Counteracting Availability: Giving Control Back to Players Within
Online Gambling
Simon Nordsvan
Department of Informatics Master thesis, 30 hp
Counteracting Availability: Giving Control Back to Players Within Online Gambling
Abstract
Gambling has become ever more available and begun to employ more techniques, such as gamification, in order to engage their players. Today, Gambling with the use of your computer or smartphone is not deemed very different from playing any other kind of game. This shortens the availability cascade, or chain of events, that in the context of Gambling can ultimately lead to severe economical or social consequences for the players. This thesis seeks to find out how the increased availability can be counteracted by giving control back to the players through the use of personal informatics as well as if, and if so how, the social aspect of Gambling affects responsibility and the perceived experience. By conducting in-depth interviews with 20 participants I gained insights into what would assist them in maintaining control as well as their perception of Gambling. In order to evaluate these findings they were, in conjunction with related research, translated into a design concept. The results of the evaluations of this design suggest that many of the same techniques used to increase engagement with Gambling can also be utilized to counteract the increased availability. The prospects for future research are outlined.
Keywords: Behavioral change, HCI, personal informatics, human-computer interaction, availability, tracking, control, social responsibility.
1. Introduction
Gambling has become more accessible. You are no longer required to physically visit the race track or casino to partake in these activities, but can now rather do so from anywhere directly through your smartphone or computer, not to mention how incentivised we are to start playing through commercials. These changes to the availability cascade (Kahneman 2011) make it ever easier to get hooked. Being able to Gamble through technology also brings about changes in the social dynamics of these activities. Hancock (2017) argues that the harm of Electronic Gambling Machines (EGM) is to be treated equally to tobacco and alcohol in terms of regulation and rehabilitation. With the introduction of EGM’s to the Gambling market the Social Corporate Responsibility has been deteriorating (Chen 2020; Hancock 2017) with the Gambling companies becoming more distant to the players utilizing their services. With EGM’s this distance not only applies to the relation between player and company, but between players as well. The participatory commitment (Clark, 2006) between players engaged in traditional Gambling is nullified by the increased digitization of Gambling but has yet to be elaborated upon in any research.
Behavioral Change and Self Determination Theory seek to provide an understanding of what motivates people to partake in activities and what alters their way of thinking (Fogg 2009a;
Cherry 2019; Mekler 2020; Fu et al. 2018). The guidelines provided by these theories are implemented through means such as gamification (Benyon 2013; Greenwald 2014; Hyrynsalmi et al. (2017). While Gambling companies often provide the necessary information for their users to reflect on their habits with Gambling, they seldom incentivise or motivate their users to change their behaviors.
Statistical information tracked and presented in a journal-like format commonly have their roots in personal informatics as part of the field of Human-Computer Interaction (HCI) (He et al. 2013; Epstein 2015). Personal informatics, as well as all of the other above mentioned topics, are double edged swords that can be used to manipulate the user. The information provided by tracking and journaling tools often serve as motivators in accordance to Fogg’s model for Behavioral Change which is where the areas of psychology and HCI, at least in the case of this thesis, intertwine. While research has been done on both positive and negative aspects of personal informatics there seems to be a lack in whether the positive can work to counteract the negative.
1.1 Aim & Research Questions
The aim with this thesis is to find out how personal informatics and the techniques that go into them can be used to balance the increased availability of Gambling in order for players to be able to engage in these activities safely. With this in mind, two research questions have been constructed.
RQ1:How can players' ability to remain in control of their Gambling be enhanced through the use of Personal Informatics?
RQ2: What, if any, role does the social aspect of Gambling play in maintaining a responsible behavior and how does it affect the experience of Gambling?
While systems conceived from personal informatics are largely focused on the self, the second research question seeks to provide an insight into how players affect one another’s behavior in terms of responsibility. With this, I have conceived the null-hypothesis that Gambling in a social contextdoes not evoke a more responsible behavior. As it is difficult to objectively state what components are part of the term ‘social’, social will throughout this thesis regard ‘engaging in activities together’. To clarify; social Gambling regards engaging in Gambling activities together with someone else. This hypothesis, together with the research questions, will serve as the foundation for further discussions later on in this paper.
In order to answer the research questions two studies were conducted. The first focuses on gathering insights through semi-structured interviews in order to determine whether a player's understanding of Gambling falls in line with the findings from the related research presented in section 2. The second study takes the findings from the interviews and implements them into a design sketch to evaluate whether the theories in section 2 together with the insights from the interviews can be translated into objective design solutions, rather than subjective understandings of emotions. Of course, some of the theories have already been proven to have their place in a design context, but I have been unable to find any documented research on a case in which the same attributes that are used to engage users with the system are utilized to prevent users from using the system excessively, which leads us to the writing of this thesis.
1.2 Terminology
To make this paper easier to follow, some clarifications are needed about the terminology. The three expressions of ‘Gambling’, ‘games of luck’, and ‘betting’ will be frequently recurring throughout this read. In accordance with the Swedish Gambling Authority (n.d.), Gambling is 1 the family name for all activities where the participant stakes money. While there seems to be no official statement that separates games of luck and betting, I have chosen to do so throughout this study as can be seen below.
Gambling (capital ‘g’): The family name for gambling and betting activities.
Games of luck : Games based on luck that cannot be affected by knowledge or otherwise, such as lottery or scratch cards.
Betting: Games where the participant has some means to predict the outcome such as betting on horse racing, football, and other sports or games where previous knowledge might play a role in predicting the outcome.
Player: A person partaking in games of luck and/or betting.
1https://www.spelinspektionen.se/en/
Entry: A statement from a respondent that has been coded. For instance, wanting Gambling companies to provide more information about their business and Gambling would be considered one Entry and coded as ‘transparency’.
1.3 Disclaimer
The sole intent behind this study is to identify means for those engaged in Gambling to be able to do so safely. Hence, the purpose is not, neither partially nor wholly, to promote Gambling as an activity.
