The effectiveness of nudging in increasing recycling behaviour

(1)

The effectiveness of nudging in increasing

recycling behaviour

Abstract

Due to an increasing pressure on the environment and natural resources getting scarce, a more sustainable economy is needed like a circular economy. Nudging is a promising policy-making tool to push people in a sustainable direction. Thaler and Sunstein (2008) define a nudge as ‘’any aspect of design that alters people’s behaviour in a predictable way without forbidding any options or significantly changing their economic incentives.’’ In this study, it is investigated if nudges can be effective in increasing recycling behaviour and how characteristics influence the effectiveness of nudges. Previous studies discussed show that different nudges can increase recycling behaviour. Results of the questionnaire show that the three nudges used are effective in increasing recycling behaviour, with a difference in effectivity. Certain characteristics increase the effectivity of the nudges. It seems from previous studies and the questionnaire that access to recycle equipment and reduced distance to recycle equipment is the most effective nudge.

Name: Simon Cornel Student number: 10811362

Course: Bachelor’s Thesis and Thesis Seminar Economics Supervisor: Konstantinos Ioannidis

Words: 8975

(2)

2

Statement of Originality

This document is written by Simon Cornel who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents

(3)

3

In a world with an increasing population, the pressure on the environment becomes too big and natural resources are getting scarce. The economy needs to become more sustainable and a way to accomplish this is to change the economy in a circular economy (Lehner et al., 2016). A circular economy is an economy where resource loops are closed, which means systems in which resources are reused and recycled and energy is conserved (Preston, 2012). This means more sustainable decisions on both the production and consumption side of the economy (Lehner et al., 2016). A promising policy-making tool to help the economy make more sustainable decisions is the principle of nudging. Thaler and Sunstein (2008) introduced this principle and define a nudge as ‘’any aspect of design that alters people’s behaviour in a predictable way without forbidding any options or significantly changing their economic incentives. To count as a mere nudge, the intervention must be easy and cheap to avoid. Nudges are not mandates.‘’ Nudging can be used to push people’s behaviour in a certain direction and, in our case, towards more sustainable decisions.

There are studies about the use of nudging in the context of environment and sustainability. In this study, the focus is laid on recycling behaviour by households and how this can be influenced by nudging. Previous studies already show the effectiveness of nudging in increasing recycling behaviour. Barile et al. (2015) found that the effectiveness of incentives like a nudge depends on the characteristics and level of environmental morale of the target individuals, with a nudge being most effective for individuals with a high level of environmental morale. The nudge in this study is providing free recycling equipment and services. In the second study discussed, Milford et al. (2015) try to increase recycling with the use of a nudge in the form of two letters. The letters contain social comparative feedback, feedback on own performance and information and advice about household recycling. The results show that the letters increase recycling, with the largest increase seen directly after the letter and the effect of the letter becoming less after some months. The goal of the study of Bernstad (2014) is to find out the effect of two nudges on food waste separation by households. The two nudges are written information about food waste separation which is distributed as leaflets to households and the installation of waste separating equipment in kitchens of households. The results show that the written information has no significant effects on the amount of average separately collected food waste and the source-separation ratio, while the installation of recycling equipment increased both significantly. In the last study discussed, Samuelsen et al. (2016) want to show the effectiveness of three nudges in increasing recycling behaviour. One nudge is an informational nudge and the two other nudges are system nudges, in which the distance to the collection point is reduced or people get free access to waste bags. These nudges are used to test two different hypotheses. The results show that all three nudges are effective in increasing recycling behaviour, thus supporting the hypotheses of the study. The goal of this study was to investigate if nudges are effective in increasing recycling behaviour and how different characteristics influence the effectiveness of nudges.

Now the structure of the study is described. The second section explains the principle of nudging, when to use it, important aspects and some examples of nudges. The third section describes briefly a circular economy and shows previous studies about the use of nudge in the context of environment and sustainability. The fourth section discusses extensively four previous studies about the use of nudges for recycling and their effects on recycling behaviour. The fifth section describes the questionnaire and shows the motivation for the variables used in the questionnaire. The sixth section shows the results of the questionnaire and explains the results that are found. The seventh section is a summary and conclusion of the results and answers the research question. The eighth section discusses the questionnaire and its results and criticism on the use of nudges.

2. Nudge

There have been some advances in psychology and behavioural economics in the last decades, which showed insights that we systematically fail in our decision-making to achieve our preferred outcomes (Hansen et al., 2013). There is a need for policymakers to see citizens as ‘homo sapiens'

(5)

5 instead of ‘homo economicus' and ‘humans' instead of ‘econs' (Thaler & Sunstein, 2008). Policymakers must stop seeing citizens as perfectly rational creatures that always process information effectively in order to make perfect decisions, but rather see them as imperfect decision-makers driven by social norms, pressures and cognitive short-cuts.

Also in the area of public policy, these advances in psychology and behavioural economics demonstrate that not paying attention to these insights in decision-making can be responsible for public policy not having the desired effects (Hansen et al., 2013). Using these insights in decision-making more can be important for public policy to achieve better results in dealing with worldwide problems like global-warming, obesity and bad economic decision-making in general. Insights in the decision-making context and choice architecture might help in achieving the desired behaviour. This can be done by designing the choice architecture in a way that gently pushes people in the direction of making better decisions for their health, wealth and happiness (Thaler & Sunstein, 2008).

In their book ‘Nudge: Improving Decisions About Health, Wealth, and Happiness’ Thaler and Sunstein (2008) call this pushing of people in a certain direction ‘nudging’. Thaler and Sunstein (2008) define a nudge as: ‘’any aspect of design that alters people’s behaviour in a predictable way without forbidding any options or significantly changing their economic incentives. To count as a mere nudge, the intervention must be easy and cheap to avoid. Nudges are not mandates.‘’ Thaler and Sunstein (2008) also discuss the five situations when a nudge is needed. First, when choices have a delayed effect, which in short mean that their benefits arrive before the costs. Second, when choices are difficult. Third, when choices are infrequent and thus learning is hard. Fourth, when the feedback on decisions is bad. Fifth, when people cannot predict the outcome of a choice so the relation between both is ambiguous.

