Faculty of Technology Department of Informatics

(1)

The loot box dilemma

A qualitative thesis of video game players

interactions and mitigation of unethical aspects in

loot boxes.

Author: Simon Bergstedt Author: Pontus Skoglund Supervisor: Mexhid Ferati Examiner: Patrik Elm

(2)

Abstract

In-game purchase options are referred to as microtransactions and have become a common revenue for modern video games. Microtransactions refer to small payments made inside of video games to acquire game content. The phenomenon of microtransactions is often related to "loot boxes". The definition of loot boxes are items in video games that players can buy with real-world money, the loot boxes, when opened contain randomised rewards. The rise of microtransactions has led to concerns for vulnerable users overspending with currently no regulation or consumer protection. The chance-based nature of loot boxes has arisen discussion regarding the similarities of loot boxes and gambling. The apparent similarity is that the player is risking the loss of real-world money for a small chance of getting a valuable reward.

The thesis aims to research video game players motivations, experiences and gameplay habits, to establish guidelines for loot box systems later. Empirical methods, such as surveys and semi-structured interviews, were used to collect data.

Established theories Hedonic motivation-system acceptance model (HMSAM) and Theory of planned behaviour (TPB) were used as a theoretical framework to understand the collected data. The results of the empirical data display negative attitudes towards loot boxes. Themes were made to make sense of the attitudes and provide an understanding of the negative attitudes. A design solution in the form of guidelines was developed based on the themes. The guidelines were implemented in a design prototype and thereafter evaluated by participants with a user experience questionnaire. The results of the user evaluation illustrate a positive attitude towards the design, but not convincing enough. Participants raised concerns about a few features of the design and had difficulties to see the features be implemented in the future. Therefore, arguments for further research of the guidelines were established.

Key words:

Microtransactions; Loot box; Gambling; design guidelines; behavior intentions.

(3)

Abstrakt

Alternativ för köp i videospel kallas mikrotransaktioner och har blivit en vanlig inkomstkälla för moderna videospel. Mikrotransaktioner avser små betalningar inuti videospel för att skaffa spelinnehåll. Fenomenet med mikrotransaktioner är ofta relaterat till "loot boxes". Definitionen för loot boxes är föremål i videospel som spelare kan köpa med riktiga pengar. Loot boxes, när de öppnas innehåller slumpmässiga belöningar. Ökningen av mikrotransaktioner har lett till oro för utsatta användare som överkonsumerar med för närvarande ingen lagstiftning eller konsumentskydd för mikrotransaktioner. Den slumpmässiga karaktären hos loot boxes har lett till diskussioner om likheterna med loot boxes och gambling. Den uppenbara likheten är att spelaren riskerar förlusten av riktiga pengar för en liten chans att få en värdefull belöning.

Denna uppsats syftar till att undersöka videospelspelares motivationer, erfarenheter och spelvanor, för att senare fastställa riktlinjer för loot box-system. Empiriska metoder, såsom enkät och semistrukturerade intervjuer, användes för att samla in data. De etablerade teorierna Hedonic motivationssystem acceptance model (HMSAM) och theory of planned behaviour (TPB) användes som ett teoretiskt ramverk för att förstå den insamlade data. Resultaten av empiriska data visar negativa attityder för loot boxes. Teman gjordes för att skapa förståelse till de positiva och negativa attityderna. En designlösning i form av riktlinjer utvecklades utifrån dessa teman. Riktlinjerna implementerades i en designprototyp och utvärderades därefter av deltagare med ett frågeformulär för användarupplevelse. Resultaten från användarutvärderingen illustrerar en positiv inställning till designen, men inte övertygande nog. Deltagarna tar upp bekymmer över några funktioner i designen och hade svårt att se funktionerna implementeras i framtiden. Därför upprättades argument för ytterligare forskning av riktlinjerna.

Nyckelord:

Microtransaktioner; Loot box; Gambling; Design riktlinjer; Beteende avsikt.

(4)

Acknowledgments

We want to thank you all participants of this thesis, without everyone it would not have been possible. A special thanks to our supervisor Mexhid Ferati who have provided many hours of amazing guidance, discussions, and inspiration. Thank you to all our friends and families that have been incredibly supportive during this thesis.

Simon Bergstedt Pontus Skoglund

Kalmar, 11-06-2020 Kalmar, 11-06-2020

(5)

Appendices

Appendix A - Question of the survey Appendix B – Interview questions

Appendix C - Confidentiality Agreement and Interview invitation Appendix D - UEQ questionnaire

Appendix E - Assignments for user evaluation

(7)

1 Introduction

In-game purchase options are becoming increasingly monetised within video games and are referred to as microtransactions (King & Delfabbro 2018a). Zackariasson and Wilson (2010) mentions that the term microtransactions refer to small payments made inside of video games to acquire game content. The phenomenon of microtransactions is often referred to as "loot boxes". Loot boxes are items in video games that players can buy with real-world money, the loot boxes, when opened contain randomised rewards (Zendle & Paul 2019). The chance-based nature of loot boxes has led to a discussion of the apparent similarities between loot boxes and gambling. One potentially suitable comparison would be to that of a roulette wheel where the consumer risks real-world money in return for a small chance of gaining a more valuable reward (ibid). The rewards within loot boxes are usually aesthetic changes to in-game models which are commonly referred to as "cosmetics". Cosmetics are used to change the appearance of the user's avatar or objects for instance weapons.

Some cosmetics are rarer than others, and therefore the prices vary from next to nothing to very valuable (Brooks & Clark 2019). The loot boxes are usually purchased with virtual money (Tomić 2017). The term virtual money is defined as:

"unregulated, digital money, which is issued and controlled by the developers, and the virtual money is used as a payment method in a given virtual community" (Tomić 2017, p. 247).

According to King and Delfabrro (2018a) the rise of microtransactions as an option in video games has led to concerns for vulnerable users overspending. There are currently limited regulatory or consumer protection frameworks for video game monetisation schemes (ibid)

Various regulatory organisations have recently discussed and decided whether they consider loot boxes to be a form of gambling; this resulted in a variety of decisions.

