Household Food Waste Reduction: The role of thriftiness and biospheric values on the effectiveness of food waste interventions

(1)

1 RIJKSUNIVERSITEIT GRONINGEN – FACULTY OF ECONOMICS AND BUSINESS

Household Food Waste Reduction:

The role of thriftiness and biospheric

values on the effectiveness of food

waste interventions

A Field Study Among Dutch Households

Sytze Boskma s2036681 MSc Marketing Management

January 2017

First supervisor: dr. J. van Doorn Second supervisor: M. Drijfhout

(2)

2

Abstract

Households are found to produce over 50% of the total food waste in Europe, which causes both moral injustice and environmental damage. Previous literature on the topic of food waste has described the motives people have to prevent food waste. Among these motives, financial and environmental motives are found, which are linked to the consumer characteristics thriftiness and biospheric values. Less is known about the role of these consumer characteristics on the effectiveness of initiatives to reduce food waste. This study therefore tries to make a first step in filling this gap by researching the effect of thriftiness and biospheric values on the effectiveness of two types of food waste interventions. A field study was conducted in which participants kept track of their wasting behaviour via a food waste diary for a week, accompanied with a questionnaire. Analysis showed that both thriftiness and biospheric values have a negative impact on the effectiveness of an intervention in the after-purchase stage and food waste reduction. This research shows the importance of taking consumer characteristics, such as thriftiness and biospheric values, into account when designing food waste interventions and when targeting to whom the intervention will be distributed. More theoretical and managerial implications, along with the limitations of this study are presented at the end of this paper.

Keywords: household food waste, food waste prevention, food waste interventions, biospheric

(3)

3

1. Introduction

While 2.2 billion people live in a state of (near) poverty, a rough estimate of one third of the annual produced food worldwide is wasted, on a yearly basis (Porpino, Parente, & Wansink, 2015; UNDP, 2014). Besides the moral injustice that is induced by wasting food, the environmental damage is evenly important to consider (Gustavsson, Cederberg, & Sonesson, 2011; Koivupuro et al., 2012). Environmental damage occurs when edible food is not consumed, ‘all the environmental impacts of producing the raw materials and processing them into food products are for nothing, and all the energy and resources used to produce the food have been used in vain’ (Koivupuro et al., 2012). Furthermore, waste has financial implications for households, whereas viable food is discarded (Papargyropoulou, Lozano, K. Steinberger, Wright, & Ujang, 2014). Food waste has become a highly covered topic in the literature since the end of last century (Prothero et al., 2011). Over the past few years, research has shown the severity of the food waste problem in tonnes of food waste per capita, econometric measures and in lost resources across the supply chain (Cox et al., 2010; Doorn, 2016; Evans, 2011; Koivupuro et al., 2012; Porpino, Parente, & Wansink, 2015, among others).

Earlier, waste management focused mainly on pollution prevention. Lately a shift is apparent to a holistic approach (Papargyropoulou et al., 2014) and even though it is widely accepted that food waste is one of the major problems that need to be tackled, no consensus is found on the exact antecedents of the problem (Evans, 2011). As mentioned before, food waste is found to take place throughout all the stages of the food chain, and causes of food waste are found to differ in each stage and are product group specific (Göbel, Langen, Blumenthal, Teitscheid, & Ritter, 2015). Even though it is argued that not a single entity can be blamed for the problem of food waste, and there is not ‘a single solution to bring forth notable changes’ (Göbel et al., 2015; Stancu, Haugaard, & Lähteenmäki, 2016) this research will focus on the households. Rational behind this choice is that each stage of the food chain requires different solutions (Göbel et al., 2015) and household are found to produce more than 50% of the total food waste in Europe (Kummu et al., 2012; Stancu et al., 2016).

(6)

6 found to decrease when household size increases (Koivupuro et al., 2012; Mallinson et al., 2016; Visschers et al., 2016), whereas single person household have less economies of scale (due to inadequate size of packages) and therefore more leftovers.

Different routines are outset in the literature and can be summarized into the following stages of the ‘domestic food cycle’ (Flower & Collet, 2015; Mallinson et al., 2016). First stage is the planning stage, which takes place before or during the shopping trip and is followed by the actual shopping. The other stages take place at home and are, in succession, storage, preparation and the consumption of food (Flower & Collet, 2015). The field of literature already consists of research on the impact of different routines on the different stages of the food cycle in relation to food waste. A solid planning routine (in the planning and shopping stage) is found to be a predictor of (less) food waste (Stefan, van Herpen, Tudoran, & Lähteenmäki, 2013). For example, making a shopping list, or the planning of meals in advance are found to decrease the amount of unplanned purchases of consumers, and to limit food waste (Bell, Corsten, & Knox, 2011). Furthermore, a proper storage routine is found to decrease food waste, whereas proper storage increases the durability of food (Koivupuro et al., 2012; Porpino et al., 2015). The preparation routine in the domestic food cycle is found to have an impact on food waste, whereas overcooking is related to more food waste (Graham-Rowe, Jessop, & Sparks, 2014). Finally, the consumption routine has an influence on the other routines, whereas consumption habits, for example amount and size of portions eaten, has an influence on the amount of food bought, prepared etc. (Cox et al., 2010).

Waste prevention methods, or so called food waste interventions, focus on these specific household routines. In this research two types of interventions are distinguished, interventions before the point of purchase (targeted at planning and shopping routines) and interventions after the point of purchase (targeted at the storage, preparation and consumption routines). Rationale behind this choice is that both the planning and shopping routine, and the other three routines are intertwined. The planning routine can occur during the shopping routine, and the consumption routine has influence on the storage and preparation routine. Therefore measures taken to prevent such routines, are inherently focussed on multiple routines.

(7)

7 waste, rather than the environmental impact of food waste (Otterbring & Gustafsson, 2012; Parizeau, von Massow, & Martin, 2015). Secondly, environmental motivations to reduce food waste stem from a concern for the environment as opposite to the financial motives. People who tend to base their decision to act pro-environmental or not on the benefits and costs for the ecosystem are found to score high on biospheric values (de Groot & Steg, 2009). People with these biospheric values will therefore be more involved with green choices rather than per se healthy or financial choices. Although the direct effect of these two motivations are well described in the literature, their influence on initiatives (interventions) to change people’s food related routines are unknown. Therefore, the following research question will be central in this research:

What is the effect of point-of-purchase and after-purchase food waste interventions on food waste reduction and what is the role of thriftiness and biospheric values on this relationship?

The following sub-questions will be addressed, in order to answer the research question:

Sub-question 1: What is the effect of point-of-purchase and after-purchase food waste interventions on food waste reduction?

Sub-question 2: What is the role of thriftiness on the effectiveness of point-of-purchase and after-purchase food waste interventions?

