Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play

April Tyack Bachelor of Mathematics Bachelor of Information Technology (Honours)

Written under the supervision of

Professor Peta Wyeth & Professor Daniel Johnson

Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

School of Electrical Engineering and Computer Science Science and Engineering Faculty Queensland University of Technology 2019

Keywords

Video games; wellbeing; self-determination theory; player experience; mixed methods; need frustration.

Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play i

Publications

The following publications are based on key results from this research:

Tyack, A., Wyeth, P., & Johnson, D. (under review). Restorative Play: Videogames Improve Player Wellbeing After a Need-Frustrating Event. (Based on Study 1; submitted for review on September 21, 2019)

Other publications include:

Tyack, A. (2019). Splendid Isolation: Optimistic Relations Towards Virtual Experience. In Proceedings of DiGRA Australia 2019. [Peer-reviewed extended abstract]

Wyeth, P., Hall, J., Carter, M., Tyack, A., & Altizer, R. (2018). New Research Perspectives on Game Design and Development Education. In Proceedings of the 2018 Annual Symposium on Computer-Human Interaction in Play Companion Extended Abstracts (pp. 703-708). ACM.

Tyack, A., Wyeth, P., & Klarkowski, M. (2018). Video Game Selection Procedures For Experimental Research. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems. ACM.

Tyack, A., & Wyeth, P. (2017). Exploring Relatedness in Single-Player Video Game Play. In Proceedings of the 29th Australian Conference on Computer-Human Interaction (pp. 422-427). ACM.

Tyack, A., & Wyeth, P. (2017). Adapting Epic Theatre Principles for the Design of Games for Learning. In Proceedings of DiGRA 2017.

Tyack, A., Wyeth, P., & Johnson, D. (2016). The Appeal of MOBA Games: What Makes People Start, Stay, and Stop. In Proceedings of the 2016 Annual Symposium on Computer-Human Interaction in Play (pp. 313-325). ACM.

Tyack, A., & Wyeth, P. (2016). Reflective experiences in videogame play. In Proceedings of DiGRA Australia 2016. [Peer-reviewed extended abstract]

ii Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play

Abstract

Understanding the ways that videogame engagement influences player wellbeing merits engagement with the range of potential players, videogames, and social contexts implicated in the player experience (PX). While research showing positive links between videogame play and wellbeing has flourished, the means by which play can improve player wellbeing is less clearly understood. Understanding how play improves wellbeing may benefit the design of interactive technologies for this purpose. Recent work in self- determination theory (SDT) literature has indicated that studying experiences of need frustration – feeling incompetent, controlled, and socially rejected – alongside need satisfaction – experiences of competence, autonomy, and relatedness – may produce new insights into the restorative qualities of videogame play. However, studies examining the ways that both need satisfaction and frustration may influence player wellbeing are yet to emerge in experimental PX research.

Two experimental mixed-methods studies with pre-post measurement were therefore conducted to understand how and to what degree videogame play could improve player wellbeing following a negative experience. In these studies, participants were randomly assigned to engage with a need-satisfying or need-frustrating activity prior to a thirty-minute session of videogame play. Manipulations were focused on competence (Study 1) and autonomy (Study 2) only. Post-experiment interviews aimed to explore the ways that specific experiences of need satisfaction, need frustration, wellbeing, and motivation emerge in and through the player experience.

In general, results showed that participants whose needs were frustrated, rather than satisfied, prior to play experienced greater improvement in need satisfaction, need frustration, short-term wellbeing, and intrinsic motivation during play. Experiences of need satisfaction during play were shown to effectively predict positive short-term wellbeing outcomes and intrinsic motivation; conversely, only competence frustration was shown to consistently predict negative short-term wellbeing outcomes. Although prior PX research had suggested that feelings of immersion (a reduced awareness of the mediated quality of a virtual environment) could moderate the effects of in-game experiences on post-play outcomes, evidence for this position was not observed. Study results similarly contrasted with research indicating that need frustration could influence

Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play iii

intrinsic motivation for a successive activity, although aspects of study design differed. Interview analysis produced elaborated positions on the ways that experiences of competence and autonomy emerge through virtual interactions, and identified three ways that players use videogames to support wellbeing – through self-suspension, wholehearted engagement in play, and as part of an ongoing practice of wellbeing maintenance.

In aggregate, the present work demonstrates that videogame play can reverse the short-term wellbeing deficits associated with need-frustrating events. Moreover, study results establish the utility of need frustration as a means to understand negative experiences in play. These results contribute to SDT and PX literatures alike, and may find further application in the design of videogames aiming to facilitate specific experiential or behavioural outcomes.

iv Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play

Table of Contents

Keywords ...... i Publications ...... ii Abstract ...... iii Table of Contents ...... v List of Figures ...... viii List of Tables ...... ix List of Abbreviations ...... xii Statement of Original Authorship ...... xiii Acknowledgements ...... xiv 1 Introduction ...... 1 1.1 Background ...... 1 1.2 Research Aims ...... 2 1.3 Contributions ...... 3 1.4 Thesis Outline ...... 4 2 Literature Review ...... 5 2.1 Player Experience ...... 5 2.1.1 Short-Term Wellbeing Benefits of Videogame Play ...... 6 2.1.2 Mood Management Theory ...... 7 2.2 Self-Determination Theory ...... 10 2.2.1 Wellbeing in Self-Determination Theory ...... 11 2.2.2 Need Satisfaction in Videogame Play ...... 13 2.2.3 Need Frustration in Videogame Play ...... 15 2.3 Summary and Implications...... 19 3 Research Design and Methodology ...... 21 3.1 Research Structure and Scope ...... 21 3.1.1 Hypotheses and Research Questions ...... 22 3.2 Study 1 Design ...... 23 3.2.1 Measures ...... 24 3.3 Study 2 Design ...... 25 3.3.1 Measures ...... 27 3.4 Approach ...... 29 3.4.1 Procedure ...... 29 3.4.2 Videogame Selection ...... 30

Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play v

3.4.3 Analysis ...... 31 3.4.4 Research Ethics ...... 32 4 Study 1 ...... 35 4.1 Method ...... 36 4.1.1 Recruitment ...... 36 4.1.2 Measures ...... 37 4.1.3 Videogame Stimulus ...... 39 4.1.4 Procedure ...... 39 4.2 Results ...... 41 4.2.1 Assumption Checks ...... 41 4.2.2 Quantitative Analyses ...... 42 4.2.3 H1 – Need Satisfaction and Frustration ...... 44 4.2.4 H2 – Short-Term Wellbeing ...... 46 4.2.5 RQ1 – Intrinsic Motivation ...... 47 4.2.6 H3 – Need Satisfaction and Frustration as Predictors ...... 48 4.2.7 RQ2 – Thematic Analysis ...... 50 4.2.8 Habitualisation of Virtual Activity...... 50 4.2.9 Aligning with Design Intent ...... 52 4.2.10 Play as Self-Abeyance ...... 54 4.2.11 Authentic Self-Engagement ...... 55 4.3 Discussion ...... 56 4.3.1 H1 – Changes in Need Satisfaction ...... 56 4.3.2 H2 – Changes in Short-term Wellbeing ...... 56 4.3.3 RQ1 – Ongoing Effects of Need Frustration ...... 56 4.3.4 H3 – Need Satisfaction and Frustration as Predictors ...... 57 4.3.5 RQ2 – Player Perspectives on Need Satisfaction, Need Frustration, and Wellbeing in PX ...... 57 4.3.6 Competence in the Experimental Play Session ...... 58 4.3.7 Videogame Play in Ordinary Life ...... 59 4.4 Directions ...... 60 5 Study 2 ...... 63 5.1 Method ...... 64 5.1.1 Recruitment ...... 64 5.1.2 Measures ...... 65 5.1.3 Videogame Stimulus ...... 67 5.1.4 Procedure ...... 68 5.2 Results ...... 69 5.2.1 Assumption Checks ...... 69 5.2.2 H1 – Changes in Need Satisfaction and Frustration ...... 73

vi Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play

5.2.3 H2 – Change in Positive Short-term Wellbeing Outcomes and Intrinsic Motivation ...... 74 5.2.4 H2 – Change in Negative Short-term Wellbeing Outcomes ...... 76 5.2.5 H3a – Need Satisfaction as Predictor ...... 77 5.2.6 H3b – Need Frustration as Predictor ...... 80 5.2.7 H4 – Moderation By Immersion ...... 81 5.2.8 RQ1 – Ongoing Effects of Need Frustration ...... 85 5.2.9 RQ2 – Thematic Analysis ...... 86 5.2.10 Autonomy Support in the Player-Game Relation ...... 87 5.2.11 Threats to the Desired Play Situation ...... 90 5.2.12 Wellbeing as Sufficient Functioning ...... 92 5.3 Discussion ...... 95 5.3.1 H1 – Changes in Need Satisfaction and Frustration During Play ...... 95 5.3.2 H2 – Changes in Positive Short-term Wellbeing and Intrinsic Motivation During Play...... 96 5.3.3 H2 – Changes in Negative Short-term Wellbeing During Play . 97 5.3.4 H3 – Need Satisfaction and Frustration as Predictors ...... 97 5.3.5 H4 – Moderation by Immersion ...... 99 5.3.6 RQ1 – Ongoing Effects of Need Frustration ...... 99 5.3.7 RQ2 – Autonomy, Wellbeing, and Motivation in PX ...... 100 5.3.8 Autonomy Support in the Player-Game Relation ...... 100 5.3.9 Threats to the Desired Play Situation ...... 101 5.3.10 Wellbeing and Motivation in Videogame Play and Ordinary Life ...... 102 5.4 Coda ...... 103 6 Discussion ...... 105 6.1 Limitations / Future Research ...... 112 6.2 Contributions ...... 114 6.3 Final Comments ...... 116 References ...... 119 Appendices ...... 131 Appendix A ...... 131 Appendix B ...... 138 Appendix C ...... 142

Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play vii

List of Figures

Figure 1 (adapted from Allen & Anderson, 2018): A cubic model of competence satisfaction in day-to-day life, competence satisfaction in videogame play, and wellbeing...... 19 Figure 2: Mark of the Ninja (Klei Entertainment, 2012) provides a wealth of contextual information to help players succeed in any situation (or quickly try again if they fail)...... 23 Figure 3: The tangram game played by participants in the experimental condition. The hint text box and surrounding shapes have been moved closer to the tangram outline (using image editing software) for clarity...... 25 Figure 4: Spore (Maxis, 2008), the stimulus videogame selected for Study 2. The cell stage (shown here) features a number of game elements that collectively support autonomy...... 26 Figure 5: 2-way ANOVA plots of competence and autonomy satisfaction, assessed before and after playing Mark of the Ninja. (○) → control condition; (●) → experimental condition. Error bars represent 95% CIs. The range of values shown on the y-axis varies for each plot...... 46 Figure 6: 2-way ANOVA plots of short-term wellbeing measures (positive affect (PA), negative affect (NA), and vitality), assessed before and after playing Mark of the Ninja. (○) → control condition; (●) → experimental condition. Error bars represent 95% CIs. The range of values shown on the y-axis varies for each plot...... 47 Figure 7: Participant autonomy satisfaction, autonomy frustration, and competence frustration measured before and after videogame play. (○) → control condition; (●) → experimental condition. Error bars represent 95% CIs. The range of values shown on the y-axis varies for each plot. ... 74 Figure 8: Participant vitality and intrinsic motivation measured before and after videogame play. (○) → control condition; (●) → experimental condition. Error bars represent 95% CIs. The range of values shown on the y-axis varies for each plot...... 76 Figure 9: Participant anger and tension measured before and after videogame play. (○) → control condition; (●) → experimental condition. Error bars represent 95% CIs. The range of values shown on the y-axis varies for each plot...... 77 Figure 10: Predicted happiness and intrinsic motivation as a product of autonomy satisfaction and immersion. Low and high levels of immersion correspond to 1st and 3rd quartile values. The range of values shown on the y-axis varies for each plot...... 83

viii Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play

List of Tables

Table 1: Zero-order correlations (Pearson's r) of T1 measures (p-values in parentheses). CS = competence satisfaction, CF = competence frustration, AS = autonomy satisfaction, AF = autonomy frustration, BMPN = Balanced Measure of Psychological Needs, PA = positive affect, NA = negative affect...... 42 Table 2: Zero-order correlations (Pearson's r) of T2 measures (p-values in parentheses). C/AS = competence/autonomy satisfaction, C/AF = competence/autonomy frustration, BMPN = Balanced Measure of Psychological Needs, PENS = Player Experience of Need Satisfaction, P/NA = positive/negative affect, Int/Enj = Interest/Enjoyment...... 43 Table 3: Descriptive statistics. Scale reliability estimates and 95% CIs represent categorical omega (ω) bootstrapped with 1000 replications. C/AS = competence/autonomy satisfaction, C/AF = competence/autonomy frustration, BMPN = Balanced Measure of Psychological Needs, PENS = Player Experience of Need Satisfaction, P/NA = positive/negative affect, Int/Enj = Interest/Enjoyment...... 44 Table 4: One-sided Welch's t-tests on need satisfaction and frustration measures, assessed after the manipulation task (T1). C/AS = competence/autonomy satisfaction, C/AF = competence/autonomy frustration, BMPN = Balanced Measure of Psychological Needs...... 44 Table 5: 2-way ANOVAs on measures of competence and autonomy satisfaction, assessed before and after playing Mark of the Ninja. C/AS = competence/autonomy satisfaction, BMPN = Balanced Measure of Psychological Needs...... 45 Table 6: 2-way ANOVAs on measures of short-term wellbeing, assessed before and after playing Mark of the Ninja...... 47 Table 7: Linear regression model of post-play (T2) positive affect. C/AS = competence/autonomy satisfaction, PENS = Player Experience of Need Satisfaction...... 49 Table 8: Linear regression model of post-play (T2) negative affect. C/AF = competence/autonomy frustration, BMPN = Balanced Measure of Psychological Needs...... 49 Table 9: Linear regression model of post-play (T2) vitality. C/AS = competence/autonomy satisfaction, PENS = Player Experience of Need Satisfaction...... 50 Table 10: Scale characteristics when measured after playing the tangram game (T1) and Spore (T2). Composite reliability estimates and 95% confidence intervals are represented by categorical omega (ω) bootstrapped with 5000 replications. Int/Enj = interest/enjoyment, C/AS = competence/autonomy satisfaction, C/AF = competence/autonomy frustration...... 67

Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play ix

Table 11: Zero-order correlations (Kendall's tau-b) at T1 in the control condition (p-values in parentheses). Int/Enj = interest/enjoyment, C/AS = competence/autonomy satisfaction, C/AF = competence/autonomy frustration...... 70 Table 12: Zero-order correlations (Kendall's tau-b) at T1 in the experimental condition (p-values in parentheses). Int/Enj = interest/enjoyment, C/AS = competence/autonomy satisfaction, C/AF = competence/autonomy frustration...... 71 Table 13: Zero-order correlations (Kendall's tau-b) at T2, aggregated across conditions (p-values in parentheses). Int/Enj = interest/enjoyment, C/AS = competence/autonomy satisfaction, C/AF = competence/autonomy frustration...... 72 Table 14: Two sided t-tests (20% trimmed means) on need satisfaction and frustration measures. (Note: ξ is bounded at zero because its equation is enclosed by a square root.) ...... 73 Table 15: 2-way between-within ANOVAs on manipulated need satisfaction and frustration variables. Significant interactions were followed by dependent t-tests (T2-T1) on 20% trimmed means...... 74 Table 16: 2-way between-within ANOVAs on positive wellbeing outcomes and intrinsic motivation. Significant interactions were followed by dependent t-tests (T2-T1) on 20% trimmed means...... 75 Table 17: 2-way between-within ANOVAs on negative wellbeing outcomes. Significant interactions were followed by dependent t-tests (T2-T1) on 20% trimmed means...... 76 Table 18: Linear regression model of post-play happiness ...... 78 Table 19: Linear regression model of post-play calmness ...... 78 Table 20: Linear regression model of post-play vitality ...... 79 Table 21: Linear regression model of post-play intrinsic motivation ...... 79 Table 22: Linear regression model of post-play anger ...... 80 Table 23: Linear regression model of post-play depressed affect ...... 80 Table 24: Linear regression model of post-play tension ...... 81 Table 25: A multiple regression model of post-play happiness with a significant interaction between autonomy (satisfaction) and immersion...... 82 Table 26: A multiple regression model of post-play intrinsic motivation with a significant interaction between autonomy (satisfaction) and immersion. ... 83 Table 27: A multiple regression model of post-play happiness with a significant curvilinear effect of autonomy (satisfaction)...... 84 Table 28: A multiple regression model of post-play intrinsic motivation with a significant curvilinear effect of autonomy (satisfaction)...... 84 Table 29: Zero-order correlations (Kendall's tau b; p-values in parentheses) of autonomy frustration and intrinsic motivation...... 85

x Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play

Table 30: A multiple regression model of post-play (T2) intrinsic motivation, as predicted by experiences of autonomy frustration in both games, accounting for T1 intrinsic motivation...... 86 Table 31: Overview of results for H1 and H2 across Study 1 and 2. While ξ is bounded at zero, it has been assigned positive and negative values in this table for consistency with δ – in other words, to clarify whether experiential variables increased or decreased during play...... 107 Table 32: Overview of results for H3a and H3b across Study 1 and 2...... 109

Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play xi

List of Abbreviations

ANOVA: analysis of variance

BMPN: balanced measure of psychological needs

BRUMS: brunel mood scale

CI: confidence interval

GDC: game development convention

HCI: human-computer interaction

IGD: internet gaming disorder

IMI: intrinsic motivation inventory

µ (mu): sample mean

MMT: mood management theory

PENS: player experience of need satisfaction

PX: player experience

RTS: real time strategy (genre)

σ (sigma): sample standard deviation

SDT: self-determination theory sr2: squared semi-partial correlation

xii Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play Statement of Original Authorship

The work contained in this thesis has not been previously submitted to meet requirements for an award at this or any other higher education institution. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made.

Signature: QUT Verified Signature

Date: ______31 October 2019______

Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play xiii

Acknowledgements

Although all theses are self-indulgent – in that they are written knowing that "agreement" to review is barely volitional – I am glad to have a space here to write exactly what I want without fear of substantive critique. I have so many people to thank:

I have to thank my mum and dad, of course, for making my existence possible at all, and raising me in a supportive environment – circumstances for which I also have my older brothers to thank, both directly (in that they were literally present) and indirectly (in that my parents had already been through everything twice before it would apply to me). Of course, the family members with whom I did not share a house growing up are just as important – my nan and grandad, my aunt, and my sister-in-law, for whom "-in-law" is retained for clarity rather than distance. It would be remiss to conclude this paragraph without highlighting the support mum provided – most tangibly in the form of cheap rent – for the duration of my studies.

I greatly appreciate the Commonwealth's contribution in providing financial support for the work described in this thesis through an Australian Government Research Training Program Scholarship.

Thanks to the GRID Lab students who graduated before me: Kellie Vella, Zac Fitz- Walter, David Conroy, Madison Klarkowski, Cody Phillips, Nicole Vickery, and Mitchell McEwan. I wouldn't be who I am now (and this thesis wouldn't be what it is) without our scattered conversations across the last three years. Louis Van Dyke, Vanessa Cheng, Filip Birčanin, and Andy Bayor, for being cool people I ended up talking to a lot later on; Liam Morrow and Jesse Sharp, for having stuck around much longer. truna, for being truna. Joel Harman – materially, for doing the hard work of developing the tangram game(s), and (more importantly) for being a long-time friend.

Belinda Spratt, for taking the time to comment on this thesis with the same boundless enthusiasm and thoughtfulness she demonstrates towards her own work. Natassja Bertram, for helping at crucial times, and refusing to wear out. I learned a lot about writing and analysis from Joshua Trevett and Dakota Barker during our time together at Haywire. I couldn't have asked for a better editorial team. Thank you both.

xiv Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play

Brendan Keogh, for all the times you provided advice, encouragement, and coffee during my candidature. Thanks for providing feedback on thesis chapters from an outside perspective, given such short notice.

Everyone I know in Melbourne who makes it a place I want to be: Goldie, Terry Burdak, Maize Wallis, Zachariah Chandler, Leura Smith, T-Dog eXtreme, Penelope Brightflower, Jackson Simpson, and Dan Golding. I'm so glad to see you every time.

I couldn't mention Melbourne without writing about my friends in the Western Videoball League – Ian MacLarty, Rylie J. Thomas, Sam Crisp, Zeiya Speede, and Dylan Schneider. I want to acknowledge Zeiya and Dylan in particular – not because they let me stay with them in Melbourne as often as I liked (a couple dozen times, as it turned out), but because they became some of my closest friends. I've lived for, and in, the time we've spent together. We're all headed for something new; something better. I hope we get to live in the same city one day.

Finally, I have to thank my supervisors – Daniel Johnson, and Peta Wyeth. Both have given me immense support over the last three years. I won't forget the ways they helped in the first month of candidature and tutoring especially. I'm grateful to Dan for his regular engagement during the crucial first year of my candidature, our discussions around study design and analysis later on, and finding time to provide feedback on my written work in a truly monstrous schedule. I have a lot of reasons to thank Peta: for consistently supporting my "extra-curricular" work (in research and tutoring); for providing honest and considered advice when I needed it; for reading (and critiquing) a great deal of unnecessarily difficult writing. It's been an incredible experience to learn from you both, and more importantly, it's been the best time of my life.

Need Frustration and Short-Term Wellbeing: Restorative Experiences in Videogame Play xv

1 Introduction

1.1 Background The increasing variation in activities and experiences facilitated through videogame play (e.g., Bruxner, 2018; Kanaga, 2017) merits an inclusive approach to the study of player experience (PX). The growing interest in studying affectively mixed or nuanced PX (e.g., Birk, Iacovides, Johnson, & Mandryk, 2015; Bopp, Mekler, & Opwis, 2016; Cole, Cairns, & Gillies, 2015; Mekler, Rank, Steinemann, Birk, & Iacovides, 2016) reflects an ongoing need to understand how players are influenced by videogame engagement; to consider the interplay of qualities specific to players, videogames, and social contexts (D. Johnson, Wyeth, & Sweetser, 2013). In particular, this thesis is informed by a view in human-computer interaction (HCI) that research should be grounded in positive psychology, "deal[ing] with human strengths and the promotion of wellbeing rather than with human weaknesses and healing alone" (Hassenzahl & Tractinsky, 2006). The inherent appeal of videogame play, alongside its potential to improve wellbeing, complement positive psychology's focus on the individual. It is therefore prudent for PX researchers to consider "when, how, and for whom there is a positive impact of videogame play" (D. Johnson, Jones, Scholes, & Carras, 2013) to improve support for human wellbeing.

PX research on wellbeing and play is often conducted through the lens of self- determination theory (SDT), a psychological theory of human motivation, growth, and wellbeing (Ryan & Deci, 2017c). Central to SDT are three basic psychological needs, whose satisfaction is crucial for human nourishment – competence (the feeling of having an effect), autonomy (a sense of volition in activity), and relatedness (social intimacy and belongingness). Likewise, experiences of need frustration obstruct healthy functioning (e.g., Chen et al., 2015). Research in the SDT-PX literature has shown that experiences of need satisfaction in play can benefit player wellbeing (Ryan, Rigby, & Przybylski, 2006; Vella, Johnson, & Hides, 2015), and indeed may support restoration following negative events (Radel, Pelletier, Baxter, Fournier, & Sarrazin, 2014; Reinecke et al., 2012).

However, research on restorative play – PX characterised by wellbeing improvement – in SDT or elsewhere (Collins & Cox, 2014; Rieger, Frischlich, Wulf,

1 Introduction 1

Bente, & Kneer, 2015) is yet to examine the consequences of need frustration on player wellbeing. Studies discussing need frustration do appear in the PX literature (Koban et al., 2018; Przybylski, Deci, Rigby, & Ryan, 2014; Reinecke et al., 2012); however, in the absence of a PX-specific instrument, experimental studies only infer need frustration from low need satisfaction. Such an approach is conceptually problematic (Van den Broeck, Ferris, Chang, & Rosen, 2016): need satisfaction and frustration occur independently (Ryan & Deci, 2017a), and influence wellbeing in different ways (Bartholomew, Ntoumanis, Ryan, & Thøgersen-Ntoumani, 2011; Chen et al., 2015). As yet, there exists no experimental PX research that simultaneously assesses experiences of need satisfaction and frustration, or their divergent effects on player wellbeing. This research would contribute to an elaborated position on the ways that videogames can support player wellbeing. Understanding how play improves wellbeing may benefit the design of interactive technologies for this purpose.

1.2 Research Aims The present work approaches restorative play from the perspective of SDT, using the parallel assessment of need satisfaction and frustration to reach a more developed understanding of when and how videogame play can improve wellbeing. In particular, this thesis aims to extend the application of SDT in PX research by exploring the ways that need frustration can influence experiences of videogame play. In this way, this research aims to further understand the extent to which videogame play can improve short-term wellbeing, and the means by which it may occur.

A research plan comprising two experimental mixed-methods studies was constructed to achieve these aims. Both of these controlled experiments employed manipulation tasks to alternatively satisfy or frustrate participants' competence (Study 1) or autonomy (Study 2), before engagement in a session of videogame play intended to induce restoration. Pre-post measures of need satisfaction, need frustration, and short- term wellbeing were applied to assess change within and between conditions during play. Post-experiment interviews aimed to extend present knowledge regarding experiences of competence, autonomy, wellbeing, and motivation as they relate to videogame play and ordinary life.

2 1 Introduction

1.3 Contributions The present research investigates the restorative effects of videogame play after need-frustrating events. This work contained in this thesis responds to a call for further SDT research on the ways that need satisfaction and frustration may affect human experience (Van den Broeck et al., 2016). While need frustration remains an ongoing topic in research on physical activity (Bartholomew et al., 2011; Gunnell, Crocker, Wilson, Mack, & Zumbo, 2013), education (Bartholomew et al., 2018; Cordeiro, Paixão, Lens, Lacante, & Sheldon, 2016), the workplace (Bartholomew, Ntoumanis, Cuevas, & Lonsdale, 2014; Unanue, Rempel, Gómez, & Van den Broeck, 2017), and day-to-day life (Chen et al., 2015), this thesis represents (to the author's knowledge) the first experimental work to simultaneously assess need satisfaction and frustration in PX literature. This thesis therefore contributes to SDT and PX literatures, demonstrating that the short-term deficits in wellbeing and motivation associated with a need-frustrating event can be reversed by a successive experience of (need-satisfying) videogame play.

The present work provides an extended view on autonomy frustration, emphasising the intensity of autonomy frustration in modifying the negative affects it may produce. It is therefore recommended that empirical need frustration researchers select outcome measures that correspond with the intensity of need frustration expected to occur.

This thesis contributes to design practice as a reminder that players can often clearly articulate the ways that individual game elements influence their experience of play. While metric-based approaches, for example, are increasingly popular in videogame development (e.g., Giovannetti, 2019), this thesis shows that variations in the approach to play can influence the ways that players identify opportunities to experience competence, autonomy, motivation, and wellbeing during play. Specifically, players identify competence support in design elements according to their literacy of videogame conventions and performances, producing varied scenes of potential satisfaction and frustration. These findings inform the conceptualisation of competence in interactive system design. An interactive system's usability and potential expressiveness varies as a function of user literacy: design elements that support expressiveness for more literate users may go unnoticed by novices, or obstruct use.

Finally, this thesis makes a methodological contribution through the development of experimental stimuli (i.e., two versions of a tangram game) that were shown to reliably

1 Introduction 3

frustrate participants' basic needs. These tangram games are likely to support the ongoing development of experimental need frustration research.

1.4 Thesis Outline Chapter 2 comprises a literature review containing the foundational literature of the thesis. Research on PX is described in the context of mood management theory (MMT) and SDT, the latter of which forms the theoretical basis of the present work. In particular, SDT concepts related to wellbeing, need satisfaction, and need frustration are elaborated in the context of videogame play.

Chapter 3 outlines the research aims, hypotheses, and research questions guiding the present work. The chapter also describes the general approach to study design and analysis that remained consistent across Study 1 and Study 2, before providing further information specific to each.

Chapter 4 pertains to Study 1, which assessed the restorative qualities of videogame play following an experience of competence frustration. This chapter elaborates the research method before presenting and interpreting quantitative results on need satisfaction and frustration, positive and negative affect, vitality, and intrinsic motivation. Qualitative interview findings that relate to the experimental play session and day-to-day life are likewise discussed, with reference to the ways that players apply prior knowledge to new videogames, and the varying degrees of engagement during play.

Chapter 5 details Study 2, which assessed the restorative qualities of videogame play following an experience of autonomy frustration. This chapter illustrates the development of the refined method before turning to report and interpret quantitative results pertaining to need satisfaction and frustration, positive and negative affect (happiness, calmness, depression, tension, and anger), vitality, intrinsic motivation, and immersion. Qualitative interview findings are likewise described and analysed with respect to experiences of autonomy, wellbeing, and motivation, as they relate to videogame play and ordinary life.

Chapter 6 presents a discussion of Study 1 and 2 in the context of prior research, before turning to acknowledge the present work's limitations, and opportunities for further study. The chapter concludes with a review of the main contributions of the research.

4 1 Introduction

2 Literature Review

This chapter elaborates on the foundational literature upon which the present work was conducted, focusing on research that implicates videogame play in wellbeing processes. The literature review begins with an outline of short-term wellbeing research in the player experience (PX) domain. This is followed by research pertaining to mood management theory (MMT), which views videogame play as an active means of maximising positively-valenced affect and moderating aroused affect. Finally, self- determination theory (SDT) – which identifies videogame play as one of many domains of human experience – is reviewed, with emphasis on the ways that need frustration appears to influence (negative) wellbeing outcomes, and how this may inform more developed models of PX.

2.1 Player Experience Player experience research originated in human-computer interaction (HCI) as "user experience with entertainment technologies" (Mandryk, Inkpen, & Calvert, 2006), a position that coincided with the growing interest in affective experiences with technology (e.g., relative to ergonomics or usability) in third-wave HCI (Bødker, 2006; Harrison, Tatar, & Sengers, 2007). Borrowing from HCI, PX research frameworks (e.g., D. Johnson, Wyeth, et al., 2013; Nacke & Drachen, 2011) aim to represent the constituent forms of videogame play across contexts – in other words, the ways that players, videogames, contexts, and their interactions contribute to an experience. If maximising player enjoyment was once seen as central when studying an entertainment medium (Sweetser & Wyeth, 2005), it has become clear that videogames support more nuanced experiences – positive and negative, physical and psychological – and a growing body of PX literature has emerged on this topic (Birk, Iacovides, et al., 2015; Mekler et al., 2016). In this thesis, these variations are primarily relevant to the extent that players experience short-term psychological benefits or detriments as a result of videogame engagement – more specifically, the extent to which play can improve short-term wellbeing, and the means by which it occurs. While PX research may be conducted using many theories (e.g., reversal theory, mood management theory) and measures (e.g., the Immersive Experience Questionnaire), this thesis employs SDT and its associated measures, due to the theory's prevalence and utility in the PX literature (section 2.2.2).

2 Literature Review 5

2.1.1 Short-Term Wellbeing Benefits of Videogame Play Qualitative studies highlight the common understanding among the general (videogame-playing) public that videogames can, and often are, used for stress reduction and affective improvement. For example, in a qualitative survey on meaningful PX (Oswald, Prorock, & Murphy, 2014), respondents also described playing videogames for catharsis, relaxation, or to "zone out". Similarly, a content analysis of online forum posts (Bourgonjon, Vandermeersche, De Wever, Soetaert, & Valcke, 2016) observed prominent themes of self-development (e.g., changed worldview), personal wellbeing, and catharsis in players' understanding of the benefits of play. Finally, interview studies on immersion (Jennett, 2010) and escapism (Kuo, Lutz, & Hiler, 2016) indicate that play for stress relief is a form of "active escapism", a process of separation from present issues that is facilitated by the fantastic quality of virtual worlds (which are less likely to include referents to the player's stressors).

Videogame play's capacity to promote affective improvement and stress reduction is equally well-established in cross-sectional and experimental literature. A large-scale survey of predominantly male (96.3%) videogame players (Reinecke, 2009) found that playing for recovery was positively associated with work-related strain and daily hassles, and more strongly associated with emotion-focused coping strategies (e.g., self- distraction, denial; Folkman & Moskowitz, 2004) than problem-focused strategies (e.g., active coping, planning; Folkman & Moskowitz, 2004). A similar study (Collins & Cox, 2014) that surveyed a larger proportion of women (48.9%) found that recovery was positively associated with weekly hours played, and this correlation was partially mediated by the quality of in-game relationships, emphasising the importance of social experiences in multiplayer games for player wellbeing.

A controlled experiment investigated the differential effects of three videogames – Bejeweled 2 (PopCap Games, 2004), Bookworm Adventures (PopCap Games, 2006), and Peggle (PopCap Games, 2007) – on physiological stress (operationalised as heart rate variability) and player affect (Russoniello, O’Brien, & Parks, 2009). In general, results indicated that participants assigned to play videogames experienced greater affective improvement than those assigned to the control condition (who worked on a computer task instead). Playing Bejeweled 2 also appeared to reduce physiological stress more effectively than control. In a follow-up study, participants assigned to play their choice of Bejeweled 2, Bookworm Adventures, or Peggle over a one-month period reported lower

6 2 Literature Review

stress (Fish, Russoniello, & O’Brien, 2014) and depressive symptoms (Russoniello, Fish, & O’Brien, 2013), relative to control (who were asked to abstain from videogame play). Similar results were also found in research on the effects of screen size on PX (Hou, Nam, Peng, & Lee, 2012). While playing Tomb Raider 2 (Core Design, 1997) for thirty minutes in either condition promoted affective improvement, participants assigned to play on a larger screen experienced greater benefits.

Using qualitative methods, cross-sectional surveys, and controlled experiments, the reviewed literature demonstrates that players intentionally play videogames for wellbeing – achieving catharsis (Bourgonjon et al., 2016; Oswald et al., 2014), separating from present stressors (Jennett, 2010; Kuo et al., 2016), strengthening online relationships (Collins & Cox, 2014), or attenuating depressive symptoms (Russoniello et al., 2013). To some extent, these publications emphasise particular qualities of videogames or the play context (e.g., large screens, fantasy worlds, other players), rather than more foundational precursors of wellbeing. In many ways, PX research benefits from the breadth of theoretical perspectives brought (in)to the field – the studies described above, for example, investigate a range of PX qualities according to the interests of their home fields of media studies, HCI, health, and communication. Equally, however, work conducted under an established theoretical agenda affords a consistency that affords depth in research on PX concepts. For this reason, it is useful to discuss PX research conducted under the MMT paradigm, a theory of media effects with a substantial empirical literature on videogame play.

2.1.2 Mood Management Theory Applications of MMT to videogame play attempt to understand how and why players use videogames to produce short-term psychological changes (Bowman & Tamborini, 2012, 2015; Reinecke, Klatt, & Krämer, 2011; Reinecke et al., 2012; Rieger et al., 2015; Rieger, Wulf, Kneer, Frischlich, & Bente, 2014; Tamborini et al., 2011). MMT (Zillmann, 1988) is a media effects theory whose central position is that human behaviour (however conscious) is directed towards extending desired affective states, while avoiding or limiting undesirable affects. Media consumption represents one domain whereby individuals can readily predict their affective response to stimuli – James Bond films, for example, can generally be expected to feature espionage in foreign countries, conventionally attractive women, action sequences, and high-end product placement.

2 Literature Review 7

MMT posits that individuals select media stimuli likely to moderate affective arousal (i.e., between boredom and stress), induce positively-valenced affects (e.g., happiness), and unlikely to feature material reminiscent of salient negative events (Zillmann, 1988). According to the theory (Knobloch-Westerwick, 2006), media vary in the degree to which they can increase arousal (excitatory potential), consume attention (absorption potential), feature contextually-relevant material (semantic affinity), and increase positive affects (hedonic valence). Applications of MMT in PX research have focused on the first two qualities, on the basis that the "interactive" nature of videogame play is both more exciting and requires more attention than engagement with other media (Bryant & Davies, 2006). These studies often employ manipulation tasks – for example, threading metal washers onto string (Bowman & Tamborini, 2015), solving logic puzzles under time pressure (Bowman & Tamborini, 2012), or watching a sad scene in a film (Rieger et al., 2014) – to induce particular affects prior to participants' exposure to the target stimulus. An experimental study (Bowman & Tamborini, 2012) that induced boredom or stress (in separate conditions) prior to videogame play found a quadratic relationship between in-game task demand and post-play affect in both conditions, such that affective improvement was maximised at a moderate level of task demand. A similar study (Bowman & Tamborini, 2015) that allowed participants to select their preferred level of task demand prior to play instead found that mood repair after a boring or stressful event was greatest when high levels of in-game task demand were chosen. In these studies (and Reinecke et al., 2012, described in Section 2.2.3), participants played Lock On: Modern Air Combat (Eagle Dynamics, 2003), a flight simulator, for five minutes per session. Game selection was predicated on the modularity of its auto-pilot feature, which simplified operationalisation of task demand; in other words, "low", "medium", and "high" task demand conditions could be clearly delineated by the extent to which auto- pilot features were assigned (or chosen by participants).

