The Impact of Food Branding On Children's Eating Behaviour and Obesity
The Impact of Food Branding On Children's Eating Behaviour and Obesity
Kathleen L. Keller
a, b,
, Laura G. Kuilema
a, b
, Norman Lee
a
, Joyce Yoon
a
, Brittany Mascaro
a
,
Anne-Laure Combes
a, d
, Bryan Deutsch
a
, Kathryn Sorte
a, c
, Jason C.G. Halford
e
a
Department of Research Medicine, New York Obesity Research Center, St. Luke's Roosevelt Hospital, Institute of Human Nutrition at Columbia University College of Physicians and Surgeons,
New York, NY 10025, USA
b
Department of Nutritional Sciences, Pennsylvania State University, University Park, PA 16802, USA
c
Mailman School of Public Health, Columbia University College of Physicians and Surgeons, New York, NY 10025, USA
d
Institut Polytechnique LaSalle Beauvais, Beauvais, France
e
School of Psychology, Eleanor Rathbone Building, University of Liverpool, Liverpool, UK
a b s t r a c t a r t i c l e i n f o
Keywords:
Food branding
Eating behavior
Childhood obesity
Branding is a technique used by the food industry to create a recognizable image to attract consumers and
hopefully boost sales of the product. Children recognize food brands from a young age, but their impact on
the development of eating behaviors and obesity is unclear. In addition, the notion that some branding tech-
niques may be used to increase intake of healthful foods, like fruits and vegetables, has not been rigorously
investigated. Three laboratory-based intake studies designed to test the impact of common food brands on
children's eating habits are presented. In the rst study, four to six year-old children (n=43) were exposed
to ad libitum test-meals where foods were presented either with or without their associated branding. In the
second study, a novel food brand based Stroop task was developed and tested to assess children's cognitive
response to food brands, and following this procedure, seven to nine year-old children (n=41) ate ad libitum
test-meals consisting of foods packaged with or without a logo from a popular fast food restaurant. Finally, a
pilot intervention was conducted with four to ve year-old children (n=16) to demonstrate the efcacy of
using licensed (spokes) characters to package and promote intake of fruits and vegetables. These studies
demonstrate that branding is an important inuence on what and how much children eat, but some children
may be more susceptible to these inuences than others. Future studies are needed to better understand the
inuence that child age, sex, and obesity has on response to food branding and marketing.
2012 Elsevier Inc. All rights reserved.
1. Introduction
The prevalence of childhood obesity has increased to epidemic
proportions over the last three decades [1,2]. Increased obesity has
been associated with a concomitant increase in incidence of chronic
diseases in children, including cardiovascular disease, type II diabetes,
and hypertension [3]. Curbing these trends has proven to be one of
the most challenging public health problems of recent times. One of
the reasons these efforts have been challenging is because the food
environment is highly obesogenic and has been described as toxic
due to the ubiquitous nature of palatable, high-energy foods and the
lack of physical effort needed to acquire these foods [4]. One of the
most important contributors to this toxic environment is food
advertising. Children are bombarded with advertisements for foods
and beverages telling them what, where, and how to eat, and they
are exposed to these messages through not only television, but
many other media outlets. It is estimated that children may view as
many as 40,000 advertisements for food each year [5], and 98% of
these are promoting foods high in fat, sodium, and sugar [6]. Several
key regulating bodies in the United States and Europe have reviewed
the available evidence and have concluded that food advertising plays
a powerful role in shaping children's eating habits and may be causal-
ly linked to the development of obesity [5,7].
There are a variety of forms of food advertising directed at chil-
dren, but the impact each of these forms may have individually is un-
known. One example is food branding, a form of marketing directed
commonly at children that is intended to develop strong emotional
attachments with a product name or logo [8]. The importance of
branding can be illustrated by examining marketing campaigns like
the Pepsi challenge. Despite the fact that in blind taste tests, the
majority of Americans report a preference for Pepsi, follow-up tests
done without blinding consumers to the cola's brand showed that the
majority preferred Coke [9]. In a follow-up neuroimaging study,
Physiology & Behavior 106 (2012) 379386
A presentation given at the Society for the Study of Ingestive Behavior Annual
Meeting, Clearwater, FL.
Corresponding author at: Department of Nutritional Sciences and Department of Food
Science, Pennsylvania State University, 321 Chandlee Laboratory, University Park, PA,
16802, USA. Tel: +1 814 863 2915; fax: +1 814 863 6103.
E-mail address: [email protected] (K.L. Keller).
0031-9384/$ see front matter 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.physbeh.2012.03.011
Contents lists available at SciVerse ScienceDirect
Physiology & Behavior
j our nal homepage: www. el sevi er . com/ l ocat e/ phb
researchers demonstrated differences in how the brain responds to
Coke versus Pepsi that reinforce the idea that branding of a product
can override the product's sensory attributes [10]. Clearly, branding
plays a powerful role in forming our hedonic reactions to foods, and
more critically, our reactions to brands may not be under conscious
control.
Children learn to recognize brands early in life [8,11], but until
recently, the impact of branding on children's food preferences and eat-
ing behavior had rarely been studied under controlled laboratory condi-
tions. Robinson et al. conducted a laboratory study with preschool
children to determine the impact of placing McDonald's packaging
onfoods that were not actually fromMcDonald's. Inadditionto French
fries and hamburgers, they also included healthier options, like carrots
and apples. They found that children reported liking the foods more if
they were presented with McDonald's branding, compared to when
the foods were presented inplain packaging [12]. Intwo related studies,
similar aged children reported that they liked snack foods better when
they were in containers that contained licensed cartoon characters on
them, compared to when they were presented in plain containers
[13,14]. The mechanisms behind these effects are unclear. It is possible
that popular brands, like McDonald's, decrease children's level of
neophobia for an associated food, thus making it a safer and more fa-
miliar option. Another possibility is that the addition of familiar brands
or cartoon characters may simply serve as a colorful, fun visual stimu-
lus that enhances the overall eating experience. Additional research is
needed to better understand how changes in food packaging can im-
pact children's preferences and intake. This research is particularly
needed at a time when children consume less than the recommended
amounts of many fresh fruits and vegetables (F&V) [15]. Techniques
used by the food marketing industry may ultimately provide sugges-
tions for making healthful foods more appealing to children.
