More in boxes, less in the bin? The Effect of Meal-kit Services on Food Waste

(1)

More in boxes, less in the bin? The Effect

of Meal-kit Services on Food Waste

Rijksuniversiteit Groningen

Faculty of Economics and Business

MSc Marketing: Marketing Management

Master’s Thesis

17

th

of June 2019

Research topic:

From the shopping cart into the trash bin:

Understanding and preventing food waste.

Author:

T.Q. Melching

Student number: S3751945

thommelching@gmail.com

(2)

Abstract

Food waste has become an increasingly important problem in modern society due to its economic, environmental, and societal consequences. Globally it was estimated that around a quarter of the total produced food supply is wasted, and about half of this total wastage of the produced food supply comes from the household-level. However, academic literature on how to prevent or reduce food waste on the household-level is scarce. Therefore, this research study examines whether a meal-kit service could reduce the amount of food wasted by households. Furthermore, it was studied whether three moderating variables (cooking skills, planning ability, and perceived time-pressure) would influence the potential success of a meal-kit service in reducing food waste. Participants in this study filled in two questionnaires, recorded their food waste quantities for two weeks with a food waste diary, and received a meal-kit service during one of those weeks. Surprisingly, the results indicated that there was a significant increase rather than the expected decrease in the amount of food waste when the meal-kit service was used. Possible explanations for this surprising result are the small sample size of this research study, the short duration, and that a meal-kit service was still an unknown concept for participants not leading to an immediate positive effect. The proposed moderators had no significant influence on this main result following the analysis conducted. However, the directional tests found higher cooking skills and perceived time-pressure, and lower planning ability associated with the weakened effect of the meal-kit service increasing food waste. Future research is necessary in order to further understand the true effect of meal-kit services on food waste, verify the main result of the meal-kit service on food waste, and investigate whether the proposed moderators really affect its success.

Keywords: food waste, household food waste, meal-kit services, meal-kits, cooking skills,

(3)

Acknowledgements

(4)

List of figures, tables, and appendices

Figure 1: Conceptual model ... 19

Figure 2: Plot of the main effect of the meal-kit service on food waste... 36

Figure 3: Johnson-Neyman plot of cooking skills on meal-kit service ... 36

Figure 4: Johnson-Neyman plot of planning ability on meal-kit service ... 36

Figure 5: Johnson-Neyman plot of time-pressure on meal-kit service ... 37

Table 1: Cooking skills scale ... 29

Table 2: Planning ability scale ... 29

Table 3: Time-pressure scale ... 30

Table 4: Satisfaction meal-kit service: content-wise scale ... 30

Table 5: Satisfaction meal-kit service: portion sizes scale ... 31

Table 6: Descriptive statistics ... 34

Table 7: Overview of the tested hypotheses and additional findings ... 41

Appendix 1: Photograph of the contents of the delivered meal-kit service... 53

(6)

1. Introduction

Food waste is a serious problem in modern society. Based on a study conducted by Kummu et al. (2012), approximately a quarter of the produced food supply is lost within the food supply chain. In higher income countries this is an even bigger issue as in Europe 29% of food supplies are wasted, and in North America & Oceania 32% of food supplies are wasted. Furthermore, when looking at the different steps within the food supply chain it appears that in both these areas most food is wasted within the (final) consumption stage. Of this quarter of the produced food supply that is ultimately wasted, a tremendous amount of over 50% is wasted at the household-level (Stenmarck et al., 2016). There is thus a lot of potential to improve in this particular stage.

There are multiple negative consequences related to the waste of food, such as economic consequences, as well as environmental and societal. According to a factsheet published by the Dutch Voedingscentrum (food centre), the average consumer wastes around 41 kilograms of food per individual person every year, which corresponds to about €145 (Van Dooren & Mensink, 2018). On a global level, it was estimated by the Food and Agricultural Organisation of the United Nations that food waste leads to $750 billion of losses each year, concerning all sectors except seafood (FAO, 2013). It can thus be concluded that food waste has a negative economic impact both on an individual, as well as a global level.

Food waste is also associated with negative environmental impacts such as the wasteful use of resources like water, cropland, and fertilisers (Kummu et al., 2012; Stancu, Haugaard, & Lähteenmäki, 2016; Vanham, Bouraoui, Leip, Grizzetti, & Bidoglio, 2015). Furthermore, it contributes to the great increase in emissions of greenhouse gases (Bio Intelligence Service, 2010; Stancu et al., 2016; WRAP, 2009). Lastly, food waste also has a huge impact on society as a whole, as the world remains to face challenges in feeding the increasing global population (Godfray et al., 2010). All in all, it can be argued that food waste is a big issue with negative consequences, and thus needs to taken seriously.

(7)

meals beforehand might lead to a lot of leftovers, thus consumers have to balance between the ideals of keeping these leftovers of meals and using them for subsequent meals, or opt for more convenient choices in which leftovers are (mostly) non-existent, such as ready-to-eat meals. Wansink, Brasel, and Amjad (2000) describe this phenomenon of being left with a bunch of ingredients due to it being used for only one specific recipe. These leftovers are subsequently very often wasted, although in an ideal situation they could still be used. Furthermore, Hebrok and Boks (2017) describe that differences in household sizes and lifestyles of consumers may influence the probability of consumers opting for a more convenient choice, and that the level of food risk that consumers perceive in a food product may also lead to them preferring convenient choices in order to regain security and manage the potential food risk.

A possible way to counter food waste on the household-level may be meal-kit services. Several meal-kit service companies such as Hello Fresh, Blue Apron (exclusively in the United States), Marley Spoon, and others have popped up in the last few years. These companies provide a home-delivered box to consumers, filled with (almost) all the necessary ingredients and fitting quantities to cook three to five (healthy) meals every week, effectively preventing the problem of leftovers as described by Wansink et al. (2000). Along with the ingredients, detailed recipes are also included which provide the consumer with the necessary instructions. The meal-kit service effectively takes care of planning-related tasks and may provide more feelings of security relating to food risk than when food is bought from a regular supermarket. It thus provides convenience to a consumer, with one consumer even stating that someone else planning out their meals was “incredibly liberating” (Whyte, 2018).

Because of convenience gaining a more prominent role for households and individual consumers, meal-kit services may thus provide a way in saving time for a consumer by outsourcing planning-related obligations. Subsequently, this saved time can then be invested in other tasks which the consumer might feel is more valuable. It can thus be argued that meal-kit services relate to the convenience needs of consumers because of it saving the precious time of consumers. Furthermore, the meal-kit services might also appeal to the economic incentives of consumers since it only provides the necessary ingredients, which may be considered a benefit since the portion sizes offered by stores are most often not the portion size that is required for a specific meal, thus leaving the consumer to buy more in quantity and spend more while they only need a smaller portion of the product(s).

