University of Groningen
Faculty of Economics and Business
Research Master in Economics and Business Marketing Department
July 11, 2019
Supervisors:
Prof. dr. ir. K. (Koert) van Ittersum Prof. dr. J. (Jenny) van Doorn
Author: Amber Werkman Kouwe 34 9103 PE Dokkum (06) 21453114 A.Werkman.2@student.rug.nl Student number S2754274
THE UNIT-SIZE EFFECT
REVISED: A WIN-WIN SOLUTION
FOR REDUCING BOTH FOOD
2
INTRODUCTION
Obesity is one of the biggest health concerns in communities across the world, it has nearly tripled since 1975 (World Health Organization, 2017). With the enormous attention now being focused on this issue, there is a transition from blaming the individual (in terms of personal responsibility) to blaming society (Senauer, 2003). In particular, the food industry is accused because they have done their work all too well. Under pressure from the ever-competitive food industry, researches together with managers perfected not just the preferred tastes of prepared foods, but also their packaging and advertising (Powell, 2013). Consequently, dietary patterns have substantially changed during the last 30 years (Cecchini & Warin, 2016). On the one hand, food availability has been increasing worldwide, take for example the massive growth of the supermarket and grocery store industry. At the same time, calorie consumption has significantly risen in the USA and a number of European countries (Cecchini & Warin, 2016). A substantial part of this rise can be attributed to increased consumption of ultra-processed products (i.e. ready-to-consume foodstuff) that now account for up to more than a half of total calorie intake in high-income countries. Lastly, the economic crisis has further exacerbated dietary patterns; people tend to switch to lower-priced (per calorie) and less healthy food (Cecchini & Warin, 2016).
Fortunately, communities can do something about this major public health concern, because overweight and obesity are largely preventable. Supportive environments and communities are fundamental in shaping people’s choices, by making the choice of healthier foods the easiest choice (the choice that is the most accessible, available and affordable), and therefore preventing overweight and obesity (World Health Organization, 2017). Health interventions proposed by the government or other initiatives by enterprises can empower consumers and facilitate healthier food choices by providing information or increasing consumer awareness (Cecchini & Warin, 2016). Food labelling, for example, is increasingly considered a crucial component of comprehensive strategies to tackle unhealthy diet and address the increasing prevalence of overweight and obesity in the United States (Cecchini & Warin, 2016).
3
Birau, Grinstein, ... & Moscato, 2016). In the UK alone, it has been estimated that households generated 7.2 million tonnes of food and drink waste a year, which was for a large extent avoidable (Graham-Rowe, et al., 2014). This is not only a waste of money, but it also has a large impact on the environment. There is a growing awareness of the positive impact of reducing the amount of wasted food on greenhouse gas emissions, energy use, food and water security, and land use (Quested, Marsh, Stunell & Parry, 2013). To guide consumers in reducing household food waste, a better alignment between their purchasing and actual consumption behaviour should be emphasized.
Much evidence suggests that people overconsume because they eat mindlessly. Painter and his colleagues (2005) discussed in their paper that visual cues of portion size influence consumption norms and expectations and it lessens one’s reliance on self-monitoring. In addition, most individuals do not consciously make negative choices to engage in behaviours that lead to the wastage of food (Evans, 2011). According to the behavioural sciences, environments can be structured in such a way that they activate self-control and subsequently facilitate healthy and prudent behaviour. In particular, Schwartz, Riis, Elbel & Ariely (2012) highlight the potential importance of portion-control interventions that specifically activate consumers’ self-control. Wansink and Van Ittersum (2013) provide such an intervention; in their article they present a win-win solution for reducing both food intake and waste. They examined the impact that plate size has on one’s consumption norms and on one’s behaviour. The authors found that the best approach to reducing or eliminating the perils of large portions is to simply encourage people to replace larger plates and bowls with smaller ones. So, portion size decisions relate to overconsumption and food waste and both issues can be solved by decreasing serving/ portion size themselves, leading to a win-win situation.
4
issue, we propose that the unit-size effect is even more impactful when people are given the opportunity to determine portion sizes themselves rather than offering them pre-portioned amounts.
Past research has established the importance of healthy eating and food waste as a research topic in the field of consumer behaviour and food consumption (Cecchini & Warin, 2016; Graham-Rowe, et al., 2014; Wansink & Van Ittersum, 2013). However, relatively little is known about the mutual connection between these two concepts. As a result, this paper aims to expand this research and build further on the attempt of Wansink and Van Ittersum (2013) to combine insights about healthy food consumption and food waste in order to tone down both issues of obesity and food waste. Therefore, this paper will address the following research question: What is the impact of unit size on food consumption and waste, and is this effect
moderated by portioning leading to a win-win solution for reducing both food intake and waste?
We present two studies that compare the effects of eating food in small vs. large unit size. We offer a new perspective, showing that the unit-size effect of food can be even more impactful when people are given the opportunity to determine portion sizes themselves, resulting in a win-win solution for reducing both food intake and waste. The provision of unportioned instead of pre-portioned foods facilitates eating according to internal rather than external cues and makes people more involved in the process of selecting food, which in turn restricts their portion size selection. We organise the remainder of this article as follows. The purpose of study 1 is to replicate the unit-size effect and to investigate the moderating role of portioning (unportioned vs. pre-portioned) on food consumption and waste. That is, participants are given the opportunity to consume treacle waffles in either small or large unit size from a portion that was either composed by themselves or by the researcher. In study 2, we test the unit-size effect on portion size selection for different portioning conditions. Moreover, we uncover the underlying process that drives this unit-size effect. That is food unit size shifts perceptions of the size of foods resulting in a bias in estimations, in line with the numerosity heuristic. In this study, participants are instructed to select a desired portion of gingerbread in a digital environment and reflect on it.
THEORETICAL BACKGROUND
5
Smaller portion sizes can help consumers to control their food intake because it influences perception, behaviour and satisfaction (Wansink & Van Ittersum, 2013). In addition, smaller sizes and packages may help individuals increase their self-control and manage their consumption (Scott et al., 2008). In particular, unit size research suggests that consumption and thus caloric intake is effectively reduced when unit sizes are smaller (compared to larger-sized units) (Ello-Martin, Ledikwe, and Rolls, 2005; Cheema and Soman, 2008). The unit size of food can be explained as the number of units in which a portion of food is divided (Van Kleef et al., 2014). For example, two portions of pizza may be equal in terms of content (e.g. in weight, calories or volume), but can be different in unit size (four slices vs. one). In one pizza lab and one pizza field experiment, Davis, Payne and Bui (2016) found that people compensate for smaller-sized pizza slices by eating more slices compared to regular-sized pizza slices, but the aggregate of calories people consume was still less. This suggests that a reduction in food size may result in increased unit choice but still fewer total calories to be consumed, in line with the traditional unit-size effect (Davis, Payne and Bui, 2016; Scott et al, 2008; Cheema and Soman, 2008).