2. Related research
While part of the end goal of this study is to identify deciding factors in decision making and behavioral change in an online context, a large amount of the research presented in this section is not solely about topics directly related to HCI. In order to identify said factors, we must first understand what causes motivation and a person to reflect on their actions. In other words, it is necessary to understand the fundamental, psychological aspect that we aim to affect through the use of tools that are within the field of HCI. Seeing as online Gambling is the focal area of this thesis, this grants a direct insight into various tools and techniques that are used to affect players and their behavior. However, in order to use these tools, or find their counterparts, we need to first understand how they affect the players. In this section, I seek to bring an understanding to how various techniques and theories are being utilized within the context of HCI to promote engagement and behavioral patterns.
2.1 Behavioral Change & Gambling
Behavioral change (BC) is dependent on three categories; motivation, ability and triggers (Fogg, 2009a). ‘Motivation’ consists of three subcategories; Pleasure/Pain, Hope/Fear, and Social Acceptance/Rejection. ‘Ability’ regards a far greater amount of properties and is dependent on the target behavior and individual striving towards said behavior. When designing persuasive technologies that, in essence, are aimed to induce behavioral change (Fogg 2009b, Morris 2012), there are many factors to consider. For one, the target behavior ought not to be too vast. For instance, as in the case of this study, completely annulating the risk of attracting a Gambling addiction might be deemed a tremendous undertaking, whilst ‘spend X less amount of money per month on Gambling’ is more approachable.
Identifying what prevents the target behavior is the next prerequisite and is always due to lack in either motivation, ability or triggers (Fogg 2009a, Fogg 2009b). Figure 1 shows Fogg’s model for behavioral change. Motivation and ability can trade off with one another, meaning that someone with severe lack of motivation might still reach the target behavior should their ability be of great quantity. This, however, requires that the right trigger presents itself at the right moment. Triggers can present themselves in many different forms such as alarms, text messages and practically whatever else that serves to grasp the attention of the person intended to be triggered. As with ability and motivation, triggers of different characters have varying
effects on the receiver. Pop-ups are one example given by Fogg (2009a) that are often interpreted as annoying as they cause an interruption in the current activities. Finding the right trigger for the task is therefore crucial.
Figure 1. Fogg’s Behavior Model (FBM).
As part of motivation, Morris (2012) points to two points that arguably are applicable in the context of Gambling; showing people what they could lose and putting the message where the action is. This falls in line with the findings of Egger (2002) who found informational content to be the most effectful means to gain trust amongst users within Gambling sites. In other words, increasing transparency by providing information about the risks of Gambling, should subsequently increase trust which aligns with the hope/fear-motivator in Fogg’s model.
As online-Gambling is the main context targeted for this tesis, some of the elements tied to ability and enabling BC are not preliminary relevant. For instance, physical effort is not particularly applicable when focusing on an online context as it requires no direct physical effort other than basic bodily functions such as typing on a keyboard or moving a mouse. The three key components in regards to ability are rather brain cycles, non-routine, and money. Brain cycles entails the amount of mental effort that is required to change a behavior, whereas the greater effort required, the less able to change a person is. Non-routine suggests that a person is more able to behave in a certain way that has been established as a routine, whereas the farther the target behavior is from the person’s routine, the less able they are to achieve it. Lastly, money regards the sheer financial resources that a person has. Loss chasing is the notion where a player tries to recover losses by increasing the quantity of their Gambling (Inagaki et al. 2019). If a player commits to loss chasing for an extended period of time, it is possible that they do not deem themselves financially able to stop Gambling before having recovered from the loss.
2.1.1 Self Determination Theory
In contrast to BC as described above, there is Self Determination Theory (SDT). SDT argues that human beings are motivated by nature and prone to partake in activities they find interesting and enjoyable with the degree of motivation being decided by three innate psychological factors;
autonomy, competence and relatedness (Cherry 2019; Mekler 2020; Fu et al. 2018). Autonomy refers to our willingness to partake, competence how large of an impact our partaking has, and relatedness regards a reciprocal sensation of belonging in social circumstances.
The fundamental difference between SDT and BC is external leverage. In SDT, the motivation to partake in an activity, or changing a habit, is solely supposed to emerge from within the person’s own psyche. Instead, BC seeks to influence the person’s psyche by adding external leverage. An example of such, in relation to Gambling, could be to offer new players a bonus when creating an account, effectively increasing their motivation to do so. It is rare that an action is driven solely by intrinsic or extrinsic motivation (Cherry 2019). With this in mind, SDT and BC are more like yin and yang than polar opposites. Aspects from BC can be emphasized in order to promote one’s self determination, and vice versa.
2.1.2 Risk & Availability
The commonly known definition of risk suggests that an activity can be objectively identified as one with a potential repercussion. Kahneman (2011) refers to the work of Slovic who challenges the idea of risk being objective, and rather means that humans have invented the notion of risk as a means to gain an objective lens on dangers and uncertainties in life. The dangers are very much real, but the concept of objective risks are not (ibid).
An availability cascade is a self-sustaining chain of events (ibid). In relevance to this thesis, such a chain of events could be a person being exposed to a Gambling commercial, effectively leading to them becoming bankrupt from excessive Gambling. In reference to risk as mentioned previously in this section, there really is no apparent danger at first sight seeing as the person in question can simply choose not to partake in these activities. However, there are certain psychological phenomena that work in the Gambling companies favor. Should a player decide to start Gambling, they are likely very uneasy about their decision until the die is cast, but regain their confidence the second it is too late to back out by convincing themselves that the right decision was made (Cialdini, 1984). Cialdini explains this behavior by the player, in this case, falling victim to a weapon of social influence, being our desire to maintain a certain self presentation, thus forcing ourselves to be content with our decision regardless of the outcome to radiate a sort of consistency in our behavior. The pitfall is how a sudden large, unexpected sum of money increases a person’s willingness to spend (Thaler 2008). In other words, a player’s mindset might remain somewhat consistent while losing, but can drastically change if winning a large amount, leading to more aggressive Gambling..
Availability has an additional meaning in the realm of HCI. Perhaps more commonly spoken of in terms of ourselves being available to one another through social media (Walker Rettberg 2014; DeVito et al. 2018), the same applies to services and systems being more easily available to us directly through our smartphones. Gambling, of course, is no exception to this, where saying
that it is more available is quite a literal statement; one is no longer required to physically visit a casino or race track in order to Gamble. In relation to the availability cascade as mentioned priorly, the increased availability of Gambling, theoretically, increases the density of the links in the chain, making it more likely for one to partake in Gambling once introduced to the first link of the chain.
