Stockholm Resilience Centre: Exploring Tools To Promote Pro-Environmental Behaviour

Noah Linder Master´s Thesis 2015/2016
Stockholm Resilience Centre

Research for Biosphere Stewardship and Innovation
Master’s Thesis, 60 ECTS

Social-ecological Resilience for Sustainable Development Master’s
programme 2013/15, 120 ECTS
Exploring tools to promote pro-

environmental behaviour:
Can behavioural insights from psychology and behavioural economics be utilized to
increase the recycling of food waste in Hökarängen, Sweden?
Noah Linder
Exploring tools to promote pro-environmental behaviour:

Can behavioural insights from psychology and behavioural economics be utilized to
increase the recycling of food waste in Hökarängen, Sweden?
Author: Noah Linder

Supervisors: Sara Borgström and Therese Lindahl
Abstract
Human behaviour need to change in order to avoid potentially catastrophic global
environmental changes, thus promoting pro-environmental behaviour amongst individuals is
one of today’s greatest challenges. This thesis presents a methodological approach for
developing and designing implementations to promote pro-environmental behaviour, using
behaviour insights from psychology and behavioural economics, mainly by combining
theories from nudging and social marketing. The method was tested on a case study in
Hökarängen, - a suburb of Stockholm (Sweden) and was used to design an information hand-
out aimed at encouraging the pro-environmental behaviour of recycling food waste. The effect
of the implementation was then tested through a natural field experiment and evaluated using
a difference-in-difference analysis. The results indicate a significant increase in food waste
collected and a significant decrease in unsorted household waste compared to a control group
in the research area. This study provides a promising methodological approach for fostering
behaviour change: and contributes to advancing the scientific fields of green-nudges and
social marketing to nurture pro-environmental behaviour. Insights from this study could be
used to guide development of policy tools to help Stockholm and Sweden reach
environmental policy goals.
Key word: Pro-environmental behaviour, Nudging, Social Marketing, Food waste recycling,
Behaviour change
TABLE OF CONTENTS
1 Introduction ............................................................................................ 1
1.1 Behaviour change as leverage points to accelerate a sustainable transformation ................. 1
1.2 Unsuccessful approaches to behavioural change .................................................................... 2
1.3 The scientific aim of the thesis ................................................................................................. 3
2 Theoretical Framing ................................................................................. 4

2.1 Pro-environmental behaviour change in psychology .............................................................. 4
2.2 Social marketing ....................................................................................................................... 5
2.3 Behavioural economics and nudging ....................................................................................... 6
2.3.1 Critique of nudging and social marketing ........................................................................ 7
3 Methodology ........................................................................................... 8
3.1 Study location and case............................................................................................................ 8
3.1.1 Stockholmshem ................................................................................................................ 8
3.1.2 Hökarängen city district ................................................................................................... 8
3.2 Methodological approach to promote pro-environmental behaviour.................................. 11
3.2.1 Pilot study....................................................................................................................... 15
3.2.2 Experimental design ....................................................................................................... 16
3.2.3 Evaluating the implementation...................................................................................... 17
4 Results ................................................................................................... 18
4.1 Findings from the Pilot study ................................................................................................. 18
4.2 Information Hand-out design ................................................................................................. 20
4.2.1 First page ........................................................................................................................ 20
4.2.2 Second page ................................................................................................................... 21
4.2.3 Third Page....................................................................................................................... 22
4.3 Measuring the effect of the information hand-out ............................................................... 23
4.3.1 The data set .................................................................................................................... 23
4.3.2 The information hand-out effects on food waste collected .......................................... 23
4.3.3 The information handout effects on household waste collected.................................. 25
4.3.4 Regression Analysis ........................................................................................................ 27
5 Discussion.............................................................................................. 29
5.1 Critical interpretation of the results. ..................................................................................... 29
5.2 Reflecting on the methodology approach and the Information hand out. ........................... 30
5.3 Potential of merging Social marketing and nudging theory .................................................. 31
5.4 A wider transformation context ............................................................................................. 32
Literature cited ............................................................................................ 33

Appendix ..................................................................................................... 38
Appendix A ......................................................................................................................................... 38
Appendix B ......................................................................................................................................... 39
Appendix C ......................................................................................................................................... 40
Appendix D ......................................................................................................................................... 41
Appendix E .......................................................................................................................................... 42
Appendix F .......................................................................................................................................... 43
Appendix G ......................................................................................................................................... 44
ACKNOWLEDGMENTS
Before we start, I want to send a big thank you to Stockholmshem for making this study
possible, especially Åsa Stenmark and Patrik Westman for taking this project on board, being
interested, supportive, and providing me with all the tools I needed to make this happened. I
want to thank Magdalena Östin at Stockholm Vatten and Fredrik Reinholdt at Ragn-Sells for
enduring endless e-mail conversations and helping me out with the data struggle, you saved
the thesis!
I want to thank my supervisors Sara Borgström and Therese Lindahl for all your guidance
throughout this whole project, without your feedback and support I would have gotten lost.
My wonderful class and thesis group for all the support and feedback, you made these last 2
very challenging years feel like a walk in the park, you are the best! Already missing those
long study sessions down in the “dungeon”.
Lastly but definitely not least I want to thank my girlfriend Diana for all your support, advice
and never-ending proof reading. My friends (you know who you are!) and my family for just
always being there for me.
1 INTRODUCTION
Different scholars and policy makers suggest a variety of pathways to a sustainable future.
But whether sustainability is achieved through political change, building resilience, creating
adaptive governance, promoting socio-technical transition, social-ecological transformation,
or relying on technological advances, one thing is most certainly true; to succeed, whichever
pathways chosen will require large scale behaviour changes. The link between environmental
problems and human behaviour is so strong that any sustainable solution will require people
to do things differently (Midden et al. 2007).
Many of the major environmental problems we are facing today are rooted in human
behaviour, global warming, biodiversity loss, deforestation, water and air pollutions, ocean
acidification and overfishing, to name a few (see e.g., Vlek and Steg 2007). In light of these
global challenges, the world united in an unprecedented way in the year of 2015, by signing
off on the COP 21 agreement at the Paris climate conference, merely months after United
Nations member countries signed off on the new Sustainable Development Goals. Global
agreements like these have the potential to become a powerful political vision that can support
the urgently needed global transformation to a sustainable future (Hajer et al. 2015). The
goals in these agreements are clearly stated, however, the road on how to reach them is not. It
is becoming increasingly clear that policy efforts today are not sufficient to solve these
problems and need to be complemented by measures aimed at changing individuals'
behaviour (Mont et al. 2014).
1.1 BEHAVIOUR CHANGE AS LEVERAGE POINTS TO ACCELERATE A SUSTAINABLE

TRANSFORMATION
Changing individual behaviour could be used as a leverage point to achieve some of the goals
in the global agreements. Even seemingly small behaviour changes can have big aggregated
impacts. A recent study estimates that an emissions reduction of 123 million tons of carbon
dioxide per year over 10 years (7,4% of US national emissions) can be achieved by altering
the behaviour in United States households and among non-business travellers (Dietz et al.
2009). The estimation was done by analysing different approaches to promote 17 different
household actions (e.g. switching to low-flow showerheads, efficient water heaters or more
fuel-efficient vehicles) only considering a switch to readily available technology with either
low or zero cost, attractive returns on investment, and without appreciable changes in
lifestyle, also accounting for the likelihood that the behaviour can be changed. Altering food
1
habits is considered to be the greatest leverage point for climate mitigation (IPCC 2014)and
replacing bottled water with tap water in the US has the potential to save 50 million barrels of
oil per year, including the energy consumed by shipping and refrigeration in the US alone
(Kazdin 2009). Furthermore, behavioural changes such as the ones mentioned above have the
potential to be achieved quickly and without new policies, technologies, and with little to no
reduction in well-being (Schultz and Kaiser 2012)
1.2 UNSUCCESSFUL APPROACHES TO BEHAVIOURAL CHANGE

Changing human behaviour has proven to be a problematic task, policy tools like information
based campaigns and advertising are commonly used to promote behaviour changes.
Unfortunately, when evaluated, they often fall short of achieving their goal (Mckenzie-mohr
2000). Large scale information and advertising campaigns also tend to be very expensive, in
one rather extreme example the American state of California launched a 200-million-dollar
advertising campaign promoting the installation of energy-efficient devices in households,
along with suggestions on behaviour changes that save energy (like closing windows on
sunny days). However when evaluated, no meaningful behaviour change was observed,
despite the big budget invested (Costanzo et al. 1986). Scholars in psychology and
behavioural economics highlight the fact that behaviour insights1, are often not utilized in the
design of campaigns and information strategies (Kazdin 2009; McKenzie-Mohr 2013). Even
though we are surrounded with messages trying to promote certain social and environmental
friendly behaviour today, it is clear that many of them still do not apply or misuse scientific
knowledge (Kazdin 2009) A common error is misusing descriptive norms (the perceptions of
which behaviours are typically performed) in a way that actually risks increasing the very
behaviour it is set out to discourage. For example park administration in Arizona’s National
Forest, tried discouraging visitors with signs that stated “Your heritage is being vandalized
every day by theft losses of petrified wood of 14 tons a year, mostly a small piece at a time.”
When researchers evaluated the use of these signs it became clear that they were actually
encouraging visitors to steal by displaying the descriptive norm of “a lot of people are
stealing souvenirs from the park”(Cialdini et al. 2006).
In the past decades policy makers and governments have turned their eyes towards using
behavioural insights as tools for policy making, and theories that utilise these, like nudging
1Insights from research in behavioural economics and psychology on how to influence decision-making and
promote behaviour changes
2
and social marketing, have become something like buzzwords amongst policy makers and
governments in the developed world (Thorpe et al. 2008; Lindahl and Stikvoort 2015) (see
below).
1.3 THE SCIENTIFIC AIM OF THE THESIS

