FACULTY OF ECONOMICS AND BUSINESS
Amsterdam School of Economics
Novelty Seeking and Inertia Bias in the Decision Making Process
by
Maria Muravina
11087439, ECTS 15Thesis for the degree of Master of Science
Behavioural Economics and Game Theory
This document is written by Maria Muravina who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
ABSTRACT
FACULTY OF ECONOMICS AND BUSINESS Amsterdam School of Economics
Thesis for the degree of Master of Science
NOVELTY SEEKING AND INERTIA BIAS IN THE DECISION MAKING PROCESS by Maria Muravina
Given that both novelty seeking and inertia bias play a crucial role in the decision making process when it comes to novel stimuli, there is very little research in the literature that looks at both of these factors together. The study offers a new experimental design that facilitates exploring the relationship between novelty seeking and inertia. It consists of two questionnaires aimed at each factor, and an actual choice of an experience that has varying degrees of intrinsic levels of novelty and status quo framing between 3 treatments. The results were not statistically significant, but there were some interesting findings: the correlation between inertia and novelty seeking was surprisingly very low and there was also a preference reversal observed.
Title Page i
Declaration of Authorship ii
Abstract iii
Table of Contents iii
1 Introduction 1
1.1 The Experiment and Findings . . . 2
1.2 Dissertation outline . . . 2
2 Literature Review 3 2.1 Introduction of the Main Concepts . . . 3
2.1.1 Novelty Seeking . . . 3
2.1.2 Inertia . . . 4
2.2 Previous studies . . . 5
2.2.1 Priming for novelty or loyalty . . . 5
2.2.2 Optimal distinctness . . . 7
2.2.3 Theorem of interior maximum: Reference point and loss aversion 8 2.2.4 Curse of innovation . . . 9
2.2.5 Risk, Uncertainty, Novelty . . . 10
2.2.6 Insiders versus Outsiders . . . 10 v
2.3 Inertia experiments . . . 11
2.3.1 Ball lottery . . . 11
2.3.2 Questionnaires . . . 12
2.3.2.1 Natural experiments . . . 13
2.3.2.2 Vignette questionnaires . . . 14
2.3.2.3 Choice experiments to elicit preferences in consumption bundles . . . 17
2.3.3 Product Choices . . . 18
2.4 Novelty seeking tests . . . 18
2.4.1 Tri-dimensional Personality Questionnaire . . . 19
2.4.2 Novelty Seeking Scale . . . 19
2.4.3 Sensation Seeking Scale . . . 20
2.4.4 Tourism . . . 21
3 Experimental design 24 3.1 Protocol . . . 25
3.2 Status Quo Questionnaire . . . 26
3.3 Novelty Seeking Questionnaire . . . 28
3.4 Pen Change treatments . . . 29
3.5 The order . . . 30 3.6 Research Question . . . 31 3.7 The Hypothesis . . . 31 4 Results 32 4.1 Correlation . . . 32 4.2 Regression analysis . . . 32 4.3 Graphs . . . 37 4.4 Between-Treatment Analysis . . . 38 4.5 Demographics . . . 39 5 Discussion 41 6 Conclusions 46 .1 Appendix 1 . . . 48 .2 Appendix 2 . . . 48 .3 Bibliography . . . 57
1
Introduction
This project was inspired by the observation of the inconsistencies between people fanta-sizing about using science fiction- like technology and their lack of tendency to adopt it in practice. Why have projects like Segway and Google Glass failed, when audiences have been mesmerized by products alike in the futuristic films? According to [Gourville, 2005], new products fail to gain market acceptance 40%-90% of the time. The more innovative the product is, the more likely it is to fail. Why have we been drawn to and jealous of Iron Man’s house artificial intelligence assistant, but when Mark Zuckerberg created an analogous Jarvis, the public took it quite skeptically and didn’t show much eagerness in adopting something like that in their homes? To understand such dilemma it would be necessary to analyse the prospects of the product on the market, including the public’s interest in the technology and the willingness to adopt it.
This study is aiming to look at the paradox of the two polarizing forces of inertia and novelty seeking. This juxtaposition occurs both in innovations in technology and beyond. People face choices to experiment with something new or to stick with the old and known every day in every area of their lives. Would you order the same dish in the restaurant or would you try something new occasionally? Would you try a fashionable but drastically different haircut? Would you take a job opportunity in a foreign country or stay where you
are? Should you mingle with strangers or chat with your friends at a gathering? These dilemmas occur all the time having varying impacts on our lives. Even though most of us hope for a change for the better, we are rarely open to the change when opportunities present themselves. These reservation and reluctance go very much in conflict with our aspirations.
1.1 The Experiment and Findings
The experiment of this study looks at the relationship between individuals’ levels of inertia and novelty seeking from questionnaires that is compared to an actual choice of experiences. For an observation of the interaction between these two factors the choice has varying degrees of intrinsic novelty and status quo between treatments. The experimental design is original, and there has been a lack of research that looks at both of these factors simultaneously. The study had 63 participants, and the results did not show statistical significance. There were however insightful observations including low correlation between novelty seeking and inertia, and preference reversal between treatments. The study offers a new setup to investigate these phenomenon that can be further looked into with a bigger sample size.
1.2 Dissertation outline
Chapter 2 draws together relevant literature and theories suggesting methods for analysis of novelty seeking and inertia both separately and jointly. We will look at the roots of these forces and what other heuristics biases they are related to. We will investigate how the decision making process in relation to novel stimuli can be affected and what techniques firms can use to their advantage to overcome set bias. We will later move on to investigate experimental setups and questionnaires that explore novelty seeking and inertia. Chapter 3 methodology behind the experiment. Chapter 4 presents the results. Chapter 5 discusses results and their relevance to other studies.
2
Literature Review
The following part of the analysis will define the two key factors- inertia and novelty seeking, put them into the context of the scientific research and explain how they affect people’s attitudes towards novel stimuli.
2.1 Introduction of the Main Concepts
2.1.1 Novelty Seeking
The internal and motivating force that drives an individual to seek out new information has been known as Novelty Seeking [Hirschman, 1980]. Novelty seeking behaviours are ob-served from birth [Flavell, 1977]. Its evolutionary advantage lies in learning and exploring the world. Consuming even presently useless information, builds up the ’bank’ of infor-mation which may show to be useful in the uncertain future scenarios [Hirschman, 1980]. Cloninger et al. [1991, 1993] has referred to novelty seeking as a heritable personality trait that entails excitement and the feeling of exhilaration from novel stimuli. This branch of
research suggests that the extent of novelty seeking depends on individual differences of the neurotransmitter dopamine, which affects behaviour of approaching and consuming stimuli through the reward system [Schultz, 2015]. The individual differences in dopamine have been linked to specific genes [Prolo & Licinio, 2002], which suggests that there is a certain genetic predetermination of novelty seeking. Rogers et al. [1971] however found innovativeness (which they define as easiness and fast speed of adopting innovation) to be highly correlated with sociological factors like urbanization, occupational status and education. Additionally we can expect social norm and peer pressure to affect individual’s openness to new stimuli. Priming has been shown to affect the level of individual novelty seeking [Fishbach et al., 2011], which suggests that even immediate slight changes in the environment can sway people towards or away from novelty seeking. Novelty seeking has been linked with higher risk preference, sparking the same neural activation [Wang et al, 2015]. Hence, there is extensive research that looks at how genetics, personality traits, social and urban factors as well as immediate environment determine the level of nov-elty seeking. Naturally, novnov-elty seeking is also related to all of the heuristics and biases of the reluctance to act upon novel stimuli, otherwise known as inertia. However, their relationship is underexplored.
