Do We Really Eat Less with Smaller Bowls and Warning Labels?

(1)

Do We Really Eat Less with Smaller Bowls

and Warning Labels?

A Lab Study Examining the Effect of Bowl Size Reductions,

Warning Labels and Their Interplay on Food Consumption Quantity

(2)

Master Thesis

University of Groningen Faculty of Economics and Business

Department of Marketing M.Sc. Marketing Management

Do We Really Eat Less with Smaller Bowls

and Warning Labels?

A Lab Study Examining the Effect of Bowl Size Reductions,

Warning Labels and Their Interplay on Food Consumption Quantity

Author: Nadine Brach Finkenweg 14 29352 Adelheidsdorf Germany Student number: S3895025 E-mail: n.brach@student.rug.nl Phone: +4915734597863

First supervisor: M.Sc. M.T. (Martine) van der Heide Second supervisor: Prof. dr. ir. K. (Koert) van Ittersum

(3)

Abstract

Across the globe, an increasing number of humans are being affected by obesity, overweight and non-communicable diseases caused by overeating. This in turn leads to rapidly increasing expenditures, such as in the health care sector. Therefore, in order to tackle this obesity epidemic, it is to explore interventions for decreasing the food consumption quantity.

In a laboratory study with a 2 (bowl size reduction: absent vs. present) x 2 (warning label: absent vs. present) between-subject design, bowl size and package were manipulated while keeping the amount of food provided constant. The findings suggest a significant main effect of warning label on consumption quantity. Bowl size reduction combined with a warning label did not influence the consumption quantity, with the results instead indicating that the warning label’s effect is suppressed by bowl size reduction. This study further revealed that the most effective method for reducing food consumption quantity involves displaying a warning label on the package and consuming the food out of a standard sized bowl. Furthermore, the effects were not moderated by a consumer’s level of health consciousness.

(4)

Preface

This master thesis was written as a final project to complete my Master’s Degree in Marketing Management at the University of Groningen (RUG). I have possessed an interest in promoting healthier lifestyles for some time now. As such, the topic of this research, namely nudge interventions to influence food consumption, caught my interest. I consider it important to learn more about how subtle nudges, especially in combination, can help people to consume less, in turn serving to promote a healthier life. Furthermore, I was offered the unique opportunity to study consumption quantity in the lab, and thus measure actual consumption, which is less common than studying food choice or intended consumption.

During this thesis process, I found great pleasure in reviewing literature, conducting my own lab study, and writing. In particular, performing a lab study with food consumption presented a unique and exciting adventure. Although the thesis process generally offered a pleasant experience, it was also a challenging one. At times, acquiring in-depth knowledge on the topic, especially regarding the combined effects, proved to be highly demanding, as did planning the lab study. Overall, however, the thesis process was a rewarding journey that taught me numerous valuable lessons.

(5)

I

List of Figures

Figure 1: Visualization of the Delboeuf illusion ... 6

Figure 2: Chilean “high in calories” warning label ... 8

Figure 3: Conceptual model ... 12

Figure 4: Bar chart of frequencies of highest educational qualification ... 13

Figure 5: Warning label used for manipulation ... 15

Figure 6: Bar chart of frequencies of consumption quantity ... 20

Figure 7: Consumption quantity as a function of bowl size reduction and warning label ... 23

Figure 8: Consumption quantity in grams as a function of warning label ... 24

Figure 9: Moderation model depicting the effect of bowl size reduction and warning label... 25

List of Tables

Table 1: Study design ... 14

Table 2: Results of factor and reliability analyses ... 20

Table 3: Descriptive statistics for consumption quantity, health consciousness, healthiness of past eating behavior, hunger, taste, attitude, purchase intention, perceived healthfulness and nutrition information ... 21

(7)

1

1. Introduction

Not only does overconsumption of food represent a leading cause of overweight and obesity, accounting for 2.8 million deaths each year (World Health Organization, 2017), but it is also a key risk factor in non-communicable diseases, such as certain types of cancer, cardiovascular diseases and type II diabetes (Global Food Security, 2016). Yearly disease and deaths related to overweight and obesity account for approximately US$2 trillion in expenditures (Global Food Security, 2016). Over the past 40 years, the prevalence for obesity has also increased tremendously, rising from 1% to 6–8% in children, 3% to 11% in men and 6% to 15% in women (Jaacks et al., 2019).

In order to tackle the issues of decreasing consumption and promoting healthier eating, the interest in nudge interventions has grown among policymakers, food marketers and researchers (Cadario & Chandon, 2019). Thaler and Sunstein (2009, p. 6) defined nudges as “any aspect of the choice architecture that alters people’s behavior in a predictable way (1) without forbidding any options or (2) significantly changing their economic incentives.” Furthermore, a nudge “must be easy” (Thaler & Sunstein, 2009, p. 6), making nudges a helpful way to fight obesity with low costs and easy implementation. As such, nudges demand further attention by researchers, policy makers, marketers and consumers. Literature differentiates between three types of nudges, namely affectively, behaviorally and cognitively oriented nudges (Cadario & Chandon, 2019). This paper focuses solely on behaviorally and cognitively oriented nudges, since previous research in the field of promoting healthier eating has identified these two intervention types as especially promising due to their effectiveness and implementation possibilities (Cadario & Chandon, 2019; Ikonen, Sotgiu, Aydinli, & Verlegh, 2019). In particular, interaction effects between behavioral and cognitive nudges seem to offer a promising area of research in influencing consumers toward healthier consumption choices (Cadario & Chandon, 2019).

(8)

2 kcal). People consume more than 70% of their daily calories using aids such as bowls, cups or plates (Wansink, Painter, & North, 2005). As such, it is of utmost importance to better understand how tableware size can be utilized to reduce consumption quantity. Therefore, the question arises concerning how interventions aiming at the opposite direction, namely reducing bowl size, affect consumption quantity. Consumption quantity reflects how much food a person eats in total independent of the food’s classification as healthy or unhealthy.

Cognitively oriented nudges comprise interventions such as nutrition fact labels, traffic light labels and placing healthier options visibly at eye level (Cadario & Chandon, 2019). While considerable research has been conducted concerning nutrition labeling, little research has investigated warning labels, especially regarding their influence on consumption (Ikonen et al., 2019). Chile was the first country to implement mandatory warning labels for products excessive in calories, saturated fat, sodium and sugar (Reyes et al., 2019). A qualitative analysis after the first year of their implementation suggested that these labels are effective in changing consumer’s knowledge about product healthfulness and can ultimately lead to behavioral changes in purchase intention of unhealthy products (Correa et al., 2019). Currently, politicians in numerous other countries, such as Brazil and India, are discussing the implementation of warning labels (Reyes et al., 2019). Because of this, warning labels are highly relevant for policy makers as well as consumers, who can profit from increased knowledge and changing behaviors. To support the policy changes, and thus eventually help to reduce overweight and obesity, researchers are tasked with supporting these processes by providing evidence regarding warning label’s effectiveness and the conditions under which such labels are effective. Accordingly, this paper extends previous research by investigating warning labels’ effects on food consumption.

(9)

3 combining simultaneously executed strategies based on nudging and social marketing, such as product, price, placement and promotion, leads to an increase in healthy food choices and sales (Velema, Vyth, Hoekstra, & Steenhuis, 2018). Despite this, no research is available (to our awareness) concerning how combining interventions influences consumption quantity. Thus, research needs to further explore how combining different interventions contributes to reducing consumption quantity.