Soaring rates of childhood obesity throughout the world [16] ne-
cessitate studies that examine the impact of food marketing on obesity.
Although the evidence is still controversial, work from several investi-
gators supports the notion that obese children may be more susceptible
to external food advertising cues than are their lean counterparts. In
previous work from Halford et al. [17,18], children were exposed to ad-
vertisements for either foods or non-foods in school-based settings, and
following, they were allowed to consume ad libitum from an assort-
ment of snacks. In addition, a brand awareness test was administered
to determine whether overweight/obese (OW) children would be able
to recognize more food advertisements than non-overweight (non-
OW) children. They found that all children consumed more after expo-
sure to food advertisements, compared to after viewing non-food
advertisements. However, the increase in intake was greater in the
OW children than in the non-OW children. Furthermore, OW children
were able to recognize more food brands in the brand awareness test
thanwere non-OWchildren. The authors hypothesizedthat OWchildren
were more responsive to external food cues, an idea rooted in the theory
of externality proposed by Stanley Schachter [19]. In Schachter's work
with adults and children, he manipulated external cues related to eating,
such as meal times, and he noted that OW individuals were more likely
to eat in response to changes in these external cues than were non-OW
individuals. Several follow-up lines of evidence support this notion in
children, including work from Jansen et al. [20], Carnell et al. [21], and
others [22,23]. However, whether greater externally motivated eating
is a causal factor in the development of obesity is not known, and the
idea remains controversial.
Given the powerful impact that branding and packaging have on
food preference, we hypothesized that it might be possible to increase
children's intake of foods in a controlled laboratory setting simply by
manipulating brand and packaging cues. In addition, because of the
previously discussed evidence that OW children might be more sus-
ceptible to food cues, we hypothesized that they might be more respon-
sive to food branding and packaging manipulation, compared to non-
OW children. To address these questions, we conducted a series of
pilot studies from2006 to 2011 that were designed to assess the impact
of manipulating food brand packaging on children's intake and eating
behaviors. These studies are examined below.
2. Laboratory studies testing the role of food branding on
children's eating behavior
2.1. Study one: Do familiar food brands affect test-meal intake in four to
six year-old children?
The purpose of this study was to determine if the presence of
familiar food brands alone would result in increased intake of a multi-
item buffet meal compared to a control condition where foods were
served in plain packaging (unbranded). In order to answer this question,
we tested 43 healthy, four to six year-old children, mean age 5.9
0.9 years, from diverse ethnic backgrounds living in New York City
[24]. Seventeen (n=17) of the children were boys (39.5%) and the
remaining 26 were girls (60.5%). Twenty-three (n=23) children
were non-OW and 20 were OW, dened as BMI-for-age85th %
[25]. There were no differences in the breakdown of sex, ethnic or in-
come group as a function of child weight status. In order to be enrolled
in the study, children had to be both familiar with and like at least ve
out of the seven foods used in the study based on parental report
during the screening procedure. Prior to their rst visit, children
were randomized to the order in which they received four test-
meals, two unbranded meals and two branded meals. In the un-
branded meals, all foods were served in plain, white plastic containers.
In the branded meals, all foods were served in their original brand
packaging. For the test-meals, foods that are commonly advertised to
children (e.g. Lunchables, Trix Yogurt, Oreos, etc.) were selected.
In addition, prior to the test-meal, we administered a brand awareness
test similar to the protocol employed by Halford et al. [17].
2.1.1. Results: Study one
Results showed no differences in intake at the branded vs. unbranded
condition when all children were combined (p=0.80). However, the
difference in intake between the branded and unbranded conditions
was signicantly different between OW and non-OW children and
this difference remainedafter adjusting for childage, sex, andethnicity.
OWchildren consumed about ~41 cal more when foods were branded,
while non-OWchildren consumed about ~45 cal less when foods were
branded (p0.05) compared to when they were unbranded. We also
found a signicant effect of sex on intake at the two conditions, with
boys tending to eat more at the branded condition than girls (pb0.05),
however, this was likely due to two boys who were outliers. When
removed from the database, the effect due to sex was no longer signif-
icant (p=0.4). It was unclear why non-OWchildren actually consumed
fewer calories when foods were branded, but we hypothesized that
children may have been less willing to open packages of some of the
foods (e.g. potato chips) if they were only going to have a few bites. In
contrast, OW children may have been more familiar with receiving
foods in their packaging, or they may have been more motivated to
open this packaging to sample the contents. Despite these differences
in intake, we found no differences in level of brand awareness or
reported hours spent watching television in OW and non-OWchildren.
For additional details, the reader is referred to the published report [24].
In a related study, Halford et al. did nd a positive association with
branded food selection and quantity of TV viewing in older children
(unpublished results). One important confound that should be investi-
gated in future studies is the food packaging itself. As noted above, the
fact that children had to open packages of foods in the branded condi-
tion may have inuenced the response, and future studies should adjust
for this effect.
