
P3. Recognizable taste and texture variations
Routines
R1. Lunch box supplement
R2. Temporal structure
R3. Socializing and sharing
Competences
C1. Continuous negotiation and appropriation
3. Development of new snacking solutions
Development of new snacking solutions consisted of an iterative
creative idea and concept generation process followed by concept
testing and prototype development. The ideas and concepts drew
on the outcomes of the cross-disciplinary research summarized in
the key requirements and researcher-industry partner collaboration to generate both technically feasible and e for the target group
e attractive solutions.
3.1. Idea and concept generation
Researchers and industry partners brainstormed in groups to
generate snack ideas and concepts taking the key COOL SNACKS
requirements into account. The ideas and concepts were then
presented and discussed in a plenary session resulting in a total of

K.G. Grunert et al. / Trends in Food Science & Technology 47 (2016) 82e92

13 snack ideas and concepts. One of the ideas was The Shaker e a
beaker with premixed fresh fruit or vegetables with a free choice of
sprinkle which had to be shaken into the beaker before consumption. Another idea was Two-in-one e a combination of something
healthy and something unhealthy and self-indulgent such as apple
slices combined with a dip. Three of the snack ideas contained
solely fruit, either fresh or frozen. One idea contained stir-fried
vegetables, and the remaining nine snack ideas combined fresh
fruit and vegetables with one or more food components such as
bread, dip and/or sprinkle. Five ideas and concepts involved a
personal touch in the composition of the snack which satised
adolescents' demand for individuality and also socializing potential
for sharing the snack with peers. This work yielded a set of recommendations providing for the key requirements for COOL
SNACKS (codes in parentheses refer to the key requirements):

Satisfying a broad target group of adolescents with various demands: Combo snack solutions consisting of various snack
products (P3, R2, R3, C1)

Individuality and interaction: Personal touch and nish in the
composition (R3, C1)

Unstructured meal settings in a social peer-group context: Social
events including sharing experience (R1, R2, R3, C1)

Balanced energy density and healthiness: Various snack products including bread, dip and sprinkle (P1, P2, P3)

Attractiveness in a novel and familiar way: Various tastes and
textures (P1, C1)

Fresh and ready to eat on the go (R1, R2)

Constant novelty to maintain loyalty (C1)

87

boys requirement for energy. The dip and sprinkle components

were included to add texture and taste variety to the snack solution
and to give it a personal touch.
The COOL SNACK concept was tested in several variations by
means of an online survey (for details, see Nrgaard, Srensen, &
Bruns, 2014). The aim of this test was to uncover whether the
concept and its variations were acceptable to adolescents and thus
a feasible steppingstone for further prototype development. A webbased, at-home survey including 163 adolescents (10e16 years old,
50% girls and 50% boys) was applied. Informed consent was obtained from both parents and adolescents and parents were
instructed that they could help their child with technical aspects of
lling out the questionnaire, but should not inuence their answers. Previous research has found that from about 12 years adolescents are capable of handling most common survey techniques

on, Couronne, Marcuz, & Ko
ster, 1999), and
(Guinard, 2000; Le
within the food area successful use of self-administered questionnaires among adolescents aged 8 has been reported (e.g.,
Nrgaard, Bruns, Christensen, & Mikkelsen, 2007; Mielby,
Kildegaard, et al., 2012). In the test each variation of the concept
was presented by means of headings and small descriptions combined with visuals of the eating context. Two of the COOL SNACK
concepts are shown in Fig. 4.
The results showed that all eight COOL SNACK concept varieties
were positively evaluated by the adolescents. Buying intentions
ranged from 4.24 to 4.98 (1 denitely not buy, 7 denitely buy)
and liking from 4.79 to 5.60 (1 don't like at all, 7 like very
much) on a seven-point smiley scale. The COOL SNACK concept and
its variations were thus used in the further prototype development
process.