The consumption effect of the unit size of food has been explained by Cheema and Soman (2008) as a decision bias that may successfully constrain consumption. In their article, they propose that partitions during consumption provide decision-making opportunities which draw attention to consumption. In other words, a partition increases the amount of attention payed to the decision because it provides ‘pause moments’ at which consumers must evaluate whether to continue consumption. Such decisions points facilitate a moment of reflection on one’s own behaviour and consequently shift the consumption decision from an automatic choice to a deliberative choice (Cheema and Soman, 2008). In addition, partition includes a small transaction cost by drawing attention to the physical act and adding a short temporal delay to consumption. As a result, smaller unit sizes of food can facilitate consumption monitoring and can be used to exert self-control and stop consuming. Furthermore, it can constrain consumption, because resources are consumed at a lower rate when they are partitioned (Cheema and Soman, 2008).
6
impulsiveness. Impulsivity was defined as “acting on the spur of the moment without considering long-term consequences and relates to a failure to avoid temptations (Churchill & Jessop, 2011)” (Van Kleef et al., 2014, p. 1083). Eating various smaller units of food activates the inference of excessive consumption compared to eating the same amount of food in a larger unit size. This observation can be linked to the explanation of Cheema and Soman (2008), who suggest that encountering a partition during consumption facilitates consumption monitoring by increasing the amount of attention consumers pay to the act of consuming. As a result, through this increased salience of consumption, consumers may perceive that they have eaten more units of smaller unit sized foods, stimulating perceptions of excessiveness and impulsivity. Furthermore, the second study revealed that people overestimated their consumed grams when they eat small unit sized food. Altogether, we can conclude that consuming the same amount of food presented in smaller units is considered to be more impulsive, excessive and less appropriate than consuming the same amount presented in one large unit (Van Kleef et al., 2014). Consequently, consumers may perceive that they have eaten enough and stop eating sooner when they eat several smaller units of food.
In sum, research has shown that consumers tend to consume less food when eating smaller-sized morsels than when eating larger-smaller-sized morsels because consumption monitoring is improved and each unit consumed also increases perceptions of overindulgence and impulsivity. In addition, the unit bias explains that people consider a single entity (within a reasonable range of sizes) as the appropriate amount to engage, consume or consider (Geier, Rozin, & Doros, 2006). Consequently, people consume much greater weights of food when offered a large as opposed to a small unit size because they feel compelled to finish their bit. Consistent with prior research on the unit-size effect, we propose that small unit sizes are effective in controlling consumption.
H1: People consume fewer calories when food is presented in smaller units compared to larger units (unit-size effect).
7
demonstrate that the consumed quantity or the rate of consumption of a resource reduces when the aggregate quantity of that resource is partitioned into smaller units. The flip side of the coin is, though, a clear increase in waste of the resource and this is particularly bad when it comes to food waste. Firstly, it has an ethical impact as it contributes towards increases in global food prices, making food less accessible (Graham-Rowe, Jessop & Sparks, 2014). Secondly, it has an economic impact because not only the food but also the money that is spent on it is lost. Lastly, it has of course a big impact on the environment such as climate, water, land and biodiversity (FAO, 2013).
To resolve this issue, we propose that the unit-size effect is even more impactful from a sustainable perspective when people are given the opportunity to determine portion sizes themselves rather than offering them pre-portioned amounts. Despite the fact that pre-portioned amounts of food can be offered in smaller and larger portions, we will assume for now that portion size is fixed, termed as pre-portioned. Specifically, we define portion size as the standardized amount of a single food item served in a single eating occasion as determined by the manufacturer (Faulkner et al., 2012; Division of Nutrition and Physical Activity, 2006; Institute of Grocery Distribution, 2008). Childcare food program guidelines widely endorsed family-style meal service, which allow children to serve themselves (Branen, Fletcher & Myers, 1997). The rationale here is that self-serving increases the development of self-help skills such as autonomy and self-regulation in eating (Savage, Haisfield, Fisher, Marini & Birch, 2012). Additionally, when children are allowed to make quantitative food choices they learn to trust their individual needs and to have confidence in their ability to elicit responses to those needs (Branen et al., 1997, p. 89). In other words, self-service is nutritionally appropriate because it facilitates eating according to internal rather than external cues. For example, people are encouraged to manage their own food intake, without being guided by environmental cues that steer how much food to eat. In the view of food waste, people are more involved in the process of selecting food and this can effectively reduce waste (Berman and Fromer, 1991). In a pizza field experiment where participants could request as many slices as they liked, scholars generally observed that participants ate all of the food units they took (Davis, Payne and Bui, 2016).
8
H2: People will (a) consume fewer calories, and (b) waste fewer calories when food is unportioned compared to pre-portioned in a standard-sized portion.
Insights from food consumption studies suggest that people overconsume because they eat mindlessly. For example, people use their eyes instead of their stomach to determine consumption. Consequently, they tend to eat most of the food put before them and feel satiated only when their plates are empty, regardless of the amount of food that is actually served on their plates or bowls (Wansink, Just & Payne 2009; Wansink, Painter & North, 2005). In this way visual cues of portion size can influence consumption norms and expectations and it lessens one’s reliance on self-monitoring. According to the behavioural sciences, environments can be structured in such a way that they activate self-control and subsequently facilitate healthy and prudent behaviour. In particular, Schwartz, Riis, Elbel & Ariely (2012) highlight the potential importance of portion-control interventions that specifically activate consumers’ self-control. In three experiments, they found that a simple change in the ordering process can activate the self-control necessary to restrict portion size selection and, ultimately, decrease calorie consumption.
As described above, the provision of unportioned foods can be another strategy that involves smaller portion sizes, which makes it easier for consumers to control their intake of calories. People can better regulate their consumption because portions are composed beforehand and based on internal cues. As a consequence, portions are better aligned with the individual needs of the consumer and external cues are less pronounced. In line with this reasoning, the unit size of food will have less of an effect on consumption since unportioned food requires self-serving, which subsequently increases self-regulation and self-control in eating. In addition, when portions are better aligned with individual consumption needs, plate waste will reduce accordingly.
H3: The unit-size effect is moderated by portioning; this effect is weakened when food is unportioned compared to pre-portioned.
9
Figure 1: Conceptual model concerning food consumption and waste
STUDY 1 Participants and design
For the first study, a 2 (unit size: small vs. regular) x 2 (portioning: unportioned vs. pre-portioned) between-subjects experimental design was used. Two hundred and thirty-seven students from a European university (University of Groningen) participated in this lab study in exchange for course credits or a monetary compensation. Eleven participants were not included in the analyses as they refused to take cookies in the unportioned condition (five participants) or failed the attention check (six participants). The majority of the responses are Dutch (50.7 %), German (9.3 %) and Indonesian (8.4 %). The sample consists of 97 (42.9 %) men and 129 (57.1 %) women who were aged between 18 and 44 years (M = 21.99, SD = 2.83). On the basis of participants’ length and weight, we computed one’s BMI (Body Mass Index). This is a measurement of body fat. The average BMI of women was 23.17 (SD = 2.83) and the average BMI of men was 22.01 (SD = 2.94).