2.2 Personal Informatics as a Means to Reflect
Personal informatics (PI) has become ever more prevalent in our everyday lives through various tracking apps and devices (He et al. 2013; Epstein 2015). Traditionally, tools with their roots in PI are designed to serve as enablers for the user to reflect upon their habits and ways of life.
However, this is largely focused towards the internal health of the user, e.g. to break habits leading to obesity, reduce stress, etcetera. There seem to exist a rather slim selection of PI aimed to affect the external health, e.g. financial resources. Furthermore, systems with their roots in PI are in the vast majority of cases utilized as journaling tools while not providing any direct means for the user to take action upon having reflected on the information gathered from using the system.
He et al. continue by pointing out three critical stages of reflection that need to be considered when designing PI systems. The first being awareness of uncomfortable thoughts and feelings.
This regards thoughts and feelings when utilizing the PI system, such as the system itself generating stress due to being an additional activity for the user to consider and keep track of.
The second is to provide users with enough information to suffice for a critical analysis of their behavior and habits. If the PI does not provide sufficient information, or rather allow for detailed enough tracking, chances are that the user will give up on using the system. The same applies if too much information is provided, whereas the user might become overwhelmed which may lead to overthinking and destructive self-reflection (Morin 2006). It is therefore required that the correct amount of information is presented at the correct moment for proper self reflection (He et al. 2013).
The third and last stage is that of developing a new perspective. In essence, this is the moment where the user of the PI comes into realization about their current behavior through pondering the information presented to them during stage two. These realizations often lead to a change in mindset and thinking, but there is no guarantee that they lead a change in behavior (Boud 1985).
2.3 Gamification
Gamification uses elements commonly found in video games such as challenges, rewards, and progression (Benyon 2013; Cunningham 2011; Greenwald 2014). It has become common to implement these elements in other systems than games to promote engagement and, in many cases, serve as a motivator for behavioral change. Greenwald gives several examples of areas where gamification has been successfully implemented, whereas personal informatics is one of them. In health tracking applications, challenges and rewards are often present to incentivise the user to remain active. Hyrynsalmi et al. (2017) rather point to the negative sides of
gamification that very much go hand in hand with the positive aspects, but become negative depending on the context in which they are applied. In other words, the results produced by the gamification have mostly been identified as positive, but may cause ethical concerns in terms of what kind of engagement is promoted. For instance, with focus shifted back to personal informatics, gamifying a health tracking application where the user is rewarded for achieving certain milestones in their progress toward a healthier lifestyle may very well be deemed ethically sound. On the other hand, implementing the same reward system where the user is rewarded for spending a certain amount of money, as is common on many low-cost online markets such as Wish or Banggood, could, in accordance with the findings of Hyrynsalmi et al., be deemed unethical.
2.4 Social Relations
With the increased digitization of Gambling (Hancock, 2017) the social relations have changed between both Gambling companies and customers as well the relations between customers.
Corporate Social Responsibility (CSR) regards a companies accountability in regards to their societal impact (Chen 2020; Hancock 2017; Porter 2006). With the condensation of the availability cascade, Hancock (2017) argues that this impact is increased while Gambling companies are no longer, to a large extent, in direct contact with their customers. With Gambling activities becoming increasingly available and, as a consequence, Gambling companies becoming more distant, their CSR increases.
The social commitment (Clark 2006) of Gambling also changes drastically with activities taking place online rather than in a physical environment. The Public self-commitment, a commitment made to oneself for the public to see, such as placing a bet of certain proportions while at the race track, disappears entirely unless the online context supports social interaction.
This is closely tied together with self-presentation (Walker Rettberg 2014; DeVito et al. 2018) and the motivating part of Fogg’s (2009a) behavior model. The motivation to maintain a certain portrayal of oneself largely disappears when the self you aim to present cannot be perceived by others. Furthermore the participatory commitment made between friends and acquaintances going to the race track or casino together is also largely deteriorated while actual planning to exercise Gambling is no longer required. Put briefly, by neglecting social interactions during the translation of Gambling from a physical space into an online space the negative societal impact of Gambling has increased. Hancock (2017) continues by emphasizing that an estimated 30-50%
of net Gambling losses are contributed by people with Gambling related problems. With the previously mentioned factors in mind, the situations of these players are likely to escalate even further.
3. Study One - Interviews 3.1 Methodology
Interviews were chosen as the method for collecting data. More specifically, semi-structured interviews (Benyon 2013) were utilized. Seeing as the aim was to acquire an understanding about the respondents' subjective understanding of various aspects of Gambling it was deemed more appropriate to solely focus on a qualitative method than a quantitative one. Mostly due to the nature of quantified methods where results ought to be presented in numbers, which would severely narrow the ability for the respondents to elaborate upon their own, personal view and experience (Benyon 2013; Cooper 2014; Sharp 2011). Cooper particularly states that
“Qualitative research is almost always the most effective means for gathering the behavioral knowledge that will help designers define and design products for users” (Cooper, 2014 p. 35).
Seeing as BC is a cornerstone for this study, it would seem natural to rely on methods that keep behavior in mind.
3.1.1 Respondents
The respondents' aliases correspond to their number in order of being interviewed, gender, whether they are experienced or not as well as their age. For instance, the first respondent would be R1EM55, signifying that they are an experienced male, whereas the eighteenth respondent is written as R18IF26, signifying a 26 year old female with lack in experience within Gambling.
Twenty respondents were recruited with help from an external recruitment agency. They handled all of the initiating contact with the respondents and scheduled the interviews that were then conducted by myself. Several criteria were set for the respondents to meet. A desired mix in both experience with Gambling, geographic location and gender rendered local recruitment of participants invalid. The diversity in geographic location was set as a criteria in order to minimize the potential influence of local culture on the results. The respondents were distributed as eight females, whereof five experienced and three inexperienced and twelve male with eight experienced and four inexperienced (see Figure 2). Their age ranged from 23 to 70 with a median of roughly 43 years of age. Note that the age of R20EM was never verified, but given the information provided while conducting the interview, their age was estimated to be around 50 years old. Due to the varied geographic locations as well as the prevalence of Covid19, interviews were decided to be conducted either through phone or Skype. All interviews ended up being conducted through phone with no video transmission being utilized. No further information such as e-mail addresses were exchanged.