(8)

product purchases as well as a lack of cooking skills are both factors that are associated with higher levels of food waste, while higher levels of perceived time-pressure may lead to consumers planning their meals and potential leftovers less careful which ultimately would increase food waste (Gustavsson, Cederberg, Sonesson, Van Otterdijk, & Meybeck, 2011; Stancu et al., 2016; Stefan, van Herpen, Tudoran, & Lähteenmäki, 2013). Because of the influence these individual capabilities (in)directly have on household food waste, it is argued that they may also influence the potential success a meal-kit service might have in reducing household food waste.

Therefore, the following research question is derived: How does the usage of a meal-kit

service influence food waste? And how do cooking skills, planning ability, and time-pressure influence the effectiveness of a meal-kit service? The objective of this research study is thus to

examine whether meal-kit services can potentially impact and reduce food waste and whether individual cooking skills, planning ability, and perceived time-pressure impact this influence. This research study focuses on a group of participating households in a within-subjects research design, where their food wasting habits are observed during a week without a meal-kit service, and during a week with a meal-kit service.

(9)

2. Literature review

The following chapter explores what research has already been conducted on the topic of food waste. This literature review first begins with an attempt to define the main concept that is food waste, and what meaning this concept holds in this research study. Subsequently, drivers of food waste are analysed and discussed, followed by a section dedicated towards already existing research about the effects of meal-kit services and ready-made meals on food waste. To conclude, the last section describes which individual capabilities are argued to influence food waste.

2.1 Defining food waste

As mentioned, it is very important to first define what food waste actually entails for this research study. To start, Thyberg and Tonjes (2016) argue that the definitions of food waste are not universally agreed upon, which would subsequently make it harder to study and quantify. The Food and Agricultural Organization of the United Nation (FAO) define it as follows: “Food

loss is defined as ‘the decrease in quantity or quality of food.’ Food waste is part of food loss and refers to discarding or alternative (nonfood) use of food that is safe and nutritious for human consumption along the entire food supply chain, from primary production to end household consumer level” (FAO, 2014). The definition from the FAO thus does not distinguish

between the different stages involved in the food waste process.

On the other hand, the definition offered by Thyberg and Tonjes (2016) does make a distinction between different stages, as well as between food loss and food waste. They argue that food loss is the “decrease in edible food mass throughout the part of the supply chain that

specifically leads to edible food for human consumption”, thus meaning that food loss is mainly

involved in the production, post-harvest, and processing stages (Thyberg & Tonjes, 2016). Food waste, on the other hand, is defined as “food which was originally produced for human

consumption but then was discarded or was not consumed by humans. Includes food that spoiled prior to disposal and food that was still edible when thrown away”, thus mainly

involving the retail and consumer stages (Thyberg & Tonjes, 2016).

Cicatiello, Franco, Pancino, and Blasi (2016) also highlight the importance of the distinction between food loss and food waste. They argue that food loss can be defined as either a “qualitative or quantitative drop in the food supply due to a reduced nutrient value of the food

(10)

edible item is lost in these processes, subsequently meaning that it is directly linked to human action (Cicatiello et al., 2016).

For the sake of this research study, it makes the most sense to focus on the household component of food waste. Food waste in this research study is subsequently for the remainder of this thesis defined as: “Food items that were produced for human consumption, but that were

ultimately discarded at the household consumer level. This includes both food that was still safe to consume, and food that was no longer safe to consume due to it being spoiled”. This combines

the notion of Cicatiello et al. (2016) that it is directly linked to human actions, and the concept of Thyberg and Tonjes (2016) that both edible and inedible food discarded is considered as waste. However, this definition distinguishes itself regarding its conception of food that is not consumed by humans. It is argued that this food cannot be regarded as wasted, because although it is not consumed by its intended target group (humans), it is still in fact consumed by another living being (for example animals) and therefore nutritious value is not wasted.

2.2 What are the drivers of food waste?

Following the proposed definition of food waste, it is essential to look at the different drivers of food waste. These drivers are discussed from several different perspectives. They may describe the larger, more on community-level drivers of food waste, but they may also describe the more individual, or household-level drivers. Looking towards a larger, community-perspective, a number of drivers can be addressed. Thyberg and Tonjes (2016) define four different factors that drive food waste from an institutional, commercial or residential perspective; 1) the modernisation of food systems, 2) cultural factors, 3) policies concerning food waste generation, and lastly 4) socio-demographic factors, and generally take a more broad viewpoint regarding the drivers of food waste rather than looking at the household-level.

(11)

Khwaja, 2004). One way that this could turn out is that there is a shift from staple food products which are high in starch contents to a diet based more on food high in protein and energy such as meat, dairy, and fish (Parfitt, Barthel, & Macnaughton, 2010). This can be considered a global trend, in which most Western, developed countries have already undergone this shift, and some Asian countries are also following this ‘Western’ diet shift (Pingali & Khwaja, 2004). Lastly, the increased globalisation of the last few years is also argued to have a lot of impact on the shift from local to global food sources discussed before, particularly because of its share in the increase of food available from a wide variety of sources besides local (Pingali & Khwaja, 2004). Subsequently, this leads once again to an emotional disconnection between consumers and the food sources, as well as diet diversification away from local food products (Thyberg & Tonjes, 2016).

The effect that cultural differences have on food, eating, and nutrition habits, as well as the subsequent waste generation, should not be underestimated, as attitudes and habits based on culture shape our behaviour (Rozin, 2005). This is exemplified by for instance the different parts of food that are considered inedible in some cultures (e.g. cartilage), whereas in other cultures they are considered to be edible and are thus consumed nonetheless (Strasser, 2000). Furthermore, some cultures may have little food culture, such as the United States and Australia, whereas other countries like France do have deep food cultures which have been formed during their rich history (Gatley, Caraher, & Lang, 2014). This is demonstrated in the difference of some cultures (e.g. United States) preferring abundance or large quantities of food, whereas others (e.g. France) prefer quality over quantity (Rozin, 2005). Cultural differences regarding shopping habits are also argued to influence food waste generation. Aspects like number of products bought, frequency of shopping trips, and the amount bought with intention to be stored rather than immediately consumed may all differ between different countries and cultures (Neff, Spiker, & Truant, 2015). Consumers from developing countries buy smaller amounts of food and go on shopping trips more frequently than those in developed countries, which may subsequently reduce food waste generation (Pearson, Minehan, & Wakefield-Rann, 2013). Lastly, the cross-cultural difference in the mindset of food retailers regarding their perception of food waste being a major problem or not may ultimately also affect waste generation due to it affecting shopping habits of consumers, as well as food waste habits of the stores (Mena, Adenso-Diaz, & Yurt, 2011).