Sub-question 3: What is the role of biospheric values on the effectiveness of point-of-purchase and after-purchase food waste interventions?

Sub-question 4: Are the different types of interventions more effective for people who score high on biospheric values and / or thriftiness?

By researching the following questions, this research contributes to the field of food waste prevention. Especially the influence of thriftiness and biospheric values on the relation between food waste interventions and food waste is of added value to the literature. Practice can benefit from these insights in the process of targeting consumers for their food waste intervention and assessing the effectiveness of possible interventions and target groups.

1.1. Structure of the paper

(8)

8

2. Literature Review

This section will give an extensive overview of the existing literature that is related to the research questions provided in the introduction. First, an overview will be given on the topic of food waste, including definitions and antecedents of food waste found in the existing body of literature. By identifying the antecedents of food waste, the effectiveness of food waste reduction methods, the so called food waste interventions, can be explained. In explaining the effectiveness of food waste interventions, Defra’s (2005) framework of behavioural change and the theory of planned behaviour of Azjen (1991) will stand central, whereas behavioural change is needed to break open the food related routines of the consumer. This section will be continued with an outset of the point-of-purchase and after-point-of-purchase interventions, central to this research. The theoretical review section will be concluded with an overview of the consumer characteristics related to the interventions and food waste, which in turn will result in hypothesis.

2.1. Food waste

Although food waste is a well know concept in the literature, definitions of food waste are diverse and are not uniformly used across the literature. For example, Lipinski, Hanson, Lomax, & Kitinoja (2013) differ between food loss and food waste, and define food waste as food that is of good quality, which is still fit for human consumptions but is discarded before or after it spoils. They reason that ‘food waste is the result of negligence or a conscious decision to throw food away’ (Lipinski et al., 2013). Gustavsson et al. (2011) take a broader perspective and include all the non-use of food in the distribution and consumption stage in their definition of food waste. Others argue that food waste is the result of ‘inappropriate behaviour of producers, retailers or consumers, as well as lack of technological inputs’ and that food waste might occur at any point in the supply chain of food (Leal Filho & Kovaleva, 2015). Reviewing these different scopes of definitions it becomes evident that, in regard to our research, a definition is needed which only includes the food waste produced by households. Therefore food waste and food loss produced by other entities in the food chain, such as producers, retailers or other sub-parties are not taken into account in this research.

(9)

9 waste, and therefore unavoidable food waste will be treated as given, and not included in the food waste definition.

Tucker & Farrelly, (2015) define food waste as everything that is a leftover from meal preparations, or ‘which remains uneaten at the end of a meal, and food that is left unused or only partially used and then disposed of, and is not diverted to pets, composting, or other useful ends’. Whereas in this study no distinction will be made between household with pets, room for composting or other means of waste recycling, the definition of food waste this research need will treat composting or recycling also as waste. Rational behind is, that in principal, the food was intended to be consumed, and not to be composed, recycled or fed to pets. However, donating left over food to causes where it will be consumed in a later stage (for example the Dutch voedselbank) is not treated as food waste.

Taken all together, household food waste is defined as the non-use of edible food or raw materials that could have been consumed if it were prepared and/or stored properly.

2.2. Antecedents of food waste

(10)

10 waste are inappropriate food conservation, avoidance of leftovers, over-preparation and excessive purchases.

When reviewing the different antecedents of food waste, a split between two decision making processes becomes apparent. A distinction can be made between ‘consumers’ decision making at the point-of-purchase’ and consumers’ consumption and disposal decisions at home’ (Doorn, 2016). Linking this distinction to the integrated model of household food waste (Porpino et al., 2015), it can be argued that the buying and stocking (remembering what is in stock) stage of the food itinerary are point of purchase related. The other before mentioned antecedents are related to consumers’ consumption and disposal decisions and falling under the stages of preparation, consumption or storage of the food itinerary as presented by Porpino et al., (2015).

2.3. Food waste prevention, interventions and behavioural change

Food waste prevention can be defined as the ‘measures taken before a substance, material or product [e.g. food] has become to waste, that reduces the quantity of waste’ (Cox et al., 2010). Food waste prevention measures are often policy based, for example, focussing on the awareness of the food waste problem (Stenmarck, Hanssen, & Silvennoinen, 2011). Policy measures are often superficial in nature, targeting broad goals, for example the UN initiative to reduce food waste by 50% in 2025. Food waste interventions are initiatives focused on food waste prevention targeted to designated instigators of the food waste problem, for example producers or consumers. Whereas, in this research, the focus lies on food waste generated in households, ‘food waste interventions’ will encompass interventions targeted to the households.

(11)

11 (Tucker & Farrelly, 2015). Also the reminder to ‘hang on to new prevention habits’ is therefore missing (Cox et al., 2010).

Food waste is embedded in everyday life, and therefore needs to be understood in relation to everyday issues (Evans, 2011). Argued is that interventions ‘should not target individuals and choices; rather, they should focus on the social and material contexts through which practices are ordered and (re)produced’ (Evans, 2011; Southerton, McMeekin, & Evans, 2011). Evans (2011) suggests that interventions in the material context are key. In other words, interventions should focus on changing food related behaviour by changing the material context of food related routines rather than trying to change the individual choice to throw food away or not. (Evans, 2011; Flower & Collet, 2015). ‘Intervention projects have found that people need to be educated about the specific actions they can take, and why these are worth doing, rather than relying on general exhortations to reduce waste’ (Cox et al., 2010).

Behavioural change is the intended outcome of food waste interventions, and therefore an important concept in the food waste literature. Many studies considering food waste are built upon theories or frameworks related to the behavioural change literature stream (Cox et al., 2010; Parizeau et al., 2015; Stancu et al., 2016; Stefan et al., 2013; Thyberg & Tonjes, 2016; Visschers et al., 2016). Two frameworks of behavioural change prevail in the theory, the framework of behavioural change and the theory of planned behaviour (Azjen, 1991; Cox et al., 2010; Darnton, Elster-jones, Lucas, & Brooks, 2006; Defra, 2005). Defra’s (2005) framework of behavioural change argues that the following actions are needed in order to stimulate behavioural change.

 Enable: People need information, guidance and support that is focussed on the intended behaviour to make behavioural change easier. This includes removing barriers, giving information and providing facilities.

 Engage: Involvement is necessary early on to create a sense of personal responsibility.  Encourage: Incentives that show the right signals are found to foster behavioural change, as

does regular feedback.

 Exemplify: Giving examples of the behaviour, and the moments the behaviour can be displayed increases behavioural change.