As established, experience is subjective and cannot be quantified (Benyon 2013; Cooper 2014;
Sharp 2011). With this in mind, the respondents were asked whether they consider themselves experienced rather than having to meet certain criteria to be deemed experienced. While Benyon, Cooper and Sharp write aboutperceived experience from using a system or product, I would argue that the same logic applies to accumulated experience. Ponder that two persons use
the same system continuously for one year. Surely they will not have had the exact same perceived experience during this period, nor the same accumulated experience afterwards.
Figure 2. Distribution between male and female respondents and their experience.
3.1.2 Ethical Considerations
Participating in this study was, of course, voluntarily. During their first contact with the recruitment agency, the respondents gave their consent to partake in the study and were given a brief introduction to the subject. The conversation with myself, who performed the actual interviews, was initiated by the respondents being given the opportunity to ask any questions that they might have had before proceeding. It was made clear to each participant that a prerequisite in order for me to be able to process the interviews was that I had their permission to record it. All respondents gave their consent to being recorded, whereas the very first part of the conversation after having started the recording consisted of me further elaborating key points such as guaranteeing the respondents that their anonymity will remain intact. A few participants asked for the purpose of the study to be further clarified before starting the interview, whereas they were given more information and it was made sure that they were sufficiently informed before proceeding with the interview.
Though two pilot interviews had been conducted before talking to the respondents, some details were changed or added during the first two or three interviews. This included a disclaimer that informed the respondents that there was no prestige involved in the interview where no answer was more right than the other, and that they were not expected to answer any or all questions, but rather those that they wanted and felt comfortable to answer. Listening back to the first couple of interviews, however, does not show any apparent indication that the respondents answered any of the questions unwillingly, whereas I deem it safe to say that all respondents truly gave their informed consent.
3.1.3 Interviews
As mentioned priorly, the interviews were conducted in a semi-structured manner, meaning that there was a set script with questions that were mostly open for the respondents to elaborate
their answers freely (Benyon 2013; Cooper 2014; Sharp 2011). The script started off by defining the difference between games of luck and betting (see section 1.2) followed by asking whether the respondents considered themselves to be gamblers or betters. During the interviews
‘gambler’ referred to a player partaking in games of luck. This part was slightly altered within the first two to three interviews, where the option of identifying with neither categories was added. This was done mostly due to the questions being found rather difficult, whereas I deemed it necessary to add the option to alleviate the respondents should they find the question’s difficulty intimidating, potentially affecting their willingness or ability to answer further questions.
Following the first question, the interview was divided into three main parts. The first was mainly to get to know the respondent by asking for how long they had been playing, what first made them interested and what the reason is that they continue playing. Note that the third question was not asked in such a questioning manner as this text might suggest.
The second part revolved around more subjective questions regarding their expectations of features and safety nets as well as their thoughts on Gambling as a perilous activity and addiction.
Part three regarded the social aspect of Gambling and, partly, the respondents own social habits in conjunction with Gambling, but also regarding what kind of help should be provided to those in risk of letting their Gambling go too far. The interview was rounded off by thanking the respondents and giving them the opportunity to ask further questions should any have been thought up along the interview. In total, the interview consisted of 15 scripted questions, where follow up questions for the sake of clarification were asked with most respondents at some point throughout the interview. For a full list of the interview questions, see appendix A.
3.1.4 Coding and Analysis
Following the conclusion of the interviews, they were all transcribed and imported into the AtlasTi cloud version which is a tool for structuring and analyzing data. While reading the transcriptions, key points from all questions were given codes for the impending thematic analysis in accordance with Mortensson’s (2019) depiction. The thematic analysis consisted of grouping the snippets of code together as seen fit. Note that my own judgement is what determined the relevance that pieces of code held against one another, which leaves some room for error and biases which might have affected the results of the analysis. The 337 codes that were associated with the content of the interviews were split into twelve different groups. Some codes were allocated to more than one group when found relevant.
The groups and codes were cross referenced in an attempt to detect patterns with the respondents. The results from the interviews are elaborated upon in the following section.
3.2 Interview Results
With the interviews being conducted with an open-question approach, they resulted in a great amount of variety amongst the responses. In total, 337 entries were documented with 106 different codes. Some amount of interpretation was needed in order to properly categorize the different entries while no respondent would answer any one question the same as another
respondent had previously. A very few entries required a greater amount of interpretation as some respondents spoke very freely and elaborated their thoughts even after having given an answer to the original question. This resulted in several entries that had their own coding made specifically for them. For instance, one respondent mentioned how they could see the possibility that people who had previously exercised a particular sport would transition into betting on that sport when they were no longer able to exercise it on their own. Seeing as this entry is an anomaly amongst the replies to this particular question it could not be coded with an already existing code, but was instead given its own seeing as it is still relevant.
Though the respondents were fairly evenly divided between male (12) and female (8) and experienced (11) and inexperienced (9), no significant patterns could be found amongst the respondents. Considering the variation in age and geographic location, there are arguably too many variables to draw a statistical conclusion without a greater amount of respondents. This is elaborated upon in section 5.1.1.
3.2.1 Gambling as a Perilous Activity
The participants were asked whether they consider Gambling risky. Whilst most respondents were not considerably worried about their own habits, they were unanimous that there is at least some risk involved with Gambling. It is worth mentioning that ‘risk,’ in this case, does not regard the obvious risk of losing the money that has been staked, but rather greater consequences such as debt, negatively impacting relations and neglecting obligations as argued in section 2.1.2. This was not explicitly articulated in the question, but rather conveyed through paralinguistics and, if needed, by further elaborating the question for the respondent.
