Food Quality and Preference 68 (2018) 215–225

Contents lists available at ScienceDirect

Food Quality and Preference

journal homepage: www.elsevier.com/locate/foodqual

Comparative performance of three interpretative front-of-pack nutrition T

labelling schemes: Insights for policy making
⁎
Gastón Aresa,b, , Fiorella Varelab, Leandro Machinb, Lucía Antúneza, Ana Giméneza,
María Rosa Curutchetc, Jessica Aschemann-Witzeld
a
Sensometrics & Consumer Science, Instituto Polo Tecnológico de Pando, Facultad de Química, Universidad de la República, By Pass de Rutas 8 y 101 s/n, CP 91000
Pando, Canelones, Uruguay
b
Centro de Investigación Básica en Psicología. Facultad de Psicología, Universidad de la República, Tristán Narvaja 1674, CP 11200 Montevideo, Uruguay
c
Observatorio de Seguridad Alimentaria y Nutricional, Instituto Nacional de Alimentación, Piedras 165, C.P. 11000 Montevideo, Uruguay
d
MAPP Centre for Research on Customer Relations in the Food Sector, Aarhus University, Fuglesangsalle 4, 8210 Aarhus, Denmark

A R T I C LE I N FO A B S T R A C T

Keywords: Different interpretative front-of-pack (FOP) nutrition labelling schemes have recently been implemented in
FOP several countries but it is still unclear which is the most effective. The present work compared three inter-
Nutrition information pretative schemes (Nutri-score, health star rating and nutritional warnings) in terms of attentional capture,
Warnings processing time, influence on perceived healthfulness and purchase intention of products with different nutri-
Nutri-score
tional profile. Two studies were conducted. In the first study, attention to and processing time for interpretation
Health star rating
Food policy
of FOP labels was evaluated using a visual search task with 112 participants. In the second study, an online
survey with 892 participants was conducted to evaluate the influence of interpretive FOP labels on purchase
intention and perceived healthfulness of a series of products. A between-subjects design was implemented to
compare a control condition (without front-of-pack nutrition information) and the three interpretive FOP
schemes. The health star rating was found to perform worse than the other two schemes in terms of capturing
attention and altering perceived healthfulness and purchase intention. The latter effect depended on the degree
of healthfulness of the food products in question, but the effect on consumer behaviour towards unhealthful
product categories was more pronounced for the warning label scheme. From a nutrition policy effectiveness
point of view, results suggest that nutritional warnings may have advantages over Nutri-score and the health star
rating in the context of the current food environment, characterized by the wide availability of products with
high content of nutrients associated with non-communicable diseases.

1. Introduction have a more lasting effect on behaviour change than individual ap-
proaches to obesity (Capacci et al., 2012; Swinburn, Egger, & Raza,
The current food environment is characterized by the wide avail- 1999). Several policy actions targeted at modifying the food environ-
ability of nutrient-poor, calorie dense foods, which are usually in- ment have been suggested, including changes in the availability of
expensive and are intensively promoted (Popkin, Adair, & Ng, 2012; healthy foods, targeted subsidies and taxes, improving the quality of the
Stanton, 2015; Story, Kaphingst, Robinson-O’Brien, & Glanz, 2008). In food supply, restrictions on advertising, and imposing nutrition label-
this situation, unhealthy foods are likely the default option for con- ling standards (Hawkes, Jewell, & Allen, 2013).
sumers, given they need to invest relatively more time, effort and
money to eat healthily (Hawkes et al., 2015; Thaler & Sunstein, 2008). 1.1. Nutrition labelling
For this reason, a reduction in obesity prevalence at the population
level is unlikely to occur until the environmental influences on eating Nutrition labelling informs consumers of the nutritional properties
behaviour are tackled (Swinburn, Egger, & Raza, 1999). of food products through two components: nutrient declaration (i.e.,
Creating supportive food environments that encourage people to eat detailed qualitative information about nutrient content, such as nutri-
healthily has been recognized as the top priority for policy making tion facts panels) and supplementary nutrition information, which in-
(Hawkes et al., 2015). These policies are more cost-effective and can tends to assist consumers to understand the nutritional value of foods

⁎
Corresponding author.
E-mail address: gares@fq.edu.uy (G. Ares).

https://doi.org/10.1016/j.foodqual.2018.03.007
Received 8 January 2018; Received in revised form 4 March 2018; Accepted 12 March 2018
Available online 13 March 2018
0950-3293/ © 2018 Elsevier Ltd. All rights reserved.
G. Ares et al. Food Quality and Preference 68 (2018) 215–225