Nudges designed by choice architects must take into account five aspects because people are not aware of a nudge and it influences their decision-making (Selinger et al., 2011). First, a nudge must be inexpensive to use. Second, it must be easy to opt-out from the nudge. Third, the nudge does not change the financial incentives of the people. Fourth, a nudge has to be transparent for the people it targets. Fifth, nudges must be designed in the best interests of the people it is meant for. It is also important for choice architects that want to use a nudge if a situation is semantically variant or semantically invariant because this determines if a nudge is effective or not. With semantic variance, it is possible that the perceptions of meaning can be very different in reaction to changes in the decision-making context. With semantic invariance, these perceptions of meaning are not likely to be different in reaction to changes in the decision-making context. But the problem is how we can know or how we can determine if a situation is semantically variant or semantically invariant. Thus, it is hard to know if a nudge is going to be effective in a particular situation.

Thaler and Sunstein (2008) give some examples of nudges in their book. The first example they give is putting healthy food in school canteens at eye height, which must lead to students eating healthier food. But in contrast, banning unhealthy food is not a nudge. Another example they give is in the field of organ donor registration. Here the nudge is the use of an opt-out system instead of an opt-in system. With an opt-out system people are already enrolled as an organ donor, but they can opt-out if they actively choose that they do not want to be an organ donor. The goal of this nudge is to increase the rate of organ donors compared to the opt-in system, where you actively have to choose to be a donor. An example of a nudge in the environmental direction is the ‘environment-orb’. This is a ball which glows up red if consumers use too much energy and glow up green if the energy consumption of consumers is modest. In the first two weeks of the experiment consumers reduced their energy consumption because the orb showed the consumers what their energy consumption is.

3. Nudges for environmental purpose and sustainability

There is a need for change in the economy because due to an increasing population the pressure on the environment becomes too big and natural resources are getting scarce. A new kind of economy that can help counter these problems is the circular economy. The circular economy means that the whole economy needs to become more sustainable, so both the production and the

(6)

6 consumption side needs to change (Lehner et al., 2016). Resource loops should be closed to help the economy become a circular economy (Preston, 2012). This means that there is a need for systems where resources are reused and recycled and where energy is conserved, instead of open systems where resources are extracted and goods become waste. The use of a nudge as a policy-making tool has already been successful in for example public health campaigns and in saving for retirement (Lehner et al., 2016). It is also claimed that perhaps nudging can be a promising tool in policy-making, in order to help humans make more sustainable decisions.

The use of nudges in the context of the environment and sustainability are called by Schubert (2017) ‘green nudges’. The term ‘green nudges’ will be used in this paper to refer to nudges in the context of the environment and sustainability. The goal of green nudges is thus to promote more sustainable behaviour and they are becoming increasingly important in the area of policy-making (Schubert, 2017). Research has already been done on the use of green nudges as policy-making tools and also about the use of nudges to change the economy in a circular economy.

Ölander et al (2014) illustrates three different kind of nudges in the field of environmental policy in which the choice architecture can influence people’s choices in different ways. The first nudge is the EU’s mandatory energy label for electronic devices like refrigerators, dishwashers, and washing machines. This energy label has changed and that is likely to reduce the effectiveness of the first energy label. The label first had a classing from G to A, with A being the most energy-efficient class. This influenced consumers and producers to buy and produce more energy-efficient electronic devices, so after eight years about 90% was classed A. They re-designed the label with classes higher than A. This is likely to reduce the effectiveness of the energy label, because the class A was already the reference point or anchor and thus all A classes were seen more or less identical. The second nudge was an opt-out system instead of an opt-in system for the participation in the Smart Grid. The idea of Smart Grid is that the supply and demand of electricity can be better balanced using technological changes, so electricity consumption is remotely regulated by the electricity supplier. In this case the nudge is to make participation instead of no-participation in the Smart Grid the default, so an opt-out system instead of an opt-in system. This means you already participate and you have to make the ‘no-choice' if you do not want to participate. The results showed a higher participation rate with the opt-out system compared to the opt-in system. The third nudge can be seen as a social nudge for energy conservation. It seems that people who are uninvolved in an issue are likely to be more influenced by descriptive norms (so the behaviour of others) than people who are involved because for uninvolved people it is more an automatic process. It is showed that the behaviour of other people, the social nudge, has a significant impact on the energy conservation of uninvolved people. This is referred to as the herding effect.

Lehner et al. (2016) have made a review of research on the use of nudges as a policy tool to promote more sustainable behaviour. In their research they looked at the effect of four different nudges in the consumption domains of energy, food and mobility. The four different nudges are: 1) simplification and framing of information, 2) changes to the physical environment, 3) changes to default policy and 4) use of social norms. It is shown that most of the nudges are effective or can be effective in the three consumption domains, where one nudge is more effective than other nudges in the different domains. For example, it is shown that a change in default policy is more effective in the domain of energy than simplification and framing of information. Here the use of an opt-out system for green energy resulted in 95% to 99% staying with green energy, while feedback on energy use only resulted in about 2% energy less energy use. This difference in the effectiveness of a nudge for the different domains is shown in this review, which means different nudges should be used for different domains.