Procedure
10
participants could serve undisturbed, meanwhile the researcher made the computer ready. Participants in the pre-portioned condition were immediately shown their place and the portion was already served (see appendix 2 for a screenshot of the setting). All participants started the experiment by entering their participant number and putting on their headphones. Following this, participants watched a short movie about Dutch history while consuming as much treacle-waffles as they liked (see appendix 3 for sample fragments). After the movie, participants were kindly requested to dispose of the remaining food in the trash bin placed behind them in the cubicle because of safety reasons regarding the electronica. The experiment continued with a number of questions about the movie. All items were presented in a random order. Finally, we asked for participants’ gender, age, nationality, length and weight. At the end of the study, participants had the opportunity to comment on the study and we thanked them for their participation.
Measures
Unit size. To systematically assess the effect of unit size on food intake and waste, we selected
two different sizes of treacle-waffles (Dutch stroopwafels). The regular treacle-waffle weights 40.9 gram and contains 196 calories. The small treacle-waffle weights 8.2 gram and contains 39 calories.
Portioning. To assess the influence of the way food is portioned on consumption and waste, we
establish two different conditions (unportioned vs. pre-portioned food). In the unportioned condition, participants could select their own portion by serving the desired number of treacle-waffles on a plastic plate. The offered amount of treacle-treacle-waffles was held the same for every participant. In the pre-portioned condition, participants were offered a fixed amount of treacle-waffles (2 units in the regular unit size condition and 10 units in the small unit size condition). The portions had an average weight of 81.7 and contained approximately 392 calories. We considered this size as an acceptable portion because this amount (two regulartreacle-waffles packaged in plastic) is also offered as a snack in supermarkets or canteens (e.g. Kanjers).
Food consumption and waste. To measure participants’ waste, we weighed the total amount of
11
consumption/ waste measure in grams by the total portion size in grams, resulting in a consumption/ waste measure in percentage.
Results
The means of the variables of interest are presented in table 1.
Table 1: Descriptive statistics for portion size, consumption and waste Portion size Consumption
(grams) Consumption (%) Waste (grams) Waste (%) Mean 58.67 42.89 78.18 15.78 21.82 Standard deviation 26.18 23.90 30.31 22.58 30.31 Minimum 8.20 0 0 0 0 Maximum 123.00 123.00 100 82.00 100
Portion size. A two-way ANOVA revealed a statistically significant interaction between the
effects of unit size and portioning on portion size, F(2,222) = 41.617, p = .000. Simple main effects analysis showed that portions in the pre-portioned condition were significantly bigger in size than portions in the unportioned condition for both unit sizes (small, p = .000; regular, p = .000). In addition, there was a main effect of unit size (small vs. regular) on portion size (F(1,222) = 42.527, p = .000). Overall, participants in the small unit size condition served smaller portion sizes than participants in the regular unit size condition, and participants in the unportioned condition served smaller portions than participants in the pre-portioned condition. In terms of the interaction effect, participants in the pre-portioned condition received equal portion sizes in both small and regular
unit size conditions. However, participants in the unportioned condition served themselves less food when the unit size was small compared to regular (see figure 2 for a graphical representation). These results are in line with our expectations because people will probably consume and waste less when portion size is reduced. In the unportioned condition, we see a clear
12
decrease in portion size compared to the pre-portioned condition. Especially, portion size is reduced when unit size is small compared to regular, hinting to the unit-size effect.
Consumption. A two-way ANOVA was conducted that examined the effect of the unit size and
13
Figure 3: Estimated means of consumption in grams Figure 4: Estimated means of consumption in percentage
Waste. A two-way ANOVA was conducted that examined the effect of the unit size and the
portioning of food on waste.Simple main effects analysis showed that waste (both in grams and percentage) was significantly lower in the unportioned condition (M =2.31, SD = 7.23) compared to the pre-portioned condition (M = 30.47, SD = 24.42) (F(1,222) = 156.239, p = .000)1. These results confirm that people will waste fewer calories when food is unportioned compared to pre-portioned in a standard-sized portion, confirming hypothesis 2b. In terms of the interaction effect, participants in the unportioned condition wasted similar amounts of food (around 10% of their selected portion) in both small and regular unit size conditions (small, p = .000; regular, p = .000). Within the pre-portioned condition, we observed the opposite unit-size effect; people waste more (in percentage) when food is presented in smaller units compared to larger units. Thus also for waste, we can conclude that the unit-size effect is moderated by portioning; this effect is weakened when food is unportioned compared to pre-portioned, supporting hypothesis 3.
To conclude, we graphically summarized all our findings in a column chart (figure 5).
14
Figure 5: Summary of the research findings
Note: the vertical axis represents the portion size in grams.
Discussion
In line with earlier findings on the unit-size effect (Davis, Payne & Bui, 2016; Scott et al., 2008), we confirmed that people consume less when food is presented in smaller units compared to larger units (hypothesis 1). In addition, we extend this finding by introducing the unportioned food intervention that transfers the unit-size effect to a lower level of consumption (see figure 3). Consumption already decreases when eating smaller-sized food units as compared to larger-sized food units, but this intervention is even more impactful when people are given the opportunity to determine portion sizes themselves. Namely, regardless of the unit size people consume fewer calories when food is unportioned compared to pre-portioned in a standard-sized portion. Next to the fact that we were able to significantly reduce consumption, the analyses also showed a beneficial interaction between unit size and portioning on food waste. In the pre-portioned condition, waste increased when food was presented in smaller units compared to larger units. This can be considered as the flipside of the unit-size effect. On the other hand, waste reduced when the moderation of unportioned (vs. pre-portioned) food was introduced. The level of waste is at its lowest in the unportioned condition when food is presented in smaller units compared to larger units. Overall, we can conclude that unportioning is an effective intervention to reduce food consumption and waste, in line with hypothesis 2. Lastly, we examined if the unit-size effect is moderated by portioning and especially if this effect is weakened when food is unportioned compared to pre-portioned. The results indicated
0,00 10,00 20,00 30,00 40,00 50,00 60,00 70,00 80,00 90,00
Regular Small Regular Small
15
that there was no interaction between the unit size and the portioning of food on consumption in grams. The unit size-effect remains almost the same (in strength) when food is unportioned instead of pre-portioned. An in-depth analysis of this finding revealed that the unit-size effect shifts from consumption to portion size when people are given the opportunity to determine portion sizes themselves. We observed a clear effect of unit size in the unportioned condition when we looked at the selected portion size in grams; participants in the small unit size condition served smaller portion sizes than participants in the regular unit size condition. On the other hand, we observed the classical unit-size effect in the pre-portioned condition when we looked at consumption in percentage; participants consumed less of their portion when food is presented in smaller units compared to larger units. Together, these findings cause the insignificant interaction between the unit size and the portioning of food on consumption in grams. This is an interesting observation because it reveals a two-stage process in the unportioned food condition. People first have to decide how much they are going to serve themselves in terms of portion size (pre-selection stage) and this strongly influences their actual consumption (consumption stage). In the pre-portioned condition, people only come across the consumption stage and here consumption might be influenced by the quantity of the predetermined portion (portion size effect2; Rolls, Morris & Roe, 2002; Young & Nestle, 2002). This suggests that consumption is a function of how much is served and we can bring this to a lower level by letting people serve their own portions (unportioning). In addition, consumption already decreases when eating smaller-sized food units as compared to larger-sized food units, but this intervention is from a sustainable perspective even more impactful when food is provided in an unportioned way. Namely, a decrease in consumption will not lead to an increase in waste because of the shifted unit-size effect from the consumption to the pre-selection stage. To conclude, when we solely focus on consumption of the portion, we did find a significant interaction between the unit size and portioning of food on consumption in percentage. We found a weakened unit-size effect when food is unportioned compared to pre-portioned, in line with hypothesis 3.
