Brands and Their Influence On Our Food Choices

Brands and Their Influence On Our Food Choices

How Brands and Health Claims Can Guide Us Towards Making Healthier Food

Choices

by

CHARLOTTE VAN VUGHT

Brands and Their Influence On Our Food Choices

How Brands and Health Claims Can Guide Us Towards Making Healthier Food

Choices

by

CHARLOTTE VAN VUGHT

University of Groningen

Faculty of Economics and Business

Master Thesis

June 26th 2017

Nieuwe Ebbingestraat 31a

9712 ND Groningen

(06) 57379575

c.m.m.van.vught@student.rug.nl

S3031268

MANAGEMENT SUMMARY

The past few years a growing trend towards health has been clearly visible. Unfortunately there is also an increasing trend visible in the amount of people who have overweight. This research gives some insights into how marketing can contribute to making consumers choose more healthy products when shopping for groceries. More specifically, it gives insights into the effect health labels and brand status can have on the decision making process with regard to the healthiness of our product choices. For this research the following research question has been stated.

What is the influence of front of packaging health claims and brand status on the healthiness of consumer in-store food purchase intentions?

To be able to answer this research question several hypotheses have been made. These hypotheses are based on previous findings in literature. There are two independent variables which are examined, and two moderation effects. The independent variables are the health label presence and the general health interest. The effect of these variables on the purchase intention of healthier food products will be tested. Whereas the moderation effects consists of the effect of brands status, and the effect of general health interest on the relationship of the health label variable on the purchase intention.

The first hypothesis states that presence of a front of packaging health claim has a positive effect on the purchase intention of healthy products. Research showed that health claims result in higher purchase intentions (Levin & Levin, 2010). In this research however, no significant direct effect has been found for the presence of a health claim. The second hypothesis, which states that national brands in comparison with store brands will strengthen the relationship between the presence of a health claim and the purchase intention of healthy products, is rejected. This is because no significant moderation effects were found for national brands and store brands. Even though Erdem et al. (2004), stated that national brands are seen as brands with higher status and therefore more likely to influence the decision making process.

intention of healthy products. Even though this is not in line with previous research (Zandstra et al., 2001; Talvia et al., 2011), the hypothesis has been rejected. The fourth hypothesis stated that the general health interest will have a positive enlarging effect on the relationship between the presence of a health claim and the purchase intention of healthy products. A significant effect for the national brand with a health label was found. Because this effect was found for general health interest * national brand with health label = 0, the parameters are interpreted for when the respondent does not satisfy both the high general health interest condition and the national brand with health label condition. This showed to have a significant negative effect for the chance of choosing the national brand, when these two conditions are not met by a respondent.

Furthermore some control variables were included to see whether the effects could be influenced by other variables. Only for the percentage of national brands a respondent usually buys a significant effect was found. This effect shows that a respondent has a higher probability of choosing the healthy brand when it is a national brand, if they have indicated that the percentage of national brand groceries is usually a larger percentage than the percentage for other kinds of brands.

Finally, there was no direct influence of front of packaging health claims found on the purchase intention of healthier products. For the brand status, also no significant effect was found. However when looking at the general health interest a significant moderation effect was found when the national brand has a health label. This effect was only found for the national brand, and not for the store brand. Thus brand status does have some effect on the healthiness of in-store food purchase intentions. This effect shows that when a national brand has a health label, and consumers have a high general health interest, the purchase intention of healthier products increases. When combining everything we can conclude that a person who generally has a high interest in health, indicated to more often purchase national brands in comparison with other brands, and sees a national brand with a health label will be most likely to buy the healthy national brand choice.

PREFACE

Everybody knows we cannot avoid making choices. As did I when it was time to decide what kind of thesis subject I was, and still am, interested in. I started looking at what I want to do after my master Marketing Management. The answer was very simple, I first want to see something of the world and travel. But I also know I want to contribute to society, I want to work and earn my place. Over the past few years I became very interested in the Fast Moving Consumer Goods industry. The first boundary of my thesis subject was thereby determined. For my premaster Marketing I wrote a research paper about materialism, this triggered my interest in consumer well being. In my thesis these two interests come together as food products are a part of the Fast Moving Consumer Goods industry and health is part of consumer well being. This choice of topic is one I am still very happy with. This preface gives me the opportunity to thank the people who were of great support during the past half year in which I wrote my thesis. First of all I would like to thank my family. Especially the past few weeks I needed some moral support and they were right there when I needed them. Furthermore I want to thank my close friends Alisa Roelandse and Femke Schijff and also my study friends, Anne Engelen, Miriam Jansen and Jetty Komrij, for listening to me, discussing certain issues, and keeping up the motivation. I also have had countless conversations with other people about my thesis. All of which either gave me a boost in motivation, or sometimes even gave me new insights into my research. I want to thank you all for taking the time to listen, converse and criticize, and for simply being interested. Finally I would like to give a special thanks to my supervisors Koert van Ittersum and Martine van der Heijden. During the whole process they guided me towards what I really wanted to research, and kept me on that path. Furthermore they provided me with useful feedback and a good atmosphere to work on my thesis. With the end of my thesis, my actual journey will almost begin. For that journey choices also had to made, and I am quite certain I will be as happy with my choice of destination as I am with the choices I made during the six amazing years I had while studying in Groningen.

TABLE OF CONTENTS

2.3 General Health Interest 16

5. CONCLUSION AND RECOMMENDATIONS 35

5.1 Summary & Conclusions 35

5.2 Managerial Implications & Recommendations 38

5.3 Limitations & Further Research 40

REFERENCES 43

1. INTRODUCTION

Nowadays the topic of healthy lifestyles can definitely be labelled as a “hot” topic. Healthy cookbooks, healthy products, and facebook pages or other programs to guide you towards a healthier life are everywhere around us. Worldwide younger consumers are found to be far more concerned about their food intake. Nielsen’s Global Health and Wellness Survey from 2015 found that Generation Z, consumers under the age of 20, are the most health-centric. With 41% saying they would be willing to pay a premium for healthier products. This compared to the percentages of the Millenials, ages 21 to 34, and Baby Boomers, age 50 to mid-60s, which are respectively 32% and 21% (Horovitz, 2015). Thus an increase of interest in health among younger people is clearly visible.