This thesis is taking a problem oriented approach with the purpose of seeking solutions to the
problem statement (Creswell 2008). The study is interdisciplinary by combining behavioural
insights from different scientific fields, focusing on behavioural economics, nudging,
psychology and social marketing. The overarching scientific aim of this thesis is to:
1) Develop a methodological approach that will be used to design an implementation
aimed at promoting pro-environmental behaviour
2) Measure and evaluate the effect of the implementation on a case study using a natural
field experiment
3) Explore the compatibility of the theories and see if expanding the “toolbox” of
behavioural insights by combining these different scientific fields could enable more
adapt and contrasted strategies to promote pro-environmental behaviour
In addition to the main scientific aims presented above, the methodological approach is meant
to draft an easy step-by-step guide for policy makers and other actors aiming to foster pro-
environmental behaviour changes.
The case used to fulfil the research objectives was to design an information hand-out
promoting the recycling of food waste in Hökarängen, a city district in southern Stockholm
(Sweden) amongst residents in Stockholmshem apartments, a natural field experiment was
designed to test the following hypothesis:
I. The information hand-out increases the amount of food waste collected from the
sorting stations in the treatment group compared to the control group
Since food waste is being sorted out from household waste the following hypothesis is added
II. The information hand-out decreases the amount of household waste collected from the
sorting stations in the treatment group compared to the control group
3
2 THEORETICAL FRAMING
The choice was made to focus on research from psychology and behavioural economics to
explore individual behaviour changes and the mechanisms behind them, in the scope of the
thesis the focus was primarily on theories with the explicit goal to promote, in some way, pro-
environmental behaviour changes (see definition below).
2.1 PRO-ENVIRONMENTAL BEHAVIOUR CHANGE IN PSYCHOLOGY

Research on pro-environmental behaviour has long been on the agenda for psychology
researchers. Brewer and Stern (2005) stated that inducing pro-environmental behaviours in
individuals is one of the most important priorities for social and behavioural science research
in order to avoid potentially catastrophic global environmental change. There is a lot of
research within psychology using different terms for somewhat similar behaviour:
environmental significant behaviour, conservation behaviour, sustainability behaviour,
environmental protection, and preservation for example (Schultz and Kaiser 2012). To
incorporate the verity of term this thesis uses a broad definition of pro-environmental
behaviour as: any human behaviour that either benefits the environment, or harms it as little
as possible (Steg and Vlek 2009).
There are two main approaches to study pro-environmental behaviour in psychology: impact-
oriented approach and intention-oriented approach (Stern 2000). The intention-oriented
approach focuses on the individual’s internal motivation to engage in pro-environmental
behaviour; for example, their attitudes, moral values, connectedness to nature, or
environmental concern. It is the underlying motivation for the individual to conduct pro-
environmental behaviour, and thus the likelihood of the individual to engage in pro-
environmental behaviour in a certain situation. The impact-oriented approach focuses on the
outcome of the behaviour and makes little assumptions about the individual’s motivation
behind the behaviour. For example, promoting specific households changes like energy
conservation , and measuring the effect on energy used (see e.g., Allcott 2011) is an impact-
orientated approach, whereas using information campaigns to promote general environmental
awareness (see e.g., Maibach 1993)is an intention-oriented approach.
There are benefits and limitations of both approaches. Individual values, beliefs, and norms
have shown to have some positive influence on environmental behaviour (Karp 1996) but it
has also been illustrated in numerous studies that there can be a great discrepancy between
attitudes and behaviour. Even though people convey pro-environmental attitudes and values it
4
does not necessary mean that they are showing more pro-environmental behaviours(Kollmuss
and Agyeman 2002). Why there is such a gap is not fully understood, but we know that we
are subconsciously justifying behaviours that contradict our beliefs so we do not have to feel
guilty about doing them (Cognitive dissonance (Festinger 1962)), we have difficulties
resisting short-term rewards, despite long-term negative consequences (Oskamp 2000) and we
state that we will act more according to our values in a temporal distance compared to the
present moment (Constual level theory (Trope and Liberman 2010)). This thesis will take an
impact approach, the aim is to promote and measure a behaviour change (increase recycling
of food waste) without necessarily trying to raise environmental awareness (although that
could of course be a side-effect).
Focusing on the actual behaviour and measuring the effect to ensure that a pro-environmental
change has indeed happened e.g. measuring the actual behaviour instead of self-reported
intentions or values might seem like a sound approach. However, focusing on changing only
one selected behaviour might be successful for that particular behaviour but it does not
necessarily entail promoting a pro-environmental behaviour in another. Research has shown
that acting environmentally-friendly in one aspect could lead people to perceive to have a
“license” to act less environmentally friendly in other areas (Moral-licensing, see e.g.,
Sachdeva et al. 2009) For example: buying green products like organic foods or Fair Trade,
can result in consumers feeling justified in making unethical purchases in other areas
(Jürgens, et al. 2010).
What is evident from psychology research is that human behaviour is a complex matter not
always easy to change, and the inherent difficulty of behaviour change could be a big obstacle
in the way of transformation towards sustainable development, which is exactly why we need
more research in this area.
2.2 SOCIAL MARKETING

Social marketing was first coined by Kotler and Zaltman (1971) and presented as a
framework to incorporate traditional commercial marketing techniques and insights such as
planning, pricing, communication, distribution, and marketing to design more effective social
campaigns. Social marketing seeks to influence behaviours that benefit individuals and
communities for the greater social good (Lefebvre 2016) One of the most recent and
prominent theories in social marketing is community based social marketing (CBSM), CBSM
5
originates as a guideline on how to make psychological knowledge visible to better develop

and deliver programs that promote pro-environmental behaviours.(Mackenzie Mohr’s, 2000).
CBSM has been tested and used with promising results across the world to promote pro-
environmental behaviours such as: inspiring residents to start back yard composting, changing
travel behaviour to reduce car usage and increase curb side recycling rates (McKenzie-Mohr
2000; Haldeman and Turner 2009). It has also been successful in delivering programs to
address human health issues (Athey et al. 2012). CBSM presents 5 steps to promote
behavioural changes: 1) Selecting behaviour 2) Identifying Barriers and Benefits 3)
Developing Strategies 4) Piloting and 5) Broad-scale Implementation (McKenzie-Mohr,
2013). This study will in part follow these steps when designing the information hand-out,
the method used in this study will be further explained below.
2.3 BEHAVIOURAL ECONOMICS AND NUDGING

Behavioural economics stray away from the traditional neoclassical economic model of
humans as being perfectly capable of making rational choices and process every available
information in order to maximize personal well-being (utility) (Mullainathan and Thaler
2000). Instead behavioural economics look at human behaviour in the light of having limited
cognitive capacity and inherent systematic cognitive failings, such as using heuristics
(cognitive rule of thumb), how framing information influence choices, being loss averse, or
falling for social pressure. All these deviations from the “rational choice” have been proven
and replicated in numerous studies both within the field of behavioural economics and in
psychology(Kahneman 2003).2 Being aware of these human flaws and cognitive tendencies
when designing social campaigns, could lead to more effective strategies in order to
encourage behaviour change for social good, which is what Thaler and Sunstein (2008)argue
in their book “Nudge” - Improving Decisions about Health Wealth and Happiness´. The
concept of nudging has since the release of the book become hugely popular, exemplified by
the creation of the Behavioural Insight Team in the UK in 2010 with the purpose to
collaborate with the government and provide consultancy on policies. Similar governmental
initiatives are seen in the US, Canada, Australia and New Zealand, and in Norway and
Denmark non-profit organizations for nudging have been initiated (Mont et al. 2014)
2) Kahneman, also provides a nice overview of these systematic cognitive flaws in his best-
selling book “Thinking fast and slow”.
6
One reason nudging became so popular amongst policy makers and governments could be
because it uses the concept of libertarian paternalism (Thaler and Sunstein 2003), which
essentially means that to be considered a nudge, the aim has to be to influencing people’s
behaviour to achieve a positive outcome for the individual or society at large (paternalistic),
without intruding on people’s freedom to choose (libertarian). For example, placing healthy
food at eye level in a grocery store to promote healthy food choices is considered a nudge, but
banning all unhealthy food in the store is not (Thaler and Sustain, 2008).
Despite the seemingly interlinking research using similar behavioural insight tools (even
refereeing to the same source literature in some cases), surprisingly little cross-referencing
exists between social marketing and nudging. Some critics of nudging have been heard from
social marketers (French 2006) arguing that nudging is too narrow of an approach, whilst
some other call for collaboration (Badshah 2010); no references to social marketing in the
nudge literature were found.
2.4 CRITIQUE OF NUDGING AND SOCIAL MARKETING

Both nudging and the concept of libertarian paternalism have been critiqued, mostly centred
on the paternalistic aspect of nudging. It has been argued that nudging is manipulative and
essentially tricking people into certain choices and behaviours (see e.g.;Marteau 2011;
Goodwin 2012). In response nudging advocates usually point out that all policy tools are
paternalistic and that nudging is comparably a “soft” policy measure, which can be used
instead of banning or taxing. (Thaler and Sustain, 2008). To avoid the manipulative aspect of
nudging, Hansen and Jespersen (2013) argue that the intention of a nudge needs to be clearly
stated. However, this is not the case in most nudges implemented today, and studies have
even shown that stating the intention of the nudge can severely lower the effectiveness of the
nudge (see e.g. Bronchetti, 2011). Nudging has also been critiqued for running the risk of
serving as an easy way out for policy makers that might use nudging as a relatively
uncomplicated and convenient policy measure instead of using more difficult and
controversial policy instruments such as to introduce laws and economic measures (Bonell et
al. 2011). Similar critiques have been targeted at Social marketing, Corner and Randall
(2011), state that “Social Marketing alone is insufficient to build support for the more
ambitious policy changes and interventions that constitute a proportional response to climate
change” (p.1). The critique towards the theories will be brought up in the discussion.
7
3 METHODOLOGY
In this study qualitative and quantitative methods are combined to develop and test a
methodological approach to promote pro-environmental behaviour, the approach is then tested
by designing an implementation and then evaluate its effect on a case study through a natural
field experiment.
The methodology section will first introduce Stockholmshem and the case study area
Hökarängen, then present the methodological approach developed, and explain how it was
applied on the case.
3.1 STUDY LOCATION AND CASE