The main goal of this first study was to replicate the unit-size effect and to investigate the moderating role of portioning (unportioned vs. pre-portioned) as an effective intervention that creates a win-win solution for reducing both food consumption and waste. In a lab experiment, we measured participants’ actual consumption and waste behaviour and found that they (a)
2 The portion size effect reflects the positive relationship between portion size of food and the amount of food that
16
consumed less food and thus fewer calories, and (b) wasted less food and thus fewer calories when food is unportioned compared to pre-portioned in a standard-sized portion. When examining these results in more detail, we learned that consumption was largely driven by the amount of food that people served. In other words, the consumed quantity of a portion is a function of the selected quantity that makes up that portion. This is an interesting but above all an important finding that sheds light on an additional angle in the food consumption process. In the second study we will take a closer look at this consumption process by solely focussing on the food selection stage. Furthermore, we will expand our design by looking into the underlying mechanism that explains why people consume less food when food is presented in smaller compared to larger units. Lastly, we will adjust our moderator in two ways. Firstly, a third condition will be added to create diversity in the pre-portioned condition by means of portion size. According to the portion size effect, larger portions lead to greater energy intake regardless of subject characteristics (Rolls et al., 2002). As a result, we will include a smaller and a larger portion size to capture this influence in our model. Secondly, in all three conditions participants could select as much food as they preferred in order to make the procedures more equal among conditions. This means that portion size is no longer fixed in the pre-portioned condition.
EXTENDED THEORETICAL FRAMEWORK
17
food that is consumed is highly dependent on the decisions about portion size, before a meal begins (Brunstrom, 2011). Fay et al. (2011) conclude in their article that meal size is typically planned in advance and that satiation often plays a secondary role in determining the amount of food that is consumed. Following this line of research, we will further unfold the cognitive structures that underpin the decisions that we make about food and portion size prior to consumption by investigating the unit-size effect on portion size selection. Specifically, the goal of study 2 is to test the unit-size effect on portion size selection for different portioning conditions. Moreover, we want to uncover the underlying process that drives this unit-size effect.
Oldham-Cooper et al. (2017) tested the hypothesis that the expected satiety of a food, as a proxy for portion size selection, can be increased by presenting it in small units compared to regular or large units. Expected satiety can be defined as the expected relief from hunger between meals and independently predicts self-selected ‘ideal’ portion sizes (Brunstrom & Rogers, 2009; Brunstrom & Shakeshaft, 2009). Results revealed a consistent unit-size effect on expected satiety: foods presented in multiple smaller units were expected to deliver significantly greater satiety than when presented in few larger units. These results thus support the idea that the unit size of food impacts beliefs about a meal before it is consumed. Consequently, expected satiety plays an important role in determining the portion sizes that people select and therefore can lead to differences in portion size selection. Moreover, in study 1 we found a clear decrease in portion size selection when food was presented in smaller units compared to larger units. In the current study, we seek to confirm the anticipated impact of the unit size of food on portion size choices and later consumption intention. Based on the above reasoning, we predict:
H4: People select smaller portions when food is presented in smaller units compared to larger units.
18
unit-size information. So this research suggests that number of units is a more powerful indicator of quantity than physical size, in line with the numerosity heuristic (Pelham, Sumarta & Myaskovsky, 1994). This heuristic goes back to the prehistorical era where it evolved as a trait that helped ensure survival in an environment in which food was scarce (Van Kleef et al., 2014). “More pieces of something usually turns out to be more of that something” (Pelham et al., 1994, p. 105). For example, a house with five bedrooms is usually larger than a house with two bedrooms.
Despite the salience of number and its usefulness in making quantity determinations, the overreliance on numerosity as a cue for judging quantity can also lead to bias. In previous research, it was found that consuming five small units of chocolate was considered to be more impulsive, more excessive and less appropriate than consuming the same amount of chocolate as one single unit (Van Kleef et al., 2014). The notion that multiple smaller units are giving the impression of increased consumption is in line with the numerosity heuristic; people judge amounts on the basis of the number of units into which a stimulus is divided without fully considering other important variables such as the size of the units (Pelham et al., 1994). Yet, the authors could not consistently identify a bias in consumption estimations due to unit size as a possible indication of this heuristic across the three studies (Van Kleef et al., 2014). Additional research in this area investigated partitioning as a function of granularity of portion size (Lewis Jr & Earl, 2018). Lewis and Earl (2018) demonstrated that portion size granularity significantly influenced consumption intentions (i.e., how much people plan to eat in a given sitting, how much people feel they need to eat in order to feel satisfied) and subsequently impacted actual consumption. This process operates by shifting perceptions of the size of foods whereby fine-grained portions (e.g., “15 gummy candies”) are perceived to be bigger than gross-fine-grained portions (e.g., “one serving of gummy candies”) (Lewis Jr & Earl, 2018). Consequently, people have the feeling that they would need to eat less in order to be satisfied. These results indicate that, holding portion size constant, smaller versus larger food units may shift consumption judgments (e.g., calorie estimates, weight, cost, time to eat). Lastly, results of Oldham-Cooper et al. (2017) indicated that the positive effect of segmentation (presenting the same amount of a food in multiple smaller units) on expected satiety was attributable to an increase in perceived volume.
19
presented in larger units, resulting in a higher quantity estimation. Accordingly, if smaller unit sizes lead to higher quantity estimations, portion size selection will decrease. Stated formally:
H5: The relationship between the unit size of food and portion size selection will be mediated by quantity estimation.