BBC News (2018) reports in an article that Belgium has already taken action against the use of loot boxes. According to the Belgium Gambling Commission, video game loot boxes are in violation of gambling legislation. Failure to comply with this legislation outlined by Belgium may lead to a fine of €800,000 and up to five years in prison (ibid). This punishment will be doubled if minors are involved, as the new legislation is intended to focus on minors and the potential impact on their mental health. Meaning that Belgium has ruled that loot boxes can be paid for with real-world money is a form of gambling and have to be removed from video games in Belgium.

Furthermore, the Netherlands have ruled that some loot boxes are a form of gambling.

However, Netherlands only considers loot boxes that contain rewards that can be redeemed for real-world money to be a form of gambling (Eurogamer 2018). France's online gambling authority ARJEL have another view on loot boxes and have ruled out all loot boxes as gambling since there is no financial value to rewards that can be won within loot boxes (PC Games Insider 2018). The Australia Senate has a similar opinion as France, according to an Australian report (Environment and Communications References Committee 2018). The Australian gaming industry's view is that loot boxes can be compared to Kinder Egg and packages of playing cards.

(8)

However, they additionally include counter-arguments that when buying a Kinder Egg, for example, the concept of the value of the content is more comfortable to comprehend. Additionally, the computer game can "win" since the rewards in the loot boxes can be obtained twice, and rewards that were "won" by the user is now considered useless (ibid).

There are similarities between gambling and loot boxes, there is limited regulation or consumer protection framework (King & Delfabbro 2018a). In Sweden, gambling is regulated by The Swedish Gambling Authority (Lotteriinspektionen 2018). The Swedish Gambling Act (Spellagen) contains various requirements for gambling liability measures, including age restriction, credit ban, the possibility of suspension and offer of self-test. Players must be registered, their identity determined and in online games, they should be able to track their transactions through a gambling account. Spellagen additionally has consumer protection in the form of attacking misleading and aggressive marketing, unfair contract terms and direct purchase requests for children (ibid). Since loot boxes are not considered gambling in Sweden, spellagen does not apply to loot boxes. Countries such as China have imposed new regulations on video game companies to show the percentage chance of getting a specific reward. The new regulations are to ensure fairness and protection of consumers who participate in "the distribution of virtual items and additional services through random draws, e.g. loot boxes" (People's Republic of China 2016).

1.1 Background

The culture of buying video games today have changed over the years. Traditionally, video games have been sold as inclusive products, meaning once the consumer purchases the video game, they own or have access to all of the content the game provides. Today, this is known as a "Premium model" for selling video games. There has been an increasing trend towards creating "Freemium Model" video games, meaning the video game itself is free-to-play (FTP) but features or content within the video game is locked behind microtransactions. The freemium model has changed the video-game industry as a whole (Krook 2017) and is becoming increasingly commonplace in the gaming industry. Downloadable content (DLC) is additional content created for an already released video game; it is part of both freemium and premium businesses and is a form of video game monetisation. DLC is usually in the form of additional levels, maps, new items, character roles, aesthetic features and game modes (LifeWire 2018).

There has recently been an emergence of a new hybrid model, a combination of the premium model and the freemium model. The hybrid model requires the player to buy a game for full price while at the same time requiring microtransactions to access all features within the game. The model functions because the features previously mentioned are required to be seen as a relevant player within video games (Zackariasson & Wilson 2010). The most commonly disputed products purchasable by microtransactions are loot boxes. An example of a freemium model with microtransaction is the first-person shooter video game Counter-Strike: Global Offensive. Users can pay $2.49 to open a 'loot box'. The problem with loot boxes is the rewards are randomly generated, and some rewards are more desirable or valuable than others (Brooks & Clark 2019).

(9)

Microtransactions are highly profitable, with some estimates stating that microtransactions generated up to $30 billion for the video game industry in 2018 alone, with the upward trajectory microtransactions will reach $50 billion by 2022 (Juniper Research 2018).

Previous studies have been done to explore potential social responsibility measures for video game monetisation for further discussion and developments in this area (King & Delfabbro 2018a). King and Delfabbro discuss the current debate about loot boxes and its similarities to electronic gambling machines. They try to implement social responsibility principles and research in the field of gambling studies. In the study, they add some of the lessons and experiences from the gambling field to develop a preliminary blueprint for social responsibility measures for video game monetisation schemes. King and Delfabbro outlines four different areas, firstly "game design and in-game purchasing systems". They look over some of the features or characteristics of video games and the in-game purchasing systems, to reduce the likelihood that players will overspend in the game systems. Secondly "transparency and accuracy of game design and features". They look at the way the video games describe or convey information about the nature of its systems to the player. Thirdly

"broad consumer protection measures" where King and Delfabbro looked over what consumer protection measures that can be implemented in video games. Lastly

"consumer information and industry accountability" where they look at what video game companies can do to communicate clear and accurate information about its products to its consumers. However, the study was not based on users, and there have been no attempts to understand the users. Although they mention that the next step of the study would be to involve users (King & Delfabbro 2018a).

Xiao and Henderson (2019) performed a study that provides further game examples to illustrate, critique and extend the proposed measures by King and Delfabbro (2018a). Xiao and Henderson acknowledge the numerous practical solutions proposed by King and Delfabbro (2018a) can be immediately applied by game designers wishing to minimise the potential risk of abusing the user's psychology.

Xiao and Henderson (2019) acknowledge a widespread adoption would require proactivity from across the industry, and given the gaming industry's economic interests, it would not be beneficial. Xiao and Henderson (2019) encourages further research on actual tests on the blueprints that could be applied by video game companies, academics, regulators and users.

The mentioned studies above did not involve users in the process of making a blueprint. The approach of this thesis will take the perspective from the end-user and involve them in the process, and look at the possibility to accommodate previous research (Xiao & Henderson 2019; King & Delfabbro 2018a) in the thesis. User- centred design (UCD) is a broad term to describe a design process in which end-users influence how the design takes shape (Abras, Maloney-Krichmar & Preece 2004).

Keinonen (2008) mentions why it is useful for the design to add human needs and desires since it can help to pursue genuine and real user values.