3.1.1 Stockholmshem
This study was conducted in collaboration with Stockholmshem, Stockholm's largest real-
estate company, housing more than 50,000 residents (about 5% of total Stockholm
municipality population). Stockholmshem is owned by the municipality, they have started
providing their residents with the possibility to recycle food waste by installing curb side
sorting stations in line with Swedish policy goals (See BOX 1). Installing the sorting stations
is only half the journey, and they serve little purpose if only few people are using them. The
aim the collaboration was therefore to designing an implantation to promote the pro-
environmental behaviour of sorting food waste and to see a significant increase in food waste
collected. Stockholmshem presented a unique opportunity to evaluate the methodological
approach and test the implementation with detailed data on both food waste and household
waste collected from the area.
BOX 1. Swedish policy goals for collecting food waste
Around 7% of food waste in Sweden was collected in year 2012, the policy goal is to increase this
number to 50% before year 2018. Furthermore the municipality of Stockholm has in their waste
management plan set the target for food waste collection to 70% in year 2020, which equals an
increase from 16.000 ton collected (in 2015), to 68.000 ton in 2020. Collecting and using food
waste to produce bio gas also contributes to fulfilling Stockholm environmental objective which
aims to reduce greenhouse gas emissions to maximum of 3 tonnes of CO2 equivalents per capita
by 2015.
3.1.2 Hökarängen city district

The study was carried out in the southern part of Stockholm, in the city district Hökarängen.
Hökarängen became popular in the mid-20th century, and saw a booming of residents that
8
grew from just around 45 people in 1946 peaking at 16 000 just six years later. Today around
9000 people live in Hökarängen.
Figure 1. A map of the research area, the borough Hökarängen is highlighted in red in the left map, and marked with a red
pin on the right locating the area in Sweden.
The area is dominated by apartment complexes mainly consisting of two and three room
apartments, Stockholmhem is prominent in the area and owns around 2900 of the total 4700
apartments. Hökarängen could, a bit bluntly, be called for a stereotypical southern (south of
Stockholm) working class suburb, with strong political support for the more socialistic
leaning political parties, a slightly below average income and enrolment in education above
high school (see Table 1). Demographical traits, such as voting behaviour could be relevant to
the study as it has been documented that it can influence how susceptible people are to
nudges(Costa and Kahn 2013) . How such demographical traits might affect the effectiveness
of the implementation, and especially the generalizability of the study will be addressed
further below in the discussion section.
9
Table 1. Demographic data for Hökarängen compared to the Stockholm (municipality) average, the data is received from
Statistik om Stockhom (n.d), and collected in year 2014.
Demographic data Hökarängen Stockholm average

Unemployment rate 3,7 % 3,0 %
Mean income (16 years and over) 248 200 SEK 340 800 SEK
Higher education 40,4 % 56.9 %
Foreign background (not born in Sweden) 26 % 24 %
Politic votes in 2014 election. 68.7% 49,8%
Socialistic block
Politic votes in 2014 election. 19,7% 43,5%
Liberal/conservative block.
The majority of households in Hökarängen do not have the possibility to recycle their food
waste, however, Stockholmshem has recently started a project to provide their residents with
curb side sorting stations for recycling food waste. This initiative started 2014 in the research
area (Figure 2)
Figure 2. A satellite picture of Hökarängen, the black square shows the research area within Hökarängen.
Today all apartments in the research area have access to a sorting station (the last ones were
installed in the fall of 2015). Stockholmshem plans to eventually install sorting station for all
10
their apartments in Hökarängen. The research area consists of 474 households (about 10 % of
all households in Hökarängen) in area-typical apartment complex buildings.
3.2 METHODOLOGICAL APPROACH TO PROMOTE PRO-ENVIRONMENTAL BEHAVIOUR

Interventions aimed to promote pro-environmental behaviour are generally more effective
when systematically planned, implemented and then evaluated (Steg and Vlek 2009). In an
integrated review of the pro-environmental behaviour research four key phases were
identified as recurring suggestions for how to promote pro-environmental behaviour: (1)
identification of the behaviour to be changed, (2) examination of the main factors underlying
this behaviour, (3) design and application of implementation to change behaviour (4)
evaluation of the effects of implementation (ibid). These phases are related to the ones
presented in CBSM presented above, similar requiring phases can also be found (slightly
different) in the Nudge literature (see e.g., Ly et al. 2013). The Nudge literature especially
highlights the importance of evaluating the effect of the implementation and to adapt
strategies to different contexts (Mont et al. 2014)
The methodological approach presented in this study was highly influenced and framed by
these requiring phases (mainly from social marketing) and presented below (Figure 3). It
combines the frameworks proposed by Steg and Vlek (2009) and Mckenzie-Mohr (2013), and
expands it further by including behavioural insights found in the literature stemming from
behavioural economics and nudging in the design process. (e.g., from Sunstein 2014)
11
Figure 3. A conceptual figure of the methodological approach, the approach is mainly based on the phases presented by Steg
and Vlek (2009) and McKenzie-Mohr (2013). The italic text provides an overview on what was done in 6 phases in the case
study Hökarängen, and the arrows and boxes describe the design process of the implementation.
12
Below follows an explanation of each phase in the methodological approach 3, and a short
description of how the phase was applied in Hökarängen.
(1) Selecting behaviour to change
The first step is to identify what behaviour to change. Focusing on behaviours with
significant negative environmental impacts, taking to account the feasibility of change is
advised. (see e.g., Steg and Vlek 2009)
For this study the choice was made together with Stockholmshem to focus on promoting the
pro-environmental behaviour of recycling food waste.
(2) Examination of the main factors underlying the behaviour
To examine the main factors underlying the behaviour the method suggests two steps; a pilot
study, and reviewing of literature. The main goal of the pilot study is to learn about the area
and uncover context specific barriers (and benefits) for the desired behaviour.
“If any form of sustainable behaviour is to be widely adopted by the public, barriers to
engaging in the activity must first be identified” (c.f. Mackenzie-Mohr, p.5. 2006).
Different behaviours will have different barriers in different contexts. Barriers can be both
internal: individual motivation like moral concerns or normative influences, or external:
barriers that varies for each community such as accessibility, convenience, or cost of changing
the behaviour (Steg and Vlek 2009)
The goal of a literature review is to explore and learn about behavioural insight tools that
could be used in the implementation, suggesting to focus on the literature from social
marketing and nudging, but also looking at previous research on similar behaviour changes.
The main goal of the pilot study in Hökarängen, was identifying barriers for the residents in
the area to recycle food waste, but also to get an overview of resident’s attitudes towards
sorting food waste and to get a rough estimate of the residents already recycling food waste.
This was done using a mixed method approach including surveys, interviews and data
analysis (this process is presented in depth below under Pilot study).
3 For a more in-depth guide to each of the phases see, Steg an Vlek (2009) and Mackenzie-Mohr (2013)
13
(3) Designing the implementation
In this phase the discoveries generated in phase 2 are used to help select behavioural insight
tools suitable for addressing barriers uncovered, these insights are laying the foundation for
the design of the implementation (see Figure 3)
For this study an information hand-out, designed using the insights drawn from phase 2, was
used as an implementation. The barriers uncovered in the pilot study were addressed using
different behavioural insight tools (this process is presented in depth below, see Information
hand-out design).
(4) Executing Implementation on a smaller scale
In order to avoid expensive failures or unexpected results, like promoting unwanted

behaviour (See intro). A small scale implementation is first put in place and evaluated before
broadly implementing the strategy. This provides an opportunity to modify and adapt the
implementation if necessary, before the large scale implementation (McKenzie-Mohr 2013).
In Hökarängen, the information hand-out was sent out to 264 households in the research area.
(5) Evaluating the effectiveness of the implementation
To evaluate the effectiveness of the implementation design, a solid experimental evaluation

design with one or more ‘treatment’ groups and a comparable control group is strongly
advised (Steg and Vlek 2009). Furthermore measuring actual behaviour changes over self-
reported behaviour changes or intentions is preferable (McKenzie-Mohr 2013). Self-reports
have been shown to not always correlate well enough with observed behaviour (See e.g.,
Corral-Verdugo 1997)
To test the effectiveness of the information hand-out used in this study, a natural field
experiment was designed and implemented. The results were analysed using a difference-in-
difference method, with the outcome variable being the amount of food waste and household
waste collected in the research area (see Experimental design).
(6) Large scale implementation.
Only after testing and evaluating at a smaller scale, a large scale implementation of the
implementation is advised. This implementation should if possible also be evaluated (Steg
and Vlek 2009). This particular step is beyond the scope of this thesis but will be discussed at
the end.
14
3.2.1 Pilot study

The pilot study was carried out in two phases to meet two purposes; 1) to visit and learn about
the research area, hand out and analyse surveys and interviewing key individuals to identify
barriers and get a sense for the attitudes and norms in the area; and 2) to analyse food- and
household waste data with the purpose of deciding how to best divide the area it into a
control- and a treatment group.
A request to participate in a survey was distributed to 92 households (around 20%) in the

research area. The households targeted for the survey were evenly distributed among the
treatment group and the control group. Each apartment complex in the two groups was given
a number and then selected using a random number generator and flipping a coin (each house
had two entrances and the coin flip decided in which entrance the request would be handed
out in). The targeted households were provided with a short information about food waste and
asked to answer 9 short questions about their household’s food and household waste habbits
(see Appendix A). After about 2 weeks a reminder was handed out. (13 households responded
to the survey)
To complement the survey, two semi-structure interviews were conducted. One interview was
with a worker at Stockholmshem (with responsibility for the area), and one interview with a
local resident from the area with over 40 years of living there and with good knowledge of
both how the area works and how Stockholmshem is operating in the area. This contact was
provided through Stockholmshem.
The qualitative methods used in the pilot is of low sample size high risk for biased answers.
Note that the findings from the pilot study is not meant as scientific results, just to provide
guidelines for the development of the implementation.
Data on food- and household waste collected from each station in the research area for year
2015 was analysed. The weight of food and household waste is measured and recorded (in
kilos per station) when collected by the collection vehicle. This data is sent to Stockholmshem
on their waste collection bills. On average the household waste is collected once every week,
and the food waste once every other week. The data was summarized and plotted in Excel to
get an overview of how much food waste is being recycled, what the trends look like and to
get an estimate of how many households that are currently recycling food waste. The data was
then used to divide the area in to a treatment and a control group. Data from 2015 is also
included in the analysis when evaluating the effect of the implementation.
15
3.2.2 Experimental design