3.2. Concept development and testing

3.3. Prototype development

From the idea and concept generation process and the set of
recommendations, the researchers formed one single COOL SNACK
concept with eight variety options. The concept consisted of
different options regarding the type and number of components.
The components included: fresh fruit, fresh vegetables, bread, dip
and sprinkle. It was understood that fresh fruit and vegetables must
be used and that all components were ready-to-eat. The fresh fruit
and vegetable components were chosen to balance energy density
and healthiness while the bread component was chosen to full

The adolescents' key requirements, the recommendations from

the idea and concept generation process, and the results of the
concept test resulted in four COOL SNACK solution prototypes: two
solo snack box solution prototypes and two share snack box solution prototypes. Each prototype consisted of different numbers of
components (i.e. fresh fruit, fresh vegetables, bread, dip and
sprinkle) and types of components in each solution. For the
development of the four COOL SNACK prototype solutions, adolescents' key requirements were rst translated into new snacking

Fig. 4. Two variations of the COOL SNACK concept tested in the concept test.

K.G. Grunert et al. / Trends in Food Science & Technology 47 (2016) 82e92

88

Table 1
Translation of adolescents' key requirements into new snacking solutions.
Adolescents' key requirements
Preferences
P1. Healthy and unhealthy
P2. Gender specic preferences in taste and energy density
P3. Recognizable variations in tastes and textures
Routines
R1. Supplement to lunch box
R2. Temporal structure

R3. Socializing and sharing

Competencies
C1. Continuous negotiation and appropriation

solution features (Table 1) and then further into prototype components and elements. A total of 19 elements were developed
(Fig. 5) considering technological constraints and practical implications in terms of production, supply chain and display. The fresh
fruit component consisted of four elements (white grapes, pineapple mini sticks, apple wedges, apple wedges and blueberries);
the fresh vegetable component consisted of four elements (mini
cucumbers, red bell pepper chunks, orange and white mini carrots,
orange mini carrots); the bread component consisted of four elements (mini whole wheat loaf with rye, mini whole wheat loaf with
dried fruit, rye bread sticks with pizza avouring, bread sticks with
sour cream and onion avouring); the dip component of three elements (sour cream-based holiday dip, mild salsa, caramel); and
the sprinkle component of four elements (liquorice, chopped, unsalted peanuts, bacon crisps, pop rocks).
All elements were ready-to-eat to full the requirements of a
temporal structure in snacking, and fresh fruit and vegetables had a
shelf-life of at least 5 days at 5  C. The elements were consumable
on-the-go or could be taken on-the-go as they were individually
packaged in transparent, easy to open and re-sealable Minigrip
zipper pouches tted in a lunchbox-type carton. Also dips and
sprinkles could be taken on-the-go as they came in transparent
beakers with re-sealable transparent lids.
Overall, the elements varied in healthiness, taste and texture,
familiarity, energy content per serving, size of serving and number

Features for new snacking solutions

Both healthy and unhealthy elements
Both low and high energy content
Both berries and meat elements
Both sweet and savoury tastes
Juicy, crisp and soft textures
A snack versus a full meal
Ready-to-eat elements
Re-sealable packaging of elements
Elements in a box
Elements can be eaten on-the-go
Two sizes of the snack box
Several elements in a snack
Many pieces of each element
Many ways of mixing elements
Combination of several elements
Familiar and novel elements

