“The impact of the Nutri-Score label on the healthiness choice of food and the moderating role of consumers’ health consciousness level”

NUTRI-SCORE:

5 LETTERS AND COLORS TO IMPROVE HEALTHY FOOD CHOICES

“The impact of the Nutri-Score label on the healthiness choice of food and the moderating role of consumers’ health consciousness level”

Margriet Sebens

University of Groningen, January 13

, 2020

2

NUTRI-SCORE:

5 LETTERS AND COLORS TO IMPROVE HEALTHY FOOD CHOICES

“The impact of the Nutri-Score label on the healthiness choice of food and the moderating role of consumers’ health consciousness level”

Master Thesis – Final Version University of Groningen Faculty of Economics and Business

Department of Marketing Msc. Marketing Management

Author: M.E. (Margriet) Sebens Adress: Westerhavenstraat 3/1

9718 AJ Groningen

E-mail: m.e.sebens@student.rug.nl Phone number: +31627362550

Student number: S2756382

1

Supervisor: M.T. van der Heide (Msc.) 2

Supervisor: Prof. dr. ir. K. van Ittersum

Completion Date

13.01.2020

Abstract

This research investigates the impact of the Nutri-Score label on the healthiness of food choices and the moderation role of health consciousness (low, moderate or high). The Nutri-Score label assigns healthiness scores (dark green label A (‘healthy’) to red label E (‘unhealthy’)) to food products. An online experiment was conducted in which participants chose one product per food category (breakfast cereals, desserts and snacks). Participants in the Nutri-Score condition (intervention) were presented with the label, subsequently their health consciousness was measured. We conducted a 2x3 ANOVA and 2x3 ANCOVA analysis that analyzed all of our hypotheses, and a PROCESS Macro analysis that correctly tested the first and third hypotheses.

Results revealed that the main effect of the Nutri-Score label and interaction effect between the Nutri-Score labels and health consciousness on healthiness of food choices were both insignificant.

Furthermore, both analyses did show a significant direct (linear) main effect between health consciousness and healthier food choices. The higher individuals become in their health consciousness, the healthier their food choices. Additionally, high health consciousness compared to low health consciousness significantly leads to healthier food choices. This research contributes to the existing literature by examining the moderation effect of Nutri-Score labels and health consciousness on the healthiness of food choices. Recommendations and the need for further research are given (to insignificant hypotheses), as interest regarding this label is growing within the food domain.

Keywords:

4

Table of Content

Abstract…....………... p.3

Introduction……… p.5

Theoretical Framework………... p.9

Food Choices……….………... p.9

Nutrition Labeling……….. p.9

Nutri-Score Label………...…. p.11

Health Consciousness………...…………... p.14

Methodology………...……….…... p.17

Participants………...……….….. p.17

Research Design………...……….….. p.19

Procedures…...………...………. p.21

Results………...………....……….. p.23

General Analysis………...………... p.23

ANOVA………...……….... p.25

Process Macro...………...……….……….. p.28

ANCOVA...………...……….………. p.29

Discussion………...……… p.32

Implications………..……… p.34

Limitations & Future Research……….……… p.36

Conclusion………...……… p.39

References………...…………...…………. p.40

Appendices………...………... p.47

Appendix A Questionnaire………... p.48

Appendix B ANOVA……… p.56

Appendix C PROCESS Macro (continuous)….………..……… p.56

Appendix D ANOVA………..…………. p.57

Introduction

According to the World Health Organization (2018a), obesity rates are rising worldwide.

Unhealthy diets, overweight and obesity contribute largely to non-communicable diseases, such as cardiovascular diseases and some cancers (Kelly & Jewell, 2018). Moreover, these diseases are estimated with a public health costs of US$30 trillion in the next 20 years (Bloom et al., 2011). All together, these diseases represent the main causes of deaths in the WHO European Region (Kelly

& Jewell, 2018). In light of the rising obesity rates worldwide, an intervention should be identified and implemented by food marketers and policy makers in order to improve the healthiness of food choices by consumers (Aschemann-Witzel et al., 2013; Ikonen, Sotgiu, Aydinli & Verlegh, 2019);

Kanter, Vanderlee & Vandevijvere, 2018).

One commonly recommended approach is to nudge consumers toward healthier food choices by providing easy to perceive, understandable and interpretable information on the nutritional content of food products (Ikonen et al., 2019). A survey in the US indicated that the majority of the Americans are confused when they obtain contradicting information on what counts as a healthy food choice (Jacqueline Howard, 2017). Furthermore, supermarkets need to provide clearer information on nutrition content, in order to prevent obesity (NOS, 2018).

One of the policy tools to rebalance of the unhealthful retail food partially, is to provide clearer information by the implementation of nutrition labeling that supports the consumers to make informed healthier food choices (Kelly & Jewell, 2018). While nutrition labels are already perceived and implemented as a popular instrument among policy makers and companies, the consumer possesses positive attitudes towards these labels as well, particularly the nutrition label on the front-of-pack (FOP) (Aschemann-Witzel et al., 2013). FOP labels have increasingly received attention from public authorities and learned societies (Julia & Hercberg, 2017).

In order to tackle these society challenges, the implementation of a newly developed FOP

nutrition label, the Nutri-Score label, needs further research in order to ensure that this label

improves the healthiness of food choices by consumers. The first country that announced the Nutri-

Score label as official FOP label was France in 2017 by the Minister of Health (Julia & Hercberg,

2017). Recently, political interest has increased as countries such as Belgium, Switzerland, Spain

and Germany approved this label (NOS, 2019a). This label was initiated in order to support

consumers in making healthy food choices, and as an incentive for companies to reformulate the

6 ingredients in their products. The Nutri-Score label (fig. 1) is a color-coded FOP summary indicator providing nutritional information based on the overall healthfulness of a product (Julia

& Hercberg, 2018). An average score is calculated through a logarithm, and the product obtains one out of the five categories, ranging from dark green label A (healthy) to red label E (unhealthy).