Other research has primarily examined the effects of play on affective valence. An experimental study (Rieger et al., 2014), whereby participants completed a frustrating task prior to a solitary play session of Mario Kart Wii (Nintendo EAD, 2008), found positive relationships between in-game need satisfaction (see section 2.1.2) and enjoyment – however, only in-game success (conceptualised as a positive outcome, and operationalised as the rank obtained after two races) emerged as a relevant predictor of improved affect (measured before and after play). However, given the sample size (n=46),

8 2 Literature Review

and probable correlations between independent variables, tests of each individual predictor's contribution may have been underpowered (Darlington & Hayes, 2016, pp. 520-521). A second study (Rieger et al., 2015), with a (film-based) sadness induction and three experimental conditions – five minutes of playing Pac-Man (Namco, 1980), watching gameplay footage of Pac-Man, and "non-exposure to media stimuli" – found that happiness increased to a similar degree across all three conditions. However, participants who played Pac-Man (which induced high arousal and task demand) reported a greater reduction in depressed affect than those who watched gameplay footage or were not exposed to media stimuli. Finally, a series of three studies on restorative play (Koban et al., 2018) investigated the effects of performance in a competitive quiz game following a false-feedback manipulation task. Wins were conceptualised as positive play outcomes, and hypothesised to increase competence satisfaction and positive affect. In general, participants who won against their opponent experienced the greatest improvement in affect and competence satisfaction; these effects were amplified in the studies where participants were assigned to win all four rounds. For participants assigned to the winning condition, close wins appeared to produce larger gains than overwhelming victories; conversely, participants assigned to lose every match experienced greater competence after play when defeat was marginal. In aggregate, these results indicate that players gauge competence relative to task difficulty, and not success or failure alone. Videogames that generate scenes of potential success therefore appear more likely to improve wellbeing, at least in the immediate aftermath of play.

The MMT-PX literature reviewed here demonstrates that videogame play can improve short-term wellbeing following confrontation with negative events. This research is also informative in demonstrating study designs whereby participants' psychological states are influenced prior to media engagement without serious consequences. However, these publications are uninformative with respect to the determinants of short-term wellbeing that emerge during the player experience. These experiences – need satisfaction and frustration, intrinsic motivation, affect, vitality, and immersion – are most germane when read in conversation with SDT, which forms the theoretical backbone of the present work.

2 Literature Review 9

2.2 Self-Determination Theory Self-determination theory (SDT) is a modernist theory of human motivation, growth, and wellbeing. SDT posits an innate human propensity for inherently interesting or enjoyable activity (intrinsic motivation), and an organising tendency towards the development of increasingly complex models of the self in relation with the social environment (organismic integration). Wellbeing (a.k.a. "thriving" or "fully functioning") occurs as a result of "activity that develops and expresses one’s most reflectively valued and well integrated human potentialities" (Ryan, Curren, & Deci, 2013).

Intrinsic motivation, organismic integration, and wellbeing processes are essential elements of SDT, and each one is energised by the satisfaction of three basic psychological needs (discussed further in section 2.2.2): competence (feelings of effectiveness in interaction), autonomy (a sense of volition and self-endorsement), and relatedness (social intimacy and connection with others). Conversely, each process is undermined by experiences of need frustration – the active thwarting of basic needs (discussed further in section 2.2.3).

Intrinsic motivation is a relational quality of the human organism with their environment, characterised by the ongoing experiences of interest and enjoyment associated with a given activity (Csikszentmihalyi, 1975; Izard, 1977). Intrinsically- motivated activity is done for its own sake; the inherently rewarding experiences of need satisfaction that result from action serve to energise future behaviour. (Extrinsically motivated behaviours, in contrast, are performed for instrumental purposes.) Videogame play is generally considered intrinsically-motivated activity (Ryan et al., 2006), and applications of the SDT-based Intrinsic Motivation Inventory (IMI; McAuley, Duncan, & Tammen, 1989; Ryan, Mims, & Koestner, 1983) across the PX literature provide support for this position (D. Johnson, Klarkowski, et al., 2018; Klarkowski et al., 2016; Mekler, Bopp, Tuch, & Opwis, 2014; Schmierbach, Limperos, & Woolley, 2012; Schmierbach, Xu, Oeldorf-Hirsch, & Dardis, 2012; Tamborini, Bowman, Eden, Grizzard, & Organ, 2010). Intrinsic motivation for videogame play is mobilised by in-game need satisfaction (Ryan et al., 2006; Tamborini et al., 2010), and the potential for players to experience aspects of their "ideal self" (Przybylski, Weinstein, Murayama, Lynch, & Ryan, 2012).

10 2 Literature Review

Organismic integration represents the process by which the self works to arrange itself into coherent structures of increasing complexity and organisation. The measurable component of this process represents the degree to which extrinsically-motivated behaviour is autonomously motivated. Integrated regulation, for example, is a highly autonomous form of extrinsic motivation resulting from the wholehearted valuation of activity considered less interesting or enjoyable. While PX research on this topic is somewhat limited, a small number of studies have investigated relationships between extrinsic motivation and play, with mixed results. The Gaming Motivation Scale (GAMS; Lafrenière, Verner-Filion, & Vallerand, 2012) was developed to assess the extent to which videogame engagement is autonomously regulated, with preliminary results indicating that well-integrated extrinsic motivation most effectively predicted gaming frequency (hours played per week). However, a similar study of an online brain-training game (Burgers, Eden, van Engelenburg, & Buningh, 2015) instead found that only intrinsic motivation predicted future play behaviour. While these studies conflict, to some extent, on the types of motivational process implicated in videogame play, research linking in-game need satisfaction and hours played (D. Johnson, Gardner, & Sweetser, 2016) suggests that autonomous motivation – intrinsic or extrinsic – is likely to energise play behaviour.

While intrinsic motivation and organismic integration are essential to SDT, understanding the theory's characterisation of wellbeing is essential to the work conducted in this thesis. Elaboration on this topic first requires a more detailed explanation of the broader historicophilosophical perspective from which SDT's theoretical model of wellbeing was derived.

2.2.1 Wellbeing in Self-Determination Theory Wellbeing refers to the psychological states (e.g., affect, vitality) that indicate positive functioning over the short- or long-term (Sheldon, 2018). As noted (section 2.2), SDT posits that wellbeing emerges from "activity that develops and expresses one’s most reflectively valued and well integrated human potentialities" (Ryan et al., 2013). Central to this position is a "neo-Aristotelian" interpretation (Ryan et al., 2013) of eudaimonia, which refers to the individual's total fulfilment of their capacity as a human being. Implicit in this definition is the requirement to act, or more specifically to engage in "particular ways of thinking and / or behaving" (Sheldon, 2018; emphasis in original) whose successful enaction will result in "the good life".

2 Literature Review 11

It is important to note, however, that while SDT defines wellbeing with respect to actions, measurement occurs at the level of experience. The most recent comprehensive summary of SDT literature does not clearly define experiential wellbeing (i.e., as a formal proposition), and instead provides a general description (Ryan & Deci, 2017c, p. 241). More practically, short-term wellbeing (and wellbeing-adjacent) concepts mentioned in this general description include awareness, vitality, affect, and meaning.

Awareness refers to a mode of open perception whereby the self actively considers itself in relation with the social context; in other words, what Ryan and Deci (2017a) describe as "open, relaxed, and interested attention to oneself and to the ambient social and physical environment". Awareness is essential to wellbeing because it facilitates autonomous behaviour and the integrative process. While the relationship between awareness and PX is yet to be explored in detail, it is worth noting (and elaborated in sections 4.3.7 and 5.3.5) that immersion "can pull one out of mindful awareness" (Rigby, Schultz, & Ryan, 2014, p. 222), indicating that the two modes of engagement are antithetical.

Vitality is characterised as "energy available to the self" (Ryan & Deci, 2008), reflecting a holistic sense of "aliveness" or joie de vivre that is not entirely captured by sleep quality, (perceived) physical health, or aroused affect (Ryan & Frederick, 1997). Instead, vitality is more robustly associated with experiences of need satisfaction in autonomy-supportive social contexts. While vitality can be considered either a state- or trait-level measure (i.e., short- or long-term), it is rarely used in PX research. As previously discussed (section 2.2.2), Ryan et al. (2006) found that in-game need satisfaction predicted changes in vitality (measured before and after experimental play sessions). A large-scale survey of online videogame players (Przybylski, Weinstein, Ryan, & Rigby, 2009) likewise found that greater levels of need satisfaction experienced in day-to-day life predicted higher post-play vitality.

Affect is defined in SDT in terms of its informational quality. The valence or intensity of experienced affect is secondary to how successfully an individual can perceive, regulate, and integrate the experience into the self (Deci & Ryan, 1985, p. 233; Ryan, Deci, & Vansteenkiste, 2016). As SDT posits that affect is merely a "symptom", rather than a direct experiential measure of wellbeing (Ryan & Deci, 2017a), affective concepts and measures are generally applied to suit the research context. In PX research, for example, affect has been measured as combined positive and (reversed) negative

12 2 Literature Review

affect (Ryan et al., 2006), separated categories of positive and negative affect (Przybylski et al., 2012), and more distinct affective states (e.g., aggressive affect; Przybylski et al., 2014).

Each concept in the bricolage that represents experiential wellbeing is energised by basic need satisfaction – and, crucially, eroded by need frustration. The centrality of basic needs in this thesis reflects a position that videogames are not merely used to moderate affect (cf. section 2.1.2), but instead for the foundational experiences of need satisfaction that (appear to) lead towards "the good life".

2.2.2 Need Satisfaction in Videogame Play While SDT clearly defines basic psychological needs as distinct concepts, their applications vary across domains to reflect the salient qualities of each social context. Accordingly, SDT's approach to PX research emphasises the ways that game design may support or thwart the satisfaction of players' basic needs, rather than analysis of "specific instances of games or genres" (Ryan et al., 2006).

As noted (see section 2.2), competence refers to feelings of efficacy in interaction. SDT proposes that videogames satisfy competence primarily by displaying (immediate and disproportionally positive) feedback, conveying a clear sense of in-game progress (Burns, 2018), varying levels of challenge over time, and implementing intuitive control mechanisms (Ryan & Deci, 2017b).

Autonomy is characterised by a sense that behaviour is self-endorsed and volitional. Videogames may satisfy autonomy by allowing self-expression (e.g., via avatar customisation; Birk, Atkins, Bowey, & Mandryk, 2016; Birk & Mandryk, 2018), providing meaningful choices and reasons for in-game activity (Rigby & Ryan, 2011b), and through dynamic or "procedural" structures (Ryan et al., 2006) that facilitate a variety of player behaviour.

Relatedness refers to experiences of social intimacy and connection with others. During videogame play, relatedness satisfaction may result from social experiences with other human players or engagement with helpful AI-controlled teammates (Rigby & Ryan, 2011c). Theoretical work (Tyack & Wyeth, 2017) has also identified (hegemonic) "gamer" culture, as well as videogames themselves (as complex, responsive, and often cooperative partners in engagement) as non-human entities that regardless appear to

2 Literature Review 13

approximate social interaction – and may, therefore, satisfy relatedness – even during solitary play experiences.

Basic need satisfaction is underpinned by autonomy support (Ryan & Deci, 2017a), a contextual quality characterised by the provision of (meaningful) choice, minimising overt control, and acknowledging others' perspectives and desires (Grolnick, Deci, & Ryan, 1997). Videogames are seen to support player autonomy primarily by including opportunities to make seemingly important decisions, imparting a sense of purpose towards in-game activity, and creating expansive environments whose exploration yields a sense of novelty (Rigby & Ryan, 2011b; Ryan & Deci, 2017b).

Need satisfaction and intrinsic motivation constructs have seen extensive application in quantitative PX research, as early SDT research in this area (D. Johnson & Gardner, 2010; Przybylski et al., 2014; Przybylski, Ryan, & Rigby, 2009; Przybylski, Weinstein, et al., 2009; Ryan et al., 2006; Tamborini et al., 2010) demonstrated the relative ease of measuring experiences across genres, and with varying game elements. More specifically, SDT researchers developed a survey instrument – the Player Experience of Need Satisfaction (PENS) – which assesses experiences of need satisfaction, immersion, and the intuitiveness of controls during play, and conducted four studies (three laboratory experiments, one survey) demonstrating its successful application (Ryan et al., 2006). In general, the researchers found that need satisfaction during play could effectively predict game enjoyment and changes in short-term wellbeing (affect, self-esteem, and vitality), and that the stability of psychological measures is an asset for researchers in the videogame domain, "a creative, quickly evolving, and widely variable area" (Ryan et al., 2006, p. 360). Links between need satisfaction and short-term wellbeing also emerged from an experimental study of single- player games (Birk & Mandryk, 2013), which is otherwise notable for finding only nominal (and statistically non-significant) effects of relatedness on PX in single-player games, which were previously assumed without testing (Ryan et al., 2006). Autonomy satisfaction during solitary play has also been shown to positively predict wellbeing in day-to-day life (Vella et al., 2015).

Despite research on the positive qualities of videogame play, SDT researchers have also expressed concern about "game overuse", which is theorised to occur when need satisfaction experienced during play is used to offset low need satisfaction in ordinary life (Rigby & Ryan, 2011a; Ryan & Deci, 2017b). In this view, videogames are particularly

14 2 Literature Review

attractive because they are designed to provide immediate, consistent, and dense experiences of need satisfaction, which may be difficult to resist. Inherent to this "Need Density Hypothesis" is a position whereby the need-satisfying qualities of videogame play are not themselves problematic (as this would be antithetical to the general theory); instead, need dissatisfaction or frustration in ordinary life is made complicit, and game overuse is merely one of many disordered behaviours humans may attempt in response (e.g., Ryan et al., 2016).

Game overuse, however, may also occur as a result of experiencing "immersion" and therefore failing to recognise the passing of time. Immersion is a conceptually nebulous term, particularly in SDT literature, where it was originally described as "presence" (Ryan et al., 2006), or the "illusion of nonmediation" (Lombard & Ditton, 1997), and later referred to as "immersion/'presence'" (Rigby & Ryan, 2007) or simply "immersion" (Przybylski et al., 2012). Immersion is alternatingly interpreted as an outcome of need satisfaction (Ryan et al., 2006) or a moderating variable (Przybylski et al., 2012) during analysis, with nominal justification in either case. Relations between immersion and other SDT concepts remain unexplained, with few exceptions (see section 2.2.1). In short: it is not clear why immersion is important to SDT-PX research, how it relates to other concepts (e.g., intrinsic motivation), or what it can (or should) be used for in analysis.

The PENS instrument itself remains a point of contention in the SDT-PX literature. The PENS is a commercial instrument, which prevents full publication of scale items, consequently limiting critical discussion – or, more importantly, further development. As a result of its ongoing stasis, the PENS lacks the capacity to assess need frustration, which has since been identified as an area of emerging importance in SDT (Van den Broeck et al., 2016). Quantitative measures are discussed in greater detail later in the text (section 3.2.1, 3.3.1, 4.1.2, and 5.1.2).

2.2.3 Need Frustration in Videogame Play Need frustration is conceptualised as the active thwarting of an individual's basic needs. In this way, it is distinguished from need dissatisfaction, which instead represents a state of low need satisfaction. Despite the apparent binarism between need satisfaction and frustration, theoretical (Ryan & Deci, 2017a) and empirical (Bartholomew et al., 2011) literature indicates that they are independent; in other words, that they may co-

2 Literature Review 15

occur within individuals. Crucially, need frustration has been shown to predict negative wellbeing outcomes (e.g., negative affect, burnout, stress) more effectively than need satisfaction (Bartholomew et al., 2011; Gunnell et al., 2013).

Although the conceptual formulation of need frustration has existed in SDT's theoretical literature for some time (Ryan & Deci, 2000), its practical application largely remained latent. Its recent resurgence in the SDT literature began in sport research (Bartholomew et al., 2011) where a survey instrument was developed and successfully applied in further studies (Bartholomew et al., 2014; Bartholomew et al., 2018). The formalisation of alternative domain-specific and general measures followed (Chen et al., 2015; Longo, Gunz, Curtis, & Farsides, 2016; Rocchi, Pelletier, Cheung, Baxter, & Beaudry, 2017; Sheldon & Hilpert, 2012). In aggregate, existing research indicates that need frustration links with negative outcomes such as burnout, depression, disengagement, and general negative affect (Bartholomew et al., 2018; Bartholomew et al., 2011; Earl, Taylor, Meijen, & Passfield, 2017; Gillet, Forest, Benabou, & Bentein, 2015; Gillet, Fouquereau, Huyghebaert, & Colombat, 2015; Gunnell et al., 2013; Haerens, Aelterman, Vansteenkiste, Soenens, & Van Petegem, 2015; Jang, Kim, & Reeve, 2016; Vanhee, Lemmens, Stas, Loeys, & Verhofstadt, 2018).

Need frustration concepts are readily adapted to the context of videogame play. Competence frustration refers to feeling inadequate or incapable of success, and may result from repeated failure to overcome in-game challenges, or when game controls themselves obstruct success (Przybylski et al., 2014). Autonomy frustration represents feeling pressured or coerced to act in certain ways; during play, experiences of autonomy frustration may result from being tasked with unwanted activities (e.g., playing a tedious mini-game to progress; Tyack & Wyeth, 2017), when a preferred approach (e.g., stealth, non-violence) is unsupported, or when play becomes a social obligation (Deterding, 2016). Relatedness frustration refers to feelings of social isolation or rejection, and may emerge when other actors – friends, acquaintances, strangers, or liked virtual characters – exhibit deviant (or "toxic") behaviour, essentially betraying the player or team (Birk, Buttlar, et al., 2016).

A drawback of the existing PX literature on need frustration is the lack of a domain- specific instrument – as a result, the few existing publications on this topic assess need frustration with substitute measures. Research using need satisfaction in this way (Przybylski et al., 2014; Reinecke et al., 2012; Sheldon & Filak, 2008) introduces a

16 2 Literature Review

confound between need frustration and need dissatisfaction, which complicates interpretation of study results. For example, a series of experimental studies about the potential causes of aggression following play (Przybylski et al., 2014) used the PENS competence subscale to test hypotheses on need frustration. The authors concluded that aggressive cognitions and behaviours were robustly predicted by experiences of competence frustration during play. Similarly, an experimental study of need frustration while playing an analogue puzzle game (Sheldon & Filak, 2008) manipulated all three basic needs, demonstrating hypothesised links between competence and relatedness satisfaction and positive outcomes, and inverse associations between relatedness frustration and game performance. Finally, a third study (Reinecke et al., 2012) used a false-feedback task – in which participants received positive or negative performance feedback according to experimental condition, rather than true task performance – to successfully manipulate participants' competence and autonomy satisfaction prior to videogame play. While results showed positive relationships between post-play need satisfaction and game enjoyment, as expected, competence satisfaction appeared to decline after play (from 5.47 to 2.80, on a 7-point scale) for participants who received positive feedback in the manipulation task (although this difference was not tested). Moreover, it is not clear why mood repair was indirectly assessed (i.e., as game enjoyment), having also measured post-play affect. To summarise: although each of these study designs likely induced need frustration, the possibility remains that only dissatisfaction occurred, and such experiences were sufficient to produce the outcomes described. Need satisfaction measures used in this way also foreclose simultaneous measurement of need satisfaction and frustration, which is essential to verify their effects on positive and negative PX outcomes.

While SDT posits that need frustration undermines organismic functioning, some empirical research has complicated this view. Links between intrinsic motivation and feeling controlled were investigated in a recent experimental study (Radel et al., 2014), in which participants solved tangrams in a controlling or neutral context before playing a puzzle game in a similar or different environment. Participants assigned to the different environment condition were taken to another room by a second researcher to play the puzzle game. In this condition, participants who were first exposed to the controlling context experienced greater post-play intrinsic motivation than participants exposed to the neutral context. However, this effect was reversed for participants in the similar

2 Literature Review 17

environment condition, for whom the researcher and room were held constant: those previously exposed to the controlling context, relative to the neutral context, experienced less intrinsic motivation towards puzzle game play. Whereas SDT conceptualises need frustration as uniformly detrimental to organismic functioning, these results instead suggest that need frustration can energise intrinsic motivation for future activity. One potential point of contention with the second study arises from the "controlling context" scale, whose sample items seem face valid, but bear little resemblance to either parent scale (i.e., the Controlling Coach Behaviors Scale (CCBS; Bartholomew, Ntoumanis, & Thøgersen-Ntoumani, 2010), and the Teacher Control Questionnaire (TCQ; Jang, Reeve, Ryan, & Kim, 2009; Reeve & Halusic, 2009)). Further research on competence frustration in tertiary education (Fang, He, Fu, & Meng, 2017) has observed similar results, although they were only shown to hold for students who experienced high competence frustration in the first class (n=78). Competence frustration undermined intrinsic motivation for all other students (n=539) across classes, in line with the established theory. Why this paradoxically positive relationship with intrinsic motivation occurred only at high levels of need frustration remains unclear.

More recently, a cross-sectional survey on Internet Gaming Disorder (IGD; Allen & Anderson, 2018) directly assessed general wellbeing (a composite measure), need satisfaction, and need frustration associated with videogame play and everyday life, and the prevalence of IGD itself. In general, IGD was associated with high need satisfaction in play, low need satisfaction in everyday life, and high need frustration in both contexts. Wellbeing results are more difficult to interpret, due to the use of quadratic and cubic models. For example, relations between game- and life-related competence satisfaction with respect to wellbeing formed a cubic model, in which wellbeing is close to maximised at three points – first, where competence in both domains is maximised; second, where everyday-competence is maximised and videogame-competence is minimised; and third, where everyday-competence is minimised and videogame-competence is moderate (see Figure 1). The practical relevance of this cubic model – whereby moderately-skilled videogame players who feel otherwise incompetent regardless appear to experience high levels of wellbeing – is somewhat unclear. On the whole, however, study results indicate that everyday need satisfaction, relative to game-related need satisfaction, is more essential to wellbeing in the long-term. These results align with those of another study (Milyavskaya, Philippe, & Koestner, 2013), which indicates that wellbeing outcomes

18 2 Literature Review

largely result from need satisfaction assessed at the same level of specificity (i.e., at the general, domain-specific, or episodic level). From this perspective, videogame play is unlikely to produce substantial improvements in general wellbeing – however, players may regardless value localised (or short-term) benefits to their wellbeing (see section 5.3.10).

Figure 1 (adapted from Allen & Anderson, 2018): A cubic model of competence satisfaction in day-to- day life, competence satisfaction in videogame play, and wellbeing.

The reviewed literature underscores the present lack of PX research on need frustration in the context of videogame play. This absence coincides with the issue of adequate measurement – validated domain-specific instruments are yet to appear, and more general scales (e.g., Chen et al., 2015; Sheldon & Hilpert, 2012) may not fully apply to in-game experiences (discussed further in sections 3.3.1 and 4.3). At present, empirical PX literature on need frustration either assesses need satisfaction as a proxy (Koban et al., 2018; Przybylski et al., 2014; Reinecke et al., 2012), confounding need frustration and dissatisfaction; refers only to general patterns of videogame engagement (Allen & Anderson, 2018); or presents other potential methodological issues (Radel et al., 2014). The dearth of experimental PX studies of need frustration accordingly limits research on the (negative) wellbeing outcomes of videogame play.

2.3 Summary and Implications The reviewed literature consistently shows that videogame play can improve short- term wellbeing, and may indeed be more effective than other screen media in reducing negative affects. The fantastic virtual worlds where play takes place appear to offer a break from issues in ordinary life, coinciding with processes of "active escapism" or immersion. While people clearly seek videogames after negative life events, the

2 Literature Review 19

experiential precursors of improved short-term wellbeing during play are yet unclear. Recent theoretical developments on need frustration (and its influence on negative wellbeing outcomes) suggest that an expanded program of research in the SDT-PX framework could offer a more complete understanding of restorative play. The present difficulty of studying need frustration in the context of PX – primarily limited by the lack of a domain-specific instrument – has obstructed further SDT-PX research on negative short-term wellbeing outcomes. Alternatively, as restorative play seems largely intentional, players themselves may offer more nuanced perspectives on when and why these behaviours are most successful. Finally (and least importantly, given SDT's theoretical ambivalence (section 2.2.2)), immersion's potential to amplify PX is yet to be studied with respect to short-term wellbeing outcomes, and therefore represents another relevant avenue of exploration. A more developed perspective on the ways that videogames can support player wellbeing will ultimately benefit the innumerable players who find in videogames something that helps.

20 2 Literature Review

3 Research Design and Methodology

3.1 Research Structure and Scope The present research was undertaken to identify the influence of need frustration in the context of player experience (PX), and its influence on short-term wellbeing outcomes. As noted in the literature review (section 2.2.3), the lack of a domain-specific measure of need frustration has obstructed experimental research on the topic, and the use of need satisfaction as a proxy measure prevents the experiential precursors of positive and negative wellbeing outcomes from being distinguished. Further study of the predictive value of need satisfaction and frustration on positive and negative outcomes is a key issue in current self-determination theory (SDT) literature (Van den Broeck et al., 2016).

There are also opportunities to extend the literature on restorative play behaviours, which emphasises qualities of videogames and social contexts implicated in short-term wellbeing improvement (section 2.2.2). Less information is available on what players actually do to create scenes that support restoration, the extent to which their experiential qualities align with SDT, and how restorative play and wellbeing are understood in relation with other domains of ordinary life.

To this end, the present work was conducted with two primary aims:

Extend applications of SDT in PX research by exploring the ways that need frustration can influence experiences of videogame play.

Further understand the extent to which videogame play can improve short-term wellbeing, and the means by which it may occur.

An experimental, mixed-methods research program was devised to achieve these aims. Specifically, the present research comprised two laboratory studies, whereby need frustration was manipulated prior to a thirty-minute period of videogame play. Experimental sessions were immediately followed by post-play interviews. Crucially, survey instruments were used to assess need satisfaction and frustration as separate constructs, and accompanied by PX measures associated with intrinsic motivation and short-term wellbeing. Study designs with two conditions and pre-post measurement were essential to demonstrate the successful manipulation of need frustration between

3 Research Design and Methodology 21

conditions prior to play, and relative change in experiential outcomes within conditions during play.

Conducting two studies allowed research on competence and autonomy frustration to occur. (As relatedness measures are only effective in the context of social play (see section 2.2.2), researching relatedness frustration would have greatly complicated study design and recruitment, and – given the timeframe – was deemed outside the scope of this thesis.) Studying two basic needs also supported a more focused qualitative approach. Given the paucity of directly comparative literature, mitigating risk was critical – for this reason, pre-existing manipulation tasks were adapted for both studies.

3.1.1 Hypotheses and Research Questions To guide the research program towards accomplishing the research aims, and addressing the research gaps identified in the SDT-PX literature (section 2.2.3, 2.3, and 3.1), the following hypotheses and research questions were formulated:

H1: Need-frustrated participants, relative to need-satisfied participants, will experience greater improvement in need satisfaction during play.

H2: Need-frustrated participants, relative to need-satisfied participants, will experience greater improvement in short-term wellbeing and intrinsic motivation during play.

H3: Experiences of need satisfaction and frustration during play will distinctly predict positive and negative outcomes.

H3a: Experiences of need satisfaction during play will predict positive short- term wellbeing outcomes and intrinsic motivation.

H3b: Experiences of need frustration during play will predict negative wellbeing outcomes.

RQ1: To what extent can need frustration influence intrinsic motivation for a successive activity?

RQ2: How do players understand need satisfaction, need frustration, and wellbeing as they relate to videogame play?

The research approach outlined in the remainder of this chapter was developed with the intention of conducting, as far as possible, robust experimental studies from which

22 3 Research Design and Methodology

reliable conclusions could be derived. The application of this principle is reflected in the stability of experimental procedures across both studies, varying only to the extent required to accommodate each manipulation task (section 3.2 and 3.3).

3.2 Study 1 Design Study 1 evaluated the effect of a prior experience of competence frustration on PX. Specifically, Study 1 tested whether participants in the experimental condition, relative to control, would experience greater improvement in competence satisfaction and short- term wellbeing during play. Data collection occurred from July-October 2017.

Figure 2: Mark of the Ninja (Klei Entertainment, 2012) provides a wealth of contextual information to help players succeed in any situation (or quickly try again if they fail).

After defining hypotheses and research questions, a false-feedback word finding task (adapted from Sheldon & Filak, 2008) was selected as the experimental manipulation in Study 1. More specifically, the task involved using adjacent letters to form words in a 4-by-4 grid over a three-minute period. Participants worked at two such puzzles, and were given positive or negative performance feedback after each one, according to experimental condition. This manipulation task was chosen as it required few materials (i.e., a pen, sheets of paper, and the presence of the researcher), lowering the barrier to application; and its prior use (i.e., in Sheldon & Filak, 2008) indicated it was generally successful in manipulating individual needs. Specifically, the original study's competence manipulation also influenced relatedness; the authors reasoned that statements made from the researcher's perspective (e.g., "I believe in you!") contributed to this confound (and were therefore removed in Study 1). This manipulation task was expected to produce

3 Research Design and Methodology 23

between-condition differences in competence satisfaction and frustration alike. The experimental procedure is described in full in section 4.1.4, and Appendix A contains the scripted false performance feedback and puzzles used in each condition.

Participants played Mark of the Ninja (Klei Entertainment, 2012), a side-scrolling stealth action game for thirty minutes (discussed further in section 4.1.3). In line with the game selection processes described earlier in this chapter, Mark of the Ninja was identified as offering strong competence support for most players, and relatively limited support for autonomy and relatedness.

3.2.1 Measures In the first survey, participants responded to demographic questions – specifically, their age (open text field; integers between 17-100 inclusive) and gender (open text field; 1-15 alphabetic characters), and level of prior experience with Mark of the Ninja (7-point item anchored with "not at all experienced" and "very experienced"). Demographic measures were collected to describe the sample.

After completing demographic questions, the first survey presented respondents with experiential measures of need satisfaction and frustration (Sheldon & Hilpert, 2012), positive and negative affect (Diener & Emmons, 1984), and vitality (Ryan & Frederick, 1997). The second (post-play) survey included the same experiential measures as the first, with additional PX-specific measures: in-game competence and autonomy satisfaction (Ryan et al., 2006), and intrinsic motivation to play Mark of the Ninja (Ryan et al., 1983). All experiential measures were selected to contribute to answering H1 or H2, in that they index a dimension of PX or need frustration (although need frustration is also a measure of PX). Measures of need satisfaction, need frustration, and short-term wellbeing were also required to answer H3. Intrinsic motivation was included in Study 1 to test RQ1. Further details about all Study 1 measures can be found in section 4.1.2.

Semi-structured interview questions focused on PX of the experimental play session (e.g., contributing factors to enjoyment, challenge, or affective responses), followed by questions about their experiences of restorative play in day-to-day life (e.g., recounting experiences of playing to feel better). The semi-structured interview format was used in both studies to allow for follow-up questions where appropriate. On the assumption that some participants would not identify play as necessarily improving

24 3 Research Design and Methodology

"wellbeing", the words "feel better / worse" were used for clarity (RQ2). The full list of interview questions can be found in Appendix A.

3.3 Study 2 Design Study 2 aimed to evaluate the effects of a prior experience of autonomy frustration on PX. As in Study 1, a study design with two conditions and pre-post measurement was employed. Data collection occurred from August-October 2018.

Figure 3: The tangram game played by participants in the experimental condition. The hint text box and surrounding shapes have been moved closer to the tangram outline for clarity.

Inspired by a previous study of autonomy frustration (Radel et al., 2014), the experimental manipulation took the form of a digital tangram game, whereby players manipulate seven standard shapes into representations of ordinary objects, which featured minor variations between conditions. Both versions of the game began with three tutorial puzzles that conveyed game rules and controls prior to displaying full-sized puzzles in a pre-determined sequence. After completing the tutorial, participants in the experimental condition were directed to "follow the hints provided to complete the puzzles" (Figure 3). The hint system took the form of a large animated arrow, which pointed at a blinking region inside the tangram set to correspond with the size and colour of an unused shape. Unauthorised moves were rejected: in such an event, the game simply returned the offending shape to its original position outside the tangram. In the experimental condition, playing the tangram game is therefore reduced to following instructions. Participants in the control condition heard an extra phrase in the scripted introduction to the game ("feel

3 Research Design and Methodology 25

free to play in whatever way makes sense to you"), and played a version of the tangram game without the hint system. The two conditions were otherwise identical.

Designing the experimental manipulation to occur entirely within the tangram game was preferred to the approach used in the original study (Radel et al., 2014), which required specialist equipment (i.e., a one-way mirror), further setup time (i.e., camera placement), or were readily integrated into the game design (i.e., directives about how to play). A digital tangram game was created for the study to minimise setup time, materials used (e.g., paper), and to more readily accommodate the design requirements of inducing autonomy frustration. The tangram game manipulation was expected to produce between- condition differences in autonomy satisfaction and frustration. The game was designed by the researcher, and developed in Unity (Unity Technologies ApS, 2017) by Joel Harman between December 2017 and July 2018. During this time, three iterations of the game were designed and playtested by the supervisory team, diverging from the initial design (e.g., removing aural directives), with additional feedback from the developer. The initial design brief used in the development of the tangram game can be found in Appendix C.

Figure 4: Spore (Maxis, 2008), the stimulus videogame selected for Study 2. The cell stage (shown here) features a number of game elements that collectively support autonomy.

Participants played Spore (Maxis, 2008), a simulation game with role-playing elements, for thirty minutes (discussed further in section 5.1.3). Spore was identified as offering strong autonomy support (e.g., by offering players a range of functional and aesthetic customisation options), less support for competence and relatedness, and low

26 3 Research Design and Methodology

probability of need frustration (e.g., because avatar death is barely penalised). SDT researchers have speculated that the voluntary nature of play may be compromised in laboratory settings (e.g., Ryan et al., 2006); however, formal tests of this assertion are yet to emerge. Regardless, the laboratory setting was considered unlikely to cause issues, as the requirement to play Spore was held constant across conditions, and agreement to participate in the study itself conveys a willingness to play Spore for thirty minutes – and, indeed, engage with any task listed on the associated information sheet.

3.3.1 Measures Measures that demonstrated unacceptably low reliability (ω < 0.6) in Study 1 – autonomy satisfaction, autonomy frustration, and negative affect (see section 4.4 for further details) – were replaced for Study 2. Scale development was not deemed necessary prior to this time, as BMPN items had seemed relatively applicable to PX. As PX-specific measures of need frustration remained unavailable, and other scales featured items of varying resemblance to PX, measures were created from existing need frustration scale items that were adapted to more accurately reflect PX. While unacceptable from a strictly psychometric perspective (Clark & Watson, 1995), this approach is consistent with prior SDT practice. The PENS, for example, includes a number of items adapted from existing measures, across subscales assessing competence (La Guardia, Ryan, Couchman, & Deci, 2000; Williams & Deci, 1996), autonomy (Williams & Deci, 2001), and relatedness (Ryan et al., 1983) alike.

First, the researcher compiled a list of potentially relevant competence and autonomy measures from the SDT literature (Bartholomew et al., 2011; Chen et al., 2015; La Guardia et al., 2000; Longo et al., 2016; Rocchi et al., 2017; Ryan et al., 1983; Sheldon & Hilpert, 2012). Items were adapted from these scales to form 13- and 14-item competence and autonomy frustration PX measures. In the next stage of evaluation, six items from both scales were removed either because they were oddly-phrased (and could not justifiably be altered further), or described behaviour rather than experience. Following a 30-minute interview about individual scale items with an undergraduate game design student, and with agreement from the supervisory team, the 5-item competence and autonomy frustration subscales used in Study 2 were finalised. (The entire process is described in greater detail in Appendix B, alongside the results of exploratory factor analyses conducted with Study 2 data.) In light of recent work in PX literature that outlines the many issues associated with unvalidated (Law, Brühlmann, &

3 Research Design and Methodology 27

Mekler, 2018) and ad-hoc (Mekler et al., 2014) measures, it is worth emphasising that these scales should not be used further in PX research without independent validation.