(Codex Alimentarius, 2017). associated with non-communicable diseases (calories, sugars, saturated
Considering that people usually find it difficult to make sense of fat and sodium) (Corvalán, Reyes, Garmendia & Uauy, 2013). This
numerical information (Paulos, 1988; Peters et al., 2006), simple gra- scheme is compulsory and products should include separate black oc-
phical information has been reported to be more efficient in influencing tagonal signs for each nutrient that exceeds pre-established criteria
risk perception and behavioural intention (Peters, Dieckmann, Dixon, (Ministerio de Salud, 2015). The warning system is currently under
Hibbard, & Mertz, 2007; Fagerlin, Wang & Ubel, 2005). For this reason, consideration in various Latin American countries including Uruguay,
front-of-package (FOP) nutrition labelling schemes have been devised where the Nutrient profile model of the Pan American Health
in order to communicate supplementary information via simple gra- Organization (2016) is used for defining excessive content of nutrients
phical information (EUFIC, 2017). associated with non-communicable diseases (sugars, fat, saturated fat
FOP nutrition labelling is not an intrusive policy and is usually well and sodium) (Ministerio de Industria Energía y Minería, 2017). The
accepted by both consumers and the industry (Mazzocchi et al., 2015). graphical representation of the warnings in the Uruguayan proposal is
In addition, FOP nutrition labelling is not a mere informational mea- shown in Fig. 1c.
sure, as it can be regarded to serve as a ‘nudge’ in the choice situation Although there is general agreement on the need to provide simple
due to its simplicity and salience (Reisch & Sunstein, 2016), meaning nutrition information to empower consumers to make more informed
that it can potentially have a stronger effect on consumer behaviour in purchase decisions, consensus on what interpretational elements are
the store than conventional nutritional information. the most appropriate to encourage consumers to make more healthful
choices and improve the health status of the population is still lacking
1.2. Interpretive front-of-pack (FOP) nutrition labelling (EUFIC, 2017). Therefore, experimental evidence on the relative ef-
fectiveness of different FOP schemes is crucially important for policy
The various FOP nutrition labelling schemes developed worldwide makers considering the adoption of this public policy.
differ in the extent to which they assist consumers to evaluate product
healthfulness (Hodgkins et al., 2012). Directive or interpretive schemes 1.3. Effectiveness of FOP nutrition labelling
provide cues about product healthfulness, which can either be based on
specific nutrients or on the overall product (van Kleef & Dagevos, The effectiveness of FOP nutrition labelling schemes is determined
2015). Directive or interpretive schemes have been reported to be more by their ability to encourage more healthful dietary patterns (EUFIC,
efficient in assisting consumers to accurately evaluate product health- 2017). To be able to do so, various steps need to be influenced by the
fulness and to encourage healthy food choices than other popular FOP label, and this poses a number of requirements to the FOP label.
schemes, such as the guideline daily amounts (GDA) or the traffic-light First, FOP nutrition labelling schemes need to catch consumers’ atten-
system (Arrúa, Curutchet et al., 2017; Arrúa, Machín et al., 2017; tion (Grunert & Wills, 2007). Considering that most in-store purchase
Ducrot et al., 2016; Julia et al., 2016; Mhurchu et al., 2017). decisions are habitual choices (van’t Riet, Sijtsema, Dagevos, & De
Different variants of interpretive schemes have been recently im- Bruijn, 2011), FOP nutrition labelling schemes need to rapidly catch
plemented in several countries (EUFIC, 2017). These largely differ in consumers’ attention in order to disrupt habitual choice routines. The
the type of information they include, their graphic representation and easier it is to notice FOP labels, the more likely consumers are to notice
the underlying nutrient profiling method used to rank product health- them and, consequently, to take them into account in their decision
fulness based on nutrition composition. It is still unclear which of the making process (Bialkova & van Trijp, 2010, 2011).
interpretive FOP label schemes currently discussed is most effective. After FOP labels are attended to, the information they convey
Three interpretative schemes have been recently implemented in should be efficiently processed (Grunert & Wills, 2007). Consumers
countries of different world regions and constitute examples of ap- cannot be expected to invest large cognitive resources to analyzing
proaches that differ in their underlying rationale and design: the French nutrition information as they usually invest little cognitive effort in
Nutri-score (Ministère de l’Agriculture et de l’Alimentation, 2017), the making the great majority of their in-store food choices (Frewer & van
Australian health-star rating (Commonwealth of Australia, 2016) and Trijp, 2007). FOP labels should facilitate understanding of nutrition
the Chilean warning system (Ministerio de Salud, 2015). information and increase the speed with which product assessments can
Nutri-score and the Australian health-star rating system are volun- be performed (Pettigrew, Talati, Miller, Dixon, & Ball, 2017), as in-
tary FOP nutrition labelling schemes that provide a global overview of dicated by shorter processing time.
product healthfulness. This global assessment is based on the product’s Changes in consumers’ food choices are only expected to occur if
content of nutrients associated with increasing the risk factors for non- FOP labels modify healthfulness perception. It can be hypothesized that
communicable diseases (energy, sugars, saturated fats and sodium), as by making information about product unhealthfulness more salient,
well as the product’s content of ‘positive’ nutritional aspects (fibre, FOP labels may be able to influence consumer beliefs and behaviour
proteins, and the content of fruits and vegetables) (Commonwealth of and discourage consumption of unhealthful products (Entman, 1993;
Australia, 2016; Ministère de l’Agriculture et de l’Alimentation, 2017). Dar-Nimrod & Heine, 2011). In this sense, the information included on
Nutri-score classifies products into 5 categories of nutritional quality, FOP labels has been reported to modify consumers’ healthfulness per-
each associated with a different colour and letter: green for the highest ception and purchase intention, in particular for those products that are
nutritional quality (A) and red for the lowest nutritional quality (E) wrongfully perceived as healthful (Arrúa et al., 2017; Lima, Ares, &
(Ministère de l’Agriculture et de l’Alimentation, 2017) (Fig. 1a). Deliza, 2018; Machín et al., 2017; Maubach & Hoek, 2008).
Meanwhile, the Australian health-star rating classifies products into 10
categories of nutritional quality, using the star rating exclusively, which 1.4. Research objectives
ranges from 0.5 stars (least healthful) to 5 (most healthful)
(Commonwealth of Australia, 2016) (Fig. 1b). The main differences The aim of the present work was to compare three interpretive
between the two systems are found in their graphical representation – front-of-pack nutrition labelling schemes (Nutri-score, health star rating
as described above – and the type of comparisons among products they and nutritional warnings) in terms of attentional capture, processing
enable: Nutri-score aims at assisting consumers to evaluate product time, influence on perceived healthfulness and purchase intention of
healthfulness across food categories, whereas the health-star rating products with different nutritional profile. More concretely, FOP label
system mainly enables consumers to discriminate between foods within schemes were compared in terms of: i) time needed for consumers to
the same category, as it establishes some category-specific criteria. identify the label on packages (Study 1), ii) time needed for consumers
The Chilean warning system, on the other hand, is a nutrient-based to classify products as healthful/unhealthful based on the information
scheme, which highlights products with high content of nutrients provided on the labels (Study 1), iii) influence on healthfulness

216
G. Ares et al. Food Quality and Preference 68 (2018) 215–225

Fig. 1. Front-of-pack nutritional labelling schemes included in the study: (a) Nutri-score, (b) health star rating and (c) warning system highlighting excessive content of fat, saturated fat,
sodium and sugars.

perception (Studies 1 and 2), and iv) ability to modify purchase in- two short training tasks before the test to get familiar with the ex-
tention (Study 2). The results can serve market stakeholders and policy perimental procedure and to reduce participants’ individual variability
makers as evidence-based insights on consumer reaction to the FOP (Treisman & Gelade, 1980) The training tasks featured geometrical
labels to make informed decisions about FOP nutrition labelling figures and required participants to identify square or circles among a
schemes. series of distractors. Once the training tasks were completed, partici-
pants completed the the experimental task with food packages.
2. Materials and methods
2.1.2.1. Task 1 – Attentional capture of FOP nutrition labelling
Two studies were conducted to compare the three interpretive FOP schemes. The first task aimed at evaluating the attentional capture of
nutrition labelling schemes. In the first study, attention to FOP labels the FOP nutrition labelling schemes and required participants to look at
and processing time was evaluated using a visual search task in which food packages and to indicate whether they featured a nutrition
participants were presented with food packages on a computer screen labelling scheme or not. A set of 24 packages were used, which
and were asked to indicate whether labels contained a FOP label or not. corresponded to four variants (no FOP nutrition information, Nutri-
In addition, they were presented with a series of food packages and score, health star rating and warnings) of six products (yogurt, potato
were asked to indicate whether they corresponded to a healthful pro- chips, bread, mayonnaise, juice and breakfast cereals). All products are
duct or not. In the second study, an online survey was conducted to frequently consumed in the country and are usually high in at least one
evaluate the influence of interpretive FOP labels on purchase intention nutrient associated with non-communicable diseases. Packages of real
and perceived healthfulness of a series of products. A between-subjects commercial products but not available in the Uruguayan marketplace
design was implemented to compare a control condition (without front- were used for the purpose of the study. Packages were retrieved from
of-pack nutrition information) and the three interpretive FOP schemes. the internet and digitally modified to include the FOP nutrition
The studies were approved by the ethics committee of the School of information schemes. The specific information included in each
Chemistry of Universidad de la República (Uruguay). product package for each nutritional labelling scheme is shown in
Table 1. The four variants (no FOP label, Nutri-score, health star rating
2.1. Study 1 – Attentional capture and processing time and warnings) for one of the food product categories used in the task
are shown in Fig. 2a.
2.1.1. Participants
A total of 112 people (ages ranging from 18 to 45 years, 70% fe- Table 1
male) participated in the study. Participants were recruited from the Labels included in the visual search task aimed at evaluating the attentional capture of
three interpretive front-of-package (FOP) nutrition labelling schemes: Nutri-score, health
consumer database of the research group who authored the study,
star rating and warnings.*
based on their interest and availability to participate. All participants
reported normal or corrected-to-normal vision and full colour vision. Product FOP nutrition labelling scheme
They signed an informed consent prior to the study and received no
compensation for their participation. Nutri-score Health star rating Warnings

Breakfast cereals D 2 Sugar, sodium

2.1.2. Experimental procedure Yogurt C 2 Sugar
The experiment was conducted in the psychophysics laboratory of Orange juice D 1.5 Sugar
Bread C 3 Sodium
the School of Psychology of Universidad de la República (Montevideo,
Mayonnaise D 4 Sodium, fat
Uruguay). A visual search paradigm was used to evaluate the time Potato chips D 2 Sodium, fat
needed to find and process the information featured in the FOP nutri-
tion labelling schemes in two separate tasks. Participants performed * In addition, labels with no FOP nutrition labelling schemes were included.