(12)

12 The model argues that an individual’s intention to display a behaviour is the central determinant for that behaviour. Motivation and ‘willingness to act’, together, encompass the intention (Azjen, 1991). Intention is influenced by three factors, which thus have an indirect effect on the actual behaviour (Azjen, 1991; Visschers et al., 2016). First of all, the attitude of a person must be in favour of the intended behaviour, for the behaviour to take place. Secondly, norms and opinions of people close by the decision maker must be in line with the intended behaviour to let the behaviour take place. Finally, the person must have ‘perceived behavioural control’ over the intended behaviour, in other words, ‘the individual should have the opportunity and sufficient resources to perform the behaviour’ (Visschers et al., 2016).

Comparing the two frameworks with each other, taking into account their usability for this study, the theory of planned behaviour appears to have the best fit with this study. The theory of planned behaviour takes the intentions of the person of interest central for behavioural change, instead of the necessary actions for change in the framework of Defra (2005). Whereas attitudes are, according to the theory of planned behaviour, a predictor of one’s intention to display a behaviour, this theory is very suitable for explaining the role of thriftiness and biospheric values on the effectiveness of interventions. Furthermore, food waste is a behaviour that is partly under volitional control of a person, and therefore the theory of planned behaviour is suitable for predicting and understanding food waste (Graham-Rowe et al., 2014; Stefan et al., 2013; Visschers et al., 2016). Finally, the actions ‘encourage’ and ‘exemplify’, of the framework of Defra (2005), are in a lesser extent applicable to the concepts in this research. The theory of planned behaviour, as developed by Azjen (1991), will therefore stand central in this research.

(13)

13 to the consumer food routines, and in the end reduces food waste (Stefan et al., 2013). This leads to the following hypothesis:

• Hypothesis 1a: Food waste interventions have a positive effect on food waste reduction Two different types of interventions are distinguished in this research, as previously stated. Interventions targeted at the planning and shopping routines are classified by the term point-of-purchase intervention whereas planning routines contribute to food waste only in an indirect manner, through shopping routines (Stancu et al., 2016). Shopping routines are found to have an direct impact on food waste, mainly by excessive purchasing (Mallinson et al., 2016; Porpino et al., 2015; Stancu et al., 2016; Stefan et al., 2013). The interventions targeted at routines after the shopping routine, are classified as after-purchase interventions. Literature suggests such interventions should focus on increasing the ability of consumers to reuse leftovers, whereas consumers tend to lack the proper knowledge regarding the storage and handling of leftovers (Terpstra, Steenbekkers, de Maertelaere, & Nijhuis, 2005), for example the possibility to freeze food (Brown, Hipps, Easteal, Parry, & Evans, 2014). The lack of information and knowledge regarding storage and handling of leftovers is found to impact consumers’ food management (Wikström, Williams, Verghese, & Clune, 2014) and thereby contributing to food waste (Aschemann-Witzel, de Hooge, & Normann, 2016; Stancu et al., 2016). It is stated that ‘among the routines, the leftovers reuse routines were the most important contributors to food waste but were closely followed by shopping routines’ (Stancu et al., 2016). Furthermore it is argued that efforts to change the reuse leftover routines may have the largest effect on food waste (Stancu et al., 2016). Compared to the after-purchase routines (for example the reuse leftover routine), it can be argued that in the shopping routines interventions to change the routine might be less effective, whereas consumers are continuously exposed to in-store stimuli, such as promotions, that distract the consumer from the initial goal of the intervention (Cox et al., 2010; Porpino et al., 2015). As mentioned before, food waste is partly under volitional control of a consumer, and therefore effectiveness of interventions in the after-purchase routines can be more effective whereas these occur mostly at home, and less distraction is apparent. Therefore a stronger effect of after-purchase interventions on food waste reduction is hypothesized:

• Hypothesis 1b: After-purchase interventions have a stronger effect on food waste reduction than point-of-purchase interventions

Besides perceived behavioural control and opinions and norms of others, attitudes play a major role in causing behavioural change. The possible role of thriftiness and biospheric values on the effectiveness of interventions will therefore be outset in the next paragraph.

(14)

14 Whereas costs are cited as one of the main barriers to food waste prevention activities, the consumer characteristic thriftiness might play a significant role on the effectiveness of food waste interventions (Aschemann-Witzel et al., 2016; Koivupuro et al., 2012; Salhofer & Obersteiner, 2008). The influence of biospheric values might play a role as an important attitude regarding food waste reduction, whereas biospheric values are linked to pro environmental behaviour (de Groot & Steg, 2009). First the role of thriftiness on food waste related behaviour will be discussed based on what is found in the literature. Secondly, the role of biospheric values on the relationship between food waste interventions and food waste will be explained. Both will result in hypotheses.

2.4.1. Thriftiness

Thriftiness, or frugality, can be defined as ‘a unidimensional consumer lifestyle trait characterized by the degree to which consumers are both restrained in acquiring and in resourcefully using economic goods and services to achieve longer-term goals’ (Lastovicka, Bettencourt, Hughner, & Kuntze, 1999). Throughout the literature, the concept of thriftiness is interpreted in different ways. Some researches treat thriftiness as a personal trait, others as a single value (Todd & Lawson, 2003). Most prevailing is the interpretation of frugality as a lifestyle construct (Lastovicka et al., 1999; Todd & Lawson, 2003). Thrifty people are therefore people who are careful in their use of resources and have a tendency to avoid waste (DeYoung, 1986). Different perspectives can be taken in discussing the effect of thriftiness on the relationship between food waste interventions and food waste reduction.

(15)

15 awareness waste less food. Therefore it can also be argued that people who are thrifty probably already display food waste avoiding behaviour, and therefore interventions are less effective.

Focussing on the role of thriftiness in the planning and mainly shopping routine, it is found that people who are price oriented are triggered by low pricings for large packages and multi-item offers (Aschemann-Witzel et al., 2016). Consequence of this tendency is that large unit sizes are encouraged to be bought, and thereby often more food than needed is bought (Aschemann-Witzel et al., 2016). As discussed before, excessive purchasing is among the main antecedents of food waste (Porpino et al., 2015). Whereas the avoidance of food waste in planning and shopping routines is often found to be driven by thriftiness rather than environmental or other concerns (Graham-Rowe et al., 2014; Visschers et al., 2016), it can be expected that these people will place more importance to the economic package than the consequent waste it might produce. Considering the arguments outset above, this leads to the following hypothesis:

• Hypothesis 2a: Thriftiness weakens the effectiveness of the point-of-purchase intervention on food waste reduction.