In recent years, there has been an increase in literature comparing the effectiveness of FOP labels, however relatively little research exists on more recent FOP labels such as warning labels and summary graded (such as Nutri-Score) FOP labels. Moreover, existing literature does not consistently show the impact of the Nutri-Score label on the healthiness of food choices. Some studies found results in favor of implementing the Nutri-Score label, with reasons that the Nutri- Score label improved the ranking ability in nutritional quality of food items (Egnell, Talati, Hercberg, Pettigrew, Julia, 2018a; Egnell, Ducrot, Touvier, Allès, Hercberg, Kesse-Guyot, Julia 2018b; Egnell, Talati, Gombaud, Galan, Hercberg, S., Pettigrew & Julia, 2019). Others argued about the ineffectiveness of the label, due to inaccurate measure of the logarithm and the over generalizable aspect that could mislead the consumer (Ikonen et al., 2019; Julia & Herberg, 2017;

Southey, 2019). Moreover, some research recommends that nutrient-specific information compared to summary indicators are more easily used to identify healthier options, while other studies found the other way around (Ikonen et al., 2019). Furthermore, previous research indicated an existing gap whether knowledge of product’s relative healthfulness translates into healthier choice and purchase behavior, the second part of the FDA objectives for FOP labels (Ikonen et al., 2019). Hence, this study extends to current inconsistent research by investigating the impact of the Nutri-Score label on the healthiness of food choices.

Thus far, research on the impact of the Nutri-Score label on improvement in nutritional quality of food items has mostly been conducted in the US and other European countries; only one study investigated the impact in the Netherlands (Egnell et al., 2018a; 2019). Furthermore, research found that 76% of the consumers within the Netherlands are positive towards implementation of this label (van der Staak, 2019). Moreover, one of the largest supermarkets (Albert Heijn) already supported the introduction of this label within the Dutch market (Albert Heijn, 2019), implicating the relevance of consistent results from more than one study.

Figure 1. Nutri-Score Label

Even though inconsistent results on the Nutri-Score label exists and only one study on the Nutri-Score in the Netherlands exists, the State Secretary of Public Health Paul Blokhuis announced (on the 28th of November 2019) that Nutri-Score will be introduced and implemented in the Netherlands in 2021 (NOS, 2019b). This implicates the relevance and importance of consistent results from more than one study, before expanding the label to the Dutch market. Thus, this study will contribute to the current literature by investigating the impact of the Nutri-Score label among Dutch consumers.

Previous research mainly focused on how FOP labels work and external environment influence the consumer behavior (Ikonen et al., 2019; Wansink, 2017). They do not consider consumer characteristics such as health consciousness, in particular to the healthy choices (Ellison, Lusk & Davis, 2013; Wansink, 2017). As literature on Nutri-Score label is still in its initial phase, this study will be the first to examine the moderation effect of health consciousness on Nutri score labeling to healthier food choices. It remains unknown whether the impact of the Nutri-Score label on healthiness of food choices is similar across all levels of consumers’ health consciousness. The label might especially be helpful to individuals who lack knowledge on the healthiness of products and find it difficult to choose the healthier option, due to the simplified summarized colored indicator of Nutri-Score that makes it easy to compare. It is important to better understand how consumers’ health consciousness influences the effect of Nutri-Score labeling on healthiness of food choices. Hence, this research will contribute to current literature by investigating this moderating effect of health consciousness on Nutri-Score labeling to healthier food choices.

Therefore, this study will discuss the following research questions:

What is the impact of the Nutri-Score label on the healthiness of food choices? How will the effect

of the Nutri-Score label be moderated by the health consciousness level of the consumer?

8

To sum up, this study addresses the existing inconsistency and contributes to the relatively

few papers that have studied the impact of the Nutri-Score label, moderated by the health conscious

consumer characteristic on the healthiness of the food choices. First, this article briefly discusses

food choices, nutrition labeling, Nutri-Score labeling and the concept of health consciousness

explaining the conceptual model (fig. 2) presented at the end of the theoretical framework. Second,

a designed online experiment is provided in the methodology section, that offers insights to

different stakeholders trying to understand the impact of the Nutri-Score label on healthiness of

food choices, moderated by health consciousness, relevant to different product categories. Lastly,

the discussion part will present implications, limitations, future research and a conclusion.

Theoretical Framework

Food Choices

Consumers are confronted with food choices every day. Consumers increasingly possess healthy eating intentions and desired health outcomes (IFIC Foundation, 2018; Ogden, Karim, Choudry & Brown, 2006). Contrarily, diseases caused by unhealthy eating still contribute to rising rates of overweight and obesity. This indicates that consumers experience difficulties in making healthy food choices.

Food choices made within the store depend on different factors, of which some vary over time (Okoro, Musonda & Agumba, 2016). First, there are factors based on nature such as physiological and nutritional needs, biological determinants (e.g. hunger, taste, appetite), psychological determinants (e.g. mood, stress and guilt) and attitudes, beliefs and knowledge about food. Second, there are factors based on nurture such as physical determinants (e.g. access, education, skills and time), social determinants (e.g. culture, family, meal patterns) and economic determinants (e.g. income, cost, availability; The Factors That Influence Our Food Choices:

(EUFIC)", 2006).

Different interventions and nudges may influence these factors to stimulate healthy food choices, such as economic incentives (e.g. taxes) and nutritional education (Cadario & Chandon, 2019). Specifically, interventions based on the visibility of food package within the store are a popular instrument in nudging consumers toward healthy food choices. The initial contact between the individual and the product is the visibility of the food package, this influences the perception of the quality of the product and the attention that a product display attracts (Shepherd, Sparks, &

Raats, 1991). This in turn influences food decisions (Cohen & Babey, 2012). Consequently, the nutritional label on front of the package may positively impact the consumer by correctly informing about the nutritional content that in turn may influence the healthfulness perception of the product and consequently their food choices.

Nutrition Labeling

Informing consumers about the nutritional content of the product, while still maintaining

consumers’ freedom of choice, is considered as an important and popular instrument to stimulate

healthier eating. An example of such an intervention method is nutritional labeling. This method

provides consumers with information about what nutrients are in the product, and thus improves

10 consumers’ ability to access and process nutrition information, that in turn is required to make health-conscious food decisions (Ikonen et al., 2019).

The purpose of regulations on nutritional labeling information (NFP), content per 100g, on the back or side of the package, (in energy, simple sugars, carbohydrates, (saturated) fat, proteins and sodium, and fibres as an option) is to effectively change consumers’ behavior towards healthy food choices (Julia et al., 2015; Nathan, Yaktine, Lichtenstein & Wartella, 2012). However, literature reveals that it is mostly perceived as difficult to interpret, not easily understood and complex. Moreover, due to time pressure at point of sale, consumers struggle to understand the information presented. Also the understanding of the content differs per country (Grunert, Fernández-Celemín, Wills, genannt Bonsmann & Nureeva, 2010; Ikonen et al., 2019; Julia et al., 2015). This back-of-pack nutrition labeling is mandatory in most countries, while only a fraction of consumers uses this for their food choices (Aschemann-Witzel et al., 2013; Julia et al., 2017b;

Roberto et al., 2012). Thus, the effectiveness to change behavior toward healthy food choices might be improved by making the nutrition information less difficult to interpret.