(12)

have to be met according to these policies (initiated by the United States Department of Agriculture). However, as a negative side-effect, they are suggested to produce more waste due to children not willing to consume the newly served food products, which is also the reason why they are criticised (Jalonick, 2014). Actual research about this has produced mixed results, as one study found a severe increase in food discarded by students, one study found both substantial amounts of food wasted both prior and after the diet revision, and another study found reduced values of food waste (Byker, Farris, Marcenelle, Davis, & Serrano, 2014; Cohen, Richardson, Parker, Catalano, & Rimm, 2014; Schwartz, Henderson, Read, Danna, & Ickovics, 2015). Another legal policy that influences food waste generation is the inclusion of expiration dates and other methods to ensure food safety. While these may reduce negative consequences of consumers consuming unsafe food products, some consumers may also neglect their common sense and use expiration dates as strict guidelines, subsequently leading to more food wasted (Watson & Meah, 2012).

Food wasting behaviours may also be explained by certain socio-demographic factors such as age, family composition or household size, and income (Pearson et al., 2013; Thyberg & Tonjes, 2016). Age is argued to be a major indicator of food waste, with less amount of food products wasted by older consumers rather than younger (Stancu et al., 2016). It is imaginable that in the future when current young consumers grow older and change regarding their lifestyle, attitudes, and knowledge capacities that they would waste less, but this is still unknown (Pearson et al., 2013). Larger households are also argued to waste more in absolute terms, although this is expected due to the higher number of people involved (Jörissen, Priefer, & Bräutigam, 2015; Stancu et al., 2016), but when household size increases then the amount of food wasted per capita decreases (Williams, Wikström, Otterbring, Löfgren, & Gustafsson, 2012). In contrast, it is thus assumed that in relative value the one-person households are the most wasteful. Lastly, consumers with higher incomes are expected to waste more than consumers with lower incomes, although ultimately food is wasted across all different income classes (Baker, Fear, & Denniss, 2009; Stancu et al., 2016; Thyberg & Tonjes, 2016).

(13)

consumers value their food would be their income and age (Parfitt et al., 2010). Secondly, the lack of awareness of how much food an individual consumer is wasting, and simultaneously having an attitude that food waste is inevitable and thus acceptable within modern society might also influence the actual amount of food wasted (Lyndhurst, 2007).

The next driver discussed by Hebrok and Boks (2017) is regarding how food stocks are managed in households. Two related components are the storage and packaging of food. It is argued that storage is most crucial when trying to reduce food waste (Farr‐Wharton, Foth, & Choi, 2014). Storage may greatly influence the duration of the edibility of food. Refrigerators and freezers are two examples of how food can be stored. They enable convenience, freshness, and reduce food risk (Waitt & Phillips, 2016). Packaging also has an important role in protecting and preserving products, while simultaneously informing and seducing consumers (Hebrok & Boks, 2017). A combination of excessive packaging sizes and the expiration date may also lead to disposal of products, because consumers are ultimately unable to completely finish the product before the expiration date, and thus prioritise reducing food risk rather than reducing food waste as mentioned before (Lyndhurst, 2011). Furthermore, date labelling can be a method to create trust among consumers, and subsequently reduce food risk (Watson & Meah, 2012).

2.3 Convenience, planning, and food waste

The last driver discussed by Hebrok and Boks (2017) is the importance of convenience to consumers. The authors relate this to three different aspects, namely; 1) household composition and lifestyles, 2) planning and subsequent leftovers, and 3) food risk. It is argued that consumers constantly try to minimise inconvenience nowadays (Graham-Rowe, Jessop, & Sparks, 2014), and that this ultimately causes trade-offs between the ideals (such as accurately planning food and leftovers, minimising food waste, and eating healthy) and opting for more convenient options that might not be able to meet all of these ideals (Bava, Jaeger, & Park, 2008).

(14)

Olivier (2013), who found that only one-third of their interviewees used shopping lists to plan in advance, and these shopping lists only acted as reminders for potentially forgotten items rather than detailed plans. Simultaneously, a reduced willingness to invest time might also make less efficient planners more likely to prefer convenient options in order to save time. Another result of a lack of planning is overstocking and leftovers. Leftovers are argued to be a category of food waste that consumers are least aware of (Lyndhurst, 2007), and are often wasted due to laziness and safety issues (Mavrakis, 2014). Lastly, food risk can be of influence as the perceptions of consumers of whether food is still edible or not can lead to good food ultimately being wasted. Convenient options often include food products that are ready-to-eat and are thus not stored for longer periods of time at the home of a consumer, ultimately reducing perceived food risk. Although food waste is not something taken lightly by consumers, it is suggested that people would rather dispose of food that they have doubts about and thus avoid risk rather than avoid food waste (Graham-Rowe et al., 2014).

In a study by Stefan et al. (2013) among Romanian consumers, some background factors were deduced. Aspects like perceived behavioural control over food waste, subjective norms, and lack of concern about the consequence of food waste behaviour were found to have an indirect effect on food being wasted, while both shopping and planning routines were very important direct antecedents of food being wasted. Shopping routines often led to excessive purchasing of food products, and subsequently, are positively associated with food waste. Planning routines, on the other hand, lead to more strict control over what food products are bought, as well as excessive purchasing of food products. However, these routines initially are heavily influenced by the moral attitudes of the consumers towards food waste (Stefan et al., 2013). It could thus be argued that if planning and shopping routines are simplified, this may prevent the initial negative effect moral attitudes can have in increasing the amounts of food being wasted.

(15)

way better able to properly plan meals ultimately leading to less food wasted. Possible explanations for this could be that this age group values food in a different way due to their past (e.g. food scarcity during World War II), or that they have more spare time to plan their meals and thus ultimately waste less food.

2.4 Meal-kit services and ready-made meals influencing food waste

Although multiple authors suggest an influence of meal-kit services on the amount of food wasted by households, empirical research in this field is still lacking. A 2016 article by the New York Times already suggested that meal-kit services can decrease food waste, gasoline consumption, and the packaging related waste that is involved with regular grocery shopping (Severson & Child, 2016). Furthermore, for people living in remote areas, a meal-kit might be the only viable option in getting certain ingredients without having to stop at supermarkets that are far away from the consumers’ homes. Romani, Grappi, Bagozzi, and Barone (2018) similarly agree that meal-kit services may reduce food waste due to all ingredients being measured out, subsequently minimising leftovers, but also that empirical verification to back up this claim is still necessary.