As in Study 1, the first survey in Study 2 included demographic questions; age and gender were again recorded via open text field (with appropriate context-specific constraints). Player skill was evaluated with two questions: first, the estimated number of hours spent playing videogames in an average week; and second, their age at the time of their first experience playing a videogame. The second measure was collected due to prior research (Hartanto, Toh, & Yang, 2016) indicating its potential utility for indexing player skill. Age and gender were collected only to describe the sample. Although the two indirect measures of player skill were intended for use in exploratory analyses, significant zero-order correlations with experiential measures were not observed, and they were not considered further. These zero-order correlations can be found in Appendix C.

In Study 1, restoration was hypothesised only in terms of need satisfaction (H1), and this was amended in Study 2 to facilitate investigation into changes in need frustration during play (section 5). Specifically, H1a predicted that need-frustrated participants, relative to need-satisfied participants, would experience a greater decline in need frustration during play. Aside from the addition of H4, as discussed below, no further changes were made to the hypotheses or research questions listed in section 3.1.1.

Experiential outcome measures (H2) included intrinsic motivation, vitality, happiness, calmness, depressed affect, anger, tension, and immersion (each discussed in further detail in section 5.1.2). Each experiential measure was assessed before and after the experimental play session. As in Study 1, it was necessary to assess positive and negative outcome measures in response to H3a and H3b, respectively. Immersion was measured to conduct exploratory tests of its utility as a moderator (section 2.2.2). In particular, H4 predicted that immersion would amplify the effects of need satisfaction on positive short-term wellbeing outcomes and intrinsic motivation.

Interview questions focused on autonomy satisfaction and frustration during past experiences of videogame play (e.g., "can you tell me about a time where you stopped playing a videogame because you didn't want to do what was required to progress"; see Appendix C for the full question list). These questions aimed to develop a more complex understanding of the ways that autonomy manifests in PX, and the extent to which participants freely associated autonomy satisfaction and wellbeing (RQ2). Questions regarding Spore were not pursued because the earliest sections of the game were not

28 3 Research Design and Methodology

considered to produce enough distinct experiences to support a viable line of inquiry. The second set of interview questions focused on wellbeing – participants were asked to describe wellbeing in their own words, and the extent to which videogame play (in general, or with particular videogames) was reflectively valued. These questions served as a contrast to Study 1, prompting reflection on past experiences of play to delve further into relations between eudaimonia (section 2.2.1) and the player experience (RQ2).

3.4 Approach

3.4.1 Procedure The procedures outlined below represent an overview of details common to both studies. Occasions where procedures differed are outlined in sections 3.2 and 3.3. More specific details are provided in chapters related to each study (see sections 4.1 and 5.1).

Recruitment and data collection processes were conducted on the Queensland University of Technology (QUT) campus. The primary aim of recruitment processes was to minimise barriers to participation. As travel can represent a substantial obstacle to participation, undergraduate students were primarily approached to participate, as they were considered most likely to travel to the campus for other reasons (e.g., to study). As such, students were approached via in-class announcements, word of mouth, and direct (i.e., verbal) interactions on campus. Interested candidates were sent information sheets prior to study participation, in line with approved ethical procedures (section 3.4.4). A random PC game key (valued at up to $20) was offered as compensation; in Study 2, participants enrolled in a first-year game design class were also offered (up to) 5% course credit in place of a one-page essay on an assigned reading. Participants were given the choice of compensation after study completion, using a separate survey form. Participation in the post-experiment interview was not rewarded.

Participants were thanked upon arrival at the laboratory, and informed that almost all further communication from the researcher would be delivered via pre-recorded audio. Participants were given an overview of their involvement in the study before providing informed consent. Following this, participants were given further information about the manipulation task, relative to their assigned condition. Participants were randomly assigned to conditions via random number generator, dependent on order of arrival. The manipulation tasks are described in further detail in sections 4.1.4 and 5.1.4.

3 Research Design and Methodology 29

Following the conclusion of the manipulation task, participants were asked to use the computer at their desk to complete the first survey, which contained demographic questions and experiential measures (see sections 3.2.1 and 3.3.1 for study-specific details). After completing the first survey, participants were instructed to begin playing the target game (see sections 4.1.4 and 5.1.4 for details), which was running in a minimised state on the same computer. After being informed that the researcher would remain behind a partition for the duration of play, participants began to play the target game from the beginning. After thirty minutes of play, participants were asked to close the game and complete the second survey, which contained PX measures. Participants were thanked again on completion, debriefed, and asked to participate in an optional interview of roughly fifteen minutes. Interview participation was only mentioned when the researcher was free of successive commitments (e.g., conducting another study). Everyone who was asked to participate in the interview agreed to do so. On average, participants took an hour to complete the study; for those interviewed, the total duration ranged between sixty-five and ninety minutes (depending on participant availability).

3.4.2 Videogame Selection Considerable effort was taken to select videogame stimuli that would maximise support for the basic need of interest (i.e., competence in Study 1; autonomy in Study 2), with minimal influence on other basic needs. In this way, changes in short-term wellbeing could be attributed to changes in satisfaction and frustration of the target need. Game design frameworks (Dena, 2017; Hunicke, LeBlanc, & Zubek, 2004) were adapted for this purpose; their flexibility provided sufficient guidance to approach game selection with an origin in PX (Tyack, Wyeth, & Klarkowski, 2018). Upon describing a desired PX in sufficient detail, associated player behaviours and game elements likely to support this PX can be identified. This series of desired game elements, play behaviours, and experiences becomes the foundation for ideation of particular videogames that pertain to most or all identified factors. This method is primarily limited by a research team's knowledge of videogame history and prior experiences of play; however, unfamiliarity can be overcome through collaboration with others (e.g., as in section 4.1.3). In the present research, the process of videogame selection was guided by conceptual SDT literature, which provided further guidance in the initial stages of describing the desired experience, emphasising its manifestions in PX. (Further study-specific details can be found in section 4.1.3, 5.1.3, Appendix A, and Appendix C.)

30 3 Research Design and Methodology

3.4.3 Analysis The approach to quantitative data analysis was broadly consistent across both studies. The few changes made between studies reflect the inclusion of new measures in Study 2, as well as a more informed position towards assumptions of normality and homoscedasticity in sample data. More specifically, Study 1 analyses were conducted on transformed data when distributional assumptions were violated. However, as analyses conducted on transformed and untransformed data did not substantively differ, analyses using untransformed data were reported regardless. In Study 2, all t-tests and between- within ANOVAs were conducted with robust tests, which are less sensitive to violations of distributional assumptions (Wilcox, 2012). Robust regressions were not used, as assumptions were satisfied in each case. (As Study 1 analyses were planned and conducted prior to changes to analysis in Study 2, it would have been inappropriate to conduct a revised post-hoc analysis with robust tests.) The Type I error rate (α = 0.05) was consistent across both studies, and lowered when conducting multiple comparisons.

In both studies, analysis began with independent samples Welch t-tests on measures of need satisfaction and frustration as a manipulation check. Hypothesis tests largely consisted of 2 x 2 between-within ANOVAs on need satisfaction (H1) and experiential outcome measures (H2); these tests were primarily concerned with interaction terms, which represent differences in relative change (e.g., wellbeing improvement) between conditions. Significant interaction terms were therefore followed by (Holm-corrected) dependent t-tests for both conditions, which quantified the degree of change in each relevant measure between pre- and post-assessments in either condition (H1, H2). While simple effects analyses are often used to probe interactions, they were not suitable here: a comparison of post-play group means would not necessarily demonstrate restoration, as participants in the control condition may also experience changes in need satisfaction or wellbeing during play. Such tests would not have provided further information as to the relative change in these outcomes during play for those in each condition.

Hierarchical linear regression models were constructed to test the extent to which need satisfaction and frustration could respectively predict positive (H3a) and negative (H3b) experiential outcomes. These models were constructed for each assessed post-play outcome: in each instance, Model 1 contained only the corresponding pre-play measure; in Model 2, the additional influence of post-play need satisfaction or frustration measures

3 Research Design and Methodology 31

was identified. Tests of the influence of need frustration on intrinsic motivation for a successive activity (RQ1) were specific to each study (section 4.2.5 and 5.2.8).

Analysis of interview responses (RQ2) proceeded by way of thematic analysis, largely according to Braun and Clarke (2006), with some exceptions that reflect the quantity of data collected (e.g., clear iterations were not observed in the process of generating themes, and themes were not explicitly derived as collections of codes). All interviews were audio-recorded and transcribed into Scrivener (Version 1.9.7; Blount et al., 2016) by the researcher. Data extracts from these transcripts were then sequentially coded, adding as many codes as were relevant to each extract. Data extracts were revisited when prominent new codes were identified. In Study 1, themes were derived via induction, as compelling observations were not immediately evident with an entirely theoretical approach (i.e., with foundations in SDT). To answer RQ2, the final analysis of Study 1 discussed themes in relation to relevant SDT concepts. In Study 2, the question list was more directly linked to SDT concepts; as a result, themes were derived with respect to SDT concepts from the start. Both semantic and latent themes were identified – in general, semantic themes referred to more descriptive comments, whereas latent themes referred to experiential or inferential remarks. Thematic analysis was chosen due to the method’s relative flexibility. Interviews were conducted largely to contextualise quantitative survey data collected during the experimental studies.

3.4.4 Research Ethics The research was approved as low-risk by the QUT Human Ethics Advisory Team (#1700000214). A variation was approved prior to Study 2 to reflect the new measures and compensation option. When recruiting for Study 2 (i.e., August-October 2018), the researcher was not involved in teaching the course for which credit was offered.

Information sheets were given to participants prior to arrival, and again prior to signing the informed consent form. Information sheets outlined the study procedure, videogame stimulus (e.g., genre, age rating), example survey items, and interview topics. The pre-recorded script emphasised that participants could withdraw from the study at any time without penalty. As Study 1 involved a false-feedback manipulation, participants were required to explain the deception in their own words after the debriefing to show they understood the true nature of the study. The manipulations in both studies required deception to experimentally induce need frustration. To minimise the risk of

32 3 Research Design and Methodology

harm, participants were informed of the possibility of experiencing mild discomfort during the study, and that they could leave at any time without consequence.

Survey data were stored in a non-identifiable format. Interviews were assigned codes (e.g., P1) rather than names, rendering them potentially re-identifiable. Contact details (e.g., collected to provide compensation) were requested in a discrete survey form, and these unlinked data were stored separately. Consent forms were stored in a locked filing cabinet on the QUT campus, and all other data were stored on password-protected QUT servers. Only the research team involved in the study had access to these data.

3 Research Design and Methodology 33

34 3 Research Design and Methodology

4 Study 1

Let me assume that I am told that some cows ruminate; I cannot infer logically from this that any particular cow does so, though I should feel some way removed from absolute disbelief, or even indifference to assent, upon the subject; but if I saw a herd of cows I should feel more sure that some of them were ruminant than I did of the single cow, and my assurance would increase with the numbers of the herd about which I had to form an opinion. Here then we have a class of things as to the individuals of which we feel quite in uncertainty, whilst as we embrace larger numbers in our assertions we attach greater weight to our inferences. –John Venn, The Logic of Chance (1988)

Study 1 investigated the restorative effects of videogame play following an experience of competence frustration. More specifically, Study 1 tested the extent to which participants who experienced competence frustration, rather than satisfaction, would experience greater improvement in competence satisfaction (H1) and short-term wellbeing (H2) during a successive experience of (need-satisfying) videogame play.

Recent work in self-determination theory (SDT; Ryan & Deci, 2017a; Van den Broeck et al., 2016) has emphasised the importance of considering both need satisfaction and frustration in research. However, their concurrent assessment and analysis remains unrealised in the experimental player experience (PX) literature. Study 1, therefore, represented an opportunity to assess the extent to which need satisfaction and frustration could respectively predict positive (H3a) and negative (H3b) short-term wellbeing outcomes.

Study 1 also examined the ongoing effect of need frustration on intrinsic motivation for a successive activity (RQ1), as recent work suggests (Fang et al., 2017; Radel et al., 2014). Post-experiment interviews built on prior qualitative research (e.g., Bourgonjon et al., 2016) aim to investigate player perspectives on need satisfaction, need frustration, and wellbeing as they relate to PX (RQ2). Study 1 therefore aimed to respond to the following hypotheses and research questions:

H1: Need-frustrated participants, relative to need-satisfied participants, will experience greater improvement in need satisfaction during play.

4 Study 1 35

H2: Need-frustrated participants, relative to need-satisfied participants, will experience greater improvement in short-term wellbeing and intrinsic motivation during play.

H3: Experiences of need satisfaction and frustration during play will distinctly predict positive and negative outcomes.

H3a: Experiences of need satisfaction during play will predict positive short- term wellbeing outcomes and intrinsic motivation.

H3b: Experiences of need frustration during play will predict negative wellbeing outcomes.

RQ1: To what extent can need frustration influence intrinsic motivation for a successive activity?

RQ2: How do players understand need satisfaction, need frustration, and wellbeing as they relate to videogame play?

4.1 Method

4.1.1 Recruitment A total of 73 participants (56 men, 15 women, and 2 non-binary people, aged between 18 to 46) were recruited for the study between July and October 2017; of these, 19 participants also contributed to the post-experiment interviews. The target sample size (for quantitative analysis) was at least 30 participants in either condition to meaningfully test between- and within-condition effects. The researcher aimed to conduct at least 10 post-experiment interviews. Participants were given the option of receiving a random PC game code for their time, valued at $20. Most participants were undergraduate students from Queensland University of Technology (QUT), recruited via in-class announcements; others were recruited via word of mouth, or directly approached on campus. Three participants in the control condition were excluded from analysis, as they failed to satisfy predetermined performance criteria (i.e., finding fewer than six words in either word grid during the manipulation task). One further participant in the control condition was excluded, as they indicated during the debriefing that their survey responses were heavily influenced by unrelated issues in their personal life, which were brought to mind while responding to vitality scale items (i.e., "I am looking forward to each new day" in Ryan & Frederick, 1997). In the experimental condition, four

36 4 Study 1

participants who correctly guessed the aim of the manipulation task without prompting were excluded. The final sample therefore consisted of 65 participants (control n=33, experimental n=32). Participants in the control condition were aged from 18 to 46 (µ = 23.97, σ = 5.53; 7 women, 25 men, 1 non-binary person). Participants in the experimental condition were aged from 18 to 46 (µ = 23.09, σ = 6.38; 7 women, 24 men, 1 non-binary person). Prior experience with the target game was rated on a 7-point scale between "not at all experienced" and "very experienced". Participants in the control condition (µ = 2.00, σ = 1.68) and experimental condition (µ = 1.41, σ = 1.37) reported comparably limited experience with the target game.1 (Approval from the QUT Human Research Ethics Committee was granted prior to data collection (#1700000214).

4.1.2 Measures Need satisfaction and frustration were measured before (T1) and after (T2) videogame play, using the 3-item subscales of the Balanced Measure of Psychological

Needs (BMPN; Sheldon & Hilpert, 2012). Competence satisfaction (CSBMPN) and frustration (CFBMPN), and autonomy satisfaction (ASBMPN) and frustration (AFBMPN) subscales were included. Descriptive statistics for each subscale can be found in Table 3. BMPN items were presented on 5-point scale between “no agreement” and “much agreement”. Example items from each subscale include “I did well even at the hard things” (CSBMPN), “I struggled doing something I should be good at” (CFBMPN), “I was free to do things my own way” (ASBMPN), and “I had a lot of pressures I could do without”

(AFBMPN). The BMPN was selected due to its status as a brief, independently validated (Cordeiro et al., 2016) measure of need satisfaction and frustration intended for use across multiple domains.

The Player Experience of Need Satisfaction (PENS; Ryan et al., 2006) instrument measured competence satisfaction (CSPENS) and autonomy satisfaction (ASPENS) after videogame play only, as some subscale items refer specifically to “the game” or “play”. As the word-finding task was not perceived to be particularly game-like or playful, it was introduced to participants only as a “task”, and the PENS manual does not encourage altering subscale items, it was deemed inappropriate for use prior to videogame play.

1 Post-hoc analyses were conducted in response to a examiner's question regarding potential relationships between prior experience with Mark of the Ninja and competence during the experimental play session. Prior game experience was not significantly associated with competence satisfaction, competence frustration, or indeed any experiential dimension assessed (ps ≥ 0.128, rs ≤ 0.191).

4 Study 1 37

PENS items were measured on a 7-point scale between “Do not agree” and “Strongly agree”. Example scale items include “I felt very capable and effective when playing”, and “I did things in the game because they interested me”, assessing competence and autonomy, respectively. PENS validation studies (Brühlmann & Schmid, 2015; D. Johnson & Gardner, 2010; D. Johnson, Gardner, & Perry, 2018) have generally supported the proposed structure of these subscales when familiarity with the target game cannot be assumed.

Short-term wellbeing was assessed at T1 and T2, with respect to affect and vitality. The former was represented with the Mood Rating Scale (Diener & Emmons, 1984), which features a 4-item positive affect (PA) subscale and 5-item negative affect (NA) subscale; the two measures are conceptually and statistically independent of each other. Items in both scales were assessed on a 7-point scale between “Not at all” and “Extremely”; sample items include “joyful” (PA) and “frustrated” (NA). Although other affective measures, such as the Positive and Negative Affect Schedule (PANAS; Watson, Clark, & Tellegen, 1988), are often used in PX research, the Mood Rating Scale was chosen for this study for consistency with the earliest SDT research on videogame play (Ryan et al., 2006).

Vitality was measured by the 6-item state level version of the Subjective Vitality Scale (Ryan & Frederick, 1997), which – on the recommendation of its independent evaluators (Bostic, Rubio, & Hood, 2000) – excludes the single reverse-scored item present in the original scale. Items were measured on a 7-point scale between “Not at all true” and “Very true”. Sample items include “At this moment, I feel alive and vital”, and "At this time, I have energy and spirit".

Intrinsic motivation was measured with the full 7-item Interest/Enjoyment subscale of the Intrinsic Motivation Inventory (IMI; Ryan et al., 1983), which has been independently validated (McAuley et al., 1989) and successfully adapted by the original authors (Ryan et al., 2006) and others (e.g., Birk, Mandryk, Miller, & Gerling, 2015) for use in the context of videogame play. Consistent with its previous use in PX research (e.g., Ryan et al., 2006), intrinsic motivation was only assessed after play. Items were rated on a 7-point scale between “Not at all true” and “Very true”. An example item from the (adapted) scale is “I enjoyed playing the game very much”.

38 4 Study 1

4.1.3 Videogame Stimulus The target game, Mark of the Ninja, was selected following identification of relevant factors in the PX and SDT literatures, and a more detailed analysis of the game itself (as discussed in section 3.4.2). Mark of the Ninja is a side-scrolling stealth action game in which players use a variety of tools (e.g., grappling hook, darts) to distract, evade, or kill enemies while accomplishing objectives and progressing through each level. The final aim of this process was to select a videogame that would consistently support competence regardless of player skill, while providing fewer opportunities to satisfy other basic needs. Previous SDT research (Ryan et al., 2006) identified game controls, progression structure, and genre as particularly relevant; in this view, competence is facilitated by “game controls [that] are intuitive and readily mastered", "tasks.. [that] provide ongoing optimal challenges and opportunities for positive feedback”, and the platform genre's “relatively limited choices over actions and environments.. [focusing] mainly on progressive challenges within a linear format” (Ryan et al., 2006, p. 351). Additional review of PX literature underscored the importance of selecting videogames that represent current design practice, with features (e.g., just-in-time tutorial prompts) that minimise time spent grappling with game rules or mechanics (Elson & Quandt, 2016; Klarkowski, Johnson, Wyeth, Smith, & Phillips, 2015; Tyack et al., 2018). A list of well- designed platform games was then compiled, in conversation with independent game developers (I. McLarty, T. Burdak, B. Keogh, L. Van Dyke, J. Thyer, personal communication, April 2017). Mark of the Ninja was selected from this list due to consistently positive reviews from both professional critics (Metacritic, 2013) and players (Valve Corporation, 2018), a diegetic and well-structured tutorial, frequent checkpoints, and an interface providing consistent feedback (e.g., numeric score, enemy sight cones, progress towards optional goals). The game's mostly linear opening levels, and infrequent presence of non-player characters, were elements identified as offering only moderate support for autonomy and relatedness, respectively.

4.1.4 Procedure Participants were welcomed and thanked for their attendance upon arrival, then introduced to the manipulation task (adapted from Sheldon & Filak, 2008): a simple word-finding puzzle. The task involved forming words of adjacent letters in a 4-by-4 grid over a three-minute period. The puzzles used in each condition can be found in Appendix A. Participants worked at two such puzzles, and were given false performance feedback

4 Study 1 39

after each one. This manipulation task was expected to produce between-condition differences in both competence satisfaction and frustration.

Participants were randomly allocated to the control or experimental condition via random number generator, dependent on order of arrival. In the control condition, the scripted introduction to the first puzzle emphasised that participants would likely improve their skills with effort, despite the apparent challenge (i.e., "just do the best you can, and you'll improve quickly"). After completing the first puzzle, participants were told they had done well, and the second puzzle would be more challenging as a result. After the second puzzle, participants were informed of their score, compared to an average score for the grid, and told the grid contained a total of 74 words. In truth, the first and second grids were comparably difficult (containing approximately 250 words each), and the purported average score was calculated for each participant as half their true score plus one.

In the experimental condition, the first puzzle was introduced as being "difficult", and participants were explicitly identified as beginners who were unlikely to perform well at the task (i.e., "still, do the best you can, even if it seems hard – maybe you’ll be lucky"). After the first puzzle, participants were told they had performed poorly, and the second puzzle would be easier as a result. After the second puzzle, participants were informed of their score, compared to an average score for the grid, and told the grid contained a total of 104 words. In truth, the two grids were comparably difficult (containing approximately 90 words each), and the purported average score was calculated as each participant's true score plus seven. All fabricated totals and average score calculations were derived in conversation with the supervisory team.

All participants completed the first survey directly after the manipulation task. Following this, participants played Mark of the Ninja, a 2D action platformer emphasising stealth, from the beginning for thirty minutes; the second survey was completed immediately afterwards. The researcher remained in the room behind a partition while participants completed both surveys and played Mark of the Ninja. Following completion of the second survey, participants were debriefed, and informed of the purpose of the experiment. The debriefing was conducted in stages to differentiate participants who had truly guessed the study’s aims, without assistance, from those who simply claimed similar insight after the fact. The option to participate in the post-experiment interview was then provided when the researcher was free of successive commitments (e.g., conducting

40 4 Study 1

another study). Everyone who was asked to participate in the interview agreed to do so. These semi-structured interviews ranged from 5 to 22 minutes, addressing the experience of playing Mark of the Ninja during the experimental session, and experiences of restorative play in day-to-day life (see section 3.2.1 for further details regarding question derivation, and Appendix A for the script). Participants took between 60 to 75 minutes to complete the study, depending on interview involvement.

4.2 Results

4.2.1 Assumption Checks Normality assumptions were satisfied for all variables (i.e., all Kolmogorov- Smirnov tests failed to achieve significance; all skewness and kurtosis z-scores < 2.58;

Field, Miles, & Field, 2012). Some variables – CSBMPN, CFBMPN, AFBMPN, and negative affect – did not satisfy Levene’s test of homogeneity of variance; Box-Cox transforms (Box & Cox, 1964) were applied to these variables (λs = 0.3, 0.8, -0.2, and 0.02, respectively) as a result. Transformed variables accordingly produced non-significant results for Levene’s test. All tests were conducted on both transformed and untransformed data, and substantive differences were not observed. For the sake of clarity, all results are reported in terms of untransformed data. Assumptions of linearity, normally distributed and independent errors, multicollinearity, and homoscedasticity were satisfied in each multiple regression analysis.

Preliminary work was conducted in R (Version 3.5.0; R Core Team, 2018): Kolmogorov-Smirnov tests were conducted using the “pgirmess” package; categorical omega (Kelley & Pornprasertmanit, 2016) was calculated with “MBESS”; and “car” was used to conduct Box-Cox transformations, and calculate Levene’s test values. All other analyses were performed using JASP (Version 0.8.5.1; JASP Team, 2017).

4 Study 1 41

4.2.2 Quantitative Analyses

—

CS = CS

Vitality

—

0.340

-

NA

(0.006)

—

0.480 0.549

valuesin parentheses).

-

PA

-

gical Needs, = PA positiveaffect,

(< 0.001) (< 0.001) (<

—

0.352 0.285 0.005

BMPN

- -

(0.004) (0.021) (0.972)

AF

—

0.175 0.512 0.154 0.296

BMPN

- -

(0.163) (0.219) (0.017)

AS

(< 0.001) (<

—

0.100 0.297 0.533 0.699 0.342

BMPN

- - -

(0.430) (0.016) (0.005)

CF

(< 0.001) (< 0.001) (<

.

—

0.664 0.458 0.256 0.753 0.495 0.501

BMPN

- - -

order correlations (Pearson'sr) (p T1of measures

(0.040)

-

CS

(< 0.001) (< 0.001) (< 0.001) (< 0.001) (< 0.001) (<

: Zero :

1

PA

NA

BMPN BMPN

BMPN BMPN

Vitality

CS CF

AS AF

Table competence satisfaction,= competenceCF frustration, = AS autonomy satisfaction, = AF autonomy frustration, = BMPN Balanced Measure of Psycholo NA = negative affect

42 4 Study 1

—

Int/Enj

n, = P/NA

—

0.689

PENS

(< 0.001) (<

AS

—

0.613 0.637

PENS

0.001)

CS

(< 0.001) (< (<

—

0.333 0.278 0.493

(0.007) (0.025)

(< 0.001) (<

Vitality

competence/autonomy satisfaction,= C/AF

—

0.196 0.290 0.236 0.294

- - - -

C/AS = C/AS

(0.117) (0.019) (0.058) (0.017)

NA

. .

—

0.158 0.421 0.590 0.522 0.714

-

(0.208)

PA

(< 0.001) (< 0.001) (< 0.001) (< 0.001) (<

—

valuesin parentheses)

-

0.184 0.188 0.079 0.099 0.344 0.196

- - - - -

BMPN

(0.143) (0.134) (0.529) (0.435) (0.005) (0.118)

AF

—

0.288 0.492 0.148 0.238 0.563 0.619 0.578

- -

BMPN

(0.020) (0.238) (0.056)

(< 0.001) (< 0.001) (< 0.001) (< 0.001) (<

AS

—

0.020 0.047 0.052 0.348 0.123 0.263 0.032 0.000

- - -

BMPN

(0.876) (0.708) (0.679) (0.005) (0.330) (0.034) (0.800) (0.999)

CF

—

0.174 0.481 0.001 0.412 0.130 0.254 0.653 0.463 0.363

- -

BMPN

(0.165) (0.994) (0.304) (0.041) (0.003)

(< 0.001) (< 0.001) (< 0.001) (< 0.001) (<

CS

order correlations (Pearson'sr) (p T2of measures

-

PA

NA

PENS

PENS

BMPN BMPN

BMPN BMPN

: Zero :

2

Int/Enj

CS

AS

Vitality

CS CF

AS AF

Table competence/autonomy frustration, = BMPN Balanced Measure of Psychological = Needs,Player PENS Experience Needof Satisfactio positive/negative affect, Int/Enj= Interest/Enjoyment.

4 Study 1 43

T1 T2 Measure µ σ ω 95% CI µ σ ω 95% CI

CSBMPN 2.421 1.063 0.903 0.830, 0.941 3.354 0.743 0.787 0.677, 0.854

CFBMPN 2.795 1.214 0.822 0.735, 0.885 2.764 0.807 0.605 0.325, 0.721

ASBMPN 3.108 0.834 0.528 0.176, 0.657 3.723 0.740 0.618 0.479, 0.756

AFBMPN 2.067 0.846 0.550 0.285, 0.727 1.974 0.703 0.625 0.416, 0.782 PA 3.865 1.395 0.909 0.854, 0.939 5.508 0.876 0.885 0.825, 0.925 NA 2.658 1.280 0.837 0.669, 0.910 2.089 0.626 0.480 0.141, 0.634 Vitality 3.936 1.206 0.893 0.829, 0.929 4.672 1.039 0.876 0.767, 0.918

CSPENS - - - - 5.128 1.230 0.833 0.728, 0.896

ASPENS - - - - 5.549 1.026 0.809 0.659, 0.880 Int/Enj - - - - 5.864 0.930 0.924 0.853, 0.954

Table 3: Descriptive statistics. Scale reliability estimates and 95% CIs represent categorical omega (ω) bootstrapped with 1000 replications. C/AS = competence/autonomy satisfaction, C/AF = competence/autonomy frustration, BMPN = Balanced Measure of Psychological Needs, PENS = Player Experience of Need Satisfaction, P/NA = positive/negative affect, Int/Enj = Interest/Enjoyment.

4.2.3 H1 – Need Satisfaction and Frustration A series of one-tailed independent samples t-tests were conducted to determine the effectiveness of the experimental manipulation (Table 4), which was expected to influence competence satisfaction and frustration.

Measure t df p δ 95% CI

CSBMPN 11.906 54.80 < 0.001 2.945 2.328, ∞

CFBMPN -8.006 52.49 < 0.001 -1.993 -∞, -1.461

ASBMPN 2.061 58.81 0.022 0.512 0.094, ∞

AFBMPN -2.370 47.92 0.011 -0.590 -∞, -0.165

Table 4: One-sided Welch's t-tests on need satisfaction and frustration measures, assessed after the manipulation task (T1). C/AS = competence/autonomy satisfaction, C/AF = competence/autonomy frustration, BMPN = Balanced Measure of Psychological Needs.

Between-condition differences were observed for both competence satisfaction (δ = 2.945) and frustration (δ = -1.993), indicating that the experimental manipulation was successful. Smaller effects also emerged for autonomy satisfaction and frustration, although these results are qualified by 95% CIs whose finite bounds approach zero, as well as the relatively low internal reliability of both measures (Table 3).

44 4 Study 1

It was expected that need-frustrated participants, relative to need-satisfied participants, would experience greater improvement in need satisfaction during play (H1). Results were consistent with this expectation: a significant interaction term was observed in a 2-way mixed ANOVA on CSBMPN (Table 5), and followed up with condition-specific dependent t-tests. As predicted, participants in the experimental condition experienced improvement, t(31) = 12.574, p < 0.001, δ = 2.223, 95% CI [1.567, 2.868]; in contrast, no discernible change was observed for participants in the control condition, t(32) = 0.300, p = 0.766, δ = 0.052, 95% CI [-0.290, 0.393].

Measure Effect F p ω2 Time 71.51 < 0.001 0.329

CSBMPN Group 46.70 < 0.001 0.413 Group * Time 64.09 < 0.001 0.305

Time 21.96 < 0.001 0.132

ASBMPN Group 2.26 0.137 0.019 Group * Time 2.40 0.127 0.010

Table 5: 2-way ANOVAs on measures of competence and autonomy satisfaction, assessed before and after playing Mark of the Ninja. C/AS = competence/autonomy satisfaction, BMPN = Balanced Measure of Psychological Needs.

An additional exploratory 2-way ANOVA was conducted for ASBMPN, as autonomy was unintentionally manipulated at T1. However, only the main effect of time was significant, indicating that the degree of improvement in autonomy satisfaction was comparable across conditions (Figure 5).

4 Study 1 45

Figure 5: 2-way ANOVA plots of competence and autonomy satisfaction, assessed before and after playing Mark of the Ninja. (○) → control condition; (●) → experimental condition. Error bars represent 95% CIs. The range of values shown on the y-axis varies for each plot.

4.2.4 H2 – Short-Term Wellbeing A set of 2-way mixed ANOVAs on positive affect, negative affect, and vitality (Figure 6) tested the prediction that participants whose needs were frustrated in the manipulation task would show the greatest improvement in short-term wellbeing during play (H2). Significant interaction terms were observed for all three wellbeing variables (Table 6); dependent t-tests show that need-frustrated participants saw increases in positive affect, t(31) = 11.650, p < 0.001, δ = 2.059, 95% CI [1.437, 2.671], and vitality, t(31) = 6.273, p < 0.001, δ = 1.109, 95% CI [0.661, 1.546], and a decline in negative affect, t(31) = -5.348, p < 0.001, δ = -0.945, 95% CI [-1.358, -0.522]. Need-satisfied participants experienced smaller increases in positive affect, t(32) = 3.811, p < 0.001, δ = 0.663, 95% CI [0.281, 1.037], but significant changes were not observed for vitality, t(32) = 1.181, p = 0.246, δ = 0.206, 95% CI [-0.141, 0.549], or negative affect, t(32) = 0.772, p = 0.446, δ = 0.134, 95% CI [-0.209, 0.476].

46 4 Study 1

Measure Effect F p ω 2 Time 122.40 < 0.001 0.432 Positive affect (PA) Group 32.18 < 0.001 0.324 Group * Time 33.63 < 0.001 0.170

Time 17.23 < 0.001 0.094 Negative affect (NA) Group 17.47 < 0.001 0.202 Group * Time 24.75 < 0.001 0.131

Time 31.65 < 0.001 0.105 Vitality Group 4.95 0.030 0.057 Group * Time 17.05 < 0.001 0.058

Table 6: 2-way ANOVAs on measures of short-term wellbeing, assessed before and after playing Mark of the Ninja.

Figure 6: 2-way ANOVA plots of short-term wellbeing measures (positive affect (PA), negative affect (NA), and vitality), assessed before and after playing Mark of the Ninja. (○) → control condition; (●) → experimental condition. Error bars represent 95% CIs. The range of values shown on the y-axis varies for each plot.

4.2.5 RQ1 – Intrinsic Motivation The extent to which need frustration could influence intrinsic motivation for subsequent videogame play was also investigated (RQ1), based on existing studies on this topic (Fang et al., 2017; Radel et al., 2014). Zero-order correlations failed to reveal any significant relation between post-play intrinsic motivation and competence frustration, either before (r = 0.162, p = 0.197) or after play (r < 0.001, p = 0.999). A two-tailed independent samples t-test on post-play intrinsic motivation was likewise inconclusive, t(61.98) = 0.094, p = 0.925, δ = 0.023, 95% CI [-0.463, 0.509]; however, the effect size suggests that finding a significant difference would not be particularly meaningful. This

4 Study 1 47

result is broadly consistent with those related to short-term wellbeing, which also showed between-condition similarities after play (H2).

4.2.6 H3 – Need Satisfaction and Frustration as Predictors Analyses were performed to determine support for the value of need satisfaction and frustration as effective predictors of positive (H3a) and negative (H3b) outcomes, respectively. Regression models were fitted to each T2 short-term wellbeing variable (positive affect, Table 7; negative affect, Table 8; and vitality, Table 9).

In each instance, Model 1 isolated the effect of the respective T1 measure; Model 2 quantified the additional effects of need satisfaction or frustration measures (as relevant). PENS measures of need satisfaction were used in these models because they more directly assess the player experience. In each model, the addition of need satisfaction or frustration measures led to significant improvement in explained variance, which – in line with results from other studies, and the broader theory (see section 2.2.3) – provides evidence of their value as predictors of positive and negative wellbeing outcomes.2

The effects of individual needs were less clearly evident: while both competence and autonomy satisfaction were associated with positive affect, competence frustration alone was shown to predict negative affect, and neither measure of need satisfaction emerged as a significant predictor of post-play vitality. Post-play vitality was more effectively predicted by its value prior to play than by in-game experiences of competence and autonomy satisfaction. However, it is inappropriate to generate confident conclusions about individual predictors given a sample of this size (Darlington & Hayes, 2016, pp. 520-521). Moreover, the inclusion of need satisfaction measures in Model 2 explained a significantly greater proportion of variance in post-play vitality, indicating that their combined relevance was somewhat obfuscated by their individual failure to reach significance.

2 In response to the question of a potential inverse relation between need satisfaction and negative short- term wellbeing, an additional model of negative affect was constructed with need satisfaction and frustration as predictors. Neither competence nor autonomy satisfaction was observed to significantly predict post-play negative affect in the presence of need frustration (see Appendix A).

48 4 Study 1

Model 1 R2 = 0.045, R2(adj) = 0.029, ΔR2 = 0.045, ΔF(1, 63) = 2.937, p = 0.091 Measure B SE β sr2 p (intercept) 4.995 0.318 - - < 0.001 Positive affect (T1) 0.133 0.077 0.211 0.045 0.091

Model 2 R2 = 0.440, R2(adj) = 0.413, ΔR2 = 0.396, ΔF(2, 61) = 21.570, p < 0.001 Measure B SE β sr2 p (intercept) 2.161 0.525 - - < 0.001 Positive affect (T1) 0.142 0.060 0.226 0.051 0.022

CSPENS 0.320 0.087 0.449 0.125 < 0.001

ASPENS 0.209 0.104 0.244 0.037 0.049

Table 7: Linear regression model of post-play (T2) positive affect. C/AS = competence/autonomy satisfaction, PENS = Player Experience of Need Satisfaction.

