Inside-Out:
How a product-level FOP bar nutrition label will facilitate
product evaluation and diminish the purchase frequency and
consumption volume of unhealthy food
By
CATERINA PORRO
Inside-Out:
How a product-level FOP bar nutrition label will facilitate
product evaluation and diminish the purchase frequency and
consumption volume of unhealthy food
Master Thesis
University of Groningen Faculty of Economics and Business
MSc Marketing Management Caterina Porro Korreweg 22B 9715AA, Groningen (NL) Email: Caterinaporro13@gmail.com Student Number: 3537161
Abstract
The still existing complexity in nutritional panels, together with the extensive marketing strategies developed to attract consumer’s attention on products in-store, are compromising consumer’s ability to accurately judge products for their content and not their appearance. This leads to severe misjudgements of serving sizes and healthiness of products and contributing further to increasing rates of health problems such as obesity or diabetes. Therefore, this study presents a novel nutrition label format, displaying the entire content of the product on the front of package, with the aim to increase transparency of product’s content and enable consumers an at-a-glance evaluation. An online survey with 120 participants was conducted to measure the effectiveness of this product-level bar nutrition label on lowering the purchase frequency and consumption volume of relatively unhealthy food, with respect to consumers’ level of nutrition literacy.
Contrary to the expectations, there is no significant difference found in purchase frequencies of products. However, there is not only a negative, marginally significant effect of the bar nutrition label on the consumption volume of products with higher content of sugars and fat, but it becomes more pronounced for consumers with a lower level of nutrition literacy. Lastly, the follow-up survey demonstrated an overall positive attitude towards the novel nutrition label format and its ability to meet consumers divergent interests and the focus on products attributes.
Preface
This thesis is my final project of the Marketing Management Master at the University of Groningen. The knowledge acquired in the past years is fully reflected in this paper. I was able to gain detailed insights into consumer psychology, retail and other marketing strategies, which have only fortified my interest in this field. The aim of this thesis was not only to increase the awareness on the problematic related to misleading nutrition labels, but also to find a solution to improve the current situation. At the end of the day I am glad, I took the risk of proposing and analysing a brand-new nutrition label format.
This valuable experience helped me to understand the mechanisms behind situations we encounter on the daily basis, as for grocery shopping, attitude toward brands or companies in general, or the “simple” comprehension of any type of information we obtain.
None of this would happen without the help of my colleagues and most importantly my thesis supervisor, Martine van der Heide. I am beyond grateful for the constant support and availability during this period of my studies. Thank you for leading me through this journey, and always assure, that the whole thesis group was on track!
1. INTRODUCTION ... 1
2. LITERATURE REVIEW ... 4
2.1ESTABLISHMENT OF NUTRITION LABELS ... 4
2.2ASSESSMENT OF CURRENTLY APPLIED NUTRITION LABELS ... 6
2.3NUTRITION LITERACY ... 12
2.4CONCEPTUAL FRAMEWORK ... 13
3. METHODOLOGY ... 14
3.1PARTICIPANTS AND SURVEY DESIGN ... 14
3.2MEASURES ... 16
4. RESULTS ... 18
4.1DESCRIPTIVES ... 18
4.2PURCHASE FREQUENCY ... 19
4.2.1 Direct effect of the suggested FOP bar nutrition label ... 19
4.2.2 Perceived healthiness as mediator of the effect on condition category on purchase frequency ... 20
4.2.3 Direct effect of consumer’s level of nutrition literacy on his/her purchase frequency ... 22
4.2.4 Moderation effect of consumer’s level of nutrition literacy on the main effect of the FOP bar nutrition label on purchase frequency ... 23
4.3CONSUMPTION VOLUME ... 24
4.3.1 Direct effect of the suggested FOP bar nutrition label on consumption volume .... 24
4.3.2 Perceived healthiness as mediator of the effect of condition dummy on consumption volume ... 25
4.3.3 Direct effect of consumer’s level of nutrition literacy on his/her consumption volume ... 27
4.3.4 Moderation effect of consumer’s level of nutrition literacy on the main effect of the FOP bar nutrition label on consumption volume ... 27
5. DISCUSSION ... 30
5.1QUANTITATIVE SURVEY ... 30
5.1.1 Purchase Frequency ... 30
5.1.2 Consumption Volume ... 31
5.2FOLLOW-UP SURVEY ... 34
6. IMPLICATIONS AND RECOMMENDATIONS ... 36
APPENDIX ... 42
APPENDIX A-RESULTS OF RELIABILITY TEST,VARIABLE DESCRIPTIVES ... 42
APPENDIX B-PURCHASE FREQUENCY:SUMMARY OF VARIABLES TESTED ... 43
APPENDIX C-CONSUMPTION VOLUME,SUMMARY OF TESTED VARIABLES ... 45
APPENDIX E–PERCEIVED HEALTHINESS AS THE MEDIATOR ... 49
Section 1: Purchase Frequency ... 49
Section 2: Consumption Volume ... 53
1. INTRODUCTION
In 1994, the Federal and Drugs Administration (FDA) took the initiative to improve the Nutrition Labeling and Education Act (NLEA) by introducing a more unified and comprehensive labeling system including the daily reference value in percentage (%DVs), with the purpose of reducing the confusion in nutrition labels and take actions against rising obesity rates. Unfortunately, there were numerousrestrictions and regulations, which made the NLEA difficult for marketers to apply, but also for consumers to use or understand (Petruccelli 1996). Nowadays, marketing implications became even more complex, taking into account the overwhelming variety of assortment and an infinite number of stimuli in the food industry. Product packages are confusingly crowded (Visschers et al. 2010), and often include health claims or misleading information about the actual content of the product, compromising the overall content transparency (Carels et al. 2006; Chandon & Wansink 2012). Consequently, this leads to a higher intake in calories and underestimation of size and product content such as saturated fats, sugar or cholesterol (Provencher et al. 2009), which further contributes to obesity and various health problems.
Despite the constantly changing trends in consumption of food, one becomes clear: Whether planned, unplanned or impulsive purchases, most of the decisions making and evaluation is taking place in-store (Jae & Delvecchio 2004), and often under time pressure (Drichoutis et. al 2006; van Herpen & van Trijp 2011). Taking in consideration the current store environment, it is not surprising, that the majority of consumers have difficulties to understand and use nutrition labels during their product evaluation (Levy and Fein 1998; McArthur et al. 2001; Rothman et al. 2006).
as the individual’s ability to obtain, process, and understand provided basic health information in order to assess an appropriate health decision. Nevertheless, individuals have limited cognitive capacity to elaborate given information.
Studies demonstrated that during the evaluation process of a product, consumers tend to rely on heuristics while grocery shopping (Petty, Capoccio & Schumann 1983; Szmigin & Piacentini 2018). Especially low-literate consumers use one-to-one correspondence, such as price or size of the package in their product evaluation and base their decision on the context of the purchase (Viswanathan, Hastak & Gau 2009; Pennings et al. 2014).