2.1.2 Inertia
The opposing force that halts the decisions to go for something new is inertia. The term, initially introduced by Newton, entails that in absence of external resistance, a moving object will keep moving in the same direction. However, since then the term has been used in the fields of medicine, psychology, sociology, organization science and others to describe a certain type of human behaviour [Allos-Ferrer et al., 2016].
Inertia’s evolutionary advantage lies in the minimization of efforts, filtering cascades of information and the lack of necessity to make the same decision twice [Hekeeren et al., 2004; Cialdini, 2009; Berger, 2016]. Without the concept of the familiar every choice made would be a ”life or death” decision [Berger, 2016].
Behavioural economics link inertia with several heuristics and biases. Status quo bias [Samuelson & Zeckhauser, 1988; Ritov & Baron, 1992] goes under the umbrella of in-ertia. It stands for the preference for the default option (no matter how arbitrary or dysfunctional it might be) and the possession of some higher standards of proof for the new option. In some cases it can be explained by the sunk cost effect where the costs
of choices that have already been made are mistakenly factored into the decision making process [Potts, 2010].
Risk and loss aversion can also result in more inert behaviour under novelty [Gourville, 2005]. Risk aversion comes into play as people find it hard to estimate costs and benefits of the previously unknown stimuli. Whereas loss aversion kicks in when people view the original option as a given and a reference point with opting out of viewed as a loss. Cialdini [2009] explained that conservatism and preference for consistency are rooted in the unwillingness to spend any mental power on contemplating other options. Another contributing factor is the failure to imagine one’s future being different from the present [Earl, 1986]. Additionally, some decisions are mentally overwhelming and people opt out of making a decision by falling into a default, as in the example with organ donations which we will explore further on [Johnson & Goldstein, 2003].
2.2 Previous studies
2.2.1 Priming for novelty or loyalty
So far we have discussed why inertia and novelty seeking play a role in the decision making in face of novel stimul. However there is little research that looks at both of these factors together. [Fishbach et al, 2011] tackled the issue via conducting a series of experiments on priming- exposure to stimuli that subsequently affects future choices. The study looked at whether priming can sway people’s preferences towards variety or loyalty and sameness. Novelty falls under the umbrella of variety, as since when people seek out novelty, they are in fact looking for a different stimuli. Loyalty and sameness are connected to inertia through a preference to stick with the same choice and not to venture out.
In one of the experiments [Fishbach et al, 2011] participants had to fill in a survey that explored how much variety they would incorporate into their product choices (eg. sham-poo, songs etc). The survey was clipped to a book with a title containing either word ’bored’ or ’loyal’, with control group having untitled books 1. The study found that
sub-jects from the ’loyal’ prime group reported higher preference for consistency and lower for variety, and vice versa in the ’bored’ prime group. This showed that situational cues
1Participants were students who were individually approached at a student center- a similar setup to
have an impact on swaying the person towards to or away from the variety. Relating to our discussion, since novelty is a variation of variety, and loyalty incorporates inertia, this experiment showed that priming can sway individual preferences towards novelty and away from inertia or vice versa.
Another experiment [Fishbach et al, 2011] was on the subliminal priming- subconscious stimuli designed to implicitly affect memory and influence the response to other stimuli. Participants were given a series of words, representing sameness or variety. Each word was followed up by a 30 milliseconds exposure to a priming word that had either a positive or negative connotations. The exposure was too short for the person to read, but it would still be processed by subconsciousness, hence altering future decisions. Subjects were later asked to choose a snack (from 6 options) for each out of the 5 weeks. The amount of different snack options a participant chose put him on the variety scale. The study showed that subliminal priming into praising or devaluing sameness or variety had people choosing more of the same kind of a snack and vice versa. Since novelty is connected to variety, and inertia implies sticking with the same option, these results suggest that choice behaviour towards inertia or novelty can be impacted with stimuli that is not even consciously registered in the brain and is in a separate task category.
The last experiment in the series [Fishbach et al, 2011] had people rank products on one of the scales of good-bad or same-different, depending on the question. For instance under a question ”How good do you think soy milk tastes?” both scales were displayed, but the ’good-bad’ scale needed a tick. The priming would occur in the positioning of the labels on the scales. In the same-is-good condition, ’same’ and ’good’ would be on the same side of scales, and ’different’ and ’bad’ would be on the other side; and vice versa for the same-is-bad condition. In the second part of the experiment subjects had to pick one snack, evaluate it and pick again, be it different or identical. The results showed that linking ’sameness’ to positive concepts encourages loyalty, and linking it to negative qualities encourages variety seeking.
This body of research shows that behaviour and preferences towards variety or same-ness can be enhanced or suppressed with outside cues with both subliminal and more obvious messages. This ties to the relationship between novelty and inertia, as prefer-ence to alternate experiprefer-ences results in seeking novelty, and preferprefer-ence to stick with the same experience is explained with inertia. Hence, this research suggests that the individ-ual’s preference of familiar or novel stimuli is fluid and can be affected even using micro alternations in the environment.
2.2.2 Optimal distinctness
Optimal distinctness focuses on the product’s perception and attempts to balance out and play into humans’ preferences between novelty seeking and the inertia bias that is holding people back into the comfort of the familiar. The perfect example of radical innovations being adapted to seem more familiar is two tell-tales of the first automobile and the first light bulb. The introduction of automobiles in the late 19th century in rural America was met with backlash- they were so drastically different to horse carriages in how they looked and operated that they frightened both horses and citizens. Cars were nicknamed ’Devil’s Wagon’ and there were laws introduced to halter the ’intrusion’ [Berger, 1980]. The brilliant counteraction came from introducing ’Horsey Horseless’ in 1899- an automobile carriage with a fake horse’s head [Berger, 1980]. In a similar manner when Edison invented the light bulb he made it look similar to an already existing kerosene lamp [Hargadon & Douglas 2001]. More current examples involve TiVo’s digital video recorders being made to look very similar to VCR’s or DVD’s even though their insides have nothing in common. For the same reason veggie burgers have grill marks and the save icon on the computer has the shape of a floppy disk [Hargadon & Douglas 2001; Gourville, 2005; Rindova & Petkova, 2007; Berger, 2016]. In essence, for people with high inertia bias, this method cloaked the revolutionary technology into something more familiar and assisted the market adaptation.