(10)

1.2 Problematization

Microtransactions have been implemented in video games as a revenue model, in the form of in-game purchases, e.g. loot boxes (Zackariasson & Wilson 2010). Loot boxes are purchased with virtual money in a closed system. The virtual money is bought with real-world money, and the virtual money cannot be refunded. The usage of virtual money serves to blur the real cost of microtransactions (Tomić 2017). The concerns of loot boxes are that the rewards are randomised, and players are risking the loss of real-life money for an valuable item (Zendle & Paul 2019). It has led to a discussion of the apparent similarities between loot boxes and gambling (ibid). Given the reasonably strong parallels between loot boxes and gambling, the question that has to be asked is what motivates users to accept and interact with the revenue model.

Various regulatory organisations have discussed if loot boxes are a form of gambling, which resulted in a variety settlement (BBC News 2018). Although there are similarities between gambling and loot boxes, there is limited regulation or consumer protection framework (King & Delfabbro 2018a). Previous studies have been made to explore potential social responsibility measures for loot box systems, but with no participation of video game players. By not including participants, their perspective and experiences are left out.

1.3 Aim

What this study aims to research is video game players' experience, habits and motivations about loot boxes with a qualitative and quantitative user-centered design approach, to gain understanding and empathy of participants. With the hybrid and freemium model video games with the usage of loot box systems as a monetization will be researched. A design solution can then be developed in an attempt to add consumer protection to the video game player's interaction with loot box systems.

1.3.1 Reserch Question(s)

RQ1: What are motivations, experiences and gameplay habits of video game players towards loot boxes?

RQ2: What guidelines can improve the ethical aspects of loot box systems?

1.4 Limitation

Only participants with previous knowledge and understanding of loot boxes were included in this study. The qualitative interviews were done with video game players, having different relationships with loot boxes, e.g. users with high and low interaction in loot boxes. The study is limited for its participants; only individuals age 18 or older could participate in the study. Reasons for only including adult participants would be because of ethical reasons if adolescence were to be included. Adolescents are a part of the video game community, and this thesis will not obtain their perspective. The design prototype will be limited to a specific game. There will be a generalisation to video games monetisation revenue during the collection of empirical data. The guidelines will be implemented in a specific loot box system in a video game, by the

(11)

choice of the authors. Important to note that this is a qualitative thesis, and therefore the findings cannot be generalised.

(12)

2 Theoretical framework

Chosen theories of this thesis will be presented below. These theories are believed to be relevant for the thesis and give a good understanding of participants behaviour and motivations. The factors that are relevant to the chosen theories, and the data from the empirical data collection will then be used to process a design solution as the final result.

2.1 Theory of planned behaviour

Although loot boxes have been linked to a range of psychological problems (King &

Delfabbro 2018b; Zendle & Paul 2019; Drummond & Sauer 2018), there have been few attempts to conceptualise problems with loot boxes using established psychological theory. The purpose of the thesis is to examine motivations, attitudes and social norms regarding loot boxes using the theory of planned behaviour (TPB).

TPB is a model widely applied to explain behavioural intentions by measuring the attitude toward the behaviour and subjective norm (Azjen 1991).

2.1.1 Attitude towards the behaviour

TPB states that behaviour is directly influenced by behavioural intentions (see Figure 2.1). Behavioural intentions are shaped by three sets of considerations (Azjen 1991).

Firstly the outcome of the behaviour and evaluations of the outcome that produce an

"attitude towards the behaviour". Attitude refers to the overall positive or negative evaluation of a particular behaviour. According to Azjen:

"we learn to favor behaviors we believe have largely desirable consequences and we form unfavorable attitudes toward behaviors we associate with mostly undesirable consequences'' (Azjen 1991, p. 191).

E.g. having a positive attitude about behaviour such as playing video games means that persons favour that behaviour.

2.1.2 Subjetive norm

Secondly, the beliefs from the individual about expectations of others about the behaviour, leading to the individual's motivation to comply with the expectations

``subjective norm`` (see Figure 2.1). Haagsma et al. (2012) mentions that subjective norm refers to perceived expectations of others to perform (or not to perform) a particular behaviour. Preece (2001) mentions that when it produces a sense of recognition from others, thus positively affecting the user's attitude to using the service. An example of the subjective norm would be if family and friends would be positive or negative towards a person behaviour such as playing video games.

2.1.3 Actual behaviour control

Thirdly "actual behaviour control" states that individuals have beliefs about the level of individual control over the specific behaviour (see Figure 2.1). According to Azjen (1991) "actual behaviour control" allows for external factors to affect the individual immediate control. It is leading to influence behaviour. External factors are factors that the individual cannot impact about the behaviour, an example of this would be opening times at the gym that could impact the control for the individual.

(13)

Figure 2.1: A self-produced model of the theory of planned behaviour (based on Azjen 1991).

A study by Wu and Tang (2012) investigates the theory of planned behaviour concerning problem gambling. Wu and Tang (2012) reports that attitudes, subjective norm and actual behaviour control were significantly related to gambling intentions, where behaviour intention and actual behaviour control were significantly related to problem gambling symptoms. However, Wu and Tang (2012) focuses on gambling and not specifically on loot boxes and video games.

2.2 Hedonic-Motivation Systems Adoption Model (HMSAM)

Given that there seem to be reasonably strong parallels between loot boxes and gambling, the question the authors have to ask is what motivates individuals to accept and interact with the revenue model. To answer this, the authors will have to touch on some underlying psychological mechanisms which may give insight into what drives player behaviour. The technology acceptance model (TAM) is a theory that aims to define what causes individuals to use an information system (Davis 1989).

TAM focuses on utilitarian-motivation systems of an individual. The behavioural intention to use is caused by extrinsic motivations and therefore implying the system has to provide external benefits to the user. TAM fails to include hedonic-motivation systems (HMS); examples of such can be video games, applications or video streaming services (Lowry et al., 2013). The behavioural intention of an HMS is caused by internal motivation and experiences that are desired by individuals. The hedonic-motivation adoption model (HMSAM) has been suggested by Lowry et al.

(2013) (see Figure. 2.2), as an extension of the TAM model.

According to Lowry et al. (2013), hedonic-motivation systems are, in essence, to fulfil users' intrinsic motivations. The use of HMS is primarily used for different feelings of pleasure than for productivity. HMS such as video games, social network sites and virtual worlds can create deep immersion and devotion. Users that devote their time within HMS (will focus on video games in this thesis) are doing it to gain intrinsic rewards and not for any external reward that might be given. What users desire the most is the process within the game or the experience of use Lowry et al. (2013).