Finally, we predict that the relationship between unit size and quantity estimation is moderated by portioning (unportioned vs. pre-portioned). The majority of our meals require some form of preparation. As stated earlier, consumption is not governed primarily by satiation (within-meal events), but rather learned and expressed in the cognitive activity associated with meal planning, so before a meal begins (Brunstrom, 2011). In the unportioned condition, consumers’ cognitive activity associated with meal planning is relatively high because consumers are completely free in how much units they serve. Contrary, in the pre-portioned condition, consumers’ cognitive activity associated with meal planning is much lower because people can use the standard sized portion as a reference point (Robinson & Kersbergen, 2018). Portion size is thought to communicate information about what constitutes a normal amount of food to eat (Herman, Polivy, Pliner & Vartanian, 2015; Kerameas, Vartanian, Herman & Polivy, 2015). This probably has consequences for accurate quantity estimation.
20
In this context, it is worth distinguishing the situation that we have been considering, in which the experimenter serves a somewhat arbitrarily standard portion to participants, from the situation in which participants select their own portion. Following the portion size effect, the size of the offered portion does matter since larger initial portions lead to greater intake (Herman, Polivy, Pliner, Vartanian, 2015). As a consequence, we distinguish between a smaller pre-determined portion and a larger pre-determined portion (Diliberti, Bordi, Conklin, Roe & Rolls, 2004). In in the large pre-portioned condition, consumers’ cognitive activity associated with meal planning is relatively low because of the large serving sizes (there is not much to choose for consumers). In the small pre-portioned condition, this cognitive activity is somewhat increased because of the small serving sizes. Comparing the two pre-portioned conditions, we predict that the unit-size effect on quantity estimation and subsequently on portion size selection will be stronger in the small pre-portioned condition compared to the large pre-portioned condition. The rationale here is that smaller portions increase cognitive activity which makes unit size more salience. All in all, we propose:
H6: The effect of unit size on quantity estimation is moderated by portioning;
(a) This effect is strengthened when food is unportioned compared to pre-portioned. (b) In the pre-portioned condition, this effect is stronger for small-sized portions than for
large-sized portions.
The hypotheses are visually represented in a conceptual model (figure 6).
21 STUDY 2 Participants and design
For the second study, a 2 (unit size: small vs. regular) x 3 (portioning: unportioned vs. small pre-portioned vs. large pre-portioned) between-subjects experimental design was used. Three hundred and forty-three students from a European university (University of Groningen) participated in this lab study in exchange for course credits or a monetary compensation. Fifteen participants were not included in the analyses because the question about the selected portion size did not function properly (eight participants) or participants failed the attention check (seven participants). The majority of the responses are Dutch (33 %), Indonesian (10.9 %) and German (10.3 %). The sample consists of 129 (39 %) men and 299 (61 %) women who were aged between 17 and 35 years (M = 20.92, SD = 2.91). On the basis of participants’ length and weight, we computed one’s BMI (Body Mass Index). This is a measurement of body fat. The average BMI of women was 22.28 (SD = 3.54) and the average BMI of men was 22.90 (SD = 3.04).
Procedure
When participants entered the research lab, they were first asked to read and sign the information brochure about the purpose of the study. Subsequently, participants were guided to one of the cubicles with the instruction to carefully read the briefing. Sessions were administered on a computer using Qualtrics. First of all, participants were randomly assigned to one of the six experimental conditions in which the manipulated factors were the unit size and the portioning of food. All participants started the experiment by entering their participant number. Following this, participants were introduced to the “Pick your portion” task in which they were asked to put together a portion of gingerbread in a digital environment (see appendix 4 for a screenshot of this digital environment). Participants had to imagine that they selected their preferred portion of gingerbread for the following situation: “It is 13.00 and you just
finished your regular lunch at the university canteen. You decide to stay till 17.00 at the university to study for your ongoing courses. In between you take a break to get some coffee/tea and consume a snack. We would like you to select your preferred portion of gingerbread as the snack for that moment. The amount of this snack needs to satisfy your hunger until the next meal.” To make it more realistic, the amount of gingerbread that participants could drag at a
22
participants were asked to indicate how many pieces of gingerbread their selected portion contained and consumption intention for the given situation was measured. The experiment continued with some process measures, including consumption norm, perceived quantity, expected satiety, impulsiveness and regulation with regard to the selected portion. In the last part, participants were asked to complete a couple of scales about their values, general health interest, health motivations and dietary restraints. All items were presented in a random order. Finally, we asked for participants’ gender, age, nationality, length and weight. At the end of the study, participants had the opportunity to comment on the study and we thanked them for their participation.
Measures
Unit size. To systematically assess the effect of unit size on portion size selection, we selected
two different sizes of gingerbread (Dutch: ontbijtkoek). The regular piece of gingerbread weights 25 gram and contains 73 calories. The small piece of gingerbread weights 6.25 gram and contains 18 calories.
Portioning. To assess the influence of the way food is portioned on the relationship between
23
Table 2: Experimental conditions
UNIT SIZE PORTIONING Unportioned (single units) Small pre-portioned (small serving of 50 gram) Large pre-portioned (large serving of 100 gram) Small 6,25 gram 18 kcal Regular 55 gram 161 kcal
Portion size selection. To measure participants’ selected amount of food, we counted the total
number of food items that individuals dragged into their selected portion and multiplied this by the average weight of the food item (depending on the unit size). This resulted in a portion size selection measure in grams.
Quantity estimation. Quantity estimation of the food serving (see table 2) was operationalised
in terms of percentages. We asked participants to estimate the weight of the serving of gingerbread that is presented under the glass bell. The standard approach to calculating quantity estimation in percentage uses the following equation: (estimated quantity in grams / actual quantity in grams) × 100.
Consumption intention. After participants selected their preferred portion size, they were asked
to indicate how much of their portion they would likely eat in the given situation. We multiplied this number by the average weight of the food item (depending on the unit size). This resulted in a consumption intention measure in grams.
Results
24
whether to the effect of unit size on portion size selection can be explained by changes in quantity estimation. Lastly, a moderated mediation analysis was conducted to account for the complete model. The means of the variables of interest are presented in table 3.
