Do we value goods constructed and customized by us
more? A study of IKEA, effort justification and
customization effect
Author: Mantas Gudelevicius
Student number: 11374241
Qualification: MSc Business Administration – Marketing track
Supervisor: Joris Demmers
Date of submission: 23rd of June, 2017
UNIVERSITY OF AMSTERDAM 2
Statement of originality
This document is written by Student Mantas Gudelevicius who declares to take full
responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources
other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of
UNIVERSITY OF AMSTERDAM 3
Table of Contents
Acknowledgements ... 5 Abstract ... 6 1. Introduction ... 7 2. Literature review: ... 11 2.1. Literature review ... 112.2. Customer participation in co-production within SD-logic ... 11
2.3. The Ikea Effect ... 13
2.4. Effort justification paradigm ... 15
2.5. The customization effect ... 16
2.6. Elaboration of conceptual model ... 18
3. Methodology ... 19
3.1. Research strategy ... 19
3.2. Sample and measurements ... 20
3.3. Variables ... 22
3.3.1. Independent variable: customization... 22
3.3.2. Independent variable: construction ... 23
3.3.3. Dependent variable: Valuation of origami ... 23
3.4. Statistical procedure ... 23 4. Results ... 24 4.1. Pre-analysis ... 24 4.1.1. Age ... 24 4.1.2. Gender ... 25 4.1.3. Conclusion of pre-analysis ... 26 4.2. Analysis ... 26
4.3. Post Hoc tests ... 29
5. Discussion ... 31
5.1. General discussion ... 32
5.1.1. Customization effect ... 33
5.1.2. Effort justification and IKEA effects ... 34
5.1.3. Interaction effect ... 35
5.2. Theoretical contributions ... 36
5.3 Managerial implications ... 37
5.4. Limitations and future research ... 38
References ... 40
UNIVERSITY OF AMSTERDAM 4
Appendix B: Instructions that builders received on how to construct origami. ... 48
Appendix C: The Schedule of the experiment ... 49
Appendix D: Questionnaires provided to participants of the experiment: ... 49
List of Figures and Tables
Figure 1: Conceptual model ... 18Figure 2: Plot of the results ... 31
Figure 3: The mean values of evaluation between different conditions. ... 35
Table 1: 4 different conditions of the experiment ... 20
Table 2: Age effect on different conditions ... 25
Table 3: Chi square test ... 26
Table 4: Results of the experiment ... 28
Table 5: Summary of hypothesis results. ... 28
Table 6: Post Hoc test for builder’s condition ... 30
UNIVERSITY OF AMSTERDAM 5
Acknowledgements
First of all, I would like to express my gratitude to my thesis supervisor Joris Demmers, for
his constant help and advice. It was a tough half-year and I am very happy that I was able to
learn from such a great academic. Furthermore, I would like to thank all the lecturers and staff
of the University of Amsterdam Business School, who have contributed towards my studies. I
am grateful to all of them, for making my Master’s studies interesting and useful in professional
life outside the university. Last but not the least, I would like thank my friend Andy Fekete, for
UNIVERSITY OF AMSTERDAM 6
Abstract
Nowadays, when consumers became extremely spoilt, marketer’s lack of new ideas of how to
stimulate consumption and surprise their customers. If customers would place great value on
the end product it would be easier to manipulate consumption (Caulkins et. al., 2006). Research
of the IKEA and effort justification effects suggest that the more work one invests when
constructing something, the more he/she will value it. The literature on customization effect
also suggests that people would evaluate a product more if their preferences would be taken
into the account during the production process. Literature, however, does not provide a clear
view regarding the interaction of all of these effects on the evaluation of a product.
Consequently, the research question: ‘To what extent do customization, effort justification and
their interaction affect consumer’s evaluations of self-made goods?’ is answered to address this research gap. The present study has demonstrated that effort justification and interaction effects
can positively affect the evaluation of a product. No significant effect was found on the
influence of customization on the evaluation of a product. These results imply that marketers,
still have some room to surprise the consumer of today’s world. To do so, they must innovate ways to approach production. Furthermore, new applications of the IKEA, effort justification
and customization effects, could become the gold standard for marketers. However, more
research is needed in order to further clarify the effects of customization, IKEA and effort
justification on the evaluation of the final good.
Key words: IKEA effect, customization effect, effort justification effect, consumer behaviour,
UNIVERSITY OF AMSTERDAM 7
1. Introduction
A new trend of network economy is developing within today’s information-based society. In
this network economy, customers becoming key contributors to production and value creation
(Lusch and Vargo, 2004). Customers not only promote particular products and brands through
eWOM (Gruen et. al., 2006), but are also involved in co-creation of production itself via the
suggestion of improvements and design solutions (Füller, 2010). Co-creation is a dynamic
social collaborative and innovative process which exists between companies and their
consumers which is directed towards value creation (Piller et. al., 2010). Rapidly increasing use
of mass customization, co-design and crowd-sourcing illustrate how important co-creation has
become. According to Piller and colleagues (2010) there are two stages in which co-creation
can occur:
1. Design of the product: Consumers can suggest their own design or suggest
improvements for the current one.
2. Product distribution/ manufacturing: Consumers can construct products themselves and
may even be able to distribute them.
One of the most compelling examples of co-creation is illustrated by LEGO, which has
managed to turn their own customers into employees. People who buy LEGO are able to
construct and create new ideas and designs and, if accepted by the community, they are
rewarded by receiving a share of their product’s sales. In recent years, there have even been
success stories where people have quit their jobs, with the hope of earning money, by
accomplishing their ideas building LEGO’s (Wagner, Pasola and Helmchen, 2017). Another
great example is vinted.com, which allows customers to trade and sell clothes which they no
longer wear. In both of these examples, the historical difference between a company and the
UNIVERSITY OF AMSTERDAM 8 customers are perceived as partners or as “co-creators of value”, which is the basis of
service-dominant logic. This is consistent with the recent shift in marketing from ‘good-service-dominant’
(G-D) logic, to ‘service-dominant’ (S-(G-D). In the service-dominant approach, suppliers offer value
proposition to the customer. If customer accepts the value proposition and engages in creating
value it turns into value-in-use (Ballantyne & Varey 2006).
In order to investigate the principles of S-D logic, this study will research the underlying
mechanisms which are active when customers participate in co-creation of value. This paper
will research the evaluations of the product, when consumers have a chance to customize it
and/or when they physically participated in the creation of it. It has been proven that people
value goods differently (Dean & Biswas, 2001). Additional value of the product might be
created through many different processes; by an object, which is believed to bring luck or
security, something that reminds us of important events or perhaps a precious gift from loved
ones.