In his paper, Schubert (2017) provides an overview of the literature on pro-environmental nudges, referred to as ‘green nudges'. He distinguishes three variations of green nudges and gives representative examples of these three different green nudges. The first category of green nudges are the nudges that exploit people's desire for a pro-environmental self-image. This can be through making product information more simple or let people take notice of certain important product characteristics. The second category of green nudges are the nudges that make people want to imitate the behaviour

(7)

7 of other people. This can be done with feedback or peer comparison on pro-environmental social norms or encourage people to signal pro-environmental behaviour to other people. The third category of green nudges are the nudges where the desired pro-environmental behaviour is set as the default referred to as green defaults. Because people do not like to choose actively, setting green defaults can be used to increase pro-environmental behaviour. The effectiveness of the green nudges discussed in the paper of Schubert (2017) differ greatly, but many seem to have limited effect in promoting environmental behaviour. The green nudges that seem to be most effective in increasing pro-environmental behaviour, are the green nudges where the desired pro-pro-environmental behaviour is set as default. The effect of the different green nudges depends very much on the context in which a green nudge is used. Also, their effectiveness depends on if the green nudges are used as a complement to more traditional policy-making tools, which in most cases increases the effectiveness of green nudges.

4. Nudges on recycling behaviour

The previous section discussed studies about the use of nudges in a wide area in the context of environment and sustainability. In this section the focus is laid on a part of this area, namely recycling. As stated before, it is not only the supply-side which needs to become more sustainable but also the demand-side as consumers have to make more sustainable decisions (Lehner et al., 2016). This is because an increasing share of climate gases and other serious environmental impacts can be attributed to the consumption side of the economy (Ölander et al, 2014). Consumers are already aware that they are partly responsible for solving climate and other environmental problems. Policy makers use different kind of tools in order to change behaviour seen from an environmental perspective and thus to help humans make more sustainable decisions. The tools that have been most effective are taxes, subsidies, fee structure and legal regulation. But sometimes these tools are not effective in changing behaviour to what is desired. New insights from the behavioural economics can help policymakers better understand decision-making of humans and how this can be influenced in order for humans to make more sustainable decisions (Lehner et al, 2016). One of these new insights is the use of nudges in order to change behaviour. Recycling is also a tool for solving environmental problems and make the economy more sustainable. People need to make sustainable decisions and one of them is to increase their recycling. Thus, the focus will now be laid on the question of whether nudges are useful to increase people’s recycling behaviour. In the following subsections, the four studies stated in the introduction will be discussed extensively. First, the structure and design of the study will be described. After this, the results of the study will be presented. At last, the conclusion and discussion of the study will be discussed.

4.1. Barile, Cullis and Jones (2015)

The objective of this study is to compare the efficiency in increasing pro-environmental behaviour of different incentives and to find out if the efficiency of different incentives depends on characteristics of people. Knowing which incentives are efficient is important for policymakers to know which incentive to use to increase pro-environmental behaviour.

The first part of the paper is to describe the individuals in their willingness to recycle. The individuals used for the questionnaire are psychology and economics students from the universities of Bath (UK) and Florence (Italy). To describe the willingness to recycle they used socio-demographic/economic variables, attitudinal variables and a proxy for environmental morale in the form of willingness to protect the environment, all derived from the questionnaire. There are some important and significant results from this questionnaire.

First, the willingness to recycle tends to increase with age, but the slope becomes flatter when the second age (25-34 years) category is reached. Also, women are willing to recycle more than males and they score higher on environmental morale. Comparing economists and psychologists, the results show that psychologists are more willing to recycle and also have a higher level of environmental morale. This is similar to the hypothesis that economics students behave more like homo economicus, thus show more instrumental behaviour and free-riding. The attitudinal variables like risk aversion,

(8)

8 social responsibility and altruism all increase the willingness to recycle. Finally, the results of the regression show that environmental morale has a big positive impact on the willingness to recycle, increasing it almost 17% after a 1% increase in environmental morale. The impact of environmental morale decreases the higher the environmental morale score becomes, so individuals with a high environmental morale already contribute more to recycling.

In the next part, the authors want to show if these differences in characteristics are relevant for the impact of different incentives in increasing recycling behaviour. The two incentives used in this paper are a nudge and a shove. The nudge is providing free recycling equipment and services. The shove is the monitoring of recycling behaviour of individuals and fine them if they do not recycle household waste. Individuals were asked in the questionnaire if they would increase recycling, decrease recycling or exert the same level of recycling as a response to the two incentives. The impact of both incentives is compared to see which incentive is more effective in increasing recycling behaviour. Results show that there is invariance in the impact of the two incentives, referred to as policy invariance. This means that individuals increase their recycling for both the nudge and the shove with almost similar percentages. But when comparing the impact of the two incentives for different levels of environmental morale, the authors found differences. They found that individuals with low environmental morale increase their recycling behaviour more as a response to the shove (than to the nudge). In contrast, individuals with high environmental morale increase their recycling behaviour more as a response to the nudge (than to the shove). So the shove is more effective for individuals with low environmental value and the nudge is more effective for individuals with high environmental value.

The value of environmental morale is determined by the variables used in the questionnaire. Variables that increase environmental morale are: willingness to voluntary contribute to recycling activities, social responsibility and altruism. As stated before women have a higher environmental morale than males. The variables financial satisfaction and religion have a negative impact on the value of environmental morale. Individuals with low environmental morale are referred to as homo economicus, so they are more instrumental behaviour and free-riding. While individuals with high environmental morale are referred to as homo behavioural economicus, which can be seen as the opposite of homo economicus.