of pieces in a serving (Table 2). The fresh fruit and vegetables and
mini loaves were considered healthy elements, as they lived up to
standards of the Danish keyhole, which is a health logo sponsored
by the Nordic governments. Grapes, pineapples, apples, blueberries, mini cucumbers, bell peppers and mini carrots were chosen
to cover different tastes, textures and familiarities (Table 2). Girls'
preferences for berries were met by mixing blueberries with apples. Mini bunches of grapes and mini cucumbers and carrots were
introduced to increase the shelf-life as compared to the fresh-cut
versions. Orange and white mini carrots were mixed to increase
novelty.
All fresh fruit, vegetable and bread elements could be eaten
directly or shared after breaking. The viscosity of the dips at 5  C
and the lumping of the sprinkle were technological constraints that
made us reject several elements in the prototype development
phase. To meet boys' requirement for a savoury taste, bread sticks,
salsa and bacon crisps were included in the COOL SNACK solution.
The bread sticks were spiced with potato crisp avourings and
produced exclusively for the COOL SNACKS project. Caramel dip
and liquorice and Pop Rocks sprinkle were included to meet girls'
requirement for sweet taste. Overall, the dip and sprinkle elements
were introduced to full adolescents' requirements for taste variety
and familiarity and to increase the appropriation or coolness of the
prototypes.
The four prototypes were prepared for testing in self-mix and
pre-mix versions. In the self-mix version, all elements in Table 2
were used. For the pre-mix version, the selected elements are
shown in Table 2. In both versions, snack elements were gathered in
a printed six-point glued carton box with a window foil and
separate inserts to full the practical requirement of keeping elements together at purchase, during consumption and on-the-go.
The box was the result of a joint effort by an advertising agency
and the research group. Green was the main colour, avoiding blue
(boys) and pink (girls) suggesting a specic gender. Pictures of
teenage boys and girls drawn in a style suggesting references to
cartoons and digital game plays were used. This teenage style was
chosen because younger children identify upwards in age (Marcia,
2001; Popper & Kroll, 2005).
4. New snacking solutions in context
4.1. Prototype testing in schools

Fig. 5. Nineteen snacking elements placed in a display cooler for self-mixing in the
prototype test at schools.

The prototypes were tested in two rounds with adolescents

K.G. Grunert et al. / Trends in Food Science & Technology 47 (2016) 82e92

89

Table 2
Description of snacking components and elements for new snacking solutions.
Snacking components and elements

Sensory description Familiarity

Components and elements
selected for pre-mix test of new
snacking solutions
Solo 1 Solo 2 Share 1 Share 2 Taste

Fruit
White grapes
Pineapple mini sticks
Apple wedges
Apple wedges & blueberries
Vegetables
Mini cucumber
Red bell pepper slices
Orange & white mini carrots
Orange mini carrots
Bread
Mini whole wheat loaf w rye
Mini whole wheat loaf w dried fruit
Rye bread sticks w pizza avouring
Bread sticks w sour cream & onion avouring
Dips
Sour cream based Holiday
Mild salsa
Caramel
Sprinkle
Liquorice
Chopped, unsalted peanuts
Bacon crisps
Pop rocks
Energy of snacking solution, KJ/snack

X
X

X
X

X
X

694

682

971

1540

(aged 10e16 years old) in three municipal primary and secondary

schools in Denmark (two of these were also part of the study on
Snacking behaviour and daily school routines, see above).
In the rst round, the adolescents were introduced to the 19
snack box elements (Fig. 5) and were asked to select elements for
two different COOL SNACK solutions e either for a solo snack box,
consisting of fresh fruit or fresh vegetables, bread, and dip or
sprinkle, or for a share snack box, consisting of fresh fruit, fresh
vegetables, bread, dip and sprinkle. After having made their selection, participants were asked to rate their selection on a sevenpoint scale. Participants were then asked to taste the snack, rst
with an emphasis on the combination of the selected elements, and
then with an emphasis on the individual elements, and both overall
liking and liking for individual elements were measured. Finally,
purchase intention was measured for three situations (buying
alone/with best friend/with peers) using a 7-point scale
(1 denitely not buy, 7 denitely buy). A number of background variables (health consciousness, food neophobia, social
interaction with peers) were also measured. 546 adolescents
participated in this rst test. The purpose of this test was to
generate data that can be used in decisions about the composition
of the two nal prototypes, one for solo and one for the share snack
box.
An analysis of the choices made shows that all snack elements
were not equally likely to be chosen (Fig. 6). For the solo snack box,
there were differences between girls and boys: for the fruit
element, girls were more likely to choose grapes, while boys were
more likely to choose apples, and for the dips category, girls were
more likely to choose the holiday dip, while boys were more likely
to choose salsa and caramel. For the shared snack box, the only
signicant gender difference was choice of bacon sprinkle (which
boys were more likely to choose). No particular combination came
out as more popular than others. In the solo snack solution, out of
258 participants with complete data, no particular combination
was chosen by more than seven participants (2.7%). Likewise, for