As time passes, human nature may remain constant, yet the goods will change (Atkinson, et.
al., 1996). Additional value might be added as a result of many reasons, but why do people
overvalue goods? What are the main factors which affect the evaluation? If a consumer is given
a complete liberty to choose and construct a product according to his/her preferences, and later
constructs it successfully, would he/she then value this good the most? The phenomenon of the
IKEA effect may offer answers to these questions. The IKEA effect has many different names
in scientific literature (also known as the handmade or trophy winner effect), however, the
definition is consistent: When consumers design and construct products themselves, they tend
UNIVERSITY OF AMSTERDAM 9 Previous literature suggests that the underlying mechanism under the IKEA effect is effort
justification paradigm, which in turn creates higher evaluations for the goods that we spent time
and effort constructing (Mochon, et. al., 2012).
Researchers present conflicting evidence of how a consumer’s evaluation changes over time
and which factors play role in the evaluation process (De Groot, Wilson, Boumans, 2002). That
is why this study will look deeper into the process of evaluation and identify the factors which
affect the estimated value of a product. Given that the applications of the IKEA effect are very
diverse, more and more researchers and practitioners are becoming interested in it. Yet, some
aspects of this phenomenon retain unclear and need be tested further. (Norton, Mochon and
Ariely, 2011). Norton and colleagues suggested investigating the IKEA effect further, and
manipulate additional variables in order to see if evaluation will change.
The mission of this research is to explore on effects which together with the IKEA effect, may
better explain the evaluation of goods. Mainly, we are interested in effort justification and
customization effects. In previous studies, all of the effects were studied as a standalone
phenomena’s (Norton, Mochon and Ariely, 2001; Egan, Santos and Bloom, 2007; Tsai & Hsiao, 2004), but no one tried to find a connection between them or manipulate them together. That is
where research gap occurs. We are determined to find if evaluation can be affected the most,
when all of these effects are manipulated. There are plenty of literature about the effort
justification effect, however there is a lack of literature about the IKEA and customization
effects, mainly it is because the last two are quite new phenomena’s and requires a more detailed study. We would like to compare these effects and provide clear evidence as to which of them
has the biggest impact on evaluation. Secondly, we will investigate the willingness to pay for
the finished product. In contrast to previous studies which paid closer attention to satisfaction
(Hackman and Oldham, 1976; Frey, Luechinger and Stutzer, 2004) or likeness (Kotturu, 2014)
UNIVERSITY OF AMSTERDAM 10 interact, and what impact they would have if manipulated together before the evaluation of the
final good. While this has been suggested by previous studies (Norton, Mochon, Ariely, 2011;
Dahl, Moreau, 2007), no detailed research has been done to study the interaction effect.
Therefore, the research question of this study is: ‘To what extent do customization, effort
justification and their interaction affect consumer’s evaluations of self-made goods?’
In order to investigate the evaluation of the good (an origami pelican), a consumer experiment
will be conducted. This will provide insight not only into the variables which affect the
evaluation of the goods, but also the size of these effects. The empirical findings of this paper
will enrich the theoretical knowledge of consumer behaviour and the S-D logic framework. We
do believe that our study will serve as an incentive, to study these interesting phenomena further
in the future. Moreover, we believe that this study will enrich the theory, that both effort
justification and customization effects contribute significantly towards the evaluation of the
good, but the strongest effect could be reached by integrating them together. The results should
also provide managerial implications for marketers to consider in order to increase profitability,
effectiveness and efficiency. Our study will contribute to the success of businesses by providing
the framework of how businesses should present the goods they sell. In other words, if they
should adopt the IKEA, customization and effort justification effects in their production. In
considering the present study, managers will better understand the factors which influence the
concept of value to create a better fit between customer’s needs and preferences. Furthermore,
we will provide some strategies which may increase the effectiveness of sales.
The paper is written as follows: The literature review provides the theoretical background,
main definitions and elaboration of the hypothesis, leading to the conceptual framework.
Further paragraph of research method will explain how the research is set-up, conducted and
analysed. Than the results will be presented, following by conclusion, limitations of our study
UNIVERSITY OF AMSTERDAM 11
2. Literature review:
2.1. Literature reviewIn this paragraph of the literature review the definitions and effects mentioned in this paper
are explained. Moreover, the existing literature about customer participation in co-production,
Ikea, customization and effort justification effects is reviewed. Furthermore, hypotheses are
introduced after the relevant theory is mentioned. After reviewing the relevant literature, this
section concludes with conceptual model of this study.
2.2. Customer participation in co-production within SD-logic
Service-dominant logic is mainly market driven and customer-centric (Lusch and Vargo
2004). In the S-D approach, companies create value through value proposition. In order for the
value to be created, customer participation is mandatory. The main difference between
good-dominant and service-good-dominant logic is that the former focuses on value-in-exchange, while
the latter focuses on value-in-use. In the original paper of service-dominant logic Vargo and
Lusch (2004), presented seven major principles. However, three of them are underlying
principles which are important to the co-production dimension of S-D logic. The first principle
is: ‘the customer is always the co-creator of value’ (Vargo & Lusch, 2008). The idea of this
principle is that value creation is an interactional process; it requires the active participation of
both the supplier and the customer. The second major principle is: ‘the supplier cannot create
value itself, only offer value propositions’ (Vargo & Lusch, 2008). That is to say that companies
can offer resources, but value is only created once the customer uses them. In other words, value
is not simply delivered to the customer through the exchange as it is in G-D logic. The third
principle; ‘value is uniquely established by the recipient’ (Vargo & Lusch, 2008), suggests that
the entity receiving the benefits will develop the value derived from it based on their experience
UNIVERSITY OF AMSTERDAM 12 have been taken into account in designing the experiment and collecting the required theoretical
knowledge.