(14)

The contribution of HMSAM, according to Lowry et al. (2013), is to give better empirical support to predict behavioural intention to use (BIU) in HMS. This theoretical model was found to be a better predictor than TAM. Lowry et al. (2013) provides an improved intrinsic motivation model because it shows that joy, curiosity, control and immersion (see Figure 2.2) do not need to emerge together at the same time as well as not have the same causal mechanisms. The following subsections will describe HMSAM extensively.

Figure 2.2: A self-produced model of HMSAM (based on Lowry et al., 2013).

2.2.1 Perceived ease of use

Lowry et al. (2013) proposes in their study perceived ease of use (PEOU) will increase curiosity. An example is if a video game has a low level of PEOU, the chance of the player to become frustrated, uninterested and stop spending time on the game. It is causing a loss of curiosity. If the PEOU is high, the player has the chance to explore the game and create more possibilities. Lowry et al. (2013) proposes that PEOU makes the user feel in more control if the user feels that a system is easier to use. As Figure 2.2 shows, PEOU has no direct prediction of behavioral intention to use (BIU).

Instead, it is fully mediated by perceived usefulness (PU), joy and curiosity (Lowry et al. 2013).

2.2.2 Curiosity

According to Agarwal and Karahanna (2000), a player's cognitive and sensory stimuli are triggered by the encounter of an interaction. In video games, there are audio and visual effects that arise from the virtual world (ibid). Loot boxes often have various effects that create curiosity with visual effects and delay that creates suspense and excitement for the user. According to Lowry et al. (2013) the factor curiosity is a vital motivational state that makes individuals want to explore more and leads to further engagement, e.g. in video games. They add that joy does not have to be included to increase BIU (see Figure 2.2).

(15)

2.2.3 Perceived usefulness

Lowry et al. (2013) mentions that PU in traditional theories such as TAM is not relevant in HMS; in this context, PU is insignificant. Even if PU is insignificant in HMS, Lowry et al. (2013) kept PU in their theory model but saw it as a more contextual factor (e.g. useful for having fun) (see Figure. 2.2).

2.2.4 Joy

The user experience can be heightened with joy, and it increases the ease of use of activities involving increased enjoyment as an experience (Agarwal & Karahanna 2000). Lowry et al. (2013) states if video game player's joy is heightened during the time they play, the player will be more immersed at the specific period of enjoyment than during periods of less enjoyed gameplay. Lowry et al. (2013) further state that joy does not have to be included to increase BIU (see Figure. 2.2).

2.2.5 Control

Agarwal and Karahanna (2000) describes control as what the player feels, i.e. being in charge of a specific segment and being able to cancel an interaction and make choices of what to do. Control is the weakness of loot boxes. Players can control if they want to open the loot box or not. The loot box system is designed with rewards;

there is no control. Lowry et al. (2013) argue that control is a vital element to gain the state of immersion (see Figure 2.2). Lowry et al. (2013) continues by explaining that these states are more likely to occur when a user feels a certain level of control during their experience.

2.2.6 Immersion

According to Brown and Cairns (2004), immersion is a vital factor for enjoyment within video games and is an outcome of good game experience. Immersion does not necessarily equal an enjoyable experience, it can create unattractive gameplay experience, but neither the less it is a powerful game experience. There are three levels of immersion: engagement, engrossment and total immersion (ibid). The first level of immersion is created if there is an interest. The second level of immersion is engrossment, and this level is reached if a player's emotions are affected by the game.

Total immersion is the final stage and is a combination of the two previous levels of immersion (ibid). Lowry et al. (2013) says that immersion has reached certain stages of attention, such as control, curiosity, and joy.

2.3 Theoretical application

The presented theories were utilised to understand the motivation and experiences for users. HMSAM is used to obtain an understanding of the context to use an HMS system (in this case, loot boxes in video games) for users. HMSAM has a focus on intrinsic motivations in HMS. TPB is used to consider motivational factors, e.g. social motivation, attitudes towards loot boxes and actual behaviour control. HMSAM and TPB are used as a complement to create the questions for the survey and the interview questionnaire. Both theories are being considered when analysing the empirical data.

(16)

3 Methodology

The following chapter describes the chosen methods of this thesis and the analysis of the empirical data. Presentation of the chosen design approach, reliability and validity and ethical aspects are included.

3.1 User-centered design

'User-centered design' (UCD) is a broad term to describe several design processes in which end-users influence how a design takes shape. It is both a broad philosophy and has a variety of methods (Abras, Maloney-Krichmar & Preece 2004). There is a spectrum of ways in which users are involved in UCD. Users must be involved one way or another in the process. UCD focuses on designing for users by involving users in the design process (ibid). The advantage of a UCD approach is the more in-depth understanding of psychological and social factors of the user.

Abras et al. (2014) mentions that there are two different spectrums of involving users.

A light way of involving users is by consulting users about their needs, observing and participating in usability testing. The other spectrum can be intensive, with users being partners in the design process. In this thesis, there has been a lighter way approach of UCD, and users have been more of consultants rather than being partners in the design process. The users were involved at the beginning of the design process, in the interviews. Users then got consulted about the collected data and that their needs were fulfilled. User evaluation was the last involvement of users in this thesis, and the evaluation helped to identify measurable usability and user-experience. Abras et al. (2014) mentions that it is only through collected feedback in an interactive iterative process involving users that products can be refined. Furthermore, users help with their knowledge to identify the flaws of the design prototype in this thesis by discussing it after the evaluation.

3.2 Survey

A survey was conducted early in the thesis as a pilot study and is a common method to collect qualitative data. The purpose of the survey was to understand demographics, experiences, habits and attitudes towards loot boxes. The questions that were chosen to use in the survey were multiple choice questions as opposed to allowing the participants to have open-ended responses. There was a total of two open questions in the survey that were answered with text. As the answers were read through, sentences were colour coded and made into categories. Purpose of the colour coding was to see if participants' answers were connected and similar. The categories were named based on the empirical data, and this was an iterative process between the colour codes and categories before the final categories were complete.