Front-Of-Pack (FOP) is known as a more effective tool compared to the NFP and are created to complement traditional complex NFP, presented at the front of the package. Examples of FOP labels are the Health Star Rating, Choices and Nutri-Score labels. FOP labels help for guidance toward food evaluation and healthier food choices through its simple, easily viewable and interpretable format at the point-of-purchase (Ikonen et al., 2019; Julia & Hercberg, 2017;

Kanter, Vanderlee & Vandevijvere, 2018). Additionally, saving cognitive effort by easier-to- understand product category layouts may direct consumers to healthier choices (Wansink, 2017).

In addition, consumers often tend to ignore the extensive NFP when a FOP label is presented, since

purchase decisions in supermarkets are most of the time made quickly and consequently consumers

will not study both nutrition sources in detail (Ikonen et al., 2019). Furthermore, compared to NFP,

nutrition information presented by FOP label received more attention (Vanderlee & Hammond,

2014). However, the type of FOP labels do differ in their effectiveness based on public health

message, and nutrient focus (Ikonen et al., 2019; Kanter, Vanderlee & Vandevijvere, 2018).

Nutri-Score Labeling

This research focuses on the evaluative interpretative Nutri-Score label provided at the front of the package (FOP) (fig.1), a cognitively-oriented intervention on nutrient information which promotes healthy food choices. This label seeks to influence what consumers know (Cowburn & Stockley, 2005; Cadario & Chandon, 2019), while consumers still experience freedom of choice, as the nutrition labeling intervention method is relatively inexpensive, simple and has relatively few modifications to the choice environment (Aschemann-Witzel et al., 2013;

Cadario & Chandon, 2019). The label summarizes the overall healthiness of the food product into a single indicator, and does not display other specific nutritional information, providing easy interpretation and facilitating understanding in order to easily compare alternatives. The overall health score is computed by the principles of a scoring as well as a threshold method taking into account the qualifying and disqualifying elements (van der Bend & Lissner, 2019). Moreover, this summarizing label adds a layer by adding the component of five colors (dark green to red) and letters respectively (A to E), of which products fall into one of these five categories (Ikonen et al., 2019).

However, some studies have found mixed results on the effectiveness of Nutri-Score Label.

First, studies have questioned the current calculation of the logarithm and the over generalizable aspect of the label (Ikonen et. al., 2019; Julia et al., 2017b; Southey, 2019). For example, the score of nutritional quality was not consistent for the food groups cheese, beverages and added fats (Julia et al., 2017b). Second, the Nutri-Score appears to be less trustworthy to individuals with higher educational level or substantial knowledge, as no specific nutrient information with numerical information is provided (Egnell et al., 2019). Third, a weak point of Nutri-Score is that 28% of the population relates the green color of the label to an organic food source (Carreño, 2017).

On the other hand, even though the Nutri-Score label has not extensively been studied

before, the positive effect of Nutri-Score labeling should not be underestimated, as some studies

have proven the effectiveness of this relatively new label on the healthiness of food choices. First,

it is expected that the colored Nutri-Score label will have a positive effect on the attention to the

package, captured by consumers. Thus, this creates a perception of healthiness of food, that in turn

influences the healthiness of food choices. The green and red colors tend to capture the attention

of the consumers, because they correspond to recognized signals. This might makes more easy to

understand and interpret, as green is being related to safety and a “go” signal, and as red is being

12 related to danger and a “stop” signal (Egnell et al., 2018a, 2018b; Fehrman & Fehrman, 2004).

This effect was emphasized by the literature on Traffic Light Labeling (Balcombe, Fraser & Falco, 2010) that shows a strong resemblance with the Nutri-Score on the use of color and interpretation, based on similar effects for the Nutri-Score are expected. A priming effect of the color red exist, as consumers are used to this color in real life, with ‘red’ as a warning or stop signal, the opposite might occur for green. Studies on the Traffic Light Label measured the impact of color on the front of the package and proved the effectiveness of adding color interpretation to the label, as healthy items increased and unhealthy items decreased in shopping carts. This indicates healthier decisions at point-of-purchase, thus showing a positive spillover effect for the Nutri-Score Label (Sonnenberg et al., 2013; Thorndike et al., 2012).

Moreover, within the Nutri-Score label, letters were incorporated to ensure increased visibility of the label, in particular to individuals who have difficulties with colors (Kelly & Jewell, 2018). To conclude, the increased attention captured by the color and the summarizing component of the label making nutritional information more easily understood (knowledge) and better comprehend (interpretation) will have a positive impact on the healthiness of food choices. Second, research suggests that interpretative labels are more effective compared to other FOP label types in making healthier choices (Ikonen et al., 2019). In particular, summary indicators are associated with a diminished cognitive workload, causing faster processing, less difficulty to understand the meaning of the information and are highly effective to compare different product alternatives.

Consumers should be able to easily and quickly compare different alternatives within the same product category at point-of-purchase decisions that benefits their health (Balcombe et al., 2010;

Egnell et al., 2018a; Ikonen et al,, 2019). Compared to the nutrition labels based on specific information, summarizing components are more easily understood and limit confusion related to interpretation of nutritional terms (e.g., saturated fats, sugars, and sodium) (Egnell et al., 2018b).

These findings are supported by two studies that use an experimental design, and found that Nutri-

Score led to the highest improvement of nutritional quality items and no effect was observed in

the number of items in the shopping card, followed by the Multiple Traffic Light label and Health

Star Rating system, GDA and Check (Egnell et al., 2018a, 2018b; Julia & Hercberg, 2017; Talati,

et al., 2019). Additionally, Nutri-Score was easiest in identifying the healthiest product and was

most likely to be understood among the participants, compared with GDA, MTL and Green Tick

(Ducrot et al., 2015a, 2015b). In terms of context between different countries, Nutri-Score

produced highest improvement in ranking ability on nutritional quality across all three food categories studied, followed by MTL, HSR and Warning Symbol observed in all 12 countries (Egnell et al., 2018a). Moreover, studies have proved that Nutri-Score achieved the best improvement in households with the lowest income, additionally it led to improvement in all subgroups of the population (specifically to individuals who purchase discount brands) compared to mixed results of other formats (Julia et al., 2017b). It is easily recognized and interpreted despite of socioeconomic and demographic status, in particular compared to the Traffic Light System (Goiana-da-Silva, 2019). Nutri-Score presents the strongest evidence to promote healthy food choices among consumers, namely those individuals that obtain unhealthier eating behaviors.

Evidence indicates that Nutri-Score guides consumers to adequately classify food products according to their nutritional quality, even to consumers who do not obtain technical knowledge on nutrition (Goiana-da-Silva, 2019).