However, even though there is no shortage of attention to health and the interest in it, numbers from the World Obesity Federation (World Obesity Day Press Release, 2015) suggest there is an ongoing increase in the percentage of people who suffer from being overweight. Whereas the percentage of people above an indicated healthy weight in 2010 was 36 percent, this increased towards 39 percent in 2014. Expectations are that this increase will continue to a percentage of 46 in 2025. This is a serious problem because overweight is associated with diabetes and coronary heart disease among others. Research showed that the current overweight will likely result into large economic and health burdens (Lightwood et al., 2009). Current medical treatments will not diminish, or greatly reduce these extra costs caused by overweight.

a way to communicate with the consumer. Research showed that consumers use packaging as an extrinsic cue to infer intrinsic product attributes like the quality of the product (Underwood & Klein, 2002). Hence, the packaging helps consumers to form an image of the content. Furthermore it also helps to form perceptions about the healthiness of a food product.

In this research the focus will lay on the influence that on package health claims and brand status can have on the purchase intention of healthy food products during shopping trips. Research by Levin & Levin (2010) gives some insights into how consumers rate products. They examined, among other things, the effect of familiar brand names in comparison with unfamiliar brand names. The results of this research show that both children and parents rate products with a familiar brand name higher in terms of quality. Next to this outcome they also found that the effect of brand name for healthy products is stronger in comparison with unhealthy products. They concluded that the promotion of healthy eating habits can be best accomplished with a familiar brand. This brings us to the problem statement which will be researched in this paper.

What is the influence of front of packaging health claims and brand status on the

healthiness of consumer in-store food purchase intentions?

This problem statement results in several questions which are relevant for this research. The first marketing activity included in this research, is the health claim which packages may contain. More often packages show lines such as “high in fibre”, “low in fat”, or “no added sugars”. These promotion messages incline that these products are a healthier choice for the consumer. This raises the question what influence such an included message on the product has on the healthiness of the eventually chosen product. These health claims can be found on different kinds of brands. For this research we make a distinction between national brands and store brands. These two categories have been chosen because they include most of the brands which can be found in a store. Implemented into this research the question arises what the difference in influence of national brands versus store brands is on the effect of health labels on the healthiness of product choice.

Tuorilla, 1999). It is known that norms and values play part in the forming of our interests in life. People can have different interests in life, and thus also in how interested they are in health. A positive correlation has been found between this general health interest and the healthiness of the shopping basket (Zandstra et al., 2001; Talvia et al., 2011). Meaning that people who are generally more interested in health tend to have healthier shopping baskets in comparison with people who are less interested in health. However, in this research we are not only interested in this effect, but also in the effect of this general health interest on the influence of health claims.

Put into perspective of the increasing overweight rates, the outcomes of health claim and brand status research can provide guidance to manufacturers and marketers to help consumers to make more healthy product choices when shopping for groceries. This paper builds on several previous findings about the impact of brands and health claims on the purchase intention. Following the suggestion of the paper of Yan, Hyllegard and Blaesi (2012), who researched the influence of brand name and message explicitness on the eventual consumer purchase intention, this paper focuses on the influence of existing brands. Yan et al. (2012) proposed further research into the domain of real brand names to be able to generalize the results into real life situations. In their research they found that testing the impact of brand status could not be generalized when non existing brands were used. Furthermore this paper builds on research of Levin and Levin (2010), looking into the influence of health claims on packages of different kinds of brands and also taking into account the interest in health of consumers. This paper will build on their findings, but also look at the general health interest of consumers. Therefore it gives more insights into how brands can use health interest in their marketing.

towards this kind of products may help consumers to choose the more healthy products. Furthermore it is increasingly believed that healthy food intake is better than using medicine to stay healthy. This also has increased the marketing and development of healthy packaged food products (Kashif & Rashidi, 2013). Thus research into the domain of healthy packaged food products will become more and more relevant for marketing as consumer interest in this type of products is increasing. Next to this, the need for healthier products is also increasing.

Brands can help consumers make healthier choices and thereby guide them towards a healthier life and ultimately higher consumer well-being. This research contributes to helping brands make decisions regarding health claims in line with the health trend which is currently going on. As health and wellness is has become a top priority for consumers around the world. According to research by Nielsen (2016), 67% of global respondents have indicated to actively seek products with healthful ingredients. As demand rises, the assortment with healthy products keeps increasing, which will confront consumers more often with better choices for their health. In conclusion, this research contributes to the very current topic of health, and specifically food consumption, because consumption results from food purchases. Because it builds upon previous research it gives more insights into on the topic and provides insights which can be used in marketing activities for food and beverage products.

After this introduction previous research concerning the topics of healthier consumer choices, brand status and health claim influence, and the impact of the general health interest of consumers will be discussed to set a theoretical framework. Next the research design and used statistical methods will be explained. This will be followed by the results of the research. Finally the conclusions and recommendations will be discussed followed by the references and appendices.

2. THEORETICAL FRAMEWORK

2.1 Health Claims

forms of marketing communication. In case of health information this can be in the form of health claims, pictures, and a short story on the packaging. Because there is an increase in the amount of products which bring particular health benefits to consumers, there also is an increase of products which display claims to promote these benefits (Van Trijp & Van der Lans, 2007).