Energy [KJ/serv.] Size of serv. [G] No. of pieces

Texture

Sweet
Sweet
Sweet
Sweet

Soft
Juicy
Crisp
Crisp/soft

Familiar
Familiar
Familiar
Novel/familiar

235
186
173
155

90
80
90
80

20
8
5
11

Sweet
Sweet
Sweet

Juicy/crisp
Crisp
Crisp
Crisp

Familiar
50
Familiar
89
Novel/familiar 162
Familiarity
162

110
70
100
100

4
7
7
7

Savoury
Sweet
Savoury
Savoury

Soft
Soft
Crisp
Crisp

Familiar
Novel/familiar
Novel
Novel

686
700
328
332

70
70
20
20

1
1
25
25

Savoury Soft
Savoury Soft
Sweet
Soft

Familiar
Familiar
Familiar

192
40
280

20
20
20

20 ml
20 ml
20 ml

Sweet
Savoury
Savoury
Sweet

Novel
Familiar
Novel
Novel

119
177
151
113

7
7
7
7

Many
Many
Many
Many

Crisp
Crisp
Crisp
Crisp

the share snack solution, out of 237 participants with complete

data, no particular combination was chosen by more than ve
participants (2.3%). As there are clear differences in the likelihood
of any particular element to be chosen, this considerable variation
cannot be attributed to random choices alone, but indicates
considerable dispersion regarding preferences for variations in
taste and texture. The results also suggest that the elements chosen
were able to satisfy both boys and girls.
In the second round, adolescents were introduced to pre-mixed
snack box solutions. This second round of testing was carried out at
the same schools with a total of 635 adolescents. Two versions of
the solo box and two versions of the share box were prepared (see
Table 2) by combining those snacking elements that were chosen
most frequently in the rst round of testing. All participants got one
of these boxes randomly assigned. Again liking was rst measured
based on the visual impression of the product. Participants then
disassembled the box and tasted the elements, followed by measures of liking and purchase intention paralleling those of the rst
round of testing. The background measures were unchanged.
Fig. 7 allows comparing liking and buying intention ratings of
the concept test and the two rounds of prototype testing in schools,
where the prototype tests additionally allow distinguishing liking
based on visual impression only and liking after tasting. Both are
important, as liking based on visual impression will be linked to
initial purchase, whereas liking after tasting will be linked to
possible repeat purchase. Four conclusions can be drawn from
Fig. 7. First, liking and buying intention ratings are higher for the
prototype tests than for the concept test, indicating that the prototype development process succeeded in developing the concept
into elements that appealed to the adolescents group. Second, for
the pre-mixed boxes, likings based on the visual impression are
lower than likings when participants selected elements themselves, showing that the COOL SNACK solutions lose some of their
attraction when snack boxes are pre-mixed, even when the window foil allows (partial) visual inspection of the content.

90

K.G. Grunert et al. / Trends in Food Science & Technology 47 (2016) 82e92

Fig. 7. Liking and buying intentions in concept and prototype tests.

higher buying intention (mean diff .46, t 2.82, p .00).

Together, these results show that the development of new
snacking prototypes, by their combination of elements with
different tastes, textures, and by including less healthful elements
with healthful main elements, were successful in appealing to both
boys and girls.
4.2. New snacking solutions in school context

Fig. 6. Choice frequencies of snacking elements.