The first conclusion of the research is that, for choosing the right incentive, it is important to know what the characteristics of the target individuals are. The second conclusion is that it is important to know the level of environmental morale of individuals and the determinants of environmental morale because then the most efficient incentives to increase recycling behaviour can be chosen. For policymakers it is hard to determine the level of environmental morale of individuals, so finding the most efficient incentives is difficult. Barile et al. (2015) conclude that individuals react differently to incentives and the impact of incentives does not have to be different, so ‘one size will not fit all’. 4.2. Milford, Øvrum and Helgesen (2015)

The authors begin with looking at previous studies for drivers for increased recycling. Socio-demographic variables as higher income, larger households and higher education increase recycling. In addition, women recycle more than men. A kerbside collection system also increases recycling. They argue that policies that promote recycling are also very important, such as the distribution of information and feedback on recycling performance. Also, the development of recycling as a social norm and community pressure can increase recycling because people are more likely to recycle the more they think other people recycle.

The goal of this study is to find out what the effect of the letters is on household recycling and waste reduction. The letters contain social comparative feedback, feedback on own performance and information and advice, all about household recycling and waste reduction. The content in the letters are the nudges in this study. The field experiment took place in the city of Kristiansand in Norway. It included around 6000 treatment households divided in 3000 households for the recycle group and 3000 households for the quantity (waste reduction) group, and there was also a control group of 3000. The waste management agency of the area provided the households waste statistics for the

(9)

9 experiment. Only the effect of the letters for the recycle group compared with the control group will be discussed because that has importance for my research.

The recycle group is also divided into two groups, referred to as recycle group 1 and recycle group 2. Recycle group 1 was about one-third of the households and recycle group 2 about two-thirds of the households. The content of the first letter for recycle group 1 was information about its own waste sorting in 2012 and 2013 and the average of waste sorting in the same district in 2013. It also contained a motivation to increase waste sorting in 2014. The content of the first letter for recycle group 2 was the same as recycle group 1, but also contained a page with advice for improving waste sorting and it showed benefits of recycling. The second letter was the same for both recycle groups. It contained information about people themselves on how waste sorting had changed from January-June 2013 to January-June 2014. It also showed the average change in waste sorting in the same district for the same periods.

The results of the experiment show that the average recycling level for the recycle group increased with 2 percentage points more than the control group. But there is a declining effect over time for the recycle group which decreases the average recycling level by 1.7 percentage points. For the two-thirds of the households who received the advice and benefits of waste sorting in the first letter, the increase in recycling level was 1.9 percentage points higher than for the one-third of the households who did not receive this. There is also a difference between households in the recycle group with a recycling share below or above the median in 2013. Households above the median increased their recycling by 1.1 percentage points more compared with the control group. For households below the median, this increase was 3.4 percentage points more compared with the control group. It is also important for the effect of the letter if households had increased or decreased their recycling share in the years 2012 and 2013, with the letter being more effective for households who increased it. The largest increase in recycling for the recycle group is seen directly after they received the letter. But after some months the effect of the letter becomes less. A reason for this can be that the last two months are July and August, which is in the summer and thus waste behaviour is different than other months. Without these two months the effect of the letter does not get less. At last, the authors found that it is important that there is trust in the recycling system. If people are convinced that their sorted waste does not get burned but is used for new materials, recycling can be increased.

4.3. Bernstad (2014)

This study focusses on household source-separation of food waste and if this can be increased with two different strategies. The author starts with mentioning previous studies about different factors that can have an effect on recycling by households. The two strategies that are going to be discussed in this study are written information about food waste separation which is distributed as leaflets to households and the installation of waste separating equipment in kitchens of households. This two strategies can be seen as nudges.

A residential area in Malmö (Sweden) was chosen for the study. The number of households used for this study was 1632 and all households were rental apartments. Households collect their food waste in paper bags which were distributed by the municipality for free. The paper bags could be disposed in a recycling building with a maximum of 100 metres to the recycling building for all households. In this recycling building households could also dispose other waste that can be recycled, like paper, glass, plastic and metals. This system of foods waste recycling had started two years before this study took place. After two years of the system the recycling rate of food waste was still very low, about 25%. Because the rate was still very low after two years, a project was started with the goal to at least double the recycling rate of two years. The project contains the two strategies mentioned above, with the two strategies performed on all the households. The waste separation rates for households were measured before and after the two interventions. To evaluate the outcomes of the intervention two methods are used: Registration of the amount of food waste in the paper bags and the analysation of food waste collected in paper bags and of residual waste. The aim of the study was to find out how effective both strategies are and thus which strategy is most effective.

(10)

10 The first intervention is the distribution of written information as leaflets, this is campaign A. The focus in the brochure was laid on motivating households by showing the environmental gains of food waste sorting. It informed households about the use of food waste for making biogas to use instead of fossil energy and fertilizers for farmlands. They also gave instructions for food waste sorting. It also contained contact information and web pages of the municipal waste management department, if people want more information. The second intervention is the installation of recycling equipment in the kitchens of households to increase the accessibility and convenience for food waste sorting, this is campaign B. Lack of space for recycling equipment is a reason for not sorting food waste. They developed special sorting equipment for sorting food waste in the paper bags and this was installed in autumn 2012. The written information given with the installation only contained sorting instructions. For campaign A the results show that the average separately collected food waste increased by 12% in first 10 weeks compared to the 10 weeks before the campaign, but this was not statistically significant. For the weeks 10-20 it was 10% higher and for the weeks 20-30 it was 7% higher. A decrease in source-separation ratio and an increase in the missorting ratio is seen after analyses of the waste composition. For campaign B the results show that the average separately collected food waste increased with 49% in the first 20 weeks and increased with 44% in the weeks 20-30 after the start of the campaign, with the increase being statistically significant. Analyses of waste composition for campaign B show the opposite results than of campaign A.