217
G. Ares et al. Food Quality and Preference 68 (2018) 215–225

Fig. 2. Examples of packages included in Study 1: (a) breakfast cereals featuring different front-of-package (FOP) nutrition labelling schemes (no FOP, Nutri-score, health star rating and
nutritional warnings, from left to right); (b) yogurt packages of healthful (above) and unhealthful (below) products featuring different FOP nutrition labelling schemes (Nutri-score, health
star rating and nutritional warnings, from left to right). Brands are blurred for publication but were shown to participants.

A total of 76 trials were considered, including 16 dummy trials to 2.1.2.2. Task 2 – Time needed to classify products according to their
familiarize participants with the task, and 60 trials of the experimental healthfulness. The second task was aimed at evaluating the time needed
design. Twenty-four of the 60 trials involved food packages that did not to assess product healthfulness based on the information displayed on
feature FOP labels whereas, in the remaining trials packages featuring the three FOP nutrition labelling schemes. For this purpose, a set of 24
FOP labels were presented to participants in duplicate. Participants packages were used, corresponding to two variants (healthful vs.
were instructed to look at the package and indicate whether or not it unhealthful) of four products (yogurt, bread, juice and breakfast
featured any logo related to nutrition information. Trials were pre- cereals) featuring one of three FOP nutrition information schemes
sented on a PC screen following Williams’ Latin square experimental (Nutri-score, health star rating and warnings). The products were
design using DirectRT software (Jarvis, 2004). The packages were dis- selected to represent product categories with different nutritional
played on the screen until a key was pressed to respond. Participants composition, available in the Uruguayan marketplace, and involved
were asked to press the key S (for yes in Spanish) if the package in- both ‘healthful’ and ‘unhealthful’ products. The information included
cluded a FOP scheme or the key N (for no) if it did not. Participants on the packages was determined based on the characteristics of
were requested to respond as fast and accurately as possible. Response products available in the Uruguayan marketplace and is shown in
times and the response were recorded using the software. Table 2. Fig. 2b provides an example of the food packages used in this

218
G. Ares et al. Food Quality and Preference 68 (2018) 215–225

Table 2 Table 3
Labels included in the visual search task aimed at evaluating the time needed to classify Socio-demographic characteristics of participants in the online study (n = 892).
labels as healthful or unhealthful using interpretive front-of-package (FOP) nutrition la-
belling schemes: Nutri-score, health star rating and warnings.* Socio-demographic characteristic Percentage (%)

Product Variant Nutri- Health star Warnings* Gender

score rating Female 66
Male 34
Yogurt Healthful B 4 –
Age
Unhealthful C 2 Excess of sugar
18–25 years 24
Orange juice Healthful B 4 –
26–35 years 27
Unhealthful D 1.5 Excess of sugar
36–45 years 20
Bread Healthful B 4 –
46–60 years 19
Unhealthful C 2 Excess of sodium
Older than 60 years 10
Breakfast Healthful B 4 –
cereals Unhealthful D 2 Excess of sugar and Educational level
sodium Primary school 6
Secondary school 67
* For nutritional warnings, the healthful alternatives did not include any information. Technical education 12
University degree 11
Postgraduate studies 3
task.
The experimental procedure was similar to Task 1. A total of 60 Socio-economic level
Low 22
trials were considered, including 12 dummy trials to familiarize parti-
Medium 59
cipants with the task and 48 trials of the experimental design. The study High 19
was divided into three blocks of 20 trials, each involving a different
FOP label scheme. Each block consisted of four dummy trials and 16
trials of the experimental design (each food package were presented in to the website of the study, where detailed instructions were given.
duplicate). The order of the blocks, as well as the order of the packages Participants gave informed consent and were given the possibility of
within the blocks, were balanced across participants, following a entering into a raffle for a 70 US$ voucher. The characteristics of the
Williams’ Latin square experimental design. Participants were in- participants, in terms of age, gender, educational and socio-economic
structed to look at the food package and indicate whether or not it level, are shown in Table 3. The sample well represented the share of
correspond to a healthful products. Participants were instructed to press low, medium and high socioeconomic status in the Uruguayan popu-
the key S (for yes in Spanish) if the package corresponded to a healthful lation, and the share of gender among respondents can be regarded as
product or the key N (for no) if it did not. Participants were encouraged corresponding to the share of females majorly responsible for food
to take a 5-min break between the experimental blocks. Response times shopping tasks.
and the response were recorded using the software.
2.2.2. Stimuli
2.1.3. Data analysis Packages of eight products frequently consumed in the country and
Following standard practice in visual search tasks, response times with different nutritional composition were included in the study:
longer than three standard derivations from the mean of each experi- lentils, canned green beans, breakfast cereals, yogurt, orange juice,
mental condition were not considered for further analysis (Kruijne & bread, mayonnaise and potato chips. Packages of real commercial
Meeter, 2016; Wolfe, Palmer, & Horowitz, 2010). The influence of the products but not available in the Uruguayan marketplace were re-
independent variables on response times was evaluated using a re- trieved from the internet and were digitally manipulated to include the
peated measures analysis of variance (ANOVA). Product, FOP nutrition three FOP nutrition labelling schemes. Examples of the labels included
labelling scheme and their interaction were considered as independent in the study are shown in Fig. 3.
variables in Task 1, whereas product, type of product (healthful vs. The nutritional composition of products available in the Uruguayan
unhealthful), FOP nutrition labelling scheme and their two-way inter- marketplace was considered to classify the labels according to each FOP
actions were considered in the analysis of Task 2. When the effects were nutrition labelling scheme (Table 3). The nutrient profiling systems of
significant, Tukey’s test at a significance level of 5% was considered for Nutri-score (Ministère de l’Agriculture et de l’Alimentation, 2017) and
post hoc comparison of means. the health star rating (Commonwealth of Australia, 2016) were con-
A logistic regression was used to evaluate differences in the like- sidered for classifying product healthfulness in each system. In the case
lihood of responding “yes” (i.e., pressing the key “S”). In task 1, only of the warning system, the Nutrient Profile Model of the Pan American
data from trials featuring the FOP nutrition labelling scheme were Health Organization (2016) was considered for defining excess of su-
analyzed considering product, FOP nutrition labelling scheme and their gars, fat, saturated fat, and sodium. Table 4 shows the nutritional
interaction as independent variables. In task 2, product, type of pro- composition of the products and the information included in each of the
duct, FOP nutrition labelling scheme and their two-way interactions FOP schemes. As shown, except for lentils and canned green beans, all
were considered as independent variables. A significance level of 5% products contained high content of at least one key nutrient (sugar, fat
was used as the criterion for statistical significance. or sodium) according to the criteria of the Pan American Health Or-
All statistical analyses were performed using R language (R Core ganization and were classified in the mid-low range of the healthfulness
Team, 2017). scale of the Nutri-score and the health star rating. The only exception to
this trend was mayonnaise, which was classified as 4 stars in the health
2.2. Study 2 – Purchase intention and healthfulness perception star rating due to the specific criteria set for fats and spreads
(Commonwealth of Australia, 2016).
2.2.1. Participants
The study was conducted with 892 people (ages ranging between 18 2.2.3. Experimental procedure
and 84, 67% females), recruited using a Facebook advertisement tar- Participants were randomly allocated to one of four experimental
geted at Uruguayan users older than 18 years old. The advertisement conditions: (i) a control condition with no FOP nutrition information
was available during a two-week period in August 2017. Participants (n = 236), (ii) Nutri-score (n = 229), (iii) Australian health star rating
had to click on the link included in the advertisement to be redirected (n = 211) and (iv) nutritional warnings (n = 216). No significant