Model 1 R2 = 0.033, R2(adj) = 0.017, ΔR2 = 0.033, ΔF(1, 63) = 2.129, p = 0.150 Measure B SE β sr2 p (intercept) 1.854 0.178 - - < 0.001 Negative affect (T1) 0.088 0.061 0.181 0.033 0.150

Model 2 R2 = 0.171, R2(adj) = 0.130, ΔR2 = 0.138, ΔF(2, 61) = 5.088, p = 0.009 Measure B SE β sr2 p (intercept) 0.919 0.344 - - 0.010 Negative affect (T1) 0.071 0.058 0.146 0.021 0.221

CFBMPN 0.260 0.090 0.335 0.112 0.006

AFBMPN 0.133 0.105 0.150 0.022 0.209

Table 8: Linear regression model of post-play (T2) negative affect. C/AF = competence/autonomy frustration, BMPN = Balanced Measure of Psychological Needs.

4 Study 1 49

Model 1 R2 = 0.202, R2(adj) = 0.189, ΔR2 = 0.202, ΔF(1, 63) = 15.940, p < 0.001 Measure B SE β sr2 p (intercept) 3.147 0.399 - - < 0.001 Vitality (T1) 0.387 0.097 0.449 0.202 < 0.001

Model 2 R2 = 0.301, R2(adj) = 0.266, ΔR2 = 0.099, ΔF(2, 61) = 4.301, p = 0.018 Measure B SE β sr2 p (intercept) 1.544 0.708 - - 0.033 Vitality (T1) 0.368 0.093 0.427 0.181 < 0.001

CSPENS 0.182 0.115 0.215 0.029 0.119

ASPENS 0.135 0.137 0.133 0.011 0.329

Table 9: Linear regression model of post-play (T2) vitality. C/AS = competence/autonomy satisfaction, PENS = Player Experience of Need Satisfaction.

4.2.7 RQ2 – Thematic Analysis These interviews aimed to produce more developed insights into participant experiences with Mark of the Ninja, and their relationship with videogame play in general. Four themes pertaining to need satisfaction, need frustration, and wellbeing as they relate to PX (RQ2) were constructed from the interview transcripts. Two of these themes – habitualisation of virtual activity, and aligning with design intent – relate to the experimental play session, and the other two – play as self-abeyance, and authentic self- engagement – concern play as it occurs in ordinary life.

4.2.8 Habitualisation of Virtual Activity This theme describes the development of habitual knowledge of the augmented "playing body" (Keogh, 2018); a process of incorporating (and incorporation into) the instruments and affordances of physical and virtual interaction. More specifically, this theme describes experiences of becoming accustomed to the rules, mechanics, control layout, and (to some extent) intended experience of Mark of the Ninja (Keogh, 2018). As P19 summarised: "Learning all those things in a very short space of time, and trying to keep them all – you know, present in your mind. [...] trying to keep progressing, and killing, and finding, and saving, and all that kind of stuff." This playstyle was often described by participants who self-identified as not being a "gamer" (P17, P19), had trouble adjusting to the button mapping (P4, P5, P10, P13, P15, P18), or lacked genre

50 4 Study 1

experience (P1, P4, P11, P18). Participants discussed having to direct attention towards more foundational aspects of play (e.g., controls, P10, P17, P19; tactics, P18), which corresponded with feelings of intense concentration or focus (P15, P18, P19). However, some of these disadvantages were mitigated by Mark of the Ninja's tutorial and user interface:

P17: I think having like the base commands on the screen, that was useful, because – for a game that I haven't played before, it made it a lot easier, trying to work out what was going on. So having those base commands was awesome, but then also having a few times where.. you know, it pops up with, like "now to do this thing, you need to do this", and all that sort of thing, that also.. really helped, and then also.. when the game would sort of pop up, like "grappling hook would be useful here", like that helped!

Game elements were more often considered novel or "cool" as a result of participants' inexperience and learning-focused approach to the game. As P1 noted: "..because this is a new game, it's interesting to see how it plays out, a bit curious about the continuation, whether there will be like new tricks, right". Participants described trying new tactics in response to failure, and occasionally seemed confused as to why their actions (e.g., stealth kills) failed (P6, P17, P19).

AT: Can you give me an example of when the tutorial was confusing? P6: Oh, the sword attack? It took me a couple of goes to realise that there wasn't a specific button for the sword, like it only appeared at specific points, when you were still undetectable to the guard, that was a part that was a little bit confusing, and that's just because I was kind of like "oh, that's".. it makes sense now that I look back at it, it's just at that time, I kept pushing it [the X button], and I kept getting ki– I kept trying to do – I ended up doing the martial kick – like the martial arts and kicking the guard, and it didn't make sense as to why I wasn't just stabbing him, but that was just me being kind of bad at the game.

However, participants appreciated qualities of the game design that encouraged perseverance in the event of failure:

P5: ..there was a period of time where it took me like a minute to get into a vent because I kept forgetting the A button makes you jump and not sprint – like, obviously that makes me feel like a fucking idiot, but [laughs], but going into a room that has three guards, taking out a light, hiding, and then doing all of that, and kind of focusing on it – getting found and going "oh fuck", going down a vent and coming up somewhere else, having another crack at it, makes you feel more competent, because you achieve

4 Study 1 51

something. Where – like, you achieve something by having a different – the game actively making you think differently about it. P17: So I think the fact that.. it – if I, like, got killed by a guard or whatever, it just took me back like a couple of.. steps back, and it wasn't a big deal. I think that helped make it more enjoyable too. It wasn't frustrating, it was just like "oh ok, like I mucked that up, but I'll keep going".

For these players, the immediately salient aspects of playing Mark of the Ninja relate to the self-contained (and easily forgotten) experiences of practice, adaptation, and novelty associated with "learning how to play".

4.2.9 Aligning with Design Intent This theme reflects participants' attempts to play in accord with the (perceived) design intent, which is incentivised (explicitly, with points, and implicitly, with greater challenge), and assumed to be the most enjoyable way to experience the game. Participants' understanding of the design intent was a product of prior engagement with the game (P2, P5), or simply inference (P7, P8, P14).

AT: How did the game challenge you? P2: Well this actually plays into what we just discussed! It was my perception of that character and fulfilling that role is where I got the challenge, because it was the tutorial and everything was very easy, but it was.. the opportunities that were presented to me, so there was some degree of autonomy in play, where it's like – I could avoid this whole situation just by walking past or because I'm this person that's like losing their mind and a hired killer, that's what they're trained to do their whole life, I can put myself out on a limb and risk myself and the mission by trying to kill everyone [laughs]. And that was fun because it's all like time-based mechanics where I'm trying to line everything up, and execute and everything at the right time. That's where the fun was.

AT: Ok – I mean, I did ask about challenge– P2: –challenge. The fun and challenge. Well, there you go. I had fun when I was being challenged, and I was trying to challenge myself as much as possible.

The associated playstyle (i.e., in-game behaviours) recognises the ninja archetype as physically vulnerable, while possessing the potential for smooth or fluid movement (i.e., relative to other characters; P2, P7, P8), and the capacity to avoid detection – both of which are contingent on player skill.

52 4 Study 1

AT: Ok. What about the stealth makes it enjoyable, then, beyond like, being easy to do? P14: I guess it has this feeling that you are a ninja, and you're.. hiding and planning your attack. It's just how it makes you feel.

P13: –going back to the whole ninja thing, like yeah you do a lot of killing, but also, there's an approach to it where you – say you have the ability to infiltrate a place but not leave any traces whatsoever, like no-one would ever know that you were there, but even though you've gone in, you've taken a bunch of things, like, you've done stuff but like no-one would ever find out until like – let's say you killed someone important, they won't find out until the next morning, stuff like that.

In this theme, successful play was identified as maintaining the character of the ninja – as archetype and player character – and simply being detected could be considered failure (P2, P7, P12, P13, P14).

P2: I was joyful, but with intermittent moments of frustration, if I like fumbled something or didn't understand an instruction and later figured it out [...] I didn't feel like I was looking up to the character anymore, and I felt just as bad as everyone else [in the game world]".

AT: Ok. So how did you feel, like – I kind of got the impression that you didn't always succeed at that [going undetected]. P13: Yeah. AT: So how did that feel? P13: Definitely frustrating. I didn't feel too challenged in the way that you can – you can always rectify your mistake. So let's say if you get spotted, you can just run up and smack the dude a few times. But yeah, it's definitely frustrating, not in the sense that "oh like, oh, I lost the game", but more of like, "damn it, I like, I failed at my own code of trying to be like this perfect ninja".

For these participants, much of Mark of the Ninja is familiar; habitualisation is merely an adjustment to the already-learned fundamentals – controller use, tactics, timing – taught in other videogames. Progress is less inherently rewarding as a result – it seems almost inevitable – instead, these players are more invested in living up to the high-risk, high-reward fantasy associated with the ninja.

4 Study 1 53

4.2.10 Play as Self-Abeyance This theme reflects the (usually intentional) suspension of the self in relation with the outside world that occurs during videogame play. These experiences were most often described as a response to issues involving work (P1, P2, P3, P4, P11, P18) or formal education (P7, P9, P15). Participants who described seeking mental disengagement in response to stress elected to play games where they could "just chill out" (P18), more "passive" games that allowed engagement at the player's own pace (P2), or simulation games where "go[ing] a little bit blank" was no impediment to making visible progress (P4). As P9 explained:

P9: I do like combat-oriented games obviously, but in a stressful – if I'm in a very very stressful environment, I prefer to go to something more relaxed, and simplistic – so like, Animal Crossing (Nintendo EAD, 2005), Stardew Valley (Barone, 2016) – things that are a bit more mellow.

Escapism instead appeared to reflect an attempt to redirect attention elsewhere; to "try to concentrate on other things that are completely different" (P1), "not.. think about anything else" (P2), or engage with games that demand attention via "really brutal" difficulty (P16). For P3, escapism is not associated with any game in particular:

P3: Yeah ok, so if work was really stressful.. if I just had the kind of day where you do not want to think about work anymore, you're just so done, then there's – it [Rayman Legends (Ubisoft Montpellier, 2014)] gives me the opportunity to think about something else. And that's just the example I use, because that's what I'm playing at the moment, but there are times where I'll go home and play some Call of Duty [Modern Warfare Remastered] (Raven Software, 2016), because sometimes you just need to shoot some bad guys.

Experiences of escapism and disengagement alike were associated with a preference for more active physical (P5, P7) or social (P11) involvement in play. As P5 explained: "I had [The Legend of Zelda:] Breath of the Wild (Nintendo EPD, 2017) – Zelda game – which I find can be – is very peaceful, so I threw that on and just kind of switched off, but you know, got to do".

Whereas escapism suggests direction (i.e., away from the unpleasant situation), disengagement represents an attempt to maintain a kind of mental stasis. It is not clear, however, whether escapism and disengagement describe distinct experiences, or merely reflect semantic variation.

54 4 Study 1

4.2.11 Authentic Self-Engagement This theme pertains to wholehearted commitment to the experience of play, from emotional, personal, and intrinsically motivated perspectives. In other words: videogame play as the attempt to develop, rather than suspend, the player's relationship with the world. In this theme, participants described experiences of "satisfying catharsis" (P10), a sense of wonder at being in the virtual world (P14), and intense feelings of sadness:

AT: Can you tell me about a game that has done that [made you feel worse]? P3: Some of the storylines – again, because I watch all the cutscenes and get very invested in it. I once played through a game where – it wasn't a happy ending, you know, and that's kind of what I'm in it for, to win all of the medals, and everyone lives happily ever after. And like it was meant to be a bittersweet-type thing, but I ended up just crying, like I was sitting there sobbing, and it was a ridiculous game, and I shouldn't have been sad. But it was, but that was only because I was emotionally invested in it. And it's ok, it ended up having a happy ending, I just had to watch the next cutscene [laughs]. AT: [bemused] Great. Would you mind telling me what the game is? P3: [laughs] Yes, it's Pokémon Mystery Dungeon[: Explorers of Time/Darkness] (Chunsoft, 2008).

Play was considered personally significant when it caused participants to remember their reasons for entering game development (P2), evoked nostalgia (P9), or pushed the limits of their perseverance (P13). Finally, participants described a variety of reasons for feeling intrinsically motivated towards particular games, such as narrative investment (P8), being part of a performance of absurdity and exaggeration (P10), and long-term skill development:

P11: I really really enjoy watching my skill grow in a game, like, learning from mistakes and improving. I played counterstrike [Counter-Strike: Global Offensive] (Hidden Path Entertainment & Valve Corporation, 2012) for a couple of years, and when I started I was really shit, and when I finished I was.. technically above average this time, like I was looking at the ranking, like – I was in the high percentage of players' skill or whatever. So that – watching that difference in play was really really rewarding.

4 Study 1 55

The experiences described in this theme suggest an alternative relationship toward videogames and play; one that wholeheartedly embraces their potential to access and experience the self.

4.3 Discussion

4.3.1 H1 – Changes in Need Satisfaction Study 1 investigated the restorative effects of videogame play following an experience of competence frustration. As expected, participants in the experimental condition – whose need for competence was frustrated in the manipulation task – experienced greater improvement in competence satisfaction during play, relative to control (H1). This result is in line with SDT, which posits that need-thwarted individuals pursue opportunities to experience need satisfaction (Ryan & Deci, 2017c, p. 85). Notably, in both conditions, competence improved or remained stable during play (cf. Reinecke et al., 2012), indicating that the experience of videogame play was indeed beneficial. Although autonomy satisfaction and frustration were also manipulated, these effects were smaller (albeit "moderate" in Cohen's (1992) abstract categorisation), and both measures demonstrated somewhat mediocre reliability at T1 (ωs ≤ 0.550). It is therefore difficult to conclusively interpret these results.

4.3.2 H2 – Changes in Short-term Wellbeing Additionally, participants in the experimental condition, relative to control, experienced greater improvement in short-term wellbeing during play (H2). Given that need satisfaction and short-term wellbeing are strongly linked – theoretically and empirically (Reis, Sheldon, Gable, Roscoe, & Ryan, 2000; Ryan et al., 2006) – need- thwarted participants' substantial wellbeing gains (δs ≥ 0.945) should not be surprising. It is worth clarifying that need-thwarted participants did not experience "more" wellbeing – need frustration was not ultimately beneficial to those who experienced it – but their greater improvement meant that post-play levels of wellbeing were comparable across conditions (i.e., as shown in Figure 6).

4.3.3 RQ1 – Ongoing Effects of Need Frustration A significant between-condition difference on post-play intrinsic motivation was not observed (RQ1), and the estimated size of the effect (δ = 0.023) was close to zero. This result is consistent with SDT, which holds that need frustration is universally

56 4 Study 1

unfavourable (Ryan & Deci, 2017a), but contrasts with more recent research (Fang et al., 2017; Radel et al., 2014), which concludes that need frustration can erode intrinsic motivation for a subsequent activity in the same environment. It remains unclear why this study's results (and SDT generally) are inconsistent with the cited literature.

4.3.4 H3 – Need Satisfaction and Frustration as Predictors Multiple regression models indicated, in line with SDT, that need satisfaction and frustration measures respectively predict positive (H3a) and negative (H3b) outcomes. These results are consistent with those in other domains (Bartholomew et al., 2011; Gunnell et al., 2013), and a more recent cross-sectional survey of PX (Allen & Anderson, 2018), but were yet to be demonstrated in an experimental play context. Although the sample size did not allow for meaningful conclusions about individual predictors, the proportion of variance explained by each model was consistently improved by the addition of need satisfaction and frustration measures (ΔR2 ≥ 0.099). The total amount of variance explained in negative affect (R2(adj) = 0.130), however, was relatively small, which corresponded with the measure's mediocre post-play reliability (ω = 0.480). Despite this, competence frustration accounted for a sizeable proportion of the variance in post-play negative affect (sr2 = 0.112) – indeed, in this sense, competence outperformed autonomy in all three models. As comparative PX literature is somewhat scarce on this topic (e.g., due to widespread variation in study design and analysis in PX research; Mekler et al., 2014), determining the extent to which these results generalise requires further investigation.

4.3.5 RQ2 – Player Perspectives on Need Satisfaction, Need Frustration, and Wellbeing in PX Thematic analysis of coded interview transcripts identified distinct themes around competence and wellbeing in PX (RQ2), relative to the salient play context. The ways that experiences in the experimental play session were described seemed largely dependent on localised knowledge: participants who were less experienced with videogames, controllers, or genre conventions talked more about the process of habituating to these new forms; in contrast, more experienced participants discussed playing in ways that would fulfil particular experiences afforded by the game design. Themes identified with respect to videogames and wellbeing in ordinary life reflect two

4 Study 1 57

approaches to play – as a means to temporarily suspend, or wholeheartedly experience, the self in relation with its social context.

4.3.6 Competence in the Experimental Play Session SDT posits (as outlined in section 2.2.2) that videogames primarily satisfy competence by responding to player activity with immediate positive feedback, a clear sense of progress, variations in challenge over time, and an intuitive control scheme. These abstractions generalise across players and videogames; they assume, to some extent, a player whose prior habitualisation to games prepares them for other, similar ones; a player with "embodied literacies" (Keogh, 2018) through which videogame control schemes can be intuited. The results of this study, however, suggest that a broader view of player experience is produced by acknowledging the range of potential videogame performances (Jayemanne, 2017). In particular, this study indicates that players identify competence support in game designs in ways that reflect the application of their embodied literacies.

Less literate participants tended to focus on the functional experience of learning to operate (in) the game, and the aspects of Mark of the Ninja's design that supported the learning process. For these participants, the on-screen control prompts, clear tutorial, and frequent checkpoints elicited competence support. The competence satisfaction afforded by in-game progress was a result of these scaffolding elements. Accordingly, these players experienced competence frustration when their incomplete understanding of the game design (e.g., conditions for performing stealth kills) obstructed progress on multiple occasions.

Participants with more applicable literacies (e.g., playing 2D platformers or stealth games) described attempts to play in ways that reflected what they saw as designer intent – trying to embody the elegance and subtlety of the ninja. These participants found competence support in the fluidity of character movement, and the responsiveness of game input, which facilitated more complicated manoeuvres and helped prevent detection. These players experienced competence satisfaction when their enacted performances aligned with their idea of how a scene should play out, as informed by the ninja-as-avatar, the array of possible approaches enabled by the game design, and the limits of their own abilities. Conversely, need frustration occurred when these participants

58 4 Study 1

failed to meet these self-imposed standards for play, even when these "failed" performances resulted in further progress in the game.

This study indicates that players identify competence support in game elements that facilitate engagement at their present literacy. Whereas less literate players attend to aspects of design that convey the fundamentals of engagement, players with more closely- aligned literacies identify competence support in game elements through which performances may be altered to reflect design intent. Experiences of competence satisfaction and frustration are situated in players’ ongoing perception of success or failure at interpreting and enacting play as they understand it. These results elaborate the present characterisation of competence in SDT games literature (Ryan & Deci, 2017b), accounting for a greater proportion of the increasingly wide range of videogame players.

Crucially, the study highlights that we cannot assume that literate players and users will necessarily experience high competence satisfaction during interactive system use, or indeed be shielded from experiences of competence frustration. Whereas more literate participants were familiar with the game controls, for example, they nonetheless described in-game scenarios characterised by competence frustration. Our findings align with those observed in a recent PENS validation study (D. Johnson, Gardner, et al., 2018), which used data from a highly literate sample and derived a single Intuitive Controls and Competence factor.

4.3.7 Videogame Play in Ordinary Life The results of Study 1 indicate varied approaches to videogame play for wellbeing improvement in day-to-day life – as a means to suspend the self through experiences of immersion and escapism, or to engage the self in authentic experience. While descriptions of immersion in SDT are somewhat indeterminate (as discussed in section 2.2.2), the broader literature on PX (section 2.1.1) indicates that the experience of immersion (or escapism) is characterised by a redirection of attentional capacities towards the videogame environment (Jennett, 2010; Kuo et al., 2016). In this way, playing videogames in response to stress (to "zone out"; Oswald et al., 2014) represents an attempt to reproduce and extend life as it is, interrupting the overwhelming and undermining aspects of the external environment (e.g., work) that wear out the self (Berlant, 2011). This mode of engagement may support emotion-based coping (Folkman & Moskowitz, 2004) in the short-term, but is nonetheless difficult to characterise as "activity that

4 Study 1 59

develops and expresses one’s most reflectively valued and well integrated human potentialities" (Ryan et al., 2013) – in other words, activity that supports wellbeing. Seeking immersion in the virtual is therefore antithetical with SDT: indeed, the theory's emphasis on mindful awareness as a quality of wellbeing (section 2.2.1) more readily aligns with an approach to play characterised by open and honest self-expression. The Study 1 results therefore highlight the conceptual issues with an SDT-PX framework that contains both immersion and wellbeing.

While there were differences in the ways that participants described experiences of immersion or escapism, they shared a common impetus. Paid employment and formal education structure life such that work-related thoughts intrude into the home. In this way, these results are consistent with prior research involving videogame play as a response to work-related stressors (Collins & Cox, 2014; Reinecke, 2009). However, the consistency of the desired experience described in this study indicates that videogame play may be valuable simply because it is separate from work. Although videogames are often celebrated as a necessarily (inter)active medium, these results alternatively suggest their use as a means to engage the mind in the pursuit of mindlessness.

The experiences of authentic self-engagement in play described by some participants appear more strongly associated with wellbeing. In many ways, these results contrast with those associated with immersion: they indicate a relationship to videogame play characterised by wholeheartedly endorsed behaviour. When videogame play is considered merely one domain among others where self-endorsed action may occur, players appear to freely access and experience emotions, enact and reinforce personal values, and more fully experience the true self. While prior research has shown that experiences of the ideal self may emerge through play (Przybylski et al., 2012), these results additionally suggest that engaging the true self in virtual environments may, for some players, represent an equally gratifying experience.

4.4 Directions Study 1 demonstrated that videogame play could reverse the effects of a competence-thwarting event, and provided some insight into how players use videogames to support wellbeing in day-to-day life. While need satisfaction and frustration in play were observed to predict positive and negative wellbeing outcomes, these results were less clear for need frustration, and autonomy was only weakly related to all three

60 4 Study 1

wellbeing outcomes when considered alongside competence. Three scales – the BMPN measures of autonomy satisfaction and frustration, and negative affect – demonstrated mediocre reliability, which complicated their respective analyses. Tests involving intrinsic motivation were limited by its exclusion from the pre-play survey; similarly, as H1 only referred to need satisfaction, changes in competence frustration during play were not tested. Finally, while the false-feedback manipulation task successfully produced between-condition differences in competence satisfaction and frustration, smaller differences also emerged for autonomy satisfaction and frustration. The extent to which these unintended manipulations influenced study outcomes remains unclear. Moreover, the manipulation caused a considerable number of participants (n=7) to be excluded from analysis. Each of these results and limitations informed the design of the second study, in which autonomy became a central concern.

4 Study 1 61

5 Study 2

More weight. –Giles Corey, The Crucible (Miller 1953)

Study 2 investigated the restorative effects of videogame play following an experience of autonomy frustration. While the overall study design was unchanged from Study 1, Study 2 featured a new manipulation task, videogame stimulus, additional hypotheses (H1a, H4), and reformulated survey instruments (as described in section 3.3.1). These changes reflect the focus on autonomy in Study 2, and issues that arose in Study 1 (section 4.4). Accordingly, Study 2 tested the extent to which participants who experienced autonomy frustration, rather than satisfaction, would experience greater improvement in autonomy satisfaction (H1) and frustration (H1a) during a successive experience of videogame play. Study 2 likewise assessed the extent to which autonomy- frustrated participants, relative to control, would experience greater improvement in short-term wellbeing and intrinsic motivation during play (H2). Wellbeing improvement during play was tested with respect to increased positive short-term wellbeing and reduced negative short-term wellbeing.

As relationships between need frustration and negative outcomes in Study 1 were weaker than expected, Study 2 aimed to more thoroughly assess the extent to which need satisfaction and frustration would respectively predict positive (H3a) and negative (H3b) outcomes. Consistent with prior applications in player experience (PX) research (section 2.2.2), Study 2 additionally aimed to test the utility of immersion as a moderating variable. More specifically, it was expected that immersion would amplify the effects of need satisfaction on positive short-term wellbeing outcomes and intrinsic motivation (H4). As immersion is, theoretically, disrupted by "any thwarts to volition, efficacy, and connectedness" (Ryan & Deci, 2017c, p. 521), equivalent moderation analyses were not conducted for need frustration.

As evidence that need frustration could influence intrinsic motivation for a successive activity (RQ1) was not observed in Study 1, more thorough tests were conducted in Study 2 to more closely resemble those in existing literature (section 2.2.3). The post-experiment interviews in Study 2 were refocused into discussion of autonomy satisfaction and frustration in PX, and the ways that videogame play relates to wellbeing

5 Study 2 63

from a self-determination theory (SDT) perspective (RQ2). Study 2 therefore aimed to respond to the following hypotheses and research questions:

H1: Need-frustrated participants, relative to need-satisfied participants, will experience greater improvement in need satisfaction during play.

H1a: Need-frustrated participants, relative to need-satisfied participants, will experience a greater decline in need frustration during play.

H2: Need-frustrated participants, relative to need-satisfied participants, will experience greater improvement in short-term wellbeing and intrinsic motivation during play.

H3: Experiences of need satisfaction and frustration during play will distinctly predict positive and negative outcomes.

H3a: Experiences of need satisfaction during play will predict positive short-term wellbeing outcomes and intrinsic motivation.

H3b: Experiences of need frustration during play will predict negative short-term wellbeing outcomes.

H4: Immersion will amplify the effects of need satisfaction on positive short-term wellbeing outcomes and intrinsic motivation.

RQ1: To what extent can need frustration influence intrinsic motivation for a successive activity?

RQ2: How do players understand need satisfaction, need frustration, and wellbeing as they relate to videogame play?

5.1 Method

5.1.1 Recruitment A total of 151 participants (120 men, 30 women, and 1 person of undisclosed gender, aged between 17 and 51) were recruited for the study between August and October 2018. Eleven participants from this sample also contributed to the post- experiment interviews. The target sample size for quantitative analysis was 150 participants (i.e., 75 in each condition) to conduct more informative tests of individual regressors (H3) than Study 1 had allowed. As in Study 1, the researcher aimed to conduct at least 10 post-experiment interviews. All participants were given the option to receive a random PC game code as compensation. Participants in a first-year game design course

64 5 Study 2

could alternatively elect to receive up to 5% course credit (as discussed in section 3.3). In sum, 62 participants were compensated with a PC game code, 44 received course credit, and 45 chose to volunteer without receiving material compensation. Most participants were undergraduate students at Queensland University of Technology (QUT), recruited via in-class announcements or directly approached on campus; additionally, some members of the general public were recruited via word of mouth. Three participants were removed from the final sample, due to an error with the online survey (n=1), an unreplicated bug in the experimental stimulus (n=1), and admitting their prior participation in Study 1 (n=1). The final sample therefore consisted of 148 participants, with 74 in each condition. Participants in the control condition were aged from 17 to 45 (µ = 23.22, σ = 5.60; 14 women, 60 men), played videogames between 0 and 54 hours per week in the previous month (µ = 16.31, σ = 12.73), and first played videogames between ages 2 to 21 (µ = 6.97, σ = 2.95). Participants in the experimental condition were aged from 17 and 51 (µ = 22.49, σ = 5.37; 16 women, 57 men, 1 undisclosed), played videogames between 0 and 96 hours per week in the previous month (µ = 16.31, σ = 12.73), and first played videogames between ages 1 to 16 (µ = 6.82, σ = 3.07). Approval from the QUT Human Research Ethics Committee was granted prior to data collection (section 3.4.4).

5.1.2 Measures Intrinsic motivation was measured with the full 7-item interest/enjoyment subscale of the Intrinsic Motivation Inventory (IMI; Ryan et al., 1983). The scale has been independently validated (McAuley et al., 1989), and successfully used in prior studies of videogame play (Birk & Mandryk, 2013; Birk, Mandryk, et al., 2015; Tamborini et al., 2011). Items were rated on a 7-point scale between “Not at all true” and “Very true”, and modified to refer to prior activity (e.g., “I enjoyed playing the tangram game very much”).

Competence and autonomy satisfaction were assessed using the 3-item Mastery and Autonomy subscales of the Player Experience Inventory (PXI) instrument (Vanden Abeele, Nacke, Mekler, & Johnson, 2016; Vanden Abeele, Spiel, Nacke, Johnson, & Gerling, under review). Items were assessed on a 7-point scale between "Not at all true" and "Very true". Sample items include "I felt capable while playing the game" (competence satisfaction) and "I felt free to play the game in my own way" (autonomy satisfaction). Scale reliability was consistently high (see Table 10 for all descriptive statistics).

5 Study 2 65

Competence and autonomy frustration were measured with 5-item subscales adapted from existing instruments (as described in section 3.3.1). Items were introduced with the following text: "Please read each of the following statements carefully, thinking about how true it was for you while playing [the Tangram game / Spore]". Items were rated on a 7-point scale between “Not at all true” and “Very true”. Sample items included "I felt disappointed with my performance in the game" (competence frustration) and "My actions in the game felt like obligations" (autonomy frustration). Subscales are reported in full in Appendix B.

Vitality was measured with a 5-item version of the Subjective Vitality Scale (Ryan & Frederick, 1997). A negatively-worded item was removed on the basis of an independent validation study (Bostic et al., 2000). Separately, the item "I am looking forward to each new day" was not included because it caused issues in Study 1 (see section 4.1.1). Items were rated on a 7-point scale between “Not at all true” and “Very true”. An example item from the scale is "At this moment, I feel alive and vital". The scale has been successfully used in PX research before (Ryan et al., 2006).

A selection of 4-item subscales from the 32-item Brunel Mood Scale (Lane, Soos, Leibinger, Karsai, & Hamar, 2007) – happiness, calmness, depression, anger, and tension – were used to measure affect. Although these subscales are not commonly used in PX research, they are relatively brief and reliable measures that facilitate more specific predictions about positive and negative outcomes (H3). The scale was introduced with the text: "How did you feel while playing [the Tangram game / Spore]?", with items rated on a 5-point scale between "Not at all" and "Extremely". Sample items include "Happy", "Calm", "Depressed", "Angry", and "Anxious", respectively.

66 5 Study 2

T1 T2 Measure µ σ ω 95% CI µ σ ω 95% CI CS 4.982 1.263 0.756 0.657, 0.828 4.811 1.252 0.843 0.790, 0.880 AS 3.108 1.624 0.929 0.900, 0.950 5.869 1.203 0.905 0.852, 0.937 CF 1.973 1.157 0.864 0.807, 0.901 1.928 1.057 0.846 0.781, 0.889 AF 3.545 1.677 0.901 0.860, 0.924 2.176 1.095 0.852 0.782, 0.897 Happiness 3.301 0.760 0.761 0.679, 0.822 3.515 0.816 0.800 0.726, 0.855 Calmness 3.590 0.851 0.817 0.755, 0.861 3.341 0.878 0.833 0.764, 0.878 Anger 1.356 0.509 0.818 0.720, 0.898 1.328 0.538 0.773 0.601, 0.891 Depression 1.169 0.378 0.775 0.606, 0.896 1.137 0.342 0.816 0.613, 0.934 Tension 1.503 0.702 0.854 0.785, 0.898 1.564 0.731 0.858 0.795, 0.897 Int/Enj 4.942 1.220 0.931 0.905, 0.947 5.713 1.051 0.927 0.890, 0.947 Vitality 3.731 1.258 0.916 0.888, 0.934 4.161 1.312 0.916 0.878, 0.939 Immersion 5.140 1.308 0.796 0.726, 0.848 5.838 1.036 0.847 0.779, 0.886

Table 10: Scale characteristics when measured after playing the tangram game (T1) and Spore (T2). Composite reliability estimates and 95% confidence intervals are represented by categorical omega (ω) bootstrapped with 5000 replications. Int/Enj = interest/enjoyment, C/AS = competence/autonomy satisfaction, C/AF = competence/autonomy frustration.

Immersion was assessed using the 3-item Immersion subscale of the PXI (Vanden Abeele et al., 2016; Vanden Abeele et al., under review), a brief and face valid measure. Items were rated on a 7-point scale between "Not at all true" and "Very true"; a sample item is "I was immersed in the game".

5.1.3 Videogame Stimulus The target game, Spore, was selected to support and restore player autonomy, following a lengthy selection process (for an overview of the method, see section 3.4.2). First, conceptual definitions of autonomy from the SDT literature were collated (e.g., Deci & Ryan, 1987; Deci & Ryan, 2000), together with literature describing the ways autonomy is experienced during videogame play (Rigby & Ryan, 2007, 2011b; Ryan et al., 2006). Unfortunately, this summary proved largely unhelpful for creating a shortlist of suitable videogames – for example, an endless number of videogames "provide considerable flexibility over movement and strategies, choice over tasks and goals", and design rewards "to provide feedback rather than to control the player's behaviour" (Ryan et al., 2006). This material also indicated that a sizeable degree of experimental control

5 Study 2 67

had to be sacrificed to maximise the number of participants who would experience autonomy satisfaction during play. The game selection process remained inconclusive after specifying other relevant criteria (e.g., frequent checkpoints, simple controls) aimed at limiting confounds and maximising the potential sample size. The final game selection was made on the basis that the original conceptualisation of autonomy in PX (Ryan et al., 2006) appeared to refer largely to Spore, which was famously demonstrated (Wright, 2005) at the 2005 Game Development Convention (GDC) by its lead designer, Will Wright – described by SDT researchers as a "leading game developer" whose design approach "reflected.. a 'procedural' structure.. [that responds] dynamically to an individual's choices without constraining or anticipating them" (Ryan et al., 2006). As Spore also met the selection criteria as initially derived (see Appendix C), it was deemed suitable for Study 2.

Spore is a simulation game whereby players guide an organism's evolutionary process – from the beginning of its life as an amoeba, through to its (land-dwelling) descendents’ colonisation of outer space. The game's core feature is an adaptive character creation system, which uses procedural generation techniques to support a range of cosmetic and functional customisation options. In this mode, players use a drag-and-drop interface to change their evolving avatar (and in later stages, their society’s architecture and spaceship design), and immediately see the effects. Game controls are relatively straightforward – much of Spore can be easily navigated with just a mouse – and avatar death results in only minor penalties. Spore’s design privileges player creativity, and de- emphasises skills more traditionally associated with videogame play.

5.1.4 Procedure Participants were welcomed and thanked for their attendance upon arrival, then introduced to the tangram game, which served as the experimental manipulation (section 3.3). In both conditions, participants were informed that the game contained all necessary information to play, and their performance in the game was unimportant. Participants in the control condition heard an additional phrase emphasising autonomy ("feel free to play in whatever way makes sense to you") during this scripted introduction to the tangram game.

Afterwards, participants played a condition-specific variant of the tangram game, as described in section 3.3. This manipulation was expected to produce between-

68 5 Study 2

condition differences in autonomy satisfaction and frustration. After 10 minutes in the tangram game, participants completed the first survey (T1), which contained demographic questions and PX measures. Upon completion, participants played Spore for 30 minutes, then completed a second PX survey (T2). The researcher remained in the room, seated behind a partition, for the entire duration of the study. After completing the second survey, participants were debriefed, allowed to select their preferred form of compensation, and given the option to participate in the post-experiment interview when the researcher was free of successive commitments (e.g., conducting another study). Everyone who was asked to participate in the interview agreed. This semi-structured interview (as noted in section 3.3.1) discussed experiences of autonomy and wellbeing with respect to PX (see Appendix C for the initial list of interview questions). Participants took between 60 to 90 minutes to complete the study, depending on interview involvement.

5.2 Results

5.2.1 Assumption Checks Normality assumptions were assessed using Shapiro-Wilk tests, and satisfied for vitality at T1 and T2. Levene's test was also conducted on sample medians to assess homoscedasticity, which was satisfied for all measures except autonomy satisfaction (T1), competence frustration (T1), happiness (T1), and tension (T1).

As attempts to satisfy the assumptions of common parametric tests largely failed, all between-condition analyses were conducted using tests that are robust to departures from normality and homoscedasticity (as discussed in section 3.4.3). The WRS2 package (Mair & Wilcox, 2018) in R (Version 3.5.1; R Core Team, 2018) was used to conduct t- tests and two-way ANOVAs on 20% trimmed means, with the robust measure of effect size ξ (where ξ = 0.1, 0.3, and 0.5 may be considered small, moderate, and large effects, respectively; Mair & Wilcox, 2016). Scale reliability values were computed as categorical omega (a measure of composite reliability) using the MBESS package (Kelley, 2018). Moderation plots were created in R using the rockchalk (P. E. Johnson, 2018) package, with axis and legend labels added using GIMP (Version 2.10.6; Kimball, Mattis, & GIMP Development Team, 2018). Regression analyses were conducted in JASP (Version 0.9; JASP Team, 2019), and assumptions of linearity, normally distributed and independent errors, multicollinearity, and homoscedasticity were satisfied in all tests.