Etgar (2008) presented the relationship between customization and co-production. He
provided a model in which co-production is described as a process with five different stages
which customers are involved in. Etgar’s model consist of: circumstances (internal and external) prior to the customer’s participation, the reason for participation, the evaluation of
customer’s cost-benefit decision if one should participate, the participation and finally the evaluation and results of participation.This research is most interested in the fifth stage: results
and evaluation of customer’s participation, since this is the focus of the present study. For further research Etgar (2008) suggested the examination of all stages in more detail and under
different conditions, for example by manipulating the level of effort used. Co-creation allows
businesses to deliver more personalised and customised offers, due to the feedback collected
during the interaction with the customer in the value creation process. Terblanche (2014)
argues, that co-production is ingenuously connected to customisation and contains different
kinds of partnerships between the supplier and the customer. Therefore, co-production is a
major ingredient of creation and it should not be used interchangeably. Components of
co-creation must be studied in a great detail in order to prevent confusion. Bendapudi & Leone
(2003) focused on the economic inference of customer participation by arguing that when the
customer is engaged in production, the supplier must guarantee that the process of production
is of high quality and involving. This process should provide benefits to the consumer which
could be in the form of accomplishment, enjoyment or self-confidence. Moreover, when firms
leave assembly work to customers, they can save some cost, which in turn reduces the end price
and leaves the consumer better off in economically. In order to examine effects which affects
the evaluation of the good, we will go into more depth and isolate three of them, which might
UNIVERSITY OF AMSTERDAM 13 2.3. The Ikea Effect
According to Mochon, the IKEA effect is defined as consumers' willingness to pay more for
self-created goods than for equivalent goods made by others (Mochon et. Al. 2012). The
experiments done by Michael I. Norton of Harvard Business School, Dan Ariely of Duke
University and Daniel Mochon of Yale University, proved that self-made goods affect the
evaluation of a product. The results of these studies showed, that when people invest their own
resources into the production of a particular good (even if it’s not according to the highest
standards), they will value the end result more than if they had not put any effort into its
creation. Additional research has shown that labour leads to higher valuation only when the
labour is fruitful. When participants failed to complete a task requiring their own labour, the
IKEA effect dissipated (Norton et. al, 2011). In other words, labour leads to love only when
that labour is successful. The authors of the most prominent literature concerning the IKEA
effect also suggested investigating the effect further and with more expensive items because
perception of value between cheap and expensive goods differs.
The other researchers, who contributed a lot to IKEA effect were Dahl and Moreau, who
suggested that placing some constraints on the amount of creativity that consumers can express
leads them to be more satisfied with their eventual creations (Dahl and Moreau, 2007). They
also proposed further research aimed at identifying the boundary conditions and enhancing the
proposed theoretical framework by testing different characteristics and combinations of these
outcome and process constraints. It is very important to present clear instructions to people on
how to engage with self-service technologies. Unclear guidelines may cause consumers to fail
in achieving their desired outcomes, leading to more negative evaluations of the end product
UNIVERSITY OF AMSTERDAM 14 technology to other contexts. This area would benefit from studies in multiple contexts to
determine what relevant antecedents increase consumer readiness.
According to White and Dolan (2009) people tend to rate their jobs as the least pleasurable
activities, however they rate these activities as the most rewarding. This link between
unpleasant work and their rewarding properties was studied by academics for more than half a
century. Festinger, famous for his studies of cognitive processes and the theory of cognitive
dissonance, defined the relationship as the effort justification effect. Festingers (1957) findings
demonstrated that the more effort people put into something, the more they value it, even if the
effort does not create any satisfaction. That is why we expect that participants who will be given
a chance to construct the good themselves will evaluate it higher. This leads to the following
hypothesis:
Hypothesis 1: If participants are allowed to construct the good, then their evaluation of the value of the good should be higher (than those who do not construct the good).
Effort justification effect was found not only in human behaviour, but in the behaviour of
animals (Kacelnik and Marsh, 2002). This effect may explain many different events and
activities that we can observe daily. For example, when a person looks for a job, he/she knows
that most of the time there will be quite a few stages that he/she needs to pass in order to get
the job. Recruiters also know that if they make the recruitment process harder and more
challenging, the new employee will value his/her work more. The same logic can be applied to
clubs and fraternities where applicants need to pass challenging tasks in order to become
members. Norton and colleagues (2011) suggested that the same psychological processes are
at work in both the IKEA and effort justification effects. The effort justification effect can
UNIVERSITY OF AMSTERDAM 15 2.4. Effort justification paradigm
Effort justification paradigm, states that people value things made by themselves more than
things made by others (Egan, Santos and Bloom, 2007). In other words, people add value to the
good if they put time and effort into its creation. For example, why is it that every parent thinks
that their child is the best and the nicest kid in the world? Effort justification paradigm might
explain that; from the day the child is born, parents spend a lot of time with it and invest a lot
of resources towards its growth into a smart and good human being. Time spent with a child
also nurtures a relationship and creates biases which affect cognitive processes. The effort
justification phenomenon is frequently explained in terms of cognitive dissonance (Festinger &
Carlsmith, 1959). Cooper (2012) argues that cognitive dissonance occurs when an individual
perceives that his/her cognitions are inconsistent with his/her behaviour. For example, when
individual performs a behaviour which contradicts to his beliefs he will feel a discomfort, and
according to theory he will strive to make these cognitions consistent with his behaviour.
Festinger did not suggest why cognitive dissonance produces discomfort and changes in
behaviour and/or cognition. The action-based model of dissonance does introduce a
fundamental incentive (Jones and Jones, 2015). Effort justification is the tendency to establish
a higher value (higher than the unbiased value) to an outcome, into effort was invested. Lydall
argues that one way to investigate whether complex cognitive processing is required for effort
justification is to determine whether this effect occurs in animals. Lydall (2010) found that
similar to humans, rats place more value on the things they work hardest for. ‘Although
appealing to such a theory to account for rat’s behaviour is possible, one of the guiding
principles of comparative psychology is to interpret an animal's behaviour in the simplest
possible terms’ (Lydall, 2010, page 45). Axsom studied the effort justification affect in the
weight loss experiments. In these studies Axsom (1983) found that the notion of effort
UNIVERSITY OF AMSTERDAM 16 who perceived their freedom to participate to be high. Since an overwhelming proportion of
subjects perceived their freedom to be high, the effects found for the effort variable are
consistent with the effort justification approach (Axsom, 1983). Axsom also proposed studying
effort justification over time, because perceptions of how people allocate effort might change.
As we can see, effort justification paradigm received a lot of attention from previous literature
and indeed it is the most studied effect out of all that our study has researched. Effort
justification paradigm can be explained through cognitive dissonance theory. The effects of this
paradigm can be detected even between animals, hence effort justification paradigm, can be
adapted in different disciplines.