219
G. Ares et al. Food Quality and Preference 68 (2018) 215–225

Fig. 3. Examples of labels included in the online study featuring different front-of-pack (FOP) nutritional labelling schemes: mayonnaise without FOP information, potato chips with
Nutri-score, lentils with health star rating and orange juice with warning system (from left to right). Brands are blurred for publication but were shown to participants.

differences in gender, age, educational level and socio-economic level 3.1.2. Time required to evaluate product healthfulness based on interpretive
distribution of the three groups of participants who evaluated different FOP nutrition labelling schemes and classification of products as healthful/
FOP nutrition labelling schemes were found (p > 0.31). unhealthful
Participants were asked to imagine that they were at the super- The time needed by participants to evaluate product healthfulness
market and they wanted to purchase a series of products. They were based on the information provided by the interpretive FOP nutrition
presented with the 8 packages, one by one, following a Williams’ Latin labelling schemes was evaluated using response times. Shorter response
Square experimental design that balanced for presentation and carry- times indicate shorter processing and consequently greater ease of un-
over effects. For each of the package, participants had to rate their derstanding. The time needed by participants to respond was sig-
purchase intention using a 10-point scale (1 = I would definitely not nificantly affected by all main and interaction effects (p < 0.001). The
buy it, 10 = I would definitely buy it). After having completed the first interaction between FOP labelling scheme and type of product
task, they were presented with the 8 packages again, in a different (η2 = 0.011), was significant, which suggests that differences between
order, and they were asked to rate their perceived healthfulness using a FOP schemes depended on whether the product was healthful or un-
10-point scale (1=, not healthful, 10 = very healthful). Finally, they healthful. When data were analyzed separately for packages of
were asked to respond to a series of socio-demographic questions. healthful and unhealthful products, the effect of product (η2 = 0.008
for healthful products and η2 = 0.036 for unhealthful products), FOP
scheme (η2 = 0.027 for healthful products and η2 = 0.012 for un-
2.2.4. Data analysis healthful products) and their interaction η2 = 0.014 for healthful pro-
Analysis of variance (ANOVA) was used to evaluate the influence of ducts and η2 = 0.026 for unhealthful products) were all significant
experimental condition on purchase intention and perceived healthi- (p < 0.001), suggesting that differences between the schemes were
ness scores. Product, experimental condition and their interaction were product-dependent.
considered as fixed sources of variation. ANOVA was also performed for As shown in Table 5, when the task involved labels of healthful
each product separately. A 5% significance level was considered in all products, the warning system showed the longest response times for all
the analyses. When differences were significant, Tukey’s test was used product categories. No significant differences were found between the
for post hoc comparison of the average values. All statistical analyses Nutri-score and the health-star rating . When the task involved labels of
were performed using R language (R Core Team, 2017). unhealthful products no significant differences were found for labels of
breakfast cereals and orange juice. For bread and yogurt packages,
participants responded quicker when the packages featured the
3. Results warning system compared to the Nutri-score. The health-star rating
only significantly differed from the warning system in the case of bread,
3.1. Study 1 leading to significantly longer response times.
Regarding the classification of products as healthful or unhealthful,
3.1.1. Attentional capture of interpretive FOP nutrition labelling schemes when packages of healthful products were considered, the warning
The attentional capture of the FOP nutrition labelling schemes was system led to a significantly (p < 0.001) lower average percentage of
evaluated using response times in a visual search task. Shorter response products categorized as healthful (M = 79%) than the Nutri-score
times indicate that respondents could find the scheme relatively faster (M = 96%) and the health-star rating (M = 95%) for all products.
and, consequently, show that the FOP label had higher attentional Meanwhile, differences among the schemes were product-dependant in
capture. The time needed by participants to respond was significantly the case of labels of unhealthful products. No significant differences
affected by product (p < 0.001, η2 = 0.007) and FOP nutrition label- were found for labels of breakfast cereals and juice, for which the
ling scheme (p < 0.001, η2 = 0.005), whereas the interaction between percentage of unhealthful responses ranged between 85% and 89%. In
product and FOP nutrition labelling scheme was not significant the case of bread and yogurt, the percentage of products categorized as
(p = 0.396). The time needed to identify the presence of FOP nutri- unhealthful were significantly higher for the warning system (85% and
tional labelling scheme on the product package did not significantly 89%, respectively) compared to the Nutri-score (51% vs. 53%, respec-
differ between the Nutri-score (M = 990 ms) and the warning system tively). The health-star rating provided the lowest percentage of un-
(M = 1062 ms). Meanwhile, the health-star rating led to a significantly healthful responses for bread labels (25%), whereas it did not sig-
higher average response time (M = 1094 ms) compared to the Nutri- nificantly differ from warnings in the case of yogurt labels (84%).
score, but did not differ from the warning system. No significant dif-
ferences (p = 0.15) between the schemes were found in participants’
ability to correctly identify the presence or absence of the scheme on
the labels.

220
G. Ares et al. Food Quality and Preference 68 (2018) 215–225

Table 5

sodium

sodium

sodium

sodium
Warning system

sugar

sugar
sugar
Average response times (ms) to evaluate product healthfulness for packages of different

fat

fat
product categories with different nutritional profile and different interpretive front-of-

of
of
of
of
of
of
of
of
of
pack nutritional labelling schemes.

Excess
Excess
Excess
Excess
Excess
Excess
Excess
Excess
Excess
Nutritional Product Nutri-score Health-star Warning
–
–

profile category rating system

Health star rating

Healthful Bread 1135a 1107a 1438b

Breakfast 1069a 1145a,b 1275b
cereals
1.5 stars
Nutritional composition of the labels included in the online study and classification according to three interpretive front-of-pack nutrition labelling schemes: Nutri-score, health star rating and warning system.

5 stars
5 stars
2 stars

2 stars

3 stars
4 stars

2 stars
Juice 1449a,b 1090a 1408b
Yogurt 1216a 1167a 1464b

Unhealthful Bread 1688b 1627b 1399a

Nutri-score

Breakfast 1225a 1202a 1243a

cereals
Juice 1274a 1267a 1332a
D

D
A
A

Yogurt 1970b 1311a 1357a

Fruits and vegetables (%)

Average values within a row with different superscript letters are significantly different
according to Tukey’s test (p < 0.05).

3.2. Study 2

3.2.1. Influence of interpretive FOP nutrition labelling schemes on

100

100

healthfulness perception
60

40
0

0
0

As expected, healthfulness perception was significantly influenced

Proteins (g)

by product (p < 0.001), which had the highest effect size (η2 = 0.66).
However, experimental condition (p < 0.001, η2 = 0.007) and the
10.1

interaction between product and experimental condition (p < 0.001,

6.6
1.2

7.1

9.6

2.0
0

η2 = 0.007) also significantly influenced perceived healthfulness. This

Fibre (g)

suggests that the impact of the interpretive FOP nutrition labelling

schemes on healthfulness perception was product-dependant. As shown
8.6
6.8

4.2

1.0

in Table 6, the three interpretive schemes did not modify healthfulness

0

0
0

perception compared to the control condition (without nutrition in-

Sodium (mg)

formation) for lentils, canned green beans and potato chips. These
products correspond to the extreme values of healthfulness perception
563

176

103

600
113

170

in the product set, that is, they are regarded as clearly healthful or
32

18

unhealthful foods in consumer’s view.