It can be concluded that campaign A, the distribution of written information as leaflets, does not have the desired effects. Both the amount of average separately collected food waste and the source-separation ratio do not show a significant increase after the campaign started. Possible explanations for this are: lack of knowledge of receivers, problems with the language, the time that the campaign started, the contradiction with waste minimizing and the information not getting the attention of the receivers. However for campaign B, the installation of recycling equipment in the kitchens of households, it can be concluded that it has the desired effects. Both the amount of average separately collected food waste and source-separation ratio show a significant increase after the campaign started, even months after the start of the campaign. This shows how important accessibility to recycling equipment and thus increasing convenience are for households to participate in waste recycling, which is also shown in previous studies. This campaign can also be seen as the creation of social norms in the field of waste recycling.

4.4. Samuelsen and Støyle (2016)

In this study, the goal is to investigate if the recycling behaviour of households can be improved with the use of reverse logistics and taking the intention-action gap into account. Three different nudges are examined to investigate their effectiveness in increasing this recycling behaviour. The first nudge is an informational nudge which tries to activate social norms and thus improve people’s motivation for recycling food waste. The second nudge is a system nudge in which the distance to the collection point for glass and metal is reduced. The third nudge is a system nudge in which people get free access to waste bags for plastic, food and residual waste and reusable bags for glass and metal, which increases the level of convenience for people. Two different hypothesis are tested to discover if the three different nudges are effective in changing recycling behaviour. The first hypothesis ‘System performance improves when supplier characteristics facilitate recycling behaviour’ is tested with the informational nudge. The second hypothesis ‘System performance improves when system characteristics facilitate recycling behaviour’ is tested with the two system nudges.

The empirical setting of the study is the Stovner district in Oslo. First, the characteristics of the inhabitants of the Stovner district and the characteristics of the waste management system in Oslo are shown. Waste must be collected in separate bags: green bags for food, blue bags for plastic and white bags for residual waste. These three bags must be put in the residual waste container and are collected by the waste management system. Other waste like glass and metal must be put in different bags and can be brought to a collection point in the neighbourhood. All the waste is being recycled and the recycled materials are used for new purposes.

(11)

11 The experiment consists of two parts. One part is a pick analysis before and after the nudge is introduced, in which the waste composition is analysed to find out actual recycling behaviour and the quality of recycling. After comparing the results of the actual recycling behaviour from before and after the nudge, the effectivity of the different nudges can be shown. The other part is a questionnaire which must help represent the demographic characteristics of the control and experimental groups and also help represent the intended recycling behaviour of the participants. It also helps describe how satisfied people are with the waste management system, what people’s beliefs are about the recycling behaviour of neighbours and beliefs about their own recycling behaviour.

First, the results of the informational nudge used to test the first hypothesis are going to be discussed. Before the nudge the control group sorted 52% of their total food waste and after the nudge this was 53%. For the experimental group 39% of their total food waste was sorted before the nudge and this increased to 56% after the nudge, thus the nudge increased recycling behaviour with 17%. Comparing this with the intention to recycle and thus the reported recycling behaviour from the questionnaire it is shown that there is an intention-action gap, with the experimental group reporting 83% and the control group reporting 92% of food waste sorted. Thus it is shown that there is a big gap between actual recycling behaviour and intentional recycling behaviour, but this gap has become smaller for the experimental group. It can be concluded that the informational nudge and thus activation of a social norm can be effective in improving food waste recycling.

Now the results of the two system nudges used to test the second hypothesis are going to be discussed. First, the results of the system nudge of reduced distance to the collection point for glass and metal are shown. The authors observed a decrease of 29% of glass and metal in the residual waste after the nudge compared with the amount of glass and metal in the residual waste before the nudge, but for the control group no change was found. Also, the amount of glass and metal in the containers of the experimental group increased steadily. 44% of the experimental group also responded to the questionnaire that they sorted more glass and metal after the nudge, but this was not found for the control group. For this nudge the results of the questionnaire and the pick analysis also show an intention-action gap, with the reported intended recycling behaviour being higher than the actual recycling behaviour for both groups. Based on this results it can be concluded that the system nudge of reduced distance to the collection point increased the recycling rate of glass and metal.

Now the results of the system nudge of free accessibility to waste bags for food, plastic, glass, metal and residual waste. The authors looked how the contamination of the different bags was to show the effect of the nudge on recycling behaviour. For the experimental group the number of contaminated bags for food and plastic decreased with respectively 10% and 9%, while for the control group this was respectively 0% and an increase of 2%. There are no useful results for residual waste because only 7% used the red bags meant for the residual waste. The results for the glass and metal bags are not yet available and a reason for this can be that the bags are big and thus takes a long time to fill up. It is thought that these bags influence the convenience to recycle glass and metal and help people start recycling glass and metal. For paper and cardboard no reduction in contaminated bags was found for both groups, but this was not one of the targeted waste. Based on this results it can be concluded that the system nudge of free accessibility to waste bags improved recycling behaviour. It is said that the two system nudges help increase the level of convenience and that this results in an improvement in recycling behaviour.

The overall conclusion is that the informational nudge and two system nudges improved the recycling behaviour of the experimental groups for the different types of target waste, while the recycling behaviour of the control group stayed constant. The intentions to recycle are similar for both groups. Thus there is support for the two hypotheses that are tested. The results from this study are in line with findings from previous studies.

5. Description of the questionnaire

The studies discussed in the previous section showed the effect of nudges on recycling behaviour and how different characteristics influence recycling behaviour. To investigate what the

(12)

12 effect is of three different nudges on recycling behaviour and how different characteristics influence this, a questionnaire has been set up. The first nudge is a letter every month with information about recycling and feedback on the performance of recycling compared with the rest of the neighbourhood. The second nudge is better access to recycling equipment like different kind of recycling bins (so installation of recycling equipment) and reduced distance to this recycling equipment. The third nudge is people who recycle (and can show this) get a green (good for the environment) sticker that they have to put on the door. In the appendix, the whole description of the three nudges can be found.