5 Study 2 69

—

Immersion

—

.001)

0.280

(<

Vitality

—

0.035 0.038

- -

(0.697) (0.678)

Tension

—

0.238 0.178 0.174

- -

(0.022) (0.060) (0.067)

Depression

—

.001)

0.424 0.219 0.046 0.082

- -

(0.030) (0.620) (0.380)

(<

Anger

Int/Enj= interest/enjoyment,= competence/autonomyC/AS

—

.001) .001)

0.334 0.246 0.390 0.174 0.181

- - -

(0.011) (0.039) (0.033)

(< (<

Calmness

—

.001) .001) .001)

0.340 0.137 0.234 0.172 0.316 0.319

- - -

values invalues parentheses).

(0.149) (0.016) (0.066) -

(< (< (<

Happiness

—

0.170 0.137 0.253 0.151 0.155 0.038 0.219

- - -

AF

(0.045) (0.104) (0.006) (0.109) (0.086) (0.646) (0.008)

—

.001) .001) .001)

0.142 0.132 0.355 0.391 0.187 0.396 0.041 0.039

- -

CF

(0.110) (0.150) (0.066) (0.643) (0.664)

(< (< (<

—

b) at T1 in the at b) controlT1 the condition in (p

-

.001)

0.201 0.410 0.258 0.127 0.090 0.031 0.080 0.059 0.181

- - -

AS

(0.027) (0.003) (0.141) (0.342) (0.747) (0.391) (0.488) (0.034)

(<

—

.001) .001) .001)

0.278 0.092 0.239 0.306 0.314 0.079 0.058 0.115 0.181 0.295

- - - - -

CS

(0.001) (0.307) (0.004) (0.400) (0.542) (0.211) (0.031)

(< (< (<

—

.001) .001) .001)

0.189 0.239 0.051 0.254 0.451 0.136 0.210 0.260 0.053 0.270 0.316

- - - - -

(0.023) (0.004) (0.562) (0.002) (0.103) (0.021) (0.006) (0.556)

(< (< (<

Int/Enj

order (Kendall'scorrelations tau

-

: Zero :

11

Int/Enj CS AS CF AF Happiness Calmness Anger Depression Tension Vitality Immersion Table satisfaction,= competence/autonomyC/AF frustration.

70 5 Study 2

—

Immersion

—

0.276

(0.001)

Vitality

—

0.025 0.052

- -

(0.771) (0.558)

Tension

—

0.184 0.068 0.081

- -

(0.059) (0.468) (0.397)

Depression

—

Int/Enj = interest/enjoyment,= C/AS

0.246 0.258 0.192 0.056

- -

(0.013) (0.005) (0.029) (0.529)

Anger

—

.001) .001)

0.291 0.234 0.356 0.290 0.245

- - -

(0.001) (0.014) (0.004)

(< (<

Calmness

valuesin parentheses).

-

—

.001) .001) .001)

0.357 0.102 0.142 0.045 0.449 0.298

- - -

(0.254) (0.133) (0.608)

(< (< (<

Happiness

—

.001)

0.129 0.283 0.193 0.112 0.198 0.080 0.104

- - - -

AF

(0.127) (0.028) (0.232) (0.022) (0.334) (0.218)

(<

—

.001) .001) .001) .001)

0.228 0.267 0.409 0.352 0.438 0.378 0.172 0.114

- - - -

CF

(0.006) (0.002) (0.039) (0.177)

(< (< (< (<

—

b) at T1 in the at b) experimentalT1 the in condition (p

-

.001) .001)

0.108 0.378 0.289 0.131 0.114 0.076 0.084 0.156 0.203

- - - -

AS

(0.197) (0.122) (0.197) (0.418) (0.335) (0.061) (0.017)

(< (<

—

(Kendall'stau

.001) .001) .001) .001)

0.161 0.403 0.117 0.433 0.359 0.223 0.298 0.256 0.264 0.359

- - - - -

CS

(0.056) (0.166) (0.012) (0.002) (0.003) (0.002)

(< (< (< (<

—

.001) .001) .001) .001) .001) .001) .001)

0.379 0.378 0.206 0.311 0.424 0.283 0.166 0.105 0.127 0.315 0.368

- - - - -

(0.013) (0.059) (0.261) (0.141)

(< (< (< (< (< (< (<

Int/Enj

order correlations

-

: Zero :

12

Int/Enj CS AS CF AF Happiness Calmness Anger Depression Tension Vitality Immersion Table competence/autonomy satisfaction,= competence/autonomyC/AF frustration.

5 Study 2 71

—

Immersion

—

0.298

0.001)

Vitality

(<

—

0.035 0.024

-

(0.563) (0.703)

Tension

—

0.415 0.138 0.170

- -

0.001)

(0.037) (0.013)

(<

Depression

—

Int/Enj = interest/enjoyment,= C/AS

0.390 0.208 0.091 0.050

- -

0.001)

(0.002) (0.150) (0.443)

Anger

(<

—

0.302 0.273 0.342 0.133 0.091

- - -

0.001) 0.001) 0.001)

(0.024) (0.133)

(< (< (<

Calmness

valuesin parentheses).

-

—

0.346 0.244 0.258 0.101 0.387 0.368

- - -

0.001) 0.001) 0.001) 0.001) 0.001)

(0.107)

(< (< (< (< (<

Happiness

—

0.204 0.087 0.169 0.157 0.130 0.237 0.178

- - - -

0.001) 0.001)

AF

(0.143) (0.008) (0.019) (0.036) (0.003)

(< (<

—

0.153 0.212 0.236 0.256 0.424 0.326 0.088 0.118

- - - -

0.001) 0.001) 0.001) 0.001) 0.001)

CF

(0.011) (0.140) (0.054)

T2, aggregated conditions across (p

(< (< (< (< (<

b) at b)

—

-

0.144 0.371 0.331 0.131 0.169 0.209 0.065 0.321 0.431

- - - - -

0.001) 0.001) 0.001) 0.001)

AS

(0.020) (0.031) (0.010) (0.002) (0.305)

(< (< (< (<

—

0.446 0.311 0.296 0.406 0.242 0.188 0.270 0.148 0.333 0.291

- - - - -

0.001) 0.001) 0.001) 0.001) 0.001) 0.001) 0.001) 0.001)

CS

(0.003) (0.018)

(< (< (< (< (< (< (< (<

—

0.417 0.432 0.159 0.242 0.489 0.125 0.175 0.203 0.044 0.391 0.507

- - - - -

0.001) 0.001) 0.001) 0.001) 0.001) 0.001)

(0.008) (0.033) (0.006) (0.002) (0.478)

Int/Enj

order (Kendall'scorrelations tau

(< (< (< (< (< (<

-

: Zero :

13

Int/Enj CS AS CF AF Happiness Calmness Anger Depression Tension Vitality Immersion Table competence/autonomy satisfaction,= competence/autonomyC/AF frustration.

72 5 Study 2

5.2.2 H1 – Changes in Need Satisfaction and Frustration Autonomy satisfaction and frustration were manipulated as intended (Table 14): participants in the experimental condition experienced lower autonomy satisfaction and greater autonomy frustration after playing the tangram game, relative to control. However, competence frustration was also manipulated, likely because the tangram game played in the experimental condition contained only nominal challenge.

Measure t df p ξ 95% CI Autonomy Satisfaction (AS) 6.642 81.70 < 0.001 0.697 0.553, 0.824 Autonomy Frustration (AF) 9.007 88.63 < 0.001 0.815 0.689, 0.911 Competence Satisfaction (CS) 1.598 88.66 0.1137 0.202 0.000, 0.447 Competence Frustration (CF) 6.037 55.59 < 0.001 0.663 0.469, 0.773

Table 14: Two sided t-tests (20% trimmed means) on need satisfaction and frustration measures. (Note: ξ is bounded at zero because its equation is enclosed by a square root.)

It was expected that participants in the experimental condition would experience a greater improvement in autonomy satisfaction (H1) and a greater decline in autonomy frustration (H1a) during play, relative to control. Two between-within ANOVAs were conducted to test this claim: significant interaction terms were observed for both autonomy satisfaction and frustration (Table 15), and followed up with condition-specific dependent t-tests. Significant improvement in autonomy satisfaction was observed for participants in both conditions; however, as expected, this increase was greater for participants in the experimental condition. In contrast, significant reductions in autonomy frustration were only observed for participants in the experimental condition. H1 was therefore supported.

5 Study 2 73

Measure Effect Q p T2 - T1 t p ξ

Time 374.192 < 0.001 Autonomy Control 9.721 < 0.001 0.75 Group 29.600 < 0.001 Satisfaction Experimental 17.988 < 0.001 0.87 Group * Time 25.582 < 0.001

Time 113.132 < 0.001 Autonomy Control -1.927 0.060 0.21 Group 48.968 < 0.001 Frustration Group * Time 62.554 < 0.001 Experimental -13.124 < 0.001 0.90

Time 1.118 0.294 Competence Control -3.432 0.001 0.33 Group 16.770 < 0.001 Frustration Experimental 4.064 < 0.001 0.45 Group * Time 24.622 < 0.001

Table 15: 2-way between-within ANOVAs on manipulated need satisfaction and frustration variables. Significant interactions were followed by dependent t-tests (T2-T1) on 20% trimmed means.

An exploratory 2-way ANOVA on competence frustration was also conducted (Figure 7), due to its unintended manipulation. Results (Table 15) indicated that participants in the control condition experienced a reduction in competence frustration while playing Spore, whereas participants in the experimental condition instead experienced greater competence frustration during play.

Figure 7: Participant autonomy satisfaction, autonomy frustration, and competence frustration measured before and after videogame play. (○) → control condition; (●) → experimental condition. Error bars represent 95% CIs. The range of values shown on the y-axis varies for each plot.

5.2.3 H2 – Change in Positive Short-term Wellbeing Outcomes and Intrinsic Motivation Study 2 assessed changes in short-term wellbeing and intrinsic motivation during play, with the expectation that participants in the experimental condition would experience greater improvement in all cases (H2). As the full list of outcome variables –

74 5 Study 2

four positive, and three negative – is somewhat unwieldy, results have been separately tabulated for clarity.

It was expected that participants in the experimental condition would experience greater improvement in happiness, calmness, vitality, and intrinsic motivation during videogame play, relative to control (H2). Between-within (2 x 2) ANOVAs (Table 16) showed significant interactions for vitality and intrinsic motivation (Figure 8).

Measure Effect Q p T2 - T1 t p ξ

Time 8.935 0.004 Happiness Group 1.493 0.225 Full Sample 3.060 0.003 0.22 Group * Time 3.071 0.083

Time 11.787 < 0.001 Calmness Group 0.814 0.369 Full Sample -3.791 < 0.001 0.23 Group * Time 0.899 0.346

Time 17.111 < 0.001 Control 0.613 0.542 0.05 Vitality Group 0.044 0.835 Experimental 5.383 < 0.001 0.43 Group * Time 7.278 0.008

Time 62.306 < 0.001 Intrinsic Control 1.255 0.216 0.13 Group 10.922 0.001 Motivation Experimental 9.198 < 0.001 0.74 Group * Time 39.637 < 0.001

Table 16: 2-way between-within ANOVAs on positive wellbeing outcomes and intrinsic motivation. Significant interactions were followed by dependent t-tests (T2-T1) on 20% trimmed means.

Follow-up analyses showed that significant improvements in vitality and intrinsic motivation were only observed for participants in the experimental condition. No significant changes in vitality and intrinsic motivation were observed for participants in the control condition. Significant interaction terms were not observed for happiness or calmness; however, significant main effects of time were followed up with dependent t- tests on the full participant sample to quantify the change in each variable during play. 2 2 More commonly-used ANOVA effect sizes (e.g., ηp , ω ) were not used because their value is influenced by qualities of study design (Okada & Hoshino, 2017). The results of these exploratory tests (Table 16) indicate that participants experienced greater happiness and lower calmness during play.

5 Study 2 75

Figure 8: Participant vitality and intrinsic motivation measured before and after videogame play. (○) → control condition; (●) → experimental condition. Error bars represent 95% CIs. The range of values shown on the y-axis varies for each plot.

5.2.4 H2 – Change in Negative Short-term Wellbeing Outcomes Participants in the experimental condition were expected to experience greater decreases in anger, depressed affect, and tension during videogame play, relative to control (H2). Between-within ANOVAs were conducted on each measure, and significant interactions for anger and tension were observed (Table 17).

Measure Effect Q p T2 - T1 t p ξ

Time 0.918 0.341 Control -2.167 0.036 0.21 Anger Group 8.787 0.004 Experimental 1.439 0.157 0.14 Group * Time 6.762 0.011

Time 0.545 0.463 Depression Group 2.783 0.099 – – – – Group * Time 0.545 0.463

Time 0.282 0.597 Control -1.231 0.225 0.10 Tension Group 3.632 0.061 Experimental 2.409 0.020 0.25 Group * Time 6.053 0.016

Table 17: 2-way between-within ANOVAs on negative wellbeing outcomes. Significant interactions were followed by dependent t-tests (T2-T1) on 20% trimmed means.

Follow-up tests on tension indicated that participants in the experimental condition experienced greater tension during play (ξ = 0.25). Significant changes in tension were

76 5 Study 2

not observed for participants in the control condition. While follow-up tests on anger suggest that participants in the control condition experienced less anger during play, the result was not significant when corrected for multiple comparisons (p < 0.025). Significant changes in depressed affect were not observed.

Some evidence for H2 was observed when positive outcomes were considered: participants in the experimental condition experienced improvement in vitality and intrinsic motivation, whereas no significant changes emerged for participants in the control condition. Against expectations, however, participants across conditions experienced similar rates of increased happiness and decreased calmness during play. As further evidence against H2, only participants in the experimental condition experienced increased tension while playing Spore. In aggregate, evidence for H2 was mixed.

Figure 9: Participant anger and tension measured before and after videogame play. (○) → control condition; (●) → experimental condition. Error bars represent 95% CIs. The range of values shown on the y-axis varies for each plot.

5.2.5 H3a – Need Satisfaction as Predictor Following mixed results in Study 1 (see sections 4.2.6 and 4.3.4), Study 2 tested the extent to which experiences of need satisfaction and frustration during play could respectively predict positive (H3a) and negative (H3b) PX outcomes. A series of multiple regression analyses were therefore conducted, with the expectation that need satisfaction would predict positive outcomes – happiness, calmness, vitality, and intrinsic motivation – and, conversely, that need frustration would predict negative outcomes – anger, depressed affect, and tension. In each case, Model 1 isolated the influence of the

5 Study 2 77

corresponding pre-play measure, and Model 2 quantified the additional effects of need satisfaction or frustration.

Model 1 R2 = 0.117, R2(adj) = 0.111, ΔR2 = 0.117, ΔF(1, 146) = 19.28, p < 0.001 Measure B SE β sr2 p (intercept) 2.304 0.283 - - < 0.001 Happiness (T1) 0.367 0.084 0.342 0.117 < 0.001

Model 2 R2 = 0.402, R2(adj) = 0.390, ΔR2 = 0.286, ΔF(1, 144) = 34.43, p < 0.001 Measure B SE β sr2 p (intercept) 0.585 0.331 - - 0.079 Happiness (T1) 0.262 0.070 0.244 0.058 < 0.001 Competence Satisfaction 0.246 0.054 0.378 0.085 < 0.001 Autonomy Satisfaction 0.150 0.056 0.221 0.030 0.008

Table 18: Linear regression model of post-play happiness

Model 1 R2 = 0.225, R2(adj) = 0.220, ΔR2 = 0.225, ΔF(1, 146) = 42.463, p < 0.001 Measure B SE β sr2 p (intercept) 1.583 0.277 - - < 0.001 Calmness (T1) 0.490 0.075 0.475 0.225 < 0.001

Model 2 R2 = 0.285, R2(adj) = 0.271, ΔR2 = 0.060, ΔF(1, 144) = 6.054, p = 0.003 Measure B SE β sr2 p (intercept) 1.154 0.369 - - 0.002 Calmness (T1) 0.415 0.076 0.403 0.147 < 0.001 Competence Satisfaction 0.205 0.065 0.292 0.050 0.002 Autonomy Satisfaction -0.049 0.066 -0.067 0.003 0.455

Table 19: Linear regression model of post-play calmness

78 5 Study 2

Model 1 R2 = 0.239, R2(adj) = 0.234, ΔR2 = 0.239, ΔF(1, 146) = 45.93, p < 0.001 Measure B SE β sr2 p (intercept) 2.257 0.296 - - < 0.001 Vitality (T1) 0.510 0.075 0.489 0.239 < 0.001

Model 2 R2 = 0.440, R2(adj) = 0.429, ΔR2 = 0.201, ΔF(1, 144) = 25.87, p < 0.001 Measure B SE β sr2 p (intercept) -0.242 0.448 - - 0.590 Vitality (T1) 0.410 0.067 0.393 0.147 < 0.001 Competence Satisfaction 0.295 0.084 0.281 0.048 < 0.001 Autonomy Satisfaction 0.248 0.087 0.228 0.032 0.005

Table 20: Linear regression model of post-play vitality

Model 1 R2 = 0.013, R2(adj) = 0.006, ΔR2 = 0.013, ΔF(1, 146) = 1.883, p = 0.172 Measure B SE β sr2 p (intercept) 5.233 0.361 - - < 0.001 Intrinsic Motivation (T1) 0.097 0.071 0.113 0.013 0.172

Model 2 R2 = 0.402, R2(adj) = 0.390, ΔR2 = 0.390, ΔF(1, 144) = 46.955, p < 0.001 Measure B SE β sr2 p (intercept) 2.030 0.437 - - < 0.001 Intrinsic Motivation (T1) 0.093 0.056 0.108 0.012 0.097 Competence Satisfaction 0.248 0.069 0.295 0.053 < 0.001 Autonomy Satisfaction 0.346 0.072 0.396 0.096 < 0.001

Table 21: Linear regression model of post-play intrinsic motivation

Both competence and autonomy satisfaction were shown to predict happiness (Table 18), vitality (Table 20), and intrinsic motivation (Table 21), and the total variance explained in each of these models is relatively large (all R2 > 0.4). However, only competence emerged as a relevant predictor of calmness (Table 19), and less variance is explained by the model (R2 = 0.285).

5 Study 2 79

5.2.6 H3b – Need Frustration as Predictor

Model 1 R2 = 0.124, R2(adj) = 0.118, ΔR2 = 0.124, ΔF(1, 146) = 20.60, p < 0.001 Measure B SE β sr2 p (intercept) 0.823 0.119 - - < 0.001 Anger (T1) 0.372 0.082 0.352 0.124 < 0.001

Model 2 R2 = 0.260, R2(adj) = 0.245, ΔR2 = 0.137, ΔF(1, 144) = 13.30, p < 0.001 Measure B SE β sr2 p (intercept) 0.533 0.130 - - < 0.001 Anger (T1) 0.257 0.079 0.243 0.054 0.001 Competence Frustration 0.177 0.039 0.347 0.105 < 0.001 Autonomy Frustration 0.048 0.037 0.098 0.009 0.191

Table 22: Linear regression model of post-play anger

Model 1 R2 = 0.045, R2(adj) = 0.038, ΔR2 = 0.045, ΔF(1, 146) = 6.832, p = 0.010 Measure B SE β sr2 p (intercept) 0.914 0.090 - - < 0.001 Depression (T1) 0.191 0.073 0.211 0.045 0.010

Model 2 R2 = 0.239, R2(adj) = 0.223, ΔR2 = 0.194, ΔF(1, 144) = 18.325, p < 0.001 Measure B SE β sr2 p (intercept) 0.692 0.095 - - < 0.001 Depression (T1) 0.093 0.068 0.103 0.010 0.174 Competence Frustration 0.131 0.025 0.405 0.141 < 0.001 Autonomy Frustration 0.038 0.024 0.123 0.014 0.106

Table 23: Linear regression model of post-play depressed affect

80 5 Study 2

Model 1 R2 = 0.213, R2(adj) = 0.208, ΔR2 = 0.213, ΔF(1, 146) = 39.580, p < 0.001 Measure B SE β sr2 p (intercept) 0.842 0.127 - - < 0.001 Tension (T1) 0.481 0.076 0.462 0.213 < 0.001

Model 2 R2 = 0.277, R2(adj) = 0.262, ΔR2 = 0.064, ΔF(1, 144) = 6.378, p = 0.002 Measure B SE β sr2 p (intercept) 0.617 0.157 - - < 0.001 Tension (T1) 0.391 0.078 0.376 0.126 < 0.001 Competence Frustration 0.184 0.053 0.267 0.060 < 0.001 Autonomy Frustration 0.002 0.049 0.003 < 0.001 0.970

Table 24: Linear regression model of post-play tension

In contrast with models of positive outcome measures, only competence frustration emerged as a significant predictor of post-play anger (Table 22), depressed affect (Table 23), and tension (Table 24). Explained variance in each model (0.200 < R2 < 0.300) was closer to typical values in PX research (Birk, Atkins, et al., 2016). As in Study 1 (section 4.2.6), autonomy frustration was not observed to predict negative outcomes, despite stronger scale reliability (ωT2 = 0.852), a larger sample (n = 148), and the inclusion of 3 negative outcome measures.3

5.2.7 H4 – Moderation By Immersion Tests were conducted to assess the utility of immersion as a moderating variable, as existing SDT-PX literature suggests (section 2.2.2). In line with the results of these prior studies, immersion was expected to amplify the effects of need satisfaction on happiness, calmness, vitality, and intrinsic motivation (H4). In these tests, interaction terms were individually added to multiple regression models of each outcome variable. Significant interactions with autonomy were not observed in models of calmness and vitality. No significant interactions between competence and immersion were observed.

3 The potential inverse relation between need satisfaction and negative short-term wellbeing was again explored through models of anger, depressed affect, and tension with both need satisfaction and frustration as predictors. As in Study 1, neither competence nor autonomy satisfaction was observed as to significantly predict negative wellbeing outcomes in the presence of need frustration (see Appendix C).

5 Study 2 81

In the interest of concision, models where significant interactions did not occur are reported in Appendix C.

Model 1 R2 = 0.488, R2(adj) = 0.473, ΔR2 = 0.488, ΔF(4, 143) = 34.02, p < 0.001 Measure B SE β sr2 p (intercept) -0.191 0.346 - - 0.583 Happiness (T1) 0.231 0.066 0.215 0.044 < 0.001 Competence 0.242 0.050 0.370 0.082 < 0.001 Autonomy 0.035 0.057 0.051 0.001 0.541 Immersion 0.270 0.055 0.343 0.085 < 0.001

Model 2 R2 = 0.525, R2(adj) = 0.508, ΔR2 = 0.037, ΔF(1, 142) = 11.06, p = 0.001 Measure B SE β sr2 p (intercept) -3.537 1.060 - - 0.001 Happiness (T1) 0.265 0.064 0.246 0.057 < 0.001 Competence 0.253 0.049 0.388 0.090 < 0.001 Autonomy 0.630 0.187 0.929 0.038 < 0.001 Immersion 0.871 0.188 1.106 0.071 < 0.001 Autonomy * Immersion -0.109 0.033 -1.458 0.037 0.001

Table 25: A multiple regression model of post-play happiness with a significant interaction between autonomy (satisfaction) and immersion.

82 5 Study 2

Model 1 R2 = 0.554, R2(adj) = 0.541, ΔR2 = 0.544, ΔF(4, 143) = 44.357, p < 0.001 Measure B SE β sr2 p (intercept) 0.772 0.420 - - 0.068 Intrinsic Motivation (T1) 0.046 0.049 0.053 0.003 0.350 Competence 0.231 0.060 0.275 0.046 < 0.001 Autonomy 0.151 0.068 0.173 0.015 0.029 Immersion 0.466 0.067 0.459 0.151 < 0.001

Model 2 R2 = 0.569, R2(adj) = 0.554, ΔR2 = 0.016, ΔF(1, 142) = 5.200, p = 0.024 Measure B SE β sr2 p (intercept) -2.063 1.311 - - 0.118 Intrinsic Motivation (T1) 0.065 0.049 0.076 0.005 0.183 Competence 0.245 0.059 0.292 0.052 < 0.001 Autonomy 0.652 0.230 0.747 0.024 0.005 Immersion 0.971 0.231 0.957 0.053 < 0.001 Autonomy * Immersion -0.092 0.040 -0.955 0.016 0.024

Table 26: A multiple regression model of post-play intrinsic motivation with a significant interaction between autonomy (satisfaction) and immersion.

Although the addition of an interaction term significantly improved proportions of variance explained by models of happiness (Table 25) and intrinsic motivation (Table 26), inspection of moderation plots (Figure 10) revealed that immersion dampened, rather than amplified, the effect of autonomy on these variables.

Figure 10: Predicted happiness and intrinsic motivation as a product of autonomy satisfaction and immersion. Low and high levels of immersion correspond to 1st and 3rd quartile values. The range of values shown on the y-axis varies for each plot.

5 Study 2 83

As autonomy satisfaction and immersion were moderately correlated at T2 (τ = 0.431), exploratory tests were conducted to distinguish moderation from potential curvilinear effects in each model (as recommended in Darlington & Hayes, 2016, pp. 430- 432). In other words, models resembling those shown in Table 25 and Table 26 were constructed, except that the interaction term in Model 2 was replaced with the square of autonomy satisfaction at T2 (Table 27 and Table 28).

Model 2 R2 = 0.506, R2(adj) = 0.488, ΔR2 = 0.018, ΔF(1, 142) = 5.211, p = 0.024 Measure B SE β sr2 p (intercept) -1.935 0.837 - - 0.022 Happiness (T1) 0.269 0.067 0.250 0.056 < 0.001 Competence 0.235 0.050 0.360 0.077 < 0.001 Autonomy 0.696 0.295 1.026 0.019 0.020 Immersion 0.288 0.055 0.365 0.095 < 0.001 Autonomy2 -0.065 0.028 -0.994 0.018 0.024 Table 27: A multiple regression model of post-play happiness with a significant curvilinear effect of autonomy (satisfaction).

Model 2 R2 = 0.571, R2(adj) = 0.556, ΔR2 = 0.017, ΔF(1, 142) = 5.637, p = 0.019 Measure B SE β sr2 p (intercept) -1.411 1.008 - - 0.164 Intrinsic Motivation (T1) 0.074 0.049 0.086 0.007 0.134 Competence 0.230 0.059 0.274 0.046 < 0.001 Autonomy 0.975 0.354 1.116 0.023 0.007 Immersion 0.486 0.066 0.479 0.162 < 0.001 Autonomy2 -0.081 0.034 -0.963 0.017 0.019 Table 28: A multiple regression model of post-play intrinsic motivation with a significant curvilinear effect of autonomy (satisfaction).

These results indicate that curvilinear effects of autonomy satisfaction are also plausible in either model. SDT provides no theoretical grounds to argue that moderation is more likely (see section 2.2.2), and the empirical literature contains examples of both moderation (Przybylski et al., 2012) and curvilinearity (Allen & Anderson, 2018). Therefore, these results represent no support for H4, and only nominal evidence for immersion as a moderating factor in the context of PX.

84 5 Study 2

It is nonetheless worth noting that the main effect of immersion produced significant improvement in explained model variance for intrinsic motivation (ΔR2 = 0.151), happiness (ΔR2 = 0.085), and vitality (ΔR2 = 0.030). The conceptual relevance of these main effects for SDT remains unclear.

5.2.8 RQ1 – Ongoing Effects of Need Frustration Study 2 investigated the paradoxical results of prior research (elaborated in section 2.2.3), which found that exposure to a controlling context in one task could influence intrinsic motivation for a subsequent activity (RQ1). In the original research (Radel et al., 2014), the first study demonstrated this relationship in a regression model that also accounted for intrinsic motivation in the first activity and (perception of) exposure to a controlling context in the second task. A second study showed that participants exposed to a controlling context during one task were more intrinsically motivated for a subsequent activity than participants exposed to a neutral context, but only when the signifiers of control were no longer present. When both activities took place in the same room, with the same researcher present (i.e., the conditions of the present study), the effect was reversed. Study 2 therefore conducted similar tests to investigate the extent to which experiences of autonomy frustration during the manipulation task (T1) subsequently influenced intrinsic motivation during play (T2), and whether between-condition differences in T2 intrinsic motivation could be observed.

Intrinsic Autonomy Intrinsic Autonomy

Motivation (T1) Frustration (T1) Motivation (T2) Frustration (T2) Intrinsic — Motivation (T1) Autonomy -0.581 — Frustration (T1) (< 0.001) Intrinsic 0.113 0.048 — Motivation (T2) (0.172) (0.559) Autonomy -0.016 -0.036 -0.440 — Frustration (T2) (0.845) (0.668) (< 0.001)

Table 29: Zero-order correlations (Kendall's tau b; p-values in parentheses) of autonomy frustration and intrinsic motivation.

5 Study 2 85

Model 1 R2 = 0.032, R2(adj) = 0.019, ΔR2 = 0.032, ΔF(2, 145) = 2.421, p = 0.092 Measure B SE β sr2 p (intercept) 4.426 0.592 - - < 0.001 Intrinsic Motivation (T1) 0.183 0.086 0.213 0.030 0.036 Autonomy Frustration (T1) 0.108 0.063 0.172 0.020 0.089

Model 2 R2 = 0.218, R2(adj) = 0.202, ΔR2 = 0.186, ΔF(1, 144) = 34.222, p < 0.001 Measure B SE β sr2 p (intercept) 5.495 0.564 - - < 0.001 Intrinsic Motivation (T1) 0.162 0.078 0.189 0.023 0.039 Autonomy Frustration (T1) 0.089 0.057 0.143 0.013 0.118 Autonomy Frustration (T2) -0.414 0.071 -0.432 0.186 < 0.001

Table 30: A multiple regression model of post-play (T2) intrinsic motivation, as predicted by experiences of autonomy frustration in both games, accounting for T1 intrinsic motivation.

While these results (Table 29 and Table 30) broadly resemble those in the study of interest (Radel et al., 2014), no ongoing effect of autonomy frustration on intrinsic motivation in a later activity was observed – in other words, experiencing autonomy frustration in the tangram game did not significantly predict intrinsic motivation after playing Spore (p = 0.118, sr2 = 0.013). (These results held even when each condition was analysed separately.) Moreover, between-condition differences in intrinsic motivation were not observed, t(89.91) = 1.396, p = 0.166, ξ = 0.168, 95% CI [0, 0.388]. Indeed, although this difference was not significant, autonomy-frustrated participants actually experienced greater intrinsic motivation (µ = 5.826, σ = 0.983) than participants in the control condition (µ = 5.600, σ = 1.110), which is inconsistent with findings in the target study. While acknowledging that the present study differs in some respects from the original work, no support was observed for its results.

5.2.9 RQ2 – Thematic Analysis Three themes – concerning autonomy support in the partnership between players and videogames, the production of pressuring videogame experiences, and wellbeing in relation to play and ordinary life – were compiled (RQ2). In contrast with Study 1 (section 4.2.7), which also discussed PX in the experimental session, Study 2 focused on experiences of videogame play in day-to-day life, and participants' relationship with videogames in general. These differences reflect the focus on SDT concepts when the

86 5 Study 2

initial list of interview questions was formulated (as noted in section 3.3.1). In particular, these interviews aimed to investigate past experiences of play as a means to develop more robust links between eudaimonia (section 2.2.1), self-endorsed behaviour, and player experience.

5.2.10 Autonomy Support in the Player-Game Relation This theme considers the ways that participants identify opportunities for autonomous behaviour in videogame play. Central to this theme is an understanding of videogames, alongside their human player(s), as active contributors in the production of play. Participant responses therefore reflect the facilitating and controlling potentials of the player-videogame dialogue.

Most frequently, participants (P2, P4, P9) discussed autonomy in the context of making choices with clear effects on the game state; decisions "that really affect the story later on.. like especially Mass Effect (BioWare, 2007) is a big one for that, like things you do in the first game will affect what happens in the third one and stuff" (P9). For these participants, choice is primarily a means to influence their own experience with a game (or series), with less consideration towards how their decisions affect the virtual world. Virtual worlds and characters are regarded as objects to be acted on: in the Mass Effect series, for example, the player is faced with recurring opportunities to commit genocide on alien and machinic species (or not). P2 similarly valued the presentation of life-or- death choices in Until Dawn (Supermassive Games, 2014) for their narrative influence:

P2: ..I like games like that [Until Dawn], where like, every choice – oh, obviously every choice you make like, kind of, changes like, the final ending of the game. So yeah. AT: Ok. can you give me like, any examples that stick out from that game? P2: [pause] There's – there's like – oh, like, even like, in like, the start, with the tutorial- ish bit, like with like, the two sisters, and like whether.. you like let go of your sister's hand and take like, the stranger's hand, or you just.. don't take the stranger's hand, and you just die with like your sister or whatever [laughs]. I haven't played it in a while. AT: Yeah, no, that's fine. But you like the fact that that choice exists, even though it just like– P2: –even though it– AT: –causes your immediate death, or?

5 Study 2 87

P2: Yeah? Yeah. I like – I like.. having choices, especially in like, really story-driven games. In Until Dawn, the outcomes of this decision are almost identical: by intention or chance, the player-character and her sister fall to their gruesome deaths, and the game continues thereafter. In this way, P2 also illustrates a second position common to participant responses (P2, P5, P10), whereby opportunities for choice are seemingly valued for their own sake, regardless of their in-game outcomes. For P5, this involved a rejection of worldly influence in favour of the (ostensibly) trivial:

P5: ..dialogue trees and stuff like that in games, that might change.. the – the like, story of the game and stuff like that, but for me, like the.. the small decisions are what makes it interesting. Like in Deus Ex (Ion Storm, 2000), you can choose [...] like a pistol or like a tranquiliser rifle or something like that, and it's just a small gameplay.. thing, effectively. Like an enemy that's dead or an enemy that's stunned, in the game.. is in the exact same state, they're like in a ragdoll kind of thing, and like there's some dialogue that reflects whether or not you went like a lethal or a non-lethal approach, but just having that option of having like a.. long-ranged non-lethal [weapon] or a short-ranged lethal [weapon] or whatever.. those small decisions. Although it is possible (albeit cynical) that these participants are merely repeating the language of triple-A videogame advertising, which privileges (among other things) freedom and choice (Oliva, Pérez-Latorre, & Besalú, 2018), it seems more likely that choice serves a functional purpose in player experience design. Indeed, the presence of choice in design may represent an attempt to convey the videogame's "willingness" to cooperate in the active reciprocity of play (Tyack & Wyeth, 2017). From this perspective, opportunities for choice support autonomy by reframing play as a negotiation with the videogame, rather than a product of computational control.

Finally, some participants (P1, P3) discussed play with an awareness of its conceptual and social intersections with material reality. For P1, this emerged through self-reflection:

P1: I find a lot of the time when it comes to choice in a linear game, it's more about the choice itself than.. the impact it's going to have [in the game] – if it's done well, that is, like if it's done well, the choice isn't like, you know, "what's the impact going to be down the line", it's like "what does this mean for me right now, why am I making that decision", all that kind of stuff.

88 5 Study 2

Here, P1 deliberately rejects in-game outcomes as inherently valuable, and instead emphasises the potential meaning to be derived from awareness and interrogation of the self in conversation with the (virtual) environment. Conversely, P1 considers choices least interesting "when they're.. extremely obvious, like 'are you going to support the freedom fighters or the nazis', I'm obviously not supporting the nazis". The videogame choice supports autonomy only to the extent that it (re)produces a scene of conflict in the self, and provides the means to resolve it. In contrast, P3 describes an approach to play that centres awareness of the social context:

AT: So when are the decisions you make in games most interesting? [pause] Or like what makes a decision in a game interesting? P3: [long pause] When it specifically affects other real players directly. [...] The decisions that you make there.. extend beyond the confines of the game world. [...] AT: [referring to a previous example from P3] When you play Monopoly (Magie & Darrow, 1935), are you playing with your friends, or are you playing with.. other people who.. are represented by things on this board. [pause] Like when the boot lands on Mayfair [and you have the opportunity to bankrupt them], is that your friend's boot, or is that another player's boot? P3: Again, I think that's circumstantial, that comes into.. the person that they are, what type of day they're having, the environment, like the human emotions, and the way they're specifically feeling, so. AT: Ok. I think, then, it's the former, right? Like you're always – it sounds like you're considering the person. P3: Absolutely. So with board games specifically, I think that's usually – like you have friends, or acquaintances that are close enough, like [they] aren't just complete strangers, like on digital games, so. Those interactions and consequences directly affect people that you know, and potentially care about, as opposed to.. online people which.. you have.. no ties to, no responsibilities for. P3 considers any game played with friends as primarily cooperative, using an awareness of the play situation to maximise mutual benefit rather than individual advantage. It is this calculus of care from which autonomy support may emerge.