Third, liking after tasting is at the same level for both testing
conditions, indicating that the loss of liking due to pre-mixing of
certain snacking elements is, compared to self-mixing, limited.
Finally, we see that liking after tasting is at par with or exceeds the
level of liking based on visual inspection, suggesting that the expectations generated by visual inspection are met in the actual
tasting experience.
There were some slight gender differences. In the self-mix situation, boys had higher visual liking than girls (mean diff .27,
t 2.10, p .03), whereas girls had higher buying intention (mean
diff .37, t 2.06, p .04). There was no difference in liking after
tasting. In the pre-mix condition, there was no difference between
boys and girls in visual liking or liking after tasting, but girls had a

In analysing preferences for the new snacking solutions, the

research presented at the beginning of this paper underlines the
importance of social context. Snack preferences relate to a social
context in which the individual nds its bearings and navigates in
relation to the environment, social relations, time of the day, norms
and values (Bech-Larsen, Jensen, & Pedersen, 2010). In order to gain
insight into how the snack box was accepted or rejected by adolescents, the snack box was put on the market for sale in three
municipal primary and secondary schools in Denmark in two
rounds. A three-week ethnographic eld study using participant
observation (Spradley, 1980) was conducted at the schools by a
team of four anthropologists. The data was further explored by nine
focus group interviews at the schools, which further illuminated
themes from the eld work.
The snack boxes sold at the school canteen turned out to be
popular, mainly among the adolescents aged 12e13 years. The
success was partly due to the breaking of snacking routines which
mostly consisted of a variety of bread sold from the canteen. We
further observed that the adolescents were enthusiastic about the
high quality of the fresh fruit and vegetable elements, the different
combinations of snack elements and especially the social aspect of
the snack box solution in terms of the small, bite size pieces of fresh
fruit, vegetables, and bread sticks, which encouraged the adolescents to share the snack box with peers.
In terms of price and size, the snack box solution was based on
the assumption that adolescents would share the cost as well as the
snack box elements, which turned out to be a highly complex
phenomenon. The adolescents and especially the girls considered
food a very social matter rather than as a means to full a physiological need. In the social context, the complexity of social games
played out was reected in the use of the snack box. By offering
pieces of snack elements to their close friends, adolescents
conrmed their social relations and by refraining from offering or
rejecting pieces offered, their distance. In order to use the snack
box in a social game, they preferred to buy the snack box themselves, as sharing the cost and the snack box would prevent access

K.G. Grunert et al. / Trends in Food Science & Technology 47 (2016) 82e92

to this social game (Christiansen, Lvschal-Nielsen, & Nielsen,

2012). Sharing food in the company of others as a social activity
contributed to a sense of group belonging (Husby, Hietmann, &
O'Doherty Jensen, 2006) which seemed to be very important for
the girls.
Like the girls, the boys preferred to buy the snack box individually, but their use of the snack box differed considerably as they
did not tend to share their snack box. The boys wanted to eat the
snack themselves.
At the schools, the content of the snack box was discussed and
explored and individual preferences came into play. There were no
noticeable gender differences in preferences, but especially individual preferences were at stake. To a great extent individual
preferences were related to and inuenced by the social context. As
the preference for a specic snack box solution developed in a
specic social context, the individual preferences were subject to
change according to social relations, location, activities, time of day,
assortment of snacks, norms and values. The pre-mixed snack box
did not fully full individual and changing preferences in a social
context, which made some adolescents reject the snack box
(Christiansen et al., 2012).
In summary, the outcome of the ethnographic eldwork showed
that, in a social school context, the snack box was well received
with interest and excitement. The shareability of the snack box was
valued, especially by the girls. The pre-mixed snack boxes lacked
exibility and could not full adolescents' changing and individual
preferences. Individuality played an important role in being an
adolescent and in their social relationships, which was mirrored in
the different ways in which the aspect of sharing entered the
snacking behaviour of boys and girls.
5. Conclusions
In the project described in this paper, we set out to develop new
healthful snacking solutions that adolescents would perceive as
attractive e we believe that this aim was achieved. Based on the
comprehensive insight into adolescents' snacking behaviour obtained, we derived seven key requirements for such a snacking
solution: healthy and unhealthy features, due regard for genderspecic preferences in taste and energy density, recognizable
taste and texture varieties, usability as supplement to a lunch box,
tting temporal structure of daily organization of snacking behaviour, usability for socializing and sharing, and invitation to
continuous negotiation and appropriation. We believe that the
solution developed lives up to these requirements. Mixing recognizable product elements fullled the requirements for familiarity,
but also for individuality and continuous negotiation and appropriation. Mixing healthy with less healthy elements and including
fresh fruit specically for girls and of savoury products specically
for boys increased the attractiveness of the snacking solutions.
Combining several elements in one solution makes the concept
exible and suitable to be used in real life situations, with changing
seasons and availability of fresh fruit and vegetables and corresponding technical constraints in production and logistics.
The process leading up to the solution was characterized by a
high degree of interdisciplinarity and cross-disciplinary communication. Both in the rst and the last phases of the project, we
combined a range of different methodologies, both qualitative and
quantitative, in order rst to obtain insight into adolescents'
snacking behaviour and second to investigate how the new solution
tted into adolescents' snacking behaviour patterns. We drew on
concepts from areas as diverse as psychology, anthrropology, media
science, sensory science and food science in the design and
execution of the various studies. The challenging task of transforming the insight obtained in the rst phase to new snacking