In summary, this research addresses the following research question: “How does bowl size reduction and the presence of a warning label influence consumption quantity?” Furthermore, this paper investigates how potential synergies between bowl size reduction and warning label influence consumption quantity. Additionally, this study explores whether consumers’ level of health consciousness influences the effect of bowl size reduction and the presence of a warning label, respectively.

In order to investigate the research question, first, consumption quantity is defined and different aspects influencing the quantity one consumes are introduced. Afterwards, a short overview of nudge interventions is provided, followed by an in-depth reflection on bowl size reduction and its role in decreasing consumption quantity. In the subsequent section, warning labels and their effectiveness in influencing consumption quantity are discussed. This is followed by a definition of health consciousness and a review of literature regarding its moderating role. Subsequently, the conceptual framework is developed. After this, the methodology for the lab experiment investigating the proposed relationships is presented. The next chapter provides the analysis of the conducted data. Afterwards, the results and contributions of this master thesis are summarized. Based on the data analysis policy and managerial implications are discussed. Furthermore, this chapter reflects on limitations of this study and suggests directions for further research. Lastly, the main findings of this research are presented in the conclusion.

2. Theoretical Background

(10)

4

2.1. Consumption Quantity

Broadly defined, food consumption features two sequential parts. First, humans decide what to eat, which can be described as food choice and could, for example, be either healthy or unhealthy. After selecting a food or a meal, humans then face the decision of how much to consume, which refers to this paper’s focus on consumption quantity, also referred to as consumption volume. Defined more narrowly, consumption consists of solely the second part of the broader definition, which describes “actually eating the food selected, specifically at the appropriate serving size” (Rafacz, 2019, p. 652). A person can consume the total amount of the selected food or parts of it. It is also possible for a person to overconsume, referring to “eat[ing] more of a particular food than would be considered healthy” (Rafacz, 2019, p. 652).

An individual’s food consumption is influenced by the individual’s expectations regarding the food, which emerge from internal and external cues that make the food salient (Ordabayeva & Srinivasan, 2019). External cues refer to external aspects, such as packaging and commercials (Wansink, 1996). Internal cues consist of emotional (e.g., mood), normative and sensory cues (e.g., hunger, taste) (Gould, 1997; Wansink & Chandon, 2014; Wansink & Park, 2001). One method of decreasing the consumption of unhealthy foods, and thus decreasing total calorie consumption, involves changing behavioral cues and increasing the convenience, attractiveness and normative quality of healthy alternatives (Hanks et al., 2013). In addition, factors such as culture, demography, disposable income, geography, marketing, religion and urbanization influence consumption quantity through food availability and accessibility, as well as food choice (Kearney, 2010). An example of food availability’s influence includes increasing consumption quantity by proximity and visibility. Wansink, Painter and Lee (2006) investigated the candy consumption of secretaries over a period of four weeks. Their results indicated that consumption volume increased when the candy was visibly placed (an additional 2.2 candies daily) and proximately placed on the desk compared to two meters away (an additional 1.8 candies daily). Furthermore, they found people to underestimate their consumption of proximately placed candies, indicating mindless overeating.

2.2. Nudge Interventions

(11)

5 Based on the classification of mind, literature further differentiates between three types of nudge interventions, namely affective, behavioral and cognitive interventions. Affectively oriented nudges influence the consumers’ feelings, while behaviorally oriented nudges aim to change consumer behavior, and cognitively oriented nudges focus on influencing the consumers’ knowledge without altering feelings or behavior (Cadario & Chandon, 2019). This paper studies the effectiveness of one behaviorally and one cognitively oriented nudge, specifically, bowl size reduction and warning labels.

2.2.1. Bowl Size Reductions

Bowl size reductions belong to the category of behaviorally oriented nudges, which seek to change consumer behavior without altering their feelings or mind (Cadario & Chandon, 2019). Thus, bowl size reductions may influence people’s consumption behavior, in other words the quantity they consume. One study found participants to consume 56% more high-energy-dense snacks when they served themselves from a large bowl compared to an otherwise identical small bowl (Wansink & Cheney, 2005). Accordingly, consumption quantity was demonstrated to increase by 27% when people use 12.5-inch plates instead of 9.5-inch plates (Wansink, 2015). A bowl size reduction refers to the opposite effect (i.e., reducing the size of a bowl), so that a smaller bowl is used instead of a larger one. Therefore, we expect bowl size reductions to decrease consumption quantity.

(12)

6 Vandenbroele, van Kerckhove and Zlatevska (2019) suggested that perceived quantity plays an important role in actual food consumption. Several studies have further revealed that people consume a larger quantity of food when they perceive portions to be small, while they consumed less when they believed portions to be large. For instance, this effect occurs when larger packages are used or food is served on larger tableware (Wansink, 1996). In accordance to that, individuals have been found to consume less when using smaller dinnerware (van Ittersum & Wansink, 2012). Research among pre-school and school-age children has further demonstrated that children consumed up to twice as much breakfast cereal when eating from a 16oz bowl compared to an 8oz bowl (Wansink, van Ittersum, & Payne, 2014). Hence, it can be concluded that larger tableware sizes lead to an increase in consumption quantity. On the contrary, this suggests that a reduction in tableware size results in a decrease of consumption quantity. As such, it can be assumed that bowl size influences the volume perception of plated food. Based on the findings discussed above, we suspect that smaller bowls lead to perceiving a larger amount of food, thus reducing consumption quantity.

H1: The reduction of bowl size negatively influences consumption quantity.

2.2.2. Warning Label

Warning labels belong to the cognitively oriented eating nudges, which influence the consumer’s knowledge in various ways, such as with visibility enhancements and front-of-package nutrition labeling (Cadario & Chandon, 2019).

Front-of-package (FOP) nutrition labels, as well as labels on shelfs and menus, provide the consumer with calorie and/or nutrition information in numerous styles. This information is typically displayed on the food packaging, whether near the food in grocery stores and self-service restaurants or on menus and menu boards in restaurants (Cadario & Chandon, 2019). The goal of FOP labeling is to “increase the proportion of consumers who readily notice,

(13)

7 understand, and use the available information to make more nutritious choices for themselves and their families, and thereby prevent and reduce obesity and other diet-related chronic disease” (Food and Drug Administration, 2010, p. 22603). Academic research has investigated various types of FOP labels and their effectiveness. Considering only the most common ones, the Institute of Medicine identified 20 different types (Institute of Medicine, 2010).

(14)

8 in raising awareness of product’s unhealthiness. This is supported by the results of a meta-analysis conducted by Ikonen et al. (2019), which found food to be perceived as less healthy when a warning label was displayed. Hence, we suspect that warning labels also influence the quantity consumed when the product is already available to the consumer.