In contrast, a study by Mallinson, Russell, and Barker (2016) reported that higher amounts of food waste were positively associated with the purchase of convenience food like ready-made meals. It is imaginable that for these consumers a ready-made meal might be one of the only possible options in terms of available time in actually consuming a meal, therefore wasting other food products that they already had at home. This does, of course, leave a lot of potential, as the food from the convenient option itself is not wasted, but rather the remaining products that were already in stock (Mallinson et al., 2016).

2.5 Individual capabilities influencing food waste

(16)

McGowan et al. (2017) discuss cooking skills and food skills, and more importantly in what way these two terms differ. Cooking skills are generally person-centred, and besides mechanical skills also include perceptual understandings of the properties of food and how it reacts when it is prepared (Short, 2003). Hartmann, Dohle, and Siegrist (2013) offer some empirical evidence of the importance of cooking skills on actual consumption. They found that cooking skills are positively correlated with vegetable consumption, but negatively correlated with convenience food consumption (such as ready-made meals, or microwave meals). In a study by Mallinson et al. (2016), it was also shown that a cluster with high cooking skills was associated with the lowest amount of food wasted compared to the other cluster, further suggesting the relationship between individual cooking skills and food wasted at the household-level.

However, the term food skills also incorporates other related concepts besides cooking skills, like a consumer’s ability of meal-planning, ingredient shopping, and food budgeting. It is thus a broader term defined as “the ability to purchase, prepare and cook food materials

using available resources, to produce well-balanced and tasty meals, appropriate to the age and needs of the individuals consuming them” (Fordyce-Voorham, 2009). Fordyce-Voorham

(2009) thus essentially argues that the term cooking skills is just one component of food skills, in addition to the aspects of the ability of meal-planning, ingredient shopping, and food budgeting. In a later study, it was argued that food skills consists of four themes (Fordyce-Voorham, 2011). Information is regarded as the first theme, which would describe the information that is necessary for an individual to be found in a variety of different sources. Second, knowledge is related to the personal capability of being aware and able to understand the information offered in order to design a meal. Third, the skills of an individual describe their capability of applying the knowledge and information themes into actual practice. Lastly, resources are about all other human (such as energy or motivation), and non-human (time-pressure) assets that influence an individual’s capability in creating a good meal.

(17)

collection of inter-related knowledge, skills and behaviours required to plan, manage, select, prepare and eat food to meet needs and determine intake” (Vidgen & Gallegos, 2014).

Multiple authors build upon this definition in subsequent studies. Truman, Lane, and Elliott (2017) argue that food literacy consists of six distinct domains; skills and behaviours, food/health choices, culture, knowledge, emotions, and food systems. It is further argued that these individual domains suggest a need for multiple sub-definitions of food literacy, rather than one overarching main definition. This would then subsequently allow for differentiated study designs, and a way to implement interventions to examine the proficiency of the different domains of food literacy.

Perry et al. (2017) similarly argue that food literacy is a multi-faceted concept, consisting of different aspects. They argue that food literacy consists of; food and nutrition knowledge, food skills, self-efficacy and confidence, ecologic (beyond self) factors, and food decisions. Combining these aspects may result in a sort of step-by-step approach, where a consumer should first possess the knowledge, followed by the skills, the self-efficacy and confidence to act on these skills, being able to deal with factors beyond yourself, and ultimately making the right food decisions based on all prior factors.

Cullen, Hatch, Martin, Higgins, and Sheppard (2015) offer an even more simplified explanation of the term by arguing that food literacy is a concept allows an individual to on the one hand be able to deal with individual food skills, and on the other be able to deal with community food security. They argue that for one to be food literate, one cannot just depend on a food secure community or just on personal food skills. Lastly, it is theorised that increasing food literacy will, in turn, increase health and well-being (Cullen et al., 2015).

These differing conceptualisations of these authors of the components of food skills and food literacy further underline the difficulty in objectively defining and measuring these concepts (McGowan et al., 2017). As shown before in this review, all authors have different understandings of what aspects are included in a term like food skills or food literacy. Vaitkeviciute, Ball, and Harris (2015) also showed in their systematic review that most studies until then only use one aspect (such as cooking confidence, cooking skills, or nutrition knowledge) to measure the overarching term food literacy.

(18)

(19)

3. Conceptual framework

This chapter is dedicated towards introducing the conceptual model (see Figure 1), and further elaborates upon its specific components and related hypotheses. As shown, this conceptual framework has two main propositions. First, it suggests that the usage of a meal-kit service leads to a reduction in food waste compared to situations where a meal-kit service is not used, due to its incorporated benefits of superior planning, reduction of food risk, and subsequent reduction of leftovers due to tailored package sizes (Hebrok & Boks, 2017). Secondly, the conceptual model argues that there are three moderators that influence the effectiveness of a meal-kit service, namely; cooking skills, planning ability, and perceived time-pressure. These moderating variables were chosen because of them being person-centred, just like the food preparation process which this research study focused on.

The meal-kit service is argued to effectively reduce food waste when consumers are high in cooking skills, whereas a meal-kit service for people with low cooking skills may not reduce food waste (Mallinson et al., 2016). This was argued because people with lower levels of cooking skills might not be comfortable with the techniques that are necessary to prepare the meals. Secondly, for consumers with a high planning ability, the initial effects proposed in the main hypothesis may be mitigated (Hebrok & Boks, 2017). Lastly, consumers with high levels of perceived time-pressure who use the meal-kit service are expected to reduce their amount of food wasted more than consumers with low levels of perceived time-pressure (Dapkus, 1985). In the following sections, the hypotheses are further elaborated upon.

(20)

3.1 The effect of a meal-kit service on food waste

In the conceptual model, it is suggested that when a time-period without a meal-kit service is compared to a period with a meal-kit service, the amount of food wasted will be lower during the period that a meal-kit service is used. Food waste in this case thus refers to the definition stated earlier, that highlights that food wasted is considered at the household-level. Multiple explanations can be offered for this effect occurring.

Firstly, planning of food provisioning is often a difficult task (Hebrok & Boks, 2017). It is argued that inefficient planning is one of the main causes of food being wasted. Additionally, when household sizes increase or lifestyles are considered busier, it will become harder to plan effectively. Planning is essential, as it provides consumers with a better idea of the stock at hand, what products are still necessary, and reduces the probability of overbuying food products (Farr‐Wharton et al., 2014). Linking this to the meal-kit service, it can be argued that a meal-kit service offers to take care of all the planning tasks for the consumer, which could be considered a convenient option. Subsequently, it prevents the potential negative effects of inefficient planning occurring.