The results of this rst experiment demonstrated that food brands
are an important visual cue and some children may overeat in their
presence. The observation that OWchildren appeared more responsive
380 K.L. Keller et al. / Physiology & Behavior 106 (2012) 379386
to food branding exposure than non-OW children is supported by pre-
vious studies from Halford et al. with children aged nine to eleven
years-old [17,18]. In both these studies, the authors exposed children
to food and non-food related advertisements embedded in a cartoon,
followed by ad libitum consumption of a range of sweet and savory
low- and high-fat snacks. Under these conditions, all children ate
more after exposure to food advertisements compared to non-food ad-
vertisements, but the effects were greater in OW compared to non-OW
children. Several studies do not support these ndings, however. Half-
ord et al. did not show any differences between OW and non-OW chil-
dren's response following food advertising exposure in ve to seven
year-old children [26]. The authors hypothesize that the weight-
related response to food advertisements might be dependent on other
factors such as child age, although a mechanism for this effect has not
yet been investigated. Further, Dovey et al. found that level of child neo-
phobia predicted children's responsiveness to advertisements for
healthy foods [27]. In addition, this same group did not nd any differ-
ences in 1113 year-old children's food preferences after food advertis-
ing exposure, although food intake was not assessed in this study so it is
not possible to compare these ndings with our own [28]. Harris et al.
[29] tested intake of goldsh snack crackers during exposure to tele-
vision programming embedded with either advertisements for food or
non-foods. They found that regardless of child weight status, viewing
advertisements for foods increased intake of the crackers by 45% com-
pared to the non-food conditions. It is important to note, however,
that the latter study used only one test food, goldsh crackers, whereas
both our study and the two by Halford [17,18] offered a variety of test
foods that ranged in palatability. Offering a variety of test foods could
have stimulated a greater response in OW children than that seen by
Harris et al. [29]. Moreover, children were offered the sh crackers
to consume while they were watching the advertisements as opposed
to priming them with advertising exposure as in Halford's studies
[17,18]. Because of these differences in protocols, comparing directly
across studies is not possible. Undoubtedly, additional research is re-
quired to better understand these relationships.
The rst experiment led to several follow-up questions. First, if
some children are more susceptible to the effects of food brands,
how can one measure this susceptibility in the laboratory? Second,
the food brands selected in this rst experiment were limited and
children likely had different levels of familiarity with these brands
that could have potentially impacted the results. Would similar re-
sults be observed if we selected a single, highly recognizable brand,
for example a brand from a popular fast food restaurant?
2.2. Study two: The role of fast food branding in children's
eating behavior
In this follow-up study, the goals were to select a recognizable fast
food brand logo, similar to that used in Robinson et al. [12] and deter-
mine the impact of placing this brand on a multi-item test-meal that
contained a range of healthful and age-appropriate food items. This was
a cross-sectional study where children came to the lab for two dinner
sessions where they were randomly assigned to receive a test-meal
that was either branded with fast food logos or unbranded with
no fast food logos. While the previous study by Robinson et al. [12]
had only measured the impact of fast food branding on children's liking
of foods, the purpose of this study was to translate these effects to food
intake, a more direct link to diet and body weight regulation. A second
goal of this study was to develop a modied Stroop task to test the hy-
pothesis that OW children have an attentional bias toward food brand
images. The food industry reaches out to children through an increasing
number of avenues, including television, the internet, billboards, and
product placement in schools [5]. Because of the ubiquitous nature of
food advertising, attempts by parents or policy makers to limit exposure
to these images are often unsuccessful. As a result, developing a reliable
method to identify children who might be most susceptible to food
branding is a worthwhile research endeavor.
2.2.1. Participants
Atotal of 41, seven to nine year-old children [mean=8.40.5 years]
completed this study. Older children were selected due to the cognitive
demands required of the Stroop task. Additionally, older childrenwould
have experienced a great number and diversity of marketing messages
from a variety of media sources. Twenty-two (22) of the children had a
BMI-for-ageb85th % (non-OW) and 19 children were OW (85th %
BMI-for-age) [25]. Fifty-one (51%) of the children were boys, and ethnic
background was representative of upper Manhattan, with ~65% of the
children coming from African-American (33%) and Hispanic-American
(32%) backgrounds.
2.2.2. The Food Brand Stroop Task
As identied in the study described above by Forman et al. [24],
some children may be more susceptible to the effects of food branding
than others. In addition, several previous studies have found that OW
children are more susceptible to external food cues and more likely to
eat in their presence compared to non-OW children [20,23,30]. Al-
though Schachter's theory of externality [19] has not been widely ac-
cepted as a causal factor in the development of obesity, these studies
make a strong case that for some children, exposure to external food
cues, like advertising, may increase their likelihood of overeating. In
previous studies, external eating has been measured with the use of
scales, such as the Children's Eating Behavior Questionnaire [31] or
the Dutch Eating Behavior Questionnaire [32]. However, a more direct
approach to measuring cognitive bias toward food brands might be
achieved through a variation on the Stroop task [33]. The classic
colorwordStroop task is one of the most widely used tests of cognitive
function where participants are asked to view color words (e.g. red,
blue) and name the color the word is printed in but ignore the actual
content of the word. This task is easy when the color and word are con-
gruent (e.g. red printed in red ink), but it is more difcult when the
color and word are incongruent (e.g. red printed in blue ink). The dif-
ference in reaction time on incongruent versus congruent tasks is a re-
liable assessment of cognitive processing [33,34]. Several modications
have been made to the classic color Stroop, including the pictureword
Stroop task, where words are printed over simple drawings or pictures.
The theory behind this test is that more inuential words will result in
longer reaction times than less inuential words [35]. In addition, an
emotion-based version of the Stroop task has been used to show that
patients who have eating disorders experience delayed reaction
times in naming words related to food or shape (e.g. fat, thin,
chocolate, etc.), and this has been interpreted as a heightened cogni-
tive processing of food-related stimuli [36]. Similar methods have been
used to report biases toward food-related words among obese patients
[23]. In the rst part of this study, we developed and tested a modied
version of the Stroop task to assess children's cognitive bias toward
food brand images with the prediction that OWchildren would perform
more poorly on incongruent versus congruent trials than non-OW chil-
dren. In addition, children also completed the classic color Stroop task
[33].