In 2016, Chile implemented “simple black-&-white stop sign warning labels” indicating an excessive amount of calories, saturated fat, sugar and sodium (Reyes et al., 2019) (see Figure 2 for a visualization of the Chilean “high in calories” warning label). A qualitative study among mothers revealed that the Chilean label was known and understood by a diversity of people after the first year of implementation (Correa et al., 2019). Furthermore, the findings of the same study identified changes in purchasing behavior, especially for individuals with a higher social economic status, as well as an overall shift toward healthier eating (Correa et al., 2019). Therefore, we expect a warning label to not only benefit the healthiness of food choice, but also reduce consumption quantity. Similarly to this prediction, research investigating the effect of low- and high-fat nutrition labels has found high-fat warning labels to cause weight-concerned participants of lower-socioeconomic status to consume less popcorn compared to participants in the control condition (Crockett, Jebb, Hankins, & Marteau, 2014). Research conducted by Wansink and Chandon (2006) revealed that low-fat nutrition labels led consumers to eat up to 50% more in a single consumption occasion, because people believed the consumed snacks were healthier and felt less guilty about the consumption. As warnings labels pursue the opposite effect of low-fat nutrition labels (i.e., to make the unhealthiness of a food product salient), we propose that a warning label’s presence reduces consumption quantity.

H2: The presence of a warning label negatively influences consumption quantity.

(15)

9

2.2.3. Combining Bowl Size Reduction and Warning Label

As previously elaborated, there is reason to expect bowl size reduction and warning label’s presence to decrease consumption quantity. The question thus arises as to whether combining bowl size reduction with a warning label may increase this influence on consumption quantity. While no literature has investigated this interaction effect, other research has successfully demonstrated that combining different nudges leads to increased healthy food choices. For instance, a recent study in worksite cafeterias combined various strategies based on nudging and social marketing, such as product, price, placement and promotion, resulting in an increase in healthy food sales over a 12-week period (Velema et al., 2018). Moreover, another experiment in a high-school cafeteria found combining different changes in food convenience, attractiveness and normativeness to increase healthier food choices (Hanks et al., 2013). Warning labels raise awareness to a product’s unhealthy nutritional information (Roberto et al., 2012). By emphasizing a product’s unhealthiness warning labels also bias calorie perceptions (Ordabayeva & Chandon, 2016), in other words a healthy product (i.e., one without warning label) is perceived as less caloric than an unhealthy product (i.e., one with a warning label). Low-fat labels were demonstrated to increase consumption quantity, because consumers expect the food to have less calories, and thus consider a larger serving more appropriate (Wansink & Chandon, 2006). In turn, this suggests that warning labels influence consumer to eat less, since they would expect a food with a warning label to contain more calories, and hence a smaller serving would be considered appropriate.

We predict that bowl size reduction enhances this effect, because bowl size influences the volume perception of the presented food. Bowl size reduction combined with a warning label influences the consumer’s perception twofold. Due to a warning label the consumer perceives the food in question to be unhealthy and more caloric, when in addition, the food is presented in a smaller bowl it appears to be a considerably larger quantity of food than when it is presented in a larger bowl. Thus, when the consumer considers the food to be unhealthy, he or she consumes less food, but even more so when he or she perceives a large quantity of food to be presented.

Thus, we propose that the interaction effect can be extended to the combination of bowl size reduction and warning label, as well as to quantity consumed.

H3: The negative effect of bowl size reduction on consumption quantity is more pronounced in

(16)

10

2.3. Health Consciousness

Various theoretical concepts have identified an individual’s motivation in terms of improving health and preventing disease to constitute a main source of healthy behavior (Newsom, McFarland, Kaplan, Huguet, & Zani, 2005). As such, we suspect this motivation to play a moderating role in influencing the previously described interaction effect. Mai and Hoffmann (2012, p. 318) defined health consciousness “as the motivational component that stimulates consumers to undertake health actions.” Awareness and motivation to become involved in healthy behaviors and growing self-consciousness about maintaining and/or improving one’s health distinguishes highly health-conscious individuals from low health-conscious individuals (Kraft & Goodell, 1993; Newsom et al., 2005). Thus, consumers high in health consciousness care about their health, and so participate in healthy behaviors (e.g., making healthy food choices) to enhance or sustain their well-being (Mai & Hoffmann, 2012). High levels of health consciousness have been demonstrated to increase purchase intentions of healthy products (Lockie, Lyons, Lawrence, & Mummery, 2002). Consequently, higher levels of health consciousness are believed to increase healthy food choices, and thus to likely also decrease overconsumption.

(17)

11 foods (Gould, 1990), and so it can be assumed that high health-conscious individuals will, if avoidance is not possible, refrain from eating large amounts of calorie-dense foods.

To conclude, highly health-conscious individuals are believed to avoid or only consume limited quantities of foods that oppose their goal of enhancing their well-being. Products provided with a warning label (e.g., “high in calories”) can be considered a food that does not promote well-being, and thus would not be consumed at all or only in limited amounts. This is supported by the findings of Glanz, Basil, Maibach, Goldberg and Snyder (1998), who determined that foods beneficial for one’s nutrition or health are important to highly health-conscious individuals. In addition, highly health-conscious individuals were found to educate themselves more about nutrition and read labels in greater detail than individuals low in health consciousness (Gould, 1990). Hence, information presented on a warning label affects individuals with higher levels of health consciousness more compared to those with lower levels.

(18)

12

H4: The moderation effect of warning label on bowl size reduction is more pronounced for

highly health-conscious individuals.

2.4. Development of the Conceptual Framework

Following the definitions of the main concepts, this section summarizes the relationships and hypotheses in a conceptual framework.

The conceptual framework visualized in Figure 3 provides a foundation for the following research. As previously described, it identifies two main effects of bowl size reduction and warning label on consumption quantity, respectively. Furthermore, it depicts warning labels’ moderating roles on how bowl size reduction affects consumption quantity. Additionally, it demonstrates health consciousness moderating the interaction of bowl size reduction and warning label regarding consumption quantity.

Figure 3: Conceptual model

3. Methodology

(19)

13

3.1. Participants

A total of 212 students were recruited from the respondent pool of a large European University (University of Groningen) to participate in this lab study between November 25th_{and December}

6th_{, 2019. In return for their participation, participants received partial course credit or monetary}

compensation (8€ per hour). All participants who tasted the food and filled in the questionnaire completely were included in this study. From the initial sample of 212 participants, 15 participants had to be excluded because their responses were incomplete (n=5) or they did not sample the presented food (n=10). The final sample consists of 108 female (54.8%) and 89 male (45.2%) participants aged 17–35 years (Mage= 20.75 years, SDage= 2.97). The majority of

participants were born in the Netherlands (50.8%), Germany (7.1%) or Indonesia (5.6%). Most participants (69%) held a high school degree as their highest educational qualification. Detailed frequencies for the participants’ highest educational level are illustrated in Figure 4.

The average BMI of women equaled 21.85 (SD = 3.32), while the men’s average BMI equaled 22.65 (SD = 2.78). According to the thresholds defined by the World Health Organization (2019), participants were on average of normal weight. Three participants did not provide sufficient information to calculate their BMI, and so the BMI measurement was calculated based on a sample of 194 participants. Additionally, most participants (97%) were physically active at least once a week (e.g., playing sports, going to the gym, biking), while 40.6% of participants engaged in physical activity two to three times a week. Meanwhile, 27.6% of the sample conducted physical activity four to six times a week, and 18.8% of participants were physically active once a week.