Secondly, the problem of the offered package sizes. This is related to the earlier proposed problem of overbuying. In conventional supermarkets, products are often offered in a small variety of package sizes, and these package sizes may not always correspond to the necessary amount of food products required for a meal. Without accurate planning in advance, the package sizes may thus lead to involuntary overbuying and subsequent leftovers. Leftovers may also be the result of buying products for specific occasions or meals, but not using them afterwards because the same occasion did not come up, or the same meal was not prepared a second time (Wansink et al., 2000). This is especially important considering that leftovers are argued to be very impactful and one of the categories of waste that consumers are very unaware of (Lyndhurst, 2007). The meal-kit services take care of this food waste generating problem because they always adjust the contents of the meal-kits to the number of portions subscribed for. This would then, in turn, lead to fewer leftovers, and less food waste from that aspect.

(21)

carefully selected and constructed by professionals, this might evoke feelings of being ‘in safe hands’. These reasonings ultimately led to the first hypothesis, which is as follows:

H1: The usage of a meal-kit service by households will reduce their amount of food wasted in a period compared to a period when the meal-kit service was not used.

3.2 Individual capabilities influencing the effectiveness of the meal-kit service

As mentioned before, there are multiple individual capabilities that may influence the amount of food being wasted when looking at the regular routines by consumers. In this section, three of these capabilities and their influence on whether or not a meal-kit service may reduce food waste are discussed, namely; cooking skills, planning ability, and perceived time-pressure.

3.2.1 Cooking skills

Cooking skills is analysed closely following the definition offered by Short (2003) as cooking skills being very person-centred and related towards an individual’s capability of preparing meals in a proper way. Studies relating the importance of cooking skills to food waste already exist, suggesting that a higher level of cooking skills are associated with less food being wasted (Mallinson et al., 2016). From this, two potential moderation effects could thus be argued.

On the one hand, consumers with lower levels of cooking skills using a meal-kit service could benefit more than consumers with higher levels of cooking skills due to the meal-kit service providing detailed recipes along with the ingredients. In this way, consumers with lower levels of cooking skills may be better able to prepare meals with the aid of the meal-kit service than they usually are, while consumers with higher levels of cooking skills are already able to prepare meals in an accurate manner, thus ultimately not wasting any more or less due to the usage of the meal-kit service.

(22)

with less willingness to experiment with cooking (Bava et al., 2008). Because the consumer is forced to experiment and act out of his or her comfort zone, it is thus argued that a satisfactory level of cooking skills are a necessity for reducing food waste through the usage of the meal-kit service. From this, the following hypothesis is derived:

H2a: The effect of meal-kit services on the amount of food wasted will be moderated by the level of cooking skills of the consumer, with the effect being greater (smaller) for consumers with a higher (lower) level of cooking skills.

3.2.2 Planning ability

Planning ability is argued to have a major direct effect on the amount of food wasted (Stefan et al., 2013). Planning routines may be expressed in more control over what food products are purchased, as well as the quantities purchased. Planning ability is subsequently defined as the level of control that consumers exert over their purchases and whether they think ahead when they are going to buy food products. A lack of planning may result in an increase in leftovers, which are a part of food waste that consumers are most unaware of (Lyndhurst, 2007). Therefore, planning ability is important since people with higher levels of planning ability are expected to waste less food as they are taking into account the food products that are already in stock, and plan their meals in advance accordingly.

As mentioned before, meal-kit services may provide benefits to consumers by liberating them from their obligations of planning (Whyte, 2018). However, it is highly imaginable that consumers who already do have the capability to plan their meals in advance accurately will not benefit from the meal-kit service (a lot). If consumers would be high in planning ability, it is unlikely that they will have to deal with a lot of leftovers, so the appropriate portion sizes is another advantage that consumers high in planning ability are expected not to benefit from as much. Conversely, the consumers that are low in planning ability are expected to benefit from the meal-kit service its offered planning advantages. The following hypothesis is thus derived:

H2b: The effect of meal-kit services on the amount of food wasted will be moderated by the level of planning ability of the consumer, with the effect being greater (smaller) for consumers with a lower (higher) level of planning ability.

3.2.3 Time-pressure

(23)

everything which is limiting their choices in what they can do with the remaining available time (Dapkus, 1985). This may subsequently result in inferior planning of meals, and less willingness to deal with possible leftovers due to the perceived time-pressure. Ultimately, higher levels of perceived time-pressure could thus lead to higher amounts of food wasted, as the feeling of having too little time results in individuals preferring other activities rather than focusing on efficient food management. Therefore, it is argued that time-pressure can have two different effects on the effectiveness of a meal-kit service.

Although a meal-kit service is designed to save time with the goal of a home-cooked meal in mind, it might still be the case that for some people the cooking process itself just takes too much time, or rather that they imagine that it takes too much time. The meal-kit service liberates the consumer from the planning-related obligations (Whyte, 2018), but simultaneously the burden that the accompanied recipes from the meal-kit service offer might be too much for the consumer to handle in terms of time. This could ultimately result in higher amounts of food wasted than when this consumer would just stick with options that save even more time, like ready-made meals.

However, it is also imaginable that especially consumers who are high in levels of perceived time-pressure would benefit from the meal-kit service. Not only are the consumers exempt from planning-related obligations due to the meal-kit service, but they also have to visit conventional supermarkets less often because all products to prepare their evening meals are delivered to their homes. In this way, the meal-kit service saves time because it takes care of the thinking of what to eat away from the consumer, as well as some of the trips from and to the conventional supermarkets. Based on these reasonings, it is expected that people who are high in perceived levels of time-pressure ultimately benefit more from a meal-kit service than those who are low in perceived levels of time-pressure. The following hypothesis is thus derived:

(24)

4. Methodology

The following chapter elaborates upon the methodology that is applied throughout this research study. This chapter first addresses the research design, followed by a section about the procedure of the field study. Next, it is described in what way the data was collected and what measures were used. To conclude this chapter, the plan of analysis is explained in detail.

4.1 Research design

This research study tries to explore whether the amount of food waste differs between two conditions, namely; one condition where the participating households are not using a meal-kit service, and one condition where the participating households are using a meal-kit service. The research study therefore utilises a within-subjects research design and compares the results of each participating household of each of the two conditions. A within-subjects design was chosen as it reduces the probability of individual differences affecting the results of this research study because all participating households can serve as their own control group. The research study is conducted in a field study setting. A field study was chosen as this was most likely to closely replicate the normal food wasting behaviour from the participating households, whereas a laboratory setting might have led to different results due to the participating households acting differently than they usually would in their respective environments.

A quantitative study was selected since the aim of the research study was to find out whether food waste behaviour would be influenced for a potentially large population, therefore a quantitative study would be more reliable and objective when making subsequent assumptions. The quantitative study allows for the use of statistics to form conclusions, and thus rules out subjectivity and is arguably better in finding robust results that can be applied to other settings. The aim of this research study is thus to determine whether a meal-kit service (IV) is able to reduce the amount of food wasted by households (DV). Additionally, this research aims to determine whether cooking skills, planning ability, and perceived time-pressure influence the effect of the meal-kit service on the amount of food wasted.