Variable Variable name Possible answers

Gender Gender Man or Woman

Age Age Open question

Student Student Yes or No

Highest level of education Educ Primary school, High school, MBO,

HBO, WO

Job Job Yes or No

Gross income Income Open question

Religious or religious raised Religion Yes or No

Married or Engaged MarEng Yes or No

Size of household HouseSiz One, Two, Three, Four, Five or more

Vegetarian Veg Yes or No

Already recycle Recycle Yes or No

Access to recycling equipment RecEquip Yes or No

Table 1: Variables used in questionnaire and possible answers

First respondents need to answer questions about their socio-demographic and economic characteristics. The choice of variables included in the questionnaire was based on various students. Gender is used because previous studies show that women recycle more than men (Milford et al., 2015; Barile et al., 2015; Samuelsen et al., 2016). The willingness to recycle increases with age which becomes less after the age category of 25 to 34 (Barile et al., 2015), while in other studies it is found that younger people recycle a smaller amount of their waste than older people (Samuelsen et al., 2016). The level of education is also important for recycling because higher education levels increase recycling (Milford et al., 2015), while other studies show no effect of the level of education (Samuelsen et al., 2016). The variable student has been added because recycling can be a convenience and time issue for students, recycling is not yet important enough or students do not have the space to store different kinds of waste. The variable student and the variable working are compared because it is thought that the two variables are opposite variables and have an opposite effect on recycling behaviour because most people study or work full time. Income and recycling are found to be positively correlated, thus a higher income increases recycling behaviour (Milford et al., 2015; Samuelsen et al., 2016). Barile et al. (2015) have found a negative relationship between religion and environmental value, with people with a higher environmental morale are more willing to recycle. Another study found that Buddhists behave more sustainable than Atheists and Christians, with sustainable also including recycling behaviour (Samuelsen et al., 2016). People that are married or engaged can feel peer pressure from each other to recycle, which can increase people’s recycling behaviour (Milford et al., 2015). Previous studies show that recycling increases when the size of households increase (Milford et al., 2015). Other studies show that more storage space and thus more space for recycling equipment can increase recycling, with storage space being associated with the size of the household (Samuelsen et al., 2016; Bernstad, 2014). Vegetarian people are said to have more affection with the environment and thus will often have a high environmental morale, which has a positive relationship with willingness to recycle.

Two other variables included in the questionnaire are if people already recycle (paper, glass, plastic, etc.) and if people have access to recycling equipment in their neighbourhood. Whether people

(13)

13 already recycle is important because it can be that nudges have different effects on people who already recycle or who do not recycle (Milford et al., 2015). For example, social comparative feedback nudges can even have an adverse effect on people's recycling behaviour because knowing that other people do not recycle or recycle less can be discouraging for people that already recycle. Previous studies show that having access to recycling equipment is positively correlated with recycling, which means that accessibility to recycling equipment can increase recycling behaviour (Bernstad, 2014; Barile et al., 2015; Samuelsen et al., 2016).

So three nudges are used in the questionnaire. The first nudge (from now on the informational nudge) is information about recycling and feedback on the performance of recycling compared with the rest of the neighbourhood. The information and feedback come in the form of a letter every month. The information contains what to recycle, what the benefits of recycling are and the nearest recycle point. The feedback states the statistics how many people in your neighbourhood recycle. This nudge can be effective because information about recycling can help people think more about recycling and show the benefits can influence people in their thinking that recycling is needed. Also giving feedback on performance compared with the rest of the neighbourhood and showing how many people recycle in your neighbourhood can result in people feeling peer pressure and thus motivates them to recycle.

The second nudge (from now on the accessibility nudge) is better access to recycling equipment like different kind of recycling bins (so installation of recycling equipment) and reduced distance to this recycling equipment. This nudge can be effective because giving people equipment to recycle and reduce the distance to recycling equipment lowers the boundary for people to recycle.

The third nudge (from now on the sticker nudge) is a green (good for the environment) sticker that people who recycle (and can show this) have to put on the door. This nudge can be effective because you now can see if people recycle or not which can result in people feeling peer pressure and thus motivates them to recycle.

6. Results questionnaire

First, the effectivity of the three nudges in increasing recycling behaviour is looked at. As a response to the informational nudge, 71,0% of the respondents would increase their recycling. As a response to the accessibility nudge, 87,0% of the respondents would increase their recycling as a response to the nudge. As a response to the sticker nudge, 34,8% of the respondents would increase their recycling as a response to the nudge.

The effectivity of three nudges is compared. This is done by comparing the means of the nudges, see table 8 (Appendix A). A significance level of 10% is chosen because the number of respondents is 69 and thus is relatively small. The results show that the accessibility nudge is 52,2% more effective than the sticker nudge, with the difference in effectivity located on the interval of 41,0% to 63,4%. This difference is significant with p-value<0,00001. Further, the results show that the accessibility nudge is 15,9% more effective than the informational nudge, with the difference in effectivity located on the interval of 6,4% to 25,4%. This difference is significant with p-value is 0,007. Last, the results show that the informational nudge is 36,2% more effective than the sticker nudge, with the difference in effectivity located on the interval of 24,3% to 48,1%. This difference is significant with p-value<0,00001. It can be concluded the accessibility nudge is most effective, followed by the informational nudge and the sticker nudge being least effective.

Now the influence of the different characteristics on the effectivity of the three nudges is discussed with the help of three regressions. This is done by calculating the odds ratio (OR) with the B coefficients, where OR = 𝑒𝐵.

(14)

14 Table 2: Regression informational nudge

The only significant variable influencing the effectivity of the informational nudge is size of household because its p-value is 0,057. For size of household the OR = e0,629 = 1,876. This means that for every person added to a household the odds of the nudge being effective increases by a factor of 1,876.