For the Food Brand Stroop Task, a total of 136 different food image
photographs were printed in color on 8 1/2 in. 11 in. cards. Half of
the images were of foods without brand logos (e.g. broccoli, cookies)
and the rest were pictures of common food brand logos. All images
were matched for color intensity, luminance, length and width dimen-
sions, and pixel size. Images were printed on a high-quality laser jet
printer, placed in protective plastic folders, and presented to children
in a three-ring binder. In order to create the congruent and incongruent
tasks, each food image was superimposed with a word that either
matched the image for the congruent tasks (see Fig. 1a and b) or did
not match the image for the incongruent tasks (see Fig. 1c and d). For
the incongruent tasks, there was always the presence of either a picture
381 K.L. Keller et al. / Physiology & Behavior 106 (2012) 379386
of a food brand logo or the name of a food brand (e.g. Frosted Flakes).
This was done based on the assumption that children who had a greater
cognitive interference to food brands would take longer to respond to
the task in question (either naming the food pictured behind the
branded word or reading the word placed in front of the branded
image). In order to prevent confusion, branded images were digitally
altered so that the only words present were those involved in the
Food Brand Stroop Task (see Fig. 1b and d). Congruent and incongruent
tasks were balanced, as were pictures of branded and unbranded foods.
As in the color Stroop, images were arranged in blocks of four. Each
block contained two branded images and two unbranded images, and
was either congruent or incongruent.
Upon presentation of the images, children were asked to either read
the word or identify the food image. Children were reminded of these
instructions before each block. In the case of image identication,
more than one answer was often accepted as correct (i.e. chocolate
or candy bar were acceptable in lieu of Snickers, or vice versa). Chil-
dren were not penalized for incorrect responses, but were not shown
the next image until a correct answer was given.
2.2.3. Scoring the Food Brand Stroop Task
All Stroop sessions were recorded using a hand-held audio record-
ing device. A stopwatch was used to time children's performance on
each block of four (congruent or incongruent). To assess performance,
children's average time for both congruent and incongruent images
was calculated and a score was created corresponding to the difference
in time to respond to the incongruent compared to the congruent im-
ages. Higher scores on both the color and Food Brand Stroop suggested
greater cognitive interference. With the Food Brand Stroop, in particu-
lar, we interpreted higher scores to mean a cognitive bias toward food
brand words or logos.
2.2.4. Multi-item test-meal
After administration of the Stroop tasks, children received a multi-
item test-meal that was either branded with the logo of a popular
fast food restaurant, or unbranded served in plain white packag-
ing. The order in which children received the test-meals was random-
ized. Children were allowed 30 min to eat as much as they wanted.
They ate individually while a research assistant read books to provide
a neutral distraction. The same foods, serving equipment, and tray
placement was used across both meals. The test-meal included: turkey
and cheese, ham and cheese, and peanut butter and jelly sandwiches,
pretzels, graham crackers, apple slices, carrot sticks, pudding, plain
and chocolate milks.
2.2.5. Results: Study two
A similar pattern of response was found for both the color and
Food Brand Stroop tasks, suggesting that the latter was at the very
least successful at demonstrating a Stroop like effect. On the color Stroop,
incongruent tasks took longer to respond to than congruent tasks (t=
6.93; pb0.001). Children averaged 5.481.03 s. on congruent tasks
and 7.571.33 s. on incongruent tasks for the color Stroop. On the
Food Brand Stroop, children averaged 7.912.09 s. on congruent tasks
and 9.974.33 s. on incongruent tasks (t=2.39; pb0.05).
Children with higher BMI z-scores tended to show greater cogni-
tive interference on the Food Brand Stroop Task, but not on the
color Stroop. After adjusting for age and reading performance, the
difference between OW and non-OW children on the color Stroop
was not signicant (p=0.10). In contrast, OW children had greater re-
sponse times on the incongruent versus the congruent trials of the Food
Brand Stroop Task than non-OW children (pb0.05). On average, OW
children took approximately 2 s longer than non-OW children to re-
spond to the incongruent trials. We interpreted this result to suggest
that OW children may have a cognitive processing bias toward some
food brand images, however, the sample size was small and results
need to be conrmed in larger cohorts. Alimitation in any Stroop proce-
dure is that one cannot deduce the exact reason for the delayed reaction
time on incongruent trials. For example, children's prior experience
with a product or their degree of liking or wanting for a food might
(a) (b)
(c)
(d)
Fig. 1. Example images in the Food Brand Based Stroop Task. Image (1a) is unbranded congruent, image (1b) is branded congruent, image (1c) is unbranded congruent, and image
(1d) is branded congruent. Children were either asked to read the word or name the picture. Time to respond to each task was recorded and averaged across congruent vs.
incongruent tasks. Higher scores on incongruent tasks were interpreted as an increased cognitive bias toward food brand images.
382 K.L. Keller et al. / Physiology & Behavior 106 (2012) 379386
have also inuenced reaction times, and assessing these parameters in
future studies is warranted.
For the test-meal portion of the study, there was a trend for all
children, regardless of weight status, to eat more at the branded com-
pared to the unbranded meal, with meanstandard deviation in-
takes of 793.0302.9 kcals and 730.0326.4 kcals, respectively
(p=0.07). When test-meal foods were analyzed individually, there
were no signicant differences in intake between conditions depend-
ing on whether the food was healthy (e.g. carrots, apples) or un-
healthy (e.g. ham and cheese sandwich, and chocolate milk). Rather,
children tended to increase intake by a small amount across all test
foods at the branded compared to the unbranded meal. There was a
trend for OW children to eat more than non-OW children regardless
of brand condition, with mean (SD) intakes of 843.7314.1 kcals
and 691.7265.3 kcals, respectively (p=0.10). However, child
weight status did not interact with brand condition (p=0.80). On
the other hand, children did show different responses to the branding
condition as a function of sex. Regardless of whether the meals were
branded or unbranded, boys ate a similar amount of food (p=0.66).
In contrast, girls ate about 100 cal more when the meals were branded
732199.6 (kcals) than when they were unbranded 632197.3 kcals,
[F(1,21)=5.8; pb0.05] (Fig. 2).