Ares, Giménez & Gámbara (2009) furthermore found that the incorporation of a health claim also achieved a positive association between the product and the health effect in the mind of consumers. Other research also shows the favourable effect of health claims on products, like the research of Tuorila & Cardello (2002) which also found that health claims increase purchase intention. Furthermore the research of Krystallis and Chrysochou (2011) and Aschemann-Witzel and Hamm (2010), states that products with health claims perform better on the market.

In general consumers view a food as healthier when there is a health claim present following the research of Williams (2005). This is in line with the research of Underwood & Klein, (2002) in which they found that the packaging helps consumers form a perception of the intrinsic product attributes. Among others they found product quality to be determined this way. Research of Lähteenmäki (2013), also shows that health claims carry the message of increased product healthiness. However, even though consumers see the product as more healthy, this research did not find that health claims necessarily make the product more appealing. Furthermore, as stated in the research of Horovitz (2015), younger people are most willingly to pay a premium price for healthier products. Indicating that if a product is healthier, the price will become less important as the younger people now, will become more important for companies when they get older.

William’s (2005) research also showed that short health claims with concise wording are perceived as useful by consumers. Consumers do not like large lines of text explaining the benefits. In this research, as also found by Roe, Levy & Derby (1999), there was no distinct difference found between nutrient content and health claims. The preference for short lines follows the system 1 thinking (Kahneman, 2011) where people prefer simple things as they take less effort.

information are available. This means that both these types of on packaging information can be interpreted separately. Another finding by Ford et al. (1996) for health claims is that they create expectations regarding the health condition mentioned.

When looking at the different types of information that can be given to consumers we can make a distinction between objective nutrition cues and evaluative nutrition cues. These objective nutrition cues offer specific quantitative information about a benefit of the product. Where evaluative nutrition cues provide information about the overall healthfulness and/or nutrients of a product (Newman, Howlett & Burton, 2016). In their research Newman et al. (2016) showed that when consumers can evaluate multiple food items simultaneously, as they also have the option to do so in supermarkets, evaluative cues on products have a more positive impact. Consumers then evaluate the products more positive and a higher purchase intention of healthier products was found. For this research it is important to only use one type of information to not bias the results. Huang & Lu (2016) found in their research an influence of the perception of food healthiness on purchase intention.

Health claims showed to have a positive influence on the purchase intention because of the increase in trust in the product and a higher perceived product quality. Therefore it is expected that the presence of a health claim will increase the purchase intention for healthy products. This gives the following hypothesis for health claims.

H1: Presence of a health claim on the front of packaging has a positive effect on the purchase intention of healthy products in comparison with no health claim presence on the front of packaging.

2.2 Brand Status

distributed by the supermarket under its own brand name. The distributers brand example is also known as a store brand, or private label. In this paper we will further refer to this kind of brand as a store brand. A specific definition for this type of brand is also given by Schutte (1969), store brands are the products owned and branded by the organizations whose primary objective is distribution rather than production.

On the other hand there are the manufacturer brands. These are defined as brands which are controlled and owned by an organization whose primary commitment is production (Schutte, 1969). For example, Heinz tomato ketchup is sold in supermarkets which do not run under the Heinz brand name. Thus Heinz is the manufacturer of the brand, but not the end distributer to the consumer. The manufacturers brands are also known as national brands. The definition by Schutte (1969) states that a national brand is owned by a manufacturer and marketed to retailers and wholesalers within the channel of distribution. Because national brands and store brands are the most common types, this research will focus on these two types of brands only.

Consumers perceive national brands very differently from store brands. Following the research of Erdem et al. (2004), store brands are perceived inferior to national brands. This inferiority comes forward from the uncertainty on how satisfied a consumer will be with the purchase from the store brand. When consumers have this uncertainty on the level of satisfaction they automatically perceive national brands, which are well-known and have used a consistent communication strategy, as quality products and thus a better product (Erdem et al., 2004). Uncertainty on the quality of products is considerably more found about store brands than there is about national brands (Steenkamp, van Heerde & Geyskens, 2010). Thus when price differences are negligible consumers will prefer national brands over store brands because of certainty of quality. This preference for national brands over private brands was also found in a study about consumer preferences for national versus private brands on food products (Wyma et al., 2012).

brands may be perceived as more preferred by other, which addresses the principle of social proof (Cialdini, 2009; Lun et al., 2007). This can also lead to a preference for the national brand over the store brand, as this brand is perceived as better.

The research of Laforet (2011) about food products also showed that to influence consumer choice a brand must be seen as the leader in a particular market segment or in its product market. Following the outcomes of previous studies, it is consistently shown that national brands have a higher or better status or are perceived as having this better status. We thus consider a national brand as having a high brand status, and a store brand as having a lower brand status. As previously mentioned Levin & Levin (2010) focused on both the familiarity of brands and the difference in effect on healthy versus unhealthy products. Because the presence of a health claim gives people the perception that the product is healthier, we can build on the research by Levin & Levin (2010). They found the strongest effect in their research was for familiar brands, which can be compared to national brands, in combination with healthy products. Thus when combining the findings it appears that national brands will strengthen the effect of health labels on purchase intention more in comparison with store brands, because the status of the national brands is being linked to a higher quality of the product. This resulted into the following hypothesis with regard to brand status.

H2: National brands in comparison with store brands will strengthen the relationship between the presence of a health claim and the purchase intention of healthy products.

2.3 General Health Interest

Lifestyle is a variable which will not change rapidly, and is more than just the interest in healthy products. It has to do with the values an individual holds close. Therefore the choice to research the impact of the general health interest has been made. This makes the lifestyle of individuals measurable on their vision towards health. Following Finch (2005) we find that ideological social values together with functional economic values play important roles in shaping individuals’ retail food buying behaviour. Meaning that what we believe in also shapes our buying behaviour.

concludes that values “exist in an interconnected, hierarchical structure in which global values are related and connected to generalized consumption-related values which are, in turn, similarly associated with product attributes.” These product attributes are the basic elements which order consumers’ attitude towards a product. Thus meaning that these attitudes ultimately influence consumer behaviour. Following this we can state that people who are generally more interested in health will pay more attention to this aspect when shopping for groceries. This is confirmed in research using the General Health Interest (Zandstra, de Graaf & van Staveren, 2001; Talvia et al., 2011).