The data will be used for quantitative measuring later in the thesis. Open ended essay type answers while useful for further understanding or context, is aimed to assess in this thesis. Surveys both have advantages and disadvantages, in the sense that is best used to gather descriptive information. For instance, demographic data, which can be used to identify potential relationships among characteristics (Goodwin 2009). The disadvantage with surveys is that it rarely is useful in explaining why those relationships exist, leaving it open for interpretations of the relationships among the characteristics. Interviews were conducted to explain these relations, as a counter for

(17)

the disadvantages with surveys. To use multiple methods to collect and analyse data is called methodological triangulation (Arvola 2014).

Participants are often unwilling to complete a survey over 20 questions without some substantial reward (Goodwin 2009). Therefore, the survey had to be limited in the number of questions it contained. The questionnaire was made to consist of questions that answered characteristics, actions, experiences, opinions, motivations, attitudes and social norms. The alternative of the answers was made as complete as possible, and the options for any single-answer question mutually exclusive (ibid). Therefore, the participants knew what to pick and there will be no confusion and minimising errors. If participants cannot answer questions there will be an inclusion of options such as others, prefer not to say and do not know. The Likert scale has an odd number of options which allows for neutral answers. The survey was made to feel linear, and the sequence of questions was thought off when it was made, and the format of questions, i.e. Likert scale, open-ended questions and multiple-choice questions were mixed up to minimise the risk of participants not paying attention to what they are selecting (ibid).

The survey contained 15 questions; 42 participants of the video game community answered the survey (see Appendix A). The first section of the survey is focused on demographics and habits on gameplay and loot boxes. Questions such as time spent on video games, if participants ever purchased a loot box and their spending habits per month as well as reasons behind their habits of loot boxes. Motivations and believed harm were asked as open questions to give the participant a chance to express their opinion on loot boxes and microtransactions. The two open questions were analysed by colour-coding. Questions regarding opinion, acceptance and experiences were asked; this is in addition to work close to the HMSAM and TPB theories.

The survey was conducted via different forums. The forums chosen were the Facebook group named “Counter-Strike Global Offensive (Sweden)” as well as via select servers on Discord. The questionnaire was open for two days before it was made private to the public.

3.2.1 Selection of participants

Arvola (2014) mentions when recruiting participants is having critical participants.

Which means participants that have something to tell the designers about the problem area and are often the ones most affected by the problem. The selection of participants was video game players, this group was essential to recruit due to the previous experience with loot boxes. It meant that the perspective on loot boxes was explored and what it takes to meet the possible requirements on improving the process and to take ethics into account of loot boxes. The process of recruiting participants falls back on convenience sampling, and the selection of participants is applied to the survey and interviews.

3.3 Qualitative semi-structured interviews

The second data collection method was conducted through qualitative interviews.

Interviews is a qualitative method and is utilised for making classifications of meaning-bearing phenomena (Johansson 2011). Semi-structured interviews can give

(18)

a more in-depth understanding than surveys, interviews questions should, therefore, be open-ended. Having a semi-structured interview means that the person being interviewed can explain their reasoning through open questions. Leading questions should be avoided because it can mislead the participants and questions are then answered with expected answers (Goodwin 2009).

There is a risk that the answers may be too shallow, and it is not possible to go into depth, leading to not understanding the participant or having to interpret the participant that can lead to errors (Arvola 2014). In addition to the value of qualitative methods is to gain a good understanding of the user's attitudes, perceptions, experiences, intentions and needs. Wright and McCarthy (2004) explains semi- structured interviews are best suited to gain empathy and understanding between participants and designers since it is based on dialogues. Semi-structured interviews encourage follow up questions and discussion to arise. Therefore, semi-structured interviews were chosen to gain a deeper understanding. Löwgren and Stolterman (2004) recommends using the technique "Why-Why-Why", which builds on the original question one asks. Asking why helps with the flow of the interview and enables for a deeper understanding of the participant. Otherwise, answers can be shallow and it is hard to understand the actual reasoning behind the answer. It is essential not to force an answer but sometimes push the participant forward with the

"Why-Why-Why" technique (ibid).

3.3.1 Selection of participants

The selection of participants for the interviews was based on the same practise as in point 3.2.1.

3.3.2 Interview questions

The questions were based on the survey results, the open-ended questions in need of more context and the research questions. Secondly, the question formulation was based on the theoretical framework; HMSAM and TPB to further understand the user's context, attitudes and motivations. The interview questions' purpose was to be more design-oriented, to achieve a more UCD approach e.g by involving the participants' opinions and ideas to the design.

Questions regarding motivation and experiences from the survey were used in the interview to understand the context and relations. Asking why participants feel and think helps to understand and gain further insights. Questions regarding motivations and what participants believe is crucial for them that makes them engage in loot boxes to further deeper the sympathy and understanding for the users. The final section of the questions were the design questions. How they believe loot boxes are designed, what can be designed differently, if participants believe loot boxes are considered gambling and if so, why. The final question is a workshop inspired question and asks about the participant's future demands how they want microtransactions to be designed (see interview questions in Appendix B).

3.3.3 Conducting the interviews

Each interview attempted to be held based on the same arrangement, and it always began with a briefing explaining the arrangement and context. All four interviews

(19)

were conducted online through online communication media. The participants were asked if they accepted being recorded. The request was made before the informed consent was read out; however, the participants were reluctant to be recorded, and therefore none of the interviews were recorded. The participants signed the confidentiality agreement before the interview started (see Appendix C). On average, each interview took roughly 35-40 minutes and got transcribed; the questions were open-ended and allowed the participants to speak freely and go deep into their answers (see Appendix B). In the case of unclear or doubtful answers, the "Why- Why-Why" technique was used, the question was repeated or rephrased to ensure that the interpretation of the question was correct. It was essential to keep a relaxed atmosphere, meaning that no answer was forced. There were a couple of occasions that felt troublesome or difficult for the participant, and they answered them to the best of their abilities. An example of this occasion was a question that was focused on video game players' mindset "What do you think could increase a player's mindset of consequences when it comes to virtual money and loot boxes?". Participants were not forced to elaborate on the questions, and the interview moved on in the questionnaire. After the conducted interviews, post-it notes were categorized, the final themes could be generated together with the survey post-it notes.

3.4 Affinity diagram

There are different ways to process data. To analyse qualitative data, one can use the affinity diagram method, which is a way of categorising or thematising data. After data collection, there is a wealth of unstructured data (Arvola 2014). The first step of an affinity diagram is to read through the data and highlight relevant data. The highlighted data of the survey and interviews were made into post-it-notes. The post- it-notes were used to be categorised based on their underlying collective meaning.