The results of this study revealed two novel ndings. First, the
Food Brand Stroop Task demonstrated that OW children may have a
cognitive bias toward some food brand images, although these nd-
ings should be interpreted with caution because of the small number
of children tested and the fact that the instrument has not yet been
validated. Moreover, although the Stroop task revealed a possible
cognitive bias towards food brand images in OW children, we did
not nd that OW children consumed more when the test-meal was
served in meals branded with fast food logos compared to when it
was served in plainpackaging. Because this nding disputes our original
hypothesis, that OW children are more responsive to food cues that
non-OWchildren, additional studies are needed to address these incon-
sistencies. These studies should be done in larger cohorts, after adjust-
ing for important confounders, such as child age, sex, neophobia, and
previous experience with a particular food brand or advertisement.
Including the Stroop test prior to the test-meal was also a confounder
in the present study because exposing children to food brands before
the meal could have inuenced their subsequent intake. The second
nding from this research was that branding of test-meals with fast
food logos was not associated with increased intake in all children.
This is in contrast to reports from Robinson et al. who found that the
presence of fast food branding increased preschool children's liking
for foods [12]. Unexpectedly, we did nd that girls appeared to be
more susceptible to the hyperphagic effects of fast food branding than
were boys at this age. It is possible boys were less susceptible to the in-
uence of fast food brands because they in fact knewthe foods served in
the laboratory were not actually from a fast food restaurant. The notion
that girls appeared to be more susceptible to branding than boys at this
age should be investigated in future studies.
2.3. Study three: Can food marketing techniques be used to increase
intake of healthful foods?
Undoubtedly, food marketing elicits a powerful effect on consumer
food choice. However, given that most foods that are marketed are
rich in sugar and fat, the argument can be made that it is the high palat-
ability of these foods that is responsible for their popularity, and not the
inuence of marketing. If this were true, food marketing techniques
would not be as effective at promoting intake of healthful foods, like
fruits and vegetables. In the real-world, advertisements for fresh
fruits, vegetables, and dairy products are rare [37,38]. Part of the chal-
lenge in marketing these foods is that they are commodity products
that are perishable; the ability of industry to prot from the sales of
these goods is limited. However, understanding the impact of utilizing
marketing techniques to promote fruit and vegetable intake in children
is a worthwhile research endeavor because parents and schools could
use this information to improve the ways foods are presented and
served to children. Recent work from Nicklas et al. demonstrated mod-
erate effectiveness of creating commercials based on two characters,
Judy Fruity and Reggie Veggie at increasing children's preferences
for some vegetables [39]. Whether or not a similar strategy could be
used to increase children's intake of these foods was the focus of this
third study.
The ultimate goal of this study was to use a behavioral marketing
approach to increase the incentives for eating F&Vamong young children.
Through this approach, we employed a multitude of mechanisms com-
monly used by the food industry to market food to children, including
licensed characters, fun and colorful packaging, and the use of pre-
miums or prizes placed within packaging. The strength of this method
is that it uses a multi-faceted approach that is more naturalistic than
manipulating only one aspect of the food brand or package. Currently,
highly sweetened, energy dense foods offer large incentives for con-
sumption because they are palatable, satiating, and often marketed
in colorful, fun, and age-appropriate packaging [40]. Vegetables, on
the other hand, are often short-changed because they are naturally
less palatable, have lower satiating potential, and are typically served
without the appeal of marketing. Although changing the palatability
of vegetables is a challenge, it may be possible to package and present
themin a way that would offer greater incentives to children for eating
them. To demonstrate the efcacy of this approach, we conducted a
seven week randomized control trial with four to ve year-old children
from diverse ethnic backgrounds. All the children were at risk for
obesity, based on having at least one parent with a BMI 25 kg/m
2
,
and they had to consume fewer than two servings of F&V per day,
based on parental report during a screening phone call. A total of 19
childrenwere recruited, but due to scheduling conicts, only 16 families
nished all seven visits and the ndings from these families will be
described herein.
Prior to the rst visit, children were randomly assigned to the in-
tervention (n=7) or control groups (n=9). Baseline measures (sex,
ethnic breakdown, BMI z-score, and age) were similar across the two
groups. Families in both groups attended weekly, small-group ses-
sions with the researchers where baseline measures were taken and
family-based nutrition education was delivered. Although the re-
searchers were not blinded to who was receiving the treatment, the
400
500
600
700
800
900
1000
Boys Girls
Child Sex
M
e
a
l
I
n
t
a
k
e
(
K
c
a
l
s
)
Unbranded
Branded
Fig. 2. Intake (kcal) at the ad libitum test-meal as a function of child sex (boys vs. girls)
and brand condition (unbranded vs. branded). Overall, boys (n=21) ate more than
girls (n=20), regardless of brand condition (p=0.04). There interaction between
brand condition and sex was not signicant (p=0.20). However, analyses run sepa-
rately showed that girls ate ~100 kcals more during the branded condition than they
did during the unbranded condition (pb0.05), while boys ate similar amounts regard-
less of condition (p=0.86).
383 K.L. Keller et al. / Physiology & Behavior 106 (2012) 379386
nutrition education lessons were run in small groups of two to three
children in each, made up of children who were from both experi-
mental groups. Thus, both groups received the same nutrition
counseling. The counseling sessions were interactive and focused pri-
marily on children, although parents were actively involved in help-
ing children with activities. In addition to nutrition education, all
families received 24, eight-ounce plastic containers lled with the fol-
lowing F&V: beets, broccoli, carrots, red peppers, pineapple, and blue-
berries. All F&V were cut into bite sized pieces on the day of the study
and were packaged and served raw. In addition, a packet of light
ranch dressing was included with the packages of vegetables. Parents
were instructed to offer a fruit and/or vegetable of the child's choos-
ing three times per day at meals and once additional at snack time.