Viswanathan, alongside other researchers (Hastak, 2002; Hastak & Gau, 2009) have focused on consumers with a lower level of nutrition literacy, who demonstrated to have the most difficulties in visualizing abstract values alike fats or calories. Their suggestion to integrate graphic representations, such as ranges or rates alongside the nutrition label facts proved to significantly improve consumers’ ability to differentiate between healthful and unhealthful products.
The usage of nutrition labels has many advantages, such as moderated intake of sugar and saturated fats and overall improvements of dietary habits (Kreuter et al. 1997; Neuhouser et al. 1999; Hassan et al. 2010; Pennings et al. 2013). Researches also suggest that applying a summary on product packages will help consumers to have a better overview about their dietary intake, evaluate the product with higher accuracy and facilitate the comparison between different products or brand attributes.
In light of the above mentioned, the purpose of this study is to increase the overall salience and transparency of nutrition attributes by examining, (1) how combining visual and verbal information will improve consumer’s understanding and use of nutrition information in favour of healthier food intake. In order to fully comprehend the effectiveness of the proposed FOP bar nutrition label format, (2) the mediation effect of the product's perceived healthiness will be assessed. Furthermore, (3) the moderating effect of nutrition literacy will be analysed, (4) whereby it is assumed that consumers with lower nutrition literacy will benefit more from this graphic layout. The above-mentioned leads to the following problem statement:
“Can a product-level FOP bar nutrition label lead consumers to healthier food choices in-store and therefore, lower the consumption of relatively unhealthy food products?” This study elaborates the statement through several research questions:
1. “To what extent can the suggested FOP bar nutrition label lead to a decrease in purchase frequency and consumption volume of relatively unhealthy food products?” 2. “Does consumers’ level of nutrition literacy affect the functionality- effectiveness of the
suggested label?”
3. “Is the effect of the bar nutrition label stronger for consumers with a lower level of nutrition literacy?”
4. “Does the FOP bar nutrition label influence the perceived healthiness of food products?”
5. “Are consumers willing to use the FOP bar nutrition label as supplementary tool during their product evaluation?”
2. LITERATURE REVIEW
In the following chapter, an introduction to the back-of-package as well as front-of-package nutrition labels will be provided. In order to develop and successfully apply the suggested FOP bar nutrition label, the benefits and implications of currently applied FOP labels have to be taken into account. In the next section of the literature review, I will focus on the most relevant research papers, which discuss consumer’s perception of nutrition labels and how to mitigate the problems related to their misunderstanding. To conclude, the role of consumer’s nutrition literacy during the evaluation process of nutrition labels will be described.
2.1 Establishment of nutrition labels
In the 1970s an increasing interest in food nutritional information started to raise awareness about the missing clarity of nutrition values on packages. In order to clear the confusion, the Food and Drug Administration (FDA) applied regulations, which prohibited the use of not scientifically tested health claims. A benchmark in nutrition facts panel formatting has been the establishment of the Nutrition Labeling and Education Act (NLEA) in 1990, its purpose was to help consumers in their evaluation and recognition of healthier food options (U.S. Food and Drug Administration, 2014). Until then, it was not mandatory for marketers and manufacturers to apply accurate nutrition values on packages.
By 1995, it was compulsory for all food products to implement the FDA’s standardized nutrition panel layout, clearly stating all nutrient contents from product ingredients, health claims, to portion sizes on the back of package. Even though it was a step forward to more informed and educated consumers, it has been criticized for its complexity and difficulties for manufacturers to apply to products (Petruccelli 1996; Pappalardo 1996). They both underlined the necessity of constant improvements in nutrition labeling, considering both market and consumers’ behavioural developments.
labels are also applied as front-of-package labels in a more summarized form. They function as a supplementary tool for better product evaluation. A more detailed description is provided below.
Despite the great effort to develop and display a standardized, more comprehensive and accurate nutrition facts panel, the extensive marketing strategies are negatively influencing consumers’ cognitive capacity to process the provided information. In order to attract consumers’ attention to the product, marketers are constantly seeking to integrate new stimuli on the product package, as well as developing and promoting new variations of current products. However, this approach is creating an overwhelming environment, whereby both BOP and FOP nutrition labels and actual food ingredients are being neglected (Celsi & Olson 1988; Visschers et al. 2010).
According to the "Elaboration Likelihood Model" (ELM) by Petty et al. (1986), individuals process information based on their motivation towards the product, and their ability to perceive and process given information. If consumers perceive the product as personally relevant and therefore their involvement and motivation increase, they are likely to rely on the strength of provided arguments and elaborate it in a more conscious, analytical and reflective manner, described in the literature as the central route of processing. Nevertheless, this approach devotes considerable time and resources to the sense-making process and when depleted, individuals tend to rely on heuristics, well-learned scripts and other relatively effortless guidelines in the evaluation process described as the peripheral route. Lack of involvement or ability to process leads consumers to neglect both label formats, but especially the back-of-package labels, making them more susceptible to impulsive buying and/or poor interpretation of nutrition values (Petty, Cacioppo & Schumann 1986; Szmigin & Piacentini 2018).
with the help of available literature, I will evaluate their effectiveness, by explicitly indicating the overall benefits and shortcomings.
2.2 Assessment of currently applied nutrition labels
Guideline daily amount (GDA), Front-of-Package label
Established in 1998 by the UK government in collaboration with consumers associations and marketers, the GDA labels serve as a summary table of nutrients in relation to the suggested favourable daily intake. This format includes the actual amount and % of Daily Value mostly of critical nutrients such as calories, fats, sugars and salt with respect to the mentioned serving portion (see Figure 1) (Unilevel, Guideline Daily Amount, n.d.).
Figure 1: An example of a GDA front of package nutrition label format. Reprinted from
Foodlabel
In order to make this widely used FOP nutrition label format more intuitive, several visual adjustments were suggested over the past years, as for example:
Multiple Traffic Light (MTL) system
Taking into consideration the above mentioned there was a mutual demand from consumers and manufacturers to apply more comprehensive and user-friendly FOP labels. Therefore, since 2007, the British FSA (alongside EU regulations and based on scientific studies), has been developing a more evaluative labeling system by including traffic light colours to current GDA labels on the front-of-package.
As demonstrated in figure 2, red signals an elevated amount of that specific nutrient also seen as less healthy, yellow is applied when the percentage of that ingredient is at a medium level and lastly, green represents generally healthier product components.
Despite the possibility to compare products at-a-glance and evidence of increase in consumers’ ability to assess the healthiness of options while shopping for groceries (Borgmeier & Westenhoefer 2009; Hersey et al. 2013; Pennings et al. 2014), there is still confusion on what is supposed to be marked as high, middle or low (Talati et al. 2016).