2.5. The customization effect
Customization effect is mainly used in mass markets, where retailers have observed the
phenomenom, that consumers like a good more when they can customize (Franke, Scheier and
Kaiser, 2009). Even though most of the time self-customization cost more, it is becoming a
popular trend, especially among millennials, who tend to look for a way to stand out from the
crowd. The customization effect may also influence how people evaluate goods. The more
customization people can use, the more they should like the good, and hence value it (Tsai &
Hsiao, 2004). With this and adding the findings about the IKEA and effort justification effects
of previous literature in mind (Norton et. al. 2011; Egan, Santos and Bloom, 2007) we expect
that the valuation of the good should be the highest when people can customize and participate
in production (construct the good). This leads to the following hypothesis:
Hypothesis 3: If participants are allowed to do both construct and customize the good, then the evaluation of the good should be the highest and most significant towards the evaluation of the good.
UNIVERSITY OF AMSTERDAM 17 By allowing customers to customize goods, companies attempt to develop a unique value to
their customers (Gilmore, Pine, 1997). Customization can therefore be used as a tool to create
a good according to individual preferences. For example, when one buys a new car, he/she can
customize the car the way he likes it. In the example of a car, a buyer can customize anything,
from colour to the suspension and engine. However, these types of customization also add
additional costs, yet consumers, particularly millennials, are willing to pay these costs in order
to stand out from the crowd. Turkay and Adinolf (2015) argues, thatcustomization presents
people with decision-making authority over the product, letting them manipulate and develop
their own experiences, based on their needs and desires. The opportunity to take part in the
creation and control the outcome of a product establish an emotional connection, which leads
to psychological ownership, or the feeling that something belongs to oneself even without valid
ownership (Margalit, 2014). According to Stump (2002) product customization has become an
important feature of business-to-business marketing. In technology-based industrial markets,
the rapid pace of technological change has increased the demand for customized products.
Buyers find customized products advantageous not only because they more precisely meet their
idiosyncratic needs, but also because customized products often incorporate the latest
technologies, which can be an important source of competitive advantage. Stump (2002) also
suggests that future research should focus on simultaneously measuring sellers as well as buyers
perceptions of how various activities moderate the impact of product customization on
satisfaction and the subsequent expectations of continuity. According to Franke and colleagues
(2009), self-designed products generate a significantly higher willingness to pay. Therefore, we
expect that participants who will be able to customize the good, will value it more. This results
in the following hypothesis:
Hypothesis 2: If participants are allowed to customize the good, then their evaluation of the value of the good should be higher (than those who do not customize the good).
UNIVERSITY OF AMSTERDAM 18 Customization effect is mediated by feelings of accomplishment and moderated by the
outcome of the process as well as the individual's perceived contribution to the self-design
process. Wang and colleagues (2010) argue, that it is unclear how efforts to customize
production would affect customer satisfaction and suggest researching this in the future.
Furthermore, Hedge and colleagues (2005) argue that the mechanisms of enhanced
customization performance should be studied in a more detail. Moreover, Paiva and Teixeira
(2007) add, that although there are a number of studies about customization, there is little
empirical evidence about this effect’s effectiveness. They therefore suggest that future research
should examine customer trade-offs in evaluating customized services and goods. This will be
tested in this research.
2.6. Elaboration of conceptual model
An illustration of the hypothesized relationships between customization and customer creation
of production towards evaluation of the final good is presented in the conceptual model in figure
1.
Figure 1: Conceptual model
Customization: Y/N
Evaluation of the
final good
Customer creation of
production: Y/N
H2 H1 H3UNIVERSITY OF AMSTERDAM 19
3. Methodology
The following chapter presents the research design and the way data was collected in order to
examine the hypothesis. Moreover, variables of the conceptual model are described and the
sample is summarized. Lastly, the statistical procedure of quantitative data analysis is
presented, which was used to test the hypothesis and answer the research question.
3.1. Research strategy
In this study, a 2(constructing the origami (builders) versus inspecting it (non-builders)) x 2
(customizing the colour of origami versus get pre-selected colour) between subject, factorial
design is used. The experimental approach was selected in order to establish different groups
which would allow participants to build and/or customise the good. Participants were randomly
divided into four condition groups (see table1). In the first condition, participants were asked
to construct the origami and they were given a chance to choose between three different colours
of origami paper; this condition tested both customization and customer participation in
production effects. In the second condition, participants were asked to construct origami, but
they could not choose the colour of it. In order to prevent individual attitudes towards a
particular colour, green colour paper was selected and presented to all of the participants. In the
third condition, people were given pre-built origami, and they could choose between three
different colours (Green, blue and yellow). The fourth and last condition, was a control
condition, in which participants were given pre-built, green origami in order to estimate value
to it. After participants constructed/ inspected origami pelican, they were asked to fill in
questionnaires (See appendix). Two different versions of questionnaires were used, one for
builders and another for non-builders. In order to ensure internal validity, all questions were
UNIVERSITY OF AMSTERDAM 20 point Likert scale with the exception of one open-ended question which asked participants to
write the evaluation of origami. Norman (2010) argues that it is common practise to use Likert
scales as an interval scale, and since it was used many times in practise Likert scales were used
as interval scale in the questionnaires too.
Participants are allowed to choose the colour of
origami.
Participants are not allowed to choose colour of origami.
Participants allowed to
construct the origami. Condition 1 Condition 2
Participants cannot construct
the origami. Condition 3 Condition 4
Table 1: 4 different conditions of the experiment
3.2. Sample and measurements
Probability sampling was used in this research since sample has a known probability of being
selected (Gallup & Newport, 2009). In other words, when a prospective participant was
approached, he or she had a chance to refuse to participate in the experiment (and it happened
quite often). It can be explained by the fact that some people were working on their own projects
and did not want to be bothered by something else. In order to collect the required amount of
participants, I followed the design of Simmons and colleagues (2011) who recommended at
least 20 participants per condition. In total 128 trials were conducted, resulting in 32 per
condition. The sample was gained during period of two weeks (for location and dates see
UNIVERSITY OF AMSTERDAM 21 and times were changed daily (Saunders, et. al., 2012). Subjects were recruited by approaching
students directly and asking them to participate in the experiment.
This method of recruitment, however, potentially threatens the internal validity, of the
experiment. The biggest potential threat comes from the so-called observer effect – which
occurs when participant is influenced by the physical presence of the researcher (a participant
might pay more attention to the experiment, because there is a researcher next to him).