On the flip side if the product is too similar to the exisitng ones, then consumers are not driven to buy it. The pleasure of familiarity tends to run out at some point and gives way to boredom, and the complexity of a stimuli is directly proportionate to the time of habituation (the diminishing of an innate response to a frequently repeated stimulus) [Berger, 2016]. For instance we get bored of a meal or a song faster than we do of a restaurant or a romantic partner [Berger, 2016].
Hence, the solution is balancing characteristics that play into novelty seeking and inertia bias. This perfect combination of novelty and familiarity has been referred to as ’optimal level of product novelty’ [Rindova & Petkova, 2007] or the Goldilocks effect [Berger, 2016]. However, there is another reason to call for the optimal distinctness- the reference point.
2.2.3 Theorem of interior maximum: Reference point and loss aversion
Wathieu [2004] proposes a model of consumer habituation, that carries a similar message of moderate level of habituation/novelty. His theorem is built on an adaptive behavior of a changing S-shaped value function. Familiarity and gains and losses relative to the reference point, similar to those mentioned before, play a key role in this model. Wathieu [2004] illustrates the shape of the value function, with a hypothetical example of driving past a burger place. On day zero when one drives past the restaurant and she considers getting a burger- not having had it before, the habituation level is zero; the reference point is not having a burger; gains from getting it are big (compared to not having it); and not getting a burger bares no losses (compared to not having it). Assuming she decides to eat it on day zero, for the day one her habitual level has increased and reference point has shifted towards having a burger. This means that gains from getting a burger are slightly smaller than before; and her losses from not getting a burger increased. This is where the loss aversion bias [Kahneman & Tversky, 1979] kicks in, as losses psychologically cause more pain, than gains cause pleasure. Hence, because losses loom larger than gains, and losses have increased, the value of the burger has also increased. At some point the consumption would be driven more by the unwillingness to bear the losses, than by marginalized gains. However, according to [Stevens, 1957], there is a ’psychological law’ that states that around the habitual level the sensitivity goes down, meaning that gains and losses are perceived less significant as they become larger. The habituation at this point will be too high, and one might actually be in a state of decreasing willingness to pay. Therefore, the willingness to pay is maximized when habituation (or familiarity) is not too high and not too low [Wathieu, 2004].Another interesting point brought by Wathieu [2004] is that the periodicity of goods impacts their habituation and sensitization rates affecting willingness to pay. So for example goods that inherently have periodicity and breaks in consumption would have a much slower rate of habituation. The examples of such goods would be vacations, seasonal items, magazines and so on. Hence, it would be interesting to look whether the reference point adjusts differently between durable goods (eg. a phone) and non-durable goods (eg. a massage) and how it affects familiarity level and novelty seeking.
2.2.4 Curse of innovation
’Curse of innovation’ is another potential explanation for why highly innovative products fail [Gourville, 2005; Rindova & Petkova, 2007]. It refers to the gap between producer-intended and consumer-received values and also occurs due to the loss aversion bias and the difference in reference points between consumers and producers. Producers’ reference point is the innovation itself. They have invested time and resources into creating a certain product with certain new features and to them, not having these features in the antecedent product is viewed as a loss. Hence, due to loss aversion, the loss is overvalued and from the producer’s perspective the benefit of the innovation is overestimated. From the consumers’ point of view, as discussed earlier, the reference point is the en-trenched product they’ve been using before considering to buy the novel alternative. So for instance, when contemplating online shopping, their reference point would be a famil-iarity of a physical shop [Gourville, 2005]. Therefore, the benefits from the novel good (eg. convenience) are viewed as a gain and the features of the original good (eg. con-trol) that have to be replaced as a cost. Yet again, because of the loss aversion bias, consumers underestimate the benefits of the innovation. The gap between producers’ and consumers’ values leads to a market failure. Producers put more value on novel products than consumers. This tendency suggests that innovators do not account for inertia biases and discrepancies in reference points and loss aversion, and substantially overestimate novelty seeking in consumers.
One of the strategies to tackle this issue is targeting first-buyers that are not familiar with the existing alternative and hence their reference point is not compromised, instead of trying to break the habits. For example, snowboarding was much faster adopted by young non-skiers, than by older public that already knew how to ski [Gourville, 2005]. Or when Kodak cameras were introduced in 1888, they were targeted not at professional photographers or even sophisticated amateurs but at non-users[Utterback, 1994]. In this sense, absolute novelty can bare better response than a familiar product with novel char-acteristics that gives way to inertia. This is almost the opposite to optimal distinctness, as it assumes that inertia is lower in complete novelty, and is higher when habits are formed. The answer for this incongruity goes back to the reference point: is it absent when the product category is so new that one does not view it as substitution? This poses questions for future research of how inertia and novelty seeking interact.
2.2.5 Risk, Uncertainty, Novelty
We have looked at decisions under novelty and their relationship to inertia and the loss aversion bias. There are however more behavioural biases and heuristics involved in the process. First of all, in decisions under novelty a risk preference is involved, as you are risking not liking the novel product or experience. Wang et al. [2015] found that high novelty seekers also tend to have higher risk preferences. A step up from displaying the risk, the choice under novelty also deals with uncertainty, because the extent of risk is unknown. Thus, resulting in higher degrees of bounded rationality. Choice under novelty also has a new environment with missing rules for choice, which results in even more bounded rationality and requires one to use imagination, explore and adopt new frameworks [Potts, 2010]. According to Kay [1979] decisions under novelty also require a decision whether to commit to a new path at all, or whether to take part in the experimental learning. This implies that one also has to decide whether to make the choice at all.
2.2.6 Insiders versus Outsiders
Dealing with novelty is such a complex process that there are plenty of sources for irra-tionality and consequently innovation failure. One of those stems from kin selection in evolutionary biology, whereby humans instinctively divide themselves into insiders and outsiders [Bergstrom, 1996] . We tend to favour people within our group, rather than the outside. This potentially helps us to take care of our offsprings and exchange favours within the group to make sure if anything were to happen to us, offsprings would be taken care of by our peers. This sentiment also applies to evaluation of new ideas. Novel ideas that come within the group (be it country specific, or organization) are valued more than novel ideas coming from outside the group [Bergstrom, 1996] . This discounting of exogenous novelty can be attributed to the endowment effect, whereby things that are already owned are valued more than those that are not [Kahneman et al., 1990]. This irrational closedness to innovation leads to lower novelty seeking, and slower, less efficient adoption of innovation.
2.3 Inertia experiments
One of the objectives of this study is to measure individuals’ level of inertia. Hereby, we will overview different experimental designs that aim at identifying inertia.