5 Study 2 89

5.2.11 Threats to the Desired Play Situation Autonomy frustration occurs in response to internal pressures or external influences alike (section 2.2.3) – in this study, internal pressure appeared to reflect qualities of the player-game relation, whereas externally-pressuring events implicated social actors (and not, for example, cultural values). A number of participants (P1, P4, P6, P8, P11) described internally pressuring play experiences when confronted with risk in ambiguous circumstances. For example, P8 described feeling overcome by the sheer range of activities available in Stardew Valley, saying "it [Stardew Valley] doesn't give you enough, sort of, push towards certain directions, so you're kind of left standing there being like 'ah, I don't know if what I'm doing is correct'.. you're not doing the secret sequence". For P8, the personal stakes of playing "incorrectly" – or rather, being unable to see the correct approach to the game – produced feelings of concern: "Every day you waste is like.. going into like a death spiral within the game, right? [...] You tank your save, because you're just.. losing more days, and then you're wasting more money, and then you just can't stay alive" (P8). While Stardew Valley does not contain a fail state, the risk of poverty and death became salient for P8 in response to the game's ambiguous direction.

Moral dilemmas in which the stakes themselves are obfuscated can become overwhelming, as P1 observed:

AT: ..can you.. tell me about a time when you stopped playing a game because you didn't want to do what was required to progress? P1: I have stopped halfway through The Witcher 2[: Assassins of Kings] (CD Projekt Red, 2011), because I can't make the decision between the human guys, and the.. [...] the rebellious elves, and I think that was mostly because in the first game, I supported the rebellious elves, and in this game, they're kind of dicks [laughs]. [...] The complexity of the two factions in The Witcher 2, where there's like, extreme moral ambiguity, you don't really know entirely what's going on, because you haven't been given all the information up until that point, it's almost – it's too hard to make a decision.

The threat produced in this scenario is personal: for P1, who becomes "invested in the.. story, and characters, and wanting to see where it goes" (P1), taking the wrong side may represent a moral failure. Internal pressure emerges with self-doubt – and like P8, P1 would rather quit, without resolving the situation, than play "incorrectly".

Finally, participants described experiences of play characterised by a combination of autonomy and competence frustration (P1, P2, P4, P8, P11). These experiences often

90 5 Study 2

involved players' repeated failure to defeat a particular boss, which was seen to obstruct further play as desired. For P2, continued attempts were motivated by the pre-existing knowledge that "the start of Kingdom Hearts [II] (Square Enix Product Development Division 1, 2005) is really slow and boring, and it gets really good towards like, the, like, middle and the end of the game". Most participants, however, were more similar to P11, who described their approach to difficult boss fights as "quite stubborn":

P11: When it comes to a boss battle, like, it takes me.. a – usually a lot of tries to kind of defeat an enemy [...] and so.. you just – like I just keep.. forcing myself to do it until I get it, because I want to move on from that. [...] The other side of it [the boss fight], you know – like, I know.. will be beneficial, and it's like, rewarding to me, like, in – like, as a player to do that. Acute obstructions to autonomy may be acceptable if players can trust the game to reward them for their efforts (e.g., with further opportunities for enjoyment). The intensity and duration of internal pressure are therefore likely to vary as a function of perceived difficulty and interest in continued play. The resultant impasse may emerge in the context of conventional videogame challenges, or at the intersection of ambiguity and risk.

External pressures toward videogame play were also prevalent (P2, P3, P5, P6, P10), largely implicating friends (P3, P5, P6) or family members (P2, P6). Socially pressuring play experiences seemed relatively mild, characterised by feelings of boredom (P2, P5), annoyance (P2), and futility (P3, P10). Despite the unpleasant quality of these experiences, participants often continued to play, motivated by a sense of social obligation. For example, P5 recalled playing Delta Force: Land Warrior (NovaLogic, 2000) on a friend's suggestion, even though "..the game itself wasn't interesting. But.. a friend recommended it to me, and so I thought 'ok, I'll play it through to the end, just to see what it's about'. [...] I didn't particularly enjoy that experience [laughs]" (P5).

However, more substantial issues may arise when the controlling figure holds a position of authority. P6, for example, described how being prevented from playing videogames at home contributed to feelings of incompetence and ostracism among peers:

P6: So my eyes are pretty bad, and at the time [as a child], my mum was pretty strict on like – if I spent time on PC. [...] So before.. junior school, each week – sounds pretty bad, but it's true, like, I only had fifteen minutes on PC every week. [...] So the time I'm playing Dota [2] (Valve Corporation, 2013), everybody in my class can.. can, like, wreck me [laughs], I can't even stand a chance to win. So every player, like – they

5 Study 2 91

can't play solo [queue] with me, I always lose. But after that, yeah I say "it feels better when I don't keep losing", and after that because I come – I come overseas to Australia, and I like, spend more time on PC definitely, so.. I find that after like a year, I can beat most of my friends. Yeah cool [laughs]. Cool story. The extent to which these early experiences of social play influenced P6's later attempts to climb the Dota 2 ranked ladder remains unclear. Regardless, experiencing autonomy frustration at home coincided with frustration of all three needs in play – being rejected by friends (relatedness) due to lack of skill (competence) prevented P6 from the desired play experience (autonomy).

5.2.12 Wellbeing as Sufficient Functioning This theme highlights an understanding of wellbeing as it relates to experiences of videogame play and day-to-day life. For some participants (P1, P7, P11), wellbeing was strongly related to supporting balance in life:

AT: So wellbeing, or wellness, is becoming a more popular term in our culture, but I'm more interested in what wellbeing means to you specifically. P7: I think I tie wellbeing very.. strongly to balance in life, in that I think – for example, like, in this study, I do believe that videogames improves wellbeing, like, but to an extent, I think there.. becomes a point where it's like diminishing returns. You know, there's an unbalance with your life, if you're playing too much videogames? [pause] Yeah I'd say like wellbeing is.. I don't know, I kind of feel like saying "happiness" is just swapping one term for another. [pause] Like both mental and physical health, I suppose, it all ties into wellbeing.

The idea that wellbeing reflects physical and mental qualities also recurred (P1, P2, P5, P7, P11), with fewer participants identifying social elements (P2, P6, P10):

P11: Wellbeing to me is like, looking after your, like, physical and mental health, and – health – and having a balance of like.. good and bad things in your life, and not getting.. overwhelmed by any particular situation. So.. like, not – like – yeah. But mostly mental health, I feel like wellbeing is directed towards, and like, that kind of just.. base level of like, keeping yourself in a.. healthy state, like, exercising occasionally, and eating the right foods, and.. those kinds of.. things. [...] I think with like, work, or uni, or things like that, [not being overwhelmed is] like not getting.. totally.. like, enveloped in it, and like.. where you can't, like, think about anything else. Like I feel like, wellbeing these days is like, having that.. balance of like, being at work, and then being at home, and having your own life, and like, I.. kind of like,

92 5 Study 2

look to my mum as an example, she like, works a lot, but like, she'll like, take time out to like, go on hikes and things like that these days, and so, like.. that.. is like, something I see as looking after your wellbeing, if that makes sense.

Like P11 above, these participants describe wellbeing largely in terms of maintaining a basic level of functioning. Here, wellbeing is a largely personal responsibility; an internal practice to mitigate the effects of work (and other obligations) on organismic functioning. Engagement in leisure activities (e.g., hiking) is one way that maintenance is seen to occur. Indeed, participants (P1, P2, P5, P7, P8, P11) described using videogames in the same way:

AT: Ok. So how do you have positive experiences? P8: I think you can manufacture them, and.. I think – [laughs] yeah I know– AT: –please, tell me more!– P8: –I – I think it's great. Yeah, and I think that's games. I think games is.. so for example, today, I.. I woke up, I went to the dentist, and then, when I was in the dentist, I got so sick I had to go to the.. the bathroom, and it was just like, the most – it was the most awkward experience of my life. But then I'm here today, and even playing Spore, and even in a study for thirty minutes, it light– it makes me feel better, right? So it's manufactured good experiences. And so.. I think.. negative stuff, obviously, you're never going to try to get that, but positive stuff.. I guess, you can try and.. work your way towards, and if it happens on random that's even better, right? It feels even better.

Videogames are seen as utilities for "maintenance" (P1), and play is "a good.. way to take a break from everything else.. that's going on" (P2). For these participants, videogame play contributes to the perpetuation of life as it is, without necessarily supporting further improvement. Finally, some participants discussed past or present challenges around moderating the amount of time spent playing videogames (P1, P6, P8, P9). One participant (P7) distanced themselves from game overuse by posing it as a hypothetical situation:

P7: I feel like if you're sitting down and playing videogames, and like, you know you shouldn't be doing it, you should be doing study, I feel like that is.. I don't know, detrimental to your happiness and wellbeing, because.. I don't know, you can't really sit down and fully enjoy the videogame, because your mind's sort of – your mind's like, torn. For P1, past engagement in game overuse was contextualised as compensatory behaviour that created further obstacles in ordinary life:

5 Study 2 93

P1: ..you have to be a little bit careful if your wellbeing is quite low, because I remember, like – mostly – that's mostly from experience, because I mean, I had a rough.. period when I was a teenager, like a fairly rough period, and I would play videogames, instead of.. doing anything else. Like – and like, and they would be things like RPGs, that give you choices, and determination, and hero – like you know, like autonomy, and like, put you in the hero – like I played Dragon Age II (BioWare, 2011), notably, which is a game I – in hindsight, is a bad game, like it's not good. [...] Like I played Dragon Age II obsessively, like instead of doing like, schoolwork and stuff, and I failed classes, it was really fun. The relatively clear description provided by P1 above contrasts with P8's ambiguous account of online game play, which was (somewhat paradoxically) characterised by both sustained engagement and feeling distanced from game content:

P8: I don't like – and I played it for a long time, but I could never get into it, and it was a good experience that actually became a reflection of bad later on, it was playing World of Warcraft (Blizzard Entertainment, 2004). I – that gameplay is like licking an ashtray [...] especially when you play it, and then you think about it later on, and you're just like "I accomplished nothing", you know? Like.. AT: What is the difference there, like, between the accomplishments in [World of Warcraft].. P8: I think [pause] I think it's because the things that [World of Warcraft] was, sort of putting forward, I didn't value, but I obviously tried to make myself value them? Or I thought that I.. I could have valued them, for example.. story. [...] The way it [World of Warcraft] communicates its story is like.. [...] You stand there, and you've got a quest log, and then you're looking at that, and you're just like "I don't want to read that, that's just like a.. a paragraph". Each participant (P1, P7, P8) described these experiences as negative when seen through the lens of abstraction (P7) or hindsight (P1, P8). In contrast, P9 was more conflicted towards their recent play behaviour, seemingly unable to reconcile their preference for boundless game content (to relieve boredom) with post-play feelings of meaninglessness:

P9: If I'm like planning on playing a game for like two hours, and then it's like, end up playing.. end up doing like sidequests or little things the entire time, I feel like it was, like, a bit of a waste after that, like.. because that's what can happen in [The] Witcher [3: Wild Hunt] (CD Projekt Red, 2015) sometimes, where you just get stuck doing like, all the little, like, bounties and stuff, and.. [...] if I get off it, after that, I'll be like

94 5 Study 2

"wow, that was kind of just a waste of time". It's like I didn't really get much done in the game, but.. I still – yeah, I don't know, it's like, I still felt it was necessary though. AT: Is it because the bounties don't have like, story attached to them? P9: They'll have like a little bit, but like, I don't really know, sometimes that's – that makes it easier, because it makes it just quicker, you can quickly just go in and get it done. I don't know. Now that I say it, I'm really on like, just both sides, I'm realising now [laughs]. [...] Like.. like at the time, I feel like I'm really spending my time well, but then after, I'm like "I don't know, that was kind of a waste". AT: Ok. Do you think games need.. those.. like, The Witcher 3 bounties, or grinding, to exist. Would you enjoy games less without them. P9: I think it's a good idea to have them, because like, otherwise games finish too quickly, and then.. and they're not as replayable then, and you just keep doing the same mission – the same story over and over. [...] And [you want] something that you can just like, if you're really bored, you can just put a couple of hours into, and like.. yeah. P9's inner conflict towards their play behaviour provides the context for P7's abstracted definition of compartmentalisation ("you can't really sit down and fully enjoy the videogame, because your mind's like, torn"). When considering these inner conflicts, participants describe their play behaviours as "obsessive" (P1), "like licking an ashtray" (P8), or "necessary" (P9), suggesting a form of controlled motivation. Playing videogames for escapism appears to involve (and may indeed require) some degree of compartmentalisation. Selecting game content with poorly-conveyed (P8) or absent (P9) narrative qualities may prevent the intrusion of aspects of the self that would prompt conflict.

5.3 Discussion

5.3.1 H1 – Changes in Need Satisfaction and Frustration During Play Following Study 1 results (section 4.2.3), Study 2 assessed the restorative effects of videogame play following an experience of autonomy frustration. As hypothesised (H1), participants in the experimental condition experienced greater improvement in autonomy satisfaction and frustration, relative to control. More specifically: while participants in both conditions experienced an increase in autonomy satisfaction while playing Spore, this effect was larger for participants in the experimental condition. Moreover, a significant decline in autonomy frustration during play was observed only for participants in the experimental condition. However, an unintentional manipulation

5 Study 2 95

on competence frustration was also observed, likely due to the trivialised difficulty of the tangram game used in the experimental condition. Consequently, competence frustration declined during play for participants in the control condition, and increased during play for participants in the experimental condition.

An unintended competence manipulation – with greater competence frustration in the "control" condition – may seem to confound analysis of H2, and (accordingly) limit conclusions drawn from the study. Helpfully, this is not the case: while the tangram game was highly effective at manipulating autonomy frustration (ξ = 0.815, Table 14), T1 correlations between autonomy frustration and negative wellbeing outcomes in the experimental condition (Table 12) were either small (τANGER = 0.253, p = 0.006) or not significant (τDEPRESSION = 0.151, p = 0.109; τTENSION = 0.155, p = 0.086). Crucially, these negative outcome measures were always more strongly associated with competence frustration – for either condition, at T1 or T2. (Zero-order correlations at T2 in either condition may be found in Appendix C.) Rather than confounding analysis, these results facilitate meaningful comparison of the effects of competence and autonomy frustration on negative short-term wellbeing outcomes.

5.3.2 H2 – Changes in Positive Short-term Wellbeing and Intrinsic Motivation During Play Participants in the experimental condition, relative to control, were expected to experience greater improvement in positive short-term wellbeing outcomes and intrinsic motivation during play (H2). However, these effects only emerged for intrinsic motivation and vitality: participants in the experimental condition experienced greater intrinsic motivation and vitality during play, but no significant changes on these measures were observed for participants in the control condition. These results likely emerged as a function of autonomy satisfaction and frustration alike: while participants across conditions experienced some improvement in autonomy satisfaction during play (as noted in the previous section), a significant decrease in autonomy frustration was only observed for participants in the experimental condition. This explanation is consistent with SDT, which posits that need frustration obstructs wellbeing (as discussed in section 2.2.3).

Only a significant main effect of Time was observed for happiness and calmness, indicating that the extent to which these affects changed during play was indistinguishable between conditions. Exploratory follow-up tests on the entire sample indicated that participants experienced increased happiness and decreased calmness during play,

96 5 Study 2

although both effects were relatively small (ξMAX = 0.23). The condition-independent increase in happiness during play may reflect a weaker relationship with autonomy, or merely Type II error. It is even less clear, from a SDT perspective, why calmness declined during play, given its strong relationship with competence (section 5.2.5). When collapsed across conditions, competence frustration remained stable between measurement occasions (Table 15). However, although changes in competence satisfaction were not tested, descriptive statistics (Table 10) suggest it generally declined during play, although the difference between pre- (µ = 4.982) and post-play (µ = 4.811) means is small. While it is possible that a minor decrease in calmness during play was caused by an even smaller decline in competence satisfaction, firm conclusions cannot be drawn given the (lack of) evidence for such a position. It is perhaps more interesting that calmness was seemingly unaffected by changes in autonomy satisfaction and frustration during play (discussed further in section 5.3.4).

5.3.3 H2 – Changes in Negative Short-term Wellbeing During Play It was likewise expected that participants in the experimental condition, relative to control, would experience a greater decline in negative short-term wellbeing outcomes during play (H2). Contrary to this expectation, tension was observed to increase during play for participants in the experimental condition, and no significant changes in depressed affect or anger during play were observed in either condition. Given strong relations between competence frustration and negative short-term wellbeing measures, as noted in section 5.3.1, these results likely emerged due to the increase in competence frustration experienced by participants in the experimental condition.

In aggregate, H2 results suggest the tangram games were more effective at manipulating participants' affective arousal than valence. From this perspective, the tangram game's low difficulty in the experimental condition likely produced boredom, while participants in the control condition felt (comparatively) stimulated as a result of greater attentional requirements.

5.3.4 H3 – Need Satisfaction and Frustration as Predictors As predicted, experiences of need satisfaction and frustration during play explained relatively large proportions of variance in positive (H3a) and negative (H3b) outcomes. However, autonomy satisfaction failed to emerge as a relevant predictor of calmness, and autonomy frustration was not observed to predict any negative short-term wellbeing

5 Study 2 97

outcomes. (These issues are discussed alongside Study 2's qualitative findings in section 5.4.) Competence explained more unique variance (sr2) than autonomy in all modelled outcomes, positive or negative, with the exception of intrinsic motivation.

The minimal relation between autonomy and calmness in Study 2 may reflect broad variations in the ways that participants chose to play Spore, given the range of opportunities that emerge in the process of avatar customisation. For example, when playing Spore as a carnivore, further growth requires active conflict with other organisms (i.e., killing and eating their remains), which may not feel particularly relaxing. Conversely, herbivorous player-characters lacking defence mechanisms (e.g., speed) may feel less calm when being hunted themselves. While participants may have felt free to play Spore according to their preference, calmness appears to have emerged from feeling competent in survival – as predator or prey.

Whereas the pattern of results associated with H2 speaks to the difficulty of characterising affects as unequivocally positive or negative in the context of PX (Birk, Iacovides, et al., 2015; Mekler et al., 2016), this interpretation is complicated by their respective associations with need satisfaction and frustration (H3) – which are identified as unequivocally positive and negative in SDT (Ryan & Deci, 2017a). As noted in section 2.2.3, competence frustration in PX is characterised by feeling inadequate when faced with challenging obstacles, and autonomy frustration in PX likewise reflects feeling restricted by the game environment. Neither experience immediately appears to lend itself to producing beneficial outcomes; however, a potential resolution rests with the concept of emotional challenge – the struggle to integrate internal conflicts associated with an event, and regulate their coincident affects. Interpreting the results of a recent study (Bopp, Opwis, & Mekler, 2018) through SDT suggests that need frustration accompanies a range of emotionally challenging experiences. Autonomy frustration in particular may result from playing videogames that feature difficult or morally ambiguous choices (e.g., as popularised in The Walking Dead: Season One; Telltale Games, 2012), or unethical plot-critical tasks (e.g., torture in Grand Theft Auto V; Rockstar North, 2015). Emotionally challenging videogame experiences create opportunities for growth: players' successful resolution of inner conflict promotes movement toward more complex forms of self-organisation. Crucially, emotionally challenging events are associated with feelings of appreciation and meaning (Bopp et al., 2018), both of which are long-term wellbeing outcomes that derive from the integrative process (Ryan & Deci, 2017c, p.

98 5 Study 2

253). The possibility of meaning in play emerges from the dynamics of need frustration – real or threatened – in a social context of support.

5.3.5 H4 – Moderation by Immersion Study 2 also investigated the utility of immersion as a moderating factor. More specifically, immersion was expected to amplify the effect of need satisfaction on happiness, calmness, vitality, and intrinsic motivation (H4). Significant interactions between competence and immersion did not emerge; however, two significant interactions between autonomy and immersion were observed in models of happiness and intrinsic motivation. Contrary to expectations, however, immersion dampened the effect of autonomy satisfaction on both outcomes. Moreover, it is not clear that these effects represent moderation, as plausible curvilinear models were also observed. It is worth noting that the circumstances in which immersion should be reasonably expected to function as a moderator are not, to the author's knowledge, specified in SDT literature. Indeed, immersion's theoretical relevance in SDT is unclear (see section 2.2.2 for a review). Consequently, these results are difficult to contextualise with respect to SDT.

More practically, immersion made direct contributions to models of happiness, vitality, and intrinsic motivation, suggesting its alternative utility as a linear predictor of PX outcomes. However, this result may be inconsistent with SDT; in particular, the apparent incompatibility of immersion and mindful awareness (Brown & Cairns, 2004; Rigby et al., 2014; Sheldon, Prentice, & Halusic, 2015) provides the clearest indication that immersion may not support healthy organismic functioning. The continued presence of immersion research in SDT should, therefore, work to clarify the nature of its relations with the theory's core concepts, justify its value as an object of study, and – if identified as a positive outcome of play – clarify its ostensibly negative relations with mindful awareness.

5.3.6 RQ1 – Ongoing Effects of Need Frustration Study 2 investigated the extent to which need frustration could influence participants' intrinsic motivation for videogame play (RQ1), as reported elsewhere in the literature (Radel et al., 2014). The cited work found that experiencing autonomy frustration could limit intrinsic motivation for a second activity when the contextual elements signifying control remained salient: more specifically, when activities were introduced by the same researcher, and took place in the same room, participants who felt

5 Study 2 99

controlled in the first activity were less intrinsically motivated for the next task than participants in a neutral condition. Study 2 therefore attempted to find support for a similar pattern of results. However, no evidence in support of the target study's results was observed. Experiences of autonomy frustration in the tangram game did not significantly predict intrinsic motivation to play Spore (p = 0.118), and nor was the result practically significant (sr2 = 0.013). Moreover, no significant between-condition difference in intrinsic motivation for Spore was observed (p = 0.166, ξ = 0.168). Indeed, a significant result in this instance would have conflicted with findings from the original study. While these results are inconclusive, it nonetheless seems reasonable that restoration may occur "as long as the latter [activity] does not include subjective autonomy threats" (Radel et al., 2014). Indeed, Study 2 results – in particular, those associated with H1 and H2 – indicate that autonomy-thwarted participants identified opportunities to experience autonomy satisfaction and intrinsic motivation in Spore.

5.3.7 RQ2 – Autonomy, Wellbeing, and Motivation in PX Three distinct themes were assembled to establish a more developed position on autonomy and wellbeing with respect to the player experience (RQ2). The first relates to the ways that the player-game relation appears to influence autonomy support. The second theme focuses on obstructions to the ongoing situation of play, and experiences of pressure that may result. The third theme relates to perceptions of wellbeing as it relates to videogame play and day-to-day life, the potential for game overuse, and the motivational qualities associated with the decision to initiate play.

5.3.8 Autonomy Support in the Player-Game Relation SDT posits that videogames support autonomy by providing opportunities for self- expression, exploration, and meaningful decision-making (Rigby & Ryan, 2011b; Ryan & Deci, 2017b; Ryan et al., 2006). Autonomy in PX may also emerge from videogames' support for varied player strategies, and their use of informational (cf. controlling) feedback (sections 2.2.2 and 5.1.3). Study 2 results were broadly consistent with these existing theoretical applications; however, the ways that participants articulated experiences of autonomy during play extend its present application in PX.

Participants described three conceptualisations of games as conversational partners in play, which may vary in the extent to which they support autonomy. The first emerges from an approach to play as a means to influence the (virtual) world. From this

100 5 Study 2

perspective, the game primarily exists to satisfy the player, who should (therefore) have power over the game. This approach to play may, to some extent, support autonomy – however, because they aim merely to re-enact the controlling logics of their environment on (virtual) others, this mode of play ultimately validates and perpetuates the existence of controlling social contexts in material reality.

A second perspective was identified from participants who value choice for its own sake. This position appears to consider players and videogames as broadly equal partners in constructing play. The initiation of play is a tacit agreement by the player to accommodate the videogame; choice conveys the videogame's willingness to reciprocate. Finally, a third category was identified in those who experience play with an awareness of its broader social context, resituating the videogame as a mediation of material reality. From this perspective, the conversational quality of play is an engagement with the socially-constructed realities initialised in the videogame. The second and third positions seem likely to support autonomy to a similar degree, as both appear to value play as a practice of negotiation and mutual accommodation. However, more aware players may engage with eudaimonic motives, which are more likely to support experiences characterised by meaning and contemplation (Mekler & Hornbæk, 2016).

5.3.9 Threats to the Desired Play Situation Autonomy frustration occurs when the ongoing situation of play as desired is threatened – by an influential social actor, or the videogame itself. One such threat arises when ambiguous circumstances are perceived to have meaningful stakes. Videogames generally reward players for working to "optimise" their play behaviour within the rules and systems provided (Grimes & Feenberg, 2009), and players may feel pressured when optimal play behaviour is indeterminately specified in the game design, or eschewed altogether. Stardew Valley and The Witcher 2, then, may frustrate autonomy when players can neither read the optimal play – for efficient accumulation, or moral truth – nor consider, even momentarily, taking action with potentially sub-optimal consequences. Internal pressure emerges from this rigid conformity to self-imposed standards for successful, productive play, and the threat to contingent self-esteem that failure represents. Internal pressures resulting from more conventional videogame challenges (e.g., boss fights) are produced in the same way, although players' criteria for "success" clearly differ.

5 Study 2 101

Likewise, externally-controlling experiences occur when social actors are seen to impose on the desired play situation. Results suggest that the degree to which social actors frustrate autonomy primarily reflects their relative authority over the player. Moreover, the scope of encroachment on play behaviour was also observed as a relevant factor – while playing a disliked videogame at someone else's request may feel restricting, greater autonomy frustration occurs when in-game behaviours are directed or controlled by another (e.g., in competitive multiplayer modes), or when play itself is refused.

5.3.10 Wellbeing and Motivation in Videogame Play and Ordinary Life SDT posits that wellbeing emerges from "activity that develops and expresses one’s most reflectively valued and well integrated human potentialities" (Ryan et al., 2013). Fully functioning individuals experience an open awareness towards the self in relation to its social context; they enact the "good life" through the extension and practice of their potentialities (section 2.2.1). From this perspective, the results of Study 2 indicate that videogame play is, at best, a means to support players' continued functioning.

Participants described wellbeing as a practice of balancing time spent fulfilling commitments (e.g., in employment or formal education) with their separate interests. More generally, a balanced life was seen to reflect mental health, and (less often) physical health and social connectedness – outcomes of maintaining an equilibrium of time spent across a range of social contexts. Videogame play was seen both as a domain that could encroach on this balance, as well as one where positive experiences could be "manufactured", or wellbeing "maintained".

Considered with respect to Study 1 results (section 4.3.7), it seems likely that the experience of play as maintenance may reflect – to varying degrees – an awareness of (and engagement with) the self, or a means of distraction from ordinary life. To the extent to which either motive is salient, however, the approach described by these participants does not appear to represent a decision to play for its own sake. The initiation of play as maintenance may instead reflect well-integrated extrinsic motivation (section 2.2); a valued activity – in this context, for the purpose of restoration – that regardless may not be seen as interesting or enjoyable when the decision to play is made. The continuation of play, however, may become more intrinsically motivated as restoration becomes less pertinent.

102 5 Study 2

However, when the decision to initiate play is less well-integrated (as identified in Study 2 results), the ways that players understand the circumstances and experiences of videogame engagement may become misaligned or less coherent. To review: SDT's need density hypothesis asserts that game overuse is more likely to occur when need-satisfying qualities of ordinary life seem inadequate, relative to videogames designed to provide immediate, consistent, and dense experiences of need satisfaction (section 2.2.2). Consistent with the SDT literature on other maladaptive coping behaviours (e.g., Ryan et al., 2016), Study 2 results indicate that game overuse is compartmentalised in the self. In other words: the decision to initiate play corresponds with a more controlled form of extrinsic motivation, and although motivation to prolong the play situation is likely characterised by in-game experiences of interest and enjoyment, feelings of guilt at "wasted time" may emerge after play concludes. Although the self attempts to isolate (i.e., compartmentalise) these conflicting aspects of the self, they are brought together in awareness in the moments following play, in which the immediate memory of the play experience is made salient in the context of ordinary life.

5.4 Coda The autonomy frustration scale used in Study 2 (Appendix B) was constructed from items adapted from existing SDT measures (section 3.3.1), identified as a distinct factor at both measurement occasions (Appendix B), consistently demonstrated high internal reliability (ωMIN = 0.851; Table 10), and exhibited the largest between-condition difference of any need satisfaction or frustration measure following the experimental manipulation (ξ = 0.815; section 5.2.2). Regardless, as a predictor of post-play anger, depressed affect, and tension, autonomy frustration was neither statistically nor practically significant. This coda is a resolution – a final reading of these ostensibly inconsistent quantitative results, in conversation with qualitative study outcomes (section 5.2.10 and 5.3.8) and the theoretical literature (section 2.2.3).

To review, autonomy frustration (or heteronomy) occurs when "one's own behaviour is regulated or ruled by forces experienced as heteron or alien to one's sense of self" (Ryan et al., 2016, emphasis in original), which may arise from either "externally enforced or self-imposed pressures" (Chen et al., 2015). Indeed, both internal and external sources of pressure were observed in the interview transcripts (section 5.2.10), although a perceived threat to the desired play situation seemed essential to both (section 5.3.8).

5 Study 2 103

Notably, even when internal pressures emerged as a result of participants' rigid adherence to self-imposed standards for play, the videogame's lack of adequate direction (i.e., Stardew Valley) or supporting information (i.e., The Witcher 2) was implicated instead. Scale items broadly align with these experiences of autonomy frustration in play – consistent with SDT, autonomy frustration caused by internally-controlling processes is nonetheless experienced as being external (i.e., heteronomous) in origin.

Acute experiences of autonomy frustration were observed in interview responses when pressuring events (e.g., an ambiguous choice) obstructed the ongoing situation of play, which produced high-arousal affects (e.g., tension) and reactive behaviour (i.e., ending play) as a result. The tangram game featured no such obstruction to progress: instead, following the game's instructions likely elicited boredom – a low-arousal state – as an outcome of frustrated autonomy. The essential clarification of the qualitative analysis, then, is that the intensity of autonomy frustration modifies the negative affects that result.4 The modest correlations consequently observed between autonomy frustration and negative affects in Study 2 likely reflect their more arousing qualities.

4 Given their common status as basic needs, it follows that competence and relatedness frustration behave in the same way. What is generally known as "challenge" can also be described as a low-intensity experience of competence frustration; the more intense experience emerges from repeated failure. Likewise, a mild experience of relatedness frustration may emerge from feeling distant, whereas the acute form would result from an ostracising event.

104 5 Study 2

6 Discussion

Over two experimental mixed-methods studies, this thesis has investigated the restorative effects of videogame play following an experience of need frustration. More specifically, the present work addressed two research aims: first, it extended the application of self-determination theory (SDT) in player experience (PX) research by exploring the influence of need frustration on experiences of videogame play. Curiously, experiencing need frustration prior to playing Mark of the Ninja or Spore did not appear to influence post-play intrinsic motivation in either study (section 4.2.5 and 5.2.8), although further research is required to understand this result. Findings derived from Study 2's interview data indicate that internally- and externally-pressuring experiences alike are seen to originate from external sources in the play context (e.g., other people, game elements). Second, the present research has contributed to a more developed perspective towards the extent to which videogame play can improve short-term wellbeing, and the means by which it may occur. Both studies demonstrated that need- frustrated individuals can experience substantial improvement in short-term wellbeing outcomes over a 30-minute play session (H2). Wellbeing improvements are effectively and reliably predicted by experiences of need satisfaction and frustration during play (H3), and competence seems particularly effective in this regard.

The results of both studies demonstrated that videogame play can reverse the need- thwarting effects of a negative event (H1). As outlined in sections 4.2.3 and 5.2.2, participants who experienced competence (Study 1) or autonomy (Study 2) frustration prior to play, relative to those who experienced need satisfaction, experienced greater increases in satisfaction of the corresponding need during play. This result was extended in Study 2, which demonstrated that participants in the experimental condition, relative to control, experienced a greater decline in need frustration during play. A summary of results associated with H1 is shown in Table 31. These results are consistent with SDT's theoretical literature (Ryan & Deci, 2017a), and the conclusions of prior PX research (e.g., Reinecke et al., 2012; Rieger et al., 2015; Sheldon & Filak, 2008). However, the present research is (to the author's knowledge) the first to assess the restorative qualities of videogame play with respect to need frustration.

Although competence, autonomy, and relatedness generally co-occur (Ryan & Deci, 2017a), Study 2's experimental manipulation was instead shown to produce

6 Discussion 105

contrasting between-condition differences in competence and autonomy frustration. Specifically, participants in the experimental condition, relative to control, experienced greater autonomy frustration and (unexpectedly) lower competence frustration after playing the tangram game (T1). These results indicate that the interrelatedness of each basic need does not preclude their individual manipulation.

Across both studies, participants whose needs were frustrated prior to play, relative to those whose needs were satisfied, experienced greater improvement in positive short- term wellbeing and intrinsic motivation (H2; Table 31). These results are consistent with existing PX research (Fish et al., 2014; Hou et al., 2012; Russoniello et al., 2013; Russoniello et al., 2009), particularly studies conducted from a mood management theory (MMT) perspective (e.g., Koban et al., 2018; Rieger et al., 2015). It is worth noting that Study 2's assessment of three positive short-term wellbeing outcomes was perhaps most useful in demonstrating that videogame play does not influence "positive affect" equally across low (calmness), moderate (happiness), and high (vitality) levels of affective arousal. Even Spore, a relatively straightforward, low-stress game with few penalties for failure appears to induce some degree of excitement. The further use of specific affective dimensions in PX research would align with emerging literature (Birk, Iacovides, et al., 2015; Mekler et al., 2016) that aims to describe the qualities of PX beyond its categorisation as either "good" or "bad" (e.g., Markey, 2015).

Only in Study 1 did need-thwarted participants, relative to need-supported participants, experience greater improvement in negative short-term wellbeing during play (H2; Table 31). Indeed, Study 2 results were largely inconclusive, aside from the unexpected increase in tension during play for participants in the experimental condition. While these results are inconsistent with prior research (e.g., Rieger et al., 2015), only small correlations between autonomy frustration and negative wellbeing outcomes were observed in Study 2 (see section 5.3.1), indicating a low ceiling for any restorative effects that could emerge from play. In contrast, the negative performance feedback provided in Study 1's experimental condition was highly effective at increasing negative affect, facilitating its greater attenuation during play.

106 6 Discussion

H1 Study 1 Study 2 Condition δ p ξ p Control 0.052 0.766 0.75 < 0.001 Need Satisfaction Experimental 2.223 < 0.001 0.87 < 0.001

Control – – -0.21 0.060 Need Frustration Experimental – – -0.90 < 0.001

Competence Control – – -0.33 0.001 Frustration Experimental – – 0.45 < 0.001 (unintended) H2 Study 1 Study 2 (positive outcomes) Condition δ p ξ p Control 0.663 < 0.001 – – Positive Affect Experimental 2.059 < 0.001 – –

Control 0.206 0.246 0.05 0.542 Vitality Experimental 1.109 < 0.001 0.43 < 0.001

Happiness Full Sample – – 0.22 0.003

Calmness Full Sample – – -0.23 < 0.001

Control – – 0.13 0.216 Intrinsic Motivation Experimental – – 0.74 < 0.001

H2 Study 1 Study 2 (negative outcomes) Condition δ p ξ p Control 0.134 0.446 – – Negative Affect Experimental -0.945 < 0.001 – –

Control – – -0.21 0.036 Anger Experimental – – 0.14 0.157

Control – – -0.10 0.225 Tension Experimental – – 0.25 0.020

Table 31: Overview of results for H1 and H2 across Study 1 and 2. While ξ is bounded at zero, it has been assigned positive and negative values in this table for consistency with δ – in other words, to clarify whether experiential variables increased or decreased during play.