To conclude, literature suggests that this type of labeling has a positive effect on the healthiness of food choices. As healthfulness information is displayed in a simple way, making it easier to compare and choose the most healthy option out of all alternatives at point-of-purchase.

Regarding the impact of the Nutri-Score label on the healthiness of food choice, the first hypothesis therefore will be:

H1: The implementation of the Nutri-Score label will positively impact the healthiness of food

choices by consumers.

14 Health Consciousness

Little empirical research has been conducted on the consumer characteristic ‘health consciousness’ that may moderate the relation between the Nutri-Score and the healthiness of food choices. ‘Health consciousness’ is defined by literature as “the motivational component that predicts a variety of health attitudes and health behaviors, the quality of food attributes consumers consider and the health actions consumers undertake (Mai and Hoffmann, 2012, p.318).”

Wansink (2017) considered health consciousness as an important characteristic that contributes to the effectiveness of the intervention to increase the sales of healthy foods.

Consumers differ in how liable they are towards making healthier food decisions, which implicates the presence of different levels in health consciousness. Consequently, results in healthiness of food choices differ in their effectiveness across different levels of consumer characteristics.

Moreover, Van Herpen & Van Trijp (2011) argue that due to time constraints (at point-of- purchase) and motivation of the individual, the attention on FOP labels differ. Hence, the motivational concept relates to the health conscious level of the individual, influencing the attention on the FOP label. Thus, this affects the healthiness perception of the product, which in turn impacts the healthiness of food choices. The Nutri-Score label could therefore positively impact healthiness of food choices, however the strength of the relationship may differ across individual health consciousness levels. Implementing these consumer characteristics and measuring the influence of these characteristics on the strength of H1 adds to a more realistic view, to who this label will (not) have an impact to (Wansink, 2017).

Individuals who possess low levels of health consciousness, are characterized as health disinterested shoppers, individuals who have little interest to change their behavior due to either the effort, sacrifice, or perceived furtility (Ellison et al., 2013; Wansink, 2017). These individuals may not be interested in acting healthy, and therefore may prefer, when choosing a product, other aspects such as taste or price over the healthiness of a product. Furthermore, this group might largely ignore nutritional information, as it will not even catch their attention (Lone, Pence, Levi, Chan & Bianco-Simeral, 2009; Wansink, 2017). Hence, the group of low health conscious individuals is expected to have a small effect of the implementation of the Nutri-Score label on the healthiness of food choices.

Individuals who possess high levels of health consciousness, are characterized as ‘health

vigilant shoppers’, individuals who possess large amount of knowledge on nutrition. This group is

“highly informed, are conscious of calories, are influenced by nutrition information and already have an incentive to act healthy and choose healthy products (Wansink, 2017, p.66).” FOP labels mainly simplify search processes for those consumers who are already interested in buying healthier products (Aschemann-Witzel et al., 2013; Ikonen et al., 2019). In addition, Ellison (2013) found that menu label have minimal influences on food choice to these high health conscious individuals that obtain high levels of nutrition awareness. Thus, individuals who possess large amount of nutrition knowledge will be minimally influenced by (menu) labels. Therefore, it can be expected that this effect will spill over to front of pack labels, and in particular to Nutri-Score labels as well. Moreover, literature suggests that high health conscious consumers tend to ignore the simplified FOP information and appears to be less trustworthy, and prefer the use of the back- of-package, lengthy detailed/specific/numerical NFP information in making their food decisions, as they are more knowledgeable and understand BOP (Egnell et al., 2019; Ikonen et al., 2019).

Hence, the group of high health conscious individuals is expected to have a small effect of the implementation of the Nutri-Score label on the healthiness of food choices.

Contradictory to studies that focus on the linear spectrum of health consciousness and only differentiate between the two levels of ‘low’ and ‘high’ health consciousness (Ellison et al., 2013), it might perhaps be the group in the middle, the ‘moderate’ health conscious individuals, we should focus on. Wansink (2017), supports this view and indicates three levels of health consciousness in his “Hierarchy of Health Predisposition.” In the middle of the hierarchy are the health predisposed shoppers (moderately health conscious), who prefer to make healthier choices, but find it difficult to consistently choose the healthier food, unless they only have to make very little sacrifice (Wansink, 2017). In contrast to the low and high health conscious consumer where we expect no significant influence to H1, we do expect a more pronounced influence of the Nutri-Score label on the healthiness of food choices, when consumers are moderately conscious of their health.

The impact of the Nutri-Score label is therefore expected to differ across all levels on consumers’ health consciousness. Therefore, the second hypothesis is:

H2: The impact of the Nutri-score labels on the healthiness food choice of consumers will be more pronounced among moderately health conscious consumers compared to more and less health conscious consumers.

Lastly, literature on health consciousness implies that not only an indirect effect exists for this

variable; it may directly impact the healthiness of food choice as well. People may vary in their

16 health consciousness level, as people have different personal characteristics. Therefore, their outcome on the healthiness of food choices may differ as well. Contradictory to H2, H3 emphasizes on a linear positive relation of health consciousness on healthiness of food choices.

Higher health conscious consumers already possess a large amount of nutrition knowledge and health awareness, while lower health conscious consumers may lack the knowledge on healthfulness (Ellison et al., 2013). However, according to Ikonen et al. (2019), obtaining high knowledge on nutrition information and awareness regarding the healthfulness of a product is not always resulting in healthier choices by consumers. The underlying effect on healthiness of food choices may be the interest and motivation (consciousness) of consumers. Health uninterested individuals make food decisions differently from those who are highly interested in the healthfulness of what they eat (Lone et al., 2009). Therefore, obtaining a positive attitude towards health interest creates consumer ratings to food differently, compared to those with a negative attitude (Roininen, Lähteenmäki & Tuorila, 1999). Consumers with high health interest therefore tend to make more healthful food choices and are more health oriented (Roininen et al., 1999).

Therefore, it is hypothesized that:

H3: Health Consciousness has a direct positive influence on the consumers’ healthiness of food choices.

Fig 2. Conceptual Model

Methodology

This study investigates the relationship between the Nutri-Score label and the healthiness of food choices. In which Nutri-Score label is the independent variable, health consciousness is the moderator and healthiness of food choices is the dependent variable.