Arvola (2014) mentions that it is good to work this process in silence as the research group creates relevant categories. A problem that Arvola (2014) mentions is that the data often does not make any significant use. The data need to be put in a context where it can create categories and themes that bind them together, which gives an insight picture of the participant's thoughts (ibid). Affinity diagram was used because of the ability to analyse data and structuring it at the same time.

3.5 Methods of the prototype

A clarification of the methods, chosen for the prototype design process, will be presented in this chapter.

3.5.1 Brainstorming

Brainstorming is a method that takes place in a group where the method is used to generate ideas together. A brainstorm should occur after all qualitative data is collected and compiled. The advantage of a brainstorm is the team gets several ideas in a short time. The ideas that are produced must not be criticised, and all ideas should be encouraged (Löwgren & Stolterman, 2004; Arvola 2014).

Creative thinking is additionally expanded since, with this method, it is possible to combine and improve each other's ideas. It is therefore vital that no idea is criticised since there is always room for improvement. Initially, in this brainstorm, the problems were introduced; the idea was then not to lead to a specific solution. Löwgren and

(20)

Stolterman (2004) says that the reason for not leading to a specific solution is that ambiguity open ups for possibility. The possibility is to generate several ideas instead of locking into a particular idea. The brainstorm was based on the compiled data from the affinity diagram; the ideas were then later used for design solutions.

3.5.2 Interface sketches

Interface sketching is a technique carried out with paper and pen for sketching out a layout of interfaces and illustrating the supposed system (Arvola 2014). Interface sketches were done separately and then discussed together. Löwgren and Stolterman (2004) mentions that interface sketches stimulate creative thinking by opening up new possibilities through combinations of ideas. By sketching individual interface sketches allowed developing something different from each other and then developing a combination of ideas that were formed into a design.

3.5.3 User evaluation

User experience (UX) cannot be evaluated with the traditional toolkits due to the subjective nature of UX. Therefore, objective measures are not reliable, rather the need to understand the user's feelings, motivation, expectations and how these factors affect the design (Obrist & Roto 2009). UX evaluation methods need to provide answers about the user's feelings of a system, and this is the most common requirement, but different methods may be varied depending on the case, and UX is very context-dependent (ibid).

User experience questionnaire (UEQ) is the evaluation method of choice in this thesis (see Appendix D). UEQ allows for immediate measurement of UX and usability (Schrepp, Hinderks & Thomaschewski 2014), and the measurement is the reason UEQ is chosen to evaluate the prototype. The UEQ typically contains six scales:

Attractiveness, Perspicuity, Efficiency, Dependability, Stimulation and Novelty.

Attractiveness acts as a valence dimension, and then there are two quality aspects, Pragmatic (goal-oriented) and Hedonic (not goal-directed). Perspicuity, efficiency, and dependability belong to the pragmatic aspects, and Stimulation and Novelty belongs to the hedonic aspect (Schrepp, Hinderks & Thomaschewski 2014). As stated above, UX is context-dependent, and the UEQ of this thesis was adapted to the context. Therefore, one scale was removed from the UEQ, dependability (ibid).

The UEQ were answered by participants, after the participants had completed the questionnaire, the data were summarised. A clarification on the data, the average score was obtained for each question for all aspects. The average score presented what overall score for each question and what was positive, neutral or negative. For an overall look of the data, a cumulative average was obtained to see scores for every aspect on their own. With this data, it could make conclusions of the result and see how the participants experienced the prototype. Feedback was received in the form of a discussion after the evaluation was completed.

3.5.3.1 Selection of participant

The selection of participants was based on individuals previous experience with the specific computer game the prototype would assimilate, Counter-Strike: Global Offensive. There was a total of four participants in the evaluation process. Where only

(21)

one of the participants participated in the interviews. Therefore, a presentation of the confidentiality agreement was needed and agreed on. The new participants had no information about what the thesis was about, a briefing of the thesis was required to give context and explain their role in the evaluation. The process of recruiting participants for the UEQ falls back on convenience sampling.

3.5.3.2 Conducting user evaluation

All evaluations were conducted remotely. During the evaluation, participants were told a scenario and then a follow-up assignment to the scenario, a total of 5 assignments (see Appendix E). Each evaluation, including the answering UEQ, took 10-15 minutes. After participants completed the assignments, a link to the electronic UEQ was given to the participants. Participants additionally had a chance to give feedback and discuss the prototype.

3.6

Reliability & Validity

Qualitative research is interpretive research. The inquirer is involved in a sustained and intensive experience with participants (Locke, Spirduso, & Silverman 2013). The concern with qualitative research is the interpretive nature and concern is inquirers reflexively shape their interpretations due to biases, values and personal background (ibid). The authors have much experience with video games and microtransaction, the rise and the current state of the phenomenon. Creswell (2018) mentions that previous experiences can potentially shape the interpretations the researchers make during the study; for instance, researchers lean toward specific themes or actively look for evidence to support their position. The prior relationship between the researchers and participants in the research team made the interviews easier since they felt comfortable.

Interpretation plays a significant factor in qualitative methods. By having an objective starting point, reliability is increased. In order to open a good validity, the interpretation made of qualitative data must, to some extent be in agreement with the participant's interpretation.

To minimise the risk of the authors previous knowledge and experience to shape the study, a summary of the interviews was sent to the participants. The participants gave their feedback on the interpretations. Validity is one of the strengths in qualitative research and is based on determining whether the results are accurate from the researcher, the participants and the reader (Creswell 2018). Follow-up questions were asked to ensure that the answers were more transparent, vibrant and to minimise interpretations. The validity is found in the questions, how they are designed and what agenda it has. What is meant to be investigated should be answered through the questions during the interview (Johansson 2011). For instance, the researchers made sure not to ask any leading questions to avoid a bias in the participant's answers. To ensure that none of the questions was misunderstood, and that the questions were comprehended in a correct way. The interview was only meant for participants that have made an in-game purchase of a loot box. The participants had experience and could, therefore, give valid answers. The questions were formulated so the participants comprehended the questions correctly. These questions further add the reliability and validity of the qualitative research.