3.2.2 Risk Factors and Addiction
The most recurring risk identified by the respondents was the risk to become addicted. This was explicitly mentioned by eleven separate respondents as part of the above mentioned question regarding whether they consider Gambling to be risky in general. In twelve entries, respondents underlined that certain people are more prone to becoming addicted by having an addiction prone gene. These twelve entries, despite being coded as such, did not all regard addiction in itself. Some of the examples brought to light by the respondents were rather related to excessive or obsessive Gambling without necessarily emphasizing addiction to be the underlying cause of the behavior, such as loss chasing. Apart from this, the largest risk the respondents saw was the availability of Gambling in today's society in relation to being able to participate in Gambling through One respondent drew the parallel between alcohol induced reduction of judgement, and how this could not be monitored in a virtual space in comparison to personally appearing at an actual casino.
“It’s like… not having too much alcohol in the mix. It’s a drawback of online casinos that they have no idea about it. If you’re at [name] for instance they can choose not to let you in or ask you to leave. So there’s a dissonance” (R9EM30).
Individual responsibility was also mentioned in eleven separate entries. In some few cases, this was mentioned in conjunction with the addiction prone gene and how, regardless of any condition or situation the player might have or find themselves in, it ultimately comes down to themselves to take responsibility for their actions. An example of such an action that was brought up several times was obsessing about loss prevention. In other words trying to, in a seemingly desperate manner, win back money that you had lost. One respondent further gave an example of such obsessive behavior where if a player would win a large amount of money early in their Gambling career, it could set an unrealistic standard with the player that they desperately seek to relive.
Peer pressure was identified on five occurrences too where the main concern was associating oneself with a group of people prone to an aggressive Gambling behavior rather than the activities in and of themselves that the player would be peer pressured to take part in. In other words, the Gambling itself would not necessarily become more aggressive in the sense that the stakes were increased, but rather the extent to which the pressured person would take part in the activities would increase.
3.2.3 The Social Aspect
Upon being asked about the social aspect of Gambling the respondents would on 18 occasions speak of traditional social Gambling. To clarify, ‘traditional’ in this sense refers to playing lotto in front of the TV with family and friends, betting on horses at the local tobacco shop or race track, and so on. The untraditional social would, in contrast, be socializing online through various Gambling-related activities such as discussing on forums, chatting with others during a game, and so on. On merely five occasions did respondents suggest that online may be considered social, where two of those entries were definitive that online Gambling is social.
Furthermore, in four entries it was explicitly expressed that performing your Gambling activities online is antisocial.
Seventeen entries were made by separate respondents claiming that Gambling in a social setting was more fun. In a very few cases, peer pressure was brought up as a potential consequence of actually perceiving the experience as more fun. Not in the sense as described previously, but rather that the group of people engaged in the Gambling activities would get carried away, not wanting to cease the activities even though the time might have been right to do so.
3.2.4 Support, Restrictions, and Interventions
The respondents were at the later stage of the interview presented with the hypothetical scenario of having a friend Gamble excessively, risking to end up in a dire economic or social situation.
The follow up question being what kind of assistance they should have access to or be provided with. 17 respondents, both those with experience of online Gambling and those without, urged that restrictions should be placed on the vulnerable person from the Gambling companies point of view. Examples of such restrictions that were given were limiting the maximum deposit each month depending on the person's income, receiving continuous reports and a statistical review of the person’s Gambling habits and the money they’d spent. An additional restriction that was
explicitly mentioned seven times was how excessive Gambling should lead to an immediate suspension for a set period of time. Similar to this, one respondent explained how they had suspended themselves from further Gambling. When attempting to Gamble the next time, the suspension was still active, whereas the respondent had been thankful for the ability to restrain themself in this manner. They said;
“I intended to purchase a share with [company name] and when I was in the store they said ‘it seems as if you’re suspended’ which I’d done myself four years ago and that made me think that ‘great, then I won’t buy it.’ Then I didn’t get stuck because I’d suspended myself and I think it’s great to be able to do so. It’s good that the companies take that kind of responsibility, so to say” (R15EM55).
The most recurring, concrete example, with seven separate entries, apart from the restrictions was that the Gambling companies themselves should be able to provide some initial support. The person in question could, at the very least, receive guidance on where to turn to receive proper care from a professional psychiatrist, given that a psychological issue is the underlying reason for having ended up in the current situation. Seeing as none of the respondents had been in this particular situation, it was seemingly difficult to imagine exactly how such a first step or help service could look like. A chat service was, however, mentioned on a very few occasions whereas one respondent underlined that it was important for the one seeking help to know that they were in contact with a human being instead of an automated system.
Lastly of the more common entries is that of interventions from people that stand close to the troubled person. Eight separate respondents identified this as one of, if not the, most important help that a person can receive, be it through family or friends. The respondents seemed vexed, however, on where to draw the line between supporting and pressuring them to seek help. One respondent said;
“It’s like… a clear and straightforward confrontation from myself or someone else who’s close to them. And if that doesn’t do it, you’ll have to seek help. Although you can’t drag that person by their hair to the carecenter, but rather it has to be voluntarily” (R9EM30).
While they wanted to help their friend, they still thought that the decision to seek medical attention is ultimately their own. Interestingly, any kind of preemptive support to prevent falling too deep into Gambling was only spoken of by one of the respondents by very briefly agreeing that the social aspect of Gambling could potentially benefit responsible behavior. In other words, the respondents did not claim to rely on their peers, or vice versa, to maintain a healthy habit about their Gambling. This will be further elaborated upon in section 5.1.2.
As mentioned in section 3.1.2, the availability was identified as a risk factor with Gambling.
Coincidently, suggestions were made by four respondents to either decrease availability or increase the threshold to start Gambling. Concrete examples as to how this should be done, however, was not elaborated upon.
3.2.5 Features
When asked about features it was made apparent that this is something that the respondents had not previously considered. This made it fairly difficult to answer what their expectations were regarding features when, for instance, visiting a Gambling website, whereas, in most cases, the question needed to be revisited or further elaborated upon. Included in this set of codes are other features that the respondents mentioned as part of their answers to other questions, such as the ability to restrict one’s Gambling directly through the Gambling company or how excessive Gambling would lead to a suspension as has been discussed above. The most recurring, desired feature was the ability to restrict oneself directly through the Gambling site with 17 entries. Some respondents did mention this kind of feature being readily available at several Gambling websites already but still emphasized that it is both important and desirable.