In line of the reasoning presented above, it can be argued that thrifty people need to have good after purchase routines, whereas economic packages and excessive purchases can lead to leftovers (Aschemann-Witzel et al., 2016). As stated before, thrifty people are people who have a tendency to avoid waste, and use resources very carefully (DeYoung, 1986). Therefore it can be argued that interventions may be less effective for people who are thrifty, whereas these people already display food waste preventing behaviour, due to their thriftiness. Graham-Rowe et al. (2014) found that food management at home is important in food waste reduction (for example knowledge of food storage) and that people who behave frugally to a greater extent already displayed this behaviour. They state that ‘the household food purchasers who had financial constraints felt that behaving frugally was fundamental to avoiding waste. […] using the food that they already had at home before purchasing more food appeared to be a key technique used by some of the household food purchasers to keep food waste, and therefore food cost, to a minimum’ (Graham-Rowe et al., 2014). This leads to the following hypothesis:

• Hypothesis 2b: Thriftiness weakens the effectiveness of the after-purchase intervention on food waste reduction.

2.4.2. Biospheric values

(16)

16 with strong biospheric values are more likely to act environmentally and have more pro-environmental preferences and intentions (Ojea & Loureiro, 2007; Steg, Bolderdijk, Keizer, & Perlaviciute, 2014; Van der Werff et al., 2013). Furthermore, found is a positive relationship between people with biospheric values, and organic purchasing behaviour (Van Doorn & Verhoef, 2015). Related to food waste, it is found that people with strong biospheric values are more likely to act in sustainable consumption (Van der Werff et al., 2013). Also it is found that people who are intrinsically motivated to behave environmental friendly, sustain this behaviour also for a longer time (Steg, Mazzeo, Caballero, Rranco and Favara, 2014).

Whereas people with biospheric values base their decisions on the consequences for the environment (Ojea & Loureiro, 2007; Steg et al., 2014), these people might already be engaged in food waste reducing behaviour. Therefore, it can be reasoned that the presence of biospheric values weakens the effectiveness of the interventions. However, regarding the planning and shopping stages, the presence of biospheric values is found to be a predictor of organic purchasing behaviour (Van Doorn & Verhoef, 2015) and to ‘strengthen active commitment to pro-environmental purchase behaviour’ (Nguyen & Greenland, 2016). Subsequently it can be argued that these people might be occupied with, for instance, finding organic products during their shopping trips and therefore less occupied with food waste reduction related behaviour. Furthermore, people with high biospheric values have an attitude in line with the intended behavioural change, and therefore these people are more susceptible for change initiatives (Azjen, 1991). As a result the effectiveness of the point-of-purchase intervention is expected to be higher for people with biospheric values, as summarized in the following hypothesis:

• Hypothesis 3a: Biospheric values strengthen the effectiveness of the point-of-purchase intervention on food waste reduction.

(17)

17 will feel responsible for and committed to the problem of food waste (Steg et al., 2014) and therefore it is expected that these people are already involved in food waste reducing behaviour, and are committed to this. This will especially be in place in the after-purchase routines, whereas practices as, for instance, over-cooking and not properly storing leftovers, lead directly to waste. This will have a negative effect on the effectiveness of the after-purchase intervention, which results in the following hypothesis:

• Hypothesis 3b: Biospheric values weakens the effectiveness of the after-purchase intervention on food waste reduction.

2.5. Conceptual Model

The hypotheses are shown in the conceptual model below. The expected effects of each of the concepts on food waste reduction are shown in this figure. A moderating effect of the two consumer characteristics is proposed, and the effects are denoted by either a positive or negative sign.

Food waste reduction Point-of-purchase intervention After-purchase intervention Biospheric Values Thriftiness Consumer Characteristics H2a: - H2b: - H3a: + H3b: - H1a: + H1a: +

(18)

18

3. Methodology

In order to make replication of this study possible, the following section will cover the method of research that was used in this study. The overview of the research design will be followed by an explanation of the data collection method that is used in order to find answers for the hypotheses developed in the previous section.

3.1. Research design and data collection

The aim of this study was to explain the influence of thriftiness and biospheric values on the relationship between two different types of food waste interventions and food waste reduction. Thereby this study contributes to previous work done by, among others, Doorn (2016). Whereas this studies comprehends knowledge testing rather than knowledge generation, a theory testing approach is chosen for this study (van Aken, Berends, & van der Bij, 2012). In order to find support for the hypotheses proposed in the previous section, and an valid answer to the research questions, a quantitative data study was set up. A quantitative research design fits best with the affirmative nature of this research (van Aken et al., 2012), and therefore a field study was conducted at a local supermarket (Albert Heijn Paterswolde) among 280 respondents.

Respondents were split into three groups, each designated to a particular intervention or control group (no intervention). Each participant received a food waste diary, a questionnaire and an intervention, in case of the intervention groups. All materials were in Dutch, whereas all the respondents had the Dutch nationality. Respondents were asked to use the intervention and keep track of their food waste through the food diary. After a week the respondents were asked to hand-in their food diary, along with the completed questionnaire at the supermarket.

3.2. Measurements and food waste interventions

The data collection had two primary sources, questionnaires and a food waste diaries. The two types of interventions distinguished in the literature section were tested among the respondents. What follows is an outset of the sources of data collection and the used interventions.

3.2.1. Food waste interventions

(19)

19 consumers are made to consider their inventory before buying groceries, combined with an extended meal planning, both reducing excessive purchasing and thereby food waste.

The second intervention, the ‘sticker intervention’ constitutes of a sheet of stickers which can be pasted to leftovers or perishable food. Also some reminders in form of a sticker are present on the sheet, which can be placed in home, to remind the consumer take leftovers with them for lunch, for example. Rational behind this intervention is that by reminding the consumer of the ‘best before’ dates and the presence of leftovers and helping the consumer organizing their leftovers, more leftovers will used and consumed instead of disposed.

In the field study, a third intervention was tested, which was no integral part of this research. The intervention, called ‘eetmaatje’, is a tool designed to help people estimate their portions of rice and pasta. By helping consumers estimating their portions, leftovers due to overcooking are reasoned to reduce, and thereby contributing to the food waste problem.

3.2.2. Questionnaire

The questionnaire used in the field study constituted of a multitude of questions regarding demographics, food related behaviour and personal characteristics. Besides the demographics, the constructs of thriftiness and biospheric values were most important for this research. An overview of the measured variables in this study can be found in table 1.

The construct of thriftiness is measured by using the scale of frugality, originally developed by Lastovicka, Bettencourt, Hughner, & Kuntze (1999). The scale measures consumers frugality (synonym for thriftiness) by asking respondents to make ratings on a set of eight items, with a five point Lickert scale, ranging from 1- describes me not at all, to 5- describes me extremely well. Examples of questions are ‘if you take good care of your possessions, you will definitely save money in the long run’ and ‘if you can re-use an item you already have, there’s no sense in buying something new’ (Lastovicka et al., 1999). The full scale can be found in the second appendix

The construct of biospheric values is measured by using the scale of Steg, Dreijerink, & Abrahamse (2005). The construct measures the general value orientation (biospheric, altruistic and egoistic) of consumers using a nine-point scale, ranging from -1- not at all important to 7-of supreme importance. Example questions the respondents had to rate are ‘respecting the earth: live in harmony with other species’ and ‘preventing pollution: protecting natural sources’ (Steg et al., 2005). The complete scale can be found in Appendix 2.