Need satisfaction and frustration were consistently shown to predict positive (H3a) and negative (H3b) PX outcomes, respectively (Table 32). In particular, need satisfaction was shown to account for substantial proportions of model variance in positive short-term wellbeing outcomes and intrinsic motivation. Need frustration explained less variance in negative short-term wellbeing outcomes, across Study 1 and 2, but these results are

6 Discussion 107

regardless stronger than expected, given those normally found in PX literature (e.g., Birk, Atkins, et al., 2016; Rieger et al., 2014; Ryan et al., 2006). Notably, competence demonstrated more consistency as a predictor of outcome variables than autonomy in both studies. Autonomy frustration was particularly ineffective in this regard, relative to competence frustration (sections 4.2.6 and 0). These results are inconsistent with research conducted in other contexts (Gunnell et al., 2013; Longo et al., 2016), although the tendency to conduct analyses on composite measures of need satisfaction and frustration (Bartholomew et al., 2018; Gillet, Forest, et al., 2015; Gillet, Fouquereau, et al., 2015; Jang et al., 2016) complicates further comparison with the literature. Autonomy frustration in PX may influence other negative short-term wellbeing outcomes more strongly (although the author is not aware of any such measures with prior use in experimental PX research), become salient only in particular modes of play (e.g., social play; Deterding, 2016; Vella et al., 2015), or lose meaning in the experimental context.

108 6 Discussion

H3a Study 1 Study 2 Measure sr2 p sr2 p Positive affect Competence 0.125 < 0.001 – – R2(adj) = 0.413; N/A Autonomy 0.037 0.049 – –

Vitality Competence 0.029 0.119 0.048 < 0.001 R2(adj) = 0.266; 0.429 Autonomy 0.011 0.329 0.032 0.005

Intrinsic motivation Competence 0.073 0.002 0.053 < 0.001 R2(adj) = 0.534; 0.390 Autonomy 0.143 < 0.001 0.096 < 0.001

Happiness Competence – – 0.085 < 0.001 R2(adj) = N/A; 0.390 Autonomy – – 0.030 0.008

Calmness Competence – – 0.050 0.002 R2(adj) = N/A; 0.271 Autonomy – – 0.003 0.455

H3b Study 1 Study 2 Measure sr2 p sr2 p Negative affect Competence 0.112 0.006 – – R2(adj) = 0.130; N/A Autonomy 0.022 0.209 – –

Anger Competence – – 0.105 < 0.001 R2(adj) = N/A; 0.245 Autonomy – – 0.009 0.191

Depressed affect Competence – – 0.141 < 0.001 R2(adj) = N/A; 0.223 Autonomy – – 0.014 0.106

Tension Competence – – 0.060 < 0.001 R2(adj) = N/A; 0.262 Autonomy – – < 0.001 0.970

Table 32: Overview of results for H3a and H3b across Study 1 and 2.

Study 2 provided no evidence of immersion's capacity to amplify (i.e., moderate) relations between PX measures (H4; section 5.2.7 and Appendix C), as is otherwise indicated in the SDT-PX literature (Przybylski, Rigby, & Ryan, 2010; Przybylski et al., 2012). Although moderation effects are typically small, and their detection may require even larger samples than the thesis timeframe would permit for experimental research (Darlington & Hayes, 2016, p. 429), there is little theoretical or empirical backing in SDT for the presence of such amplifying effects. Indeed, if immersion "magnifies the effects of experiences in virtual contexts" (Przybylski et al., 2012) on post-play outcomes, Study 2 results highlight the practical difficulties associated with further research on this property. The results of Study 2 alternatively suggest that immersion has greater utility as a linear predictor of experiential outcomes (section 5.3.5), as its effects on happiness, vitality, and intrinsic motivation were considerable – even after accounting for the influence of competence and autonomy satisfaction. Further applications of immersion in

6 Discussion 109

SDT-PX research should nonetheless be accompanied by elaboration of the concept in relation with the theory's central propositions, as discussed in section 5.3.5.

Neither study found support for claims that need frustration could influence intrinsic motivation for a successive activity (RQ1), as recent work suggests (Fang et al., 2017; Radel et al., 2014). Conclusions drawn from this prior research would indicate that need-thwarted participants in both studies should have experienced lower intrinsic motivation for videogame play, relative to need-satisfied participants, due to salient contextual cues retained from the manipulation task. In the cited studies, context was operationalised with respect to physical location, the presence of social actors (i.e., a researcher or educator), and task difficulty; the first two of these factors were held constant across activities in both Study 1 and 2. However, post-play differences in intrinsic motivation between conditions were not observed in either study's results (section 4.2.5 and 5.2.8); likewise, pre-play need frustration did not emerge as a statistically or practically significant predictor of post-play intrinsic motivation in Study 2 (section 5.2.8), as the cited literature would appear to suggest. Although the experimental manipulation used in Study 2 was modelled on procedures described by Radel et al. (2014), to the extent that resources allowed (see section 3.3), selecting Spore, a commercially-developed entertainment videogame, as the second task (cf. Sokoban, a block-pushing puzzle game) may have excised contextual cues signifying need frustration. Alternatively – given an additional twenty minutes in the Study 2 play session – residual need-thwarting effects of the experimental context may have merely dissipated over time.

Analysis of interview responses involving need satisfaction, need frustration, and wellbeing as they relate to videogame play (RQ2) informed and extended the quantitative results otherwise described. Study 1 results emphasised that the range of ways that players identify competence support in a game's design reflects their application of "embodied literacies" (Keogh, 2018); their naturalised understanding of videogame conventions and performances. Competence support is identified in game elements that facilitate engagement at the player's present literacy (section 4.3.6). Whereas less literate players attend to aspects of design that convey the fundamentals of engagement, players with more closely-aligned literacies identify competence support in game elements through which performances may be altered to reflect design intent. Experiences of competence satisfaction and frustration are situated in players' ongoing perception of success or failure

110 6 Discussion

at interpreting and enacting play as they understand it. These results elaborate the present characterisation of competence in SDT-PX literature (Ryan & Deci, 2017b), accounting for a greater proportion of the increasingly wide range of videogame players.

Analysis of Study 2 results produced a more developed position on the ways that autonomy support is mediated by the approach to play itself (section 5.3.8), complementing conceptual SDT-PX work (Ryan & Deci, 2017b; Ryan et al., 2006), and prior research on the social context of play (Deterding, 2016).

When in-game choices are primarily seen as opportunities to be influential; to decide what happens (and hence prevent others from deciding), play appears to reify the systems of control that produce autonomy frustration in day-to-day life. Autonomy support may be limited, or to some extent compromised, by this position of (perceived) authority over the videogame as a result. In contrast, autonomy support may be greater when play is produced in partnership with the videogame; when in-game choices are seen as opportunities to negotiate the ongoing situation of play. Finally, when play is initiated with eudaimonic motives, with the intent to engage with the versions of reality instantiated by the videogame – in short, when playing with awareness of the virtual, the real, and the self in relation – more reflective or meaningful experiences may also result. Study 2 results also identified a range of internal and external pressures that may emerge in play, or play-adjacent contexts. More specifically, players describe experiences of autonomy frustration when they perceive a threat to the desired play situation – when the game design or another person is seen to obstruct further interested engagement. Despite variations in participants' circumstances, however, the experiential qualities of autonomy frustration remained consistent (section 5.4).

Both Study 1 and 2 investigated ways that players approach videogames for wellbeing support, with varied results. Study 1 results identified two categories of play behaviour aimed at restoration. Videogames are used as a means of self-suspension: being in the virtual world facilitates disengagement from unpleasant experiences in ordinary life. Immersion in the videogame environment permits a temporary mindlessness that interrupts the persistent suffusion of work (or alternative structurally dominant influences) into locales of respite and leisure. These results align with a growing trend in SDT's theoretical (Rigby et al., 2014, p. 222) and empirical (Sheldon et al., 2015) literature that questions the value of immersion in PX. Conversely, the second approach to play identified in Study 1 was characterised by authentic self-expression: players

6 Discussion 111

wholeheartedly engage with the virtual environment as a means to more fully experience and develop the true self. These results extend prior SDT-PX research on ideal self- expression (Przybylski et al., 2012), indicating that players are also intrinsically motivated by opportunities to experience the actual self in play.

Study 2 results (section 5.3.10) identified a general perception that wellbeing reflects a balanced life, in which the expenditure of time and effort across domains is actively moderated. Videogame play may threaten this equilibrium or contribute to the maintenance of players' wellbeing. The decision to play videogames as a form of maintenance is extrinsically motivated (to varying degrees of autonomous regulation), whereby videogame play is pursued for an instrumental outcome – wellbeing maintenance – rather than for its inherently interesting and enjoyable qualities. Consistent with SDT (section 2.2), ongoing engagement in play is nonetheless sustained by experiences of interest and enjoyment towards the play situation, but more controlled regulations for initiating play produce internal conflicts that become salient when play concludes. This work extends the current SDT-PX perspective by recognising the importance of motivation in PX beyond its emergence in the immediate play situation (Nacke & Drachen, 2011).

6.1 Limitations / Future Research The present work has extended the application of SDT in PX research, exploring the ways that need frustration can influence experiences of videogame play, and contributing to knowledge about how and to what degree videogame play can improve short-term wellbeing. Nonetheless, the program of research described in this text remains incomplete in many ways.

The measurement and analysis of intrinsic motivation varied between studies. In Study 1, intrinsic motivation was only assessed in the post-play (T2) survey, which prevented its inclusion in H2, H3a, and limited analysis of RQ1. Similarly, having specified only improvement in need satisfaction in H1, changes in need frustration during play could not be reasonably assessed. These issues were avoidable, and weaken conclusions drawn from Study 1.

Both studies adapted experimental manipulations from existing work (Radel et al., 2014; Sheldon & Filak, 2008), with the aim of introducing between-condition differences in competence (Study 1) or autonomy (Study 2). Unintended manipulations were

112 6 Discussion

nonetheless observed – more specifically, autonomy satisfaction and frustration (Study 1), and competence frustration (Study 2). These effects were largely unpredictable, given that neither cited study had assessed need frustration. However, without further details regarding the original study's word-finding task (Sheldon & Filak, 2008), it is unclear why between-condition differences in autonomy satisfaction were not observed in the original study. Separately, while the word-finding task was highly successful at manipulating competence, a considerable number of participants (n=7) were excluded from analysis as a result of its use.

Although the author of the present work focused on selecting appropriate videogame stimuli in both studies, future efforts should further attend to the development and standardisation of experimental manipulations for need frustration. To some extent, issues with both manipulations reflect the author's assumptions about both participant samples (i.e., their familiarity with word-finding tasks and tangrams) that were somewhat optimistic. Despite the researcher's attempts to conduct rigorous and generally intelligible videogame selection procedures, Study 2 processes largely failed in this regard (although Spore itself generally produced the desired results). Robust methods for selecting appropriate experimental stimuli remains an ongoing challenge in PX research, and further work in this domain would benefit the experimental literature.

Issues with survey measures were observed in both studies. In Study 1, low reliability scores were observed for autonomy satisfaction, autonomy frustration, and negative affect (as noted in section 4.4), which led to their replacement in Study 2. Issues with measures of autonomy satisfaction and frustration in Study 1 speak to the difficulty of applying general scales to a particular context. Moreover, some autonomy-related items in the Balanced Measure of Psychological Needs (BMPN; Sheldon & Hilpert, 2012) refer to behaviours and events (e.g., "I was really doing what interests me") that are, perhaps, less generalisable than experiences associated with the construct (e.g., volition; section 2.2.2). The poor reliability of negative affect (Diener & Emmons, 1984) in Study 1's post-play assessment is odd, given that generalised positive and negative affect are often measured in PX research – albeit with the Positive and Negative Affect Schedule (PANAS; Watson et al., 1988). The bias towards high-arousal affects in the PANAS (Diener et al., 2010) – which is not shared by the scale used in Study 1 – may contribute to its comparatively reliable use in PX research.

6 Discussion 113

While issues with scale reliability did not recur in Study 2 (section 5.1.2), the approach to scale development was largely ad-hoc. The resultant competence and autonomy frustration subscales are face valid, and they were distinguished from competence and autonomy satisfaction in exploratory factor analyses (Appendix B); however, the nature of their construction was less rigorous than more recently developed measures in PX research (Brühlmann, Vollenwyder, Opwis, & Mekler, 2018; Power, Cairns, Denisova, Papaioannou, & Gultrom, 2018; Vanden Abeele et al., 2016; Vanden Abeele et al., under review). Further use of these measures is ill-advised, in light of well- documented issues caused by the continued use of unvalidated and ad-hoc measures in PX research (Law et al., 2018; Mekler et al., 2014). Further PX research on need frustration would benefit from the development of validated domain-specific instruments to assess need frustration; for example, through independent validation of the measures developed for Study 2.

Finally, it is worth considering that the present work did not study relatedness in any capacity. Relatedness is less frequently studied in PX research, relative to competence or autonomy (Tyack & Wyeth, 2017), despite its influence on organismic functioning and wellbeing (section 2.2.2). Relatedness measures in present use may only apply to social videogame play (Denisova, Nordin, & Cairns, 2016; Tamborini et al., 2010; Vella et al., 2015), complicating measurement and further research. It seems likely that further conceptual work and scale development are necessary for research on relatedness satisfaction and frustration in solitary and social play to proceed. Similarly, recent conceptual work (Mekler & Hornbæk, 2019; Mekler, Iacovides, & Bopp, 2018) and scale development (Vanden Abeele et al., 2016; Vanden Abeele et al., under review) on meaningful experiences has produced opportunities for experimental research on the ways that need frustration may influence meaning in PX. While other aspects of wellbeing are similarly understudied in PX research (e.g., awareness), their investigation in the context of play may require further preliminary work (e.g., scale development).

6.2 Contributions The present work has elaborated SDT research in the context of videogame play, investigating the restorative effects of videogame play after episodes of competence and autonomy frustration, advancing knowledge regarding the extent to which play can improve short-term wellbeing, and the means by which it may occur. In doing so, this

114 6 Discussion

thesis has created a more developed understanding of "the nature and extent of the wellbeing benefits of videogames... knowledge of when, how, and for whom there is a positive impact of videogame play" (D. Johnson, Jones, et al., 2013).

This research extends present applications of SDT in PX research, investigating the influence of an episode of need frustration on successive experiences of videogame play. In doing so, the present work has contributed to a call within SDT for further research on need satisfaction and frustration in parallel (Van den Broeck et al., 2016). Studies of this kind are thriving in research on physical activity (Bartholomew et al., 2011; Gunnell et al., 2013), education (Bartholomew et al., 2018; Cordeiro et al., 2016), the workplace (Bartholomew et al., 2014; Unanue et al., 2017), and day-to-day life (Chen et al., 2015); however, fewer studies have investigated need satisfaction and frustration in the context of videogame play (section 2.2.3). Indeed, this thesis represents the first experimental work on need satisfaction and frustration in the context of videogame play, and as such contributes to the broader SDT literature. In constructing interview questions around PX in the experimental play session (Study 1; section 4.3.6), the present research has produced more localised knowledge of player perceptions and interpretations of in-game experiences. Likewise, constructing interview topics on particular experiences and behaviours associated with play – autonomy satisfaction and frustration (section 5.3.8) and wellbeing (sections 4.3.7 and 5.3.10) – have produced more specific insights on the intersection of SDT and PX in day-to-day life.

Separately, this thesis contributes to the present understanding of autonomy frustration in empirical research. Specifically, it has become clear that the intensity of autonomy frustration shapes the negative affects that result (section 5.4), whereby intensity forms a positive relationship with aroused affect. Studies of low-intensity autonomy frustration would therefore benefit from assessing negative affects characterised by low arousal (e.g., fatigue, boredom).

In contrast with the increasing popularity of analytic- or metric-driven design in game development practice (Drachen, Canossa, & Yannakakis, 2009; Giovannetti, 2019; Whitson, 2013), this thesis emphasises the value of player subjectivity as a source of knowledge about player experience. This thesis contributes to game design practice as a reminder that players themselves often clearly understand, and can readily articulate, the ways that specific in-game experiences (i.e., competence, autonomy, wellbeing, and motivation) emerge in and through play. Notably, Study 1 results showed that players

6 Discussion 115

identify competence support in particular game elements as a function of their (self- described) skill. However, literacy neither guarantees competence satisfaction, nor guards against competence frustration – rather, literacies produce new fields in which competence satisfaction or frustration may occur. These findings inform the conceptualisation of competence in interactive system design. An interactive system's usability and potential expressiveness varies as a function of user literacy: design elements that support expressiveness for more literate users may go unnoticed by novices, or obstruct use. Study 2 instead demonstrated that participants can often recall past experiences to a degree that reveals fundamental qualities of their approach towards individual videogames and play in general. Across both studies, participants described ways they intentionally use videogames – or particular game segments – as a means to elicit restorative play experiences.

Finally, the design and application of original experimental stimuli – the tangram games created for Study 2 – represent a methodological contribution of the present work. In Study 2, the tangram games were observed to successfully induce between-condition differences in autonomy satisfaction and frustration, as intended (section 5.2.2). Their additional manipulation of competence frustration is undesirable, however, and its prevention requires further design iteration. In particular, it is likely that learning the rules and controls in the tutorial was, for many players, insufficient preparation for the regular- size tangrams that followed (section 3.3). An extended tutorial that guides players through the tactics of play more gradually (e.g., with a steadily larger tangram introduced at each step) may limit between-condition differences in competence frustration in future work. This aspect of the game is under development. As portable software with low system requirements and a 10-minute runtime, future versions of the tangram game stimuli could be readily applied across a variety of study designs, which would likely benefit the field of experimental autonomy research.

6.3 Final Comments This thesis has examined qualities of restorative play from the perspective of SDT, a broadly applicable theory of human motivation, development, and wellbeing. Across two studies, the negative effects of a need frustrating episode were neutralised during a subsequent experience of videogame play. Post-experiment interviews produced more developed perspectives on the ways that competence, autonomy, motivation, and

116 6 Discussion

wellbeing emerge in and through the player experience. In aggregate, the present work represents the first experimental research to simultaneously examine the effects of need satisfaction and frustration in the context of videogame play, extending present knowledge of the extent to which videogames improve player wellbeing, and the means by which these benefits occur. As the influence of videogame production and play continues to grow across the globe (Abraham & Jayemanne, 2017; Dyer-Witheford & de Peuter, 2009; Huntemann & Aslinger, 2013), it is essential to continue investigating "how videogames in all their forms can enable a flourishing society" (D. Johnson, Jones, et al., 2013). The research contained in this thesis represents a fraction of what remains to be done.

6 Discussion 117

References

Abraham, B., & Jayemanne, D. (2017). Where are all the climate change games? Locating digital games’ response to climate change. Transformations, 30, 74-94. Allen, J. J., & Anderson, C. A. (2018). Satisfaction and Frustration of Basic Psychological Needs in the Real World and in Video Games Predict Internet Gaming Disorder Scores and Well-Being. Computers in Human Behavior, 84, 220-229. Barone, E. (2016). Stardew Valley [Videogame]. London, England: Chucklefish. Bartholomew, K. J., Ntoumanis, N., Cuevas, R., & Lonsdale, C. (2014). Job Pressure and Ill-Health in Physical Education Teachers: The Mediating Role of Psychological Need Thwarting. Teaching and Teacher Education, 37, 101-107. Bartholomew, K. J., Ntoumanis, N., Mouratidis, A., Katartzi, E., Thøgersen-Ntoumani, C., & Vlachopoulos, S. (2018). Beware of Your Teaching Style: A School-Year Long Investigation of Controlling Teaching and Student Motivational Experiences. Learning and Instruction, 53, 50-63. Bartholomew, K. J., Ntoumanis, N., Ryan, R. M., & Thøgersen-Ntoumani, C. (2011). Psychological Need Thwarting in the Sport Context: Assessing the Darker Side of Athletic Experience. Journal of Sport and Exercise Psychology, 33(1), 75-102. Bartholomew, K. J., Ntoumanis, N., & Thøgersen-Ntoumani, C. (2010). The controlling interpersonal style in a coaching context: Development and initial validation of a psychometric scale. Journal of Sport and Exercise Psychology, 32(2), 193-216. Berlant, L. (2011). Slow Death (Obesity, Sovereignty, Lateral Agency). In Cruel Optimism (pp. 95-120). Durham, USA: Duke University Press. BioWare. (2007). Mass Effect [Videogame]. Washington, USA: Microsoft Game Studios. BioWare. (2011). Dragon Age II [Videogame]. Redwood City, USA: Electronic Arts. Birk, M. V., Atkins, C., Bowey, J. T., & Mandryk, R. L. (2016). Fostering Intrinsic Motivation through Avatar Identification in Digital Games. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems (pp. 2982-2995). ACM. Birk, M. V., Buttlar, B., Bowey, J. T., Poeller, S., Thomson, S. C., Baumann, N., & Mandryk, R. L. (2016). The Effects of Social Exclusion on Play Experience and Hostile Cognitions in Digital Games. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems (pp. 3007-3019). ACM. Birk, M. V., Iacovides, I., Johnson, D., & Mandryk, R. L. (2015). The False Dichotomy Between Positive and Negative Affect in Game Play. In Proceedings of the 2015 Annual Symposium on Computer-Human Interaction in Play (pp. 799-804). ACM. Birk, M. V., & Mandryk, R. L. (2013). Control Your Game-Self: Effects of Controller Type on Enjoyment, Motivation, and Personality in Game. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (pp. 685-694). ACM. Birk, M. V., & Mandryk, R. L. (2018). Combating Attrition in Digital Self-Improvement Programs using Avatar Customization. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems. ACM. doi:10.1145/3173574.3174234

References 119

Birk, M. V., Mandryk, R. L., Miller, M. K., & Gerling, K. M. (2015). How Self-Esteem Shapes our Interactions with Play Technologies. In Proceedings of the 2015 Annual Symposium on Computer-Human Interaction in Play (pp. 35-45). Blizzard Entertainment. (2004). World of Warcraft [Videogame]. California, USA: Blizzard Entertainment. Blount, K., Powell, L., Dolapchiev, T., Petra'ka, I., Baumgartner, J., Vignoni, D., & Yoshida, K. (2016). Scrivener for Windows and Linux (Version 1.9.7): Literature and Latte. Bødker, S. (2006). When Second Wave HCI meets Third Wave Challenges. In Proceedings of NordiCHI. Bopp, J. A., Mekler, E. D., & Opwis, K. (2016). Negative Emotion, Positive Experience?: Emotionally Moving Moments in Digital Games. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems (pp. 2996-3006). ACM. doi:10.1145/2858036.2858227 Bopp, J. A., Opwis, K., & Mekler, E. D. (2018). “An Odd Kind of Pleasure”: Differentiating Emotional Challenge in Digital Games. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems. ACM. doi:10.1145/3173574.3173615 Bostic, T. J., Rubio, D. M., & Hood, M. (2000). A Validation of the Subjective Vitality Scale Using Structural Equation Modeling. Social Indicators Research, 52(3), 313-324. doi:10.1023/A:1007136110218 Bourgonjon, J., Vandermeersche, G., De Wever, B., Soetaert, R., & Valcke, M. (2016). Players' Perspectives on the Positive Impact of Video Games: A Qualitative Content Analysis of Online Forum Discussions. New Media & Society, 18(8), 1732-1749. Bowman, N. D., & Tamborini, R. (2012). Task Demand and Mood Repair: The Intervention Potential of Computer Games. New Media & Society, 14(8), 1339- 1357. Bowman, N. D., & Tamborini, R. (2015). “In the Mood to Game”: Selective exposure and mood management processes in computer game play. New Media & Society, 17(3), 375-393. Box, G. E. P., & Cox, D. R. (1964). An Analysis of Transformations. Journal of the Royal Statistical Society, 26(2), 211-252. Braun, V., & Clarke, V. (2006). Using Thematic Analysis in Psychology. Qualitative Research in Psychology, 3, 77-101. Brown, E., & Cairns, P. (2004). A Grounded Investigation of Game Immersion. In CHI '04 Extended Abstracts on Human Factors in Computing Systems (pp. 1297- 1300). ACM. Brühlmann, F., & Schmid, G.-M. (2015). How to Measure the Game Experience? Analysis of the Factor Structure of Two Questionnaires. In Proceedings of the 33rd Annual ACM Conference Extended Abstracts on Human Factors in Computing Systems (pp. 1181-1186). ACM. doi:10.1145/2702613.2732831 Brühlmann, F., Vollenwyder, B., Opwis, K., & Mekler, E. D. (2018). Measuring the “Why” of Interaction: Development and Validation of the User Motivation Inventory (UMI). In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems. ACM. Bruxner, G. (2018). The Haunted Island, a Frog Detective Game [Videogame]. Bryant, J., & Davies, J. (2006). Selective Exposure to Video Games. In P. Vorderer & J. Bryant (Eds.), Playing Video Games: Motives, Responses, and Consequences (pp. 210-226). New York, USA: Routledge.

120 References

Burgers, C., Eden, A., van Engelenburg, M. D., & Buningh, S. (2015). How feedback boosts motivation and play in a brain-training game. Computers in Human Behavior, 48, 94-103. Burns, M. S. (2018). Make Number Bigger. Retrieved from https://www.magicalwasteland.com/notes/2018/1/9/make-number-bigger CD Projekt Red. (2011). The Witcher 2: Assassins of Kings [Videogame]. Warsaw, Poland: CD Projekt. CD Projekt Red. (2015). The Witcher 3: Wild Hunt [Videogame]. Warsaw, Poland: CD Projekt. Chen, B., Vansteenkiste, M., Beyers, W., Boone, L., Deci, E. L., Van der Kaap-Deeder, J., . . . Verstuyf, J. (2015). Basic Psychological Need Satisfaction, Need Frustration, and Need Strength Across Four Cultures. Motivation and Emotion, 39(2), 216-236. Chunsoft. (2008). Pokémon Mystery Dungeon: Explorers of [Time/Darkness] [Videogame]. Tokyo, Japan: The Pokémon Company. Clark, L. A., & Watson, D. (1995). Constructing Validity: Basic Issues in Objective Scale Development. Psychological Assessment, 7(3), 309-319. Cohen, J. (1992). A Power Primer. Psychological Bulletin, 112(1), 155-159. Cole, T., Cairns, P., & Gillies, M. (2015). Emotional and Functional Challenge in Core and Avant-garde Games. In Proceedings of the 2015 Annual Symposium on Computer-Human Interaction in Play (pp. 121-126). ACM. doi:10.1145/2793107.2793147 Collins, E., & Cox, A. L. (2014). Switch On to Games: Can Digital Games Aid Post- Work Recovery? International Journal of Human-Computer Studies, 72, 654- 662. doi:10.1016/j.ijhcs.2013.12.006 Cordeiro, P., Paixão, P., Lens, W., Lacante, M., & Sheldon, K. M. (2016). Factor Structure and Dimensionality of the Balanced Measure of Psychological Needs Among Portuguese High School Students. Relations to Well-Being and Ill-Being. Learning and Individual Differences, 47, 51-60. doi:10.1016/j.lindif.2015.12.010 Core Design. (1997). Tomb Raider II [Videogame]. London, England: Eidos Interactive. Csikszentmihalyi, M. (1975). Beyond Boredom and Anxiety: The Experience of Play in Work and Games. California, USA: Jossey-Bass. Darlington, R. B., & Hayes, A. F. (2016). Regression Analysis and Linear Models: Concepts, Applications, and Implementation. New York, USA: Guilford Press. Deci, E. L., & Ryan, R. M. (1985). Intrinsic Motivation and Self-Determination in Human Behavior. New York, USA: Plenum Press. Deci, E. L., & Ryan, R. M. (1987). The Support of Autonomy and the Control of Behavior. Journal of Personality and Social Psychology, 53(6), 1024-1037. Deci, E. L., & Ryan, R. M. (2000). The "What" and "Why" of Goal Pursuits: Human Needs and Self-Determination of Behavior. Psychological Inquiry, 11(4), 227- 268. Dena, C. (2017). Finding a Way: Techniques to Avoid Schema Tension in Narrative Design. Transactions of the Digital Games Research Association, 3(1), 27-61. Denisova, A., Nordin, A. I., & Cairns, P. (2016). The Convergence of Player Experience Questionnaires. In Proceedings of the 2016 Annual Symposium on Computer- Human Interaction in Play. ACM. Deterding, S. (2016). Contextual Autonomy Support in Video Game Play: A Grounded Theory. In Proceedings of the 34th Annual ACM Conference on Human Factors in Computing Systems (pp. 3931-3943). ACM.

References 121

Diener, E., & Emmons, R. A. (1984). The Independence of Positive and Negative Affect. Journal of Personality and Social Psychology, 47(5), 1105-1117. doi:10.1037/0022-3514.47.5.1105 Diener, E., Wirtz, D., Tov, W., Kim-Prieto, C., Choi, D.-w., Oishi, S., & Biswas-Diener, R. (2010). New Well-being Measures: Short Scales to Assess Flourishing and Positive and Negative Feelings. Social Indicators Research, 97(2), 143-156. Drachen, A., Canossa, A., & Yannakakis, G. N. (2009). Player modeling using self- organization in Tomb Raider: Underworld. In IEEE Symposium on Computational Intelligence and Games. IEEE. Dyer-Witheford, N., & de Peuter, G. (2009). Games of Empire: Global Capitalism and Video Games. Minneapolis, USA: University of Minnesota Press. Eagle Dynamics. (2003). Lock On: Modern Air Combat [Videogame]. Montreuil, France: Ubisoft. Earl, S. R., Taylor, I. M., Meijen, C., & Passfield, L. (2017). Autonomy and Competence Frustration in Young Adolescent Classrooms: Different Associations with Active and Passive Disengagement. Learning and Instruction, 49, 32-40. doi:10.1016/j.learninstruc.2016.12.001 Elson, M., & Quandt, T. (2016). Digital Games in Laboratory Experiments: Controlling a Complex Stimulus Through Modding. Psychology of Popular Media Culture, 5(1), 52-65. Fang, H., He, B., Fu, H., & Meng, L. (2017). Being Eager to Prove Oneself: U-Shaped Relationship between Competence Frustration and Intrinsic Motivation in Another Activity. Frontiers in Psychology, 8. doi:10.3389/fpsyg.2017.02123 Field, A., Miles, J., & Field, Z. (2012). Exploring assumptions. In Discovering Statistics Using R (pp. 166-204). London, England: Sage. Fish, M., Russoniello, C. V., & O’Brien, K. (2014). The Efficacy of Prescribed Casual Videogame Play in Reducing Symptoms of Anxiety: A Randomized Controlled Study. Games for Health Journal, 3(5), 291-295. Folkman, S., & Moskowitz, J. T. (2004). Coping: Pitfalls and Promise. Annual Review of Psychology, 55, 745-774. doi:10.1146/annurev.psych.55.090902.141456 Gillet, N., Forest, J., Benabou, C., & Bentein, K. (2015). The Effects of Organizational Factors, Psychological Need Satisfaction and Thwarting, and Affective Commitment on Workers' Well-Being and Turnover Intentions. Le Travail Humain, 78(2), 119-140. doi:10.3917/th.782.0119 Gillet, N., Fouquereau, E., Huyghebaert, T., & Colombat, P. (2015). The Effects of Job Demands and Organizational Resources through Psychological Need Satisfaction and Thwarting. The Spanish Journal of Psychology, 18. doi:10.1017/sjp.2015.30 Giovannetti, A. (2019). 'Slay the Spire': Metrics Driven Design and Balance. Retrieved from https://www.gdcvault.com/play/1025731/-Slay-the-Spire-Metrics Grimes, S. M., & Feenberg, A. (2009). Rationalizing Play: A Critical Theory of Digital Gaming. The Information Society, 25, 105-118. doi:10.1080/01972240802701643 Grolnick, W. S., Deci, E. L., & Ryan, R. M. (1997). Internalisation Within the Family: The Self-Determination Theory Perspective. In J. E. Grusec & L. Kuczynski (Eds.), Parenting and Children's Internalisation of Values: A Handbook of Contemporary Theory (pp. 135-161). New Jersey, USA: John Wiley. Gunnell, K. E., Crocker, P. R. E., Wilson, P. M., Mack, D. E., & Zumbo, B. D. (2013). Psychological Need Satisfaction and Thwarting: A Test of Basic Psychological Needs Theory in Physical Activity Contexts. Psychology of Sport and Exercise, 14(5), 599-607. doi:10.1016/j.psychsport.2013.03.007

122 References

Haerens, L., Aelterman, N., Vansteenkiste, M., Soenens, B., & Van Petegem, S. (2015). Do Perceived Autonomy-Supportive and Controlling Teaching Relate to Physical Education Students' Motivational Experiences Through Unique Pathways? Distinguishing Between the Bright and Dark Side of Motivation. Psychology of Sport and Exercise, 16(3), 26-36. doi:10.1016/j.psychsport.2014.08.013 Harrison, S., Tatar, D., & Sengers, P. (2007). The Three Paradigms of HCI. In Alt. Chi. Session at the SIGCHI Conference on Human Factors in Computing Systems. ACM. Hartanto, A., Toh, W. X., & Yang, H. (2016). Age matters: The effect of onset age of video game play on task-switching abilities. Attention, Perception, & Psychophysics, 78(4), 1125-1136. Hassenzahl, M., & Tractinsky, N. (2006). User Experience – A Research Agenda. Behaviour & Information Technology, 25(2), 91-97. Hidden Path Entertainment, & Valve Corporation. (2012). Counter-Strike: Global Offensive [Videogame]. Bellevue, USA: Valve Corporation. Hou, J., Nam, Y., Peng, W., & Lee, K. M. (2012). Effects of screen size, viewing angle, and players’ immersion tendencies on game experience. Computers in Human Behavior, 28, 617-623. Hunicke, R., LeBlanc, M., & Zubek, R. (2004). MDA: A Formal Approach to Game Design and Game Research. In Proceedings of the AAAI Workshop on Challenges in Game AI. Huntemann, N. B., & Aslinger, B. (Eds.). (2013). Gaming Globally: Production, Play, and Place. New York, USA: Palgrave Macmillan. Ion Storm. (2000). Deus Ex [Videogame]: Eidos Interactive. Izard, C. E. (1977). Human Emotions. New York, USA: Springer. Jang, H., Kim, E. J., & Reeve, J. (2016). Why Students Become More Engaged or More Disengaged During the Semester: A Self-Determination Theory Dual-Process Model. Learning and Instruction, 43, 27-38. doi:10.1016/j.learninstruc.2016.01.002 Jang, H., Reeve, J., Ryan, R. M., & Kim, A. (2009). Can Self-Determination Theory Explain What Underlies the Productive, Satisfying Learning Experiences of Collectivistically Oriented Korean Students? Journal of Educational Psychology, 101(3), 644-661. JASP Team. (2017). JASP (Version 0.8.5.1). Netherlands: University of Amsterdam. Retrieved from https://jasp-stats.org/ JASP Team. (2019). JASP (Version 0.9). Netherlands: University of Amsterdam. Retrieved from https://jasp-stats.org/ Jayemanne, D. (2017). Performativity in Art, Literature, and Videogames. Cham, Switzerland: Springer International. Jennett, C. (2010). Is Game Immersion Just Another Form of Selective Attention? An Empirical Investigation of Real World Dissociation in Computer Game Immersion (PhD Thesis) Johnson, D., & Gardner, J. (2010). Personality, Motivation and Video Games. In Proceedings of the 22nd Conference of the Computer-Human Interaction Special Interest Group of Australia on Computer-Human Interaction (pp. 276-279). ACM. doi:10.1145/1952222.1952281 Johnson, D., Gardner, J., & Sweetser, P. (2016). Motivations for videogame play: Predictors of time spent playing. Computers in Human Behavior, 63, 805-812. Johnson, D., Gardner, M. J., & Perry, R. (2018). Validation of two game experience scales: The Player Experience of Need Satisfaction (PENS) and Game Experience