Participants

We recruited a total of 196 adult participants (18 years or older) through social media

(Facebook, Instagram and WhatsApp) to conduct our online survey. As the Nutri-Score logo

consists of and captures attention by the colors, people who are color blind are excluded from the

dataset. Moreover, people who had food related allergies such as lactose intolerance or allergies

to nuts and gluten were excluded as some of the products in the experiment contain these

ingredients. Furthermore, respondents who did not complete the survey and participants from the

pre-test were excluded as well. After screening the data of these participants, 156 usable returns

were collected to analyze. The age of respondents ranged from 19 to 71 years old, either male or

female (38.5% and 61.5% respectively). Participants were mostly located in the Netherlands, with

an average age of 28 years old (M

= 28.6 years, SD

= 11.50)(table 1). Furthermore, most

participants are educated, ranging from high school to master’s degree of which the majority

(40.8%) obtained their bachelor’s degree as their highest level of education. Data was collected

before the announcement of State Secretary of Public Health Blokhuis (28th of November 2019)

on the introduction and implementation in 2021 of the Nutri-Score label in the Netherlands and

will therefore not affect the results from this research.

18 Table 1 – Descriptive of the research population

Variable Mean SD

Age 28.26 11.50

Nutrition information knowledge

3.10 1.56

Health Consciousness 4.87 0.91

Gender Frequency Percentage

Male 60 38.5

Female 96 61.5

Residence

Netherlands 151 96.8

Outside Netherlands 5 3.2

Education

High School 7 4.5

Secondary education 8 5.1

University of applied sciences 32 20.5

Bachelor’s degree (WO) 63 40.4

Master’s degree 45 28.8

Doctorate 0 0

Other 1 0.6

Research Design

The study used a 2x1 between-subject-design. We applied a one-way factorial between- participants design, respondents were randomly assigned to one of the two conditions, either with Nutri-Score (n = 77) or without Nutri-Score (n = 79) (1-5 healthiness product) (Appendix A).

Participants presented with the Nutri-Score label were asked at the end of the study whether they noticed the Nutri-Score label. Furthermore, a general question for participants in all conditions asked whether they were familiar with the Nutri-Score label. The moderator health consciousness was included as measured variable.

The design of the Nutri-Score label (fig. 3) was replicated from existing sources already implementing the label to the products. These sources are Delhaize (one of the largest (online) supermarkets in Belgium)(https://www.delhaize.be/nl-be/), and Consumentenbond (Dutch Non- Profit Organization to protect consumers choices by informing them) who both used the calculation by a computed logarithm of the Nutri-Score (Cammelbeeck, 2019). The label consists of five colors, ranging from dark green to red and five letters, ranging from A (most healthy) to E (most unhealthy)respectively. The type of healthiness of the products used in this research depends on one of the five healthiness labels the product is designated to regarding replicated the sources used. Back of the package labels were not shown, in order to mimic the shelf and online grocery shopping when products are being compared based on the front of the package. Prices are still presented below the product, in order to keep the experimental setting as realistic as possible. The Nutri-Score label is either presented below the product when participants are assigned to the presence of the label condition, or the Nutri-Score label is not presented when participants are assigned to the control condition (absence of the label) (fig. 4).

Fig 3. Breakfast Cereal with and without Nutri-Score label

20 Conducting research with high varieties between and within product categories was an important criteria to select the products used in this study. In total three product categories were selected based on the categories used in the article of Nikolova & Inman (2015) and based on the information provided by Cammelbeeck (Consumentenbond), who mentioned that breakfast cereals are confusing to consumers in day-to-day life (2019). Furthermore, the product categories chosen in this study are ‘ready to eat’ and do not need to be prepared before consuming (e.g. baking, frying, cooking). Since the products chosen are ‘ready to eat’ at point-of-purchase, the nutritional value will remain the same, thus making the logarithm of the Nutri-Score for these products valid and relevant for this experiment. Hence, the products categories used in this study are: breakfast cereals, desserts and snack foods. The categories represent a sufficient variety of different food categories that vary in their health and their nutrition scores across the product assortment in each category (Nikolova & Inman, 2015).

The main criteria for the products chosen within a category was the variation of healthiness within the category, all five levels of healthiness (label A to E) needed to be represented within one question.

Lastly, the Nutri-Score label was pre-tested among 10 participants. The required

improvements were added to the final version of the survey.

Fig 4. Question (survey) on snack choice with and without label condition and products are randomized, displayed in the survey.

Procedures

All participants conducted a Qualtrics survey that would take approximately 5 minutes to complete and is conducted at their time of preference. First, informed consent was obtained from the participants (Appendix A). Next, participants were randomly assigned to the manipulated condition of the Nutri-Score label or to the control condition (without Nutri-Score label). After being randomly assigned, participants started with the experiment.

The Nutri-Score label intervention was implemented to find out whether Nutri-Score label

influenced the healthiness of food choice. In the experiment, participants were asked to select one

product of their preference that they would normally buy as well. Resulting in three product

choices out of the 15 options over three product categories presented, these choices together

formed the dependent measure of this research. The products within each category were

randomized in one single row, thus not presented from healthy to unhealthy, in order to simulate

a real-life shopping trip in the (online) supermarket.

22 After the experiment participants were asked “How often do you normally buy the following products?” which were rated by participants on a five-point Likert scale (1 = “Never”;

5= “Always”).

Next up, the moderator ‘Health Consciousness” was measured on nine items, using a seven point Likert rating scale where 1 = “not at all” (low health consciousness) and 7=”very well” (high health conscious)(Gould, 1988). Such as, “I reflect about my health a lot.”, and “I’m very involved with my health.” In order to prevent participants from understanding what moderator would be measured, we asked these statements after the experiment.

After this, questions on mood and price sensitivity were asked. In order to measure the influence of mood during the shopping within the experiment, the item “How do you feel at this moment?” was asked (Aarts & Dijksterhuis, 2003). A seven-point Likert scale is used, such as

‘bad - good’. In order to measure the price sensitivity of participants, they were asked to indicate on three statements to what extent they agree or disagree. Such as, “For me, price is decisive when I am buying a product.” A seven point Likert scale is used in which, 1 =”Strongly Agree”, 7=”Strongly Disagree”. The three items for price sensitivity were reverse coded to 1=”Strongly Disagree” and 7=”Strongly Agree.” Next, participants were asked how hungry they felt at the moment, seven-point Likert scale (Finkelstein & Fishbach, 2010) with anchor points 1=”Extremely Hungry”, 7=”Extremely Satisfied.” In order to measure the consideration of nutrition information on food choices, the item “I consider nutrition information when making a decision about which foods to eat.” was used. 1=”Strongly Agree, 7=”Strongly Disagree.” This item was reverse coded to 1=”Strongly Disagree” and 7=”Strongly Agree.” Moreover, frequency of physical activity was asked on a one item five scale 1=”Daily”, 5=”Never.”