69.0% 2.0% 3.6% 22.3% 3.0% 0 20 40 60 80 100 120 140 160 High school degree Secondary vocational degree Bachelor degree from University of Applied Sciences Bachelor degree from University Master degree F re que nc y

Highest educational qualification

(20)

14

3.2. Research Design

For this study, a 2 (bowl size reduction: absent vs. present) x 2 (warning label: absent vs. present) between-subject design was employed. Therefore, participants were randomly assigned to one of four conditions, namely (1) a control group without intervention and three groups with the interventions (2) warning label, (3) bowl size reduction and (4) warning label and bowl size reduction combined (see Table 1).

Table 1: Study design

Warning label

Absent Present

Bowl size reduction

Absent

(1) No label, standard bowl size (2) Warning label, standard bowl size

Present

(21)

15 Warning label

The warning label utilized for the associated conditions was designed in a manner replicating the existing Chilean “simple black-&-white stop sign warning label” (Reyes et al., 2019, p. 1). As depicted in Figure 5, this includes the text phrase “high in calories” written in white color and with capital letters in front of a black background. Based on the Chilean model, the warning label manipulation is displayed in an octagonal shape, resembling a stop sign. For the conditions with a warning label, this label was displayed on a food package (see Appendix: Figure A II).

As only one food product was used across manipulations, the goal was to find a product that could realistically be used for the “high in calories” warning label conditions and was also relevant in terms of overconsumption. In general, the category of snacks is highly relevant as snacking has increased in recent years, comprising almost 20% of the daily calorie intake (Nielsen, Siega-Riz, & Popkin, 2002). In specific, the category of “salty snacks” offers a relevant category to investigate concerning overconsumption as consumption of this category has doubled since 1977 (Dunford & Popkin, 2018).

According to Hui, Bradlow and Fader (2009), salty snacks are categorized as an unhealthy product among cookies and carbonated beverages. Since perceiving a product (category) as dietary forbidden or acceptable has been found to influence consumption behavior (Knight & Boland, 1989), the chosen product should be seen as more ambiguous in terms of perceived healthiness. Thus, for this study, pretzels were utilized as a food product to be consumed by participants. This is due to the fact that pretzels are considered ambiguous in terms of health and are, in contrast to their general category, not perceived as healthy or unhealthy by most people (Finkelstein & Fishbach, 2010). Additionally, pretzels were found to be consumed “mainly to satisfy their [participants’] appetite rather than to improve their health (e.g., by consuming vegetables) or to obtain hedonic pleasure (e.g., by consuming chocolate)” (Finkelstein & Fishbach, 2010, p. 361). Nevertheless, pretzels contain numerous calories, specifically 399 kcal per 100 gram (The Lorenz Bahlsen Snack-World GmbH & Co KG, 2019).

(22)

16 As such, according to the thresholds defined by the Chilean policy, they would need to be labeled with a “high in calories” warning label from the first stage onwards which started with a threshold of 350 kcal per 100 gram (Correa et al., 2019).

In particular, “Saltletts Brezel” from the German manufacturer “The Lorenz Bahlsen Snack-World GmbH & Co KG” were chosen as a product to be displayed. This was due to the fact that neither this product (pretzels in a single package) nor the specific brand (Saltletts) are available on the Dutch market. Therefore, participants could realistically be misled into believing that the study concerned their perception of a new salty snack. Consequently, the participants were completely unaware of the study’s true purpose (i.e., their food consumption being measured) to avoid influences on the results.

Bowl size reduction

In the conditions featuring bowl size reduction, the pretzels were presented to participants in a smaller bowl compared to the conditions lacking bowl size reduction. Following the method of Choi et al. (2019) and Wansink et al. (2014), a difference in bowl diameter of 3 centimeter (cm) was chosen. Thus, the standard bowl size possessed a diameter of 15 cm, while a bowl diameter of 12 cm was selected for the smaller-sized bowl (see Appendix Figure A III for an illustration of the two bowl sizes). Bowls with a diameter of 12 and 15 cm are typically used in people’s homes (Choi et al., 2019; Inter IKEA Systems B.V, 2019), and thus represent a realistic manipulation. Throughout the experiment, it was ensured that each participant received the exact same amount of pretzels, namely 80 grams, which exceed this product’s suggested serving size of 30 grams by 166%. This portion size was chosen in order to ensure that most participants would not be able to finish their pretzels during their participation in the lab study.

3.3. Procedures

(23)

17 In a Qualtrics survey (see Appendix B.1 for a summary of all items), participants were asked to rate how hungry they were at the present moment on a seven-point Likert scale, with anchor points of 1 = “not at all hungry” and 7 = “very hungry” (Finkelstein & Fishbach, 2010). In addition, we questioned participants about the time that had passed since their last food consumption and starting the survey (Vandenbroele et al., 2019).

Afterwards, participants were randomly assigned to one of the two warning label conditions (absent vs. present) via a computer-generated algorithm. They were presented with instructions that they were about to see the package of a salty snack called “Saltletts Brezel” from the German brand “Lorenz,” and that this is currently not available in the Netherlands, but might be introduced to the Dutch market soon (see Appendix B.1, Table A I: Overview of items and scales used in the survey). Then, depending on the participant’s condition, he or she was presented the package with or without the warning label (see Appendix A Manipulation for an illustration of both packages: Figure A I and Figure A II). Participants were also informed that they would taste this food later, which increases involvement similar to the involvement manipulation employed by Petty, Cacioppo and Schumann (1983). We also measured the time participants spent looking at the package.

Following this, participants were randomly assigned to a condition of bowl size reduction (absent vs. present) based on their cubicle (cubicles 1, 2, 5 and 6: absent; cubicles 3, 4, 7 and 8: present). All participants were provided with 80 grams of pretzels in a small vs. a standard-sized bowl depending on the condition. Then, they were told that they would be presented a movie trailer next, because the study was highly interested in modeling a realistic consumption situation (see Appendix B.3 Introduction to Consumption Part and Video Presentation). To provide participants with sufficient time to consume the pretzels, a five-minute, 48-second-long movie trailer of “Sonic the Hedgehog” (FilmSpot Trailer, 2019) was presented to participants (based on a similar approach utilized by Vandenbroele et al. (2019)). Participants were unaware of this study’s intention or the fact that their pretzel consumption would be measured. To ensure that participants watched the full video, the time they spent on the page of the video was measured.

(24)

18 cover story, participants were asked about their intention to purchase the pretzels on a three-item scale, which was similarly employed by Howlett, Burton and Kozup (2008), Keller et al. (1997) and Kozup et al. (2003). This comprised items such as, “Given the package shown previously, how probable is it that you would consider the purchase of the pretzels, if you were interested in buying a snack” (1 = “not probable”, 7 = “very probable”).

Before continuing with the survey participants were asked to return the pretzels (see Appendix B.4 Ending Consumption Part). Thus, it was ensured that the consumption time was equal between conditions (i.e., approximately eight minutes). Following a finished lab session, a research assistant collected the box with the leftover pretzels and weighed them in order to calculate the dependent variable (i.e., consumption quantity). As the portion size was held constant during this study, the dependent variable was represented by absolute consumption, which was calculated based on the initial portion size of 80 grams subtracted by the amount of leftover pretzels in grams.

The Qualtrics survey continued by asking participants to rate how frequently they usually consume salty snacks (1 = “never”, 7 = “very often”). Then, participants were asked to rate the pretzels’ perceived healthfulness on a four-item scale adapted from the scale utilized by Kozup et al. (2003), featuring items such as, “Based on the package, how important would the product you just tasted be as part of a healthy diet” which were rated on a seven-point Likert scale with anchor points of 1 = “not important at all” and 7 = “very important”.