To evaluate the effect of the information hand-out a natural fieldexperiment (NFE) was
designed, a NFE is an experiment in a field environment where the subjects are unaware of
being part of the experiment. To address the ethical concerns of NFE, and potentially creating
real world effects with the implementation (Cohen 2013), every contact with residents was
approved by Stockholmshem, furthermore the purpose of the information hand-out was
clearly stated on the front page (see Appendix B) in line with (Hansen and Jespersen 2013)
Figure 4. Satellite picture of the research area within Hökarängen, the blue area represents the treatment group, all the
houses within that area got the information hand-out delivered to them. The red area represents the control group which got
no information. The red and blue stars show where the sorting stations are located.
In total 474 households were targeted in the study, with 210 households in the control group
and 264 households in the treatment group (see Figure 4). The two groups were divided
geographically by their blocks, to avoid potential spill-over effects (see below).
16
3.2.3 Evaluating the implementation

To analyse the results a Difference in difference (DiD) analysis was used. DiD is one of the
most popular tools in applied economics research for evaluating the effects of public
interventions and other treatments (Abadie 2005). DiD has, for example, been used to
evaluate the effectiveness of policy implementation (see e.g., Finkelstein 2002), the effect of
Nudges in field experiments (Kallbekken and Sælen 2013), and the impact of natural disasters
(Tian and Guan 2015). DiD is especially useful when the aim is to estimate causal effect of an
implementation and there is no way to randomly select the population in the control and
treatment group. DiD is designed to control both for pre-treatment differences between the
treatment and the control group and for trends over time that are unrelated to the intervention
(Gertler et al. 2010).The most basic DiD design (used in this study) analyses data for two
groups (treatment and control group) over two time periods (pre and post implementation).
The treatment group is exposed to the treatment, and the control group is not. The average
gain in the control group is then subtracted from the average gain in the treatment group to
calculate the Average treatment effect. See Figure 5.
Figure 5. Visual explanation of a DiD analysis, based on Gerthler et al. 2010.
Since the control and treatment groups in this study were defined by geographical conditions
and not completely random, a DiD is arguably a suitable method to analyse the effectiveness
of the information hand-out. But only if some key assumptions holds true mainly the parallel
trend assumption; that is in the absence of the treatment, the average outcomes for the treated
and control groups would have followed parallel paths over time(Gertler et al. 2010). And that
no spill over effect occurred, that is that the implementation in the control group did not affect
the households in the control group. These assumptions will be discussed further under
Critical interpretations of the results.
17
4 RESULTS
4.1 FINDINGS FROM THE PILOT STUDY

The interviewees stated that they perceived overall attitudes in the area to be positive towards
recycling food waste, they believe there is a drive within the community to “do the right
thing”. The Stockholmshem worker estimated that around 50% were already doing it,
although he admitted that it might be an optimistic guess. The local estimated hesitantly that
maybe around 80% of the residents were sorting their food waste already. These pro-food
recycling waste norms were also supported in the Survey response (see appendix X)
The measured behaviour tells a less optimistic story; based on the data collected on food and
household waste I estimated that around 10-20% of the area household were recycling food
waste in the end of 2015 (See Appendix F for calculation). Interestingly to note is the big gap
between perceived behaviour (although from a small sample) and the measured behaviour.
The main reason for the pilot study was to identify barriers for the residents to recycle food
waste. Presented below is a summary of the main barriers uncovered, where they were found,
and why they were considered a barrier. See table 2.
18
Table 2. The barriers uncovered in the Pilot study, their sources and the reasons for being perceived as a barrier. The barriers are listed in order with the perceived biggest barriers listed first.
Barrier Source of Reasons for being perceived as a barrier

barrier
Lack of information, the Interviews, Both interviewees stated a major concern that they thought the information about the new sorting stations did not reach out to everyone.
information did not reach out. Survey This worry was supported by an answer in the survey. One respondent stated that the reason he didn’t sort food waste was because it was
too far to the sorting station, and he had to take the car (the longest possible way to a sorting station is about 200 meters).
Can’t tell the difference Interviews The station for food waste is very similar to the station for household waste. It might not be obvious to everyone what the difference
between the sorting stations. between them is, especially if the information didn’t reach out. This confusion also manifested itself in the fact that the wrong type of trash,
mostly plastic bags and household waste, was thrown in the sorting station for food waste and contaminating the waste.
The belief that the garbage Interviews There is a belief in the area that the garbage truck doesn’t sort the waste. Instead they dump it all in the same truck, and thus all the sorting
truck doesn’t sort the waste is obsolete. This is raised as a concern in both interviews The local brought this up as well. He believed himself that the garbage men did
anyway. sometimes out of laziness. He also stated that several others in the community have seen it happened. The Stockholmshem worker said that
this was indeed happening (the truck sorting everything in the same truck), but it’s not due to laziness, it’s when plastic bags and other
garbage gets thrown into the food waste stations, contaminating the food waste and thus needs to be sorted as regular house waste
Sorting food waste is an Survey Inconvenience reasons mentioned in the survey:
inconvenience. “Not enough space for another garbage bag”,
“Living alone means very small amount of food waste, need to keep track of having compostable bags”.
“I walk with a crutch, so I can only carry one bag at the time”
Laziness. Survey Laziness was stated as one of the reasons for not sorting food waste.
The need for the brown Interviews, To start recycling food waste residents need to acquire compostable bags. These are available for free in the close-by waste room, but its not
compostable bags Survey certain that everyone knows that, and they also need to remember getting one to be able to start recycling food waste.
Cultural differences. Interviews Cultural differences were brought up as a potential barrier in the interviews, but the views differed about how big this impact might be.
Mostly it came down to not understanding the language, and the information, but normative differences in what’s perceived as normal when
handling trash were also mentioned. Although the Stockholmshem worker said that he did not see the need for translating information hand-
outs.
People don’t understand the Interviews There might be a language barrier to understand the previous information sent out.
information
The sorting station stinks Interviews In both interviews this was mentioned as a major issue. The food waste sorting station starts to stink during summer due to food waste going
during summer. bad. The local profile even told me he heard of people moving just because of the stench. At Stockholmshem they were well aware of the
problem, and proclaimed that they now have solved it with more regular cleaning.
The stations are hard to open. Interviews The local had heard some people in the area complaining about the stations, that they are hard to open. A key is needed to open the station
(municipal regulation), this provided an obstacle for some of the elderly people since the lock is located pretty far down on some of the
stations.
19
4.2 INFORMATION HAND-OUT DESIGN

The resulting implementation designed was a 3-page long information hand-out (See
appendix B-D), this section will break down the design process behind the hand-out, which
barriers that were addressed, and what behavioural insight tools were used (See the boxes and
arrows in Figure 3)
The choice of designing an information strategy was made in collaboration with

Stockholmhem. The decision was mainly out of ethical consideration to the residents and their
privacy, an information hand-out was considered the least intrusive strategy. The limiting
factors of this choice is addressed under discussion.
4.2.1 First page

Recall that the barriers are listed in order, with the perceived biggest barriers listed first. The
front page of the information hand-out was designed to address the first two barriers: “lack of
information, information about the new station might not have reached out to” and “can’t tell
the difference between the two sorting stations”. The front page features a picture of a food
waste sorting station from the area. Note also that the information hand-out itself is addressing
the lack-of-information barrier.
The subtitle of the information hand-out is a descriptive social norm, to encourage recycling,
using the phrasing (translated)“Join your neighbours on Hovmästargatan recycle your food
waste” (“Hovämstargatan” is the name of the street the recipients lived on)
Such messages have shown to have great potential of promoting pro-environmental behaviour
in numerous studies Nolan et al. (2008) showed that using the descriptive norm of “Join Your
Neighbors in Conserving Energy” was more effective than using messages of Self-Interest
“Save Money by Conserving Energy” or environmental protection “Protect the Environment
by Conserving Energy”. Furthermore a large field study aiming at nudging people towards tax
compliance, showed that localised normative message of “the great majority of people in your
local area pay their tax on time” was more effective than a more general descriptive normative
message “The great majority of people in the UK pay their tax on time” (Hallsworth et al.
2014)
The source of the information has been showed to greatly influence recipient’s perception of
the information, and especially the trustworthiness of the message (Stern 1992). Craig and
McCann (1978) showed this in a classical study where they sent out a brochure to apartments
in New York trying to influence residents to cut energy use in air-conditioned apartments. The
20
brochure was sent out with the consigner of either “The New York State Public Service” or a
local Electrical company, “Con Edison”. The following month the recipients of the brochure
sent from “New York State Public Service” saved 7% on their electrical bills, while no results
at all were shown from the household receiving of the same brochure sent from “Con
Edison”. One major concern when developing the information handout was that it wouldn’t
be perceived as trustworthy or that it would be thought of as unwanted commercials and be
ignored. To prevent this the hand-out was sent out in envelopes with the Stockholmshem logo
on it, based on the assumption that the residents would perceive the information to be a)
relevant for them to read and b) more trustworthy, the logo was also on the front page next to
Stockholm Resilience Centre and the Stockholm University logo. Some sources might be
perceived as more credible than others by recipients, thus adding more increases the chance of
some of the sources to be perceived credible (Mckenzie-Mohr 2006).
4.2.2 Second page