According to Saunders and colleagues (2012) this could also lead to subject biases, or in other
words, participants might oppose their own beliefs and reasoning. To prevent these possible
threats, I tried to dress non-formally and keep a warm and friendly dialogue (both verbally and
non-verbally).
Data was collected at three different campuses of the University of Amsterdam. Participants
were asked to either build an origami pelican or inspect it. After doing it, they were asked to
complete a short questionnaire. I followed the same procedure used by Norton, Mochon and
Ariely (2011) in their original IKEA effect experiment, only with a different object and with
monetary conditions. In exchange for participating in my research a participant earned 1 euro.
However before evaluating the origami pelican, participants were informed that they can bid
this one euro on it and if successful they might win it. This was done intentionally in order to
prevent the endowment effect which suggests that people tend to value an object more once
they take ownership of it (Kahneman, Knetsch and Thaler, 1991). This is also why we
emphasized that participants do not own the origami and why they were asked to place their bid
from 0 to 100 in monetary terms (for example, if a participant wrote 70, we would assume he
is bidding 70 cents for the pelican). If the participants bidding price was higher than the random
number (a random number from 0 to 100) which the researcher had drawn, the participant could
keep the origami pelican. If however, the random number was higher, the participant would not
UNIVERSITY OF AMSTERDAM 22 mobile application ‘Random UX’. Moreover, at the end of the experiment, after participants
signed the participation agreement, they were given 1 euro anyway, so all of them received
either one euro or one euro and an origami pelican.
In total, 128 subjects participated in this study, however not all the questionnaires were
completed correctly. As a result, 5 questionnaires were discarded resulting in a total of 123 data
samples in this study. The age of participants ranged from 19 to 43 years old (M=23.78,
SD=3.85). 58% of participants were women, while the remaining 42% were men.
3.3. Variables
As was already shown in the conceptual model, three variables were tested. The following
part describes each of the variables in more detail.
3.3.1. Independent variable: customization
The first variable, which was manipulated was customization. This variable was measured by
the price that participants bid on origami pelican. In their primary work concerning the IKEA
effect, Norton, Mochon and Ariely (2011) focused only on customer participation in
production. They did however mention that customization might affect the final outcome of the
good; ‘’some labour allows for product customization (making a bear with one’s alma mater’s logo) – which might increase valuation’’ (Norton et. al., 2011, P. 3). That is why participants were able to choose either the colour of the origami themselves, or inspect an origami pelican
made with pre-selected colour; hence they were allowed to customize the origami or they were
not. As it was stated earlier, previous literature has presumed that a higher evaluation of a good
UNIVERSITY OF AMSTERDAM 23
3.3.2. Independent variable: construction
The second variable, which was manipulated in order to measure dependent variable was
construction or the condition in which participants were asked to put effort into constructing
the origami. As was stated in the previous literature written by Norton, Mochon and Ariely
(2011, page 3) ‘When people imbue products with their own labour, their effort can increase
their valuations’. On the basis of this argument, participants were asked to actually build the origami themselves (builders), or inspect pre-built origami pelican (non-builders). Literature
presumes, that the effort justification effect might play a role here. People who build the origami
should evaluate it higher than the ones who only had a chance to inspect it. This variable, similar
to customization, was measured by the price which participants bid on the origami pelican.
3.3.3. Dependent variable: Valuation of origami
Evaluation of the final good (origami pelican) was the variable which we wanted to measure.
The evaluation varied from 0 to 100 in monetary terms, where 100 was 1 euro and 0 was 0
cents. In order to measure dependent variable, both independent variables were manipulated
separately, as well as together in interaction condition. Some interesting results occurred, but
we will go deeper with them in results section.
3.4. Statistical procedure
Raw data from the questionnaires was entered into IBM SPSS (ver.22) statistical software.
Due to unreadable or incomplete questionnaires 5 of them were discarded (6% of total data
UNIVERSITY OF AMSTERDAM 24 done. To test the hypothesis, a regression analysis was conducted. Presentation of the results as
well as an analysis used by the SPSS are discussed in the Results section below.
4. Results
This chapter starts with a pre-analysis, where we tried to find possible biases which may
occurred when we allocated participants to different conditions. After that main results will be
presented following by overview of the hypothesis and follow up tests in order to strengthen
our findings.
4.1. Pre-analysis
Several preliminary analyses were conducted in order to check for biases when allocated
participants into different conditions. We wanted to examine if there were any significant
differences in age and gender between four conditions. Tests taken to test it and results are
provided below.
4.1.1. Age
A one way ANOVA was conducted to check for significant differences in age between the
four conditions. (Table 2). Apparently, age varied significantly between the condition groups
and this might caused differences between them. Age was significantly higher in the builder
condition as compared to the non-builder condition. F(1, 119)=9.822, p=.002 (Significant).
However, it had not differ significantly in customization condition F(1, 119)=1.904, p=.170
(Not significant) and interaction between the conditions F(1, 119)=.308, p=.580 (Not
significant). The results indicate that in one out of four of conditions, age was significantly
UNIVERSITY OF AMSTERDAM 25 this happened by chance. However, this result suggests that age should be included as a control
variable in the main analyses.
Source Df MSE F P Builder 1 136.159 9.822 .002 Customization 1 26.395 1.904 .170 Builder*Customization 1 4.265 .308 .580 Error 119 13.863 Total 123
Table 2: Age effect on different conditions
4.1.2. Gender
A second control variable which we tested was gender. We wanted to examine if there was a significant difference in the number of females and males between four conditions. To do so, a
Chi square test was used. Chi square test can be used when nominal and ordinal variables are
presented and one want to see if there is any association between them. (Nominal - where data
has no meaningful rank or order). Moreover, the sample should be randomly drawn from the
population and must be mutually exclusive. The results of this test are presented in table 3. The
was no significance in relationship between these variables, x2 = 22.169, df=32, p=.903. This
means there was no significant difference in the number of males and females in different
condition groups. Moreover, the likelihood ratio approves this statement: X=28.723, df=32,
UNIVERSITY OF AMSTERDAM 26
Table 3: Chi square test
Source X Df P Pearson Chi-square 22.169a 32 .903 Likelihood ratio 28.723 32 .633 Linear-by-Linear Association .938 1 N of valid cases 123
a. 60 cells (90.9%) have expected count less than 5. The minimum expected count is .42.
4.1.3. Conclusion of pre-analysis
It can be concluded, that some biases were found in the allocation of the participants across
different condition groups. This suggest that randomization was not 100% successful in all four
conditions. While gender seems to have no significance between the conditions, age appears to
have some significance on builders who were not allowed to customize condition. Most
probably this happened by the coincidence, therefore the age should be controlled for as a
potential confounding variable.