(20)

20

Question Variables Example Derived from

1-17, 21-22 Demographics Age, household size & income E.g. Silvennoinen,

Katajajuuri, & Hartikainen (2014)

29 Thriftiness ‘if you take good care of your possessions, you will definitely save money in the long run’

Lastovicka et al., (1999)

27 Biospheric Values ‘respecting the earth: live in harmony with other species’

Steg et al., (2005)

Table 1: Overview of the measured variables

3.2.3. Food waste diary

(21)

21

4. Descriptive statistics and reliability analysis

In order to gain a clear overview of the dataset, descriptive statistics will be presented in the following sections. Furthermore, reliability analysis will be conducted in order to assess the internal consistency of the used scales. Finally, the control variables used in the analysis are discussed.

4.1. Descriptive statistics

280 cases where set out in the field study at Albert Heijn Paterswolde, of which an original dataset of 116 respondents can be constituted (responds rate of 41.43%). For reasons of quality and reliability 8 cases were deleted from this dataset due to empty questionnaires and missing food waste diaries. In one case a missing respondent number was reason to delete the case from the dataset. After cleaning up the data set for missing values, the dataset consisted of 108 respondents. However there has been ambiguity among the respondents regarding the food waste diary, whereas multiple respondents did fill in their entire consumption instead of their waste. As consequence, another 39 cases where considered invalid and have been deleted from the dataset. This resulted in a total dataset of 69 cases. Out of these 69 cases, 13 people received the shopping list intervention (18.84%), 22 people the Eetmaatje intervention (31.89%), 16 people the sticker intervention (23.19%) and the control group consisted of 18 people (26.09%). Finally, Eetmaatje is excluded from the dataset for this research, leaving a workable dataset of 47 cases.

4.1.1. Intervention usage rate

Multiple people did indicate in the questionnaire that they did not use the intervention during the week that they kept track of their waste. In the point-of-purchase intervention (the shopping list), five people mentioned not using the intervention at all, six did use the intervention. One person mentioned using his or her own shopping list, and one person did not answer this question at all. With regard to the after-purchase intervention (the sticker intervention), usage rates were even lower. Out of the 16 people in the workable dataset, only two people indicate that they have actually used the stickers. Excluding the cases that did mention not using the intervention would lead to a sample size which can be considered statistically useless. Whereas people might have been influenced by receiving the intervention, labelling these people as control would bias the results. The intervention usage rate will be excluded from the initial analysis, but included in the robustness check.

4.1.2. Demographics

(22)

22 level of education 6.4% of the sample finished VMBO / MAVO as highest education. 4.3% of the respondents filled in HAVO and 17% MBO as highest degree. 36.2% of the sample has a HBO degree and 29.8% did finish a university program. Finally 6.4% of the sample indicated that they had finished another type of education, and no respondent filled in VWO as highest education. Concluded can be that most of the cases in this dataset are highly educated (either HBO or university). Almost half of the sample (46.8%) is currently employed, either employed or self-employed. 14.9% of the respondents is housewife or houseman and 6.4% is unemployed. In line with the, on average, high age of the sample, 25.5% of the respondents is retired. 6.4% of the sample filled in another type of occupation, for example volunteering work. Regarding income, most people indicated that their entire net household income equalled €4000 or more, corresponding with 29.8% of the sample. 21.3% of the sample indicated an household income of €3000-3999. 10.4% of the sample indicated that they rather did not state their income. One respondent did not fill in anything at this question. This respondent will be treated as if indicated ‘rather do not state my income’. Finally the household size was asked among the respondents, and found was that the majority of the respondents have an household size of two persons (44.4%). This is in line with the, on average, high age of the respondents in this dataset. Households with four persons were the second most mentioned, which was indicated 22.2% of the time. Households with one person (15.6%) were third most mentioned, followed by households with three persons (13.3%) and five persons (4.2%). Two people left this question blanc.

4.1.3 Food waste

Food waste was calculated in two ways, both found in table 2 below. First the total food waste (hence: food waste) was calculated by simply summing the total grams and millilitres per respondent together. However, whereas the category coffee / tea accounts for a major part of the waste of the people with high waste, this category might be biasing the results. Therefore waste without liquids is also calculated, to account for possible impact of the liquids category in the robustness check. The average amount of waste in the total waste calculation is 915.72 with a maximum of 5254 and a minimum of 0.00 for the respondents without any food waste. For the no liquid calculation all the liquids where detracted, resulting in a mean of 640.62 grams, with a maximum of again 5254 and a minimum of 0.00.

Variable Sample size Minimum Maximum Mean Std. Deviation

Total food waste 47 .00 5254.00 915.72 1083.92

Waste without liquids 47 .00 5254.00 640.62 936.72

(23)

23 4.2 Reliability Analysis

Internal consistency is reached when all the questions in a scale measure the same concept. Therefore thriftiness and biospheric values where tested on their Cronbach’s alpha, in order to assess the internal consistency of the questions used in the two scales.

4.2.1. Thriftiness

The concept of thriftiness was measured with eight questions, all on a 5-point Likert scale. A Cronbach’s alpha of 0.809 (α = .809) was found, indicating that all the questions in the scale measure the same concept. As a result all the items can be aggregated into one scale, forming a new averaged variable of thriftiness for further analysis. The mean of this new created variable is 3.62, with a minimum of 1.75 and a maximum of 5. The mean of 3.62 indicates that people are overall slightly thrifty. The descriptives of this variable can be found in the table below.

Variable Sample

size

Minimum Maximum Mean Std.

Deviation

Cronbach’s alpha

Thriftiness 47 1.75 5.00 3.62 0.64 0.809

Table 3: Descriptives thriftiness

4.2.2. Biospheric values

The concept of biospheric values was measured by four questions, originally developed by Steg et al. (2005), assessed on a 9-point Likert scale ranging from -1 to 7. A Cronbach’s alpha of 0.83 (α = .830) was found after assessing the internal consistency of the four questions. Therefore no questions had to be taken out of this concept, and all questions were averaged into one variable for biospheric values. This newly created variable has a mean of 5.41 ranging from a minimum of 3 to a maximum of 7. The mean of 5.41 indicates that overall people have a tendency for biospheric values. The descriptives for the variable of biospheric values can be found in the table below.

Variable Sample

size

Minimum Maximum Mean Std.

Deviation Cronbach’s alpha Biospheric values 46 3 7 5.41 1.38 0.830

Table 4: Descriptives biospheric values

4.3. Control variables

(24)

(25)

25

5. Results

This section describes the results following from the analysis performed to find an answer on the hypotheses developed in the literature section. First the results regarding the (main) effect of the interventions on food waste reduction will be discussed. Secondly, findings relating to the expected moderating role of thriftiness will be presented. Finally, results regarding biospheric values and its expected moderating role will be discussed.