Respondents in both conditions were asked about the familiarity with the Nutri-Score label, one item on a five point Likert scale 1=’extremely familiar’ and 5=’ not familiar at all’. This item was reverse coded to 1=’not familiar at all 5= extremely familiar. A manipulation check question was only displayed to Nutri-Score group, if they noticed the label (“yes”, “I don’t remember” or

“no”). Moreover, demographics such as gender, age, whether they currently live in the Netherlands, and their highest level of education were obtained.

Lastly, participants were debriefed about the purpose of the study (Appendix A).

Results

This section will discuss the results of the survey. First, we inspected the data on missing values, outliers and odd figures. Second, we recoded reverse coded items. Third, a reliability analysis measured a reliable overall health consciousness index (α = 0.90, n = 9 items). Next, we conducted a descriptive analysis, a 2x3 ANOVA, a PROCESS Macro analysis and a 2x3 ANCOVA.

Descriptive Analysis

To examine our first hypothesis, whether participants healthiness of food choices increased by displaying the Nutri-Score label, we created a weighted average of the nutrition score of each participant’s product choices (from 1-5, in which 1 = label A ‘most healthy’ and 5 = label E ‘most unhealthy’) across the three product categories: breakfast cereals, desserts and snacks. Participants were either assigned to the control condition (label absent) or the Nutri-Score condition (label present). Thus, we had two observations: the consumer’s individual volume-weighted average nutrition score of the food choices across three categories in either the control condition and the Nutri-Score condition.

On average participants were higher in health consciousness (M = 4.87) (on a seven-point Likert scale) and chose averaged healthy products (M = 2.80) (on a weighted average score of 1- 5, in which 1 = label A ‘most healthy’ and 5 = label E ‘most unhealthy’). Thus, the lower the weighted average score, the healthier the food choice. More specifically, across all three food categories, the participants within the Nutri-Score condition scored healthier (M = 2.74, SD = 0.76) compared to the participants within the control condition (M = 2.87, SD = 0.82) (fig.6). In table 2 the means of variables of interest are presented.

Hence, these descriptive results provide initial support for H1. However whether this effect

is significant still remains unanswered, this will be analyzed and answered in the following

analyses. First, we will present the results on the 2 x 3 ANOVA analysis that tests all our

hypotheses. In order to analyze our differences between the three groups in our moderator, we

visually binned (using two cut points) the continuous average weighted health consciousness

variable into three categories consisting an equal amount of participants per group (low, moderate

and high). After this, the results on the PROCESS Macro (regression) analysis (using the health

consciousness variable as continuous variable) will be presented. Since there is no hard evidence

24 that a linear interaction exists, we will analyze the data in ANOVA using the binned health consciousness variable and using the continuous health consciousness variable in PROCESS Macro compare findings. Moreover, the PROCESS Macro will test the group more specifically.

Table 2.

Variable Mean SD Min. Max.

Healthiness of food choices* 2.80 0.79 1.00 4.67 Nutri-Score Condition* 2.74 0.76 1.00 5.00

Control Condition* 2.87 0.82 1.00 5.00

Health Consciousness 4.87 0.91 1.78 7.00

Hunger 4.47 1.48 2.00 7.00

Nutritional information use 4.90 1.56 1.00 7.00 Frequency Purchase Breakfast

Cereals*

3.03 1.21 1.00 5.00

Frequency Purchase Desserts* 2.46 0.93 1.00 5.00 Frequency Purchase Snacks* 2.97 0.84 1.00 5.00 Average Frequency Purchase

product categories*

2.82 0.60 1.00 4.67

Price sensitivity 4.73 1.22 1.67 7.00

Mood 4.53 0.97 1.00 7.00

*All variables are based on a 7 point-Likert scale (1= negative, 7= positive), except* healthiness of food choices, Nutri-Score condition and Control Condition (1-5), in which 1=A and 5=E and frequencies that are based on a 5 point-Likert scale (1= never, 5= always).

Fig. 6 Estimated marginal means of healthiness of food choices for both label conditions 2,74

2,76 2,78 2,80 2,82 2,84 2,86 2,88

Control Condition Nutri-Score Condition Healthiness of food choices (higher = unhealthier)

Label Condition

Food choices general

2 x 3 ANOVA

Before performing a 2 x 3 ANOVA, we needed to check whether the data has met the six requirements in order to conduct this analysis. The first three assumption are met: the dependent variable is continuous (interval), the independent variables consist each of two or three categorical groups (dummy is dichotomous: control condition or Nutri-Score condition, health consciousness has three categories: low, moderate and high levels) and an independence of observations (participants are randomly assigned to the conditions).

The last three assumptions are partially met. First, we conducted boxplot analysis to examine any outliers, consequently no significant outliers were found. Second, there needs to be homogeneity of variance. The Levene’s test for equality of variances showed that there was homogeneity of variances (p > .05). Third, the dependent variable should be normally distributed for each combination with the groups of independent variables. The Shapiro-Wilk and Kolmogorov-Smirnov test of normality revealed both significant results (p < .05), consequently the null hypothesis is rejected, indicating no normal distribution. In addition, the visual inspection of the Normal Q-Q plots did not show major deviations from a normal distribution. According to Refinetti (1996) ANOVA tolerates violations to the normal distribution assumption relatively well.

However, the results of the ANOVA might be slightly inaccurate which we need to keep in mind.

We conducted the ANOVA, with using the average healthiness scores as dependent variable and independent variables the dummy variable of the label (present = 1, absent = -1) and health consciousness binned variable are used as fixed factors.

A 2 (label: Nutri-Score Condition vs. Control Condition) x 3 (level of health consciousness: low vs. moderate vs. high) ANOVA on healthiness of food choices did not reveal a statistically significant interaction effect, F(2, 150) = 1.86, p = 0.16 (fig. 8). Due to the insignificant interaction effect of health consciousness and label condition on healthiness of food choices, we need to reject hypothesis 2.

The 2 x 3 ANOVA on the main effect revealed that the label condition, with Nutri-Score

condition indicating healthier food choices (M = 2.74) than the Control Condition (M = 2.87), has

no significant influence on the healthiness of food choices, F(1, 150) = 0.41, p = 0.52. Hence,

against the prediction of hypothesis 1, the label condition does not impact the healthiness of food

choices. Therefore, hypothesis 1 needs to be rejected.

26 The 2 x 3 ANOVA revealed a statistically significant main effect of health consciousness on healthiness of food choices, F(2, 150) = 3.12, p = 0.05. The pairwise comparisons found that the high health conscious group choose significantly healthier products compared to the low health conscious group (CI

= -.364; p = 0.014). Furthermore, the estimates of the health consciousness groups indicated a linear increase from low (M = 3.01) to moderate (M = 2.82) to high (M = 2.65) levels of health consciousness (graphical representation in fig. 7). Thus, in line with hypothesis 3, health consciousness does positively impact the healthiness of food choices.