(25)

19 In order to assess changes in participants’ hunger levels by snacking on the pretzels, the item of Finkelstein and Fishbach (2010) was repeated. This was followed by a question regarding participant’s dietary restrictions (e.g., vegetarian, vegan or religious restrictions) or food-related allergies. Additionally, we were interested in whether participants were dieting in order to control their weight at the time of the study (Lowe et al., 2006). To check whether participants noticed the manipulation and to ensure construct validity, they were asked, “Did you notice a ‘high in calories’ warning label on the package shown earlier,” for which they were presented three answer possibilities: “yes”, “no” and “I don’t remember.”

Afterwards, participants reported demographic information concerning their gender, age, education, height and weight, physical activity and country of birth. By means of self-reported height and weight, the participants’ BMI was calculated and categorized into “underweight,” “normal weight,” “pre-obesity” and “obese” using the WHO thresholds (World Health Organization, 2019).

Finally, participants were asked “What do you think we were expecting to find in this study” (Best & Papies, 2019) to check their awareness of the study’s hypotheses. At the end of the survey, participants were thanked for their time and fully debriefed regarding the study’s purpose through an on-screen debriefing.

4. Results

This section presents the lab study’s results. First, the data was inspected for missing values, oddities and outliers with consideration of Mahalanobis’ distance, Cook’s distance and the centered leverage value. Furthermore, a thorough analysis of participants’ guesses regarding the studies’ goals revealed that no participant was completely aware of the study’s purpose, specifically that the food consumption was measured. Next, reverse-coded items were recoded so that the higher values indicated a more positive response.

4.1. Reliability and Factor Analyses

(26)

20 “I notice how I feel physically as I go through the day,” was removed from the health consciousness scale. Based on the results of the conducted analyses, all other items are adequate for computing unweighted average factors for attitude, health consciousness, nutrition information, perceived healthfulness and purchase intention (see Table 2).

Cronbach’s Alpha KMO measure Bartlett’s test of sphericity (Significance) Communalities Factor loadings Attitude 0.97 0.78 χ2(3) = 802.08, p < 0.01 > 0.94 > 0.97 Health consciousness 0.90 0.89 χ2(28) = 918.27, p < 0.01 > 0.48 > 0.69 Nutrition information 0.81 0.50 χ2(1) = 127.20, p < 0.01 > 0.85 > 0.92 Perceived healthfulness 0.80 0.77 χ2(6) = 269.65, p < 0.01 > 0.49 > 0.70 Purchase intention 0.89 0.75 χ2(3) = 339.71, p < 0.01 > 0.81 > 0.90 Table 2: Results of factor and reliability analyses

4.2. Descriptive Results

On average, participants consumed 30.45% of the provided 80 grams of pretzels (M = 24.36,

SD = 17.14, range = 1-80 grams) (see Figure 6 for a bar chart with frequencies). This is slightly less than the portion size recommended by the manufacturer and features roughly 100 calories (The Lorenz Bahlsen Snack-World GmbH & Co KG, 2019). Three participants (1.5%) finished the complete portion of 80 grams.

0 1 2 3 4 5 6 7 8 9 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 F re que nc y Consumption quantity

(27)

21 The means of the variables of interest are presented in Table 3. A comparison of health consciousness and healthiness of past-eating behavior indicates that the participants’ intention and past behavior are mostly similar. Furthermore, the pretzels were, on average, perceived to be somewhat tasty (M = 5.14), and participants possessed a positive attitude toward them on average (M = 5.41). As expected, the pretzels were, on average, considered ambiguous regarding healthiness, though slightly unhealthy (M = 2.81).

Mean Standard

deviation Minimum Maximum Consumption quantity (grams) 24.36 17.14 1 80 Health consciousness 4.87 0.99 2.13 7 Healthiness of past eating behavior 4.47 1.38 1 7 Hunger (pre consumption) 3.72 1.64 1 7 Hunger (post consumption) 3.11 1.55 1 7 Taste 5.14 1.42 1 7 Attitude 5.41 1.40 1 7 Purchase Intention 3.84 1.42 1 7 Perceived Healthfulness 2.81 1.03 1 5.75 Nutrition information 4.41 1.46 1 7

Table 3: Descriptive statistics for consumption quantity, health consciousness, healthiness of past eating behavior, hunger, taste, attitude, purchase intention, perceived healthfulness and nutrition information

4.3. Testing the Role of Bowl Size Reduction and Warning Label through

ANCOVA

(28)

22 (< 3 box-lengths from the edge of the box), and there was no evidence that they resulted from an error in recording, a miscalculation, a malfunction of measuring equipment, or a similar reason. Additional analyses revealed that the results remained unchanged when excluding the detected outliers. Second, the data was not normally distributed, as demonstrated by the Shapiro-Wilk and the Kolmogorov-Smirnov test of normality (p < 0.01). Additionally, normal Q-Q plots were visually inspected to confirm this finding. Despite the assumption of normally distributed data for ANCOVA being violated, we continued with the ANCOVA due to its robustness to deviations from normality (Maxwell & Delaney, 2004). Third, homogeneity of variance was reviewed by Levene’s test for equality of variances (p = 0.30).

The two-way interaction between bowl size reduction and warning label was not significant whilst controlling for hunger and taste perception, F(1,191) = 1.86, p = 0.17, partial η2 = 0.01. Simple main effect analysis revealed that bowl size reduction did not influence consumption quantity, whether for the warning label absent, F(1,191) = 1.73, p = 0.19, partial η2 < 0.01, or present, F(1,191) = 0.39, p = 0.53, partial η2 < 0.01. Furthermore, the warning label’s effect on consumption quantity was analyzed by simple main effect analysis. This disclosed a significant effect of warning label on consumption quantity in the condition bowl size reduction absent, F(1,191) = 4.80, p = 0.03, partial η2 = 0.03. However, warning label did not influence consumption quantity when bowl size reduction was present, F(1,191) = 0.06, p = 0.81, partial

η2 < 0.01. The simple main effect of warning label was further analyzed by pairwise

(29)

23 Due to the insignificant two-way interaction, the main effects for bowl size reduction and warning label were analyzed. This analysis revealed that bowl size reduction did not influence consumption quantity, after controlling for hunger and taste perception, F(1,191) = 0.22, p = 0.64, partial η2 < 0.01. Hence, in contrast to prior findings, this study’s results did not support the prediction that bowl size reduction decreases consumption quantity. Therefore, hypotheses one needs to be rejected.

There was a marginally statistically significant effect of warning label on consumption quantity while controlling for hunger and taste perception, F(1,191) = 2.94, p = 0.09, partial η2 = 0.02. To further analyze warning label’s main effect on consumption quantity, pairwise comparisons with the Bonferroni adjustment were conducted (for a graphical representation, see Figure 8). When a warning label was presented to participants, they consumed 3.83 grams less pretzels (M =22.54, SE =1.50, 95% CI [19.57, 25.50]) compared to when they did not see a warning label on the package (M = 26.15, SE = 1.48, 95% CI [23.23, 29.07]). This suggests that the presence of a warning label reduces consumption quantity. Thus, these results confirm hypothesis two, which predicted that the presence of a warning label negatively influences consumption quantity. 0 5 10 15 20 25 30

Warning label absent Warning label present

Co n su m p ti o n q u an ti ty in g ra m s

Bowl size reduction absent (p = 0.03) Bowl size reduction present (p = 0.81)

Figure 7: Consumption quantity in grams as a function of bowl size reduction and warning label, ∗ p < 0.05.