Parents were instructed not to cook the foods, and not to coerce or
pressure children into eating anything. In addition, parents were re-
sponsible for bringing leftover F&V back to the lab the following
week so they could be post-weighed.
Children who were in the control group received F&V in plain
plastic containers throughout the study. Children who were in the in-
tervention group received plain plastic containers during the two
baseline weeks and the follow-up week, but during weeks three
through six, they were given F&V in containers decorated with their
favorite cartoon characters. In addition, a sticker was included inside
each decorated container to simulate the practice of premiums used
by the food industry [41,42]; children were allowed to collect these
stickers on a game board to cash in for a prize the following week.
An example package is shown in Fig. 3.
2.3.1. Results: Study three
Children in the intervention group consumed more servings of
both fruits and vegetables across the three study time points: base-
line, treatment, and control (pb0.05 for all). It is unclear why children
in the intervention group demonstrated such high F&V consumption
at baseline, as they reportedly consumed fewer than two servings
per day before starting the study. However, we observed a trend for
an increase in daily vegetable intake of approximately 20 g per day,
approximately one serving, from baseline in the intervention group,
while the control group showed no change in intake (pb0.05). In
addition, fruit intake also increased by approximately one serving per
day from baseline in the intervention compared to the control group,
however, due to the small cohort size, this difference was not signicant
(p=0.61). Because children in the intervention group showed such
high baseline consumption, it is likely we hit a ceiling effect with re-
spect to fruit intake. In all, children in the intervention group increased
total F&V consumption by approximately 125 g over four days (~ two
servings per day), while children in the control group showed no
change (p=0.07). At follow-up, children in the intervention group
showed a 200 g increase over four days (~three servings per day) com-
pared to baseline, and this was signicantly greater than the change in
F&V intake in the control group (pb0.05) (Fig. 4). It is interesting
to note that even when packaging cues were removed at follow-up,
children in the intervention group continued to show higher F&V
intake, suggesting that once these behaviors are established, the incen-
tives are no longer needed. These ndings need to be replicated in a
larger cohort. Once replicated, packaging F&Vin age-appropriate, color-
ful, cartoon packaging may be a strategy families could use to increase
intake of these foods in young children.
An unexpected nding from this study was the decrease in BMI
z-score observed in the intervention group, while children in the con-
trol group showed anincrease across the study (pb0.05). Although pre-
vious studies in children have demonstrated that substituting low
energy foods, like F&V, for higher fat options may result in weight loss
[43], our study focused on adding F&V to the diet, as opposed to substi-
tution. In studies with adults that have looked at the impact of adding
F&V to the diet, there has generally been no effect on body weight
[44,45]. In addition, because the study lasted only seven weeks, we
did not expect changes in weight status across this time period. It is pos-
sible that children may be more responsive to the effects of adding low
energy dense foods to the diet because they are smaller and have lower
calorie requirements overall. In addition, it is not known if these effects
would persist over a longer time period, once children have an oppor-
tunity to adjust to changes in their diets. Regardless, these ndings are
intriguing and warrant follow-up.
The take-home message from this study was that it might be pos-
sible to manipulate food branding cues to promote intake of healthful
foods among children. These results should be interpreted with cau-
tion because the cohort was small and the study period was short.
In addition, there are several limitations to the study. Foremost, we
used multiple incentives and packaging cues in the treatment
group: deducing these effects to one single mechanism is not possi-
ble, but should be investigated in future studies. Finally, this was a
short-term study. Food marketers establish relationships with chil-
dren over long periods of time, with each exposure intended to rein-
force the products' appeal [11]. It is possible that with additional
exposure to the packaging created in the present study, the improve-
ments inF&Vintake might have been greater; future studies are needed
to test this.
Fig. 3. Packages created for 45 year-old children in the intervention group of Study
three. Package designs were tailored to the children's interests and contained their
favorite cartoon characters. In addition, stickers were provided inside the packages to
serve as an additional incentive for consumption.
-150
-100
-50
0
50
100
150
200
250
300
350
Treatment Follow-up
D
i
f
f
e
r
e
n
c
e
i
n
T
o
t
a
l
F
V
I
n
t
a
k
e
(
g
)
f
r
o
m
B
a
s
e
l
i
n
e
Control
Intervention
p=0.07
p=0.04
Fig. 4. Difference in fruit and vegetable intake between the treatment weeks (weeks
36) and baseline and the follow-up week (week 7) and baseline. During the treatment
weeks, there was a trend for children in the intervention group to increase intake to a
greater extent than control children (p=0.07). Children in the intervention group in-
creased intake during the follow-up week a greater amount than did control children
(pb0.05).
384 K.L. Keller et al. / Physiology & Behavior 106 (2012) 379386
3. Conclusions and recommendations for future studies
The studies in this paper illustrate the power of branding to inu-
ence children's eating behaviors. The presence of branding alone was
associated with increased intake of laboratory test-meals in some
children, particularly overweight four to six year-old children and
seven to nine year-old girls. In addition, among four to ve year-old
children, we demonstrated a possible role for use of licensed cartoon
characters in increasing intake of healthful, traditionally unmarketed,
fruits and vegetables. Although tentative, these studies demonstrate
that branding cannot only change a child's affective reaction to a
food [13,14], but can also impact how much of that food a child
eats, at least in a short-term controlled laboratory setting. The impli-
cations that branding may have on longer term eating patterns and
obesity require additional study.
Regulating the amount of exposure children get to food marketing
has been a challenge, particularly in the United States (US). Although
regulations on both quality and quantity of advertising directed at
children were debated actively in the 1970s, no statutory limits
were passed to prevent the food industry from continuing to target
young children [5]. One could argue, however, that regulations directed
at the food marketing industry are antithetical to the very notion of a
free market economy. Even if regulations are passed, they are unlikely
to happen anytime soon. Meanwhile, childhood obesity rates are con-
tinuing to increase around the globe [46], despite the fact that many
other countries have longstanding (e.g. Sweden) or more recent restric-
tions (e.g. United Kingdom) [47] on both the amount and content of
foods marketed at children. A further complicating factor is where
these regulations should be directed given the multiple avenues used
by the food and beverage industry to reach consumers, including the in-
ternet (e.g. advergames), product placement at schools, television, viral
marketing, and numerous other strategies.