Therefore, hypothesis 3 needs to be accepted.

Additionally, we analyzed marginal means using Syntax (fig. 9) in order to find the more detailed effect whether one group compared to the other group is more significant. The pairwise comparison found that when the Nutri-Score label is implemented, the high health conscious group (M = 3.08) choose significantly healthier products compared to the low health conscious group (M

= 2.45; CI

= -.632; p = 0.01). The Nutri-Score label significantly increases healthier food choices for high health consciousness (M = 2.45) compared to high health conscious participants where the Nutri-Score is not implemented (M = 2.85; CI

= -.398, p = 0.05).

Fig. 7 Estimated marginal means of healthiness of food choices per health consciousness level 2,4

2,5 2,6 2,7 2,8 2,9 3 3,1

Low Moderate High

Healthiness of food choices (higher = unhealthier)

Health Consciousness Level

Healthiness of food choices

Fig. 8 Moderation effect of health consciousness on control condition (label absent) vs. Nutri-Score condition (label present) on healthiness of food choices.

Fig. 9 Healthiness of food choices as a function of health consciousness level. (ANOVA) 0,0

0,5 1,0 1,5 2,0 2,5 3,0 3,5

Control Condition Nutri-Score Condition Healthiness of food choices (higher = unhealthier)

Label Condition

Healthiness of food choices

Low Health Consciousness Moderate Health Consciousness High Health Consciousness

0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5

low moderate high

Healthiness of food choices (higher = unhealthier)

Health Consciousness Level

Healthiness of food choices

Control Condition Nutri-Score Condition

28 PROCESS Macro

Next, the regression analyses PROCESS Macro for SPSS v3.3 (Model 1) by Andrew F. Hayes (2013) was performed. First, model 1 was conducted to validate the results of the 2x3 ANOVA analysis. Moreover, as we hypothesized three levels of health consciousness, PROCESS Macro was used to find out whether a linear interaction effect instead of a non-linear interaction effect existed. We used the dummy variable of the label (present = 1, absent = -1) as independent variable and the calculated average of nutrition score of three product categories combined as the dependent variable, and as continuous health consciousness variable is used as moderator (Appendix C). The overall model is significant, R = 0.254, R

= 0.0643, p = 0.01. The regression analysis verify the ANOVA findings that that the main effect of Nutri-Score label, p = 0.33, is not significant (fig.

10). In line with the finding in the ANOVA analysis, the interaction effect between Nutri-Score label and health consciousness on healthiness of food choices is not significant, F<1, p = 0.18. In order to understand the interaction, we elaborated on this by examining the Johnson-Neyman technique (Johnson & Neyman, 1936). We did not find any points of significance for individuals low and moderate in health consciousness (p < 0.10). We did find a marginally significant point for individual high (+1 SD) in health consciousness (p = 0.10). However, the Johnson-Neyman technique did not find any points of significance, indicating that on a larger spectrum of values (e.g. +1,5 SD) we can not a significant effect for individuals high in health consciousness. Hence, this proves that (even for values on a larger spectrum) there is no evidence for a significant interaction effect between the label and health consciousness. Consequently, hypotheses 1 and 2 are rejected.

The results of the regression confirm the significant main effect of health consciousness on

healthiness of food choices that was also found in the ANOVA, p = 0.05. Furthermore, in line with

the performed ANOVA, high health conscious levels are expected to have a marginally significant

effect on healthiness of food choices, p = 0.10. Consequently, hypothesis 3 is accepted by the

analyzed data. In line with the performed ANOVA and the PROCESS Macro using the health

consciousness as a continuous variable, health conscious levels have a significant positive effect

on healthiness of food choices, p = 0.05. Hence, in line with ANOVA results, PROCESS Macro

confirmed the significant linear effect of hypotheses 3 and is therefore accepted.

To conclude, the higher the levels of consumers’ health consciousness indicate healthier food choices made by consumers. Moreover, in line with the ANOVA results, PROCESS Macro confirmed the insignificant effect of hypotheses 1 and therefore cannot be accepted.

Fig. 10 Nutri-Score condition by health consciousness on healthiness of food choices.

2 x 3 ANCOVA

This study revealed that health consciousness increases healthiness of food choices. In order to check the robustness of the results, we included various control variables and performed the ANCOVA analysis. The control variables used are hunger, average frequency buy product category, gender, mood and price. These variables may influence food choices, and might explain their differences between individuals. First, gender is included as research found that women are more likely to make healthier food decisions and use nutritional information more often compared to men (Ikonen et al., 2019). Second, hunger is included as the degree of hungriness effects your decision on food. Research found that hunger highly influences attitudes toward high-fat foods (Lozano, Crites & Aikman, 1999). Furthermore, research found that mood and familiarity influences food choice motives. According to Sun (2008), people who are focused on their health prefer emotional enhancing and familiar food as well, next to the importance of healthy foods aspect. Lastly, research found that price sensitivity influences the use of labels, as the higher prices sensitive individuals tend to be less interested in labels and the use of them (Balcombe et al., 2010).

2,2 2,3 2,4 2,5 2,6 2,7 2,8 2,9 3,0 3,1

-1 SD moderate +1 SD

Healthiness of food choices (higher = unhealthier)

Health Consciousness

Impact of label condition on the health consciousness

Control Condition Nutri-Score Condition

30 These covariates might explain a difference in the healthiness of food choices. Thus, we performed an ANCOVA to compare the results with the ANOVA (Appendix D).

The 2 (label: Nutri-Score Condition vs. Control Condition) x 3 (level of health consciousness: low, moderate, high) ANCOVA confirmed that the interaction effect on healthiness of food choices is not significant, F(2, 145) = 1.71, p = 0.18. Thus, in line with the ANOVA analysis, hypothesis 2 is rejected.

The 2 x 3 ANCOVA confirmed that the main effect of label condition, with Nutri-Score condition indicating on average healthier food choices (M = 2.74) compared to the Control condition (M = 2.87), has no significant influence on healthiness of food choices, F(1, 145) = 0.53, p = 0.46. Thus, in line with previous analyses, the Nutri-Score label did not significantly change consumers’ behavior towards healthier food choices. Therefore, hypothesis 1 is rejected.