Figure 2: Example of a Multiple Traffic Light (MTL) system used as FOP nutrition label,
Reprinted from Foodlabel
Health star rating (HSR) labels
In Australia and New Zealand, a revised format has been proved to be the most efficient among other summary indicator systems in helping consumers determine the healthfulness of a product (Talati et al. 2016; Lundeberg et al. 2017). Despite positive results, the novelty of the Health Star Rating label (HSR) and its restricted geographical presence has implications on the overall evaluation of this format.
The layout itself includes the classical GDA nutritional references, which are further enhanced with an overall summary star based indicator, representing the healthfulness of the product on a scale from ½ to 5 stars (see Figure 3). The evaluation process becomes easier and faster across many categories.
Despite the positive assessment so far, the same restriction found by the MTL system has been detected (Maubach et al. 2014; Talati et al. 2016). The measurement depends on manufacturers’ or experts’ evaluation, which is often based on estimated adults’ daily intake, neglecting consumers individual preferences and needs.
Figure 3: An example of the Health Star Rating in Australia, Reprinted from Diabetic Health
Healthier Choices logos
Lastly, the most restrictive evaluation form of labels has been applied in a vast variety of formats and designs across many countries. In 1989 Sweden introduced the so-called "Keyhole" logo; Followed by The Netherlands, in 2006, the first health Choices logo was applied on food products (see Figure 4).
During their research about the effectiveness of different front-of-package formats, Feunekes et al. (2008) underlined the advantages of logos such as the rapid evaluation and recognition of the healthier product among alternatives. Van Herpen and Van Trijp (2011) came to the same results, nevertheless, they also found a crucial disadvantage of this format. It is mainly perceived as a health claim and advertising tool applied by manufacturers, causing less credibility and general distrust among consumers. Consumers are not perceiving this format to be transparent enough and the fact that it is not always present is causing confusion. Consumers have no proof whether the logo is missing because the product is unhealthy or only because the format is not mandatory and therefore seen as unnecessary by manufacturers to apply.
Figure 4: the Swedish National Food Agency's Keyhole symbol and the Choices logo in the
Netherlands, Reprinted from Food Navigator
Multiple traffic light formats, together with more evaluative formats such as the star systems or tick logos, have been permanently subject to misinterpretation (Scott & Worsley 1994; Hersey et al. 2013). Therefore, these formats are being applied by manufacturers only voluntarily and included on the front-of-package rather as a supplementary evaluation tool.
range from half the standard serving size up to double the serving sizes, which leads to an underestimation of actual nutritional values and further increase in obesity rates. To mitigate this problem, the FDA considered the inclusion of a full-package data on the product to supplement the per-serving nutrition panel information, however, little has been done until nowadays.
Several studies examined the distribution and layout of information on packages. Visschers et al. (2010) found that using a simplified design of FOP nutrition labels with a moderate amount of information attracts consumers’ attention and improves consumers’ ability to process the content (Pennings et al. 2013). In fact, Viswanathan (1994) found that putting numbers on packages, such as “200 calories” is difficult to be visualized and has to be compared with other subjects to derive meaning. Furthermore, this study was enhanced in 2002, by referring that summary information is principally linked to a product category, and consequently very useful in judging product relative values while comparing different brands (Viswanathan & Hastak 2002).
In sum, research suggests that labels with combined visual and verbal information have a positive effect on the overall usage of nutrition labels, from the assessment of product content to the comparison with other products considered during the evaluation. Building on these findings, I argue that providing a FOP bar nutrition label with nutrition attributes marked along a scale will facilitate the usage of labels while grocery shopping. Not only will the nutritional attributes of a product become more salient, but it also enables consumers to have a more comprehensive overview of the entire product content. This way, the transparency of the product’s content increases and will lead consumers to healthier in-store purchase decisions. A preview of the suggested FOP nutrition label is provided in figure 5. In consequence, I also propose that:
Figure 5: Example of application of the suggested product-level FOP bar nutrition label
A simplified format of nutrition labels provides a simple interpretation tool to assess the product’s overall healthiness without implications on the processing load. (Scott & Worsley, 1994; Lundeberg et al. 2018). Carels et al. (2006) stated, that regardless of consumer's health- or weight-change orientation, they use a wide range of characteristics to judge the healthiness of a product. In fact, the most commonly used attribute is fat content, followed by calories, sugars or proteins. Many other researchers analysed the relationship between perceived healthiness and the effectiveness of nutrition labels. As already mentioned during the evaluation of currently used FOP nutrition labels, product labels referring to the healthiness of a product have a positive effect on overall consumption, and consumer’s welfare (Teisl et al. 2001; Kim et al. 2001).
Taking into account the constantly increasing obesity rate, it is indispensable for manufacturers and researchers to provide a comprehensive evaluation tool to properly communicate product’s nutritional content. Congruent with previous findings, Feunekes et al. (2008) suggested applying a summary of the whole nutritional content as FOP label to offer an overall interpretation of products healthiness, especially referring to improved decision-making and no elevated requirements of consumer's nutritional literacy. Based on the results of above-mentioned studies, I propose that:
H2b: Product’s perceived healthiness mediates the effectiveness of the product-level FOP bar nutrition label on lowering the consumption volume of relatively unhealthy food products.
2.3 Nutrition literacy
Literacy and numeracy have been defined as the ability to understand, interpret, apply and engage with written information in form of text or numbers. It is indisputably essential in order to form any type of further knowledge. Taking into account the context of this study, the literature describes health literacy, also defined as nutrition literacy, as consumers’ ability to identify relevant nutrients and allocate their attributes to positive or negative outcomes such as chronic diseases (Ratzan & Parker, 2000).
In the past years, many researchers analysed the relationship between the level of health literacy and the overall functionality of nutrition labels leading to a uniform result. Lino et al. (1999) considered literacy as the key factor in the assessment of differences in dietary habits. As a result, they found salutary eating habits to be positively associated with higher nutrition literacy skills. In consequence, I assume that:
H3a: An increasing level of nutrition literacy will have a negative effect on consumers’ purchase frequency of food with high levels of sugar, calories and fat.
H3b: An increasing level of nutrition literacy will have a negative effect on consumers’ consumption volume of food with high levels of sugar, calories and fat.
In their experiment, Rothman et al. (2006) found that even participants with a higher level of health literacy have struggled to interpret values written on the nutrition labels in question, especially when engaging with calculations relative to serving size and recommended daily value in percentage. Huizinga et al. (2009) also found that a proper interpretation of nutrition labels is strongly associated with literacy and numeracy skills, explicitly in relation to correct estimations of portion sizes. Considering numeracy and literacy deficits throughout the population, this study suggested implementing a line across the package as a measurement of serving size in order to improve current labels. Based on the results of above-described studies it is assumed that:
H4a: The negative effect of the suggested product-level FOP bar nutrition label on purchase frequency is more pronounced for consumers with lower nutrition literacy.