4.2 Procedure of the field study

(25)

moment would be informed through their stated e-mail address. A total of 30 households signed up to participate. After signing up, all households received a personal e-mail stating their respondent numbers, and already some information regarding the research study. Furthermore, they received a link to the first questionnaire and were asked to fill it in before the 4th of May 2019, as well as the informed consent form which they were asked to hand in as well before the same date.

Afterwards, all of the participating households received a set of materials which were necessary to participate in this field study. They received a personal letter explaining the research study process, two A3-sized waste diaries for seven days each, a numbered plastic bag in which they could put their remainders of the meal-kit service, and a printed copy of the informed consent form. They received this set of materials on the 1st of May 2019. The meal-kit services were offered by a regional meal-meal-kit service supplier called De Buurman. Because of that, this research study was limited to only use households that ranged from two to five people, and households were unable to choose their meals since De Buurman only offered three different meals each week. In Appendix 1 a photograph can be found of what a typical delivered meal-kit service had to offer in terms of content.

On the 4th of May 2019, the field study began, and the first 15 households also received their meal-kits on that day. The other 15 households received their meal-kits a week later, on the 11th of May 2019. It was decided to first have 15 households receive the intervention in the first week of the field study, and the remaining 15 households in the last week to prevent effects coming from the order in which the intervention was offered. All of the meal-kits were delivered to the participants’ homes. The field study ran from the 4th of May until the 17th of May 2019. During that period all participants were asked to keep track of their food waste diary for a total of 14 days. After the first week of the field study was over, the plastic bags of the first 15 households that received their meal-kit service were collected. Subsequently, a week later the other plastic bags were collected as well as all the waste diaries. On the last day of the field study, all participating households received an e-mail with the link to the second questionnaire.

4.3 Data collection

(26)

instructed to be filled in by the person who was (primarily) responsible for the eating-related household tasks of their household (e.g. cooking and doing groceries). All of the data was collected between the 25th of April and the 20th of May 2019. Data from the waste diaries as well as the plastic bags were collected at the homes of the participating households, the data from the questionnaire was collected digitally through Qualtrics’ research software. Digital distribution was chosen for the questionnaires because this made it easier to collect responses, as well as making it easier to process after every response was collected.

4.3.1 Food waste diary

In order to measure food waste of the participants for a total period of 14 days, a food waste diary was used. The diary consisted of two major parts; 1) the frequency of purchasing food, and 2) the quantities of wasted food products. In the first part, the participating households could indicate on which day, and how often they 1) ordered a takeaway meal, 2) went out for dinner, 3) did large grocery shopping, and 4) did small (additional/forgotten) grocery shopping.

Subsequently, in the second part households had to indicate the quantities of wasted food products for every individual day. This part suggested a total of 24 different product categories, and one additional ‘other’ category if participants were unsure of where to put their wasted product. A distinction was made between liquid food waste (in millilitres) and solid food waste (in grams). Additionally, standard sizes of teaspoons, tablespoons, serving spoons, as well as standard portion sizes were suggested throughout the diary in order to get participants to be as accurate as possible while noting down their wasted quantities. An example of a filled in food waste diary for one of the weeks can be found in Appendix 2.

4.3.2 Plastic bag

(27)

4.3.3 Questionnaires

As mentioned, all participating households were required to fill in two different questionnaires; one before the start of the 14-day period in which they had to keep track of their food waste diary, and one after this 14-day period. It was chosen to administer two separate questionnaires because this made it possible to first gain some information about the participating households, and afterwards, ask questions about the experiences of participants regarding the meal-kit service. Furthermore, distributing the second questionnaire also allowed for adding additional questions in case this was necessary. Both questionnaires were required to be filled in by the person (primarily) responsible for the eating-related household tasks of their household (e.g. cooking and doing groceries).

The first questionnaire consisted of 32 questions in total. First of all, every respondent was once again informed that the questionnaire was supposed to be filled in by the person (primarily) responsible for the eating-related household tasks of their household. Afterwards, their participant number was asked, as well as questions regarding their cooking skills, cooking enjoyment, planning ability, portioning ability, and perceived time-pressure. To finish this questionnaire, some demographic statistics were asked, namely; gender, educational level, profession, average working hours per week, and the amount and age of the people in their household.

The second questionnaire consisted of 32 or 33 questions, depending on the previous experience of the respondent with meal-kit services. As with the first questionnaire, every respondent was first informed that the questionnaire was supposed to be filled in by the person (primarily) responsible for the eating-related household tasks of their household and to fill in their respondent number. Afterwards, questions regarding cooking skills, planning ability, satisfaction regarding the products and offered portioned sizes by the meal-kit service, and some questions for the supplier were asked. To finish off this questionnaire, the age of the respondent, the income level of the household, and whether the household had pets were asked.

4.4 Measures

(28)

4.4.1 Food waste

As mentioned, the dependent variable food waste was measured via two different measurement instruments, namely; the food waste diary, and the numbered plastic bags that were distributed to all participating households. First of all, the food waste diary allowed for detailed, per product category information on what quantities of food were wasted by each household (Koivupuro et al., 2012). It focuses mainly on the avoidable food waste categories, thus products that were edible before thrown away and could have been prevented.

The food waste diaries are self-assessed measures by the participating households, thus meaning that participants are required to measure and indicate their own food waste. However, this self-assessment measure can prove to be very demanding for participants, and therefore the results should be critically assessed as they are likely to be indicative estimates of the true amounts of food waste (Koivupuro et al., 2010). It is imaginable that participants have subconsciously or consciously forgotten to record some of their products that they have wasted either due to self-presenting bias or due to actually forgetting about it altogether. In the case that participants only wrote down the product that they threw away rather than the quality, a standardised measure was obtained from the Dutch food centre’s Eetmeter, which is a diary where one can keep track of all food products that he/she consumes (Voedingscentrum, 2019). The sum of the total waste in grams/millilitres indicated for each week of the food waste diaries is considered the measure of the dependent variable food waste for this research study. From this total waste in grams/millilitres, two different measures were also derived, namely; 1) the sum of the total solid food waste in grams which was calculated by deducting the liquid waste from the total, and 2) the waste in grams per capita which was calculated as the sum of the total waste in grams/millilitres for each week relative to the number of people in the household.