Table 3: Descriptive statistics for quantity estimation, portion size selection and consumption intention Quantity estimation
(percentage)
Portion size selection Consumption intention
Mean 252.45 74.90 53.38
Standard deviation 234.82 38.80 32.23
Minimum 20 12.50 0
Maximum 20003 200 200
Quantity estimation. A two-way ANOVA was conducted that examined the effect of the unit
size and the portioning of food on quantity estimation. There was a main effect of both unit size (F(1,322) = 24.935, p = .000) and portioning (F(1,322) = 66.796, p = .000) on quantity estimation. Moreover, we found a statistically significant interaction between the effects of unit size and portioning on quantity estimation, F(2,322) = 24.448, p = .000. Simple main effects analysis showed that quantity estimations were significantly higher in the small unit size condition compared to the regular unit size condition when food was unportioned (p = .000), but there were no differences between unit size when food was small pre-portioned (p = .854) or large pre-portioned (p = .768). These results are partly in line with our expectations. First of all, the significant interaction between unit size and portioning indicates the moderating role of portioning on the relationship between unit size and quantity estimation as we predicted. We observe a clear unit-size effect in the unportioned condition compared to the pre-portioned condition. However, the results did not indicate a difference within the pre-portioned condition; the unit-size effect was neither present in the small portioned condition nor the large pre-portioned condition. We will come back to this issue when testing the whole model (moderated mediation analysis). The effect of unit size on quantity estimation depending on the way food is portioned (moderator) is visually represented in figure 7.
3 We checked for outliers by calculating the z-score of each observation. Any z-score greater than 3 or less than
25
Figure 7: Effect of unit size on quantity estimation depending on portioning
Portion size selection. Another two-way ANOVA was conducted that examined the effect of
the unit size and the portioning of food on portion size selection. There was a main effect of unit size on portion size selection (F(1,322) = 16.909, p = .000). This result shows that people select smaller portions when food is presented in smaller units compared to larger units, consistent with hypothesis 4. In addition, there was a main effect of portioning on portion size selection (F(1,322) = 149.355, p = .000). Moreover, we found a statistically significant interaction between the effects of unit size and portioning on portion size selection, F(2,322) = 8.473, p = .000. Simple main effects analysis showed that participants selected significantly smaller portions in the small unit size condition compared to the regular unit size condition when food was unportioned (p = .000), but there were no differences between unit size when food was small pre-portioned (p = .724) or large pre-portioned (p = .283). These results provide detailed insight into the effect of unit size on portion size selection as predicted in hypothesis 4. People only select smaller portions when food is presented in smaller units compared to larger units under the condition that food is unportioned. This again confirms the moderating role of portioning, but this time on the relationship between unit size and portion size selection. The effect of unit size on portion size selection depending on the way food is portioned (moderator) is visually represented in figure 8.
26
Figure 8: Effect of unit size on portion size selection depending on portioning
Mediation analysis. The following analysis tests the proposed mediational pathway examining
the effects of unit size on portion size selection via quantity estimation. In this analysis, unit size (small vs. regular) is the independent variable, portion size selection (in grams) is the dependent variable, and quantity estimation (in percentage) is the mediator. We used PROCESS for SPSS v3.3 Model 4 for testing mediation with 1000 bootstrap samples (Hayes, 2017). The analysis supports the predicted pathway from unit size to portion size selection via quantity estimation as evidenced by the bias corrected 95% confidence interval for the indirect effects excluding zero (95% CI [4.3354, 9.7285]; see figure 9 for point estimates from the mediation model). Participants presented with small unit sizes overestimated the quantity, and as a consequence selected a smaller portion size than participants presented with regular unit sizes. So, we could confirm that the differential effect of unit size (small vs. regular) on portion size selection was caused by a change in quantity estimation. Overall, these results provide support for the mediational pathway of quantity estimation as proposed in hypothesis 5.
27
Figure 9: Mediation model depicting the process by which unit size influences portion size selection. Coefficients are unstandardized regression coefficients from the PROCESS model. * p < .05. ** p < .01. *** p < .001.
Moderated mediation analysis. In order to test the whole model including moderation, we
employed conditional process analyses using a series of moderated mediation models (PROCESS Model 7; Hayes, 2017) with 5000 bootstrap samples. Our model compared differences between unportioned, small pre-portioned and large pre-portioned food. Specifically, it tested whether the relationship between unit size (small vs. regular) and portion size selection was mediated by quantity estimation and whether this mediation depended on the way food was portioned. The 95% confidence interval for the conditional indirect effect of unit size onto portion size selection through quantity estimation did not include zero among participants in the unportioned condition (95% CI: 14.4419, 27.0268). Thus, when food is unportioned, those in the small unit size condition estimated quantity significantly higher than those in the regular unit size condition, and this increased quantity estimation led to the selection of a smaller portion size. Based on these results, we could confirm that the effect of unit size on quantity estimation is moderated by portioning. This effect is strengthened when food is unportioned compared to pre-portioned and thereby confirming hypothesis 6a. However, the 95% confidence interval included zero both in the small pre-portioned condition (95% CI: −2.6139, 1.7202) and large pre-portioned condition (95% CI: -.6218, 2.1237). Therefore, when food is pre-portioned (either in small or large servings), increased quantity estimation did not mediate the effect of unit size on portion size selection, rejecting hypothesis 6b.
Discussion
28
those in the regular portion size condition because they overestimated the quantity of the serving size, but only when food was provided in an unportioned way. When food is presented in smaller units, and subsequently provided in single units (unportioning), people overestimate the quantity of a serving, and this increased quantity estimation leads to a smaller portion size selection, supporting hypotheses 4 and 5. In terms of the moderator, we find a significant interaction between the unit size and the portioning of food on quantity estimation. Contrast analyses confirm that this effect is predominantly driven by the unportioned condition; estimations were significantly higher in the small unit size condition compared to the regular unit size condition when food was unportioned, but no differences were found when food was small pre-portioned or large pre-portioned (see figure 7). In other words, the effect of unit size on quantity estimation was strongly present when food was unportioned, but switched off in case food was pre-portioned, consistent with hypothesis 6a. This finding supports our notion that the unit-size effect shifts from food consumption to food selection when people are given the opportunity to determine portion sizes themselves. We observed a clear effect of unit size on portion size selection in the unportioned condition. On the other hand, we observed the classical unit-size effect on consumption intention in both pre-portioned conditions. In line with the numerosity heuristic (Pelham et al., 1994; Van Kleef et al., 2014; Lewis Jr & Earl, 2018), this process operates by shifting perceptions of the size of foods whereby food presented in smaller units is perceived to be bigger than food presented in larger units. Consequently, people have the feeling that they would need to select less in order to be satisfied.
GENERAL DISCUSSION
29
rationale here is that unportioning increases the development of self-help skills such as autonomy and self-regulation in eating (Savage et al., 2012).