The second page of the guide starts with an image that clearly demonstrates the actual
behaviour that is being promoted, which might help with the cultural barrier (‘people don’t
understand the information’). The information in the hand-out was phrased to be vivid,
tangible, and relatable using phrases such as (translated);
“If all households in Hökarängen would sort their food waste it would be enough biofuel to
support 15 garbish trucks for a year”
“A buss can drive 2.5 kilometres on only one bag of food waste”,
“Every Swede produces on average 100 kilos of food waste per year”.
Presenting information in a vivid and tangible way increases the likelihood that a message
will be attended to initially and more likely to be remembered. (McKenzie-Mohr 2013). Using
phrasing like “if you were to add up all the cracks around and under your doors, you’d have
the equivalent of a hole the size of a football in your living room wall” to promote home
insulation (Gonzales et al. 1988) has been shown to be an effective tool.
The attitudes uncovered in the survey responses and in the interviews were very pro-
recycling, and highlighted in the information hand-out as follows (translated):
“In a survey recently sent out to households in Hökarängen around 8 out of 10 residents
stated that they considered recycling food waste to be “very important”
21
Combining the community injunctive norms such as “people in Hökarängen believe recycling
food waste is the right thing to do” with the descriptive norms “join your neighbours […] sort
your food waste” could be a persuasive way to frame information (Schultz et al. 2007).
4.2.3 Third Page

The third page includes a picture of the desirable behaviour, but more importantly it shows
the brown garbage bag more clearly. The purpose here is to ensure that people use the correct
bags to prevent contamination of the food waste. But mainly this page is designed to address
and overcome the barrier “The belief that the garbage truck doesn’t sort the waste anyway” by
explaining that (translated);
“If household waste or plastic bags are thrown into the food waste station the food waste is
considered contaminated. When this occurs, the garbage truck picks up food waste in the
same truck that pick up the household waste, and all the effort of recycling food waste is lost.
It is therefore very important to use the right bag when you sort your food waste”
This gives the residents a reason to why the same garbage truck sometimes collect both
households and food waste, (it’s not out of laziness) and highlights the importance of using
the right bag.
To address the barriers “Sorting food waste is an inconvenient, “Laziness” and “The need for
the brown compostable bags” the recipients were also provided with two recyclable garbage
bags in the envelope.
The bags were phrased like a gift in the hand-out. Helping people or giving something makes
people wanting to give something back (Gouldner, 1960). Friedman and Herskovitz (1990)
showed that giving costumers a free gift when entering a store resulted in customers spending
significantly more money in the store. The natural way to give something back for the
households in Hökarängen after receiving the free bags could be to try and start recycling
food waste. More importantly people might already be willing or even have the intention to
start recycling food waste, but still fail to do so. If we want to establish a new behaviour we
have to practice it, and people might be perfectly willing to change behaviour but dont
because they are not persistent enough in practicing the new behaviour until it has become a
habit (Kollmuss and Agyeman 2002). Providing garbage bags to recipients might not ensure
that a new habit takes place, but if it increasing the chance of getting people to start recycling
food, the chances are definitely higher that they will continue to do so.
22
4.3 MEASURING THE EFFECT OF THE INFORMATION HAND-OUT

4.3.1 The data set
The data set consists of data on food and household waste collected from the nine sorting
stations in the research area from January 2015 up until June 15, 2016. The data for year 2015
was received from Stockholms Vatten (sent on bills to Stockholmshem) and the data from
2016 was received from Ragn-Sells database (a waste managing company employed by
Stockholm Vatten). A total of 863 times waste was reported collected, (weight on household
waste was reported 560 times and food waste 302 times). In total 125 191 kilos of household
waste was reported collected and 12 689 kilos of food waste.
In 2015 there were 22 occasions where waste collection was reported but no data on waste
weight was recorded, these were registered as missing data in the data set.
In the year-end of 2015 Stockholm Vatten employed a different company in charge of waste
management in the area, the transition did not go smoothly 4, no data was able to be recorded
because of truck failures and resulted in a data loss for the first two and a half months of
2016.The implication of this data loss is discussed further in the discussion.
4.3.2 The information hand-out effects on food waste collected

Figure 6 shows an overview of the total amount of food waste collected before and after the
information hand-out was sent out.
Figure 6. Aggregated data on food waste collected from all stations within the control group (N = 4) and treatment group (N
= 5), over 35 two-week time periods, spanning from 2015-01-01 to 2016-05-30.
4)This incident was reported on in media, see e.g., : www.svd.se/sopkaos-i-stockholm-sedan-arsskiftet-en-

sanitar-olagenhet/om/sverige (Swedish article) “garbage chaos in the city of Stockholm”
23
Every data point is aggregated data on food waste registered for all stations in the treatment
and control group respectively. Each point represents waste collected in a certain time period,
where each period consists of half a month starting from 2015-02-01 to 2015-02-15 (time
period 3) up and till 2016-05-01 to 2016-05-15 (time period 35). Note that food waste data
from December 2015 (time period 1-2) was removed from the analysis since it was the first
collection point for the newly installed food waste stations in the control group area and there
is no data on exactly how long they waited before this first pick up. Those weights could
therefore be misleading and not comparable to the other time periods.
Note also that the periods 25-28 are missing due to complication from switching of companies
mentioned above. Some of the variation that occurs in the figures is due to different number
of collections in the time periods, this variation is accounted for later in the regression
analysis. Time periods containing missing data points are removed from the plotted data (but
included in the regression analysis).
The information hand-out was sent out to all apartments in the treatment group on April 14,
2016, just after the last collections in time period 31. Visually it looks like a treatment affect
might have occurred, there seems to be a bigger difference between the two groups after the
treatment compared to before. To explore the results further, Table 3 list the average amount
of food waste each station collected in the time period before the treatment, and after the
treatment, and the difference between the two groups.
Table 3. Average amount of food waste collected per sorting station in the control group (5 stations) and in the treatment
group (4 stations) before and after treatment and the difference between the groups before and after treatment. Standard
deviations in brackets. The Difference-in-difference (DiD), or the average treatment effect (ATE) is presented in italics
Group Pre-treatment(T1) food Post-treatment(T2) Difference

waste food waste (T2-T1)
(kg, average per station) (kg, average per
station)
Control group 37.5 28.16 -9.34
(CG) (29.4) (15.9)
Treatment group 58.16 57,9 -0.26
(TG) (55.35) (34.55)
Difference 20,66 29,74 DiD(ATE) = 9.08
(TG-CG)
The difference in average amount of food waste collected between the two groups is higher
after the treatment; before the treatment the average amount of collected food waste in the
treatment group was 20.66 kilos more than the control group (58.16 kilos compared to 37.5
kilos), and after the treatment the difference between the control and the treatment group is
24
29.74 kilos, which means an estimated average treatment effect (ATE) of 9,08 kilos of food
waste collected per station, every two weeks .
If we present the same data in a box-plot looking at the distribution and the median of food
waste collected before and after the treatment we can see an increase in the treatment group
and a relative unchanged control group (comparing medians). See figure 8.
Figur 8. A box plot of the amount of food waste collected in the control group and the treatment group pre and post
treatment.
Note that this is an opposite trend compared to when we evaluate the means, nevertheless, the
positive treatment effect still holds true.
The mean values showed a decrease in average weight collected in the control group while the
average weights collected in the treatment group remained the same. This difference is due to
outlier values in the pre-treatment sample which drive up the mean values. But it is also a
reflection of the relatively small sample size.
4.3.3 The information handout effects on household waste collected

The total amount of Household waste collected over time is displayed bellow, see Figure 9.
The plotted data visualizes a potential treatment effect also on the household waste. The
difference between the groups appears to be bigger after the treatment compared to before.
25
Note that the there is a weight spike that occurs in time period 31 just before the
implementation, this spike is partly because of a rare occasion of 3 household waste collection
Figure 9. Aggregated data on household waste collected from each station for the control group (N= 4) and the treatment
group (N = 5), over 35 two-weeks periods, starting 2015-01-01 to 2016-05-30
during the same time period in the control group, on the 1th the 8th and the 15th (usually there
is only two for household waste). Such variation is accounted for in the regression analyses. A
cause for concern is that the trend seems to start before the implementation, this is addressed
in the discussion.
The difference in average amount of household waste collected before the treatment and after
the treatment, and the difference between the two groups is presented below, see Table 4.
Table 4. Average amount of household waste collected per sorting station in the control group (5 stations) and in the
treatment group (4 stations) before and after treatment and the difference between the groups before and after treatment.
Standard deviations in brackets. The Difference-in-difference (DiD), or the average treatment effect (ATE) is presented in
italics
Group Pre-treatment (T1) Post-treatment (T2) Difference

Household waste (kg, Household waste (kg, (T2-T1)
average per station) average per station)
Control group 418.5 796.1 +377.6
(CT) (198.3) (77.7)
Treatment group 471.4 633.8 +162.4
(TC) (166.2) (187.3)
Difference 52.9 -162.3 DiD(ATE) = - 215.2
(CT-TC)
26
The average amount of household waste collected was 52.9 kilos more in the treatment group
compared to the control group before the information hand-out. After the information hand-
out were sent out an average of 162.3 kilos less household waste was collected in the
treatment group compared to the control group, which leaves a negative difference in the
difference estimate (ATE) of - 215.2 kilos. The distribution and the median of household
waste collected before and after the treatment is displayed in the box plot below (Figure 10)
Figur 10. A box plot of the amount of household waste collected in the control group and the treatment group pre and post
treatment.
This box plot highlights the fact that there’s an increase of household waste collected in both
groups, but the increase is significantly higher in the control group.
Overviewing data shows a possible treatment with an increase in food waste collected, and a
decrease in household waste. To see if the DiDs between the groups observed are statistically
significant a regression analyses was performed.
4.3.4 Regression Analysis

A multivariate linear panel regression was executed using STATA 13. A regression analysis
does not only allow us to control for other variables that may influence the results (such as the
number of collection in each period), by employing a panel data model we can also control for
potential serial correlation and heterogeneity in the error terms. Results from the regression
analysis can be found in Table 5 below. A fixed effect model was used (as opposed to a
random effects panel data model) to account for the fact that the population was not randomly
selected (Baltagi 2008)To account for potential serial correlation and heterogeneity we use
cluster robust standard errors.
27
Table 5. Impact Estimating average treatment effects for food waste and household waste through a panel data regression
model robust standard error in brackets.
Food waste
Coefficient Household waste
Coefficient
(Robust St. error) (Robust St. error)
p value p value
Constant 15,67149 – 52,83332