5.1. The effect of food waste interventions on food waste reduction

Goal of the first hypothesis (1a) was to find the effect between food waste interventions (in general) and food waste reduction. Furthermore, hypothesis 1b compares the two types of interventions on its effectiveness on food waste reduction. In order to test the first hypotheses, the means (food waste incl. liquids) of the different intervention groups were compared and tested with an ANOVA and T-tests. Figure two shows the different means per group, table 5 depicts the results of the statistical tests.

(26)

26

Figure 2: Mean comparison between intervention and control groups All tests p > 0.1

Descriptives

Group Sample size Mean Std

Deviation Minimum Maximum Control 17 1140.53 1456.19 60.00 5254.00 Intervention 30 788.33 804.51 .00 3375.00 Point-of-purchase intervention 13 805.38 648.53 .00 1990.00 After-purchase intervention 17 775.29 925.90 .00 3375.00

Statistical tests DV: food waste

F-stat Sig. Intervention vs control 1.149 .289 Point-of-purchase vs control .594 .447 After-purchase vs control .762 .389 Point-of-purchase vs After-purchase .010 .921

Table 5: Empirical results hypothesis 1

1140,53 788,33 805,38 _775,29 0 200 400 600 800 1000 1200 Hypothesis 1 Foo d was te in cl. liq u id s

(27)

27 5.2. The Moderating role of thriftiness and biospheric values

Based on the literature and the conceptual model discussed in the literature section, a moderating effect of thriftiness was predicted. More specifically it was expected that, for people high on thriftiness, interventions in the point-of-purchase stage (h2a) and after purchase stage (h2b) would be less effective in reducing food waste. Regarding biospheric values a similar moderating effect was expected for after-purchase interventions (h3b). A strengthening effect of biospheric values on the relationship between point-of-purchase interventions and food waste was expected in hypothesis 3a. In order to find empirical results for these hypotheses, regression models were computed for thriftiness and biospheric values both for the point-of-purchase and after-purchase interventions.

5.2.1. The moderating role of thriftiness DV Food waste

Independent variable B Std. Error T-stat P-value Tolerance VIF

Constant 3272.250 5016.754 .652 .526 Age -19.411 22.003 -.882 .394 .569 1.758 Gender -880.084 535.879 -1.642 .124 .867 1.153 Education 808.756 389.387 2.077 .058* .649 1.541 Household size 87.803 335.391 .262 .798 .465 2.151 Income -18.505 189.368 -.098 .924 .674 1.483 Point-of-purchase intervention (PPI) dummy -1335.930 3836.057 -.348 .733 .019 53.043 Thriftiness -1126.270 736.761 -1.529 .150 .369 2.708 PPI * Thriftiness 271.005 1039.995 .261 .798 .017 58.824 *** Significant at 0.01 level ** Significant at 0.05 level * Significant at 0.10 level

Table 6: The moderating role of thriftiness on the relationship between point-of-purchase interventions and food waste (control variables included)

N = 22, R²-adjusted = .220 , F-stat = 1.740, p-value = .180

(28)

28 the relative low sample size of this study. No significance was found for the interaction variable between thriftiness and the point-of-purchase intervention. Therefore hypothesis 2a is not supported. Regarding control variables, education was found to be significant on a 90% confidence interval (B = 808.756), indicating that higher educated people waste more food compared to lower educated people. However, whereas the overall model is found to be insignificant, this result should be interpreted with caution.

DV Food waste

Independent variable B Std. Error T-stat P-value Tolerance VIF

Constant 5097.964 3898.066 1.308 .208 Age -2.852 18.733 -.152 .881 .468 1.513 Gender -921.809 541.936 -1.701 .107 .637 1.570 Education 959.828 302.230 3.176 .006*** .661 1.513 Household size 53.617 289.704 .185 .855 .433 2.311 After-purchase intervention (API) dummy -6031.298 2675.375 -2.254 .038** .021 47.472 Thriftiness -1609.407 611.037 -2.634 .017** .266 3.755 API * Thriftiness 1839.586 747.185 2.462 .025* .019 47.472 *** Significant at 0.01 level ** Significant at 0.05 level * Significant at 0.10 level

Table 7: The moderating role of thriftiness on the relationship between after-purchase interventions and food waste (control variables included)

(29)

29 used. The interaction term, between the after purchase intervention and thriftiness is found to be significant (p = .025, B = 1839.586), indicating that thriftiness moderates the relationship between the after purchase intervention and food waste. The positive B indicates that people high on thriftiness waste more food when using the after purchase intervention, thereby making the intervention less effective. Therefore it can be concluded that hypothesis 2b is supported. A graphical representation of the interaction effect between thriftiness and the after-purchase intervention can be found in figure 3. Similar to the previous tested model, education is found to be a significant predictor of food waste (p = 0.006, B = 959.828), implying that higher educated people waste more food. Finally, high VIF scores and low tolerance scores are found for the API and the interaction term and are accounted for in the robustness check.

Figure 3: Interaction effect between thriftiness and the after-purchase intervention

5.2.2. The moderating role of biospheric values DV Food waste

Independent variable B Std. Error T-stat P-value Tolerance VIF

Constant _-191.666 _2658.533 _-.072 _.944 Age _17.958 _19.707 _.911 _.379 _.460 _2.173 Gender _-632.377 _440.547 _-1.435 _.175 _.832 _1.201 Education _695.298 _278.782 _2.494 _.027** _.822 _1.217 Household size _422.650 _248.965 _1.698 _.113 _.548 _1.826 Income _175.421 _135.053 _1.299 _.217 _.860 _1.163 0 200 400 600 800 1.000 1.200 1.400 1.600 1.800

Low thriftiness High thriftiness

Foo

d

was

te

(30)

30 Point-of-purchase intervention (PPI) dummy -5048.586 2682.213 _-1.882 _.082* _.025 _39.962 Biospheric values _-766.515 _226.665 _-3.382 _.005*** _.591 _1.692 PPI * Biospheric values _786.121 _476.004 _1.652 _.123 _.025 _40.508

*** Significant at 0.01 level ** Significant at 0.05 level * Significant at 0.10 level

Table 8: The moderating role of biospheric values on the relationship between point-of-purchase interventions and food waste (control variables included)

In order to test if there is a moderating effect of biospheric values on the relationship of the point-of-purchase intervention (PPI) and food waste, the regression model presented in table 8 was computed. The overall model is found to be significant (p= .021, F = 3.561), with an R² of .687, explaining 68.7% of the variance in food waste. The moderating effect is found to be insignificant (p = .123, B = 786.121), and therefore hypothesis 3a is not supported. However, a significant effect between biospheric values and food waste is found (p = .005, B = -766.515), indicating that people high on biospheric values waste less food in general. Similar to the previous discussed regression model, a (moderately) significant relationship is found between the intervention and food waste (p = .082, B = -5048.586). The high B and the insignificant previous tests makes that this result has to be interpreted with caution. Education is, again, found to be a significant predictor of food waste (p = .027, B = 695.298). The high VIF scores and low tolerance scores of the interaction term and the intervention variable are accounted for in the robustness check.