Transparency, mainly in regards to information about addiction, was brought up by seven respondents at eight separate occasions. These occasions were fairly widely spread between the different questions of the interviews, suggesting that the respondents might perceive the value of transparency differently depending on the context in which it is presented. The two most common occurring occasions where transparency was spoken of was either when talking about the expectations that the respondents had of Gambling companies, or when asked what kind of relief Gambling companies should offer those who had let their Gambling habits run out of control.
3.2.6 Miscellaneous Comments
Three respondents, who had experience with both betting and games of luck, commented on the investment that was needed to participate. Routing back to the question of availability, games of luck were identified as more risky whilst the threshold to begin playing was perceived as lower.
Once your online profile had been registered, starting a Gambling session according to R20EM 2 was no more than a few clicks away. Using their child as an example when talking about availability they said;
“They are not allowed to play on any internet casinos or similar even though they are so easy to get access to through your mobile phone. You make an account and then it’s just a matter of sitting down and pushing a few buttons” (R20EM).
In contrast, betting was argued to be less available while a bigger effort is required to participate in a bet rather than games of luck. For instance, when betting on a horse race you have to make active choices and create a system for how you predict the outcome of the race. As such, three respondents, apart from R20EM, deemed betting more intellectually demanding than games of luck and that betting is more fun because of this and has more potential as a hobby while it requires a larger effort from the practitioner.
R13IM36 drew parallels between Gambling and gaming communities in regards to the social aspect. They pointed to gamers who invest money in free-to-play games, such as Fortnite or
2 The age of R20 was never verified
League of Legends, to alter the appearances of their characters to make them feel more unique and create their own way of presenting themselves through their characters. R13IM36 suspects that similar personalization would be desirable within the Gambling industry, seeing as Gambling is, and already has, established itself online. During the evaluation mentioned in the previous section, this was also briefly looked at whereas no apparent options to customize one’s profile in terms of self presentation were given.
As can be seen in previous sections, availability has been recurring and has exclusively been about availability online. R19EM54 did make a remark about availability as a prerequisite to exercise Gambling in a traditional manner. They said;
“It is an advantage, if we’re talking horse racing, if you actually live in a city that has a race track that actually arranges interesting events. [...] If we take ATG, then I’d wish that the Tillsammans-feature would perhaps be further developed to be3 even more like a community than it is today. To me the social part is more online than in real life” (R19EM54).
Another remark that was made by R19EM54 was on information provided by the player to the company. They said;
“I don’t know I’m… a bit allergic to, you know… providing too much personal information. [...] I don’t really know what to say. A level that works and a good supply (of services on Gambling sites)” (R19EM54).
While this is the only entry that speaks more explicitly of not wanting to provide an excessive amount of personal information, there was another comment given by R10EF49 who explained that they had received a contract from the company whose services they employed where they declared how the personal information of R10EF49 was handled. This transparent act gave R10EF49 a sense of security in that the company was perceived as conducting their business seriously.
4 Study Two - Concept Design and Assessment
4.1 Methodology
With the related research presented in section 2 together with the results of the aforementioned interviews, a basic design sketch was created. The goal with the sketches was to evaluate whether my interpretation of the interview results and their relation to the related research was correct. Two participants were recruited, consisting of two males, one being an experienced
3“Tillsammans” is a betting service provided by ATG in Sweden that combines betting with elements from both gamification and social media.
player and the other having no experience at all. They will be presented in similar fashion as the respondents in the interview, namely P1EM36 and P2IM27.
The reason for the evaluation consisting of only two participants is largely due to Covid19.
The delays that the spread of the virus caused in Sweden at the time of this study caused the design process to be severely delayed, limiting both the allotted time to do the evaluations, but also limited the ability to recruit participants. Therefore, both P1EM36 and P2IM27 are personal friends of mine.
Nielsen (1993) suggests that any design process should include no less than three iterations.
However, due to the above mentioned reasons, only one iteration of the design was tested. Focus was largely oriented toward the features of the design rather than the experience they provided.
For this first iteration, as mentioned, I wanted to evaluate the findings from the interviews and research rather than provide the participants with the most enjoyable experience possible.
Future iterations would, naturally, have been further developed in both areas.
4.1.1 Concept Design
The sketches of the concept design were created using Balsamiq, a tool to create wireframes and basic prototypes. Due to this being the first iteration of the concept design, I chose to utilize Balsamiq due to its ease of use and ability to create sketches with a fidelity of your choice. Lau (2019) points out that the chosen fidelity largely decides how the participant of the evaluation perceives the prototype. In short terms, focus when evaluating a lower fidelity prototype often lies with functionality. A prototype of higher fidelity is expected to have the functions largely in place while it at first glance is perceived as more polished which, in turn, makes the evaluator rather focus on their experience when using the system. With this in mind, creating a low-fidelity prototype seemed like the most appropriate route seeing as the prototype seeks to provide a brief evaluation of the implemented functions in relation to the findings of section 2 and results of study 1. Having the participants focus on their experience would hence divert them from the intended focal area of this study. The mockups were made to represent windows in a web browser as can be seen in figure 3 and 4 as well as appendix B. With the apparent desire for self restriction and how sufficient information is required for self reflection (He et al. 2013;
Epstein 2015; Morin 2006) the main functionality of the system to-be revolved around acting as a hub from which the user can apply both global and specific restrictions whilst being provided with an accumulation of statistical data from all Gambling sites on which they are active. It was, quite fittingly, given the name ‘GambleManager’.
As stated in section 2.2, journaling tools can provide the information necessary for someone to reflect upon their behavior. While Gambling sites do generally provide this possibility to an extent, there are currently no means to track one’s behavior across all Gambling sites collectively. With this in mind, a large focal point for the first iteration was to provide the participants of the evaluation extensive journaling capabilities. Graphs and charts that show the users current Gambling balance, their win-loss rate, time spent and other similar statistics are presented both for individual sites as well as collectively.