When values change, this could also result in a change in for example purchase behaviour. Consumers also hold values towards the products they buy. When looking at the purchase intention for healthier food products, the value a consumer attaches to health is of importance. If a consumer is concerned with health as an outcome, they will engage in protective health behaviours (Hayes & Ross, 1987). Thereby Steptoe et al. (1995) and Wardle (1993), also find health beliefs to be an influencing factor on behaviour. The general health interest of a consumer gives insights into these health believes.

There are also several factors which influence the intention to eat healthy (Sapp & Jensen, 1998; Deshpande, Basil & Basil, 2009). For this research the importance of a healthy diet is the most important influencer, but both studies also reported the influence of gender, food features, dietary status and cues to action. These factors will also influence the general health interest of a consumer. In the research of Cook (2016) the motivation of consumers, social support and food features all have significant effects on the healthful consumption likelihood. These factors are all of importance for this research, because for the consumption of food, it first has to be purchased and thus a purchase intention can be found.

information. The middle segment, predisposed people, consists of consumers who only want to change towards healthier behaviour if it is easy and does not involve much sacrifice. The bottom segment, disinterested people, have no interest in changing their behaviour towards more healthy. These consumers do not believe the benefits outweigh the effort or the sacrifice or do not think it is important. These findings are in line with other research. Two separate studies both found more favourable attitudes and greater behavioural intentions for consumers with high product involvement in comparison with consumers with low product involvement (Kim, Haley, & Lee 2008; Limbu & Torres 2009). High interest in a product is known to lead to high product involvement. Thus when put into perspective of the general health interest, a higher interest, will lead to higher product involvement.

In previous research the general health interest was used as an independent variable to research its influence on product choice. When following previous research which used the general health interest we find several indications that a high score on this scale is associated with the purchasing of healthier products. In comparison this means that a lower score on this general health interest is associated with fewer purchases of healthier products. An association with a lower fat intake, lower high-fat oils and high-fat savoury snack consumption, and a consumption increase of vegetables and fruit with the general health interest were found (Zandstra, de Graaf & van Staveren, 2001). This finding is supported by research of Talvia et al. (2011). As an addition they also highlight in their research the association with higher fibre intake. When taking into account the influence of health claims, previous research showed that these health claims mainly have relevance for consumers who are already interest in health. These consumers are generally looking for more information when shopping (Garnett et al., 2015; Grunert & Wills, 2007; Hawkes et al., 2015). To my knowledge no previous research has been conducted about the explicit expectations of general health interest on a health claim impact. When combining these findings with the positive influence a health claim has on the purchase intention of a product, several expectations can be formed.

related product communication stronger in comparison with consumers who are not interested in health. This would result in a stronger impact of health related marketing communication. Because the general health interest is expected to have both a direct effect on the purchase intention of healthier products and to make the relationship between health claims and purchase intention of healthy products stronger, two separate hypotheses have been formed. First it is hypothesised that higher general health interest in comparison with a lower interest will have a positive direct effect on purchase intention of healthy products, as can be found in H3. This hypothesis takes into account that the expected moderation effect of health labels on purchase intention of healthy products will become stronger when a consumer scores higher on the general health scale. Taking this into account H4 has been formulated.

H3: The general health interest will have a direct positive effect on the purchase intention of healthy products.

H4: The general health interest will have a positive enlarging effect on the relationship between the presence of a health claim and the purchase intention of healthy products.

2.4 Conceptual Framework

FIGURE 1

The effect of health claims on healthy food purchase intentions

3 RESEARCH DESIGN

3.1 Data Collection

For this research adults from the Dutch market were questioned. The Dutch market is a developed market in which there is plenty availability of products which include health claims. Respondents are likely to have been confronted in real life to choose between a product with a health claim or without a health claim. Thus they can relate to the decision making process. The products used in the questionnaire consist of national brands and store brands known by the Dutch population, as they are also sold at a large Dutch retailer. Adults were selected as respondents because the questionnaire involves decisions considering health related claims. This is something children are not as involved in as adults, as for most of them their parents will do the weekly groceries. The questionnaire has been distributed using the social media platform Facebook. In order to get a representative sample size for each of the conditions, at least 30 respondents are required per condition.

Health Claim: - Present - Not Present Brand Status: - National brand - Store brand Purchase Intention Healthy Products

General Health Interest

H4:  H1:  H2: +

3.2 Experimental Design

For this research a between subjects factorial design will be used for the condition question. The rest of the questions are the same for everyone. Each respondent will be randomly assigned to one of the four conditions of the research. For each condition the health claim will either be not present, or present but for products with different brand status. These conditions are specified in table 1. The research design will be a 2 X 2 design, with general health interest as measured variable.

TABLE 1 Research conditions

Health Claim Brand Status General Health Interest Condition

Not Present High & Low Scale Condition 1

Present High Scale Condition 2

Present Low Scale Condition 3

Present High & Low Scale Condition 4 3.3 Procedure

has a health label. In condition 4 both the national brand and the store brand have a health label. During the questionnaire each respondent is confronted only once with their assigned condition. The rest of the purchase decisions are there to create a shopping trip experience. Figures 2 and 3 show two of the conditions. Figure 2 shows the image respondents of condition 1 are confronted with, where no health label is present. Figure 3 shows this for condition 2, where there is a health label present for the national brand.