H4b: The negative effect of the suggested product-level FOP bar nutrition label on consumption volume is more pronounced for consumers with lower nutrition literacy. When assessing the impact of nutrition literacy on health-dietary outcomes, Carbone and Zoellner (2012) underlined the still present gap in research, especially in the case of consumers with lower levels of literacy. The main concern is the lack of evidence and generalizability of available studies, inflicting the overall evaluation of health literacy in the context of nutrition. In sum, many nutrition labels were suggested and tested over the past years. Nevertheless, without a consistent collaboration between manufacturers and policymakers, and without the consideration of all potential factors moderating the effectiveness of nutrition labels, the development of a functional nutrition label format is still at its beginning. In order to test the efficiency of the FOP bar nutrition label and address the suggested hypotheses, an experimental study has been designed and presented in the following part of the paper.
2.4 Conceptual framework
age and level of hunger. Furthermore, the nutrition literacy was included as moderator in the model. To conclude, the mediation of the product’s perceived healthiness will be tested.
Figure 6: Conceptual model
3. METHODOLOGY
3.1 Participants and survey design
A total of 156 participants from 17 different countries, mostly European, engaged in the online survey. The study itself was structured into three main parts. First, the between-subject experiment was presented, followed by the nutrition literacy assessment test and concluded by a follow-up study. The final sample consists of 120 respondents who completed all tasks required for further analysis. The distribution of the study occurred through several social media channels.
After a brief introduction, participants were asked to evaluate three products usually consumed for breakfast; a 1,5L orange juice, a 500g strawberry yogurt and a 350g package of Nutella. The three products were purposely ordered according to their increasing content in sugars and saturated fats. The entire survey can be seen in Appendix F.
between-subjects experiment. In the control group, participants obtained a picture of the three products displayed as regular package formats, representing the currently used product packages. An equal number of participants evaluated the three products with the suggested bar nutrition label displayed on the front of the package. An example of one of the breakfast products in both representations can be seen in figure 7.
Figure 7: Standard product package and modified packaged including the suggested FOP bar
nutrition label
The next section of the study was applied to assess the participant's level of nutrition literacy. Participants obtained two nutrition label panels and were asked to provide answers according to the values written in the panels. Three questions were taken from the Nutrition Literacy Assessment Instrument (NLit), developed by Rothman et at. (2006).
When used as a whole, it has a good internal consistency (Cronbach’s a= 0.84) and high internal reliability (Kuder-Richardson= 20, Coefficient= 0.87). Yet, considering the length of the survey and the context of this study paper, it was not manageable to include all 24 measurement items at disposition. In this case, the focus was on the Food Label and Numeracy section (see in Appendix D).
followed by their consideration about future usage of this nutrition label while grocery shopping. Lastly, an open question to their overall impressions of the suggested FOP bar nutrition label was included. At the end of the survey, various demographic questions were asked to get insights about the sample.
Figure 8: Müsli package with implemented FOP bar nutrition label
3.2 Measures
Regardless of the condition participants were assigned to, in the breakfast section, all participants answered three questions to measure the consumption volume, purchase frequency and perceived healthiness per each product separately. In total, they answered each question three times. This layout was set in order to measure the differences in the effectiveness of the suggested nutrition label by a decreasing healthiness of the displayed product.
flavour, how often they would purchase that type of product per month.
After completing the first task, participants answered four questions related to nutrition labels. Two Back-of-Package nutrition panel tables were presented. Participants obtained the BOP nutrition panel of a peanut butter cookies package and had to indicate how many grams of sugar they would consume by eating three cookies in total. The correct answer was 27 grams. The next part was taken from the Nutrition Literacy Assessment Instrument (NLit), which can be found in the appendix. Participants were asked to look at the BOP nutrition panel of Macaroni and cheese and answer the questions below. The first question was related to calories: “How many calories will you eat if you eat the whole container?” With four options: 250, 500, 700 and 750 calories. The correct answer was option B, 500 calories. The following question referred to saturated fats: “If you are trying to limit your intake of saturated fat to 7 grams per meal, how many cups of macaroni and cheese can you eat if you eat nothing else at the meal?” with four options: 1- 4 cups. The correct answer was 2 cups, equal to 6 grams of saturated fats. Lastly, a question related to a more specific nutrient was presented with two possible answer options of yes and no: "If you are advised to increase your fiber intake, is macaroni and cheese a good food choice?" Considering the nutrition panel indication, where the product has 0 grams of fiber, the correct answer was no. With every correct answer, participants gained 1 point. In total, the score varies from 0 (no correct answers) to 4 (all answers are correct). A detailed description of the coding of the nutrition literacy score is provided in the upcoming chapter. In the last part of this survey, the comparison task was presented to all participants regardless of the condition they were assigned to in the previous section. This section was included not only, to let participants experience a situation, where they would use the bar label to evaluate products from the same category. But also, to understand how participants react and use the suggested label, while asked to make a healthier decision between similar products. The following questions were applied: First, “Regardless of your regular consumption of this specific type of product, use the presented label to determine which of the two products you consider to be healthier.” and "On what information have you based your answer?" With possible options: Sugars, Proteins, Fat, Fiber, Calories, and Others.
approach was applied to the next question. "Would you be using the suggested format of nutrition label in the future while grocery shopping?" (1 – "Definitely not to 5 – "Definitely yes").
Since this label is not currently applied on product packages, an open question was included at the end of the comparison part to obtain participants opinion toward the suggested label. It was explicitly stated, that the displayed FOB bar nutrition label might not be the final design. Few opinions will be included in the discussion part of this paper. The survey was concluded by stating the age, gender, and nationality, as well as the level of hunger on a scale from 0% to 100%.
4. RESULTS
4.1 Descriptives
The final sample consists of 120 participants. Overall, participants were 25% Italian, 21% Czech, 19% German, 13% Dutch followed by 13 other minorities. The sample consists of approximately 60% female and 40% male respondents with the average age of 27.1 years. Eight participants have not provided information regarding their age and level of hunger, hence it has been not accounted in certain models of the analysis.
First of all, a reliability test was conducted to analyse, whether the three items (orange juice, strawberry yogurt, and Nutella) used for measuring the effect of the bar nutrition label (IV) on the purchase frequency (DV1) and consumption volume (DV2), can be considered as one variable in the following analysis. Resulting with Cronbach`s a= 0.273, it implicated to proceed with the analysis by testing each item separately (see Appendix A). There is variation in the means across products, whereby the orange juice has a mean of 20.2 (St. Dev.: 15.2), strawberry yogurt has a mean of 31.3 (St. Dev. 30.5) and Nutella a mean of 9.3 (St. Dev.: 12.8).