(29)

4.4.2 Cooking skills

Cooking skills was measured by using a validated scale as proposed by Poelman et al. (2018). The respondents were asked to rate six statements on a seven-point Likert scale, ranging from 1 “completely disagree” to 7 “completely agree”. The internal consistency was sufficient following the executed reliability analysis with α = 0.773 after the second (reversed) item was deleted. The items that were analysed can be found in Table 1 below.

Table 1: Cooking skills scale

adopted from Poelman et al. (2018)

Cooking skills Moderator

*Excluded for further analysis

1. I am able to prepare fresh vegetables in different ways (for example cooking, steaming or stir-frying, or in different dishes).

2. I find it difficult to prepare a meal with more than five fresh ingredients.*

3. I am able to alter a recipe myself (for example if I am missing one of the ingredients).

4. I am able to prepare fresh fish in different ways (for example grilling, pan frying or stewing, or in different dishes).

5. I am able to prepare a meal using fresh ingredients (so without pre-packed and processed foods).

6. I am able to see, smell or feel the quality of fresh foods (for example of meat, fish or fruit).

7-point Likert scale Completely disagree 1 – 2 – 3 – 4 – 5 – 6 – 7 Completely agree

4.4.3 Planning ability

Planning ability was measured by using a validated scale as proposed by Stefan et al. (2013). The respondents were asked to rate six statements on a seven-point Likert scale, ranging from 1 “completely disagree” to 7 “completely agree”. A reliability analysis was once again conducted to see whether all the items were internally consistent. This resulted in α = 0.674, which is on the brink of the satisfactory threshold of α > 0.7. However, because of the relatively small sample size (N = 30) and a low number of items this lower alpha was expected. Furthermore, the alpha could not be increased by deleting any of the items, and this would also negatively affect the content validity. Therefore, it was decided to retain all items and use this scale for further analysis. The items that were analysed can be found in Table 2 below.

Table 2: Planning ability scale

adopted from Stefan et al. (2013)

Planning ability Moderator

1. I make a list of the food I want to buy prior to my shopping trip.

2. I check my food inventories prior to my shopping trip. 3. I already plan my meals several days in advance.

(30)

4.4.4 Time-pressure

Perceived time-pressure was measured by using a validated scale as proposed by Brunner, Van der Horst, and Siegrist (2010). It contained five statements in total, and respondents were asked to rate those on a seven-point Likert scale, ranging from 1 “completely disagree” to 7 “completely agree”. Following the executed reliability analysis, the internal consistency gave a value of α = 0.786. Because this is satisfactory, it was decided to retain all different items from this scale. The items that were analysed can be found in Table 3 below.

Table 3: Time-pressure scale

adopted from Brunner et al. (2010)

Time-pressure Moderator

1. ‘‘So much to do, so little time’’; this saying applies very well to me. 2. I need more hours in the day to get my work done.

3. I feel like I have a lot of time on my hands.

4. I feel like no matter how hard I work, I’ll never get caught up. 5. I am always in a rush.

7-point Likert scale Completely disagree 1 – 2 – 3 – 4 – 5 – 6 – 7 Completely agree

4.4.5 Satisfaction about the meal-kit service: content-wise

This satisfaction scale was measured by using a 14-item scale concerning the variety, quality, taste, and other related variables. Respondents were asked to rate those statements on a seven-point Likert scale, ranging from 1 “very unsatisfied” to 7 “very satisfied”. Following the executed reliability analysis, the internal consistency gave a value of α = 0.956. Because this is highly satisfactory, it was decided to retain all different items from this scale. The items that were analysed can be found in Table 4 below.

Table 4: Satisfaction meal-kit service: content-wise scale

Satisfaction meal-kit service: content-wise Moderator

1. The variation among the available products. 2. The freshness of the products.

3. The shelf life of the products. 4. The flavour of the products. 5. The appearance of the products. 6. The quality of the products.

7. How healthy the meal-kit service is. 8. How nutritious the meal-kit service is. 9. The naturalness of the meal-kit service. 10. The colour of the products.

11. The shape of the products.

12. The size/quantity of the products. 13. The structure of the products. 14. The ease of the products.

(31)

4.4.6 Satisfaction about the meal-kit service: portion sizes

This portion size satisfaction scale was measured by using a 3-item scale concerning how fitting the participant found the offered portion sizes for their household. Respondents were asked to rate those statements on a seven-point Likert scale, ranging from 1 “completely disagree” to 7 “completely agree”. Following the executed reliability analysis, the internal consistency gave a value of α = 0.941. Because this is highly satisfactory, it was decided to retain all different items from this scale. The items that were analysed can be found in Table 5 below.

Table 5: Satisfaction meal-kit service: portion sizes scale

Satisfaction meal-kit service: portion sizes

1. The meal-kit service contained very appropriate portion sizes. 2. The meal-kit service contained appropriate portion sizes for my household.

3. I am satisfied with the offered portion sizes.

7-point Likert scale Very unsatisfied 1 – 2 – 3 – 4 – 5 – 6 – 7 Very satisfied

4.4.7 Control variables

A number of socio-demographic variables were included as control variables in this research study, namely; age, gender, household size and ages of all individual members, educational level, average working hours per week, the income level of the household, and whether the household owned pets. These control variables were added based on previous research findings that they may significantly affect household food waste (Koivupuro et al., 2012). Additionally, from the household size combined with the individual ages of all members of the household, an extra control variable was computed, namely; household composition. This variable consisted of four different types: 1) couple without children (anymore), 2) couple with child(ren) aged 15 and above, 3) couple with child(ren) aged 14 and below, and 4) single-parent households.

The control variables were all tested one by one on a General Linear Model with repeated measures in SPSS. The dependent variable, in this case, was the sum of the total waste in grams/millilitres indicated for each week in the food waste diaries, and a control variable was the independent variable. The results turned out that none of these control variables were significantly interacting with the dependent variable, and therefore were not regarded necessary to control for in the further analyses to test the hypotheses.

(32)

4.5 Methods of analysis

For the first hypothesis H1, it was chosen to analyse with a paired samples t-test in order to determine whether there was a significant difference in the amount of food wasted while the participating households either had a meal-kit service in one week, or did not have one in the other week. Furthermore, the paired samples t-test also allowed checking what direction a potentially significant difference would be. The dependent variable would be the sum of the total waste in grams/millilitres indicated for each week in the food waste diaries. Additionally, robustness checks were done by performing the same paired samples t-test with only the solid food waste in grams, and the waste in grams/millilitres per capita as the dependent variables.