Saturated fat (g)

Significant differences among the experimental conditions were

established for the remaining 5 products with intermediate healthful-
ness scores (Table 6). The health-star rating showed the lowest ability
to modify healthfulness perception. This interpretive scheme sig-
0.1

5.0

0.5
0.5

1.5

nificantly decreased healthfulness perception relative to the control

0

only in the case of orange juice. On the contrary, the warning system
Total fat (g)

had the greatest influence on healthfulness perception, significantly

reducing scores compared to the control condition in 5 products:
0.3
0.4

8.5

2.8
5.0

breakfast cereals, yogurt, orange juice, bread and mayonnaise. The

10
0

0
Sugars (g)

Table 6
Average healthfulness# scores for product packages in different experimental conditions:
12

28
24

control (no nutritional information), Nutri-score, health-star rating and warning system.
0
0

5
0

0
Calories (kcal)

Product Control Nutri-score Health-star Warning

(n = 236) (n = 229) rating system
(n = 211) (n = 216)
146

109

217

252

160

7.6 a 7.9a 7.9a 7.8a

90

98

47

Lentils
Canned green 6.4a 6.7a 6.8a 6.4a
beans
Portion size (g)

Breakfast 6.3a 5.5b 5.9a,b 5.4b

cereals*
Yogurt* 6.2a 5.6a,b 6.0a 5.2b
Orange juice* 5.6a 4.7b 5.0b 5.0b
130
100

200
200
100
30

12

28

Bread* 5.2a 4.8a,b 5.0a,b 4.5b

Mayonnaise* 3.5a 2.9b,c 3.3a,b 2.7c
Canned green beans

Potato chips 1.9a 1.8a 1.6a 1.6a

Breakfast cereals

Orange juice

#
Potato chips

Healthfulness was evaluated using a 10-point scale (1 = not healthful, 10 = very

Mayonnaise

healthful).
Product

Yogurt
Lentils

* Indicates products for which significant differences among the experimental condi-
Table 4

Bread

tions were established (p < 0.05). Average scores within a row with different superscript
letters are significantly different according to Tukey’s test (p < 0.05).

221
G. Ares et al. Food Quality and Preference 68 (2018) 215–225

Table 7 nutrition labelling are expected to influence their salience in the con-
Average purchase intention# scores for product packages in different experimental con- text of the food package, and consequently their attentional capture
ditions: control (no nutritional information), Nutri-score, health-star rating and warning
(Yantis, 2000). Results from the present work showed that the health-
system.
star rating had a significantly lower ability to catch consumers’ atten-
Product Control Nutri-score Health-star Warning tion than Nutri-score. While the exact reason cannot be explained by
(n = 236) (n = 229) rating system the analysis, it might be related to the fact that the health star is smaller
(n = 211) (n = 216) in size and has a more homogeneous colour distribution (Antúnez,
Lentils 7.5 a 7.2a 7.1a 7.1a Giménez, Maiche, & Ares, 2015; Bialkova & van Trijp, 2011; Cabrera
Canned green 6.4a 6.7a 6.8a 6.4a et al., 2017). The Nutri-score, in turn, uses the traffic-light colour
beans system, which thus may have contributed to better capture consumer
Breakfast 5.5a 4.7b 4.9a,b 4.4b attention. It has been widely reported that colour increases the salience
cereals*
of the stimuli and thus improves its attentional capture (Green &
Yogurt* 4.7a 4.5a,b 4.7a 4.1b
Orange juice 5.5a 5.1a 5.1a 4.9a Anderson, 1956; Williams, 1966). In line with this, Antúnez et al.
Bread* 4.7a 4.3a,b 4.2a,b 3.8b (2015) found that polychromatic FOP nutritional labelling schemes
Mayonnaise* 4.4a 3.8b 4.0a,b 3.7b better catch consumer attention compared to monochromatic schemes.
Potato chips 3.6a 3.3a 3.5a 3.4a
However, Nutri-score and the warning system performed similar in
#
Purchase intention was evaluated using a 10-point scale (1 = ‘I would definitely not
terms of attentional capture. The latter result may be explained by the
purchase it’, 10 = ‘I would definitely purchase it’). darker colour of the warning and the fact that the warning label size
* Indicates products for which significant differences among the experimental condi- depends on the number of nutrients exceeding the pre-established cri-
tions were established (p < 0.05). Average scores within a row with different superscript teria, as separate warnings are included for each nutrient (Ministerio de
letters are significantly different according to Tukey’s test (p < 0.05). Salud, 2015).

Nutri-score scheme showed an intermediate performance, leading to 4.2. Time needed to classify products as healthful/unhealthful and
significant differences relative to the control in the scores of 3 products: aggregate healthfulness perception
breakfast cereals, mayonnaise and orange juice.
The present work also evaluated differences in the time needed by
3.2.2. Influence of interpretive FOP nutrition labelling schemes on purchase consumers to classify products as healthful or unhealthful using the
intention information provided by the three interpretive FOP labels. In the case of
Purchase intention was significantly influenced by product healthful products, consumers required the longest time when the
(p = < 0.001, η2 = 0.19) and experimental condition (p < 0.001, product package featured the warning system compared to the other
η2 = 0.004), whereas the interaction between product and experi- two interpretive schemes. This difference can be attributed to the fact
mental condition was marginally significant (p = 0.078). This suggests that nutritional warnings only highlight products with excessive con-
that at the aggregate level, the three interpretive schemes significantly tent of key nutrients and therefore provide no specific cues for product
reduced purchase intention compared to the control condition healthfulness when considering healthy food products, which conse-
(M = 5.3a). The lowest average purchase intention scores were found quently do not feature any warning. On the contrary, Nutri-score and
for the warning system (M = 4.8c) and the Nutri-score (M = 4.9b,c), the health star rating are also included on the packages of healthful
followed by the health-star rating (M = 5.0b). However, when differ- products. In this sense, previous research has shown that consumers
ences were analyzed on a product-by-product basis, significant differ- respond faster when they have to identify the presence of a logo on the
ences between the experimental conditions were found for 4 of the 8 label than when they have to identify its absence (Antúnez et al., 2015;
products (Table 7). Bialkova & van Trijp, 2010; Yantis, 1992).
The simplified version of the health-star rating did not significantly In addition, when considering packaging of healthful products, both
modify purchase intention compared to the control condition in any of Nutri-score and the health star rating increased the percentage of par-
the evaluated products. The warning system, in turn, led to significantly ticipants who regarded the products as healthful compared to nutri-
lower purchase intention scores in 4 products: breakfast cereals, yogurt, tional warnings. This might be related to the fact that only products
bread and mayonnaise. Meanwhile, the Nutri-score was associated with with high nutrient content feature warnings and it may be easier for
a reduction in purchase intention in 2 of the products: breakfast cereals consumers to make inferences on product healthfulness when specific
and mayonnaise (Table 7). information is available on the package. Therefore, results from the
present work suggest that the warning system may have greater po-
4. Discussion tential to discourage consumers from making unhealthy food choices
than highlighting healthy products. In the case of processed products,
FOP nutrition labelling is receiving increasing attention worldwide FOP schemes that highlight ‘healthful’ products have raised concerns as
as a strategy to guide consumers towards healthier food choices (EUFIC, they can create healthful associations with products that do not ne-
2017). However, an in-depth understanding of how consumers perceive cessarily have a balanced nutritional composition, particularly if the
different FOP nutritional labelling schemes is still lacking in order to criteria underlying the schemes are not stringent enough (Nestle &
support policy making (van Kleef & Dagevos, 2015). In this context, the Ludwig, 2010). In particular, the Dutch government has recently for-
present work compared three interpretive schemes which have been bidden the use of the industry-led Dutch healthy eating logo, known as
recently implemented worldwide, in terms of attentional capture, pro- Dutch tick, on food packaging in response to a consumer right group
cessing time, and effect on consumers’ perceived healthfulness and campaign (Michail, 2016). Complainants against the Dutch tick argue
purchase intention. that it confuses consumers rather than helping them to make informed
decisions on healthy food choice. Given the voluntary nature of this
4.1. Attentional capture initiative not all manufacturers participate in the system and therefore,
several products available in the marketplace do not feature the tick,
The attentional capture of FOP labels, i.e. their ability to catch but are healthy food options. On the other hand, one of the main ad-
consumers’ attention, is expected to increase the likelihood of con- vantages of positively rated systems, like the Dutch logo, is that they are
sumers using this information in their decision making process expected to provide a good incentive for the industry to reformulate
(Bialkova & van Trijp, 2010). Differences in the graphic design of FOP their product toward healthier nutritional profiles.