In order to make up for the loss of the participatory commitment (Clark, 2006) the users are given the option to establish a social network through ‘trustees’ who are given insight into the
Gambling habits of the user, whilst also giving the user insight into their own. By introducing achievements into the design, users would be incentivised both for their own sake (Benyon 2013; Cunningham 2011; Greenwald 2014), but also given direct increments to which they could commit together with their trustees. While not intended to function as a social media, achievements would also enable a sense of self-presentation as explained by DeVito et al. (2014) and Walker Rettberg (2018) by allowing the users to present themselves through what they have accomplished when using the system. Furthermore, should limits be surpassed by the user, their trustees will receive an alert and be given the option to directly contact the person in question.
This largely builds upon the reflections given by respondents in study 1 where interventions seemed more natural than providing preemptive social support. However, seeing as some damage would have already befallen the person by the time that the intervention takes place, prevention rather than intervention should minimize loss. By providing users with this function they should have greater ability, in relation to Fogg’s (2009a) behavior model, to assist in proactively preventing rather than reactively intervening.
Due to time constraints gamification was not elaborately implemented. Rather, the headline for the systems’ intended gamification element was implemented into the navigational menu, but was not made clickable. It was decided to be included despite not being elaborate for the sake of finding the initial reactions of the participants upon encountering achievements in this context. It was my assumption that the participants would reflect upon this increment as they are both vivid gamers and familiar with the concept of achievements as part of a gamified context.
Figure 3. The first slide of the prototype, showing an overview of recent activities and statistics.
Figure 4. The second slide of the prototype, showing the page where the user can manage their general Gambling restrictions.
4.1.2 Think Aloud
The evaluation was performed as a Think Aloud session (Barnard, 1994) . The participants were asked to verbally reflect upon their thoughts and actions while interacting with the early prototype. This approach was chosen in order to gain insight into their cognitive understanding of the prototype, as well as to enable the finding of nuances in their reflections that could, hopefully, be connected to the findings of the interviews and related research. Worth noting is that neither P1 or P2 had any insight in the study apart from its general purpose.
P1 and P2 were each presented with a hypothetical user scenario, where P1 was given a scenario from an experienced player’s point of view, and P2 that of an inexperienced (see appendix C). The session started with each participant reading their scenario out loud before beginning the evaluation of the design sketches. The participants were encouraged to ask questions apart from reflecting if there was something that they did not fully understand. Both participants were given a clarification in excess of the presented scenario, stating that the visible content of the prototype was generated from having used it for a couple of days, having added people to their contacts list, and so on. Furthermore, P2 was given a few pointers on features to look at if they got stuck as this was their first time participating in an evaluation of this kind.
Some features, such as ‘achievements,’ were explained when necessary in a manner that did not instill biases with P2.
The evaluation with P1 was carried out in my home sitting by my personal computer, whereas the evaluation with P2 was carried out through Discord, which allowed us to communicate verbally and for P2 to share their screen as they were interacting with the prototype.
4.2 Results of Design Concept Evaluation
In this section, the results of the evaluations are presented. Seeing as the participants had rather different views on the prototype, they will each be presented in their own section. Figures 3 and 4 below are presented to provide a better understanding of what the participants were presented with. Scaled up images that are more readable can be seen in appendix B. As mentioned in the previous section the participants were tasked with a think aloud where they communicate their reflections verbally as they go through the prototype.
4.2.1 P1EM36
As previously stated, the first participant is experienced with Gambling and has been practising a systematic way of Gambling for some time, hence gaining great insight into a vast variety of Gambling sites. Already when reading their scenario (see appendix C). they reflected over the fact that, according to them, money was sitting idle in their bank account due to Covid19, and the lack of matches to bet on that came with it, rather than being invested through Gambling.
Some initial questions asked by P1 indicated that it was not made perfectly clear to them what the purpose of the system before them was or why they should use it. This was clarified as necessary.
When looking at the ‘overview’ page, P1 pointed out that;
“I can instantly tell that the purpose of this tool is to limit one’s Gambling. It says here that you can lock yourself out. I mean… What seemed tempting for me was the ability to keep track of, especially the ‘outcome this week.’ [...] that’s what I’m mostly after, like, ‘how am I actually doing?’” (P1EM36).
They continued by elaborating upon the ability to lock oneself from further Gambling and mentioned how Gambling sites often have such features implemented already and that the functions of this design could probably be connected somehow to these already existing features.
It was also made clear that, from the experience of P1, the already implemented limiters usually do not have any roof or recommended deposit, meaning that a player could effectively enter any sum of money as their max deposit, rendering the function redundant. To emphasize, this reflection was made toward already established systems, not the prototype for this study.
Upon inspecting the trustees that can be seen in figure 3, P1 stated that they felt pointed out as if trustee 2 should not pay any attention to P1 and their Gambling and that it reminded them of buddy programs that are commonly used as part of rehabilitating addictive behaviors in regards to drugs or alcohol. While not explicitly stated, it is my interpretation of the paralinguistics in this statement that leads me to believe that this was viewed as something negative by P1.
Throughout the evaluation, P1 drew parallels between the prototype and a Swedish self restriction service that is available to everyone and connected to all Gambling sites that have a Swedish license to conduct business with Gambling for money. These reflections seemed to rather acknowledge the value of the implemented features for the sake of self restriction rather
than criticizing them seeing as some of them are already implemented into a well established system.
My assumptions regarding the acknowledgement of achievements mentioned in section 4.1.1 were proven to be true. At first, P1 seemed a bit vexed regarding whether the achievements were tied to the activities that users engaged in that involved Gambling, or actions they took to restrain themselves from playing too much. Upon reading the scenario a second time, P1 was determined that the achievements were tied to one's self restriction, time not spent Gambling, etcetera. Furthermore, P1 gave an example of a Gambling site that uses gamification elements in abundance which diminished their credibility while the excess made the intent behind them too obvious according to P1. This will be further discussed in section 5.2.1. Availability in terms of access to information was also briefly reflected upon. P1 said;
“[...] it’s obvious that it’s supposed to be a tool to enable healthy practices of Gambling. That makes me think that there should be some kind of way to get support. What if I need help? Where do I turn? I guess there could be different solutions such as phone numbers or chats. [...] and it could pop up if I choose to lock myself out like ‘Alright. Is it possible that you are in need of some help?’”