91

ideas and concepts and subsequently to physical prototypes was

made possible by extensive cross-disciplinary communication between project members with competences in the social sciences
and humanities, partners with food science competence and industry partners.
The methodological protocols developed and used in the various
studies that formed part of the project are codied and ready for
future use. However, we believe that especially the process in
which we combined diverse sets of skills and competencies to
achieve the innovation goal can be a model for similar ventures in
the future.
We view this project as a proof of principle. We have shown that
it is possible to develop a snack for adolescents that is based on
healthy ingredients and that adolescents accept as part of their daily
school routine. We do not claim that the solution developed here
would have appeal beyond the context in which and for which it
was developed, or that any specic results on preferences for
various aspects of the tested COOL SNACK solution necessarily
generalizes. But we believe that we have demonstrated the value of
interdisciplinary cooperation when it comes to nding healthier
snacking solutions for young people. User-driven innovation is by
no means a new concept in the food industry (Grunert et al., 2008),
and consumer-driven product innovation is by denition interdisciplinary, at it requires both competence in analysing consumers and in developing and producing new products. However,
in spite of massive attention to the need for collaboration across
functions and disciplines (Jacobsen et al., 2014), there are few
documented cases showing how different disciplines can play
together in the successful development of new products. We believe
that our case documents the promise of feeding psychological, sociological, sensory and technological competence into a joint project aimed at developing a new product that appeals to the target
group and is also endorsable from a public health perspective.
Acknowledgements
This project was supported by a grant from the Danish Council
for Strategic Research grant no. 0603-00197B. We are also indebted
to a number of commercial partners who participated throughout
the project through their advice and opinions and made various
contributions in the development and production of prototypes.
References
Bech-Larsen, T., Jensen, B. B., & Pedersen, S. (2010). An exploration of adolescent
snacking conventions and dilemmas. Young Consumers, 11, 253e263.
Belk, R. W., Ger, G., & Askegaard, S. (2003). The re of desire: a multisited inquiry
into consumer passion. Journal of Consumer Research, 30, 326e351.
Benner, M., Linnemann, A. R., Jongen, W. M. F., & Folstar, P. (2003). Quality Function
Deployment (QFD)dcan it be used to develop food products? Food Quality and
Preference, 14, 327e339.
Brock, S., & Kayser-Nielsen, N. (2009). Brns spisevaner - En undersgelse om 4. Og 6.
Klasse. Working paper. Aarhus University, Department of Culture and Society.
Brown, K., McIlveen, H., & Strugnell, C. (2000). Nutritional awareness and food
preferences of young consumers. Nutrition & Food Science, 30, 230e235.
Buckingham, D. (2009). 'Creative' visual methods in media research: possibilities,
problems and proposals. Media, Culture & Society, 31, 633e652.
Christensen, P. H. (2003). Brn, mad og daglige rutiner Barn 2-3. Tema: Barn og mat.
Smakebiter fra aktuell forskning (pp. 119e135). Trondheim: Norsk Senter for
Barneforskning.
Christiansen, T., Lvschal-Nielsen, P., & Nielsen, C. V. (2012). En snack! Jamen, det er
jo frugt: Et antropologisk studie af brn og unges valg af en alternativ snack.
Aarhus University, CFK - Folkesundhed og Kvalitetsudvikling.
Cooke, L., & Wardle, J. (2005). Age and gender differences in children's food preferences. British Journal of Nutrition, 93, 741e746.
de Graaf, C. (2006). Effects of snacks on energy intake: an evolutionary perspective.
Appetite, 47, 18e23.
Dolan, P., Hallsworth, M., Halpern, D., King, D., Metcalfe, R., & Vlaev, I. (2012).
Inuencing behaviour: the mindspace way. Journal of Economic Psychology, 33,
264e277.
Fagt, S., Matthiesen, J., Biltoft-Jensen, A., Groth, M., Christensen, T., Hinsch, H.-J.,