(13,10176) (26,82654)
0,266 0,084*
Avrage Treatment effect (ATE) 11,65304 – 212,1389
(6,061314) (58,54588)
0,091* 0,007***
Number of collections 24,16204 245,9077
(9,97981) (13,92568)
0.042** 0.000***
P value for the regression (Prob > F) 0.0428** 0.0000***
*significant on the p < 0.1 level, **= Significant on the P < 0.05 level, ***=Significant on the P <
0.01 level.
The ATE of the implementation had a significant effect on both food and household waste.
The information handout had an estimated effect of 11.65 more kilos of food waste per
stations per time period collected (which equals an average of 46.6 kilos more food waste
collected every other week in the control group area compared to the treatment group area).
The implementation had negative ATE on household waste of 212.13 kilos per station. Note
that these numbers are very similar to the ones estimated above in the overviews, even after
accounting for amount of collections and potential serial correlation within the stations.
The number of collections in each period (how many time the garbage truck collected waste
in the given time period) was accounted for in the multivariable regression, and had, as
suspected, a significant impact on the waste collected in each time period. (Food waste P >
0.05, household waste P > 0.001).
These results support both hypotheses:
I. The information hand-out increases the amount of food waste collected from the
sorting stations in the treatment group compared to the control group.
II. The information hand-out decreases the amount of household waste collected from
the sorting stations in the treatment group compared to the control group.
28
5 DISCUSSION
The main scientific aim of this theses was to develop and test a methodological approach to
promote pro-environmental behaviour, the method was tested by designing an information
hand-out that was distributed to a research area in Hökarängen. To evaluate the method and
the information hand-out a field experiment was executed to measure differences in waste
collected. The results shows statistically significant support for the hypotheses that the
information hand-out created a desired behavioural change.
5.1 CRITICAL INTERPRETATION OF THE RESULTS.

Even though the results were significant there is a need for a critical reflection about their
reliability before discussing the relevance.
The study design and execution was interrupted by the change of waste managing companies
in the beginning of 2016 that prohibited data collection over 5 time periods and resulted in
missing data. Missing data, high variance, and small samples, which is the case in this study
especially for the food waste data, not only lower the reliability of the results, but also make it
hard to justify the assumption of equal trends in DiD. Even though the trends look similar
before the implementation (see e.g. figure 6), the DiD analysis attributes any trends changes
in the treatment and comparison groups that happen from the time period after the treatment
to the treatment. If any other variable is present that effect the difference in trends between the
two groups, the estimation will be invalid or biased (Gertler et al. 2010). There is no way to
guarantee that these trends remained parallel during the time of the missing data, or would
have been without treatment. But in an effort to control for this a robustness test (Wooldrig
2007)was performed, the same regression analysis was used but with a “fake” treatment time
period which also covered the weight spike observed (for regression details and figure see
appendix G). The regression showed no significant treatment effect in the fake treatment
period. Although it shows a relatively strong trend (P= 0.166). The variable for “Number of
collection” was still highly significant ( P < 0.001), which indicated that the weight
differences in the weight spike is mainly due to different numbers of collections. The
robustness test reinforces the parallel trend assumptions, which increases the likelihood that
the changes observed was caused by the information hand-out.
Other important circumstances were the fact that the data collection was done by two different
companies, this might affect the results, though this is assumed to affect both groups equally
and thus accounted for in the regression analyses. Natural field experiments have inherent
29
flaws, they are not as reliable as fully controlled experiments field and there is always a risk
for unknown confounding factors affecting the outcome (Harrison and List 2004).
These concerns should be taken into account when evaluating the results from this study.
Nevertheless, significant results in line with both hypotheses are promising results for the
methodical approach and a good indication that the information hand-out in fact created a
desired pro-environmental behavioural change. To further investigate the effect, following up
with a regression analysis on more post treatment data is suggested, it would also give
insights to the long term effect of the hand-out, this is easily feasible but unfortunately not in
the scope of this thesis.
5.2 REFLECTING ON THE METHODOLOGY APPROACH AND THE INFORMATION HAND OUT.
This study proposes a strategic way to interlink science and practise and a potential tool for
policy makers and other actors aiming to promote pro-environmental behaviour changes. The
methodological approach is also a systematic way of learning and evaluating strategies,
highlighting the complexity of human behaviour and thus the importance of testing and
adapting implementations to the context. Note that some of the barriers found in the pilot
study are seemingly highly contextual, and demographical traits in Hökarängen like a political
leaning towards the socialistic block and pro-environmental norms, might have been
contributing factors to implementations like the information hand-out designed being
successful. However, the approach is designed to take advantage of such pro-environmental
norms, other strategies might have been developed in areas with less “environmental friendly
traits”. This only enforces the importance of a Pilot study, and the importance of adapting
implementations to the specific context and your target populations.
The decision to only communicate with the residents though written information was a
limiting factor: instead of a survey other potentially more insightful and often suggested
options, like focus group workshop with residents, door-to-door knocking, or telephone
surveys (McKenzie-Mohr 2013), could have been used. It also restricted the options in the
design of the implementation, for example, using community role models or so-called block
leaders, emphasizing personal contact, making people commit to a cause, are all tactics from
the fields of social marketing and nudging that previously have shown promising
30
results5.Including such methods or other behavioural insight tools might have created a bigger
behaviour change.
Informational strategies are often criticized as ineffective for fostering behaviour changes
(McKenzie-Mohr and Schultz 2014), nonetheless they can be effective tools when pro-
environmental behaviour is relatively convenient (Steg and Vlek 2009) these strategies also
tend to be cost effective, cheap to produce and distribute, and easy to scale up (ibid).
The results from this study (although they should be taken with caution) do indicate that if
carefully designed – an information hand-out can have a significant effect, while being easy to
adapt, scale up and cost effective. There is often an economic incentive for housing
companies to promote the recycling of food waste amongst their residents. Stockholmshem
for example, pays 610,00 SEK (64 Euros) per ton for the collection of household waste and
350,00 SEK (37 Euros) per ton for food waste. Considering the low costs involved in the
execution of this study, and the positive results, it seems the information hand-out is both an
environmentally and economically sound strategy for housing companies, presenting a
possible win-win situation.
The encouraging results motivate further research, testing the methodological approach on
other behaviours and in other contexts. Since the method only measured the behavioural
impacts of the information hand-out, future research could follow up with qualitative research
exploring how implementations like this affects attitudes, norms, and values to get deeper
insights behind the behavioural changes.
5.3 POTENTIAL OF MERGING SOCIAL MARKETING AND NUDGING THEORY

This study aimed to explore the compatibleness of nudging and social marketing theoretical
frameworks, since there has been concerns about how to integrate these frameworks (Badshah
2010). Based on the experiences gained during the design of the information hand-out where
both theories were applied, no obvious obstacles in combining the frameworks was
encountered, rather the contrary. Social marketing has a more holistic view on behaviour
change and often present guidelines in line with the ones used in this study. The field of
nudging, based on behavioural insights from lab and field experiments could be an especially
important contribution to the phase “Designing the implementations”. Adding behavioural
5 ) See
e.g,; www.cbsm.com, for a good overview of behavioural insights from social
marketing, and (Ambler et al. 2011) for an overview of insights from behavioural economy
31
insights from the field of nudging consequently expanding the toolbox for changing behaviour
is therefore something that should be further investigated, and the methodological approach
developed in this study could serve as a model for doing this. On the other hand, behavioural
economics and nudging could potentially learn a lot from social marketing approaches.
Especially when it comes to the strategic planning process e.g. how to find and address
barriers, or how to select which behaviour to change.
5.4 A WIDER TRANSFORMATION CONTEXT

There is a multitude of promising conceptual frameworks studying transformations toward
sustainability, two prominent ones are transition management (Loorbach 2010) and
sustainability transformation in social-ecological systems (Westley et al. 2011) they
emphasise the need for fundamental and large scale transformation of society to ensure
staying in a safe operating space within our planet’s boundaries (Rockström et al. 2009).The
tools and theories discussed in this thesis are meant to complement the community and
societal level framework mentioned above by addressing the potential (and the complexity) of
individual behaviour change. Simply focusing on promoting pro-environmental behaviour
amongst individuals is not going to be enough to create the large scale transformation needed,
for example, only around one third of the energy consumed in the United States is used by
individuals and households while the remainder is used by business, industry, the military,
and other players (Gardner and Stern 2002), thus behaviour change needs to be integrated and
operationalized across sectors, organizations, policy, and among individuals.
”Nudges can be seen as a helpful part of the solution but not a magic bullet.”
(French, 2011, p.157)
Nevertheless looking at quick achievable behaviour changes with the potential of big
aggregated impacts without reduction in human wellbeing is arguably a good way to
accommodate other measures. Two important aspects of behaviour change are highlighted in
this thesis (1) the potential positive impact of the behavioural change (2) the plasticity of the
behaviour change (how hard it is to change) and by combining these aspect there is a
possibility to find behavioural leverage points where small and easily made behavioural
changes have big positive impacts for society. The methodological approach presented in this
thesis could potentially help guide the development of tools to target theses leverage points,
and accelerate the urgently needed transformation towards a sustainable future.
32
LITERATURE CITED
Abadie, A. 2005. Semiparametric Estimators. Review of Economic Studies 72: 1–19.

doi:10.1111/0034-6527.00321.
Allcott, H. 2011. Social norms and energy conservation. Journal of Public Economics 95.
Elsevier B.V.: 1082–1095. doi:10.1016/j.jpubeco.2011.03.003.
Ambler, T., S. Braeutigam, J. Stins, S. Rose, S. Swithenby, B. Bates, A. Lennox, C. Bates, et

al. 2011. MINDSPACE: Influencing behaviour through public policy. Marketing Letters
33: 1–50. doi:10.1111/j.1753-4887.2009.00206.x.
Athey, V. L., R. J. Suckling, A. M. Tod, S. J. Walters, and T. K. Rogers. 2012. Early diagnosis
of lung cancer: evaluation of a community-based social marketing intervention. Thorax
67: 412–7. doi:10.1136/thoraxjnl-2011-200714.
Badshah, A. 2010. Behavioral Economics and Social Marketing: New Allies in the War on
Absent Behavior. Information Technologies & International Development 6: 72–77.
doi:10.2105/AJPH.81.10.1339.
Baltagi, B. 2008. Econometric analysis of panel data. John Wiley & Sons.
Bonell, C., M. McKee, A. Fletcher, P. Wilkinson, and A. Haines. 2011. One Nudge Forward,
Two Steps Back: Why nudging might make for muddled public health and wasted
resources. British Medical Journal 342: 241–242. doi:10.1136/bmj.d408.
Brewer, G. D., and P. C. Stern. 2005. Decision making for the environment: Social and
behavioral science research priorities. National Academies Press.
Cialdini, R. B., L. J. Demaine, B. J. Sagarin, D. W. Barrett, K. Rhoads, and P. L. Winter.