2.3.1 Ball lottery
There are several ways used in the literature for the testing of inertia, one of which is a lottery. It has been performed by [Charness & Levin, 2005; Alos-Ferrer et al., 2016; Sautua, 2016]. These experiments have their differences, but the core of the design is the same. In the [Sautua, 2016] experiment participants were presented with a bag that had 10 bi-coloured balls inside (red and blue). Participants knew that there would be 8 balls of one colour and 2 balls of another, but they did not know which colour had the majority. In other words they did not know which state of the world they were in. Participants were also randomly assigned (with a 50% chance) a Blue or a Red lottery ticket 2. In the experiment there was a total of two draws of the balls, and if the second
draw corresponded with the colour of the ticket, then the participant would win $10. If not, they would not win anything. The experiment had several stages. In stage one a ball was drawn out of the bag and then replaced with the same colour. This was meant to give participants an idea in which state they were in. Afterwards subjects had to fill in a personality test. After the test they were offered $0.10 to swap their lottery ticket to the opposite colour3. Lastly the second ball was drawn that determined the payoff.
Based on the Bayes’ rule, the optimizer’s strategy is to choose the ticket with the highest expected payoff, given the updated probability of the state of the world. Hence, the rational player should choose to keep the ticket if the first draw’s colour corresponded to the ticket (since there was higher chance that drawn ball belonged to majority), and to swap the ticket if it did not correspond with the colour of the first draw. In the scenario of the draw being different colour than the ticket, additional payment for swapping increased the rationale to swap even more. Decision-making under the influence of inertia however,
2In some treatment groups subjects drew a ball at random out of a bag with one red and one blue balls
and wrote its colour as their lottery ticket, to ensure the trust of participants that the ticket distribution was not rigged.
3In some treatment groups subjects knew that they will get a chance to swap their tickets. This was
done to see whether endowment hypothesis (having a preference towards original ticket) or expectation hypothesis (with the possibility of switch anticipated, the final ticket choice being the reference point) is .
would result in people staying with the same lottery ticket no matter what the colour of the first ball was. The study found that 70% of participants displayed inertia by keeping the original ticket.
[Charness & Levin, 2003; Alos-Ferrer et al., 2016] had a slightly different design. Their experiments were done on computers, included two urns (instead of a bag) and had multiple (60) choices made. The two urns contained 6 bi-coloured balls. There were also two world states with a 50/50 chance of each state and every two draws the states of the world were randomized4. The composition in the urns and between states also
differed between the experiments. In some cases, Bayes’ rule prescribed the ’win-stay lose-shift’ strategy similar to the one in [Sautua, 2016]. Some matrices entailed ’win-shift lose-stay’ strategies. This helped researches see the difference between positive and negative reinforcement. As intuitively expected, error rates (not getting the right ball) significantly increased, as inertia drives people to repeat once successful choice, which would be a logical fallacy in the ’win-shift lose-stay’ Bayesian updating strategy.
Similarly to our experiment, these studies also included questionnaires. Alos-Ferrer et al. [2016] had its participants fill the Preference for Consistency questionnaire from [Cialdini et al, 1995]. Their results showed that a higher score in Preference for Consistency was associated with more errors in the presence of inertia. Sautua [2016] also integrated a personality questionnaire ’Big Five’ [Oliver et al., 1991] that looked at traits such as open-ness/closedness to experience, excitement-seeking, compliance and more. Unfortunately, Sautua did not publish the results of the questionnaire and possible correlations between personality traits and levels of inertia. The test was used mainly as a distraction for the participants.
2.3.2 Questionnaires
There are however questionnaires, both that occur naturally and are designed by experi-menters, that are targeted at researching the inertia and the status quo bias.
4Participants were paid (e0.18 in [Alos-Ferrer et al., 2016] and around $0.30 in [Charness & Levin,
2.3.2.1 Natural experiments
There are two types of questionnaires that measure inertia: natural experiments and lab-oratory experiments. Governments, public agencies and others have been running natural experiments. In this research experimental and control conditions were determined natu-rally, via conjuncture of circumstances. The most famous status quo natural experiment was analyzed by [Johnson & Goldstein, 2003 ; Gimbel et al.,2003] on organ donation. They looked at European countries and divided them into two groups based on their default policy of organ donations. Countries such as Netherlands and Germany had an explicit-consent policy, meaning that people were presumed to not be organ donors, unless they registered to be one. Whereas countries like France and Portugal had a presumed-consent policy, meaning that all citizens were donors, unless registered out of it. If the decisions were made in accordance with classical economics theory, the default parameter would have little effect on the behaviour. However, the research found that5having
dona-tion set as a default increased donadona-tions by more than 16%. Gimbel et al. [2003] included more Eastern European countries in his analysis and found that presumed-consent default option increased donations by 56.5%.
Possible explanations include the unwillingness to put the effort to fill in the form. Con-trary to this explanation, an example with the Netherlands6 (explicit-consent country)
showed that efforts to facilitate this process had little effects [Oz et al., 2003]. Another explanation (among the aforementioned loss aversion, reference point factors and so on) is trying to avoid the unpleasantness and stress related with thinking about organ donations [Johnson & Goldstein, 2003]- the drive to not think too much. Other examples of natural experiments include insurance policy between Pennsylvania and New Jersey in the early 90s [Johnson et al., 1993]. The two states at approximately the same time introduced a law that included the option for motorists to have a reduced right to sue (if not seriously injured) that also implied lower interest rates. However, in Pennsylvania the default was the full right to sue at a higher price, and in New Jersey the default was the reduced right to sue at a lower price. The options were the same, but the framing differed. The data showed that 75% of Pennsylvanians kept their full right to sue, whilst only 20% of New Jerseyans opted to acquire the full right to sue. This illustrates a substantial bias towards
5[Johnson & Goldstein, 2003] study was conducted for years 1991-2001. It ran multiple regression and
time-series analysis of number of corpse donations made per million, with default as a predictor variables. The study controlled for countries’ difference in transplant infrastructure, propensity to donate, religion, education level and a time variable.
6In 1988 there was a creation of national donor registry with extensive campaigns in education and
the status quo option. Similarly, Madrian and Shea [2001] found a dramatic change in the savings behaviour of employees of a large U.S. corporation7 after the default option
was changed from electing participation in the savings plan to automatic enrollment upon hire.
2.3.2.2 Vignette questionnaires
So far we have looked at natural opt-in/opt-out experiments that looked at inertia and the status quo bias. The questions had people make actual decisions about their lives. And even though it provides the external validity, there are some limitations to it. Not being able to test other relevant factors like novelty seeking or test individuals’ strength of the inertia bias is the key disadvantage for our purposes.
Another angle of looking at inertia is to conduct an experiment in a controlled environment by giving people hypothetical vignettes with multiple choice options, one framed as a status quo. This method allows for measuring inertia across several scenarios. Whilst it can provide data to show whether inertia exists for a certain case across population, it can also rank people through their multiple answers on a scale of susceptibility to being swayed towards status quo options. Schweitzer [1994] conducted a case study experiment looking at the status quo bias -the preference for the current state of affairs, and the omission bias- the preference for inaction. Both these biases fall under the umbrella of inertia, but it is hard to distinguish between the two- most of the times inaction leads to remaining in the current state. Schweitzer designed a questionnaire with hypothetical scenarios covering the following topics: choosing a car, a health care insurance, care for infection and deciding on the damage control at a dam. The differences between the status quo, omission and neutral treatments were in the framing of the last paragraph. In the example of the health care insurance plan, this was the neutral treatment:
This was added for the status quo bias treatment:
In this framing no action is framed in such a way that in order to remain in the current state you have to act. The following was also added for the omission bias treatment:
In this framing no action is needed to remain in option A, but the default option is not strongly associated with habituation.