92

K.G. Grunert et al. / Trends in Food Science & Technology 47 (2016) 82e92

et al. (2004). Udviklingen i danskernes kost 1985-2001: Med fokus p sukker og

alkohol samt motivation og barrierer for sund livsstil. Frederiksberg: Department
of Nutrition, The Danish Institute for Food and Veterinary Research.
Grunert, K. G. (2002). Current issues in the understanding of consumer food choice.
Trends in Food Science & Technology, 13, 275e285.
Grunert, K. G., Jensen, B. B., Sonne, A. M., Bruns, K., Byrne, D. V., Clausen, C., et al.
(2008). User-oriented innovation in the food sector: relevant streams of
research and an agenda for future work. Trends in Food Science & Technology, 19,
590e602.
Guinard, J.-X. (2000). Sensory and consumer testing with children. Trends in Food
Science and Technology, 11, 273e283.

, R. (1991). The discursive production of selves. Theory & Psychology, 1, 51e63.
Harre
Husby, I., Hietmann, B. L., & O'Doherty Jensen, K. (2006). Meals and snacks from the
child's perspective: the contribution of qualitative methods to the development
of dietary intervention. Public Health Nutrition, 12, 739e747.
Im, S., Nakata, C., Park, H., & Ha, Y. W. (2003). Determinants of Korean and Japanese
new product performance: an interrelational and process view. Journal of International Marketing, 11(4), 81e112.
Jacobsen, L. F., Grunert, K. G., Sndergaard, H. A., Steenbekkers, B., Dekker, M., &
hteenma
ki, L. (2014). Improving internal communication between marketing
La
and technology functions for successful new food product development. Trends
in Food Science & Technology, 37, 106e114.
ster, E. P. (2003). The psychology of food choice: some often encountered falKo
lacies. Food Quality and Preference, 14, 359e373.
ster, E. P. (2009). Diversity in the determinants of food choice: a psychological
Ko
perspective. Food Quality and Preference, 20, 70e82.
Krogager, S. (2012). When girls and boys use media and food. An experimental study of
media use in 4th and 9th grade (PhD thesis). Aarhus University.
Kuenzel, J., & Musters, P. (2007). Social interaction and low involvement products.
Journal of Business Research, 60, 876e883.
Le Bigot Macaux, A. (2001). Eat to live or live to eat? Do parents and children agree?
Public Health Nutrition, 4, 141e146.