In this paper, we studied the effect of unit size on portion size selection, food consumption and food waste. The unit size bias is the phenomenon that smaller units of food tend to reduce consumption (Van Kleef et al., 2014). The purpose of two between-subjects experiments was to examine how altering the unit size of food impacts eating behaviour under different portioning conditions. Across two studies, we provide beneficial evidence for the unportioned food intervention both from a health and sustainability perspective. Basically, with a small unit size, people select smaller portion sizes and subsequently consume and waste less, but only when food was provided in an unportioned way. In the first study, participants consumed fewer calories when food was presented in smaller units compared to larger units (unit-size effect). Participants’ consumption behaviour showed a similar pattern when provided with the opportunity to determine portion sizes themselves but on a lower level of consumption. Moreover, this portioning intervention significantly reduced food waste up to seven percent of the selected portion. Examining these results in more detail, we learned that portion size selection is an important determinant in the food consumption process. Specifically, the results indicate that the unit-size effect shifts from food consumption to food selection when food is unportioned and in this way prevent an increase in food waste.
30
our plate (and therefore in our stomachs). We found that participants selected smaller portion sizes when food was presented in smaller units compared to larger units, but only when food was unportioned. This supports our notion that the unit-size effect shifts from food consumption to food selection when people are given the opportunity to determine portion sizes themselves. As noted earlier, previous studies have shown that the unit size of food that is consumed has a robust and independent effect on satiation and ad libitum eating (Cheema & Soman, 2008; Van Kleef et al., 2014; Geier, Rozin, & Doros, 2006). Our results complement these findings and they extend this literature by showing that food unit size has a significant effect on decisions about portion size, before a meal begins – under the condition that food is provided unportioned. Moreover, this process operates by shifting perceptions of the size of foods whereby food presented in smaller single units is perceived to be bigger than food presented in larger single units. Consequently, people have the feeling that they would need to select less in order to be satisfied. From this we can conclude that foods of equal volume but divided into different unit sizes are not expected to be equally satiating, consistent with previous research on perceptual judgments of the amount of food (Brunstrom, et al., 2010; Oldham-Cooper et al., 2017; Lewis Jr & Earl, 2018). So food volume, specifically perceived volume, plays an important role in understanding the unit-size effect in eating behaviour.
In summary, both food intake and waste can be controlled by processes that operate before a meal begins. Previous research found that participants consumed less food when it was presented in multiple small units compared to one or few large unit(s). Our findings extend this work by showing that the unit size of food impacts beliefs about a meal before it is consumed, leading to a win-win solution for reducing both consumption and waste. An important boundary condition in this regard is the portioning of food. For pre-portioned food, the unit-size effect does not hold on portion size selection, but is only effective on ad libitum intake. Unfortunately, this results in negative implications for food waste. For unportioned food, people select smaller portion sizes and subsequently consume and waste less when food units are small compared to large. In conclusion, the unit-size is most effective in meal planning provided that food units are perceived as single units from which a portion can be composed.
Limitations and future research
31
investigation. For example, one relationship is that unportioned instead of pre-portioned food reduces both consumption and waste. Questions for this relationship do arise. Did people feel more self-control when they got the opportunity to determine portion sizes themselves and subsequently selected less food? Did people satisfy their needs in the unportioned condition? In the unportioned situation, we may well ask why participants selected the particular portions they do. Second, lab sessions were scheduled throughout the day, which could influence participants’ consumption behaviour. In other words, what a participant consumed before entering the research lab could have an impact on how much (s)he ate/ selected during the session. For example, it is likely that people consume/select less food after lunchtime. In the future, it is recommended to focus on a specific consumption moment, for example lunch or dinner, to minimize this fluctuation. Third, this study only focussed on consumption and waste in terms of the amount of one type of food. In the future, it would be useful to include different types of food in order to determine if people choose different sources within food groups when served family-style. In this context, it would be more appropriate to conduct controlled experiments in the field (e.g. in a canteen setting). Also in the future, other predictors of portion-selection might be considered. For example, it would be interesting to incorporate measures of food-specific restraint and other beliefs about the health consequences of consuming food. Lastly, the way will be open to further uncover the underlying mechanisms that explain how portioning influences both food consumption and waste.
CONCLUSION
32
REFERENCES
Berman, C., & Fromer, J. (1991). Meals without Squeals: Child Care Feeding Guide and
Cookbook. Palo Alto, CA: Bull Publishing Company.
Block, L. G., Keller, P. A., Vallen, B., Williamson, S., Birau, M. M., Grinstein, A., ... & Moscato, E. M. (2016). The squander sequence: understanding food waste at each stage of the consumer decision-making process. Journal of Public Policy & Marketing, 35(2), 292-304. Branen, L., Fletcher, J., & Myers, L. (1997). Effects of pre-portioned and family-style food service on preschool children's food intake and waste at snacktime. Journal of Research in
Childhood Education, 12(1), 88-95.
Brunstrom, J. M. (2011). The control of meal size in human subjects: a role for expected satiety, expected satiation and premeal planning. Proceedings of the Nutrition Society, 70(2), 155-161. Brunstrom, J. M., Collingwood, J., & Rogers, P. J. (2010). Perceived volume, expected satiation, and the energy content of self-selected meals. Appetite, 55(1), 25-29.
Brunstrom, J. M., & Rogers, P. J. (2009). How many calories are on our plate? Expected fullness, not liking, determines meal‐size selection. Obesity, 17(10), 1884-1890.
Brunstrom, J. M., & Shakeshaft, N. G. (2009). Measuring affective (liking) and nonaffective (expected satiety) determinants of portion size and food reward. Appetite, 52(1), 108e114. Chandon, P., & Ordabayeva, N. (2017). The accuracy of less: Natural bounds explain why quantity decreases are estimated more accurately than quantity increases. Journal of
Experimental Psychology: General, 146(2), 1-60.
Chandon, P., & Wansink, B. (2007). Is obesity caused by calorie underestimation? A psychophysical model of meal size estimation. Journal of Marketing Research, 44(1), 84-99. Cheema, A., & Soman, D. (2008). The effect of partitions on controlling consumption. Journal
of Marketing Research, 45(6), 665-675.
Chernev, A., & Chandon, P. (2010). Calorie Underestimation Biases in Consumer Choice. In R. Batra, P. A. Keller, & V. J. Strecher (Eds.), Leveraging Consumer Psychology for Effective
33
Cecchini, M., & Warin, L. (2016). Impact of food labelling systems on food choices and eating behaviours: a systematic review and meta‐analysis of randomized studies. Obesity reviews, 17(3), 201-210.
Churchill, S., & Jessop, D. C. (2011). Too impulsive for implementation intentions? Evidence that impulsivity moderates the effectiveness of an implementation intention intervention. Psychology and Health, 26(5), 517-530.
Ello-Martin, J. A., Ledikwe, J. H., & Rolls, B. J. (2005). The influence of food portion size and energy density on energy intake: implications for weight management. The American journal
of clinical nutrition, 82(1), 236S-241S.
Evans, D. (2011). Blaming the consumer–once again: the social and material contexts of everyday food waste practices in some English households. Critical Public Health, 21(4), 429-440.
Food and Agriculture Organization of the United Nations. (2013). Food Wastage Footprint:
Impacts on Natural Resources: Summary Report. FAO.