FIGURE 2 FIGURE 3

Condition 1 question example Condition 2 question example

Research of Romero & Biswas (2016) showed that to be congruent with consumers’ mental organization of food items healthy products should be positioned on the left and unhealthy products on the right. This enhances ease of processing, which in turn enhances self-control. When there is higher self control, the likelihood of choosing healthy options is higher. To not let this interfere with the variables in this research, the products within a purchase decision should always be randomized. The overall purchase decision was also randomized to exclude other effects. Except for the purchase decision in which the condition is examined. The position in the survey of this purchase decision is set to the fifth decision, to make sure that the point in shopping trip for this decision has no influence.

3.4 Operationalization of Variables

Here all variables examined in this research will be described as in how they will be tested and which scales will be used for this testing. First the product category which will be used for the health claim condition in the questionnaire is discussed. Second the measurement of the general health interest will be discussed. Next, a distinction between brand status of products will be made. Finally we assess how the product choice for healthy products will be examined.

Product Category and Health Claim Selection

This research began with stating that overweight is an increasing problem globally. More healthy consumption choices are a good start of fighting this problem. To get to healthier consumption, people first have to buy the right products. These switches between food choices are mainly made between alternative products from the same food category (Ross & Murphy, 1999).

For the questionnaire nine food categories with each three alternative products were selected. These are the following categories: low-fat quark, breakfast cereals, soup, breadsticks, butter, fish fingers, pasta sauce, apple sauce and biscuits. Where the butter category has four different conditions, which all consist of the same three products. But each of these conditions has a different health label condition as specified earlier on in table 1. Respondents will be confronted with only one of these four conditions.

The health claim has to relate to the product category butter. As butter consists of fat, the health label attached in conditions 2, 3 and 4, will have the health claim: “less fat”. The health claim presence will be treated as the independent variable for this study.

Brand Status

and health label will be simultaneously shown. Brand status will be displayed in the form of a product picture and some information about the product. It is clearly visible if the product is a national brand or a store brand. The third option is also a store brand, but with a much lower status. This product is not an equivalent for the national brand. It is merely there to provide respondents with a third choice. The control condition, condition 1, will indicate what the normal choice distribution is without the health label. Thus in this condition respondents have to choose between the different brands without being confronted with a health claim, while in condition 2, 3 and 4, the respondents are confronted with the health claim.

General Health Interest

To measure the general health interest, this research will use the General Health Interest Scale as introduced by Roininen & Tuorilla (1999). This scale is part of the Health and Taste Attitude Scales (HTAS) Questionnaire and consists of eight questions. The respondents are requested to answer these questions on a seven point likert scale, where 1 is “disagree strongly” and 7 is “agree strongly”. All respondents are asked to fill in the General Health Interest Scale to assess their general interest in health. The complete scale can be found in appendix II. An example question of this scale is: “I am very particular about the healthiness of foods.”

The general health interest is both a moderator and expected to have a direct effect on purchase intention, which also makes it an independent variable. Because the general health interest will be measured with a separate scale, both effects can be measured with the help of this scale.

Purchase Intention Healthy Products

should be visible because condition 1 has no health label. Meaning that no healthier choice can be made by a respondent in condition 1.

3.5 Other Variables

Next to the variables which are to be tested for the hypotheses, also some control variables are included in the questionnaire. There are four control questions and four demographic questions. The control questions are about the average percentage of national brands and store brands respondents usually buy, how hungry respondents are at the moment of participation, and two questions with which their BMI can be calculated. The demographic questions include gender, age, and country of residence now, and country of residence in the past. The questions about country of residence are there to control for product familiarity. Because products from the Dutch market are used in the questionnaire, these questions are a check to see if people know the products. They can be familiar with the products because they either live, or have lived in the Netherlands.

3.6 Research Methods

The results of the questionnaires will be analyzed with the help of the statistical program SPSS. First, some descriptive were made and to check the reliability of the variables a Cronbach’s alpha test was done. As stated by Bagozzi (1988) the Cronbach’s alpha should have a score of .60 to be reliable. In order to check whether the hypotheses hold, a multinomial logistic regression analysis has been conducted.

4. RESULTS

4.1 Sample Characteristics

FIGURE 4

Age distribution respondents

Furthermore 94.3% of the sample has a Dutch nationality and 97.9% either currently lives in the Netherlands or has lived there in the past. As there is a small percentage which have never lived in the Netherlands, there is a possibility that they are not familiar with the products. There will be checked if this group has any effects on the results in further analyses.

4.2 Reliability Tests

Next the dataset was checked for several factors which could influence the statistical tests. The general health interest consists of eight questions which were transformed into a total score on the general health interest scale. Before calculating the total score, first four questions are reversed. These reversed questions replace the original questions for the scores, and therefore also for the Cronbach’s Alpha test. For this scale a Cronbach’s Alpha of .849 was found, and there was no indication that the scale becomes better when items are removed. Indicating that the general health interest total score can be used for further analyses.

the use of two slightly different models. For this dummy variable the median (4.50) was taken as the cut off point, dividing the sample into two groups. The group with higher scores than the cut off point can be seen as respondents with a high general health interest, where as respondents with a score of 4.50 or lower in comparison with higher, have a low general health interest. Because both the total scores scale of the general health interest and the dummy scores are used, the multicollinearity test was conducted twice. Once with each of the two general health interest variables. Table 2 shows the VIF values for both models where column one shows the scores for the total general health interest scores, and column two shows the scores for the dummy variable.

TABLE 2

Multicollinearity check

VIF Scores with Total Scale VIF Scores with Dummy Scale

National Brand Health Label 2.076 3.321

Store Brand Health Label 2.057 3.175

General Health Interest 2.805 3.075

National Brand * Store Brand 1.843 2.812

General Health Interest * National Brand

1.883 2.717

General Health Interest * Store Brand

3.128 3.168

Percentage National Brand 2.148 2.172

Percentage Store Brand 2.174 2.196

The scores for both of the models indicate no multicollinearity between the variables, because all values are between one and ten.