Subjects then were presented with four alternatives, which they had to rank on the scale from 1 to 7 of how likely they were to choose each option.
The study found statistical significance of both the omission and the status quo effects. This also suggests that the combination of these factors would have a higher effect on behaviour.
Despite being significant, the effects however were not very strong, which can be explained by design issues. Each participant faced all 3 treatments consecutively, with the neutral scenario appearing first. That could have possibly anchored participants towards the option they had chosen first. The effects could have been stronger if subjects had in each scenario only one treatment, for instance, just the status quo in the health care insurance case. And of course the effects would have been stronger if those were not hypothetical, but actual choices.
Samuelson and Zeckhauser [1988] composed a case study questionnaire to look at the effects of the status quo bias (they did not separate status quo from omission bias). They had a sample of 486 participants and found evidence of the status quo effects. The questionnaire contains two parts: the first part has a lot of similarities with [Schweitzer, 1994]; the second part is a sequential decision task, and the first choice the participant makes becomes a status quo and guides the next question.
The first part is a hypothetical scenario whereby subjects have to pick between four alternatives. They were either all neutrally framed, or one had the position of status quo. The options were varied between treatment groups and, unlike in [Schweitzer, 1994], each participant saw only one version of the question. The case studies were hypothetical and participants had to imagine themselves in the position of ministers, governors, inheritors and professors. The results from this part showed that people choose status quo options at a higher rate, than alternative options to status quo8, and that status quo options were
also preferred to the neutral options in the neutral treatment.
Part two of the questionnaire in [Samuelson and Zeckhauser, 1988] was designed to test whether status quo bias is still present in scenarios when the status quo option was not predetermined by an outside force, but when subjects chose it themselves. The scenario was as follows- as a top manager of a regional airline, the participant had to decide on the type and number of aircrafts for the coming year (there was a varying amount of alternatives). The decision was made twice: once with a good forecast for demand and air fares, and once with a bad forecast. The order of forecasts was changed between treatments. Based on the decision a participant makes in year one, he is handed a sheet where the said decision is framed as status quo. So for instance if for the good year a big fleet is chosen, in the following bad year it is more rational to swap for a small fleet. If the change has not occurred, then it is evidence of inertia’s presence. The results showed some evidence in the status quo bias, which was more prominent in cases with more alternatives9.
8The null hypothesis of status quo option being indistinguishable from its alternatives was rejected at
10% significance level in 31 out of 54 cases.
9The weak presence of the status quo bias means that using this part to test individual’s tendency
2.3.2.3 Choice experiments to elicit preferences in consumption bundles
Another angle to study the status quo bias is through the observation of human behaviour over different proportions of consumption bundles with and without options framed as endowments (the status quo option).
Masatlioglu and Uler [2013] ran a computer choice experiment that elicited participants’ preferences over two or three alternatives. Subjects had 16 questions that had them pick between different consumption bundles of money and Belgian chocolates. At the end one of the choices was randomly picked and given as payment.
Hereby see the screenshot of the experiment as an example:
The experiment had several parts, the answers in the first part determined the questions in the consecutive parts. In the first part participants had to choose between two consump-tion bundles. The following parts included reference point opconsump-tions that were dominated by at least one of the alternatives.
One of the goals of the experiment was to disentangle the status quo effect (which was framed as an endowment) and the decoy effect. Decoy effect occurs when ”the binary comparison of two objects, contrary to normative theory, is affected by the introduction of an asymmetrically dominated alternative to an existing choice set” [Masatlioglu & Uler, 2013]. This entails that introduction of an inferior alternative (clearly inferior towards one of the options) affects the preference between two options through comparison and reference effect. Experimenters wanted to see whether framing the third alternative (the inferior reference point) as an endowment would be statistically significant. The results found that when a decoy option was introduced 18% switched to the decoy option, whereas
in the case of status quo, 29% switched to the status quo option. And even though the endowment was very weak (since subjects did not actually own the status quo option), the effect was still visible.
2.3.3 Product Choices
An experiment done by Tversky and Schafir [1992]10 has also looked at the inclination
towards the default option. This design has served as an inspiration for our experiment and we will also discuss it later on. Subjects were asked to fill a questionnaire and were paid $1.50. After the questionnaire11, subjects in one group were given the opportunity
to exchange their payment for a prize of a metal Zebra pen. Subjects in the other group were offered to exchange money for either a Zebra pen or a pair of plastic Pilot pens. Participants were informed that the prices of each prize were around $2. After the decision was made, participants got whichever payment they chose: the default option was money and the alternatives were pens. The study found that when only one alternative was available (Zebra pen), 75% of participants chose to exchange money for the pen. However, when both Zebra pen and Pilot pens were available, most people decided to stick with the default option- $1.50. Authors concluded that the study showed that conflict (having more options) increased the tendency towards the status quo.
The aforementioned designs of testing inertia and status quo have certain benefits and limitations12. However, for the purpose of our study, we have used the [Samuelson &
Zeckhauser, 1988] questionnaire as a personal scale of inertia, and we have used the experimental set up by Tversky and Schafir [1992] as an inspiration for our main treatment. There are some crucial differences in our design, that we will discuss in the next chapter.
2.4 Novelty seeking tests
Another objective of this study is to measure individuals’ level of novelty. Hereby, we will overview different questionnaires that tackled this issue.
There are many examples and tests that aim to measure individual novelty seeking:
Nov-10An almost identical study was also done by [Simonson & Tversky, 1992]
11The authors do not talk about the questionnaire, so it is assumed it was solely served as a distraction.
elty Seeking scale in the Tri-Dimensional Personality Questionnaire [Cloninger, 1987], the Novelty Seeking Scale [Hirschman, 1984], the Specific Loss of Interest and Pleasure Scale [Salem et al., 2017], the Arousal Seeking Tendency Scale [Mehbrain & Russell, 1974], the Exploration and Stimulation Seeking Scale [Driver, 1977], the Desire for Novelty Scale [Pearson, 1970], Leisure and Boredom Scale [Iso-Ahola & Weissinger, 1990], the Tourist Seeking Scale [Lee & Crompton, 1992] and the Sensation Seeking Scale [Zuckerman, 1984, 1996].
We will overview some of the tests that are used prominantely in the literature across fields in economics, marketing, sociology, psychology and medical science.
2.4.1 Tri-dimensional Personality Questionnaire
Tri-dimensional Personality Questionnaire was developed by Cloninger [1987] as a part of psychobiological theory. It has three temperamental subscales: the Novelty Seeking Scale, Harm Avoidance and Reward Dependence. It later had another trait added (Persistence) and was renamed as Temperament and Character Inventory [Cloninger et al.1993]. The Novelty Seeking Scale has 34 true or false questions, with higher score representing higher Novelty Seeking. According to Cloninger13 individuals that score high on the novelty
seeking scale tend to be ’quick-tempered, curious, easily bored, impulsive, extravagant, and disorderly’14. This ties novelty seeking to many other personal qualities, all of which
however carry negative connotations. For our purposes, harm avoidance and reward dependence are not of interest to our study.