(30)

24 Furthermore, the ANCOVA results demonstrated that hunger, F(1,191) = 14.26, p < 0.01,

partial η2 = 0.07, and taste perception, F(1,191) = 41.14, p < 0.01, partial η2 = 0.18,

significantly affect consumption quantity. Consequently, the amount consumed by participants was influenced by their hunger and their taste perception of the pretzels.

4.4. Testing the Role of Bowl Size Reduction and Warning Label through

Regression

Next, regression analysis using Hayes’ Process macro (2017) for SPSS v3.4 (model 1) was performed to verify the results of the conducted ANCOVA. Thus, consumption quantity was used as the dependent variable and bowl size reduction as the independent, while warning label was used as moderator. Additionally, the model controlled for hunger and taste perception. The overall model achieved a high goodness-of-fit, R2_{= 0.27 (Cohen, 1988), and bowl size}

reduction and warning label were able to predict consumption quantity to a statistically significant degree, while controlling for hunger and taste perception, F(5,191) = 14.43, p < 0.01. The regression results confirm the insignificant interaction effect found in the performed ANCOVA (ß = 1.45, SEß = 1.06, t = 1.37, p = 0.17, 95% CI [-0.65, 3.55]). The regression

analysis supports the ANCOVA findings that the main effect of bowl size reduction is not significant (ß = -0.50, SEß = 1.06, t = -0.47, p = 0.64, 95% CI [-2.59, 1.60]). Moreover, the

regression replicated warning label’s significant effect on consumption quantity (ß = -1.81, SEß

= 1.06, t = -1.71, p = 0.88, 95% CI [-3.89, 0.27]). Thus, participants consumed less when a warning label was presented. Consequently, hypothesis one and three have to be rejected and

0 5 10 15 20 25 30 Co n su m p ti o n q u an ti ty in g ra m s

Warning label absent Warning label present

(31)

25 hypothesis two was confirmed by the analyzed data. The point estimates from this moderation model are depicted in Figure 9. In addition, we found hunger (ß = 2.47, SEß = 0.65, t = 3.78, p

< 0.01, 95% CI [1.18, 3.76]) and taste perception (ß = 4.88, SEß = 0.76, t = 6.41, p < 0.01, 95%

CI [3.37, 6.37]) to predict consumption quantity, and thus it is important to control for these.

Figure 9: Moderation model (1) depicting the effect of bowl size reduction and warning label Coefficients are unstandardized regression coefficients from the PROCESS model. † p < 0.10.

In order to check for multicollinearity, variance inflation factor (VIF) scores were calculated. All calculated VIF scores equal one, suggesting that multicollinearity is not an issue.

4.5. Testing the Role of Health Consciousness

In the following, model three of Hayes’ Process macro (2017) was employed to test the moderated moderation effect predicted in hypothesis four. Therefore, consumption quantity serves as the dependent variable and bowl size reduction as the independent, whereas warning label (W) and health consciousness (Z) are both moderators. In addition, the model controls for hunger and taste perceptions. The overall model is significant, R2_{= 0.29, F(9,187) = 8.31, p <}

0.01.

The regression revealed that the two-way interaction of bowl size reduction and warning label was not significant (ß = -3.53, SEß = 5.66, t = -0.63, p = 0.53, 95% CI [-14.69, 7.63]).

Additionally, the main effects of bowl size reduction (ß = -7.83, SEß = 5.67, t = -1.38, p = 0.17,

95% CI [-19.02, 3.35]) and warning label (ß = -0.96, SEß= 5.63, t = -0.17, p = 0.87, 95% CI

(32)

26 The three-way interaction of bowl size reduction, warning label and health consciousness was insignificant (ß = 0.96, SEß= 1.13, t = 0.84, p = 0.40, 95% CI [-1.28, 3.19]), suggesting that

health consciousness does not moderate the moderation effect of warning label. Thus, hypothesis four regarding the moderated moderation has to be rejected. Furthermore, the interaction term between warning label and health consciousness was not significant (ß = -0.11,

SEß= 1.13, t = -0.98, p = 0.92, 95% CI [-2.34, 2.12]). Hence, health consciousness does not

moderate the main effect of warning label regarding consumption quantity. In addition, the regression demonstrated that health consciousness does not moderate bowl size reduction’s main effect on consumption quantity (ß = 1.44, SEß= 1.14, t = 1.27, p = 0.21, 95% CI [0.80,

3.68]). Moreover, the main effect of health consciousness regarding consumption quantity was not significant (ß = -1.17, SEß= 1.14, t = -1.03, p = 0.31, 95% CI [-3.40, 1.07]). As such, health

consciousness does not influence consumption quantity.

Replicating the results of the previous analyses, participants’ hunger (ß = 2.46, SEß= 0.67, t =

3.68, p < 0.01, 95% CI [1.14, 3.79]) and taste perceptions (ß = 4.95, SEß= 0.77, t = 6.48, p <

0.01, 95% CI [3.44, 6.463]) predict consumption quantity.

Multicollinearity was not a concern in this analysis, as all VIF scores remained below the cut-off value of four.

4.6. Robustness Tests

(33)

27 as covariates), model G achieved a lower goodness-of-fit (R2_{= 0.03) and was only marginally}

significant (p = 0.08) rather than significant (model A). This means that model G is less suitable for explaining how bowl size reduction, warning label and their interplay influence consumption quantity. This is due to the fact that the covariates (hunger and taste perception) explain a considerable amount of variance. This regression (model G) demonstrated results similar to the previously conducted regression. However, contrary to previous regressions, the interaction effect between bowl size reduction and warning label was significant in this model (ß = -2.48, SEß= 1.21, t = 2.05, p = 0.04, 95% CI [0.09, 4.86]). In contrast to the previous

models, warning label was not significant (ß = -1.95, SEß= 1.21, t = -1.62, p = 0.11, 95% CI

[-4.34, 0.43]). In line with our previous findings, bowl size reduction did not significantly influence consumption quantity (ß = 0.16, SEß= 1.21, t = 0.13, p = 0.90, 95% CI [-2.23, 2.54]).

The varying results can be explained by the fact that the covariates (hunger and taste perception) significantly influence consumption quantity and therefore skew the results.

5. Discussion

This section summarizes and discusses the findings of this investigation into consumption quantity and elucidates whether the hypotheses are supported. Furthermore, contributions and implications of this study’s findings are highlighted. Additionally, limitations and recommendations for future research are presented. This research aimed at providing insights into bowl size reduction, warning label, their interaction and health consciousness as a moderator. The main research question to be answered was as follows: “How does bowl size

reduction and the presence of a warning label influence consumption quantity?”

(34)

28 and it seems that bowl size reduction suppressed the warning label’s effect on consumption quantity. Thus, we reject hypothesis three. As no significant main effect of bowl size reduction on quantity consumed was found, hypothesis one also has to be rejected. This resembles findings of Ayaz, Akyol, Cetin, and Besler (2016) and Robinson et al. (2016) that tableware size did not influence consumption quantity. In line with our expectation, we found statistical evidence of a main effect of warning label on consumption quantity. Therefore, hypothesis two is accepted. In contrast to our expectation, health consciousness was not a moderator of the interaction effect of bowl size reduction and warning label. Hence, hypothesis four was rejected. An overview over the acceptance and rejection of specific hypotheses is provided in Table 4.