4.2. Analysis
A two-way factorial ANOVA was performed to examine the effect of customer participation
in production and customization on the evaluation of the good (See table 4). The outcome
variable was found to be normally distributed and equal variances are assumed based upon
results of Levene’s test (F (119) = 3.171, p=.127). There was found to be statistically non-significant interaction between the effects of customer participation in production and
customization on evaluation of the good, F(1, 123) = 2,801, p=.097, η²=.023. However, since
the level of significance in interaction between the independent variables groups is below 0.010
UNIVERSITY OF AMSTERDAM 27 assume these results to be marginally significant. Even though, this practise is being criticized
by literature, it is still being used by many researchers (Pritchet et. al., 2016).
Hypothesis 3 proposed thatif participants are allowed to do both construct and customize the
good, then the evaluation of the good should be the highest and most significant towards the
evaluation of the good. Results indicate that interaction condition does not significantly affect
the evaluation, however a marginal level of significance was reached (p=.097). Furthermore,
the evaluation in interaction condition was indeed highest between all of the conditions
(M=.6452, SD=.24219). As a result, hypothesis 3 is partially supported.
The customization effect was not found to be statistically significant towards the evaluation
of the good, F(1, 123) = 2,453, p=.120, η²=.021. Hypothesis 2 proposed that if participants are
allowed to customize the good, then their evaluation of the value of the good should be higher
(than those who do not customize the good). Results indicate, that even though there was no
statistically significant effect of customization on evaluation of the good, the final outcome was
valued more compared to the control group. In customization with no construction condition
evaluation was higher (M=.5160, SD=.18097) compared to control group (M=.5043,
SD=.28668). Even bigger difference occurred in customization and construction condition, where evaluation was higher (M=.6452, SD=.24219) compared to control group (M=.5043,
SD=.28668). Considering these findings, hypothesis 2 is partially supported.
Customer participation in production outcome showed no significance towards the evaluation
of origami, F(1, 123) = 2,792, p=.0.97, η²=.023. Unfortunately, there were no simple main
effects found. However, since the level of significance between the independent variables
groups is below 0.010 (p=.097) we assume this output to be marginally significant. Hypothesis
1 proposed that if participants are allowed to construct the good, then their evaluation of the
UNIVERSITY OF AMSTERDAM 28 that a marginal level of significance was reached, however the evaluations between builders
who did not customize (M=.4887, SD=.25146) was lower than the evaluations of the control
group (M=.5043, SD=.28668). The evaluation of builders who were able to customize the
origami (M=.6452, SD=.24219) was higher than the evaluations of control condition. As a result
hypothesis 1 is partially supported.
For interaction and builders condition a post hoc analysis will be done in order to do a follow
up for marginally significant effect.
Table 4: Results of the experiment
Source SS Df MS F P η² Corrected Model .640a 5 .128 2.168 .062 .085 Intercept 1.486 1 1.486 25.180 .000 .177 Age .126 1 .126 2.129 .147 .018 Gender .016 1 .016 .266 .607 .002 Builder .165 1 .165 2.792 .097 .023 Customization .145 1 .145 2.453 .120 .021 Builder*Customization .165 1 .165 2.801 .097 .023 Error 6.903 117 .059 Total 43.561 123 Corrected Total 7.543 122
a. R squared = .085 (Adjusted R squared = .046).
Table 5: Summary of hypothesis results.
Hypothesis Verdict
H1
If participants are allowed to construct the good, then their evaluation of the value of the good should be higher (than those who do not construct the good).
Partially supported
H2
If participants are allowed to customize the good, then their evaluation of the value of the good should be higher (than those who do not customize the good).
Partially supported
H3
If participants are allowed to do both construct and customize the good, then the evaluation of the good should be the highest and most significant towards the evaluation of the good.
Partially supported
UNIVERSITY OF AMSTERDAM 29 4.3. Post Hoc tests
According to Norusis (2006) Post hoc tests are created for cases in which the researcher has
already achieved a significant omnibus F-test. Post Hoc tests will be done to better understand
the marginally significant effect of builders and interaction (builders*customization) terms. The
results of this tests is provided in plot and tables below. In builders condition a post hoc test
revealed that the difference in evaluation of the origami was statistically significant between
builders who customized the origami themselves and builders which were presented a
pre-selected colour of origami (p=.023). Builders who customized (M=.6452, SD=.24219) valued
origami more, than builders who did not customize (M=.4887, SD=.25146). Moreover, there
was no evidence of statistical significance of difference between non builders who customized
the origami and non-builders who did not customize the origami (p=.965). Non-builders who
customized (M=.5160, SD=.18097) valued origami pelican less than non-builders, who did not
customized (M= 5043, SD=.28668).
In the interaction condition outcome showed similar results. The evaluation of the origami
pelican was statistically significant between people who customized the origami and were
allowed to construct it versus the ones who were allowed only to inspect it. (p=.019).
Participants who customized and constructed origami (M =.6452, SD=24219) evaluated it more
than participants who customized but did not construct the origami (M=.5160, SD=.18097).
Participants who were asked to only inspect the origami showed no statistically significant
difference in evaluation, compared to participants who were asked to construct origami but no
customization were allowed. (p=.968). Non-builders, who did not customize (M=.5043,
SD=28668) evaluated origami more, than builders who did not customize (M=.4887, SD=.25146). This is an interesting outcome, since people who actually put effort and
UNIVERSITY OF AMSTERDAM 30 constructed the origami evaluated it less than people who did not put any effort into constructing
the origami, customization effect remaining constant in both of the conditions. As we can see
from the follow up tests, the results do follow the consistency predicted by literature reviewed
in the beginning of the paper. We found a marginally significant effects in customer
participation in production (IKEA) and interaction effects towards the evaluation of the origami
pelican.
Table 6: Post Hoc test for builder’s condition
95% Confidence Interval for Differenceb
Customization: (I) Builder (J) Builder (I-J) Std. Error Sig.b Lower
bound
Upper bound No
Customization
Non builders Builders -.003 .064 .968 -.129 .124 Builders Non builders .003 .064 .968 -.124 .129 Customization Non builders Builders -.150* .063 .019 -.275 -.025
Builders Non builders .150 .063 .019 .025 .275
* . The mean difference is significant at the .05 level.
b. Adjustment for multiple comparisons: Least Significant difference (equivalent to no adjustments).