2006. Managing social norms for persuasive impact. Social Influence 1: 3–15.
doi:10.1080/15534510500181459.
Cohen, S. 2013. Nudging and informed consent. The American Journal of Bioethics 13.
Taylor & Francis: 3–11.
Corner, A., and A. Randall. 2011. Selling climate change? The limitations of social marketing
as a strategy for climate change public engagement. Global Environmental Change 21.
Elsevier Ltd: 1005–1014. doi:10.1016/j.gloenvcha.2011.05.002.
Corral-Verdugo, V. 1997. Environmental Dual “ Realities ” of Conservation Behavior : Self-

Reports Vs. Journal of Environmental Psychology 17: 135–145.
doi:10.1006/jevp.1997.0048.
Costa, D. L., and M. E. Kahn. 2013. Energy conservation “nudges” and environmentalist
ideology: Evidence from a randomized residential electricity field experiment. Journal of
the European Economic Association 11: 680–702. doi:10.1111/jeea.12011.
Costanzo, M., D. Archer, E. Aronson, and T. Pettigrew. 1986. Energy conservation behavior:
The difficult path from information to action. American Psychologist 41: 521–528.
doi:10.1037/0003-066X.41.5.521.
33
Craig, C. S., and J. M. McCann. 1978. Assessing communication effects on energy

conservation. Journal of consumer research 5. The Oxford University Press: 82–88.
Creswell, J. W. 2008. Editorial: Mapping the Field of Mixed Methods Research. Journal of
Mixed Methods Research 3: 95–108. doi:10.1177/1558689808330883.
Dietz, T., G. T. Gardner, J. Gilligan, P. C. Stern, and M. P. Vandenbergh. 2009. Household

actions can provide a behavioral wedge to rapidly reduce US carbon emissions.
Proceedings of the National Academy of Sciences of the United States of America 106:
18452–6. doi:10.1073/pnas.0908738106.
Festinger, L. 1962. A theory of cognitive dissonance. Vol. 2. Stanford university press.
Finkelstein, A. 2002. The effect of tax subsidies to employer-provided supplementary health

insurance: Evidence from Canada. Journal of Public Economics 84: 305–339.
doi:10.1016/S0047-2727(00)00155-9.
French, J. 2006. Why nudging is not enough. Journal of Social Marketing 1, no. 2: 154–162.
doi:http://dx.doi.org/10.1108/09564230910978511.
Gertler, P. J., S. Martinez, P. Premand, L. B. Rawlings, and C. M. J. Vermeersch. 2010.
Impact Evaluation in Practice. The World Bank. doi:10.1596/978-0-8213-8541-8.
Gonzales, M. H., E. Aronson, and M. a. Costanzo. 1988. Using Social Cognition and
Persuasion to Promote Energy Conservation: A Quasi-Experiment. Journal of Applied
Social Psychology 18: 1049–1066. doi:10.1111/j.1559-1816.1988.tb01192.x.
Goodwin, T. 2012. Why We Should Reject “Nudge.” Politics 32: 85–92. doi:10.1111/j.1467-
9256.2012.01430.x.
Hajer, M., M. Nilsson, K. Raworth, P. Bakker, F. Berkhout, Y. de Boer, J. Rockström, K.

Ludwig, et al. 2015. Beyond cockpit-ism: Four insights to enhance the transformative
potential of the sustainable development goals. Sustainability (Switzerland) 7: 1651–
1660. doi:10.3390/su7021651.
Haldeman, T., and J. W. Turner. 2009. Implementing a Community-Based Social Marketing

Program to Increase Recycling. Social Marketing Quarterly 15: 114–127.
doi:10.1080/15245000903154618.
Hallsworth, M., J. A. List, R. D. Metcalfe, and I. Vlaev. 2014. The Behavioralist As Tax
Collector : Using Natural Field Experiments to Enhance Tax Compliance. NBER
Working Paper Series: 1–44. doi:10.3386/w20007.
Hansen, P. G., and A. M. Jespersen. 2013. Nudge and the Manipulation of Choice.
European Journal of Risk Regulation 1: 3–28. doi:10.2307/2489305.
Harrison, G. W., and J. A. List. 2004. Field experiments. Journal of Economic literature 42.
American Economic Association: 1009–1055.
IPCC. 2014. Climate change 2014: synthesis Report. Contribution of working groups I, II and
III to the fifth assessment report of the intergovernmental panel on climate change.
IPCC.
Kahneman, D. 2003. Maps of bounded rationality: Psychology for behavioral economics. The
34
American economic review 93. JSTOR: 1449–1475.
Kallbekken, S., and H. Sælen. 2013. “Nudging” hotel guests to reduce food waste as a win-
win environmental measure. Economics Letters 119. Elsevier B.V.: 325–327.
doi:10.1016/j.econlet.2013.03.019.
Karp, D. G. 1996. Values and their Effect on Pro-Environmental Behavior. Environment and
Behavior 28: 111–133. doi:10.1177/0013916596281006.
Kazdin, A. E. 2009. Psychological science’s contributions to a sustainable environment:

extending our reach to a grand challenge of society. The American psychologist 64:
339–356. doi:10.1037/a0015685.
Kollmuss, A., and J. Agyeman. 2002. Mind the Gap: Why do people act environmentally and
what are the barriers to pro-environmental behavior? Environmental Education
Research 8: 239–260. doi:10.1080/13504620220145401.
Kotler, P., and G. Zaltman. 1971. Social marketing: an approach to planned social change.
The Journal of Marketing. JSTOR: 3–12.
Lefebvre, C. R. 2016. A Consensus Definition of Social Marketing: retrieved
from:socialmarketing.blogs.com/r_craiig_lefebvres_social/2013/10/a-consensus-
definition-of-social-marketing.html accessed on: 2016-05-15.
Lindahl, T., and B. Stikvoort. 2015. Nudging – det nya svarta inom miljöpolicy?
Loorbach, D. 2010. Transition management for sustainable development: a prescriptive,

complexity‐based governance framework. Governance 23. Wiley Online Library: 161–
183.
Ly, K., M. Zhao, and D. Soman. 2013. A Practitioner’s Guide to Nudging. Rothman School of
Management: University of Toronto: 1–28. doi:Research Report Series Behavioural
Economics in Action.
Maibach, E. 1993. Social marketing for the environment: using information campaigns to
promote environmental awareness and behavior change. Health Promotion International
8: 209–224. doi:10.1093/heapro/8.3.209.
Marteau, T. 2011. Judging nudging: can nudging improve population health? Bmj 342: 263.
Mckenzie-mohr, D. 2000. Promoting Sustainable Behavior: An Introduction to Community-

Based Social Marketing. Journal of Social Issues 56: 543–554. doi:10.1111/0022-
4537.00183.
McKenzie-Mohr, D. 2000. Quick reference: community-based social marketing. Retrieved

March 26: 2002. doi:10.4269/ajtmh.2011.11-0095.
Mckenzie-Mohr, D. 2006. Fostering Sustainable Behavior. New society publishers.
McKenzie-Mohr, D. 2013. Fostering sustainable behavior: An introduction to community-

based social marketing. New society publishers.
McKenzie-Mohr, D., and P. W. Schultz. 2014. Choosing Effective Behavior Change Tools.
Social Marketing Quarterly 20: 35–46. doi:10.1177/1524500413519257.
35
Midden, C., F. Kaiser, and T. Mccalley. 2007. Technology’s four roles in understanding
individuals' conservation of natural resources. Journal of Social Issues 63: 155–174.
doi:10.1111/j.1540-4560.2007.00501.x.
Mont, O., M. Lehner, and E. Heiskanen. 2014. Nudging: Ett verktyg för hållbara beteenden?
Mullainathan, S., and R. Thaler. 2000. Behavioral Economics. Cambridge, MA.

doi:10.3386/w7948.
Nolan, J. M., P. W. Schultz, R. B. Cialdini, N. J. Goldstein, and V. Griskevicius. 2008.

Normative social influence is underdetected. Personality & social psychology bulletin 34:
913–23. doi:10.1177/0146167208316691.
Oskamp, S. 2000. Psychology of Promoting Environmentalism: Psychological Contributions

toAchieving an Ecologically Sustainable Future for Humanity. Journal of Social Issues
56: 373–390. doi:10.1111/0022-4537.00173.
Rockström, J., W. L. Steffen, K. Noone, Å. Persson, F. S. Chapin III, E. Lambin, T. M.

Lenton, M. Scheffer, et al. 2009. Planetary boundaries: exploring the safe operating
space for humanity.
Sachdeva, S., R. Iliev, and D. L. Medin. 2009. Sinning saints and saintly sinners the paradox
of moral self-regulation. Psychological science 20. SAGE Publications: 523–528.
Schultz, P. W., and F. G. Kaiser. 2012. Promoting Pro-Environmental Behavior.

doi:10.1093/oxfordhb/9780199733026.013.0029.
Schultz, P. W., J. M. Nolan, R. B. Cialdini, N. J. Goldstein, and V. Griskevicius. 2007.