2.4.2 Novelty Seeking Scale
Hirschman [1984] composed a questionnaire that looked at four characteristics. One of them was Novelty Seeking Scale which focused on individual’s propensity to try new products, be involved with many organizations and vary the media consumed as well as ’geographical venturesomeness’. The questionnaire asked subjects to rate on a scale to seven how willing they are to try new things in the following categories: dances,
13http://psychobiology.wustl.edu/what-does-the-tci-measure accessed 03.06.17
14Examples could not be illustrated as neither TCI or TDPQ were accessible online. They required
substantial payment and approval process for access. The questionnaires are still very popular among researchers [Wang et al., 2015; Fruyt et al., 1999]
places to shop, apparel, home furnishings, movies, books, magazines, vacation spots, foods, restaurants, political ideas, religious ideas, types of transportation, hairstyles, sport activities. Hirschman [1984] then ran the answers against a demographical questionnaire. She found that high novelty seekers tend to be firstborn or the only child, are not socially isolated and see their values in correspondence to societal trends. The limitation of such format of this scale is that placing preference for certain product classes does not address the psychological constructs of novelty. This means that its applicability is questionable. Lee and Crompton [1992] found that this scale is not related to novelty in a tourism context (even though one of the questions it features is on vacation spots). More abstract and personal questions would work better in assessing strength of one’s personality trait.
2.4.3 Sensation Seeking Scale
Sensation Seeking Scale has been developed by Zuckerman [1984, 1996] and is used as a questionnaire in our study. According to him, sensation seeking is ”a trait defined by the seeking of varied, novel, complex and intense sensations and experiences, and the willingness to take physical, social, legal and financial risks for the sake of such experi-ence”. It has several versions, but the most popular (V) is divided into four subscales: thrill and adventure seeking, boredom susceptibility, experience seeking and disinhibi-tion. The test proved to be reliable across cultures (in Israel, China, France, UK and Canada) [Birenbaum, 1986; Wang et al., 2000; Carton et al., 1992; Rowland&Franken, 1986; Gray&Wilson, 2007]. The test demands participants to pick out of two polarized statements to which they relate the most- one exhibiting tendency towards high sensation seeking and another towards low.
Here is the example of such question:
Fife-Schaw [2000] has criticised this format as it does not provide an indication of the extent to which participants agree or disagree with the items. However it does give a clear-cut data of binary choices that are easy to analyze and allow for simple comparisons between individuals.
Some of the questions on the scale are culturally outdated, especially the ones that talk about ’hippies’ and ’queer’. In general the age of the questionnaire raises a question if it is an appropriate format to use, despite its popularity in research. To answer this question Gray and Wilson [2007] conducted a UK based experiment using the Sensation Seeking Scale and run it through analytical methods of Facet Theory- which is used to research questionnaires that have items included because of theoretically based arguments [Donald, 1994]. The questionnaire was changed into a 7 point Likert-type scale, which helped address the issue of subjects being able to report how much they agree or disagree with items. This in turn has helped identify questions which did not differentiate between people’s responses and were unreliable. Gray and Wilson [2007] have revised the Sensation Seeking scale using empirical data. Unreliable questions, that also happened to talk about homosexuality, hallucinogenic drugs and swingers, were excluded. The up-to-date version is a reliable questionnaire to measure sensation and novelty seeking, which we will incorporate in our study.
2.4.4 Tourism
Novelty Seeking tests have been widely used in the marketing research of tourism [Lee&Cromp-ton, 1992; Hernández-Lobato et al., 2006; Assaker, Vinzi & O’Connor, 2011; Assaker&Hal-lak, 2013]. [Albaity & Melhem, 2017] paper stands out for the purpose of our study as it looks at the relationship between novelty seeking, loyalty (that can be otherwise seen as status quo) and the destination image. They have collected questionnaires from 508 tourists visiting United Arab Emirates. The questionnaire contained five parts: three factors mentioned earlier, tourist satisfaction and demographic questions. Participants had to rank each item on a 5 point Likert scale from strongly disagree to strongly agree. The novelty seeking part had the following statements:
• This destination offers an unusual experience. • This destination offers new discoveries.
• This destination offers new experiences. • This destination is new for me.
• I intend to visit UAE in the future.
• UAE would be my first choice as a vacation in the region.
• UAE provides more benefits than other destinations in the region. • I would advise/recommend other people to visit the UAE destination.
The study found a strong relationship between novelty seeking and destination loyalty. It showed that when tourist’s satisfaction increases due to novelty and destination image, her propensity to return and recommend the destination increases- an intuitive finding. However, if we see destination loyalty as status quo or default option this poses certain problems to our theory. We have assumed that novelty seeking and inertia are opposing factors: the more you seek novelty, the less inert you have to be. In most cases seeking novelty implies taking action (reducing omission bias) and purposefully shifting from the original state to a new state (reducing status quo bias). In the case of tourism however, we see the opposite trend- the more novelty a destination offers, the more likely one is to return.
There are several possible explanations to such discrepancy, one being that this particular test potentially carries a misleading name of novelty seeking scale. It does not ask if a person prefers or seeks out new experiences, but merely if he has experienced them on the vacation. Even if there is a correlation between novelty experienced and satisfaction, it does not necessarily imply that novelty was actively seeked out, or if a participant consciously looks out for novelty in his choice of vacations. For instance, a person may have experienced new things, but the motivation to return will be tied to satisfaction rather than novelty. Secondly, the study itself pointed out that the longer time a tourist spends at a destination which she considers novel, the less likely she is to return [Albaity & Melhem, 2017]. Hence, the possible explanation is that upon arrival to a new environment, a novelty seeker spots new potential activities that he does not have time to fulfill within his stay. Thirdly, the answer might lie in the reference point. With vacation destinations, the reference point might be home rather than the previous destination. Since tourism is a nondurable good, the habituation level falls during the time spent back at home. Finally, we can question the validity of these findings, due to the existing contradict-ing research- [Niininen et al., 2004] found that tourists that have high novelty-seekcontradict-ing propensity are less likely to return to the same destination and have more varied touristic destinations. They used Arousal Seeking Tendency Scale questionnaire
[Mehbrain&Rus-sell, 1973] on 123 UK residents. The sample is not big enough to make conclusive remarks, but it does put into question the reliability of the [Albaity & Melhem, 2017]’s results.