The ANCOVA revealed that the main effect of health consciousness becomes marginally significant when including control variables, F(2, 145) = 2.42, p = 0.09. However, still in line with previous analyses, health consciousness has an impact on healthiness of food choices. The pairwise comparison found that the high health conscious group choose marginally significant healthier products compared to the low health conscious group (CI

= -.330, p

0.09).

Furthermore, the estimates of health consciousness groups did still indicate, after inserting control variables, a linear increase from low (M = 2.98) to moderate (M = 2.80) to high (M = 2.67).

Therefore, in line with expectations, health consciousness does positively impact the healthiness of food choices, consequently hypothesis 3 is accepted. To conclude, these results confirm the robustness of our main findings.

Additionally, we analyzed marginal means using Syntax (fig. 11) in order to find the more detailed effect whether one group compared to the other group is more significant. The pairwise comparison found that when the Nutri-Score label is implemented, the high health conscious group (M = 2.47) choose significantly healthier products compared to the low health conscious group (M

= 3.07; CI

= -.591; p = 0.02). The Nutri-Score label significantly increases healthier food choices for high health consciousness (M = 2.48) compared to high health conscious participants where the Nutri-Score is not implemented (M = 2.86; CI

= -.387, p = 0.05).

Lastly, multicollinearity of the control variables was checked. This linear regression used

the mean of healthiness scores as dependent variable. We checked whether two or more

independent variables within our model were highly correlated. Results indicated that VIF scores ranged from 1.01 to 1.09. Hence, no concerns for multicollinearity.

Fig. 11 Healthiness of food choices as a function of health consciousness level. (ANOVA) 0

0,5 1 1,5 2 2,5 3 3,5

low moderate high

Healthiness of food choices (higher = unhealthier)

Health Consciousness Level

Healthiness of food choices

Control Condition Nutri-Score Condition

32

Discussion

This chapter provides a discussion and summary on the findings of this study about the Nutri-Score label and health consciousness. Furthermore, implications for consumers, public policy makers and marketers are discussed. Additionally, limitations and future research recommendations are discussed and lastly a conclusion is presented. The purpose of this research was to investigate the recently developed Nutri-Score label, and its’ interaction with the consumer characteristic health consciousness on the consumers’ healthiness of food choices. Consequently, the main research questions that were investigated are: What is the impact of the Nutri-Score label on the healthiness of food choice? How will the effect of the Nutri-Score label be moderated by the health consciousness level of the consumer?

Results of an online experiment among Dutch participants revealed that consumers’ health consciousness has a significant positive linear effect to the healthiness of food choices. Thus, in line with what we expected in hypothesis 3, we found that the higher people are in their health consciousness, the healthier their food choices are. Specifically, this effect was significantly more pronounced for high health consciousness consumers compared to the moderate and low levels of health consciousness. This research adds to current literature on consumers’ health consciousness, by indicating that the level of health consciousness does positively impact the healthiness of food choices.

Furthermore, contradicting to what we expected, we did not find a significant increase in

healthiness of food choices when the Nutri-Score label was displayed. Therefore, hypothesis 1 is

rejected. Little literature exists on the Nutri-Score label, however these studies found an

improvement in nutritional quality and ranking of Nutri-Score compared to other type of FOP

labels. Contradicting to what we expected in hypothesis 2, our results showed that high health

conscious participants had a more pronounced effect in making healthier choices compared to

moderate and low levels. Previous literature indicated that FOP labels would have a significantly

higher effect to consumers who lack knowledge on nutritional information compared to high in

health consciousness (who do have high levels of nutritional knowledge), due to the ease of the

FOP understanding and the fact that high conscious consumers would even neglect FOP label due

to their knowledge nutrition information NFP (Ellison et al., 2013; Egnell et al., 2019; Ikonen et

al., 2019). Thus, our results are surprising, however, both hypothesis 1 and 2 were not significant.

Consequently, we did not find results in line with what we expected for both hypotheses. These insignificant results could be explained by the relatively small range of our research sample group of 156 participants. Additionally, the majority of our sample is highly educated and aged between 20-30 years old, not showing a representative sample of society in which lower socioeconomic and older participants would have been integrated as well. Out of our three hypotheses, only one was fully supported. An overview of the acceptance and rejection of the hypotheses is presented in table 7.

Hypotheses Accept Reject

H

: The implementation of the Nutri-Score label will positively impact the healthiness of food choices by consumers.

X

H

: The impact of the Nutri-Score label on the healthiness food choice of consumers will be more pronounced among moderately health conscious consumers compared to more and less health conscious consumers.

X

H

: Health Consciousness has a direct positive influence on the consumers ‘healthiness of food choices.

✓

Table 7. Overview of hypotheses.

34 Implications

Our study has found important implications for consumers, public policy makers and marketers. Our research contributes to current literature by revealing a main effect of consumers’

health consciousness on healthiness of food choices (hypotheses 3). In particular, the consumers who indicate high levels of health conscious, showed a significant increase in healthiness of food choices across all analyses. Thus, higher health conscious people are focused on choosing the healthier option that is available and confirms the direct main effect of health consciousness on healthiness of food choices. Moreover, this result is in line with previous literature indicating that high health conscious consumers have an intrinsic motivation to educate themselves on nutrition and are better able to read it compared to lower levels of health consciousness (Mai and Hoffmann, 2012). This finding is relevant for different stakeholders such as policy makers and insurance companies who strive for a healthier population, thus preventing the nutrition related diseases (e.g.

obesity), that in turn reduces costs for society as well. This research found evidence that higher health conscious people make healthier food decisions, therefore policy makers are advised to find an intervention that stimulates low and moderate levels of health consciousness toward the high levels of health consciousness. This could be done by educating society on nutritional daily intake.

Moreover, marketers should be aware of the target groups they are focusing on and the product they are marketing, especially since high health conscious consumers are knowledgeable on nutrition information and make healthier choices.

Contradicting to previous research and against our expectations, the results in this study

revealed no significant impact of implementing the Nutri-Score on the healthiness of food choices

(hypothesis 1). Thus, participants did not significantly make healthier choices when the label was

displayed. As research on the Nutri-Score is still limited this research is the first in finding an

insignificant impact of the Nutri-Score on healthiness of food choices. This result needs to be

further verified by future research, as consistent results on this topic is important, due to increasing

interest from countries that already implemented the Nutri-Score label (NOS, 2019a). Specifically,

announced by state secretary Blokhuis (28

of November 2019) the Nutri-Score label will be

implemented within the Dutch market in 2021. Since data of this study was collected before the

announcement it did not affect the results of this study. Results of this study revealed that a high

percentage, namely 56,4% of the sample of Dutch participants (fig. 13) were not familiar at all

with the Nutri-Score label. Public policy makers within other countries that already implemented