Research Article: The Constructive, Destructive, and Reconstructive Power of Social
Norms. Psychological Science 18: 429. doi:10.1111/j.1467-9280.2007.01917.x.
Steg, L., and C. Vlek. 2009. Encouraging pro-environmental behaviour: An integrative review
and research agenda. Journal of Environmental Psychology 29. Elsevier Ltd: 309–317.
doi:10.1016/j.jenvp.2008.10.004.
Stern, P. C. 1992. What psychology knows about energy conservation. American

Psychologist 47: 1224–1232. doi:10.1037/0003-066X.47.10.1224.
Stern, P. C. 2000. Toward a Coherent Theory of Environmentally Significant Behavior.

Journal of Social Issues 56: 407–424. doi:10.1111/0022-4537.00175.
Sunstein, C. R. 2014. Nudging: A Very Short Guide. Journal of Consumer Policy: 583–588.
doi:10.1007/s10603-014-9273-1.
Thaler, R. H., and C. R. Sunstein. 2003. Libertarian paternalism. The American Economic
Review 93. JSTOR: 175–179.
Thaler, R. H., and C. R. Sunstein. 2008. Nudge: Improving Decisions about Health, Wealth,
and Happiness. Yale University Press.
Thorpe, A., R. Merritt, D. McVey, and A. Truss. 2008. What Next for Social Marketing –
Developing “Superman” or a Sustainable System? Social Marketing Quarterly 14: 63–
71. doi:10.1080/15245000701849187.
36
Tian, X. L., and X. Guan. 2015. The Impact of Hurricane Katrina on Students’ Behavioral
Disorder: A Difference-in-Difference Analysis. International Journal of Environmental
Research and Public Health 12: 5540–5560. doi:10.3390/ijerph120505540.
Trope, Y., and N. Liberman. 2010. Construal-level theory of psychological distance.

Psychological Review 117: 440–463. doi:10.1037/a0018963.
Westley, F., P. Olsson, C. Folke, T. Homer-Dixon, H. Vredenburg, D. Loorbach, J.

Thompson, M. Nilsson, et al. 2011. Tipping Toward Sustainability: Emerging Pathways
of Transformation. AMBIO 40: 762–780. doi:10.1007/s13280-011-0186-9.
Vlek, C., and L. Steg. 2007. Human Behavior and Environmental Sustainability: Problems,
Driving Forces, and Research Topics. Journal of Social Issues 63: 1–19.
doi:10.1111/j.1540-4560.2007.00493.x.
Wooldrig, I. 2007. Difference-in-Differences Estimation. retrieved from:

http://www.nber.org/WNE/lect_10_diffindiffs.pdf. accessed on: 2016-07-25.
37
APPENDIX
APPENDIX A
38
APPENDIX B
39
APPENDIX C
40
APPENDIX D
41
APPENDIX E
Kan du
Sorterar tänka Hur många
Respons Antal Antal du sig Övriga Hur av dina
ID Kön rum personer matavfall? börja ? Största anledningen/hindret anledningar/hinder viktigt? grannar?
1 Kvinna 2 2 Ja x x
Mycket Några få
2 Kvinna 2 1 Ja viktigt (15-35%)
Ungefär
Mycket hälften (35-
3 Man 2 2 Ja viktigt 65%)
Ungefär
4 Kvinna 2 2 Ja viktigt 65%)
Inte Ungefär
särskilt hälften (35-
5 Man 2 2 ja viktigt 65%)
Mycket Ingen
6 Kvinna 3 1 ja viktigt uppfattning
Ganska
Mycket Många (65-
7 Kvinna 2 3 ja viktigt 85%)
Nej. Sorterar
ordentligt hos
mina föräldrars Ungefär
Lathet, brist på utrymme hem i Södertälje Ganska hälften (35-
8 Hen 2 2 Nej Ja för flertal sophinkar/kärl. dock. vikigt 65%)
Ungefär
9 Man 3 3 Ja viktigt 65%)
Ungefär
Små mängder i ensamhushåll, kräver att jag
ser efter statusen på pappåsar. Känsla av frihet Ganska Några få
11 Kvinna 2 1 Nej Nej när man slänger iväg saker regelbundet. vikigt (15-35%)
jag har så långt till jag går med
sorteringen.jag måste ta kryckor kan bara Ganska Några få
12 Man 4 2 Nej Nej bilen till sorteringen. ta en kasse vikigt (15-35%)
Ganska
Mycket Många (65-
42
APPENDIX F
A total of 74 450 kilos of household waste was reported to be collected over the year, and a
total of 8600 kilos of food waste, According to Stockholm Vatten, a government owned
company in charge of Stockholm city waste management, around 40% of unsorted household
waste is recyclable food waste 6. Assuming that 40% of all waste (household and food waste
combined) in Hökarängen is food waste, I estimated that over the year of 2015 around 33 200
kilos of food waste, and thus about 26% of it was recycled. Although there is a declining trend
in sorting food waste in the area (see figure) and in the end of 2015 based on the waste data I
estimate around 10-20% of households was currently recycling food waste in the end of 2015.
Note that this estimate is more of a best-guess than an actual fact.
The total amount of waste collected in the control and treatment group for every time period of year 2015. Each time period
represents half a month (Time period 1 = January 1st -15th 2015, Time period 24 = December 16th-31st 2015), time periods
that contains missing data points were removed from the graph.
http://www.stockholmvatten.se/globalassets/pdf1/informationsmaterial/avfall/privatkund/bros
chyrer/matavfall_villa_webb_2015.pdf
43
APPENDIX G
Tabell Impact Estimating average treatment effects for food waste and household waste through a panel data regression
model robust standard error in brackets, Robustness test using a Fake treatment period.
Household waste
Coefficient
(Robust St. error)
p value
Constant -34.11708
(37.14578)
0.385
Avrage Treatment effect (ATE) -108.3028
FAKE TREATMENT PERIOD (71.14719)
0.166
Number of collections 243.8532
(18.66247)
0.000***
P value for the regression (Prob > F) 0.0000***
Aggregated data on hosuhehold waste collected over time, showing the gap of missing data in time period 25-28, and the fake treatment
period used to strengthen the parallel trend assumption in DiD. Note that the fake period incorparets the weight spike in timeperiod 31
44

Stockholm Resilience Centre: Exploring Tools To Promote Pro-Environmental Behaviour

Uploaded by

Copyright:

Available Formats

You might also like

Stockholm Resilience Centre: Exploring Tools To Promote Pro-Environmental Behaviour

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Stockholm Resilience Centre: Exploring Tools To Promote Pro-Environmental Behaviour

Uploaded by

Copyright:

Available Formats

Noah Linder Master´s Thesis 2015/2016

Stockholm Resilience Centre

Master’s Thesis, 60 ECTS

Exploring tools to promote pro-

Exploring tools to promote pro-environmental behaviour:

Author: Noah Linder

2 Theoretical Framing ................................................................................. 4

Literature cited ............................................................................................ 33

1.1 BEHAVIOUR CHANGE AS LEVERAGE POINTS TO ACCELERATE A SUSTAINABLE

1.2 UNSUCCESSFUL APPROACHES TO BEHAVIOURAL CHANGE

1.3 THE SCIENTIFIC AIM OF THE THESIS

2.1 PRO-ENVIRONMENTAL BEHAVIOUR CHANGE IN PSYCHOLOGY

2.2 SOCIAL MARKETING

originates as a guideline on how to make psychological knowledge visible to better develop

2.3 BEHAVIOURAL ECONOMICS AND NUDGING

2.4 CRITIQUE OF NUDGING AND SOCIAL MARKETING

3.1 STUDY LOCATION AND CASE

BOX 1. Swedish policy goals for collecting food waste

3.1.2 Hökarängen city district

Demographic data Hökarängen Stockholm average

3.2 METHODOLOGICAL APPROACH TO PROMOTE PRO-ENVIRONMENTAL BEHAVIOUR

(1) Selecting behaviour to change

(2) Examination of the main factors underlying the behaviour

(3) Designing the implementation

(4) Executing Implementation on a smaller scale

In order to avoid expensive failures or unexpected results, like promoting unwanted

(5) Evaluating the effectiveness of the implementation

To evaluate the effectiveness of the implementation design, a solid experimental evaluation

(6) Large scale implementation.

3.2.1 Pilot study

A request to participate in a survey was distributed to 92 households (around 20%) in the

3.2.2 Experimental design

3.2.3 Evaluating the implementation

Figure 5. Visual explanation of a DiD analysis, based on Gerthler et al. 2010.

4.1 FINDINGS FROM THE PILOT STUDY

Barrier Source of Reasons for being perceived as a barrier

4.2 INFORMATION HAND-OUT DESIGN

The choice of designing an information strategy was made in collaboration with

4.2.1 First page

4.2.2 Second page

4.2.3 Third Page

4.3 MEASURING THE EFFECT OF THE INFORMATION HAND-OUT

4.3.2 The information hand-out effects on food waste collected

4)This incident was reported on in media, see e.g., : www.svd.se/sopkaos-i-stockholm-sedan-arsskiftet-en-

Group Pre-treatment(T1) food Post-treatment(T2) Difference

4.3.3 The information handout effects on household waste collected

Group Pre-treatment (T1) Post-treatment (T2) Difference

4.3.4 Regression Analysis

Constant 15,67149 – 52,83332

These results support both hypotheses:

5.1 CRITICAL INTERPRETATION OF THE RESULTS.

5.3 POTENTIAL OF MERGING SOCIAL MARKETING AND NUDGING THEORY

5.4 A WIDER TRANSFORMATION CONTEXT

Abadie, A. 2005. Semiparametric Estimators. Review of Economic Studies 72: 1–19.

Ambler, T., S. Braeutigam, J. Stins, S. Rose, S. Swithenby, B. Bates, A. Lennox, C. Bates, et

Cialdini, R. B., L. J. Demaine, B. J. Sagarin, D. W. Barrett, K. Rhoads, and P. L. Winter.

Corral-Verdugo, V. 1997. Environmental Dual “ Realities ” of Conservation Behavior : Self-

Corral-Verdugo, V. 1997. Environmental Dual “ Realities ” of Conservation Behavior : Self-