3
Experimental design
The study had 63 subjects (21 per treatment) in their twenties, students of the Univer-sity of Amsterdam. Subjects were approached individually in the univerUniver-sity cafeteria and were asked to fill in the questionnaire. The experiment lasted for approximately 15 min-utes per person. The experiment consisted of four parts: status quo questionnaire, pen change treatment, novelty seeking questionnaire and demographic questionnaire. Novelty seeking and status quo questionnaires were taken from previous literature [Samuelson & Zeckhauser, 1988; Zuckerman 1984, 1996; Gray & Wilson, 2007]. The pen change was divided into 3 groups: offer to swap an ordinary pen for a feather pen (high novelty, high status quo); offer to swap an ordinary pen for a good quality pen (low novelty, high status quo); ordinary pen being taken away and an offer presented between a feather quill and a good quality pen (novelty to be determined, low status quo). Pens were collected at the end of the experiment. After all subjects had participated in the experiment, one was randomly selected for a 50 euro Albert Heijn gift card.
3.1 Protocol
For this experiment people who were sitting alone in the university cafeteria were ap-proached. This was done to reduce the effect of peer judgment on their choices. The opening question was: ”Would you like to take part in a questionnaire? It takes around 15 minutes and you get a chance to win a e50 Albert Heijn gift card.” If the subject agreed, he or she was handed an ordinary pen and the first part of the study, which entailed general instructions1 and the status quo questionnaire. Subjects were
identi-fied in experimenter’s notes with serial numbers and were randomly assigned into three treatments that determined the pen offer and the wording at the end of the first part of questionnaire. The experimenter took a seat nearby (usually one seat away, to give participant his/her space and privacy but also be easily available after they finish with the first part). As the participant started reading and filling in the questionnaire, the experimenter discreetly prepared the pen appropriate for the treatment and the second part which entailed the novelty seeking questionnaire.
In the treatments 1 and 2 at the end of the first part it read:
”This is the end of Part 1. Please give the handout to the experimenter”
After the participant passed the first handout, he/she was offered a handout with the second part. Next, another pen was shown by the experimenter that was held in a half-bent arm. The experimenter said: ”This was part one. Here is part two and would you like to fill it in with this pen [pointing at the good quality pen if in treatment 2, or a feather and a bottle of ink if in treatment 1], or the one you’ve got?” The participant would vocally state his/her preference. If the ordinary pen was chosen, then the feather quill (in Treatment 1) or the good quality pen (in Treatment 2) was put back into the folder. If the newly offered pen was chose, then the experimenter would take away the ordinary pen, and hand over the new pen. In case of a feather and a bottle of ink, experimenter sometimes helped open the bottle and explained that the subject just had to lightly dip the feather in (sometimes tissues were offered in case of ink blobs).
In the treatment 3, the end of first part read: ”This is the end of Part 1. Please give the handout and the pen to the experimenter.”
1For the handouts, see Appendix1. The wording of the instructions was taken from the original
After the participant passed the handout and the ordinary pen, he/she was offered a handout with the second part. Simultaneously, the experimenter showed two pens a good quality pen and a feather with a bottle of ink. The experimenter said: ”This was part one. Here is part two and would you like to fill it in with this pen [pointing at the feather with ink] or this pen [pointing at good quality pen]?” After the subject stated their preference, their pen of choice was given to them and the other pen was put back in a folder.
As the participant was marking their answers in part 2, the experimenter would discretely put a note on the personal phone, which pen the subject chose.
After the participant finished part 2 and handed it in, he/she was given a sheet with demographic questions (on age, gender and major subject). It was a separate handout, so that participants’ personality test answers were not affected by them seeing further demographic questions (even though everything was anonymous). After all three handouts were collected, the participant was offered a sheet with a list of emails, where they could put their email to participate in the lottery for the e50 Albert Heijn gift card2 3.
3.2 Status Quo Questionnaire
The Status Quo questionnaire was taken from the paper [Samuelson & Zeckhauser, 1988] on the behavioural experimental economics. There are several reasons for such choice. First of all, given the constraints, the goal of the experiment was to be short, concise and neat. Secondly, it was crucial to have it being completed with pen and paper, as the pen use was the main part of the treatment. Therefore a computer based experiment had to be excluded. This also ties together with the third requirement- the need to exclude studies that used sequential decisions- where previous choice determines the next question [Herne, 1988; Masatlioglu and Uler 2013; Dean, 2008]. Lastly, the purpose of this part was to put participants on a scale of inertia and compare them between each other. Most studies were designed to check for the existence of inertia in the population, instead of looking at each case individually. The former requires having a control group
2The lottery for the e50 Albert Heijn gift card was used as a payment for the participation in the
experiment. The reason for making one payment of e50 through lottery, instead of e0.80 to each participant, was made with behavioural heuristics and biases in mind. People are optimistic and they overestimate small probabilities [Reyna & Brainerd, 2008] The gift card was unlikely to have produced any feelings of gratitude that could have influenced the behaviour, since the chances of actually winning a lottery were small.
3After all the data was collected, the lottery winner was picked randomly using the random number
unexposed to inertia allowing for treatments. Our study however assumed the presence of inertia (based on the evidence from literature), and aimed to measure the strength of inertia’s presence in each participant, and how it affected their choices in the pen treatment. Combining it with the fact that our experiment did not have a big enough sample to exclude half of the participants, we arrived at a necessary condition of having only inertia treatment in the first part of the experiment. Hence, all studies that had random assignment of inertia and control treatments had to be eliminated. An example would be the ball lottery [Charness&Levin, 2003; Alos-Ferrer et al., 2016; Sautua, 2016], as there would be participants whose lottery ticket colour matched the colour of the first ball picked- the best predictor of the following ball and therefore the victory, meaning that it was a rational choice to stick with the ticket and not strictly inert.
[Samuelson & Zeckhauser, 1988] questionnaire from the experimental behavioural field, mostly fulfilled our requirements, as it was designed for a pen and paper execution; we could only use the status quo treatment; it also had multiple vignette questions, which allowed us to differentiate people that were very inert (that preferred mostly the status quo options), from people that displayed milder inertia bias (with mixed answers of status quo and other options).
Nevertheless, there were certain problematic aspects of the questionnaire, that had to be addressed. It consisted of two parts, the second one involving sequential decisions that, as discussed earlier, were troublesome for our setup. For this reason and to shorten the questionnaire, we included only the first part of the questionnaire. Secondly the question-naire was outdated and aimed at the American audience and required modifications from our part to fit the time and geographical environment of our subjects. For example, we replaced: ”For your car, you require air conditioning and a stereo radio with rear speak-ers.” with ”For your car, you require navigation and cruise control.” We also replaced ”the U.S. Armed Forces” with ”the Royal Netherlands Army”, amongst other examples. The questionnaire ended up consisting of 5 vignettes each with 4 ambiguous choices4 one
of which was being framed as status quo- an option to keep things in their original state. Hereby is an example of the questionnaire:
3. On behalf of your company, you are in charge of preparing a price bid to supply a fixed quantity of mattresses to the Royal Netherlands Army. The Army will select
4With a bigger sample we would also vary the alternative that is framed as status quo between subjects
to control for possible general preferences of one alternative over the other; or have a control group that did not have any alternatives framed as status quo, and compare the status quo treatment to that.
the lowest of the sealed price bids submitted. Your company’s cost of fulfilling the contract (should it win it) is estimated to be e100,000. In the past, a common practice of your firm in bidding for contracts of this type is to apply a 15% markup to cost in setting the bid. In this case, although you suspect your company may have lower costs, you are aware of a number of competitors who are eager to obtain the contract. Your estimate is that a bid of e115,000 has a 70% chance of winning the contract. You are also considering other bids. Your choices are:
• a) Bid e115,000. Your chances of winning the contract are 70%. • b) Bid e125,000. Your chances of winning the contract are 50%. • c) Bid e120,000. Your chances of winning the contract are 60%. • d) Bid e130,000. Your chances of winning the contract are 40%.