Table 7: Post Hoc test for interaction condition
95% Interval for differenceb
Builder (I) Customization (J) Customization (I-J) Std. Error Sig.b Lower bound Upper Bound Non builders No customization Customization .003 .063 .965 -.123 129 Customization No customization -.003 .063 .965 -.129 .123 Builders No Customization Customization -.145* .063 .023 -.269 -.020
Customization No customization .145 .063 .023 .020 .269
*. The mean difference is significant at the .05 level.
UNIVERSITY OF AMSTERDAM 31
Figure 2: Plot of the results
5. Discussion
In order to answer the research question of this studythe findings of our study will be discussed
in this chapter. First the results will be discussed about how different effects affects the
evaluation of the good. Furthermore, the outcome of this research will be compared to the
existing literature. Subsequently, theoretical and managerial implications will be presented,
following by limitations and a number of suggestions for future research.
UNIVERSITY OF AMSTERDAM 32 5.1. General discussion
The purpose of this paper was to contribute to existing literature about the effect of customer
customization and participation in production on the evaluation of the good. This study about
the evaluation of the good is part of growing trend in academia where researchers are paying
devoting increased attention to the psychology underlying consumer involvement. This
direction has become popular as a result of more and more businesses shifted from viewing
customers as recipients of the product, to that of a co-founders of its value (Vargo & Lusch,
2004; Prahalad & Ramaswamy, 2000). Previous literature has considered all of the effects
separately, however, the combination of customization effect and customer involvement in
production together has not yet been studied.
This research was done by observing, how people would evaluate the goods differently when
all of the effects were used separately and together. The highest evaluation of the origami
pelican was hypothesized under the interaction (customization*building) condition, however
results indicated only a marginal significance which lends only marginal support to the
hypothesis. Furthermore, higher evaluations of the good were hypothesised in the conditions
where participants had the opportunity to only construct or only customize the good. The data
collected in the customization condition did not prove to be significant, while data from the
customer participation in product condition did. Furthermore, evaluations varied between the
conditions, resulting in partial support of the hypotheses. A more detailed overview of each of
UNIVERSITY OF AMSTERDAM 33
5.1.1. Customization effect
Even though, the results of present experiment (N=123) did not show the required level of
significance in order to state that we found significantly main effect, we observed a higher
evaluation of origami which was customized. This means, that findings are in part consistent
with those previously obtained in the literature, stating that customization should increase the
actual evaluation of the good (e.g. Tsai and Hsiao, 2014; Turkay and Adinolof, 2015; Hedge et
al., 2005). Participants who were able to customize, valued the origami pelican higher versus
participants who were not able to do so (M= 58,3 vs. M=49,6). As you can see in the table
below, these results partially support the thought that people who can customize the good, will
value it more. This can be explained by the idea that individual’s idiosyncratic needs are met
and that therefore they would value the good more (Stump, 2002). Moreover, we have a clear
difference between the two conditions of customization and no customization, providing us
insights of importance of customization when creating the good. According to Ostrom and
Iacobucci (1995) customization may be valued more because it may show high quality or an
accepted fit for ones preferences. That is the main reason why it should be allowed for
customers to customize the good prior to their purchase. Customization does not only increases
customer’s satisfaction and evaluation of the production, but also allows customers to express themselves, and quite often for this possibility, customers are ready to pay extra (Svensson &
UNIVERSITY OF AMSTERDAM 34
5.1.2. Effort justification and IKEA effects
Builder’s conditions (who were testing customer participation in production and IKEA effects)
showed marginal statistical significance towards the evaluation of the good. Our results
partially go in line with the outcome predicted by previous literature which stated that people
tend to value goods more, when they add effort and time into constructing something (Norton
et al., 2011, Dahl and Moreau, 2007). In interpreting the data, an additional interesting result
appeared. Builders who did not customize evaluated the origami pelican less, than participants
who were able only to inspect it (M=48,8 vs M=50,4). This might happened, because labour
leads to love only when the good is made succesful, because otherwise IKEA effect dissapear
(Norton et al., 2011). This finding is in line with the findings of Savitsky, Medvec and Gilovich
(1997) which argue that failing to complete tasks (in this case completion of origami pelican)
has unfavourable consiquences. Indeed, during the experiment, we noticed, that some people
were disappointed of the final outcome of the origami and that is why they valued it less. A fail
to build nice origami created mistrust and dissapointment in ones work, resulting in low
evaluations. One more explanation could be, that participants who were asked to construct the
origami, did not spend enough time to attach to their creation. The average time to build the
origami was around 10 minutes, which is not enough to tie up to your creation. We assume, that
a longer time interval of construction would have different results. For example, when people
try to asembly original IKEA furniture, they tend to work for at least a few hours, which leds
them to attach to their creation more. Eventhough our effort to show the importance and
effectiveness of customer participation in production and IKEA effects was unsuccesful, we do
believe that a different setup of the study would bring different results. We will elaborate on
UNIVERSITY OF AMSTERDAM 35
Figure 3: The mean values of evaluation between different conditions.
5.1.3. Interaction effect
As it was in the builder’s condition, the interaction condition showed marginal significant
effect towards the evaluation of the good. It comes in line with literature, which presumes that
the evaluation of good which is constructed and customized by the individual himself, should
be valued the most (Norton et al., 2011). Since previous literature had never studied these
effects in an interaction condition, but solely hypothesised that evaluation in the interaction
should be the highest, we decided to research this. Indeed, the mean value of interaction effect
was the highest, but unfortunately not high enough to reach the main level of significance
(M=64.5 compared to M=50.4). Since we discussed our findings, we are now able to answer
our research question: To what extent do customization, effort justification and their interaction
affect consumer’s evaluations of self-made goods? The answer is that customization effect and customer participation in production do effect evaluations of self-made goods and in our
experiment they increased the evaluation. However only customer participation in production
and interaction conditions resulted in marginally significant effect towards the evaluation, while
UNIVERSITY OF AMSTERDAM 36 condition where valuation was the lowest, but according to previous literature it supposed to be
at least higher than the evaluations in the control group (Mochon et. al. 2012). We assume this
happened due to unsuccessful labour, when people did not manage to reach the wanted result
and depreciated their own work. Most of the findings of our study go in line with previous
literature and promote the use of customization and customer participation in the production/
IKEA effects in practise.