Furthermore, the level of nutrition literacy among participants was tested through four separate questions. In order to simplify the interpretation of the scores, each answer was measured to be either correct (1) or wrong (0). The participant could have gained a minimum of 0 points and a maximum of 4 points in total. The overall mean of the nutrition literacy test score is 3.31. In total, 67 (55,8%) participants answered correctly to all four questions. Another 25% managed to answer 3 out of 4 questions correctly, 13,3% 2 out of 4, and 5,8% of the respondents answered correctly one question (see Graph 1). Among the questions, the highest percentage of incorrect answers occurred by the question related to saturated fats with 23,3%, followed by the question about sugars with 20,8% of incorrect answers. All tables related to the Nutrition Literacy test can be found in Appendix B.
Graph 1: Overview of correct answers scored in the nutrition literacy test
We proceed with the overall analysis by assessing each hypothesis separately, with respect to the dependent variable.
4.2 Purchase Frequency
4.2.1 Direct effect of the suggested FOP bar nutrition label
Following the structure mentioned before, each product will be analysed separately. Results show, that the overall model explaining the effect of the suggested FOP bar nutrition label on the purchase frequency of orange juice is not significant (R2= 0.047; p-value= 0.265), whereby the impact of the FOP bar nutrition label on the purchase frequency is also insignificant (b= 1.94; p-value= 0.268).
In the case of strawberry yogurt, similar results occurred. Although the overall model is not significant (R2= 0.042; p-value= 0.561), the effect of the FOP bar nutrition label shows marginal significance (b= -2.688; p-value= 0.079). The finding suggests, that by applying the bar nutrition label, the purchase frequency of the strawberry yogurt declines. Congruent with the two items above, the overall model stating the effect of the FOP bar nutrition label on the purchase frequency of Nutella is equally not significant (R2= 0.036; p-value= 0.697) and the condition variable is insignificant (p-value= 0.636). For each item, R2 is very close to zero, suggesting that the variables do not explain enough overall variance. (Malhotra et al., 2007). Taking in consideration all three items, the hypothesis 1a, the presence of the FOP bar nutrition label has a negative effect on consumers’ purchase frequency of food with high levels of sugar, calories and fat, must be rejected.
4.2.2 Perceived healthiness as mediator of the effect on condition category on purchase frequency
Starting with the healthiest product used in the survey, the overall summary of the model is significant (R2= 0.102; p-value= 0.034), whereby the direct effect of perceived healthiness, (described as b(MY) in table 1), on purchase frequency of orange juice is also significant (b= 0.088; p-value= 0.005). Nevertheless, the direct effect of the condition dummy (c) on the purchase frequency resulted insignificant, as already demonstrated in chapter 4.2.1. When the perceived healthiness has been included in the model, it maintained its insignificant coefficient c’= (b= 1.7921; p-value= 0.271).
Looking at the values regarding whether the purchase frequency of the strawberry yogurt is mediated thought participant’s perception of product healthiness, we can conclude that the overall model is significant (R2= 0.124; p-value= 0.0007). Also in this case, perceived healthiness, described as b(MY), is positive and significant (b= 0.0879; p-value= 0.0008). The condition dummy demonstrates insignificant values regardless of whether the effect is direct or mediated.
Lastly, the overall model explaining the mediation effect for the least healthy item, Nutella, proved to be equally significant (R2= 0.164; p-value= 0.007), whereby the perceived healthiness had a significant effect on purchase frequency (b= 0.264; p-value= 0.0035).
The above-mentioned draws a conclusion, hypothesis 2a, that product's perceived healthiness mediate the effectiveness of the product-level FOP bar nutrition label on lowering the purchase frequency of relatively unhealthy food products, is rejected.
Figure 9: Mediation effect of perceived healthiness, framework
Table 1: Effects and significance of tested variables
*Significance by p-value <0.05
a= direct effect of condition dummy on perceived healthiness b= direct effect of perceived healthiness on purchase frequency c= direct effect of condition dummy on purchase frequency
c`= direct effect of condition dummy on purchase frequency when the mediator is included 4.2.3 Direct effect of consumer’s level of nutrition literacy on his/her purchase frequency
In order to assess the direct effect of consumer’s level of nutrition literacy on the purchase frequency of relatively unhealthy food products, a regression analysis with nutrition literacy score regressed on consumers’ purchase frequency was conducted. Again, each item will be analysed separately. The model regarding the purchase frequency of orange juice is overall insignificant (R2=0.047; p-value= 0.265). Therefore, we can conclude that there is not enough variance explained by the variables.
In addition, the variable predicting the negative effect of consumer’s nutrition literacy level on purchase frequency is also insignificant (b= 0.355; p-value= 0.667). In case of consumer’s purchase frequency of strawberry yogurt, the overall model emerged to be insignificant (R2= 0.042; p-value= 0.384). The nutrition literacy score, demonstrated to have no effect on consumer’s purchase frequency of strawberry yogurt (b= 0.549; p-value= 0.499). Finally, for Nutella, nutrition literacy score presented insignificant values as well (b= 0.025; p-value= 0.985).
In summary, hypothesis 3a, that an increased level of nutrition literacy will have a negative effect on consumers’ purchase frequency of food with high levels of sugars, calories and fat, is not supported.
Outcome Orange Juice Strawberry Yogurt Nutella
Coefficient b p-value b p-value b p-value
4.2.4 Moderation effect of consumer’s level of nutrition literacy on the main effect of the FOP bar nutrition label on purchase frequency
The interaction effect between consumers' level of nutrition literacy and the condition dummy proved to have no significant effects. There is no interaction effect of the purchase frequency of orange juice (b= -0.748; p-value= 0.663) and strawberry yogurt (b= -0.462; p-value= 0.786), values provide no significant effects. In case of Nutella, the interaction effect is insignificant (b= 1.051; p-value= 0.50). There is no support for a moderation effect of consumer’s level of nutrition literacy on the functionality of the proposed bar nutrition label on lowering the purchase frequency of relatively unhealthy food products.
Yet, it has been hypothesized, that the effectiveness of the suggested bar nutrition label varies depending on consumers’ level of nutrition literacy. Whereby it is suggested, that consumers with lower nutrition literacy will benefit more from the applied label. The conditional effect of the independent variable on the dependent variable at the values of the moderator was inspected. Regardless of consumers’ level of nutrition literacy, none of the items provided significant values, therefore, no further analysis is conducted. A detailed overview for each tested item can be found in Appendix B.
In consequence, hypothesis 4a, that the negative effect of the suggested product-level FOP bar nutrition label on purchase frequency is more pronounced for consumers with low nutrition literacy, is rejected.