(22)

Creswell (2018) mentions that in order to prove the reliability of qualitative research, the researchers must fully document each step of their case study. The documentation must contain as many steps as possible of the process, and in such detail that the reader can be able to follow how the data was retrieved. The documentation of the thesis has been done and is shown as Figures, Tables and Appendices. The survey captured patterns to support the interviews in the way that it helped to investigate what was intended to, thus increasing the validity through method triangulation. The reliability was increased in the way that the results from qualitative methods supported each other.

3.7

Ethical aspects

In consideration to the Swedish research councils’ principles (Good research practice 2017), this thesis has at the collection of empirical data followed the council's information requirement, consent requirement, utilisation requirement and confidential requirement.

The information requirement states that researchers should inform participants about their role they will have within the research as well as on what terms, regarding the participation. Consent requirement means that participants always have the right to be a part of the research or not. Additionally have the right to determine their participation. Utilisation requirement comprises the information collected is used only for the stated research purpose. The confidentiality requirement means that the participant will not be able to be identified by outsiders taking part in the research study (Good research practice 2017).

Participants received an invitation before the conducted interview. The invitation included information about the thesis, and the purpose of the interview (see Appendix C). With the invitation, a confidentiality agreement was attached in the e-mail and needed to be signed by the participant (see Appendix C). Authors added in the agreement that only audio would be recorded, but it would be optional because the studied area may be sensitive to some. If participants did not want their information recorded, it would be fully respected. The reason behind making the audio recording optional is not only due to the sensitive aspect of the study, in addition because it may lead the participant to feel uneasy and would not act naturally (Arvola 2014).

The potential adverse effects and risks that may arise in connection with this thesis have been during interviews and user evaluation (ALLEA 2013). The interviews were conducted online, and it still had an emphasis on creating an environment that made the participants feel comfortable, and participants dared to express themselves and that no answers were considered inadequate. The participants never should feel compelled to answer the questions either, which admittedly is hard to tell since the interview was conducted online, and it had no visual feedback in the form of a webcam.

(23)

4 Empirical results and thematic analysis

In this chapter, the authors present the result of the empirical data that has been collected by a quantitative survey and qualitative interviews.

4.1

Survey results

The first questions of the survey regarded some demographics (see Figure 4.1 and Figure 4.2). Figure 4.1 is illustrating the gender identification of the participants and the majority are 92,9% male.

Figure 4.1: Gender identification of participants

Figure 4.2 illustrates the result of the age groups in the survey. The biggest age group was 25-29 years old participants with 40,5% and the second-largest age group were 30-40 years old participants with 35,7%.

Figure 4.2: Participants age of the survey

90,5% of the participants said they have purchased microtransactions before (see Figure 4.3). 28 participants answered they spent €0-10 every month on microtransactions or loot boxes (see Figure 4.4). There is no connection for the amount of hours participants play every month nor age, and it varied a lot.

(24)

Figure 4.3: Purchase of loot boxes

Four participants spent more money per month than others, a total of two participants spent €21-50 a month and two participants €51 or more, these participants feel enjoyment when spending on microtransaction or loot boxes. The four participants have a positive view of microtransactions/loot boxes and spend a high amount of hours every week on their gameplay. Age factor has no connection to spending

behaviour at all

Figure 4.4: Pie chart of monthly spending.

Table 4.1 displays participants' opinions regarding the microtransactions (loot boxes were specified in the survey). 25 participants had in common a negative opinion out of 42. 16 out of 25 of these participants believe microtransactions are socially accepted or neutral in the video game community. 12 participants responded it is accepted within their group of friends to purchase it, while 8 participants answered they do not know. 18 out of 42 participants think it is socially acceptable, 24 participants see problems regarding loot boxes and microtransactions towards the video game community.

(25)

Table 4.1: Table of participants overall opinion of microtransactions.

The two open-ended questions of the survey were colour coded to create categories to be able to see patterns of the participant's answers. Question number eight (see appendix A) display participants believe the motivation for purchasing microtransactions is feeling unique with cosmetic items, time-saving, and the experience of opening loot boxes (see table 4.2). Question nine (see appendix A) displays three believed harmful effects loot boxes have on video game players and the community: Overspending, gambling-like behaviour and it ruins the game as a whole (see table 4.2).

Table 4.2: Result of the two open ended questions in the survey.

There are three major experiences participants believe video game players experience when purchasing loot boxes. 27 of the participants chose thrill, 24 participants chose curiosity and 22 participants chose joy (see figure 4.5)

(26)

Figure 4.5: Believed experiences when purchasing microtransactions.

The final question of the survey was regarding what participants believe are appealing with loot boxes, 81% of participants answered the rewards, 11,9% of participants answered the experience (see figure 4.6).

Figure 4.6: Result of what participants believe are appealing with loot boxes.

(27)

4.2

Interview results

The participants' answers of the interview were written down on post-it notes and categorized in patterns. These patterns were later made into themes as Figure 4.7 illustrates.

Figure 4.7: Findings and themes of the interviews.

The themes from the interview results are explained extensively below.

4.2.1

The participants acceptance of loot boxes and motivations

All participants say that loot boxes have become a norm in video games, especially free-to-play games. Loot boxes as a revenue model was not the norm before, with time it has become socially accepted. Participants feel it has become a norm since most users open loot boxes.

All participants believed that being aesthetically appealing is a motivational factor for many users. It is nice to stand out from the crowd and feel unique but additionally be a part of the community. Three out of four participants say it feels good to have certain cosmetics that not everyone has; it shows that users have played for a long time and not be seen as a beginner player. One participant said that personally, cosmetics is out of interest.

4.2.2

Participants have a negative view on loot boxes

All participants view loot boxes as something negative for video game players. Two of the participants specify that users lack control of their spendings. Another participant specified it from minors' perspective that it is unhealthy for users under the age of 18 to spend real-world money on loot boxes. Participants think loot boxes are exploiting video game players because users cannot always afford it but engage in loot boxes regardless.

(28)

Participants have a negative view of loot boxes. In the past, participants have engaged with loot boxes. The experience of opening loot boxes was varied, but what was shared between them was the knowledge of the low chance of obtaining a desirable reward. Other experiences varied, one participant did it because it is fun to gamble with a mind of, whatever happens - happens. One participant said experiencing disappointment but liked the thrill of the chance of getting something rare.