The next variables seem to increase the odds of the nudge being effective: Being a woman, getting a year older, MBO, HBO, WO, not having a job, not being religious or religious raised, not being married or engaged and not being vegetarian. The next variables seem to decrease the odds of the nudge being effective: Not being a student, increasing gross income per year with thousand euros, if people do not already recycle and if people do not have access to recycle. All these variables are insignificant, so it cannot be said with certainty that they affect the odds of the nudge being effective.

(15)

15 Table 3: Regression accessibility nudge

The significant variables influencing the effectivity of the accessibility nudge are religious or religious raised and size of household, with respective p-values of 0,049 and 0,031. The variable age is also quite interesting because it is marginally significant, with a p-value of 0,104. For religious or religious raised the OR = e2,930 = 18,728. This means that the odds of the nudge being effective is 18,728 greater for people that are not religious or religious raised than for people that are religious or religious raised. For size of household the OR = e1,748= 5,743. This means that for every person added to a household the odds of the nudge being effective increases by a factor of 5,743. For age the OR = e0,115 = 1,112. This means that for every year someone gets older the odds of the nudge being effective increases by a factor of 1,112.

The next variables seem to increase the odds of the nudge being effective: Not being a student, MBO, HBO, not having a job and if people do not already recycle. The next variables seem to decrease the odds of the nudge being effective: Being a woman, WO, increasing gross income per year with thousand euros, not being married or engaged, not being vegetarian and if people do not have access to recycle. All these variables are insignificant, so it cannot be said that they affect the odds of the nudge being effective.

(16)

16 Table 4: Regression sticker nudge

The only significant variable influencing the effectivity of the sticker nudge is gross income because its p-value is 0,046. For gross income per year in thousands of euros the OR = e0,032 = 1,033. This means that for every increase of thousand euros in gross income per year the odds of the nudge being effective increases by a factor of 1,033.

The next variables increase the odds of the nudge being effective: Getting a year older, MBO, HBO, WO, not having a job, increasing gross income per year with thousand euros, not being religious or religious raised, not being married or engaged and not being vegetarian. The next variables decrease the odds of the nudge being effective: Being a woman, not being a student, if people do not already recycle and if people do not have access to recycle. All these variables are insignificant, so it cannot be said that they affect the odds of the nudge being effective.

7. Discussion

This study has some drawbacks that need to be discussed. First, it cannot be said that the questionnaire gives factual results. People can answer whatever they want to questions in a questionnaire, but it cannot be said that they do not lie and would do the same in real life. To find out what people would do in real life a field experiment has to be done, in which the three nudges used in the questionnaire are tested in real life. Another drawback of the questionnaire is that the number of respondents is only 69 which is relatively small. A small number of respondents has an influence on the significance of the regression results of the questionnaire and has therefore perhaps ensured that

(17)

17 so few variables are significant. The last drawback of the questionnaire is that it is not known how much the respondents really recycle, like in percentage of total household waste. Now it is only shown if people would increase their recycling as a response to the nudge and not with what percentage they change their recycling. It can be important to know the actual recycling data of the respondents to get more significant results. For further research it is thus important that besides the questionnaire an field experiment is done, a larger number of respondents is used and to know the actual recycling data of the respondents.

It is also important to take into account the ethical criticisms of nudges. A point of criticism is whether the freedom of choice is really preserved with nudging because people are unaware that their decisions are influenced by a designed choice architecture (Selinger et al., 2011). For people that value their freedom of choice in a high degree, it can be a matter that their choice architecture is designed in a particular way. It is thus important that nudges are designed in such a way that the freedom of choice is preserved. Also, it can be that choice architects fail to design the choice architecture in a good way, which can lead to pushing people into making wrong decisions. An example is that choice architects can select their values and preferences when designing choice architecture, while it can be that these values and preferences are not in line with the values and preferences of the people getting nudged. Another criticism is that people make decisions in another way when their decisions are nudged compared to when people need to make decisions themselves (Selinger et al., 2011).. This can result in people making the right decisions when nudges are used in a choice architecture, while they make wrong decisions for the same issue when their choice architecture is not influenced by nudges. A drawback of this is that people do not learn to make good decisions and expect others to push them in the direction of good decisions. This can lead to people being less bothered that their decisions are controlled by other people, which can be a bad thing because then they simply follow without thinking.

8. Conclusion

The goal of this study was to investigate if nudges are effective in increasing recycling behaviour and how different characteristics influence the effectiveness of nudges. First four studies about nudging and recycling behaviour were discussed, after which a questionnaire was done to investigate the effectivity of three nudges and which characteristics influence the effectivity of these three nudges.

Barile et al. (2015) found that a nudge is most effective for individuals with a high level of environmental morale. Milford et al. (2015) showed that a nudge in the form of letters increases recycling, with the largest increase seen directly after the letter and the effect of the letter becoming less after some months. Bernstad (2014) found that a nudge in the form of written information did not have significant effects on recycling behaviour, while a nudge in the form of the installation of recycling equipment increases recycling behaviour significant. Samuelsen et al. (2016) showed with their study that an informational nudge, a nudge in the form of reduced distance to the collection point and a nudge in the form of free access to waste bags are all effective in increasing recycling behaviour.

The results of the questionnaire show that all three nudges used in the questionnaire are effective in increasing recycling. The most effective is the accessibility nudge (87% would increase recycling), followed by the informational nudge (71% would increase recycling) and the least effective is the sticker nudge (34,8% would increase recycling). The effectivity of the three nudges is compared and it is shown that the differences in effectivity between the nudges are significant.