on, F., Couronne, T., Marcuz, M. C., & Ko
ster, E. P. (1999). Measuring food liking in
Le
children: a comparison of non-verbal methods. Food Quality and Preference,
10(2), 93e100.
Marcia, J. E. (2001). Identity in childhood and adolescence. In J. S. P. Neil (Ed.), International encyclopedia of the social & behavioral sciences (pp. 7159e7163).
Amsterdam: Elsevier.
Mielby, L. H., Edelenbos, M., & Thybo, A. K. (2012). Comparison of rating, best-worst
scaling, and adolescents' real choices of snacks. Food Quality and Preference, 25,
140e147.
Mielby, L. H., Jensen, S., Edelenbos, M., & Thybo, A. K. (2013). An approach to
measuring adolescents' perception of complexity for pictures of fruit and
vegetable mixes. Journal of Sensory Studies, 28, 66e75.
Mielby, L. H., Kildegaard, H., Gabrielsen, G., Edelenbos, M., & Thybo, A. K. (2012).

Adolescents' and adults' visual preferences for pictures of fruit and vegetable
mixes e effect of complexity. Food Quality and Preference, 26, 188e195.
Mielby, L. H., Nrgaard, M. K., Edelenbos, M., & Thybo, A. K. (2012). Affective
response of adolescents toward fruit- and vegetable-based snacks and the role
of neophobia, gender and age. Journal of Sensory Studies, 27, 425e438.
Nicklaus, S., Boggio, V., Chabanet, C., & Issanchou, S. (2004). A prospective study of
food preferences in childhood. Food Quality and Preferences, 15, 805e818.
Nrgaard, M. K., Bruns, K., Christensen, P. H., & Mikkelsen, M. R. (2007). Children's
inuence on and participation in the family decision process during food
buying. Young Consumers, 8, 197e216.
Nrgaard, M., Hansen, N. K., & Grunert, K. G. (2013). Peer inuence on adolescent
snacking. Journal of Social Marketing, 3, 176e194.
Nrgaard, M. K., Srensen, B. T., & Bruns, K. (2014). A concept test of novel healthy
snacks among adolescents: Antecedents of preferences and buying intentions.
Food Quality and Preference, 33, 17e26.
Nu, C. T., MacLeod, P., & Barthelemy, J. (2007). Effects of age and gender on adolescents' food habits and preferences. Food Quality and Preference, 7, 251e262.
Piernas, C., & Popkin, B. M. (2010). Snacking increased among U.S. adults between
1977 and 2006. The Journal of Nutrition, 140, 325e332.
Popper, R., & Kroll, J. (2005). Conducting sensory research with children. Oxford:
Blackwell.
Reckwitz, A. (2002). Towards a theory of social practices. European Journal of Social
Theory, 5, 243e253.
Reynolds, K. D., Hinton, A. W., Shewchuk, R. M., & Hickey, C. A. (1999). Social
cognitive model of fruit and vegetable consumption in elementary school
children. Journal of Nutrition Education, 31, 23e30.
Savige, G., MacFarlane, A., Ball, K., Worsley, A., & Crawford, D. (2007). Snacking
behaviours of adolescents and their association with skipping meals. International Journal of Behavioral Nutrition and Physical Activity, 4, 36.
Schatzki, T. R. (1996). Social practices: a Wittgensteinian approach to human activity
and the social. Cambridge: Cambridge University Press.
Schatzki, T. R., Knorr-Cetina, K., & von Savigny, E. (2001). The practice turn in
contemporary society. London: Routledge.
Spradley, J. P. (1980). Participant observation. New York: Holt, Rinehart and Winston.
Thomson, P. (2008). Doing visual research with children and young people. London:
Routledge.
van't Riet, J., Sijtsema, S. J., Dagevos, H., & De Bruijn, G. J. (2011). The importance of
habits in eating behaviour. An overview and recommendations for future
research. Appetite, 57, 585e596.
Warde, A. (2005). Consumption and theories of practice. Journal of Consumer Culture, 5, 131e145.
Weijzen, P. L., de Graaf, C., & Dijksterhuis, G. B. (2008). Discrepancy between snack
choice intentions and behavior. Journal of Nutrition Education and Behavior, 40,
311e316.