4.2.3

Participants concerns about loot boxes

Participants believed that loot boxes do not necessarily have to be a problem if there is a balance. Loot boxes is a big revenue model but as the current state of today with loot boxes, it is problematic. Rewards are hidden behind tedious goals, and instead of reaching them, users pay for it instead.

Virtual money is a problem according to the participants. Participants can personally see the value of virtual money. Participants believe it can cause problems for minors, as minors might not understand the real-world money behind it. One participant gave an example of a child asking their parents for 2500 v-buck (virtual money in the game

"Fortnite"). It does not sound like real-world money, and therefore the value is getting lost in translation. Two participants state an understanding of the virtual money, although the perception can be skewed at times when spending on loot boxes. Because the loot box's price is hidden behind virtual money and not real-world money.

All participants experience that video game companies are making users feel the need to keep spending on loot boxes and microtransactions. Participants mention that video game companies are only focusing on what generates the most money and being selfish. Which is made through daily deals, time-limited offers of certain items, designed to look fun and exciting such as casino themed designs and effects. All participants agree that loot boxes are associated with gambling because of its nature.

Even if users always receive a reward, if it is not a rare one or desirable - users feel that it is a loss.

4.2.4

Participants thoughts of improvement for loot boxes

The need for improvement regarding loot boxes was a common answer from all participants. Increasing transparency was the most common answer, as in showing the actual odds of certain rewards, so users understand what the chances are to obtain rewards. Participants say that virtual money should not exist at all and display the real-world money instead, to understand the value of rewards better. Show users spending history and clarify how much users are spending. Users under the age of 18 should have parental control, spending limits or allowance settings from their parents.

Users above the age of 18 could have their own spending limits to keep track of it better. Users should not be able to obtain the same reward twice from a loot box.

Virtual money should be obtainable through playing the game, in that way users can open loot boxes if users spend enough time in the actual game. The effects of loot boxes should be taken away completely, both visual and audio effects as it can add suspense and thrill in the opening.

(29)

4.3 Merging survey results and interview results

To make sense of the empirical data, cumulative themes were generated and structured from the two results of the chosen methods. This aims to create clarity of the result and ensure that the results from the participants will be used for the planned design prototype. By grouping the different results into color boxes with post-it-notes (see figure 4.8), six themes could be generated and are explained extensively below.

Figure 4.8: The six themes generated by merging the results of the survey and interviews.

(30)

Design solution: Participants want to play the actual game and get rewarded with battle passes and more events from the video game companies. There is a pattern that participants want to see more reliable parental control and can set a limit of how much can be spent inside the game, as in a monthly allowance. Common answer from participants is a change of the nature of loot boxes. Example is that users cannot obtain the same reward twice and the loot boxes only contain a set amount of items. In the end, users will only be able to purchase a set amount of loot boxes before obtaining all items.

Lack of transparency: Participants said that the loot box system needs to be more transparent with their users. Participants want to see video game companies showing the percentage of the odds to obtain specific rewards. As well as show how much the player has spent (i.e. for the past month). To be able to see real-world money, one participant said that showing the real-world money value of the purchased virtual money at all times is believed to help. If the video game companies would be more transparent and provide more feedback to the users, it would be less dishonest.

Gameplay habits and loot box engagement: Loot boxes are experienced sometimes as fun and a sense of thrill when opening them. 90% of the participants have opened a loot box at least once, and the prominent reason is the rewards, specifically cosmetics. 28 of the participants spend between €0-10every month. Amount of gameplay hours have no connection with the amount of real-world money spent every month, except for the few participants that spend the most real-world money every month. The participants that spend a high amount of real-world money tend to spend more hours in video games.

Motivations and experiences of loot boxes: What motivates participants to use loot boxes according to the interviews are the desire to obtain specific rewards that stand out and make them unique. Certain rewards are limited edition, and it can show off how long users have played the game compared to beginner players.

Experience of loot boxes according to the participants varies. Similar answers are that it is fun and a mindset of "whatever happens, happens" as well as participants experience the thrill and enjoyment. One participant said that they do not feel they are losing real-world money.

Unethical design can lead to unhealthy behaviour: Participants said that loot boxes are perceived as casino gambling and are predatory for minors in video games. The transparency is low, and it is suggested from some participants that it is psychologically manipulative. Loot box systems make users feel close to obtaining a desirable rare reward. Virtual money can be considered unethical as well, it is not perceived as of any value, but it is purchased with real-world money, the value gets lost in translation.

Participants open loot boxes even with the knowledge of "loss" (e.g. high chance of not receiving the desired reward). Participants feel the loot boxes are exploitive and can create problematic behaviour for some users. The problematic behaviour may lead to overspending, and minors ask their parents for virtual money or even use their

(31)

credit card on their own. In addition, participants said loot box rewards can be perceived as an investment, but is in most cases, a net loss.

Social norms of loot boxes: Loot boxes are acceptable in moderation, if the game is good the users want to support the developers and therefore buy loot boxes or other microtransactions. Loot boxes are socially accepted and somewhat appealing, and the comment from participants is that everyone has done it. Social aspects, such as being a part of the gaming community, is a motivation for interacting with loot boxes.

Participants have said that cosmetics that make them stand out and are unique is appealing. Being unique and standing out gets approval and recognition from the community.

Faculty of Technology Department of Informatics

The loot box dilemma

A qualitative thesis of video game players

interactions and mitigation of unethical aspects in

loot boxes.

Abstract

Key words:

Abstrakt

Nyckelord:

Acknowledgments

Table of contents

Appendices

1 Introduction

1.1 Background

1.2 Problematization

1.3 Aim

1.4 Limitation

2 Theoretical framework

2.1 Theory of planned behaviour

2.2 Hedonic-Motivation Systems Adoption Model (HMSAM)

2.3 Theoretical application

3 Methodology

3.1 User-centered design

3.2 Survey

3.3 Qualitative semi-structured interviews

3.4 Affinity diagram

3.5 Methods of the prototype

Reliability & Validity

Ethical aspects

4 Empirical results and thematic analysis

Survey results

Interview results

The participants acceptance of loot boxes and motivations

Participants have a negative view on loot boxes

Participants concerns about loot boxes

Participants thoughts of improvement for loot boxes

4.3 Merging survey results and interview results