Given the challenges inherent in regulating the food industry,
there is a clear need to develop alternative solutions. This is under-
scored by the fact that marketers are currently developing more pow-
erful and efcient methods of reaching consumers through the use of
brain imaging techniques like functional magnetic resonance imaging
fMRI [48]. The studies presented within this paper reveal some possi-
ble strategies for applying marketing techniques to healthier food
options. The appeal of these methods is that they can be instituted
on a smaller scale basis, within individual families or at schools, and
they can be enacted immediately. However, before these strategies
can be promoted at the public health level, additional studies are
needed. First, the impact of child age on response to food branding
and advertising is unclear. It is believed that before age of seven, chil-
dren are unable to understand the intention of advertising, and thus
may be more vulnerable to its inuence [49]. If advertising strategies
are used to promote healthful foods, it is possible that a window of
opportunity exists where these methods are more likely to work than
others. Second, additional studies are needed to determine the charac-
teristics that might make a child susceptible to advertising cues. For ex-
ample, child weight status may be one factor that is associated with
increased susceptibility to these cues [17,24]. Investigating additional
characteristics, such as food cue responsiveness [20,23] and eating
related disinhibition [31] might also shed light on why some children
are more susceptible to food advertising than others. Third, the mecha-
nisms by whichfood advertising works are still unknown. Brain imaging
techniques, like fMRI, may serve as a novel approach to understanding
how the brain perceives brands. The food industry has embraced these
techniques, but researchers have yet to widely apply these methods to
studying the effects of advertising in the laboratory.
In conclusion, these studies demonstrate that branding is an im-
portant visual cue that impacts how much children eat, although some
children appear to be more susceptible to these cues than others. Future
studies that can pinpoint the exact mechanisms by which branding
works are necessary. Parents should be advised to continue to work
toward limiting children's exposure to advertising, while at the same
time developing strategies to present healthier foods to children that
are attractive, appealing, and age-appropriate.
Disclosure
There are no conicts of interest on behalf of any of the authors.
Acknowledgments
The authors kindly thank Emma Boyland, PhD, fromthe University
of Liverpool for her helpful feedback on drafts of this manuscript.
Funding for this study came from NIH grant K01DK068008 and a
St. Luke's Roosevelt Hospital Pilot Award. Additional support came
from the Obesity Research Center Grant (NIH grant 5P30DK026687-27).
References
[1] Anderson R, Crespo C, Bartlett S, Cheskin L, Pratt M. Relationship of physical activity
and television watching with body weight and level of fatness among children. JAMA
1998;279:93842.
[2] Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal KM. Prevalence of
overweight and obesity in the United States, 19992004. JAMA 2006;295:
154955.
[3] Weiss R, Kaufman FR. Metabolic complications of childhood obesity: identifying
and mitigating the risk. Diabetes Care 2008;31(Suppl. 6).
[4] Brownell KD, Horgen KB. Food ght: the inside story of the food industry, America's
obesity crisis, and what we can do about it; 2004.
[5] McGinnis M, Gootman J, Kraak V, Committee on Food Marketing and the Diets of
Children and Youth. Food marketing to children and youth: threat or opportunity.
Washington, DC: The National Academies Press; 2005.
[6] Powel LM, Szezpka G, Chaloupka FJ, Braunschweig CL. Nutritional content of tele-
vision food advertisements seen by children and adolescents. Pediatrics
2007;120:57683.
[7] Hastings G, McDermott LL, Angus K, Stead M, Thomson S. The extent, nature and
effects of food promotion to children: a review of the evidence. Geneva: World
Health Organization; 2006.
[8] Connor SM. Food-related advertising on preschool television: building brand
recognition in young viewers. Pediatrics 2006;118:147885.
[9] Woolfolk ME, Castellan W, Brooks PI. Pepsi versus Coke: labels, not tastes, prevail.
Psychol Rep 1983;52:1856.
[10] McClure SM, Li J, Tomlin D, Cypert KS, Montague LM, Montague PR. Neural corre-
lates of behavioral preference for culturally familiar drinks. Neuron 2004;44:
37987.
[11] Valkenburg PM, Buijzen M. Identifying determinants of young children's brand
awareness: television, parents and peers. Appl Dev Psychol 2005;26:45668.
[12] Robinson TN, Borzekowski DLG, Matheson DM, Kraemer HC. Effect of fast food
branding on young children's taste preferences. Arch Pediatr Adolesc Med
2007;161:7927.
[13] Lapierre MA, Vaala SE, Linebarger DL. Inuence of licensed spokescharacters and
health cues on children's ratings of cereal taste. Arch Pediatr Adolesc Med
2011;165:22934.
[14] Roberto CA, Baik J, Harris JL, Brownell KD. Inuence of licensed characters on chil-
dren's taste and snack preferences. Pediatrics 2010;126:8893.
[15] Guenther PM, Dodd KW, Reedy J, Krebs-Smith SM. Most Americans eat much less
than recommended amounts of fruits and vegetables. J Am Diet Assoc 2006;106:
13719.
[16] de Onis M, Blossner M, Borghi E. Global prevalence and trends of overweight and
obesity among preschool children. Am J Clin Nutr 2010;92:125764.
[17] Halford JCG, Gillespie J, Brown V, Pontin EE, Dovey TM. Effect of television adver-
tisements for foods on food consumption in children. Appetite 2004;42:2215.
[18] Halford JC, Boyland EJ, Hughes GM, Stacey L, McKean S, Dovey TM. Beyond-brand
effect of television food advertisements on food choice in children: the effects of
weight status. Public Health Nutr 2008;11:897904.