222
G. Ares et al. Food Quality and Preference 68 (2018) 215–225

4.3. Effect on healthfulness perception of food products the health star rating was the least effective, as it did not significantly
modify consumers’ purchase intention respect to the control for any of
In the case of packages of unhealthful products, differences between the evaluated products. This result can be related to the fact that the
the schemes were product-specific along the distinction of healthful health-star rating showed the lowest ability to modify healthfulness
versus non-healthful product categories. When products were clearly perception. In this sense, one of the mechanisms by which FOP labels
identified as unhealthful (C-D scores in Nutri-score and less than 3 stars may modify consumers’ eating behaviours is by making product un-
in the health star rating), the three schemes behaved similarly. healthfulness more salient (Entman, 1993; Dar-Nimrod & Heine, 2011).
However, in the case of products with scores in the mid-range of the In the present work, the Nutri-score and the warning system simi-
healthfulness scale of Nutri-score or the health star rating, warnings led larly decreased purchase intention of some products with intermediate
to faster responses and to a higher percentage of participants identi- healthfulness perception. Interestingly, in the case of the product with
fying the products as unhealthful. These results suggest that scores in the lowest perceived healthfulness (potato chips), no significant de-
the middle of the scale may be difficult to interpret for consumers. In crease in purchase intention was found compared to the control. This
this sense, Hamlin & McNeill (2016) reported that although the health result might be explained by the fact that consumers have been re-
star rating provides ordinal information about product healthfulness, ported to ignore health information on labels of unhealthful products
consumers tend to convert it in binary cue (healthful/unhealthful) to (Balasubramanium & Cole, 2002). Regarding the comparison of both
support their decision making. schemes, warnings had a significant effect on consumers’ purchase in-
Interpretive FOP nutrition labelling schemes tended to modify tention for a larger share of products than Nutri-score, which can be
healthfulness perception in the case of products with intermediate explained considering its higher effect on perceived healthfulness due
perceived healthfulness, whereas they did not change consumer per- to its focus on individual nutrients.
ception in the case of products that were clearly identified as healthful
(e.g. lentils) or unhealthful (e.g. potato chips). These results are in 4.5. Implications for policy making
agreement with previous studies that report that FOP nutrition labelling
schemes tend to only influence healthfulness perception of products A number of implications and considerations for policy decisions
that are wrongfully perceived as healthful (Arrúa et al., 2017; Lima, can be drawn from the results of the present work, in particular in light
Ares, & Deliza, 2018; Machín et al., 2017; Maubach & Hoek, 2008). of the objectives of FOP labelling schemes in the context of different
Interestingly, although Nutri-score and the health star rating system market environments and consumer behaviour.
highlighted healthful products, in the present work they did not lead to According to Hawkes et al. (2015), nutrition labelling policies
changes in consumers’ perceived healthfulness. Previous research has should maximize the response of the food system stakeholders in the
shown that consumers are usually aware of nutritional recommenda- intended direction. From this point of view, a FOP labelling scheme
tions (Brown et al., 2008). This stresses that the main objective of FOP with a stronger effect on consumer attention, processing time, health-
nutritional labels should decreasing perceived healthfulness of products fulness perception and purchase intention should be the preferred
with high content of nutrients associated with non-communicable dis- choice for policy makers, as it would more effectively change the
eases instead of increasing perceived healthfulness of healthful pro- structure of consumer demand. In addition, a FOP label with a strong
ducts (Hawkes et al., 2015). demand effect, in particular when it is compulsory on all products,
In the present work, the health-star rating had the smallest effect on would encourage the supply side to react accordingly, for example by
healthfulness perception, which can be attributed to the fact that it does reformulating food products across the whole assortment, instead of
not use the effect of colour cues on healthfulness perception (Braun & only certain products that are then positioned as healthful alternatives
Silver, 1995). Nutri-score uses orange/red to highlight the lowest nu- for a health-aware and -interested consumer group. Research conducted
tritional quality, whereas nutritional warnings use black octagonal in Australia and New Zealand confirms that voluntary labels might be
signs to highlight products with excessive content of nutrients asso- used to that purpose, as approximately 67% of the products including
ciated with non-communicable diseases. Both black and red have been the health star rating in 2016 were labelled with 3.5 stars or more
shown to be associated with hazard and unhealthfulness (Cabrera et al., (Health Star Rating Advisory Committee, 2017).
2017; Griffith & Leonard, 1997; Leonard, 1999), whereas octagonal It is an established fact that the majority of consumers do not look
signs have been shown to be associated with the concept of un- for nutrition information when shopping for food, due to time pressure
healthfulness (Cabrera et al., 2017). These results suggests advantages and other distractions (van Kleef & Dagevos, 2015), as well as low
of using learned associations between signs and symbols (e.g. traffic health motivation overall (Grunert & Wills, 2007). Thus, policy deci-
light colours or “stop” signs) in the design of FOP labels, which have sions on FOP labels should take the behaviour of such consumers as
been previously stressed in warning design (Wogalter, Dejoy, & well as vulnerable groups into special consideration . In this context, an
Laughery, 2005). essential function of a FOP label is attentional capture on the one hand,
In two of the eight products, bread and yogurt, the warning system and the correction of misperceived healthfulness perceptions on the
had higher ability to reduce healthfulness perception than Nutri-score. other hand. The results indicate that nutritional warnings appear to
This difference can be attributed to the nutrient-specific vs. global ap- perform better in this regard than the other two schemes.
proach to product healthfulness. According to the Nutrient Profile Furthermore, the choice of FOP label needs to be considered in the
Model of the Pan American Health Organization (2016) these products context of the market environment, in particular the extent to which
had excessive content of key nutrients, whereas according to the Nutri- unhealthful products are largely available and promoted to a greater
score criteria these products received a “C” rating due to their content extent (Popkin et al., 2012). The results show advantages of nutritional
of “positive” nutrients. The criteria of the Nutri-score scheme consider warnings over the Nutri-score and the health star rating schemes in
the global nutritional profile of products and give less focus to the terms of its ability to modify healthfulness perception and purchase
content of nutrients associated with non-communicable diseases than intention of products with high content of sugar, fat and sodium.
the Nutrient Profile Model. From a food policy point of view and in light of the consumer be-
haviour and market environment, one of the risks of the mere voluntary
4.4. Effect on purchase intention implementation of Nutri-score and the health star rating system is that
it might be used to highlight products with an intermediate position or
The ability of FOP labels to modify consumers’ purchase decisions healthful profile rather than contributing to improve consumer ability
and encourage more healthful choices is the key determinant of their to identify unhealthy food products. However, it should be highlighted
effectiveness (EUFIC, 2017). Results from the present work showed that that although nutritional warnings appeared more effective in helping