For the remaining hypotheses (H2a, H2b, and H2c) it was decided to utilise the SPSS macro called MEMORE by Montoya (2019). MEMORE allows for testing moderation in a within-subjects design as well as performing directional tests in order to better understand not only whether a moderator its effect is different from zero, but also whether its effect is positive or negative which is necessary for testing these hypotheses (Montoya, 2019). These tests were conducted with the moderators (cooking skills, planning ability, and perceived time-pressure) one by one, and had the sum of the total waste in grams/millilitres indicated for each week in the food waste diaries as the dependent variable, and a moderator as the independent variable. Lastly, a section was dedicated towards looking at the calculated waste percentage and the numbered plastic bags with the remainders of the meal-kit service. This section was added to further understand what specific variables may influence the amount of food wasted from the meal-kit service itself rather than the bigger scope of the research which compared a week without a meal-kit service to a week with a meal-kit service. Furthermore, the waste amount in grams of the plastic bags is compared to the self-assessed waste amount from the waste diary in order to detect potential over- or underreporting by participants.

(33)

5. Results

In this chapter, the results of this research study are discussed. The first section is dedicated to descriptive statistics, in order to further elaborate on the details of how the sample of this research study looks like. Afterwards, the hypotheses are tested as explained before in section 4.5, the satisfaction about the meal-kit service in terms of its contents and portion sizes is analysed, and lastly, a section is dedicated to the findings about the plastic bag in which the remainders of the meal-kit service were collected and the related waste percentage.

5.1 Descriptive statistics

Ultimately, a total of 30 households participated in this research. Of those 30 households, everyone filled in the required questionnaires, the waste diaries, and had handed over their numbered plastic bags. There was one household that had not handed in any peels in their numbered plastic bag but only their remaining edible products. Because of the small sample of this research, and because of the fact that this research study focuses mainly on edible and avoidable waste it was decided to still include their data for the analyses. This leads to our sample of N = 30.

(34)

Table 6: Descriptive statistics

Income level Frequency Percent

€1000 – €1499 1 3.3%

€1500 – €1999 1 3.3%

€3000 – €3999 7 23.3%

€4000 or more 10 33.3%

Prefer not to say. 11 36.7%

Gender Male 4 13.3% Female 26 87.7% Educational level MAVO/VMBO 3 10% MBO 6 20% HBO 17 56.7% University 4 13.3% Profession Employed 21 70% Self-employed/Freelance 6 20% Houseman/Housewife 1 3.3% Jobless 1 3.3% Retired 1 3.3% Household composition

Couple without children 9 30%

Couple with 15+ aged child(ren) 11 36.7%

Couple with 14- aged child(ren) 7 23.3%

Single-parent household 3 10% Pet ownership No. 18 60% Yes. 12 40% Range Mean SD Age 26 – 67 51.47 8.57

Average working hours per week 0 – 45 28.70 11.12

Household size 2 – 5 3.07 0.91

Cooking skills 2.40 – 7.00 5.93 0.89

Planning ability 1.00 – 7.00 4.93 1.33

Time-pressure 1.20 – 6.00 4.03 1.03

Satisfaction meal-kit service: content-wise 3.79 – 7.00 6.04 0.75

Satisfaction meal-kit service: portion sizes 1.33 – 7.00 5.19 1.66

Waste percentage 0.12% - 44.82% 11.28% 12.05% 5.2 Main effect of the meal-kit service on food waste

(35)

service but instead acted as they normally would. As mentioned before, the first hypothesis of this research is as follows;

H1: The usage of a meal-kit service by households will reduce their amount of food wasted in a period compared to a period when the meal-kit service was not used.

The sums of the total waste in grams/millilitres indicated for each week in the food waste diaries were used as the dependent variables. Results from the paired samples t-test indicated that there was a significant difference in the scores for with meal-kit service (M = 1867.67, SD = 1852.67) and without meal-kit service (M = 1307.67, SD = 1612.26) conditions;

t (29) = 2.61, p = 0.014. These results suggest that a meal-kit service really does have a

significant influence on food waste generated by households. However, unlike the expectations as explained in the hypothesis, the results suggest that when a household uses a meal-kit service, the amount of food waste generated by a household increased rather than decreased.

For the robustness tests, it was decided to also conduct two additional paired samples t-tests. To start, the solid food waste in grams indicated for each week in the food waste diaries were selected as the dependent variables. This was done as the liquid food waste may heavily influence the previous results, even though the meal-kit service does not supply liquid food products. Therefore it makes sense to see whether the previous results also hold when only solid food waste is analysed. Results from the paired samples t-test indicated that there was again a significant difference in the scores for with meal-kit service (M = 1504.17, SD = 1401.11) and without meal-kit service (M = 841.07, SD = 760.50) conditions; t (29) = 3.57, p = 0.001.

Next, the waste in grams/millilitres per capita indicated for each week in the food waste diaries were selected as the dependent variables. This was done as household size is argued by literature to influence food waste (Koivupuro et al., 2012), therefore a per capita measurement was chosen as the dependent variable to see whether the previous results hold. Results from the paired samples t-test indicated that there was again a significant difference in the scores for with meal-kit service (M = 561.28, SD = 508.22) and without meal-kit service (M = 381.10, SD = 463.61) conditions; t (29) = 2.57, p = 0.016.

Although the results indicated that there was a significant difference between the week with a meal-kit service and the week without a meal-kit service, the significant difference suggested that the amount of food waste generated by a household was increased rather than decreased due to the usage of a meal-kit service. Therefore, hypothesis H1 was rejected. In

(36)

Figure 3: Johnson-Neyman plot of cooking skills on meal-kit service

5.3 Moderating effects of cooking skills, planning ability, and time-pressure

This section elaborates upon how the second set of hypotheses of this research study were tested. This is done in order to determine whether the proposed moderators cooking skills, planning ability, and perceived time-pressure significantly influence the effect of the meal-kit service on food waste. To test these hypotheses, the SPSS macro called MEMORE by Montoya (2019) was used in order to first determine whether there is a significant interaction effect, followed by directional tests to determine whether the effect is positive or negative.

To test hypothesis H2a the sums of the total waste in grams/millilitres indicated for each week in the food waste diaries were used as the dependent variables (Ydiff = Waste diary with meal-kit service – Waste diary without meal-kit service), and cooking skills was used as the

0 200 400 600 800 1000 1200 1400 1600 1800 2000

With Meal-Kit service Without Meal-kit service

F o o d w a st e in g /m l

Plot of the main effect of the meal-kit service on food waste

Food waste in g/ml Solid food waste in g Food waste in g/ml per capita

Figure 2: Plot of the main effect of the meal-kit service on food waste

500,00 550,00 600,00 650,00 700,00 750,00 2 3 4 5 6 7 Y d if f Cooking skills 250 300 350 400 450 500 550 600 650 700 750 1 2 3 4 5 6 7 Y d if f Planning ability

More in boxes, less in the bin? The Effect of Meal-kit Services on Food Waste