Table 2: Effects and significance of tested variables Purchase
Frequency Orange Juice Strawberry Yogurt Nutella
Model summary R
2=0.047 R2=0.042 R2=0.036
Sig.=0.265 Sig.=0.561 Sig.=0.697
Coefficients b p-value b p-value b p-value
Condition dummy 1.939 0.268 -2.688 0.079* 0.574 0.636 NLScore 0.355 0.667 0.549 0.499 0.025 0.985 Condition dummy x NLScore -0.748 0.663 -0.462 0.786 1.051 0.50 Level of hunger 0.061 0.031* 0.024 0.482 0.034 0.193 Age -0.002 0.982 -0.030 0.802 -0.042 0.434 * Significance of p-value <0.05
b= the impact of the predicting variable on the dependent variable NLScore= Nutrition literacy assessment test score
4.3 Consumption Volume
4.3.1 Direct effect of the suggested FOP bar nutrition label on consumption volume The same analysis structure was used to determine the main effect of the FOP bar nutrition label on consumers’ consumption volume of all three items. In total three analyses were performed using the PROCESS tool in Spss 25, developed by Andrew F. Hayes, and specifically with Model 1. The condition dummy (No label/FOP bar nutrition label) was regressed on the consumption volume, whereby the participant's age and level of hunger were included as control variables. All coefficients are summarized in table 4. The overview of the estimated marginal means per each product can be found in graph 2.
The overall model summary of the consumption volume anent the orange juice is significant (R2= 0.127; p-value= 0.000). However, there is no significant effect of the suggested FOP bar nutrition label on the consumption volume of orange juice (b= 3.1; p-value= 0.299).
The overall model related to the strawberry yogurt proved to be insignificant (R2= 0.050; value= 0.417) and also the main effect of the condition dummy was insignificant (b= 2.14; p-value= 0.735).
The overall model for Nutella proved to be insignificant (R2= 0.119; p-value= 0.144) and the condition dummy demonstrated to have no significant effect on the consumption volume of Nutella (b= -1.866; p-value= 0.321).
Graph 2: Differences in consumption volume between the items across the two conditions
4.3.2 Perceived healthiness as mediator of the effect of condition dummy on consumption volume
In order to assess, whether perceived healthiness mediates the effectiveness of the bar nutrition label on lowering consumers’ consumption volume of relatively unhealthy food products, the same analysis structure as seen in Chapter 4.2.2 will be applied. The category dummy (No label vs. FOP bar nutrition label) was set as the independent variable (X), the total consumption volume per item was used as dependent variable (Y) and finally, the perceived healthiness per item was the mediator variable (M). To remain consistent with the previous model, level of hunger and participant's age were included as control variables to evaluate the possible mediation effect. The analysis was conducted through the Process tool in Spss, specifically Model 4. All coefficients can be found in Appendix E, section 2 and further summarized in table 3.
First, the mediation effect of perceived healthiness on consumption volume of orange juice was tested, providing highly significant values (R2= 0.165; p-value= 0.000). Although the model as a whole proved no significance, with the direct effect of the condition dummy on the dependent variable insignificant (b= 2.997; p-value= 0.303), the effect of M on Y is significant (b= 0.156; p-value= 0.017). While evaluating the results related to the consumption volume of strawberry yogurt, the overall summary of the mediation effect is significant (R2= 0.096; p-value= 0.007), whereby also in this case, perceived healthiness shows to be significant (b= 0.276; p-value= 0.017). Thus, the model summary explaining the direct effect of the X on Y is insignificant (p-value= 0.407). Proceeding with Nutella, the mediation model summary is equally significant (R2= 0.211; p-value= 0.0005) and the effect of perceived healthiness on the consumption
volume is also significant (b= 0.514; p-value= 0.0002). Despite the significance of the overall model explaining the direct effect of X on Y (R2= 0.037; p-value= 0.046) the condition dummy was not significant as well (b= -2.221; p-value= 0.266).
Considering the above stated, hypothesis 2b, product’s perceived healthiness mediates the effectiveness of the product-level FOP bar nutrition label on lowering the consumption volume of relatively unhealthy food products, is rejected. Yet, there is a positive significant effect of perceived healthiness on lowering the consumption volume of relatively unhealthy products, which can be seen in the difference between the c and c’ relationships in table 3.
Figure 10: Mediation effect of perceived healthiness, framework
Table 3: Effects and significance of tested variables
Outcome Orange Juice Strawberry Yogurt Nutella
Coefficient b p-value b p-value b p-value
c (XY) 2.997 0.303 1.53 0.806 -2.221 0.266 a (XM) 2.681 0.565 -3.791 0.472 1.123 0.507 b (MY) 0.156 0.017* 0.276 0.017* 0.514 0.0002* c’ (XY)’ 2.58 0.373 2.576 0.668 -2.799 0.134 Age -0.444 0.0001* -0.697 0.108 -0.91 0.114 Level of Hunger 0.122 0.042* 0.072 0.449 0.337 0.554 * Significance of p-value <0,05
a= direct effect of condition dummy on perceived healthiness b= direct effect of perceived healthiness on purchase frequency c= direct effect of condition dummy on purchase frequency
4.3.3 Direct effect of consumer’s level of nutrition literacy on his/her consumption volume
As the effect of consumer’s nutrition literacy level was also regressed on the consumption volume of relatively unhealthy food products, the same analysing procedure was followed. First, the overall model related to the orange juice proved to be significant (R2= 0.127; p-value= 0.000). Thus, the direct effect of the nutrition literacy score was insignificant (b= -0.261; p-value= 0.847). In the case of the strawberry yogurt, the nutrition literacy score proved to be insignificant with (b= -2.481; p-value= 0.561). Concluding with Nutella, the direct effect of nutrition literacy was equally insignificant with (b= -1.290; p-value= 0.412).
Considering the above listed, hypothesis 3b, stating that an increasing level of nutrition literacy will have a negative effect on consumers’ consumption volume of food with high levels of sugars, calories, and fats, is rejected.
4.3.4 Moderation effect of consumer’s level of nutrition literacy on the main effect of the FOP bar nutrition label on consumption volume
In order to assess whether consumer's level of nutrition literacy moderates the effect of the proposed bar nutrition label on consumer's consumption volume of relatively unhealthy food products, the results of the above-executed moderation analysis will be considered. The same procedure as for other variables was applied, whereby each item was evaluated separately due to their nonconformity.
Here, the moderation effect in the case of orange juice is marginally significant (b= 4.594; p-value= 0.091), while the moderation effect regarding the strawberry yogurt is not significant (b= 3.716; p-value= 0.652). The distribution of the varying consumption volume of orange juice, with regard to the nutrition literacy score, is pictured in the scatterplot below (graph 3). Different results have been found by the product with a higher amount of sugars and fats such as Nutella. Here, the interaction effect between the condition dummy and the nutrition literacy score is negative and marginally significant (b= -1.866; p-value= 0.062).
list. For participants with higher nutrition literacy, the effect of the label is not significant (b= 2.482; p-value= 0.417). In contrast, for participants who scored low in the nutrition literacy test, the effect of the bar nutrition label on diminishing the consumption volume of Nutella was proved to be marginally significant (b= -7.26; p-value= 0.054). Despite the overall model summary being not significant (R2= 0.119; p-value= 0.144), the results provide accountable proof for the suggestion, that the lower consumer’s nutrition literacy level is, the more they benefit from the applied bar nutrition label. The distribution of the varying consumption volume of Nutella, with regard to the nutrition literacy score, is pictured in the scatterplot below (Graph 4).