4.3 Hypotheses Tests

In each condition participants had to indicate their preferred choice for butter. The conditions each have approximately the same amount of respondents. Namely, condition 1 (71), condition 2 (68), condition 3 (71) and condition 4 (73). To get an indication of the preferences of the respondents, the percentages for each brand option can be found in figure 5.

FIGURE 5

Product choice for butter per condition

The data was also checked for significant differences for the control question, percentage of national, store, or other brands a respondent normally puts into their shopping basket. The outcome for the ANOVA test was significant for both national brands and store brands (p = .00 for both), but not for other brands (p = .188). In further analyses therefore the percentage of national brands and store brands respondents usually buy will be taken into account.

Next a multinomial logistic regression was done. This test shows whether the two models and variables are significant, and what the parameters of the individual variables are. First the tests were done with the complete sample, and to see if there are significant differences also with a subsample excluding the respondents who have never lived in the Netherlands. This was the case for 16 respondents, leaving the subsample on 267 respondents. This subsample showed to have more significant results, therefore the subsample will be used for further analyses.

The first analysis was done for the model with the total general health interest score. The complete output of this model can be found in appendix III. The Pearson Goodness of Fit test is insignificant (p = .378). This means that the model fits the data well. Another indication of a good fit to the data is shown in the model fitting information for the total model. This indicates if the model with variables is a better predictor than the intercept alone. The model is highly significant (p = .00), thus meaning that the full model significantly predicts the dependent variable better than the intercept alone. The parameters in this model indicate what would happen with the chance of choosing the national brand.

Furthermore the percentage national brands shows, when people have indicated to generally buy a higher percentage of national brands, they are more likely to choose the national brand. The full model parameter estimates can be found in table 3.

TABLE 3

Parameter estimates for model with total general health interest score

B Exp(B) P-value

Intercept _{-.607 .655}

National Brand Health Label = 0 .027 1.028 .947

Store Brand Health Label = 0 .193 1.213 .653

General Health Interest _{-.136 .873 .563}

National Brand * Store Brand =0 _-.131 .877 .826

General Health Interest * National Brand Health

Label .596 1.815 .036

General Health Interest * Store Brand Health Label .022 1.022 .939

Percentage National Brand _{.031 1.031 .002}

Percentage Store Brand _{-.012 .988 .286}

is zero. Thus indicating that either the general health interest is low, or no health label is present for the national brand, or a combination of the both. This estimate indicates a negative effect for choosing the national brand when one of the previously mentioned things is at order. Thus when there is no health label present for the national brand, or when there is a low general health interest. Furthermore the percentage national brands shows the same outcome for both models. When people have indicated to generally buy a higher percentage of national brands, they are more likely to choose the national brand. The parameters for all variables of the model with the dummy variable of general health interest can be found in table 4.

TABLE 4

Parameter estimates for model with general health interest dummy variable

B Exp(B) P-value

Intercept -.633 .494

National Brand Health Label = 0 .659 1.933 .207

Store Brand Health Label = 0 _{.260 1.297 .615}

General Health Interest =0 _.589 1.802 .242

National Brand * Store Brand =0 -.161 .852 .787

General Health Interest * National Brand Health

Label = 0 -1.359 .257 .022

General Health Interest * Store Brand Health

Label = 0 -.146 .864 .807

Percentage National Brand _{.030 1.031 .002}

Percentage Store Brand _{-.012 .988 .290}

TABLE 5 Hypotheses overview

Hypothesis Confirmed

H1: Presence of a health claim on the front of packaging has a positive effect on the purchase intention of healthy products in comparison with no health claim presence on the front of packaging.



relationship between the presence of a health claim and the purchase intention of healthy products.



purchase intention of healthy products.



relationship between the presence of a health claim and the purchase intention of healthy products



4.3.1 Hypothesis 1

Hypothesis 1 states that presence of a health claim on product packaging will positively influence the purchase effect of healthy products. When looking at the multinomial logistic regression no main effect of health labels is found. Both for national brand and store brand with health label no significant effect is found (p = .207 and p = .615). These variables indicate whether the addition of a health label have an effect on the purchase intention of healthier products. Where the healthy product is seen as the product with the label. Because no significant effects have been found in this research, it has to be concluded that the health labels did not have a significant influence on the product choice. Therefore hypothesis 1 has to be rejected.

4.3.2 Hypothesis 2

(p = .207, p = .615). The model does indicate that there is a significant relationship found for national brands. However, this is in combination with the moderation effect of general health interest (p = .022). Therefore it is not an effect of the national brand itself. Because no direct moderation effect was found for either national brands or store brands it has to be stated that hypothesis 2 is rejected.

An effect which has been found for national brands is the significant effect on choosing national brands when people have indicated to normally buy more national brands (p = .002). This shows that when people normally buy more national brands, the probability is higher that they will choose for the national brand here as well (B = .030, Exp(B) = 1.031). This is not a surprising result, but nevertheless worth mentioning because it does show that a good national brand selection was made.

4.3.3 Hypothesis 3

A direct effect of the general health interest on purchase intention of healthier products is stated as hypothesis 3. The regression analysis did not indicate a significant effect for this direct effect for the national brand (p = .242). Therefore the parameters cannot be interpreted. Thus hypothesis 3 has to be rejected. However, when looking at the parameters for the other brand choice an interesting almost significant effect was found (p = .075). This effect will be further discussed in 4.4, the additional results section.

4.3.4 Hypothesis 4

FIGURE 6 FIGURE 7

Effect for national brand condition Effect for store brand condition

Figure 6 shows the moderation effect the general health interest has on the product choice when the national brand has a health label. Figure 7 shows this effect for the condition in which the store brand has a health label. It is clearly visible that the respondents with a high general health interest in the condition where the national brand has a health label had the highest percentage of choosing for the national brand. In figure 6 a clear difference is visible with the group who has a low general health interest. When looking at figure 7 it is clearly visible for the group who have high general health interest that they choose more for the product with a health label in this condition, namely the store brand. Summarizing, the figures give us a visualization of the effect found in the results. There is a negative effect found when either there is no health label for the national brand, or when there is a low general health interest (B = -.1359, Exp(B) = .257). Thus more people choose for the national brand when it has a health label attached and they have a high general health interest. This indicates that hypothesis 4 can be accepted.