[19] Schachter S, Gross LP. Manipulated time and eating behavior. J Pers Soc Psychol
1968;10:98106.
[20] Jansen A, Theunissen T, Slechten K, Nederkoorn C, Boon B, Mulkens S, et al. Over-
weight children overeat after exposure to food cues. Eat Behav 2003;4:197209.
[21] Carnell S, Wardle J. Appetite and adiposity in children: evidence for a behavioral
susceptibility theory of obesity. Am J Clin Nutr 2008;88:229.
[22] Braet C, Van Strien T. Assessment of emotional, externally induced and restrained
eating behavior in nine to twelve-year-old obese and non-obese children. Behav
Res Ther 1997;35:86373.
[23] Braet C, Crombez G. Cognitive interference due to food cues in childhood obesity. J
Clin Child Adolesc Psychol 2003;32:329.
[24] Forman J, Halford JCG, Summe H, MacDougall M, Keller KL. Food branding inu-
ences ad libitum intake differently in children depending on weight status: re-
sults of a pilot study. Appetite 2009;53:7683.
[25] Cole TJ, Bellizi MC, Flegal KM, Dietz WH. Establishing a standard denition for
child overweight and obesity worldwide: international survey. Br Med J
2000;320:16.
385 K.L. Keller et al. / Physiology & Behavior 106 (2012) 379386
[26] Halford JCG, Boyland EJ, Hughes G, Oliveira LP, Dovey TM. Beyond-brand effect of
television (TV) food advertisements/commercials on caloric intake and food
choice of 57-year-old children. Appetite 2007;49:2637.
[27] Dovey TM, Taylor L, Stow R, Boyland EJ, Halford JCG. Responsiveness to healthy
television (TV) food advertisements/commercials is only evident in children
under the age of seven with low food neophobia. Appetite 2011;56:4406.
[28] Halford JCG, Boyland EJ, Cooper GD, Dovey TM, Smith CJ, Williams N, et al. Chil-
dren's food preferences: effects of weight status, food type, branding and televi-
sion food advertisements (commercials). Int J Pediatr Obes 2012;3:318.
[29] Harris JL, Bargh JA, Brownell KD. Priming effects of television food advertising on
eating behavior. Health Psychol 2009;28:104413.
[30] Carnell S, Wardle J. Measuring behavioural susceptibility to obesity: validation of
the child eating behaviour questionnaire. Appetite 2007;48:10413.
[31] Sleddens EFC, Kremers FPJ, Thijs C. The children's eating behavior questionnaire:
factorial validity and association with body mass index in Dutch children age 67.
Int J Behav Nutr Phys Act 2008;49:19.
[32] Van Strien T, Frijters JER, Bergers GPA, Defares PB. The Dutch Eating Behavior
Questionnaire (DEBQ) for assessment of restrained, emotional and external eat-
ing behavior. Int J Eat Disord 1986;5:295315.
[33] Stroop JR. Studies of interference in serial verbal reactions. J Exp Psychol 1935;8:
64362.
[34] Macleod CM. Half a century of research on the Stroop effect: an integrative re-
view. Psychol Bull 1991;109:162203.
[35] Rosinski RR. Pictureword interference is semantically based. Child Dev 1977;48:
6437.
[36] Dobson KS, Dozois DJA. Attentional biases in eating disorders: a meta-analytic re-
view of Stroop performance. Clin Psychol Rev 2004;23:100122.
[37] Kotz K, Story M. Food advertisements during children's Saturday morning televi-
sion programming: are they consistent with dietary recommendations? J Am Diet
Assoc 1994;94:1296300.
[38] Story M, French S. Food advertising and marketing directed at children and ado-
lescents in the US. Int J Behav Nutr Phys Act 2004;1:117.
[39] Nicklas TA, Tsue-Goh E, Goodell S, Acuff DS, Reiher R, Buday R, et al. Impact of
commercials on food preferences of low-income, minority preschoolers. J Nutr
Educ Behav 2011;43:3541.
[40] Elliot C. Assessing fun foods: nutritional content and analysis of supermarket
foods targeted at children. Obes Rev 2008;9:36877.
[41] Institute of Medicine Committee on Food Marketing and the Diets of Children and
Youth. Food marketing to children and youth: threat or opportunity? Washing-
ton, DC: The National Academies Press; 2006.
[42] Kelly B, Halford JC, Boyland EJ, Chapman K, Bautista-Castano I, Berg C, et al. Tele-
vision advertising to children: a global perspective. Am J Public Health 2010;100:
17306.
[43] Epstein LH, Paluch RA, Beecher MD, Roemmich JN. Increasing healthy eating vs.
reducing high energy-dense foods to treat pediatric obesity. Obesity 2008;16:
31826.
[44] Pierce JP, Natarajan L, Caan BJ, Parker BA, Greenberg ER, Flatt SW, et al. Inuence
of a diet very high in vegetables, fruit, and ber and low in fat on prognosis fol-
lowing diagnosis with breast cancer: the women's healthy eating and living
(WHEL) randomized trial. JAMA 2007;298:28998.
[45] Smith-Warner SA, Elmer PJ, Tharp TM. Increasing vegetable and fruit intake: ran-
domized intervention and monitoring in an at-risk population. Cancer Epidemiol
Biomarkers Prev 2000;9:30717.
[46] Wang Y, Lobstein T. Worldwide trends in childhood overweight and obesity. Int J
Pediatr Obes 2006;1:1125.
[47] CIAA. Codes of conduct for food and beverage marketing communication; 2006.
[48] Perrachione TK, Perrachione JR. Brain and brands: developing mutually informa-
tive research in neuroscience and marketing. J Consum Behav 2008;7:30318.
[49] Wilcox B, Cantor J, Dowrick P, Kunkel D, Linn S, Palmer E. Report of the APA task
force on advertising and children; 2004.
386 K.L. Keller et al. / Physiology & Behavior 106 (2012) 379386