DV Food waste

Independent variable B Std. Error T-stat P-value Tolerance VIF

(31)

31 Biospheric values _-768.599 _227.872 _-3.373 _.004*** _.339 _2.954 API * Biospheric values _639.853 _286.076 _2.237 _.039** _.055 _18.107

*** Significant at 0.01 level ** Significant at 0.05 level * Significant at 0.10 level

Table 9: The moderating role of biospheric values on the relationship between after-purchase interventions and food waste (control variables included)

(32)

32

Figure 4: Interaction effect between biospheric values and the after-purchase intervention

5.3. Robustness Checks

As mentioned before, liquids might account for a disproportional part of the total food waste, and therefore additional analysis has been performed to check for the impact of liquids. First of all the main effects were tested, with waste without liquids as dependent variable. No significant differences in the results were found (F = 1.479, p = .230). Additional analysis on the regression models, with food waste without liquids as dependent variable did not lead to new, mentionable, differences in results. As mentioned before, usage rates of the interventions were rather low. Excluding these cases for the initial analysis would lead to a non-workable data set. In the robustness check, the difference between users and non-user (but receivers) of the intervention were tested. First the difference in food waste in the point-of-purchase intervention group was measured. A non-significant difference (F = 1.120, p = .315) was found between the means of the users and non-users (users 1044.17, non-users 641.67). Excluding the non-users in the after-purchase group would lead to a sample size of 2, and therefore no statistical analysis were performed on this group.

5.3.1. Multicollinearity check

High VIF (Variance Inflation Factor) scores, along with low tolerance scores were observed in all the regression models, which can be an indication of multicollinearity between the concepts in the study. In order to account for multicollinearity, all the regression models were recalculated with the use of standardized variables for the interventions, thriftiness and biospheric values. No VIF scores above the threshold of four and no tolerance scores lower than the threshold of .02 were observed after the

0 500 1000 1500 2000 2500

Low biospheric values High biospheric values

Foo

d

was

te

(33)

33 recalculation of the regression models. The significance of the models and concepts did not differ from the initial models. Therefore it can be concluded that the high VIF scores and low tolerance scores in the initial models were no indication of multicollinearity.

5.3.2. Additional analysis

Effects between thriftiness and food waste and biospheric values and food waste were found when testing for moderation. These two variables were tested in a single model, with the same control variables, but no interaction variables. The regression model (F = 2.315, p = .059) showed that these direct relationships do not hold when the interaction variables are removed from the model.

(34)

34

6. Discussion

This study was conducted with the intent to research the relationship between food waste interventions and food waste reduction. More specifically, researching the role of thriftiness and biospheric values on this relationship was among the goals of this study. Different regression models were calculated in order to assess the role and impact of these two concepts on the relationship between food waste interventions and food waste reduction. This section will summarize the results presented in the previous sections. Furthermore, the derived insights from the results will be related to the theory identified in the literature section and possible explanations for the not supported hypotheses will be presented. Table 10 shows the hypotheses that stand central in this study, and whether support is found or not.

H1a Food waste interventions have a positive effect on food waste reduction

Not supported

H1b After-purchase interventions have a stronger effect on food waste reduction than point-of-purchase interventions

Not supported

H2a Thriftiness weakens the effectiveness of the point-of-purchase intervention on food waste reduction

Not supported

H2b Thriftiness weakens the effectiveness of the after-purchase intervention on food waste reduction

Supported

H3a Biospheric values strengthen the effectiveness of the point-of-purchase intervention on food waste reduction

Not supported

H3b Biospheric values weakens the effectiveness of the after-purchase intervention on food waste reduction

Supported

Table 10: Overview of the hypotheses

6.1. The effect of food waste interventions on food waste reduction

(35)

35 artefact of the low sample size used in this study. Regarding hypothesis 1b, a stronger effect was expected for the after-purchase intervention than for the point-of-purchase intervention (Aschemann-Witzel et al., 2016; Brown et al., 2014; Stancu et al., 2016; Terpstra et al., 2005; Wikström et al., 2014). Again, the means of the different groups did not differ significantly, even though the mean of the after-purchase group was lower than the point-of-after-purchase group. The lower mean of the after-after-purchase group would be in line with Stancu et al. (2016) who state that efforts taken to change the way people handle leftovers may have the largest effect on food waste reduction. Similar to the previous hypothesis, the high standard deviations, caused by the low sample size, might be an explanation to the insignificant results. An elaboration on the causes of the low sample size and other problems with the data will be given in the limitations section.

6.2. The role of thriftiness and biospheric values

It was expected that the effect of the point-of-purchase interventions would be weaker the higher the thriftiness of a person, whereas people high on thriftiness are prone to buying large economic sized packages, leading to food waste (Aschemann-Witzel et al., 2016). No support was found for this hypothesis, however the positive B value of the interaction term was in line with what was expected. Different explanations can be thought of regarding the insignificance of the effect of thriftiness. The overall model was found to be insignificant, and therefore it can be concluded that the model was not a good fit for the data. This can be due to, on the one hand, the low sample size of the study, and on the other hand the low usage rate of the interventions. Both will be elaborated on in the limitations section of this paper.

Regarding the after-purchase intervention, it was expected that for respondents with high thriftiness, the effectiveness of the intervention on food waste reduction would become weaker. This is supported by the results of this study, however the effect is stronger than expected. Based on the literature it was expected that people who behave in a thrifty way already display food waste reducing behaviour at home (Graham-Rowe et al., 2014). Therefore the effectiveness of the intervention was expected to be less strong, however, the results show that people high on thriftiness waste more food when receiving the after-purchase intervention than when they did not receive this intervention. A possible explanation for this phenomena can be the low usage rate and sample size of the after-purchase intervention, which might have biased the results.

(36)

36 did not indicate that this effect was in place, however the results did indicate that, although insignificant, that the relationship between the intervention and food waste reduction would become weaker for people with high biospheric values.

The last hypothesis expected that high biospheric values would weaken the effect of the after-purchase intervention on food waste reduction. The results of this study did support this hypothesis, which is clearly shown in figure 4. This is in line with the theory, which expected people who score high on biospheric values to already display food waste reducing behaviours in their after-purchase routines (Graham-Rowe et al., 2014; Ojea & Loureiro, 2007). As mentioned before, these results should be interpreted with caution, whereas the usage rate of the after-purchase intervention was low.