Hypotheses Accept/Reject H1: The reduction of bowl size negatively influences consumption

quantity. Reject

H2: The presence of a warning label negatively influences consumption

quantity. Accept

H3: The negative effect of bowl size reduction on consumption quantity is

more pronounced in the presence of a warning label. Reject

H4: The moderation effect of warning label on bowl size reduction is more

pronounced for highly health-conscious individuals. Reject

Table 4: Overview of hypotheses

5.1. Contributions and Implications

(35)

29 Additionally, this would lead people to consume less of those products with a warning label (if they are purchased despite the label) as demonstrated by this study.

Second, this study adds to the research by revealing no significant interaction effect between bowl size reduction and warning label. This provides room for future research combining nudges, though further analysis is required to verify this finding and to investigate whether the combination of a warning label with other nudges may be effective (e.g., combining warning label with package size). Since simple main effect analysis revealed that warning labels combined with a standard-sized bowl (no bowl size reduction present) cause people to consume less, retailers and manufacturers should be aware of this effect when offering products with a warning label. They might wish to provide consumers with normative cues regarding the bowl size to use when consuming their product, which would also aid in avoiding the opposing effect on consumption quantity revealed for the combination of smaller bowls with warning labels. For instance, they could place a message such as “These pretzels are best consumed from a bowl with a diameter of 15 cm” on their package or shelves.

(36)

30 to investigate reduced package sizes, as many food items, such as snacks, are consumed directly from the package and not placed in a bowl before consumption. Additionally, many food manufacturers have begun offering smaller packages in order to serve health-conscious consumers.

Fourth, contrary to expectations, this study’s findings did not support a moderating role of health consciousness. While this requires further investigation, it may imply that health consciousness only influences nutrition interest and food choice (Glanz et al., 1998; Gould, 1990; Howlett et al., 2009; Visschers et al., 2013), but is not predictive of consumption behavior, and thus does not play a role in consumption quantity.

5.2. Limitations and Directions for Future Research

This research faced several limitations that could, in turn, provide directions for future research. First, the data was collected in lab sessions scheduled throughout the day, which may have influenced participants’ consumption behavior. Certainly, what a participant ate before entering the lab influenced how much (s)he ate, but time of day could also influence consumption behavior by providing the participant with normative cues regarding how appropriate eating a snack during that time is. For instance, participants might not perceive the morning hours as appropriate for eating a salty pretzel, and thus consume less compared to what they would potentially consume during the evening hours. As such, future research should focus on specific consumption occasions, such as an afternoon or evening snack, so as to avoid time-of-day differences in consumption. Time of day was used as a covariate to control for this effect, but due to analytical limitations we only differentiated between before- and after-noon, which might not be specific enough to capture this effect.

(37)

31 Third, this study solely investigated the effect of pre-portioned food consumption. On the one hand, individuals mostly serve themselves on a day-to-day basis, especially with snacks, which are also often eaten directly out of the package rather than taking the time to portion the snack into a bowl. On the other hand, the process of portioning the food by oneself could function as a filter to reduce consumption quantity in the sense that the individual first needs to portion a certain amount that (s)he proposes to be able to eat, after which the individual may opt to leave food in the bowl. Even though previous research has suggested that most people only feel satisfied when finishing their plate, this may not hold when combined with a warning label. Additionally, opposed to previous research, this study found that bowl size reductions did not significantly influence food consumption, which raises further questions, such as whether the small bowl led to the misperception that the amount of food presented was smaller, or whether individuals were misled by the fullness of the bowl (compared to the standard-sized bowl). Moreover, this raises the questions whether the perfect bowl size exists. Future research can fruitfully investigate these questions.

6. Conclusion

(38)

32

References

Acton, R. B., & Hammond, D. (2018). The impact of price and nutrition labelling on sugary drink purchases: Results from an experimental marketplace study. Appetite, 121, 129–137. https://doi.org/10.1016/j.appet.2017.11.089

Ayaz, A., Akyol, A., Cetin, C., & Besler, H. T. (2016). Effect of plate size on meal energy intake in normal weight women. Nutrition Research and Practice, 10(5), 524–529. https://doi.org/10.4162/nrp.2016.10.5.524

Berry, C., Burton, S., Howlett, E., & Newman, C. L. (2019). Understanding the calorie labeling paradox in chain restaurants: Why menu calorie labeling alone may not affect average calories ordered. Journal of Public Policy & Marketing, 38(2), 192–213. https://doi.org/10.1177/0743915619827013

Best, M., & Papies, E. K. (2019). Lower socioeconomic status is associated with higher intended consumption from oversized portions of unhealthy food. Appetite, 140, 255–268. https://doi.org/10.1016/j.appet.2019.05.009

Billich, N., Blake, M. R., Backholer, K., Cobcroft, M., Li, V., & Peeters, A. (2018). The effect of sugar-sweetened beverage front-of-pack labels on drink selection, health knowledge and awareness: An online randomised controlled trial. Appetite, 128, 233–241. https://doi.org/10.1016/j.appet.2018.05.149

Cadario, R., & Chandon, P. (2019). Which healthy eating nudges work best? A meta-analysis of field experiments. Marketing Science, 130, 1–22. https://doi.org/10.1287/mksc.2018.1128

Choi, J., Jessica Li, Y., & Samper, A. (2019). The influence of health motivation and calorie ending on preferences for indulgent foods. Journal of Consumer Research, 46, 606–619. https://doi.org/10.1093/jcr/ucz002

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates.

Correa, T., Fierro, C., Reyes, M., Dillman Carpentier, F. R., Taillie, L. S., & Corvalan, C. (2019). Responses to the Chilean law of food labeling and advertising: Exploring knowledge, perceptions and behaviors of mothers of young children. International Journal

(39)

33 https://doi.org/10.1186/s12966-019-0781-x

Crockett, R. A., Jebb, S. A., Hankins, M., & Marteau, T. M. (2014). The impact of nutritional labels and socioeconomic status on energy intake: An experimental field study. Appetite,

81, 12–19. https://doi.org/10.1016/j.appet.2014.05.024

Delboeuf, J. (1865). Note sur certaines illusions d’optique: Essai d’une the´orie psychophysique de la manie`re dont l’oeil appre´cie les distances et les angles [Note on certain optical illusions: Essay on a psychophysical theory concerning the way in which the eye evaluates distances and angles]. Bulletins de l’Acade´mie Royale Des Sciences, Lettres et

Beaux-Arts de Belgique, 19(2), 195–216.

Dunford, E. K., & Popkin, B. M. (2018). 37 year snacking trends for US children 1977–2014.