223
G. Ares et al. Food Quality and Preference 68 (2018) 215–225

consumers to identify unhealthful products, they can be expected to Sectorial de Investigación Científica (Universidad de la República,
raise greater industry opposition than Nutri-score or the health star Uruguay) for financial support. Pablo Velazquez and Jesse Witzel are
rating if they were implemented on a compulsory basis. In this sense, thanked for the design of the stimuli for the experiments.
the food industry has been reported to oppose schemes that present
products in an unfavourable frame (BEUC, 2006; Magnusson, 2010). References
The health-star rating had a rather weak effect. Thus, it does not
seem to be the scheme of choice for policy makers. Although this system Antúnez, L., Giménez, A., Maiche, A., & Ares, G. (2015). Influence of interpretation aids
has been previously reported to be well-understood and preferred by on attentional capture, visual processing, and understanding of front-of-package
nutrition labels. Journal of Nutrition Education and Behavior, 47, 292–299.
consumers over other schemes (Pettigrew et al., 2017), concerns over Arrúa, A., Curutchet, M. R., Rey, N., Barreto, P., Golovchenko, N., Sellanes, A., et al.
their effectiveness have been raised previously. According to Hamlin & (2017). Impact of front-of-pack nutrition information and label design on children's
McNeill (2016), the health star rating failed to cause changes in con- choice of two snack foods: Comparison of warnings and the traffic-light system.
Appetite, 116, 139–146.
sumer preferences of food products based on their nutritional compo- Arrúa, A., Machín, L., Curutchet, M. R., Martínez, J., Antúnez, A., Alcaire, F., et al. (2017).
sition. Warnings as a directive front-of-pack nutrition labelling scheme: Comparison with
Although results of the present work suggest advantages of nutri- the Guideline Daily Amount and traffic-light systems. Public Health Nutrition, 20,
2308–2317.
tional warnings over Nutri-score and the health star rating system, the Balasubramanium, S. K., & Cole, C. (2002). Consumers’ search and use of nutrition in-
findings do not allow to make assumptions about its effects on actual formation: The challenge and promise of the nutrition labeling and education act.
consumer dietary choice, especially not over time, and does not look at Journal of Marketing, 66, 112–127.
BEUC (2006). Discussion group on simplified labelling: Final report. Brussels: BEUC.
vulnerable groups in particular, nor compare across countries. Thus,
Bialkova, S., & van Trijp, H. (2010). What determines consumer attention to nutrition
further comparative research is necessary on the ability of these FOP labels? Food Quality and Preference, 21, 1042–1051.
nutrition labelling schemes to modify eating behaviour across different Bialkova, S., & van Trijp, H. C. M. (2011). An efficient methodology for assessing at-
consumer segments in different countries, as well as on their real im- tention to and effect of nutrition information displayed front-of-pack. Food Quality
and Preference, 22, 592–601.
pact on diet and health once implemented in the marketplace. Braun, C. C., & Silver, N. C. (1995). Interaction of signal word and colour on warning
In closing, it is worth highlighting that results from the present work labels: Differences in perceived hazard and behavioural compliance. Ergonomics, 38,
should be interpreted bearing aware of its limitations. Firstly, the study 2207–2220.
Brown, K. A., Timotijevic, T., Barnett, J., Sheperd, R., Lähteenmäki, L., & Raats, M. M.
was not conducted under externally valid conditions. Consumers did (2008). A review of consumer awareness, understanding and use of food-based
not face a real purchasing situation in which they are exposed to several dietary guidelines. British Journal of Nutrition, 106, 15–26.
distractors; instead they were presented with food packages, one by Cabrera, M., Machín, L., Arrúa, A., Antúnez, L., Curutchet, M. R., Giménez, A., et al.
(2017). Nutrition warnings as front-of-pack labels: Influence of design features on
one, on a computer screen. Therefore, results may not properly reflect healthfulness perception and attentional capture. Public Health Nutrition, In press,.
the actual consumers’ purchase behaviour in a real-life setting. Future http://dx.doi.org/10.1017/S136898001700249X.
research to explore consumer perception of imperative front of pack Capacci, S., Mazzocchi, M., Shankar, B., Macias, J. B., Verbeke, W., Perez-Cueto, F. J. A.,
et al. (2012). Policies to promote healthy eating in Europe: A structured review of
nutrition labelling schemes in real or simulated environments con- policies and their effectiveness. Nutrition Reviews, 70(3), 188–200.
sidering also time restrictions is warranted. Codex Alimentarius (2017). Guidelines on nutrition labelling. CAC/GL 2-1985. Rome: FAO/
In addition, in contrast with the great variety of products to which WHO.
Commonwealth of Australia. Health Star Rating System. (2016). Available at: http://
consumers are exposed in a real market setting, the present study in-
healthstarrating.gov.au/internet/healthstarrating/publishing.nsf/content/home Last
volved a limited number of products categories. However, the selected accessed March 14th, 2018 .
product categories comprised product with different levels of health- Corvalán, C., Reyes, M., Garmendia, M. L., & Uauy, R. (2013). Structural responses to the
fulness in order to provide an overview of the diverse alternatives obesity and non-communicable disease epidemic: The Chilean law of food labelling
and advertising. Obesity Reviews, 14, 79–87.
available in the marketplace. Research considering a wider range of Dar-Nimrod, I., & Heine, S. J. (2011). Genetic essentialism: On the deceptive determinism
categories and products within each category is still necessary to gather of DNA. Psychological Bulletin, 137, 800–818.
more in depth insights on the impact of imperative FOP nutritional Ducrot, P., Julia, C., Méjean, C., Kesse-Guyot, E., Touvier, M., Fezeu, L. K., et al. (2016).
Impact of different front-of-pack nutrition labels on consumer purchasing intentions:
labels on consumer behaviour. Finally, it should be mentioned that the A randomized controlled trial. American Journal of Preventive Medicine, 50(5),
consumer sample that participated in the study could not be regarded as 627–636.
representative of the Uruguayan population, even though it well re- Entman, R. M. (1993). Framing: Towards clarification of a fractured paradigm. Journal of
Communication, 43, 51–58.
presented socioeconomic groups and gender of food shoppers. EUFIC (2017). Global Update on Nutrition Labelling. The 2017 edition. Brussels: EUFIC.
Fagerlin, A., Wang, C., & Ubel, P. A. (2005). Reducing the influence of anecdotal rea-
5. Conclusions soning on people’s health care decisions: Is a picture worth a thousand statistics?
Medical Decision Making, 25, 398–405.
Frewer, L., & van Trijp, H. C. (2007). Understanding consumers of food products.
The objective of this research was to compare three interpretive FOP Cambridge, UK: Woodhead Publishing.
label schemes in terms of attentional capture and processing time, as Green, B. F., & Anderson, L. K. (1956). Color coding in a visual search task. Journal of
Experimental Psychology, 51, 19–24.
well as their effect on healthfulness perception and purchase intention,
Griffith, L. J., & Leonard, D. (1997). Association of colors with warning signal words.
because these represent different crucial steps in consumer food choice International Journal of Industrial Ergonomics, 20, 317–325.
behaviour. Across these different steps, the health star rating was found Grunert, K. G., & Wills, J. M. (2007). A review of European research on consumer re-
to perform worse than the other two schemes, while the warnings sponse to nutrition information on food labels. Journal of Public Health, 15(5),
384–399.
tended to perform best. The effect that the FOP label schemes exerted Hamlin, R., & McNeill, L. (2016). Does the Australasian “health star rating” front of pack
depended on the degree of healthfulness of the food products, both for nutritional label system work? Nutrients, 8. http://dx.doi.org/10.3390/nu8060327.
the Nutri-score and the warning labels, but the effect on consumer Hawkes, C., Jewell, J., & Allen, K. (2013). A food policy package for healthy diets and the
prevention of obesity and diet-related non-communicable diseases: The
behaviour for unhealthful product categories was more pronounced for NOURISHING framework. Obesity Reviews, 14, 159–168.
the warning label scheme. This suggests that nutritional warnings may Hawkes, C., Smith, T. G., Jewell, J., Wardle, J., Hammond, R. A., Friel, S., et al. (2015).
have advantages over Nutri-score and the health star rating in the Smart policies for obesity prevention. The Lancet, 385, 2410–2421.
Health Star Rating Advisory Committee. (2017). Two year progress review report on the
context of the current food environment, characterized by the wide implementation of the Health Star Rating system – June 2014 – June 2016. Available
availability of products with high content of nutrients associated with at. http://www.healthstarrating.gov.au/internet/healthstarrating/publishing.nsf/
non-communicable diseases. Content/145A3C2B51C5A741CA25803D0006D3B4/$File/Health-Star-Rating-
Progress-Review-Report.pdf Accessed December 11th, 2017.
Hodgkins, C., Barnett, J., Wasowicz-Kirylo, G., Stysko-Kunkowskac, M., Gulcand, Y.,
Acknowledgement Kustepeli, Y., et al. (2012). Understanding how consumers categorise nutritional la-
bels: A consumer derived typology for front-of-pack nutrition labelling. Appetite, 59,
806–817.
The authors are indebted to Espacio Interdisciplinario and Comisión