3.3 Novelty Seeking Questionnaire
The Novelty Seeking questionnaire was taken from the Sensation Seeking Scale [Zucker-man 1984, 96]. It has been successfully used in economics research [Eckel & Gross[Zucker-man, 2002] and in marketing research [Lee & Crompton, 1992; Gilchrist et al., 1995; Lepp & Gibson, 2008; Pizam et al., 2004].
Gray and Wilson [2007] revised the questionnaire using empirical data and deleted un-reliable and outdated questions. In our study we also had to exclude questions on sex and drugs because they would be inappropriate in the setting of approaching individual students outside confined laboratory environment. Additionally, for the sake of shortening the time of the experiment (both to have a higher response rate and have more partic-ipants) repetitive questions were also excluded. For instance we removed the question about skiing but left in a question about scuba diving. There were in total 15 items left, each consisting of two statements: one expressing novelty seeking and the other expressing reservations for novelty. The subjects were asked to pick a statement to decide on which statement they gravitate towards the best. Here is an example from the questionnaire:
• 1. A. I am not interested in experience for its own sake
a little frightening, unconventional, or illegal
Such format of the questions provided binary results, which were simple to analyse.
3.4 Pen Change treatments
Between the two questionnaires there is a pen change designed to reveal tendencies towards Status Quo or Novelty Seeking. Its purpose was to demonstrate how subjects exhibit Novelty Seeking and Status Quo biases in an actual choice setting and how it compares to their relative positioning on the scales from the questionnaires. It was inspired by the status quo experiment in [Tversky & Schafir, 1992] where subjects were asked to fill a questionnaire5 and were paid $1.50 (a default option). After the questionnaire, subjects
in one group were given the opportunity to exchange their monetary for a prize of an elegant metal Zebra pen; subjects in the other group were offered to exchange money for either a Zebra pen or a pair of plastic Pilot pens. Participants were informed that the prices of each prize were around $2. After the decision was made, participants got whichever payment they chose: the default option was money and the alternatives were pens.
In our design, the pen change stage was different between 3 treatments that had varying intrinsic strength of novelty and status quo framing6. In Treatment 1 the choice set was
high in novelty and high in status quo; in Treatment 2 the choice was low in novelty and high in status quo; in Treatment 3 the choice was low in status quo and the degree of novelty was determined by the subject’s decision.
At the beginning of the experiment participants were given an ordinary pen to fill the first questionnaire. After the first questionnaire, subjects in the Treatment 1 group were offered to swap their pen for a feather pen. This treatment featured both strong status quo factor (as subjects had a choice to hold onto their original pen) and novelty seeking (as subjects also had a choice to swap for a feather quill -a more intriguing, strange-looking option that they probably had not used before).
In Treatment 2 subjects were offered to swap their pen only for a better quality pen. It was chosen, so that it would be a superior option to the ordinary pen, but it was much
5The authors do not talk about the questionnaire, so it is assumed it was solely served as a distraction.
lower in novelty compared to the feather quill. The strength of status quo framing would be close to the 1st treatment (as participants also had an option to stick with the first pen), but the novelty factor was much weaker (as it was still a normal pen that subjects had used many times in the past and it did not differ much from the original pen). In Treatment 3 the original pen was collected by the experimenter (thus decreasing the status quo factor as subjects no longer could hold onto their original pens), and the participants were given a choice to write either with a good quality pen or a feather quill (hence increasing the Novelty factor as subjects would face a choice between something new and bizarre and something commonly used and seen).
The differences between our experiment and [Tversky & Schafir, 1992] are that in our case the default option was a pen as opposed to money (which made alternatives more comparable). We also took the default option in Treatment 3 away and it was not part of the choice set. The goal was to reduce status quo bias in the 3rd treatment and focus more on the novelty effect. Furthermore, in our experiment participants did not get to keep their prize but could merely use it within the experiment (for two reasons: the pens had different monetary value; as there was not enough budget available to give out pens given that the ”good quality” type of pens used costed 20 euros).
3.5 The order
The order of the experiment was designed in such a way, that the pen change was in the middle, so that participants could get used to the initial pen and see it as a status quo after the first questionnaire. The second questionnaire provided a reason to use a pen after the pen decision was made. The novelty questionnaire had to be used last, because it is clear from its questions that novelty preference is being tested. We did not want this knowledge to have affected the choice of pen. The status quo questionnaire was much more ambiguous, and only people familiar with the status quo bias would be able to figure out the nature of testing (which could still be possible in some cases, since our subjects were students). There was a concern that writing with a feather quill or a pen would affect the responses in the novelty questionnaire. With the larger sample we would swap the order of questionnaires to control for possible effects.
3.6 Research Question
The experiment was designed to show how the relative positioning on the Novelty Seeking scale and the Inertia scale would reflect in making an actual choice, in which either both factors are strongly present or either of them. Hence we would be able to see how these two forces interact in people that report different levels of status quo bias and novelty seeking, and under different choice sets where novelty or status quo framing were manipulated to be of different degrees.
3.7 The Hypothesis
Hereby, you can see the hypothesis of our study:
H1: people that score high on the novelty seeking questionnaire are more likely to choose
a more novel pen option: in Treatment 1 choose a feather over an ordinary pen; in Treatment 2 choose a good quality pen over an ordinary pen; in Treatment 3 choose a feather over a good quality pen.
H2: people that score high on the inertia questionnaire are more likely to choose a more
status quo option: in Treatment 1 choose an ordinary pen over a feather ; in Treatment 2 choose an ordinary pen over a good quality pen; in Treatment 3 choose a good quality pen over a feather.
Additional exploratory data: which of the effects of novelty or inertia has a stronger effect on the decision of the pen choice.
4
Results
4.1 Correlation
First of the correlation levels of inertia and novelty seeking scores from the questionnaires was calculated. The assumption of the study was that inertia and novelty seeking are polarizing forces, swaying decisions into opposite directions. Hence one would expect a large correlation value. However, after analysing our experiment we have found that the scores from novelty and status quo questionnaires were not correlated with a correlation value of (-0.03785). This was not expected, as it seems to point to the direction that people can have unrelated strengths of these two drives. For instance high novelty seekers might also be highly reserved to make a leap and go for something new.
4.2 Regression analysis
Moving on to the analysis of the treatments the analysis binary qualifications were used. In the Novelty Seeking questionnaire ’1’ stood for an answer that represented high novelty