(P1EM36).
As a last remark, P1 pointed to their overall understanding of the prototype, whereas it was that the primary function of the system was to enable self restriction. As touched upon earlier in this section, P1 was primarily interested in the statistical analysis that the system sought to provide and did not necessarily deem themselves in need of the restrictions. With this in mind, they argued that it could serve the end product well to have a more neutral stance and not so heavily focus on self restrictions, but rather provide this service as part of a system that allows a highly informative overview of the player’s habits with Gambling. Thus, should a player find themselves in need, it would be a short step to initiate self restrictions seeing as they would, ideally, already be using the platform for the sake of its traits inherited by deriving from personal informatics.
4.2.2 P2IM27
The evaluation with P2 started out with them reading their given scenario and the clarification as described in section 4.1.2. Some further clarifications were needed as it had not been made clear to P2 that the prototype was supposed to represent a third party addition to one's web browser and not provided through the Gambling companies directly. When this misunderstanding had been cleared up, the general opinion about the first mockup as seen in figure 3 was positive. Continuing to the second mockup, represented by figure 4, P2 expressed skepticism toward the idea of having the system give you recommendations based on details about the user’s economy. They said;
“I don’t feel like… [...] since it’s an add-on, I don’t know who has access to it and so on. I don’t feel as if what I make every month is anyone else’s business. Even if it
would potentially help me by, in this case, recommending me with ‘about this much is as much as you should spend,’ I still wouldn’t enter the information. It’s so personal…” (P2IM27).
The evaluation continues with P2 briefly mentioning the ability to limit one’s time spent on Gambling sites through the prototype as something positive. Furthermore, they explored the
‘self suspension’ function, where the player can suspend themselves with immediate effect for 30 days, which was also appreciated.
Further exploring the ‘manage’ page led to P2’s discovery of the function to restrict oneself from specific Gambling sites which was also appreciated while P2’s understanding of the Gambling industry, as shared with many of the respondents from study 1, is that there are more and less serious actors on the market. The ability to directly lock any activity from taking place on certain sites perceived as less credible therefore, again supposing that my interpretation of the paralinguistics is correct, induced a sense of calm with P2.
Next, the exclamation marks that can be seen in both figure 3 and 4 were brought to attention by P2. P2 expected to be able to hover their mouse cursor over the exclamation marks to receive further information about what they wanted to convey, similar to question marks that are often used for the same purpose. While this was not a working function in the prototype, the intended function was exactly that which P2 had been expecting. They further pointed out that it would not make sense for the exclamation marksnot to provide further information when hovered by stating that“[...] as a player or user you would want to know what they mean without having to sit around guessing [...].” (P2IM27). Furthermore, they argued that, due to the familiarity with such information bubbles, users would likely instinctively hover over them for more information.
Achievements were again brought up for discussion. P2 thought in line with P1 when first spotting the achievements tab in the navigational menu, being slightly conflicted regarding the intent behind including achievements in the system, depending on from which point of view they were implemented; as a means to make players play more, or more restricted. P2 was asked to read the introductory scenario again, whereas they grew confident that achievements regarded restrictions about their Gambling rather than the other way around.
Lastly, a closer look went into the trustees. In contrast to P1, P2 seemed to highly appreciate the idea of having those around them know about their Gambling habits. In the prototype, two examples of interactions with trustees were given, where in the first case, the user is met with a message stating that one of their trustees, which would be a two-way commitment, have surpassed one of their Gambling limits and have been suspended from further Gambling. The user, P2 in this case, is given the option to contact the trustee directly through the message bearing the information. Similarly, P2 appreciated having received a message from one of their other trustees, wanting to make sure that all is well when P2 has surpassed one of their allotted limitations. They said;
“A good friend, I suppose, who got the same kind of message as the first one. So he contacted me to ask ‘hey, is everything okay?’ So it’s actually just a good thing that he used the message
that popped up for him, I guess, at his own profile, so he was being a good guy and wrote to me.” (P2IM27).
5. Discussion
Throughout this section the results of the two studies will be discussed. This includes the interviews both in regards to respondents and the choice thereof as well as the structure of the interviews might have influenced the results and the prototype that was conceived with the data gathered from the interviews as well as the limitations of these studies and what future studies might entail.
The following discussions are conducted bearing the two research questions in mind, seeking to provide an answer to each.
5.1 Study One - Interviews
5.1.1 RespondentsExperience did not seem to have a large impact on the respondents and their ability to answer the questions. However, some of the more experienced respondents had a tendency to elaborate more, providing further miscellaneous insights, giving more nuance to their replies.
The geographic location seemed to have less of an impact. However, while diverse, most respondents live in areas where they could fairly easily access either casinos, race tracks, or other traditional forms of Gambling. It is possible that this fact has contributed to the, mostly, shared mentality regarding online Gambling and the apparent bias towards the traditional.
While the respondents did not suffer from any apparent problems or addiction related to Gambling it is important to keep in mind that one of the intended user groups is unrepresented in this study. Interviewing people with current issues related to Gambling could have yielded different results and it is possible that some aspects that could have been desirable to include in the design have gone unnoticed as part of the absence of this particular user group.
5.1.2 Interview Results
While the respondents were informed that there were not right or wrong answers during the interview the expected results of the interviews were vastly different from what was actually found. The results were expected to revolve to a greater extent around how the social aspect could serve as a means to maintain and shape a responsible behavior within Gambling.
However, as mentioned in section 3.2.4, the kind of social support that the respondents mainly saw was that of intervening if someone let their habits get out of hand, rather than preventing this from happening in the first place. Considering the emphasis on the individual’s responsibility, it is likely that the respondents would argue that they cannot take responsibility for someone else not to make bad decisions, but their solidarity makes it so that they would help someone in need should the occasion present itself. This tendency does also fall in line with relatedness serving as a motivator in accordance with Cherry (2019,) Mekler (2020,) and Fu et