4.4 Additional Results

scale of general health interest. However, because this effect was not found for the other choice options hypothesis 3 cannot be accepted.

The other control variables were also tested, but they did not result into significant effects. Therefore they are left out of the original model. This goes for the variables Gender (p = .526), Hungry (p = .487), BMI (p = .687) and Age (p = .364) The output of the multinomial logistic regressions for the model with these variables can be found in appendix IV.

5. CONCLUSION AND RECOMMENDATIONS

This chapter will answer the research question for which the hypotheses are stated. The conclusions are based on previous research and the results of this research. Furthermore managerial implications and recommendations will be formulated. Finally the limitations of this research and further research suggestions will be discussed.

5.1 Summary & Conclusions

A growing trend towards health has been clearly visible over the past years. Unfortunately there is also a global increase in the amount of people who have overweight. This research gives some insights in to how marketing can contribute to making consumers choose more healthy products when shopping for groceries. The focus is on how health labels and brand status can influence consumers to choose the healthier product options.

A survey was distributed in which four conditions regarding health labels and brand status were examined. With the results of this survey four hypotheses have been researched and possible relationships between the variables are analyzed.

leads to the expected perception in the mind of the consumer. This would be a healthier and qualitative better perception of the product with the health label.

The second hypothesis, which states that national brands in comparison with store brands will strengthen the relationship between the presence of a health claim and the purchase intention of healthy products, is rejected because no significant moderation effects were found for national brands and store brands. However there is a significant effect for national brands when combined with the general health interest, but this is a moderation effect of the general health interest variable. Following the research of Laforet (2011), it makes sense that this effect was only found for the national brand. His research shows that a brands must be seen as the leader in its market to influence consumer choice. Where previous studies showed that national brands, which have a familiar brand name are considered to have a higher or better quality (Levin & Levin, 2010).

For the third hypothesis no significant relationship has been found. This hypothesis states that the general health interest will have a direct positive effect on the purchase intention of healthy products. This is not in line with previous research. Following this research a higher interest in health, which is measured with the general health interest, should lead to a higher purchase intention of healthy products (Zandstra, de Graaf & van Staveren, 2001; Talvia et al., 2011). An explanation for this different outcome could be the manipulation of the health claim condition. As stated before the respondents may not have interpreted the products with a health claim as a healthier option than the other product options. However, for the other brand option an almost significant effect was found. Showing that respondents are more likely to choose the store brand over the other brand, when they have a higher general health interest.

condition and is placed in the national brand with health label condition. This showed a significant negative effect for the chance of choosing the national brand, when these two conditions are not met by a respondent.

Furthermore some control variables were included to see whether the effects could be influenced by other variables. Only for the percentage of national brands a respondent usually buys a significant effect was found. This effect shows that a respondent has a higher probability of choosing the healthy brand when it is a national brand if they have indicated that the percentage of national brand groceries is usually a larger percentage than for other kinds of brands. The other control variables, which are hunger, gender, age, and a combination of two questions to calculate BMI, did not have a significant effect on the purchase intention of healthier products. Hunger might not show a significant effect because the choice respondents had to make was between products of the same product category. Furthermore, the category in which the different conditions were implied was butter. With logical reasoning it can be stated that butter is normally not a product people crave for when they are hungry. Gender and age might not have shown significant effects because of the sample distribution. The largest part of the respondents were younger females. An insufficient amount of older respondents or, in case of gender, males might have caused the insignificant findings for these variables.

For the BMI also no significant effect was found to have an influence on purchase intention of healthier products. As butter is not used in great amounts the effect of the ‘less fat’ label might not have been big enough to be seen as a healthier choice. Also, no correlation has been found between the general health interest, or any of the separate questions of this scale, with the BMI. The general health interest scale also includes a question regarding fat, ‘it is important for me that my diet is low in fat’. This question could indicate an extra interest in the health claim, but no significant effect was found so no inferences can be made about this variable.

Lastly the research question will be answered. The research question is as follows:

With the help of the hypotheses the research question can now be answered. Firstly there is no direct influence of front of packaging health claims found on the purchase intention of healthier products. However, as stated before, no pre-tests were done to examine whether the health claim attached to the product was really perceived as a healthier choice. Thus even though this research states that there is no direct influence of health claims on the healthiness of consumer in-store food purchase intentions, this finding might have to do with the manipulation.

When looking at the brand status, is a product a national brand, or a store brand, no significant effect was found. However when taking the general health interest into account, a significant moderation effect was found on the national brand with a health label. This effect was only found for the national brand, and not for the store brand. Thus brand status does have an effect on the healthiness of in-store food purchase intentions. This effect shows that when a national brand has a health label, and consumers have a high general health interest, the purchase intention of healthier products increases. The single effect of national brands with a health label was however not significant. Thus only a combined effect of brand status and general health interest showed to have influence on the purchase intention. In conclusion it can be said that a brand alone has no influence on the purchase intention of healthy products. This also goes for a health label alone and general health interest alone. However, the moderation effect of general health interest on a national brand with a health label does have a significant effect. Thus only combined significant effects are found. In real life you also have these factors combined. When combining everything we can conclude that a person who generally has a high interest in health, more often purchases national brands in comparison with other brands and sees a national brand with a health label will be most likely to buy the healthy national brand choice.

5.2 Managerial Implications & Recommendations

were to be repeated, the method of manipulation should be tested before conducting the questionnaire.