5.2. Theoretical contributions
The results of the research have some theoretical implications. The study contributes to
literature about effort justification, customization and IKEA effects, as well as how differently
they do affect the evaluation of the good. Having in mind, that research on the interaction effect
of customization and effort justification effects is rare, this study is one of the first in researching
this and its affect towards the evaluation of the good. Furthermore, this study contributes to
research about S-D logic and most importantly to co-production. Given the fact that S-D logic
is becoming the very important in marketing studies, there is still not enough up to date research
on this topic. This paper enriches current literature with a few findings.
Firstly, the highest evaluation of the final good can be reached when the effects that we
examined are manipulated together. This has not been studied before, and should add new ideas
to academia and marketing science. Secondly, we found that effort justification/ IKEA effect
not always guarantees a higher willingness to pay. If construction of the product is unsuccessful,
the IKEA and effort justification effects disappear leading to lower evaluations. This idea goes
in line with findings by (Norton et. al., 2011) that customers tend to value their creations only
when their labour is successful. Thirdly, our results imply, that when compared between
themselves. Effort justification effect is more significant towards the evaluation of the final
UNIVERSITY OF AMSTERDAM 37 people’s psychology than customization effect. However if used unsuccessfully, effort
justification may lead to worsen results, while it is not the case where customization is used,
since participants externalize their preferences.
5.3 Managerial implications
Companies need to learn how to be more flexible and reactive in their reaction to customers
in order to co-create value with them. This calls for a change in firm’s structure and strategies
(Leavy 2012). Managerially, results of this research are important, because they do help to
identify levers, which if manipulated correctly, may increase the effectiveness of business. We
would like to update the findings by Ostri and Iacobucci (2015) that managers who seek
satisfaction should make sure they excel at customization. As our results indicate, customization
increases the value of the final good too, hence the willingness to pay increases (as well the
willingness to pay extra). That is why, customization of goods can be a smart strategy, which
would results in happier customers, who will spend more. As it was with the IKEA paradox and
its underlying mechanism, the effort justification effect, we would suggest to use the production
method, where people could put their own effort into constructing something and by that, get
attached to their creations. This is a useful strategy because of two reasons. Firstly, self-made
goods may signal that a person is competent enough to build something by himself. It would
make a person proud of himself and it also shows to others what he is capable of (Spence, 1973).
Secondly, goods that require some assembly are in many cases a little bit cheaper than already
pre built ones. This is because the costs of production are decreased. Buying these cheaper
goods may lead to a more positive feeling, since the customer can associate himself as being a
‘smart shopper’. (Schindler, 1998). However as we observed, businesses should create the tasks which are not too hard to carry out and they also should provide clear instructions, because as
UNIVERSITY OF AMSTERDAM 38 it was stated by Norton and colleagues (2011) labour only leads to love when it is successful.
Managers should try to create tasks that creates highest value while remaining inside the scope
of the majority of consumers abilities. Even though we did not reached the significant effect
level in all conditions, the results of this study do follow the previous literature and interpret
results as such that the value of the goods increases when customer gets involved in the
production process and is allowed to customize the product. To conclude, by letting customers
to customize and assembly the good, businesses may not only save some costs, but also are able
to charge premium prices and in the end, customers will be happier and more satisfied. That is
why it is recommended to use customization and encourage to let customers participate in the
production of goods.
5.4. Limitations and future research
The first limitation of this study is the sample used for collection and analysis of data. The
results are based on evaluations of origami pelicans only created by students of the University
of Amsterdam. This limits the generalization for other groups of people. The feasibility in
approaching this group of people is the central reason to collect data from this sample, however
some potential bias may have appeared. The second limitation is our method of data collection.
We ran our experiment and data collection in a short time interval, limiting the possible analysis
of changes that may occur in the participants or environment. Additionally, there is a lack of
control in the field experiment and again additional biases may occur. Third limitation that may
have affected the final output was the range of evaluation of the origami that participants used
in our experiment. We used a range between 0 and 100 in monetary terms (0 as a minimum and
UNIVERSITY OF AMSTERDAM 39 since participants had much more options to choose from and a greater differences might be
observed.
Future research should include items which are more expensive and require more time to
assembly. Moreover it would be interesting to see how people evaluate the goods which they
have no possibility to obtain, compared to the goods that people are confident of receiving. In
other words, it would be interesting to see the endowment effect role in similar design study as
ours. Finally, more research should be done on the unsuccessful labour or on instructions that
leads to challenging construction work. It would be very interesting to see the evaluations, when
people would be asked to evaluate their work in the middle phase of construction (for example
until it was still successful) and then in the end of the construction if the work was not
successful. We suggest that evaluation in the mid phase would be higher than in the end phase,
however this should be further researched. Last suggestion for future research would be to allow
people to customize the good the way they like, and not to limit them with possibility to choose
out of just a few colours. When customizing people could choose shape, ornaments and full
design of the good. We suggest that a full consideration of individual preferences would
UNIVERSITY OF AMSTERDAM 40
References
Alvesson, M., Sandberg, J., (2011). Generating Research Questions through Problematization.
Academy of management. Vol. 36 no. 2. Pp. 247-27.
Atkinson, R. L., Atkinson, R. C., Smith, E. E., Bem, D. J., Hoeksema, S. N. (1996). Hilgard's
Introduction to Psychology. Harcourt College Publishers. Thirteen edition. Pp. – 462-468.
Axsom, D, Cooper, J, (1985). Cognitive Dissonance and Psychotherapy: The Role of Effort
Justification in Inducing Weight Loss. Journal of experimental social psychology, 21,
149-160.
Ballantyne, D.,Varey, R.J., (2006). Creating value-in-use through marketing interactions: The
exchange logic of relating, communicating, and knowing. Marketing Theory, 6(3), pp.
335–348. doi:org/10.1177/1470593106066795
Bendapudi, N. & Leone, R.P. (2003), ‘Psychological implications of customer participation in co-production. Journal of Marketing, 67(1), pp. 14–28. http://dx.doi.
org/10.1509/jmkg.67.1.14.18592
Buhren, C, Plebner, M, (2014), The Trophy Effect. Journal of Behavioural Decision Making,
Volume 27, pp. 363–377.
Caulkins, J. P., Feichtinger, G., Haunschmied, J., Trangler, G. (2006). Quality Cycles and the
Strategic Manipulation of Value. Operations, Research, Vol. 54. Issue. 4. Pp.- 665-667.
Cooper, J. (2012). Cognitive dissonance theory. The handbook of theories of social