In consequence, hypothesis 4b, that the negative effect of the suggested product-level FOP bar nutrition label on consumption volume and purchase frequency is more pronounced for consumers with lower nutrition literacy, is rejected.
Graph 3: Differences in Consumption Volume of orange juice upon the condition dummy
Graph 4: Differences in Consumption Volume of Nutella upon the condition dummy varying
across different levels of nutrition literacy
Table 4: Effects and significance of tested variables Consumption
Volume Orange Juice Strawberry Yogurt Nutella
Model summary R
2=0.127 R2=0.050 R2=0.119
Sig.=0.000* Sig.=0.417 Sig.=0.144
Coefficients b p-value b p-value b p-value
Condition dummy 3.104 0.299 2.136 0.735 -1.866 0.321 NLScore -0.261 0.847 -2.481 0.561 -1.290 0.412 Condition dummy x NLScore 4.594 0.091** 3.716 0.652 -1.866 0.062** Level of Hunger 0.116 0.054* 0.062 0.562 0.029 0.576 Age -0.474 0.000* -0.738 0.106 -0.183 0.053*
* Significance of p-value <0.05, **marginally significant with p-value <0.10 R2= total variance in the model explained by the variables
5. DISCUSSION
This chapter serves as a general discussion of the results, by assessing the hypotheses and providing a detailed explanation of the occurred effects found in the analysis. The results of the follow-up study will be also mentioned at the end of this chapter.
The overall aim of this study was to investigate the effectiveness of nutrition labels in providing consumers sufficient information about the product, to enable them healthier purchase decisions and a more conscious overall food consumption. The focus was on consumers with a lower level of nutrition literacy, whereby the literature already underlines many problems these consumers may face while evaluating a product from its package (Rothman et al. 2006; Viswanathan et al. 2009). Despite the increasing awareness of the problem over the past decades and extensive researches related to it, there are only few suggestions and solutions proposed to mitigate the issue.
5.1 Quantitative survey
In line with the expectations, the study demonstrated that the suggested FOP bar nutrition label can be potentially applied as supplementary evaluation tool on product packages. Nevertheless, none of the nine hypotheses resulted in being significant with mostly too little variance explained. An overview of the hypotheses and their results can be found in table 5. The two dependent variables will be described separately.
5.1.1 Purchase Frequency
Despite the mediation model being insignificant, the mediated relationship of the FOP bar nutrition label on decreasing purchase frequency of relatively unhealthy food provided interesting insights. The results of a=XM and b=MY demonstrated that when the suggested label was applied on products, participants perceived Nutella and orange juice to be less healthy than when using regular product packages (see Table 1). Consequently, the purchase frequencies of the two products were lower.
There was a significant direct effect of perceived healthiness on purchase frequency, meaning that if participants considered the product to be less healthy as such, their purchase frequency decreased. Overall, the finding is congruent with the results of the study conducted by Teisl et al. (2001), stating that providing health-related information do not imply consumers to change their purchasing behaviour from relatively unhealthy to healthy alternatives.
Lastly, the effect of participants’ nutrition literacy test score was tested. There was no direct effect of the nutrition literacy score on the purchase frequency of any of the displayed products. Which implies, that participant’s level of nutrition literacy had no influence on lowering products purchase frequencies. Furthermore, the moderation effect of nutrition literacy on the effectiveness of the FOP bar nutrition label resulted to be insignificant. In other words, there was no difference in the purchase frequency of the items, regardless if participants scored high or low in the nutrition assessment test. This result can be also explained by the nature of the product packages, larger package sizes are generally bought less frequently per month than smaller packages.
5.1.2 Consumption Volume
The second dependent variable tested, was the consumption volume of orange juice, strawberry yogurt, and Nutella. There was no direct effect of the condition dummy on the consumption volume. Meaning that, whether participants received products with or without the bar nutrition label at the front of package, there was no significant difference in their consumption volume. In order to maintain the consistency throughout the analysis, age and participant’s level of hunger were included in the models as control variables. Both control variables were significant by the orange juice, none by strawberry yogurt, and only age by Nutella. As result, the older participants were, the less orange juice and Nutella they would consume and the hungrier they were, the more orange juice they would consume.
The more participants perceived the product to be healthy, the higher their consumption was. Nevertheless, there was no mediation effect, considering the insignificant relationship between the predictor and the outcome. Here, it is important to acknowledge the difference in the average consumption volume. Compared with orange juice and strawberry yogurt, it is very low for Nutella. In fact, people usually do not eat a large amount of Nutella on average, compared to the consumption of orange juice and strawberry yogurt. This finding is also considered to be the cause of the low variance demonstrated in the models, and most importantly, the incompatibility of the items represented by the low Cronbach’s a in the reliability test. In line with results of Borgmeier et al. (2009), despite the influence of different label formats on product’s perceived healthiness, there is no major impact on the actual food intake or food choice.
In the case of the second dependent variable, participants’ level of nutrition literacy demonstrated to be a relevant predictor of the changes in consumption volume. While, there was no direct effect of the nutrition literacy score on the consumption volume, the interaction effect of the nutrition literacy score and the suggested FOP bar nutrition label on the outcome was marginally significant for orange juice and Nutella. This relationship was tested at different levels of nutrition literacy. For orange juice with the bar nutrition label on the package, the consumption volume increases the more literate participants are, but it is on the contrary for less healthy products such as Nutella. This finding underlines the outcome of the study by Lino et al (1999), where literacy was considered the main factor of distinctive dietary habits among consumers. No differences were found for the third item. As expected, the suggested FOP bar nutrition label had a more pronounced impact on relatively unhealthy food or products with a higher amount of sugars, fat or calories. This finding confirmed the objective of this study, whereby the increased salience of the product’s content positively influenced consumer’s ability to evaluate the product for its content and not its appearance.
bar nutrition label as such, is able to lead consumers to healthier decisions when evaluating food products.
When comparing the results of the two dependent variables, divergent outcomes were found. The focus of the study was to assess the difference in the effectiveness of the suggested bar nutrition label by products with different nutritional contents. While for purchase frequency, no moderation effect and also no relevant difference throughout the varying nutrition literacy levels occurred, the suggested bar nutrition label had an impact on the consumption of the less healthy product of this study. One explanation for this finding is that the overall level of participant’s nutrition literacy was relatively high, and therefore not enough dimensional variance was provided. This implication can be also considered as the cause of insignificant values in the overall models, where not enough variance emerged within the sample.