Also no significant effect of brand status was found, meaning that there is no significant difference between the influence of solely a national brand or store brand on the relationship between health labels and purchase intention of healthier products. However, a significant effect has been found on the moderation effect of a higher general health interest on a national brand with a health label on purchase intention of healthier products. This effect was not significant for store brands. For firms this means they should take into account if they have a national brand, or store brand when thinking about implying a health claim. This research shows that, only for national brands, the purchase intention of healthy products increases. Thus when firms have national brands, they could consider to use health labels as it leads to positive results. Provided that the products are allowed to have health labels of course. For firms which hold store brands, this effect was not found and as it is hard to comply to the norms set for health labels, it might not be worth the effort and costs to put a health label on a store brand. Because the found effect was for people with a higher general health interest, it is advised for national brands to target the people who are concerned with their health. Currently a health trend is clearly visible and there are numerous bloggers and vloggers online who show which products to use and which not. Collaborating with these people will draw attention to the product and increase brand awareness. This will in turn result into higher purchase intentions among consumers.

national brands and/or are known to buy healthy products can be send a personal offer for the national brand with health label to increase the awareness for this product.

Because this research is held in the Netherlands, and almost all respondents are Dutch (94.3%) or live or have lived in the Netherlands (97.9%), it is not generalizable to other countries. Furthermore, the sample consists mainly of young adults (18-29). It cannot be stated if the results would also account for older generations. More about generalizability can be found in the limitations and further research section of this chapter.

5.3 Limitations & Further Research

A first limitation of this study is that it is not generalizable to other countries. There are several things which may cause trouble for generalizability. First, the interpretation of the health claim may not be the same in each country. Research of Van Trijp & Van der Lans (2007) showed differences in how people interpret health claims between countries. This means that the results of this research could be different in other countries. Furthermore countries can have different rules when it comes to legislation for health labels on food packages (Leathwood et al., 2007). It is therefore recommended to conduct further research in other countries before generalizing these results.

Secondly, a limitation of this research is that some respondents might not shop at Albert Heijn. This would mean that they are not familiar with the products which are store brands of Albert Heijn and are more inclined to choose for the national brand because it is familiar to them. This research did not test for product familiarity and thus no conclusions can be made if unfamiliarity of the products played a part in the decision making. For further research it is therefore recommended that product familiarity is also tested to exclude the influence which unfamiliarity can have on the decision making process.

different store brand and if this has an influence on the decision making process when health labels are included.

The fourth limitation is the unevenly distributed age of the sample and the smaller part of male respondents. The largest part of the sample are young female adults. To give a clear image of the results for the Netherlands more respondents of higher age and more male respondents should also be reached. Previous research showed that there are differences of what product attributes individuals find important depending on gender, age, lifestyle, cultural background, race, socioeconomic characteristics and education level (Prescot et al., 2002). This research only asked about gender, age and partially about lifestyle with the general health interest. Further research could focus more on the other characteristics and aim for a sample with sufficient respondents to make inferences about the different groups. Only the product category butter was researched. This category was chosen because it is a very basic product category. However this also limits the generalizability towards other product categories in which the difference in healthiness of products can be much larger. Also differences in usage can be found between product categories. Butter for example is a product which can have a much higher usage than for example pasta sauce which is only used when making pasta. The usage rate can differ per consumer, but overall there are product categories which are used more often than other products. Because all product categories are different, following Ascheman-Witzel and Hamm (2010) also a different approach of communicating health claims is needed for each product category. This gives us another path for future research, namely to research the differences between different product categories.

Also a limitation of this research is the limited amount of products respondents had to choose of. Only three products were available for each product category, where in real life there are many more options available. When there are much more options available, the health label will not stand out as much as it did in this research. Research towards a more real life shopping experience can be done to find out if health labels have any influence when confronted with in a supermarket with all the other brands around it.

possible to check whether full attention was paid to the survey. To eliminate this factor research should be conducted in real life situations during a shopping experience. Another interesting research direct is the restaurant environment. This research can also serve as a framework which can be used for similar research in the environment of restaurants.

REFERENCES

Ares, G., Giménez, A., & Gámbaro, A. (2009). Consumer perceived healthiness and willingness to try functional milk desserts: Influence of ingredient, ingredient name and health claim. Food Quality and Preference, 20(1), 50-56.

Aschemann-Witzel, J., & Hamm, U. (2010). Do consumers prefer Foods with nutrition and health claims?: Results of a purchase simulation. Journal of Marketing Communications, 16, 47-48.

Cialdini, R. B. (2009). Influence: Science and practice (5th Ed.). Boston: Allyn and Bacon Coates, S. L., Butler, L. T., Berry, D. C. (2004). Implicit memory: a prime example for brand consideration and choice. Applied Cognitive Psychology, 18, 1195–211.

Cook, L. A. (2016). Health Belief Model and Healthy Consumption: Toward an Integrated Model. Journal of Food Products Marketing, 1-17.

Deshpande, S., Basil, M. D., & Basil, D. Z. (2009). Factors influencing healthy eating habits among college students: An application of the health belief model. Health marketing

quarterly, 26(2), 145-164.

Fajardo, T. M., & Townsend, C. (2016). Where you say it matters: Why packages are a more believable source of product claims than advertisements. Journal of Consumer

Psychology, 3(26), 426-434.

Finch, J. E. (2005). The Impact of Personal Consumption Values and Beliefs on Organic Food Purchase Behavior. Journal of Food Products Marketing, 11(4), 63-74.

Ford, G. T., Hastak, M., Mitra, A., & Ringold, D. J., (1996) Can Consumers Interpret Nutrition Information in the Presence of a Health Claim? A Laboratory Investigation. Journal of Public

Policy & Marketing Vol. 15 (1) 16-27.

Garnett, T., Mathewson, S., Angelides, P., & Borthwick, F. (2015). Policies and actions to shift

eating patterns: What works? A review of the evidence of the effectiveness of interventions aimed at shifting diets in more sustainable and healthy directions. London, UK: Chatham