Davis, B., Payne, C. R., & Bui, M. (2016). Making small food units seem regular: How larger table size reduces calories to be consumed. Journal of the Association for Consumer
Research, 1(1), 115-124.
Diliberti, N., Bordi, P. L., Conklin, M. T., Roe, L. S., & Rolls, B. J. (2004). Increased portion size leads to increased energy intake in a restaurant meal. Obesity research, 12(3), 562-568. Division of Nutrition and Physical Activity (2006) Research to Practice Series No. 2: Portion
Size. Atlanta, GA: Centers for Disease Control and Prevention.
Faulkner, G. P., Pourshahidi, L. K., Wallace, J. M., Kerr, M. A., McCrorie, T. A., & Livingstone, M. B. E. (2012). Serving size guidance for consumers: is it effective?. Proceedings
of the Nutrition Society, 71(4), 610-621.
34
Graham-Rowe, E., Jessop, D. C., & Sparks, P. (2014). Identifying motivations and barriers to minimising household food waste. Resources, conservation and recycling, 84, 15-23.
Geier, A. B., Rozin, P., & Doros, G. (2006). Unit bias: A new heuristic that helps explain the effect of portion size on food intake. Psychological Science, 17(6), 521-525.
Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis:
A regression-based approach. Guilford Publications.
Herman, C. P., Polivy, J., Pliner, P., & Vartanian, L. R. (2015). Mechanisms underlying the portion-size effect. Physiology & Behavior, 144, 129-136.
Institute of Grocery Distribution (2008) Portion Size: A Review of Existing Approaches. England: Institute of Grocery Distribution.
Kerameas, K., Vartanian, L. R., Herman, C. P., & Polivy, J. (2015). The effect of portion size and unit size on food intake: Unit bias or segmentation effect?. Health Psychology, 34(6), 670. Ledikwe, J. H., Ello-Martin, J. A., & Rolls, B. J. (2005). Portion sizes and the obesity epidemic. The Journal of nutrition, 135(4), 905-909.
Lewis Jr, N. A., & Earl, A. (2018). Seeing more and eating less: Effects of portion size granularity on the perception and regulation of food consumption. Journal of personality and
social psychology, 114(5), 786.
Oldham-Cooper, R. E., Wilkinson, L. L., Hardman, C. A., Rogers, P. J., & Brunstrom, J. M. (2017). Presenting a food in multiple smaller units increases expected satiety. Appetite, 118, 106-112.
Pelham, B. W., Sumarta, T. T., & Myaskovsky, L. (1994). The easy path from many to much: The numerosity heuristic. Cognitive Psychology, 26(2), 103-133.
Powell, A. (2013). The whys of rising obesity. Retrieved from
https://news.harvard.edu/gazette/story/2013/09/the-whys-of-rising-obesity/ (accessed 26 June 2018).
35
Robinson, E., & Kersbergen, I. (2018). Portion size and later food intake: evidence on the “normalizing” effect of reducing food portion sizes. The American journal of clinical
nutrition, 107(4), 640-646.
Rolls, B. J., Morris, E. L., & Roe, L. S. (2002). Portion size of food affects energy intake in normal-weight and overweight men and women. The American journal of clinical nutrition, 76(6), 1207-1213.
Schwartz, J., Riis, J., Elbel, B., & Ariely, D. (2012). Inviting consumers to downsize fast-food portions significantly reduces calorie consumption. Health Affairs, 31(2), 399-407.
Savage, J. S., Haisfield, L., Fisher, J. O., Marini, M., & Birch, L. L. (2012). Do children eat less at meals when allowed to serve themselves? The American journal of clinical nutrition, 96(1), 36-43.
Scott, M. L., Nowlis, S. M., Mandel, N., & Morales, A. C. (2008). The effects of reduced food size and package size on the consumption behavior of restrained and unrestrained eaters. Journal of Consumer Research, 35(3), 391-405.
Senauer, B. (2003). The Obesity Crisis: Challenge to the Food Industry. Food Industry Center, University of Minnesota.
Vandenbroele, J., Zlatevska, N., Van Kerckhove, A., & S. Holden, S. (2017). On the overconsumption of food portions: is the problem in the size or the number? Presented at The La Londe Conference.
Van Kleef, E., Kavvouris, C., & van Trijp, H. C. (2014). The unit size effect of indulgent food: How eating smaller sized items signals impulsivity and makes consumers eat less. Psychology
& health, 29(9), 1081-1103.
Wansink, B., Just, D. R., & Payne, C. R. (2009). Mindless eating and healthy heuristics for the irrational. American Economic Review, 99(2), 165-69.
36
Wansink, B., & Van Ittersum, K. (2013). Portion size me: Plate-size induced consumption norms and win-win solutions for reducing food intake and waste. Journal of Experimental
Psychology: Applied, 19(4), 320.
World Health Organization. (2018, February) Obesity and overweight. Retrieved from http://www.who.int/mediacentre/factsheets/fs311/en/ (accessed 26 June 2019).
37
APPENDIX 1
In the unportioned condition, participants could select their own portion by serving the desired number of treacle-waffles on a plastic plate. The offered amount of treacle-waffles was held the same for every participant (6 for regular units and 25 for small units).
Condition 1: Unportioned food with regular units Condition 2: Unportioned food with small units
APPENDIX 2
In the pre-portioned condition, participants were offered a fixed amount of treacle-waffles (2 units in the regular unit size condition and 10 units in the small unit size condition).
Condition 3: Pre-portioned food with regular units Condition 4: Pre-portioned food with small units
38
APPENDIX 3
The short movie about the Dutch history “A royal tour - 200 years Kingdom of the Netherlands” has been taken from YouTube (https://www.youtube.com/watch?v=9AcGhjBtp3Y&t=244s). The video guide viewers past the most beautiful and important sites in Holland that are associated with the royal family and give the visit a royal touch. Viewers will be seeing beautiful palaces, such as Paleis Noordeinde and Huis ten Bosch, where the royal family lived and worked. And they will also see the small bakery where King Willem-Alexander got his sandwich as a little boy.
39
APPENDIX 4
In the “Pick your cookies” task participants were asked to put together a portion of gingerbread in a digital environment. Participants could select their own portion by dragging the desired number of gingerbread on a virtual plate. The amount of gingerbread with which participants could drag varied between conditions.
The unportioned condition with small units The unportioned condition with regular units
40
The large pre-portioned condition with small units The large pre-portioned condition with regular units
APPENDIX 5
In order to make the “Pick your cookies” task more realistic, we visualised the amount of gingerbread that participants could drag at a time (dependent on the condition) under a glass bell next to the computer screen of each cubicle.
The unportioned condition with small units The unportioned condition with regular units
41
The small pre-portioned condition with small units The small pre-portioned condition with regular units
The large pre-portioned condition with small units The large pre-portioned condition with regular units