Table 5: Overview of study hypotheses
Hypotheses Results
H1a: The suggested product-level FOP bar nutrition label has a negative effect on
consumers’ purchase frequency of food with high levels of sugar, calories and fat.
Not supported H1b: The suggested product-level FOP bar nutrition label has a negative effect on
consumers’ consumption volume of food with high levels of sugar, calories and fat.
Not supported H2a: Product’s perceived healthiness mediates the effectiveness of the
product-level FOP bar nutrition label on lowering the purchase frequency of relatively unhealthy food products.
Not supported
H2b: Product’s perceived healthiness mediates the effectiveness of the
product-level FOP bar nutrition label on lowering the consumption volume of relatively unhealthy food products.
Not supported
H3a: An increasing level of nutrition literacy will have a negative effect on consumers’ purchase frequency of food with high levels of sugar, calories and fat.
Not supported H3b: An increasing level of nutrition literacy will have a negative effect on
consumers’ consumption volume of food with high levels of sugar, calories and fat.
Not supported H4a: The negative effect of the suggested product-level FOP bar nutrition label on
purchase frequency is more pronounced for consumers with lower nutrition literacy.
Not supported
H4b: The negative effect of the suggested product-level FOP bar nutrition label on
consumption volume is more pronounced for consumers with lower nutrition literacy.
5.2 Follow-up survey
Considering the novelty of the suggested bar nutrition label, a qualitative study was included at the end of the survey to get primary insights about consumer's perception of this particular label. The aim of this task was to enable participants to experience the evaluation of two similar products of the same category while using the bar nutrition label applied on the front of package. As result, 95 participants (80,51%) selected the Müsli package displayed as "B" and based their decision mostly on the sugar content of the product (see Graph 5). The graphic representation manifests the diversity of arguments people base their decision on while grocery shopping, discussed by Carels et al. (2006).
Graph 5: Argument of choice of the Müsli package in the comparison task
Graph 6: Participants evaluation of ease of understanding of the suggested FOP bar nutrition
label
Graph 7: Participants evaluation of future use of the FOP bar nutrition label
misjudgement of the portion people actually consume, which is also a part of the study conducted by Seiders and Petty (2004) and Huizinga et al. (2009). To conclude, one participant particularity matched the aim of this study with his comment, and therefore deserves to be fully mentioned:
"I love the fact that this new label would make the amount of ingredients very visible. Always when people want to show how much sugar there is in Coca-Cola and how bad it is for you,
they stack little cubes of sugar. This works because you only then realize how much sugar you're actually consuming. And this package does the same type of thing. Knowing that I would consume physically this amount of sugar in a package, I would definitely reconsider
buying and consuming it." Thomas, 22
6. IMPLICATIONS AND RECOMMENDATIONS
Despite the clear and short structure of the questionnaire, the novelty of the suggested bar nutrition label increased the difficulty of the survey. A more extensive introduction of the suggested format could have helped participants in their evaluation, nevertheless, it was not included not to bias their responses. Overall consumers evaluated the suggested FOP bar nutrition label to be an improvement of currently used nutrition labels. More than 50% of them would use the bar nutrition label as a supplementary tool for a more accurate product evaluation while grocery shopping. This positive approach to the suggested label underlines the importance of the experimental study.
with products or nutrition labels in general, but also time pressure in store, could have been relevant to acknowledge.
Speaking of time pressure, one participant underlined another possible limitation of this study, which was not included in the model. Often, people do not take the time to read the labels thoroughly, as they mostly buy according to their habits. This argument matters also when looking at the products chosen for the survey. Restricting the survey to breakfast products was practical for congruency of the study, but quite restrictive for a better understanding of consumers' consumption and purchase intentions of the three products. Despite the consistent evaluation scales and questions among all products, the products themselves proved to be independent, which might have influenced the results of the analysis. In addition, providing another type of less healthy product instead of Nutella could have provided better fitting values with the remaining two products regarding the consumption volume and purchase frequency.
Another limitation relevant to the outcome was the overall high level of nutrition literacy among participants, where almost 60% obtained all four points from the test and another 25% answered three out of four questions correctly, which again provided not enough variance for the moderation analysis. Future researches should extend the nutrition literacy assessment test in their survey in order to obtain more accurate results of consumer’s nutrition literacy. Here, only four questions were asked to assess the level of nutrition literacy, which might have mitigated the reliability of the used measurement tool (NLit), by Rothman et al. (2006).
Overall, the study serves as a guideline for further research in the effectiveness of similar formats. The format may be also more effective when respective colours would be incorporated. Thus, there are several aspects to be analysed, such as the effect of the used colours and the disposition on the package itself. In many cases, it is recommended to design the package in a simple way and highlight more relevant product information such as its content.
REFERENCES
Borgmeier, I., & Westenhoefer, J. (2009). Impact of different food label formats on healthiness evaluation and food choice of consumers: a randomized-controlled study. BMC public health, 9(1), 184.
Burton, S., Biswas, A., & Netemeyer, R. (1994). Effects of alternative nutrition label formats and nutrition reference information on consumer perceptions, comprehension, and product evaluations. Journal of Public
Policy & Marketing, 36-47.
Carbone, E. T., & Zoellner, J. M. (2012). Nutrition and health literacy: a systematic review to inform nutrition research and practice. Journal of the Academy of Nutrition and Dietetics, 112(2), 254-265.
Carels, R. A., Harper, J., & Konrad, K. (2006). Qualitative perceptions and caloric estimations of healthy and unhealthy foods by behavioral weight loss participants. Appetite, 46(2), 199-206.
Celsi, R. L., & Olson, J. C. (1988). The role of involvement in attention and comprehension processes. Journal of consumer research, 15(2), 210-224.
Chandon, P., & Wansink, B. (2012). Does food marketing need to make us fat? A review and solutions. Nutrition reviews, 70(10), 571-593.
Diabetic Health Clinic, The Australian Health Star Rating, does it work? (2016) Retrieved October 12 from https://www.diabetichealthclinic.org/the-australian-health-start-rating-does-it-work/
Feunekes, G. I., Gortemaker, I. A., Willems, A. A., Lion, R., & Van Den Kommer, M. (2008). Front-of-pack nutrition labelling: testing effectiveness of different nutrition labelling formats front-of-pack in four European countries. Appetite, 50(1), 57-70.
Foodlabel, Front of pack labeling (2018). Retrieved October 12, 2018 from http://www.foodlabel.org.uk/label/front-of-pack-labelling.aspx
Foodwatch, Red, amber and green for understandable information (2017). Retrieved October 12, 2018 from https://www.foodwatch.org/more-information/how-traffic-light-labeling-works/
Food Navigator, Healthy Logo: Netherlands "choice" logo confirmed as first government-backed scheme in
EU (2013). Retrieved October 12, 2018 from
https://www.foodnavigator.com/Article/2013/04/16/Netherlands-Choices-logo-confirmed-as-first-government-backed-scheme.jpg