Pediatric Obesity, 13(4), 247–255. https://doi.org/10.1111/ijpo.12220

El Observatodo. (2019). Ley 20.606: “El etiquetado de mis alimentos sí importa.” Retrieved October 20, 2019, from http://www.elobservatodo.cl/noticia/sociedad/ley-20606-el-etiquetado-de-mis-alimentos-si-importa

Fernandes, A. C., Oliveira, R. C., Proenca, R. P. C., Curioni, C. C., Rodrigues, V. M., & Fiates, G. M. R. (2016). Influence of menu labeling on food choices in real-life settings: A systematic review. Nutrition Reviews, 74(8), 534–548. https://doi.org/10.1093/nutrit/nuw013

FilmSpot Trailer. (2019). Sonice the hedgehog: 5 minute trailers (4K ultra HD) new 2020. Retrieved November 20, 2019, from https://www.youtube.com/watch?v=GxlVSQyuhRU Finkelstein, S. R., & Fishbach, A. (2010). When healthy food makes you hungry. Journal of

Consumer Research, 37(3), 357–367. https://doi.org/10.1086/652248

Food and Drug Administration. (2010). Front-of-pack and shelf tag nutrition symbols (75 FR 22602, April 29, 2010). Federal Register, 75(82), 22602–22606.

Glanz, K., Basil, M., Maibach, E., Goldberg, J., & Snyder, D. (1998). Why Americans eat what they do: Taste, nutrition, cost, convenience, and weight control concerns as influences on food consumption. Journal of the American Dietetic Association, 98(10), 1118–1126. https://doi.org/10.1016/S0002-8223(98)00260-0

(40)

34 perspective. The Journal of Consumer Affairs, 22(1), 96–118. https://doi.org/10.1111/j.1745-6606.1988.tb00215.x

Gould, S. J. (1990). Health consciousness and health behavior: The application of a new health consciousness scale. American Journal of Preventive Medicine, 6(4), 228–237. https://doi.org/10.1016/S0749-3797(18)31009-2

Gould, S. J. (1997). An interpretive study of purposeful, mood self-regulating consumption: The consumption and mood framework. Psychology and Marketing, 14(4), 395–426. https://doi.org/10.1002/(SICI)1520-6793(199707)14:4<395::AID-MAR6>3.0.CO;2-4 Hanks, A. S., Just, D. R., & Wansink, B. (2013). Smarter lunchrooms can address new school

lunchroom guidelines and childhood obesity. Journal of Pediatrics, 162(4), 867–869. https://doi.org/10.1016/j.jpeds.2012.12.031

Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis,

second edition: A regression-based approach (2nd ed.). New York, NY: Guilford Publications.

Howlett, E. A., Burton, S., Bates, K., & Huggins, K. (2009). Coming to a restaurant near you? Potential consumer responses to nutrition information disclosure on menus. Journal of

Consumer Research, 36(3), 494–503. https://doi.org/10.1086/598799

Howlett, E., Burton, S., & Kozup, J. (2008). How modification of the nutrition facts panel influences consumers at risk for heart disease: The case of trans fat. Journal of Public

Policy and Marketing, 27(1), 83–97. https://doi.org/10.1509/jppm.27.1.83

Hui, S. K., Bradlow, E. T., & Fader, P. S. (2009). Testing behavioral hypotheses using an integrated model of grocery store shopping path and purchase behavior. Journal of

Consumer Research, 36(3), 478–493. https://doi.org/10.1086/599046

Ikonen, I., Sotgiu, F., Aydinli, A., & Verlegh, P. W. J. (2019). Consumer effects of front-of-package nutrition labeling: An interdisciplinary meta-analysis. Journal of the Academy of

Marketing Science. https://doi.org/10.1007/s11747-019-00663-9

Institute of Medicine. (2010). Examination of front-of-package nutrition rating systems and

symbols: Phase I report. Washington, DC: The National Academies Press.

(41)

35 Jaacks, L. M., Vandevijvere, S., Pan, A., McGowan, C. J., Wallace, C., Imamura, F., Mozaffarian, D., Swinburn, B., & Ezzati, M. (2019). The obesity transition: stages of the global epidemic. The Lancet Diabetes and Endocrinology, 7(3), 231–240. https://doi.org/10.1016/S2213-8587(19)30026-9

Kearney, J. (2010). Food consumption trends and drivers. Philosophical Transactions:

Biological Sciences, 365(1554), 2793–2807. https://doi.org/10.1098/rstb.2010.0149 Keller, S. B., Landry, M., Olson, J., Velliquette, A. M., Burton, S., & Andrews, J. C. (1997).

The effects of nutrition package claims, nutrition facts panels, and motivation to process nutrition information on consumer product evaluations. Journal of Public Policy and

Marketing, 16(2), 256–269. https://doi.org/10.1177/074391569701600206

Knight, L. J., & Boland, F. J. (1989). Restrained eating: An experimental disentanglement of the disinhibiting variables of perceived calories and food type. Journal of Abnormal

Psychology, 98(4), 412–420. https://doi.org/10.1037/0021-843x.98.4.412

Kozup, J. C., Creyer, E. H., & Burton, S. (2003). Making healthful food choices: The influence of health claims and nutrition information on consumers’ evaluations of packaged food products and restaurant menu items. Journal of Marketing, 67(2), 19–34. https://doi.org/10.1509/jmkg.67.2.19.18608

Kraft, F. B., & Goodell, P. W. (1993). Identifying the health conscious consumer. Journal of

Health Care Marketing, 13(3), 18–25.

Lockie, S., Lyons, K., Lawrence, G., & Mummery, K. (2002). Eating “green”: Motivations behind organic food consumption in Australia. Sociologia Ruralis, 42(1), 23–40. https://doi.org/10.1111/1467-9523.00200

Lowe, M. R., Annunziato, R. A., Markowitz, J. T., Didie, E., Bellace, D. L., Riddell, L., Maille, C., McKinney, S., & Stice, E. (2006). Multiple types of dieting prospectively predict weight gain during the freshman year of college. Appetite, 47(1), 83–90. https://doi.org/10.1016/j.appet.2006.03.160

Mai, R., & Hoffmann, S. (2012). Taste lovers versus nutrition fact seekers: How health consciousness and self-efficacy determine the way consumers choose food products.

Journal of Consumer Behaviour, 11(4), 316–328. https://doi.org/10.1002/cb.1390

Maxwell, S. E., & Delaney, H. D. (2004). Designing experiments and analyzing data: A model

Do We Really Eat Less with Smaller Bowls and Warning Labels?

Do We Really Eat Less with Smaller Bowls

and Warning Labels?

A Lab Study Examining the Effect of Bowl Size Reductions,

Warning Labels and Their Interplay on Food Consumption Quantity

Do We Really Eat Less with Smaller Bowls

and Warning Labels?

A Lab Study Examining the Effect of Bowl Size Reductions,

Warning Labels and Their Interplay on Food Consumption Quantity

Abstract

Preface

Table of Contents

List of Figures

List of Tables

1.

Introduction

2.

Theoretical Background

2.1.

Consumption Quantity

2.2.

Nudge Interventions

2.3.

Health Consciousness

2.4.

Development of the Conceptual Framework

3.

Methodology

3.1.

Participants

Highest educational qualification

3.2.

Research Design

3.3.

Procedures

4.

Results

4.1.

Reliability and Factor Analyses

4.2.

Descriptive Results

4.3.

Testing the Role of Bowl Size Reduction and Warning Label through

ANCOVA

4.4.

Testing the Role of Bowl Size Reduction and Warning Label through

Regression

4.5.

Testing the Role of Health Consciousness

4.6.

Robustness Tests

5.

Discussion

5.1.

Contributions and Implications

5.2.

Limitations and Directions for Future Research

6.

Conclusion

References