224
G. Ares et al. Food Quality and Preference 68 (2018) 215–225

Jarvis, B. G. (2004). DirectRT research software, version 2004 [Computer Program]. New Paulos, J. A. (1988). Innumeracy: Mathematical illiteracy and its consequences. New York
York: Emprisoft. NY: Hill and Wang.
Julia, C., Blanchet, O., Méjean, C., Péneau, S., Ducrot, P., Allès, B., et al. (2016). Impact of Peters, E., Dieckmann, N., Dixon, A., Hibbard, J. H., & Mertz, C. K. (2007). Less is more in
the front-of-pack 5-colour nutrition label (Nutri-score) on the nutritional quality of presenting quality information to consumers. Medical Care Research and Review, 64,
purchases: An experimental study. International Journal of Behavioral Nutrition and 169–190.
Physical Activity, 13, 101. http://dx.doi.org/10.1186/s12966-016-0416-4. Peters, E., Vastfjall, D., Slovic, P., Mertz, C. K., Mazzocco, K., & Dickert, S. (2006).
Kruijne, W., & Meeter, M. (2016). Implicit short- and long-term memory direct our gaze in Numeracy and decision making. Psychological Science, 17, 407–413.
visual search. Attention, Perception and Psychophysics, 78, 761–773. Pettigrew, S., Talati, Z., Miller, C., Dixon, Kelly, B., & Ball, K. (2017). The types and
Leonard, S. D. (1999). Does color of warnings affect risk perception? International Journal aspects of front-of-pack food labelling schemes preferred by adults and children.
of Industrial Ergonomics, 23, 499–504. Appetite, 109, 115–123.
Lima, M., Ares, G., & Deliza, R. (2018). How do front of pack nutrition labels affect Popkin, B. M., Adair, L. S., & Ng, S. W. (2012). Global nutrition transition and the pan-
healthfulness perception of foods targeted at children? Insights from Brazilian chil- demic of obesity in developing countries. Nutrition Reviews, 70, 3–21.
dren and parents. Food Quality and Preference, 64, 111–119. R Core, & Team (2017). R: A language and environment for statistical computing. Vienna,
Machín, L., Cabrera, M., Curutchet, M. R., Martínez, J., Giménez, A., & Ares, G. (2017). Austria: R Foundation for Statistical Computing.
Healthfulness perception of ultra-processed products featuring different front-of-pack Reisch, L. A., & Sunstein, C. R. (2016). Do Europeans like nudges? Judgment and Decision
nutritional information schemes across two income levels. Journal of Nutrition Making, 11, 310–325.
Education and Behavior, 49, 330–338. Stanton, R. A. (2015). Food retailers and obesity. Current Obesity Reports, 4, 54–59.
Magnusson, R. (2010). Obesity prevention and personal responsibility: The case of front- Story, M., Kaphingst, K. M., Robinson-O’Brien, R., & Glanz, K. (2008). Creating healthy
of-pack food labelling in Australia. BMC Public Health, 10(1), 662. food and eating environments: Policy and environmental approaches. Annual Review
Maubach, N., & Hoek, J. (2008). The effect of alternative nutrition information formats on of Public Health, 29, 253–272.
consumers’ evaluations of a children’s breakfast cereal. University of Wollongong. Swinburn, B., Egger, G., & Raza, F. (1999). Dissecting obesogenic environments: The
Partnerships, proof and practice. International Nonprofit and Social Marketing Conference development and application of a framework for identifying and prioritizing en-
2008 Jul 15–16. Wollongong: University of Wollongong. vironmental interventions for obesity. Preventive Medicine, 29, 563–570.
Mazzocchi, M., Cagnone, S., Bech-Larsen, T., Niedźwiedzka, B., Saba, A., Shankar, B., Thaler, R. H., & Sunstein, C. R. (2008). Nudge. Improving decisions about health, wealth, and
et al. (2015). What is the public appetite for healthy eating policies? Evidence from a happiness. New York: Yale University Press.
cross-European survey. Health Economics, Policy, and Law, 10(3), 267–292. Treisman, A. M., & Gelade, G. (1980). A feature-integration theory of attention. Cognitive
Mhurchu, N., Volkova, E., Jiang, Y., Eyles, H., Michie, J., Neal, B., et al. (2017). Effects of Psychology, 12, 97–136.
interpretive nutrition labels on consumer food purchases: The Starlight randomized van Kleef, E., & Dagevos, H. (2015). The growing role of front-of-pack nutrition profile
controlled trial. American Journal of Clinical Nutrition, 105, 695–704. labeling: A consumer perspective on key issues and controversies. Critical Reviews in
Michail, N. (2016). Dutch industry-backed nutrition label ticked of with mounting criti- Food Science and Nutrition, 55, 291–303.
cism. Avaiable at: https://www.foodnavigator.com/Article/2016/04/22/Dutch- van’t Riet, J., Sijtsema, S. J., Dagevos, H., & De Bruijn, G. J. (2011). The importance of
industry-backed-nutrition-label-ticked-off-with-mounting-criticism# Accessed March habits in eating behaviour. An overview and recommendations for future research.
4th, 2018. Appetite, 57, 585–596.
Ministère de l’Agriculture et de l’Alimentation. (2017). Nutri-score. Dossier de presse. Williams, L. G. (1966). The effect of target specification on objects fixated during visual
Available at: http://agriculture.gouv.fr/telecharger/87510?token= search. Perception & Psychophysics, 1, 315–318.
537276c9f22122955add7292abbcce0d Accessed December 10th, 2017. Wogalter, M. S., Dejoy, D. M., & Laughery, K. R. (2005). Organizing theoretical frame-
Ministerio de Industria Energía y Minería, Consulta pública sobre proyecto de decreto work: A consolidated communication – Human information processing (C-HIP)
relativo a Rotulado de alimentos envasados Available at: http://www.miem.gub. model. In M. S. Wogalter, D. M. Dejoy, & K. R. Laughery (Eds.). Warnings and risk
uy/-/consulta-publica-sobre-proyecto-de-decreto-relativo-a-rotulado-de-alimentos- communication (pp. 13–21). London: Taylor & Francis.
envasados 2017 Accessed August 17th, 2017. Wolfe, J. M., Palmer, E. M., & Horowitz, T. S. (2010). Reaction time distributions con-
de Salud, Ministerio (2015). Decreto número 13, de 2015. Santiago de Chile: Ministerio de strain models of visual search. Vision Research, 25, 1304–1311.
Salud. Yantis, S. (1992). Multielement visual tracking: Attention and perceptual organization.
Nestle, M., & Ludwig, D. S. (2010). Front-of-package food labels: Public health or pro- Cognitive Psychology, 24, 295–340.
paganda? JAMA, 303(8), 771–772. Yantis, S. (2000). Goal-directed and stimulus-driven determinants of attentional control.
Pan American Health Organization (2016). Pan American Health Organization nutrient In S. Monsell, & J. Driver (Vol. Eds.), Control of cognitive processes: Attention and
profile model. Washington, DC: Pan American Health Organization. performance: Vol. XVIII. Boston, USA: MIT.

225