6.3. The effect of the control variables

(37)

37

7. Conclusions

The goal of this study was to study the relationship between two different food waste interventions and food waste reduction. Furthermore this study tried to explain the role thriftiness and biospheric values play in this aforementioned relationship. Found was that thriftiness and biospheric values moderate the relationship between the after-purchase intervention and food waste reduction in such a way that the intervention becomes weaker, the higher the thriftiness or biospheric values. The low sample size, in combination with the low intervention usage rates might have had a major influence on the results of this study. This section will present the academic and managerial contributions of this study and the limitations of this study will be discussed. Finally, possible avenues for further research will be identified.

7.1. Academic and managerial contributions

This study adds to the academic field whereas it is the first study, to the best of the author’s knowledge, to study the role of thriftiness and biospheric values on the intervention effectiveness regarding food waste reduction. Thereby this study tries to make the first step in filling that gap within the literature, and contributes to the academic field by helping explain the concept of household food waste prevention. This research can therefore be used as a starting point for further research in this direction. This research shows the importance of understanding consumer characteristics and the need to take these into account when designing food waste interventions. By understanding the role of thriftiness and biospheric values, interventions can be adapted to people with (or without) certain characteristics, to be extra effective. Furthermore, by understanding the role of consumer characteristics on the intervention effectiveness, geographical targeting can be used in order to find an target group with a potential for high effectiveness. For example, with the current after-purchase intervention, distribution of this intervention in front of a cheap supermarket will result in less effectiveness, whereas it can be expected that a lot of the visitors will score high on thriftiness. The insights provided by this research can also help in creating new or adopting existing interventions to people high on thriftiness or biospheric values in order to effectively target these people as well. Finally this research gives insight in the effectiveness of the used interventions, for the Voedingsbank (Dutch Nutrition Centre). Consumer responses on the used intervention showed that an intervention should be easy to use and self-explainable.

7.2. Limitations

(38)

38 of waste for a week, complemented with a questionnaire of fourteen pages) might have been a reason for people to withdraw from the study. The low sample size is one of the main limitations of this study. This may have a large influence on the generalizability of this study, and the significance of the results. Therefore this study should be replicated on a larger scale in order to test the whether the results of this study hold.

The second observed problem regarding the data, were the low usage rates of the interventions. Excluding the cases that did not use the intervention was, although preferable, not feasible in this study. The low number of left cases in the study would have made the entire database statistically unworkable. Different explanations can be identified which could have caused the problem of the low intervention rate. It could be that the intervention was not clear for the respondents. In the end of the questionnaire, feedback was asked about the intervention. About the point-of-purchase intervention many people indicated that they found it hard to use the intervention, and doubts existed to what extent the intervention would be effective. Ease of use of the intervention was furthermore scored very low (most respondents scored this question two out of seven, seven indicating very easy to use). Same was observed for the after-purchase intervention, where usage rates were even lower. Almost all the people indicated that they did not see the use and effectiveness of the intervention. Also the intervention was not found to be pleasant to work with, which has a negative effect on the usage rate. Furthermore, people were asked to what extent they would like to use the intervention in the long run. Almost all people indicated that they would not like to use the intervention in the long run. Finally, it could have been the case that the researchers might not have been clear enough in explaining the necessity to use the intervention for the respondents. The low intervention usage rate has had a large influence on the significance of the results. Without using the intervention people might still have been primed towards sustainable consumption, however this is an assumption, and will never be the case for all the respondents. Therefore, the results of this study should be interpreted with caution, and results should be replicated in other studies before academic value can be derived.

7.3. Future research

(39)

39

References

Aschemann-Witzel, J., de Hooge, I., & Normann, A. (2016). Consumer-Related Food Waste: Role of Food Marketing and Retailers and Potential for Action. Journal of International Food &

Agribusiness Marketing, 4438(September), 1–15.

http://doi.org/10.1080/08974438.2015.1110549

Azjen, I. (1991). The theory of planned behaviour. Organizational Behaviour and Human Decision

Processes, 50, 179–211.

Bell, D. R., Corsten, D., & Knox, G. (2011). From Point of Purchase to Path to Purchase: How Preshopping Factors Drive Unplanned Buying. Journal of Marketing, 75(1), 31–45. http://doi.org/10.1509/jmkg.75.1.31

Bolderdijk, J. W., & Steg, L. (2015). Promoting Sustainable Consumption: The Risks of Using Financial Incentives. Handbook of Research on Sustainable Consumption, 328, 1–33.

Brown, T., Hipps, N., Easteal, S., Parry, A., & Evans, J. (2014). Reducing domestic food waste by freezing at home. International Journal of Refrigeration, 40, 362–369.

Cox, J., Giorgi, S., Sharp, V., Strange, K., Wilson, D. C., & Blakey, N. (2010). Household waste prevention — a review of evidence. Waste Management & Research, 28(3), 193–219. http://doi.org/10.1177/0734242X10361506

Darnton, A., Elster-jones, J., Lucas, K., & Brooks, M. (2006). Promoting Pro-Environmental Behaviour : Existing Evidence to Inform Better Policy Making. Defra, (October 2016), 84. Retrieved from

http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Promoting+Pro-Environmental+Behaviour+:+Existing+Evidence+to+Inform+Better+Policy+Making+by#0 de Groot, J. I., & Steg, L. (2009). Mean or green: which values can promote stable pro-environmental

behavior? Conservation Letters, 2, 61–66. http://doi.org/10.1111/j.1755-263X.2009.00048.x Defra. (2005). UK Sustainable Development Strategy, Securing the Future.

DeYoung, R. (1986). Encouraging Environmentally Appropriate Behaviour: The Role of Intrinsic Motivation. Journal of Environmental Systems, 15(4), 281–291.

Doorn, J. van. (2016). Commentary : Why Do We Waste So Much Food ? A Research Agenda, 1(1), 53–56. http://doi.org/10.1086/684462

Doorn, J. van, & Verhoef, P. (2015). Drivers of and Barriers to Organic Purchase Behaviour. Journal of

Retailing, 91(3), 436–450.

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. http://doi.org/10.1080/09581596.2011.608797

Flower, T., & Collet, K. (2015). Survey of existing consumer products and services which reduce food waste. WRAP & Shift. Retrieved from Retreived from www.

Göbel, C., Langen, N., Blumenthal, A., Teitscheid, P., & Ritter, G. (2015). Cutting food waste through cooperation along the food supply chain. Sustainability (Switzerland), 7(2), 1429–1445. http://doi.org/10.3390/su7021429

Household Food Waste Reduction: The role of thriftiness and biospheric values on the effectiveness of food waste interventions