After this, there is looked at the influence of different characteristics on the effectivity of the three nudges. The only significant variable influencing the effectivity of the informational nudge is size of household, with for every person added to a household the odds of the nudge being effective increases by a factor of 1,876. The significant variables influencing the effectivity of the accessibility nudge are religious or religious raised and size of household are significant, while the variable age is marginally significant. For people that are not religious or religious raised the odds of the nudge being effective is 18,728 greater. For every person added to a household the odds of the nudge being effective increases by a factor of 5,743. For every year someone gets older the odds of the nudge being

(18)

18 effective increases by a factor of 1,112. The only significant variable influencing the effectivity of the sticker nudge is gross income, with for every increase of thousand euros in gross income per year the odds of the nudge being effective increases by a factor of 1,033. It can be that other variables also influence the effectivity of the nudges, but that due to the small number of respondents these results are not found.

It can be concluded that it seems that nudges can be effective in increasing recycling behaviour, while still a lot of further research is needed. There is a difference in the effectivity of different nudges, with previous studies and this study showing a nudge in the form of access to recycle equipment and reduced distance to recycle equipment being the most effective nudge. Also, it seems that certain characteristics influence the effectivity of different nudges, so it is maybe better to use certain nudges for people or groups of people with the same certain characteristics.

(19)

19

References

Barile, L., Cullis, J., & Jones, P. (2015). Will one size fit all? Incentives designed to nurture prosocial behaviour. Journal of Behavioral and Experimental Economics, 57, 9-16.

Bernstad, A. (2014). Household food waste separation behavior and the importance of convenience. Waste management, 34(7), 1317-1323.

Hansen, P. G., & Jespersen, A. M. (2013). Nudge and the manipulation of choice: A framework for the responsible use of the nudge approach to behaviour change in public policy. European Journal of Risk Regulation, 4(1), 3-28.

Lehner, M., Mont, O., & Heiskanen, E. (2016). Nudging–A promising tool for sustainable consumption behaviour?. Journal of Cleaner Production, 134, 166-177.

Milford, A. B., Øvrum, A., & Helgesen, H. (2015). Nudges to increase recycling and reduce waste. Ölander, F., & Thøgersen, J. (2014). Informing versus nudging in environmental policy. Journal of Consumer Policy, 37(3), 341-356.

Preston, F. (2012). A Global Redesign?: Shaping the Circular Economy.

Samuelsen, A. G., & Støyle, R. V. (2016). The power of nudging; Using reverse logistics to improve recycling behaviour in household waste management while taking the intention-action gap into account.

Schubert, C. (2017). Green nudges: Do they work? Are they ethical?. Ecological Economics, 132, 329-342.

Selinger, E., & Whyte, K. (2011). Is there a right way to nudge? The practice and ethics of choice architecture. Sociology Compass, 5(10), 923-935.

Thaler, R. H. & Sunstein, C. R. (2017). Nudge: Improving decisions about health, wealth and happiness (7th edition). Amsterdam, Belgium: Business Contact.

(20)

20

Appendix A: Tables

Nudge 1: Information about recycling and performance feedback

Frequency Percent Valid Percent

Cumulative Percent

Valid Yes 49 71,0 71,0 71,0

No 20 29,0 29,0 100,0

Total 69 100,0 100,0

Table 5: Frequencies informational nudge

Nudge 2: Better access and reduced distance recycling equipment

Cumulative Percent

Valid Yes 60 87,0 87,0 87,0

No 9 13,0 13,0 100,0

Total 69 100,0 100,0

Table 6: Frequencies accessibility nudge

Nudge 3: People that recycle get green sticker to put on the door

Cumulative Percent

Valid Yes 24 34,8 34,8 34,8

No 45 65,2 65,2 100,0

Total 69 100,0 100,0

Table 7: Frequencies sticker nudge

(21)

21

Appendix B: Questionnaire

Beste bezoeker, welkom bij deze vragenlijst. U bent uitgenodigd om deel te nemen aan een onderzoek. Uw medewerking is erg belangrijk voor mij, omdat uw deelname mij helpt in het

voltooien van mijn scriptie. De verzamelde gegevens worden anoniem geanalyseerd. Bovendien is uw deelname volledig vrijwillig. Wanneer u nog vragen, opmerkingen of moeilijkheden hebt, kunt u contact opnemen met Simon Cornel via simoncornel@live.nl. Ook als u de resultaten van het onderzoek wilt ontvangen kunt u mij een mail sturen. Het invullen van de vragenlijst duurt minder dan 5 minuten. Vanaf nu gaat de vragenlijst door in het Engels.

1. What is your gender? A Man

B Woman

2. What is your age? ……….

3. Are you a student (WO, HBO, MBO, etc.)? A Yes

B No

4. What is your highest level of education? A Primary school B High school C MBO D HBO E WO

5. Do you have a job? A Yes

B No

6. What is your average gross (bruto) income per year (in euro's)? ……….

7. Are you religious or religious raised? A Yes

B No

8. Are you married or engaged? A Yes

B No

9. What is the size of your household? A One B Two C Three D Four E Five or more

10. Are you a vegetarian? A Yes

(22)

22 B No

11. Do you already recycle (paper, glass, plastic, etc.)? A Yes

B No

12. Do you have access to recycling equipment like bins (near your house)? A Yes

B No

13. Imagine you get information about recycling and get feedback on the performance of recycling compared with the rest of the neighbourhood. The information and feedback comes in the form of a letter every month. The information contains what to recycle, what the benefits of recycling are and the nearest recycle point. The feedback states the statistics how many people in your neighbourhood recycle. Would you increase your recycling behaviour?

A Yes B No

14. Imagine you get better access to recycling equipment like different kind of recycling bins (so installation of recycling equipment) and reduced distance to this recycling equipment. Would you increase your recycling behaviour?

A Yes B No

15. Imagine that people who recycle (and can show this) get a green (good for the environment) sticker that they have to put on the door. You now can see if people recycle or not. Would you increase your recycling behaviour?

A Yes B No

The effectiveness of nudging in increasing recycling behaviour