References 123

Questionnaire (GEQ). International Journal of Human-Computer Studies, 118, 38-46. doi:10.1016/j.ijhcs.2018.05.003 Johnson, D., Jones, C., Scholes, L., & Carras, M. C. (2013). Videogames and wellbeing: A comprehensive review. Melbourne, Australia: Young and Well Cooperative Reseach Centre. Retrieved from https://eprints.qut.edu.au/105915/1/2013%20CRC%20Report%20Videogames_a nd_Wellbeing.pdf Johnson, D., Klarkowski, M., Vella, K., Phillips, C., McEwan, M., & Watling, C. N. (2018). Greater rewards in videogames lead to more presence, enjoyment and effort. Computers in Human Behavior, 87, 66-74. doi:10.1016/j.chb.2018.05.025 Johnson, D., Wyeth, P., & Sweetser, P. (2013). The People-Game-Play Model for Understanding Videogames’ Impact on Wellbeing. In Proceedings of the IEEE International Games Innovation Conference (pp. 85-88). IEEE. Johnson, P. E. (2018). rockchalk: Regression Estimation and Presentation (Version 1.8.117). Kanaga, D. (2017). Oikospiel Book I [Videogame]. Kelley, K. (2018). The MBESS R Package (Version 4.4.3). Kelley, K., & Pornprasertmanit, S. (2016). Confidence Intervals for Population Reliability Coefficients: Evaluation of Methods, Recommendations, and Software for Composite Measures. Psychological Methods, 21(1), 69-92. doi:10.1037/a0040086 Keogh, B. (2018). With Thumbs in Mind. In A Play of Bodies: How We Perceive Videogames (pp. 75-108). Cambridge, USA: MIT. Kimball, S., Mattis, P., & GIMP Development Team. (2018). GNU Image Manipulation Program (Version 2.10.6). Klarkowski, M., Johnson, D., Wyeth, P., McEwan, M., Phillips, C., & Smith, S. (2016). Operationalising and Evaluating Sub-Optimal and Optimal Play Experiences through Challenge-Skill Manipulation. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems (pp. 5583-5594). ACM. Klarkowski, M., Johnson, D., Wyeth, P., Smith, S., & Phillips, C. (2015). Operationalising and Measuring Flow in Video Games. In Proceedings of the Annual Meeting of the Australian Special Interest Group for Computer Human Interaction (pp. 114-118). ACM. Klei Entertainment. (2012). Mark of the Ninja [Videogame]. Washington, USA: Microsoft Studios. Knobloch-Westerwick, S. (2006). Mood Management: Theory, Evidence, and Advancements. In J. Bryant & P. Vorderer (Eds.), Psychology of Entertainment (pp. 257-272). New York, USA: Routledge. Koban, K., Breuer, J., Rieger, D., Mohseni, M. R., Noack, S., Bente, G., & Ohler, P. (2018). Playing for the thrill and skill. Quiz games as means for mood and competence repair. Media Psychology. Kuo, A., Lutz, R. J., & Hiler, J. L. (2016). Brave New World of Warcraft: A Conceptual Framework for Active Escapism. Journal of Consumer Marketing, 33(7), 498- 506. La Guardia, J. G., Ryan, R. M., Couchman, C. E., & Deci, E. L. (2000). Within-Person Variation in Security of Attachment: A Self-Determination Theory Perspective on Attachment, Need Fulfillment, and Well-Being. Journal of Personality and Social Psychology, 79(3), 367-384. doi:10.1037/0022-3514.79.3.367

124 References

Lafrenière, M.-A. K., Verner-Filion, J., & Vallerand, R. J. (2012). Development and validation of the Gaming Motivation Scale (GAMS). Personality and Individual Differences, 53, 827-831. Lane, A. M., Soos, I., Leibinger, E., Karsai, I., & Hamar, P. (2007). Validity of the Brunel Mood Scale for use with UK, Italian and Hungarian athletes. In A. M. Lane (Ed.), Mood and Human Performance: Conceptual, Measurement, and Applied Issues (pp. 119-130). New York, USA: Nova. Law, E. L.-C., Brühlmann, F., & Mekler, E. D. (2018). Systematic Review and Validation of the Game Experience Questionnaire (GEQ) – Implications for Citation and Reporting Practice. In Proceedings of the 2018 Annual Symposium on Computer- Human Interaction in Play (pp. 257-270). ACM. doi:10.1145/3242671.3242683 Lombard, M., & Ditton, T. (1997). At the Heart of It All: The Concept of Presence. Journal of Computer-Mediated Communication, 3(2). Longo, Y., Gunz, A., Curtis, G. J., & Farsides, T. (2016). Measuring Need Satisfaction and Frustration in Educational and Work Contexts: The Need Satisfaction and Frustration Scale (NSFS). Journal of Happiness Studies, 17(1), 295-317. Magie, L., & Darrow, C. (1935). Monopoly [Board Game]: Hasbro. Mair, P., & Wilcox, R. (2016). Robust Statistical Methods in R Using the WRS2 Package. Report. Harvard University. Mair, P., & Wilcox, R. (2018). WRS2: A Collection of Robust Statistical Methods (Version 0.10-0). Mandryk, R. L., Inkpen, K. M., & Calvert, T. W. (2006). Using psychophysiological techniques to measure user experience with entertainment technologies. Behaviour & Information Technology, 25(2), 141-158. doi:http://dx.doi.org/10.1080/01449290500331156 Markey, P. M. (2015). Finding the Middle Ground in Violent Video Game Research: Lessons From Ferguson (2015). Perspectives on Psychological Science, 10(5), 667-670. Maxis. (2008). Spore [Videogame]. Redwood City, USA: Electronic Arts. McAuley, E., Duncan, T., & Tammen, V. V. (1989). Psychometric Properties of the Intrinsic Motivation Inventory in a Competitive Sport Setting: A Confirmatory Factor Analysis. Research Quarterly for Exercise and Sport, 60(1), 48-58. doi:10.1080/02701367.1989.10607413 Mekler, E. D., Bopp, J. A., Tuch, A. N., & Opwis, K. (2014). A Systematic Review of Quantitative Studies on the Enjoyment of Digital Entertainment Games. In Proceedings of the 32nd Annual ACM Conference on Human Factors in Computing Systems (pp. 927-936). ACM. Mekler, E. D., & Hornbæk, K. (2016). Momentary Pleasure or Lasting Meaning? Distinguishing Eudaimonic and Hedonic User Experiences. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems (pp. 4509-4520). ACM. doi:10.1145/2858036.2858225 Mekler, E. D., & Hornbæk, K. (2019). A Framework for the Experience of Meaning in Human-Computer Interaction. In Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems. ACM. Mekler, E. D., Iacovides, I., & Bopp, J. A. (2018). "A Game that Makes You Question...": Exploring the Role of Reflection for the Player Experience. In Proceedings of the 2018 Annual Symposium on Computer-Human Interaction in Play (pp. 315-327). ACM. doi:10.1145/3242671.3242691 Mekler, E. D., Rank, S., Steinemann, S. T., Birk, M. V., & Iacovides, I. (2016). Designing for Emotional Complexity in Games: The Interplay of Positive and Negative

References 125

Affect. In Proceedings of the 2016 Annual Symposium on Computer-Human Interaction in Play Companion Extended Abstracts (pp. 367-371). ACM. Metacritic. (2013). Mark of the Ninja Critic Reviews for PC. Retrieved from https://www.metacritic.com/game/pc/mark-of-the-ninja/critic-reviews Milyavskaya, M., Philippe, F. L., & Koestner, R. (2013). Psychological Need Satisfaction Across Levels of Experience: Their Organization and Contribution to General Well-being. Journal of Research in Personality, 47(1), 41-51. Nacke, L., & Drachen, A. (2011). Towards a Framework of Player Experience Research. In Proceedings of the Second International Workshop on Evaluating Player Experience in Games at FDG. ACM. Namco. (1980). Pac-Man [Videogame]: Namco. Nintendo EAD. (2005). Animal Crossing: Wild World [Videogame]. Kyoto, Japan: Nintendo. Nintendo EAD. (2008). Mario Kart Wii [Videogame]. Kyoto, Japan: Nintendo. Nintendo EPD. (2017). The Legend of Zelda: Breath of the Wild [Videogame]. Kyoto, Japan: Nintendo. NovaLogic. (2000). Delta Force: Land Warrior [Videogame]. California, USA: NovaLogic. Okada, K., & Hoshino, T. (2017). Researchers’ choice of the number and range of levels in experiments affects the resultant variance-accounted-for effect size. Psychonomic Bulletin & Review, 24(2), 607-616. Oliva, M., Pérez-Latorre, Ó., & Besalú, R. (2018). ‘Choose, collect, manage, win!’: Neoliberalism, enterprising culture and risk society in video game covers. Convergence, 24(6), 607-622. doi:10.1177/1354856516680324 Oswald, C. A., Prorock, C., & Murphy, S. M. (2014). The Perceived Meaning of the Video Game Experience: An Exploratory Study. Psychology of Popular Media Culture, 3(2), 110-126. doi:10.1037/a0033828 PopCap Games. (2004). Bejeweled 2 [Videogame]. Seattle, USA: PopCap Games. PopCap Games. (2006). Bookworm Adventures [Videogame]. Seattle, USA: PopCap Games. PopCap Games. (2007). Peggle [Videogame]. Seattle, USA: PopCap Games. Power, C., Cairns, P., Denisova, A., Papaioannou, T., & Gultrom, R. (2018). Lost at the Edge of Uncertainty: Measuring Player Uncertainty in Digital Games. International Journal of Human-Computer Interaction. Przybylski, A. K., Deci, E. L., Rigby, C. S., & Ryan, R. M. (2014). Competence-Impeding Electronic Games and Players' Aggressive Feelings, Thoughts, and Behaviors. Journal of Personality and Social Psychology, 106(3), 441-457. Przybylski, A. K., Rigby, C. S., & Ryan, R. M. (2010). A Motivational Model of Video Game Engagement. Review of General Psychology, 14(2), 154-166. Przybylski, A. K., Ryan, R. M., & Rigby, C. S. (2009). The Motivating Role of Violence in Video Games. Personality and Social Psychology Bulletin, 35(2), 243-259. Przybylski, A. K., Weinstein, N., Murayama, K., Lynch, M. F., & Ryan, R. M. (2012). The Ideal Self at Play: The Appeal of Video Games That Let You Be All You Can Be. Psychological Science, 23(1), 69-76. Przybylski, A. K., Weinstein, N., Ryan, R. M., & Rigby, C. S. (2009). Having to Versus Wanting to Play: Background and Consequences of Harmonious Versus Obsessive Engagement in Video Games. CyberPsychology & Behavior, 12(5), 485-492.

126 References

R Core Team. (2018). R: A Language and Environment for Statistical Computing (Version 3.5.1). Vienna, Austria: R Foundation for Statistical Computing. Retrieved from https://www.R-project.org Radel, R., Pelletier, L., Baxter, D., Fournier, M., & Sarrazin, P. (2014). The paradoxical effect of controlling context on intrinsic motivation in another activity. Learning and Instruction, 29, 95-102. doi:10.1016/j.learninstruc.2013.09.004 Raven Software. (2016). Call of Duty: Modern Warfare Remastered [Videogame]. Santa Monica, USA: Activision. Reeve, J., & Halusic, M. (2009). How K-12 teachers can put self-determination theory principles into practice. Theory and Research in Education, 7(2), 145-154. Reinecke, L. (2009). Games and Recovery: The Use of Video and Computer Games to Recuperate from Stress and Strain. Journal of Media Psychology, 21(3), 126-142. Reinecke, L., Klatt, J., & Krämer, N. C. (2011). Entertaining Media Use and the Satisfaction of Recovery Needs: Recovery Outcomes Associated with the Use of Interactive and Noninteractive Entertaining Media. Media Psychology, 14(2), 192-215. Reinecke, L., Tamborini, R., Grizzard, M., Lewis, R., Eden, A., & Bowman, N. D. (2012). Characterizing Mood Management as Need Satisfaction: The Effects of Intrinsic Needs on Selective Exposure and Mood Repair. Journal of Communication, 62(3), 437-453. Reis, H. T., Sheldon, K. M., Gable, S. L., Roscoe, J., & Ryan, R. M. (2000). Daily Well- Being: The Role of Autonomy, Competence, and Relatedness. Personality and Social Psychology Bulletin, 26(4), 419-435. Rieger, D., Frischlich, L., Wulf, T., Bente, G., & Kneer, J. (2015). Eating Ghosts: The Underlying Mechanisms of Mood Repair Via Interactive and Noninteractive Media. Psychology of Popular Media Culture, 4(2), 138-154. Rieger, D., Wulf, T., Kneer, J., Frischlich, L., & Bente, G. (2014). The Winner Takes it All: The Effect of In-Game Success and Need Satisfaction on Mood Repair and Enjoyment. Computers in Human Behavior, 39, 281-286. Rigby, C. S., & Ryan, R. M. (2007). The Player Experience of Need Satisfaction (PENS): An Applied Model and Methodology for Understanding Key Components of the Player Experience. Immersyve, Inc. Retrieved from http://immersyve.com/download/pens-white-paper/?wpdmdl=8283 Rigby, C. S., & Ryan, R. M. (2011a). Dangerous Waters: The Addictive Undertow of Games. In Glued to Games: How Video Games Draw Us In and Hold Us Spellbound (pp. 97-118). Santa Barbara, USA: Praeger. Rigby, C. S., & Ryan, R. M. (2011b). Games and the Need for Autonomy. In Glued to Games: How Video Games Draw Us In and Hold Us Spellbound (pp. 39-64). Santa Barbara, USA: Praeger. Rigby, C. S., & Ryan, R. M. (2011c). Games and the Need for Relatedness. In Glued to Games: How Video Games Draw Us In and Hold Us Spellbound (pp. 65-80). Santa Barbara, USA: Praeger. Rigby, C. S., Schultz, P. P., & Ryan, R. M. (2014). Mindfulness, Interest-Taking, and Self-Regulation: A Self-Determination Theory Perspective on the Role of Awareness in Optimal Functioning. In A. Ie, C. T. Ngnoumen, & E. J. Langer (Eds.), The Wiley Blackwell Handbook of Mindfulness: Volume I (pp. 216-235). West Sussex, England: Wiley. Rocchi, M., Pelletier, L., Cheung, S., Baxter, D., & Beaudry, S. (2017). Assessing need- supportive and need-thwarting interpersonal behaviours: The Interpersonal

References 127

Behaviours Questionnaire (IBQ). Personality and Individual Differences, 104, 423-433. Rockstar North. (2015). Grand Theft Auto V [Videogame]. New York City, USA: Rockstar Games. Russoniello, C. V., Fish, M., & O’Brien, K. (2013). The Efficacy of Casual Videogame Play in Reducing Clinical Depression: A Randomized Controlled Study. Games for Health Journal, 2(6), 341-346. Russoniello, C. V., O’Brien, K., & Parks, J. M. (2009). The Effectiveness of Casual Video Games in Improving Mood and Decreasing Stress. Journal of CyberTherapy and Rehabilitation, 2(1), 53-66. Ryan, R. M., Curren, R. R., & Deci, E. L. (2013). What Humans Need: Flourishing in Aristotelian Philosophy and Self-Determination Theory. In A. S. Waterman (Ed.), The Best Within Us: Positive Psychology Perspectives on Eudaimonia (pp. 57- 75). Washington, DC: American Psychological Association. doi:10.1037/14092- 004 Ryan, R. M., & Deci, E. L. (2000). The Darker and Brighter Sides of Human Existence: Basic Psychological Needs as a Unifying Concept. Psychological Enquiry, 11(4), 319-338. Ryan, R. M., & Deci, E. L. (2008). From Ego Depletion to Vitality: Theory and Findings Concerning the Facilitation of Energy Available to the Self. Social and Personality Psychology Compass, 2(2), 702-717. Ryan, R. M., & Deci, E. L. (2017a). Basic Psychological Needs Theory: Satisfaction and Frustration of Autonomy, Competence, and Relatedness in Relation to Psychological Wellness and Full Functioning. In Self-Determination Theory: Basic Psychological Needs in Motivation, Development, and Wellness (pp. 239- 271). New York, USA: Guilford. Ryan, R. M., & Deci, E. L. (2017b). Motivation and Need Satisfaction in Video Games and Virtual Environments. In Self-Determination Theory: Basic Psychological Needs in Motivation, Development, and Wellness (pp. 508-531). New York, NY: Guilford. Ryan, R. M., & Deci, E. L. (2017c). Self-Determination Theory: Basic Psychological Needs in Motivation, Development, and Wellness. New York, USA: Guilford Press. Ryan, R. M., Deci, E. L., & Vansteenkiste, M. (2016). Autonomy and Autonomy Disturbances in Self-Development and Psychopathology: Research on Motivation, Attachment, and Clinical Process. In D. Cicchetti (Ed.), Developmental Psychology: Theory and Method (pp. 385-438). Hoboken, USA: Wiley. Ryan, R. M., & Frederick, C. (1997). On Energy, Personality, and Health: Subjective Vitality as a Dynamic Reflection of Well-Being. Journal of Personality, 65(3), 529-565. Ryan, R. M., Mims, V., & Koestner, R. (1983). Relation of Reward Contingency and Interpersonal Context to Intrinsic Motivation: A Review and Test using Cognitive Evaluation Theory. Journal of Personality and Social Psychology, 45(4), 736- 750. doi:10.1037/0022-3514.45.4.736 Ryan, R. M., Rigby, C. S., & Przybylski, A. (2006). The Motivational Pull of Video Games: A Self-Determination Theory Approach. Motivation and Emotion, 30(4), 344-360. Schmierbach, M., Limperos, A. M., & Woolley, J. K. (2012). Feeling the Need for (Personalized) Speed: How Natural Controls and Customization Contribute to

128 References

Enjoyment of a Racing Game Through Enhanced Immersion. Cyberpsychology, Behavior, and Social Networking, 15(7), 364-369. Schmierbach, M., Xu, Q., Oeldorf-Hirsch, & Dardis, F. E. (2012). Electronic Friend or Virtual Foe: Exploring the Role of Competitive and Cooperative Multiplayer Video Game Modes in Fostering Enjoyment. Media Psychology, 15(3), 356-371. Sheldon, K. M. (2018). Understanding The Good Life: Eudaimonic Living Involves Well-Doing, Not Well-Being. In J. P. Forgas & R. F. Baumeister (Eds.), The Social Psychology of Living Well (pp. 116-136). New York, USA: Routledge. Sheldon, K. M., & Filak, V. (2008). Manipulating Autonomy, Competence, and Relatedness Support in a Game-Learning Context: New Evidence that All Three Needs Matter. British Journal of Social Psychology, 47(2), 267-283. Sheldon, K. M., & Hilpert, J. C. (2012). The Balanced Measure of Psychological Needs (BMPN) Scale: An Alternative Domain General Measure of Need Satisfaction. Motivation and Emotion, 36(4), 439-451. Sheldon, K. M., Prentice, M., & Halusic, M. (2015). The Experiential Incompatibility of Mindfulness and Flow Absorption. Social Psychological and Personality Science, 6(3), 276-283. Square Enix Product Development Division 1. (2005). Kingdom Hearts II [Videogame]. Tokyo, Japan: Square Enix. Supermassive Games. (2014). Until Dawn [Videogame]. California, USA: Sony Computer Entertainment. Sweetser, P., & Wyeth, P. (2005). GameFlow: A Model for Evaluating Player Enjoyment in Games. Computers in Entertainment, 3(3). Tamborini, R., Bowman, N. D., Eden, A., Grizzard, M., & Organ, A. (2010). Defining Media Enjoyment as the Satisfaction of Intrinsic Needs. Journal of Communication, 60(4), 758-777. Tamborini, R., Grizzard, M., Bowman, N. D., Reinecke, L., Lewis, R. J., & Eden, A. (2011). Media Enjoyment as Need Satisfaction: The Contribution of Hedonic and Nonhedonic Needs. Journal of Communication, 61, 1025-1042. Telltale Games. (2012). The Walking Dead: Season One [Videogame]. San Rafael, USA: Telltale Games. Tyack, A., & Wyeth, P. (2017). Exploring Relatedness in Single-Player Video Game Play. In Proceedings of the 29th Australian Conference on Computer-Human Interaction (pp. 422-427). ACM. Tyack, A., Wyeth, P., & Klarkowski, M. (2018). Video Game Selection Procedures For Experimental Research. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems. ACM. doi:10.1145/3173574.3173760 Ubisoft Montpellier. (2014). Rayman Legends [Videogame]. Montreuil, France: Ubisoft. Unanue, W., Rempel, K., Gómez, M. E., & Van den Broeck, A. (2017). When and Why Does Materialism Relate to Employees’ Attitudes and Well-being: The Mediational Role of Need Satisfaction and Need Frustration. Frontiers in Psychology, 8. doi:10.3389/fpsyg.2017.01755 Unity Technologies ApS. (2017). Unity (Version 2017.3) [Game Engine]. California, USA: Unity Technologies ApS. Retrieved from https://unity3d.com/get- unity/download/archive Valve Corporation. (2013). Dota 2 [Videogame]. Bellevue, USA: Valve Corporation. Valve Corporation. (2018). Mark of the Ninja on Steam. Retrieved from https://store.steampowered.com/app/214560/Mark_of_the_Ninja/

References 129

Van den Broeck, A., Ferris, D. L., Chang, C.-H., & Rosen, C. C. (2016). A Review of Self-Determination Theory's Basic Psychological Needs at Work. Journal of Management, 42(5), 1195-1229. Vanden Abeele, V., Nacke, L. E., Mekler, E. D., & Johnson, D. (2016). Design and Preliminary Validation of The Player Experience Inventory. In Proceedings of the 2016 Annual Symposium on Computer-Human Interaction in Play Companion Extended Abstracts (pp. 335-341). ACM. Vanden Abeele, V., Spiel, K., Nacke, L., Johnson, D., & Gerling, K. (under review). Development and Validation of the Player Experience Inventory: A Scale to Measure Player Experiences at the Level of Functional and Psychosocial Consequences. Vanhee, G., Lemmens, G. M. D., Stas, L., Loeys, T., & Verhofstadt, L. L. (2018). Why are couples fighting? A need frustration perspective on relationship conflict and dissatisfaction. Journal of Family Therapy, 40, S4-S23. Vella, K., Johnson, D., & Hides, L. (2015). Playing Alone, Playing With Others: Differences in Player Experience and Indicators of Wellbeing. In Proceedings of the 2015 Annual Symposium on Computer-Human Interaction in Play (pp. 3-12). doi:http://dx.doi.org/10.1145/2793107.2793118 Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and Validation of Brief Measures of Positive and Negative Affect: The PANAS Scales. Journal of Personality and Social Psychology, 54(6), 1063-1070. Whitson, J. R. (2013). Gaming the Quantified Self. Surveillance & Society, 11(1/2), 163- 176. Wilcox, R. (2012). Introduction to Robust Estimation and Hypothesis Testing (3rd ed.). Amsterdam, Netherlands: Elsevier. Williams, G. C., & Deci, E. L. (1996). Internalization of biopsychosocial values by medical students: a test of self-determination theory. Journal of Personality and Social Psychology, 70, 767-779. Williams, G. C., & Deci, E. L. (2001). Activating patients for smoking cessation through physician autonomy support. Medical Care, 39(8), 813-823. Wright, W. (2005). Will Wright and Spore [Video]: GDC. Zillmann, D. (1988). Mood Management Through Communication Choices. The American Behavioral Scientist, 31(3), 327-340.

130 References

Appendices Appendix A Study 1 Material Descriptive Statistics Age Experience 0 1 0 1 Valid 33 32 33 32 Missing 0 0 0 0 Mean 23.97 23.09 2.000 1.406 Std. Deviation 5.531 6.382 1.677 1.365 Minimum 18.00 18.00 1.000 1.000 Maximum 46.00 46.00 7.000 7.000 0 = control condition, 1 = experimental condition Gender Control condition Experimental condition Gender Count % Gender Count % F 7 21.2 F 7 21.9 M 25 75.8 M 24 75.0 Non-binary 1 3.0 Non-binary 1 3.1 Total 33 100.0 Total 32 100.0

Videogame Selection Criteria Confirmation notes o “selection criteria of previous studies... represent an informal consensus [on relevant game elements]” o MDA adaptation o genre archetypicality o accessibility (e.g., limited possibility space/control options) o reasonable experience of the game within the allotted session o current; up-to-date (recent release) o strong competence support, “medium to low” support for other needs MDA o Mechanics: Avoid dialogue options / skill trees Limit travel options (i.e., not open world) Provide clear feedback (e.g., score) Clearly indicate player progress Avoid multiplayer games o Dynamics Avoid provision of choice Make linear elements seem (relatively) desirable Challenge/skill balance Game elements resist player progress Avoid player-NPC communication o Aesthetics Limit discovery, expression, narrative, fellowship

Appendices 131

Support challenge, competition The extent to which Mark of the Ninja meets these criteria: • high metacritic rating (91%) and positive user reviews (97% positive on steam on 14 june, 2018) • accessible controls • limited mechanics (but not overly simple for more experienced players) o 2D platformer (following justification by Ryan2006, even if it doesn't apply there) • structured tutorial • can be played in 30 minutes to a reasonable degree • relatively recent (in terms of genre/mechanic standards) MDA: limited choices (pathing only), dialogue (1-way), clear progress (moving left to right, UI elements update), numeric score (used frequently), feedback (noise circles, sight cones, score, text (e.g., stealth kill), etc.)

Shortlist super mario 3d world (wii u) +well-known game/series +relatively accessible (tutorial prompts etc.) +generally pretty easy (at least for the first half-hour) +no-one can argue (rightfully or not) that it’s too old/unrepresentative of the industry +almost zero story setup /i’d probably need to fix character choice (i.e., by telling participants to play as mario) /peak-end effects?? -less linear than i’d like

yoshi’s woolly world (wii u) +well-known character +very accessible (tutorial prompts etc.) +very easy +linear +very little story setup /peak-end effects?? /no idea if people will want to play it -i can imagine the “it’s a kids’ game” reviewer criticism already -i’d probably want to get the good wii u controller

the floor is jelly (pc; “partial” controller support) +avoids pre-existing knowledge effects +intuitive (you can only jump, by the looks of things) +minimal setup time +linear +probably less room for autonomy support than other games listed +no issues with peak-end effects; there are no ‘bosses’ etc. /somewhat unusual by normal standards (reviewers might hate it) /i’d need to play it to tell how hard it is (note: I did play it; it was hard) -you can go backwards -possible performance issues

mark of the ninja (pc; controller support) +may avoid pre-existing knowledge effects

132 Appendices

+relatively accessible (on-screen button prompts, tutorial) +well-received (probably safe from reviewer criticism) /stealth mechanics create slightly different px /not entirely linear; may support autonomy pretty well /more violence than any other game /peak-end effects?? -more story content than other games listed

H3 Multiple Regression Analysis (Non-significant Need Satisfaction Predictors) Negative Affect Model Summary Model R R² Adjusted R² RMSE R² Change F Change df1 df2 p 0 0.359 0.129 0.086 0.598 0.129 3.012 3 61 0.037 1 0.468 0.219 0.153 0.576 0.090 3.405 2 59 0.040 Note. Null model includes NA1, PENSC, PENSA

ANOVA Model Sum of Squares df Mean Square F p 0 Regression 3.231 3 1.077 3.012 0.037 Residual 21.811 61 0.358 Total 25.042 64 1 Regression 5.488 5 1.098 3.312 0.011 Residual 19.554 59 0.331 Total 25.042 64 Note. Null model includes NA1, PENSC, PENSA

Coefficients Model Unstandardized Standard Error Standardized t p 0 (Intercept) 2.780 0.442 6.294 < .001 NA1 0.097 0.059 0.199 1.662 0.102 PENSC -0.126 0.077 -0.248 -1.636 0.107 PENSA -0.054 0.092 -0.089 -0.591 0.557 1 (Intercept) 1.819 0.600 3.029 0.004 NA1 0.082 0.057 0.168 1.432 0.157 PENSC -0.075 0.079 -0.148 -0.953 0.345 PENSA -0.065 0.097 -0.107 -0.673 0.504 CF2 0.228 0.094 0.294 2.421 0.019 AF2 0.086 0.112 0.097 0.770 0.444

Experimental Script (Italicised sections were spoken aloud; plain text was recorded) Step 1: Set up the game Run Mark of the Ninja up to the final menu prior to the first level Minimise the game Place the consent form, participant info sheet, and a pen on the desk Make sure your desk has the illustrated example grids, both relevant word grids, both the experiment and interview scripts, a pen, and a book

Appendices 133

Step 2: Explain the study and receive informed consent Before we start -- Just so you know, for most of this experiment, you’ll be given instructions from a pre-recorded audio track. I’ll be in the room the whole time, but my actual involvement will be fairly minimal. Hi. Thanks for coming in today to take part in this experiment. This audio recording has been prepared to provide a consistent experience for all participants. It’s not very long, so please wait until the end before asking questions. Feel free to read the participant consent form in front of you, but to summarise: Your participation will involve working on 2 word finding tasks, each taking 3 minutes, and then playing Mark of the Ninja for 30 minutes. Before and after you play Mark of the Ninja, you’ll fill out some surveys. The session should take around one hour. Please make sure to turn your phone, or any other devices, to silent mode, then place them on the desk to your right. If you need to use the bathroom, you should go now, as there won’t be any opportunity to go until we’re finished. If you have any questions, please ask them now. Otherwise, please sign and date the consent form, place both pages on the desk to your right, then let the researcher know that you’re ready to continue. Step 3: Begin the word finding task Control Condition Preamble: Ok, it’s time to begin the word-finding task. First, here’s some general information. The puzzle involves finding as many words as possible, and writing them down on the same page. You can make words of three or more letters, but each letter must connect to the previous one. You may not use the same letter square twice in a single word, but using two squares with the same letter is allowed. Outside of these basic rules, you are welcome to work on this task in whatever way makes sense to you. After the three minutes are up, the researcher will collect the paper, count the number of valid words you found, and let you know how well you did. One thing to keep in mind is that this puzzle is quite challenging, and beginners find that they only scratch the surface of the possible words. Just do the best you can, and you’ll improve quickly. If you have any questions, please ask them now; otherwise, let the researcher know that you are ready to begin. [Go behind a partition while the participant works on the task for 3 minutes] Your time is up. The researcher will now collect and score your paper. [Sit at a desk not clearly visible to the participant for ~30 seconds to “grade” it] Well done. I’ll give you a harder grid this time, since you obviously had no problems with the first one. Otherwise, the task is exactly the same; write down as many words as you can in three minutes. Please begin the task now. [Go behind a partition while the participant works on the task for 3 minutes] Your time is up. The researcher will now collect and score your paper. [Sit at a desk not clearly visible to the participant for ~30 seconds to “grade” it] This grid contains 74 words, and you found [X] of them. The average 3-minute score for this particular grid is [0.5X + 1], so you’ve performed well above the average.

Experimental Condition Preamble: Ok, it’s time to begin the word-finding task. First, here’s some general information. The puzzle involves finding as many words as possible, and writing them

134 Appendices

down on the same page. You can make words of three or more letters, but each letter must connect to the previous one. You may not use the same letter square twice in a single word, but using two squares with the same letter is allowed. Outside of these basic rules, you are welcome to work on this task in whatever way makes sense to you. After the three minutes are up, the researcher will collect the paper, count the number of valid words you found, and let you know how well you did. One thing to keep in mind is that this puzzle is quite difficult, and beginners (like you) usually don’t find very many words. Still, do the best you can, even if it seems hard. Maybe you’ll be lucky! If you have any questions, please ask them now; otherwise, let the researcher know that you are ready to begin. [Go behind a partition while the participant works on the task for 3 minutes] Your time is up. The researcher will now collect and score your paper. [Sit at a desk not clearly visible to the participant for ~30 seconds to “grade” it] Hmmmm.. I’ll give you an easier grid this time, since it seems like you had trouble with the first one. Otherwise, the task is exactly the same; write down as many words as you can in three minutes. Please begin the task now. [Go behind a partition while the participant works on the task for 3 minutes] Your time is up. The researcher will now collect and score your paper. Sit at a desk not clearly visible to the participant for ~30 seconds to “grade” it] This grid contains 104 words, and you found [X] of them. The average 3-minute score for this particular grid is [X + 7], so you’ve performed well below the average. Step 4: Participant completes the first set of surveys Now, please move over to the desk to your left. [beat] Once you’re there, you’ll notice three icons in the centre of the desktop. Please double-click the one on the left, labelled “survey 1”. [beat] Your three-digit participant number can be found on the post- it note attached to the monitor. Please note that the researcher is unable to answer any questions about the survey — you should interpret each question with your own judgement. Now, fill out this survey until you get to the screen telling you to stop. Step 5: Game play As I mentioned before, you’ll be playing Mark of the Ninja today. Mark of the Ninja is already running, so please select it from the taskbar at the bottom of the screen to bring it up. If you’ve never played the game before, that’s ok; the game will teach you how to play itself. Please put on the headphones now, as they won’t prevent you from hearing this recording. Make sure the headphones and chair height are comfortable for you before you begin playing. The researcher will be behind the partition the whole time; just play through as you normally would until you’re told to stop. If you have any questions, please ask them now; otherwise, let the researcher know that you are ready to begin. [30 minutes pass] Thirty minutes have passed since you began playing. Please pause the game using the Start button, and then exit the game. [beat] You should now be looking at the desktop. Please double-click the centre link labelled “survey 2”, and answer the survey questions until you reach the end. Step 6: Debriefing

Appendices 135

Ok, that’s the end of the experiment. Before I begin the debriefing, what did you think the aim of this experiment was? [pause for response; ask probing questions if necessary] The purpose of this experiment is to determine the extent to which videogames can improve player wellbeing after a negative event. To this end, the feedback you received about your performance, relative to others, in the word-finding task was completely false. You were told that the number of words that you found within the time limit was [more/less] than the average, but this was entirely arbitrary. I’m required to make sure that you understand this before you leave, to avoid any long-term effects of this manipulation. For this reason, can you please explain back to me your understanding of the word-finding task feedback? If you would like to receive a random game key as reward for your participation today, please open the third survey link on the desktop and enter an email address that you currently use. Prizes will be sent out periodically while the experiment is being run. Do you have any questions? Thanks! I need to ask you not to discuss this experiment with other people, as the feedback is less effective when participants know that it’s untrue. Does that make sense? Step 7: Interview (optional) I’d now like to invite you to participate in a short interview about your play session, and your thoughts about games and wellbeing more generally. This interview will take roughly fifteen minutes. Thank you; have a great day.

Puzzle Rule Sheet Each letter must connect to the previous one (this example shows a viable way of forming “LAPS”):

You may not use the same square twice to form a single word (the red arrow here demonstrates the illegal move in trying to form “PEP”):

However, the use of two squares that share the same letter is allowed (forming “PASS” in the example below):

136 Appendices

Puzzles (Control Condition)

Puzzles (Experimental Condition)

Interview Questions This first part of the interview is about your experience playing Mark of the Ninja during the experiment. We’re interested in knowing what’s most important to your experience of the game. So first, I’d like to know what kinds of things you noticed when you were playing. (This is probably too vague/open-ended for a lot of people, but I don’t want to give specific prompts unless they ask. If they do: “things like the stealth, audio triggers, the visual style... any aspect of the game, really.” The backup questions may also help if this prompt fails.) Potential follow-up questions: -Why do you think you noticed that? -What was important about that?

Backup Questions: -What did you like most about the game? (get specific) (aiming to determine the game elements that make MotN enjoyable) -How did you feel while you were playing? -What was it about the game... (aiming to determine the game elements that support affect/vitality) -How did the game challenge you? -What was it about the game... (aiming to determine the game elements that support competence)

Intentional Play for Wellbeing -Tell me about a time you played a videogame to feel better. -What was it about the game that helped? (try to get specific with game features) -Do games ever make you feel worse after you’re finished playing? -Why do you think that happens/could happen? -Do you think some games are better than others at making you feel better? Why?

Appendices 137

Appendix B Scale Development Final Subscale Items Competence Frustration 1) I felt disappointed with my performance in the game 2) I felt insecure about my abilities in the game 3) I felt like a failure because of the mistakes I made 4) I doubted if I could properly complete tasks in the game 5) I often felt inadequate or incompetent Autonomy Frustration 1) My actions in the game felt like obligations 2) I felt pushed to play in certain ways 3) I felt forced to follow directions regarding what to do in the game 4) I felt pressured to play the game in a particular way 5) I felt controlled and pressured

Exploratory Factor Analysis (EFA) Results Pre-Play (T1) Component Loadings RC 1 RC 2 RC 3 RC 4 Uniqueness AF1 0.783 . . . 0.371 AF2 0.808 . . . 0.250 AF3 0.753 . . . 0.237 AF4 0.829 . . . 0.330 AF5 0.627 . . . 0.500 AS1 . . 0.850 . 0.159 AS2 . . 0.846 . 0.181 AS3 . . 0.836 . 0.218 CF1 . 0.663 . . 0.403 CF2 . 0.763 . . 0.349 CF3 . 0.768 . . 0.504 CF4 . 0.694 . . 0.479 CF5 . 0.836 . . 0.375 CS1 . . . 0.671 0.438 CS2 . . . 0.729 0.286 CS3 . . . 0.756 0.481

Chi-squared Test Value df p Model 81.334 62 0.050

Additional fit indices

138 Appendices

RMSEA 90% CI TLI BIC Model 0.052 [0 - 0.072] 0.972 -228.493

Scree Plot

Post-Play (T2) Component Loadings RC 1 RC 2 RC 3 RC 4 Uniqueness AF1 0.724 . . . 0.494 AF2 0.766 . . . 0.413 AF3 0.772 . . . 0.455 AF4 0.789 . . . 0.335 AF5 0.661 . . . 0.411 AS1 . . . 0.612 0.324 AS2 . . . 0.870 0.150 AS3 . . . 0.838 0.206 CF1 . 0.711 . . 0.365 CF2 . 0.827 . . 0.443 CF3 . 0.843 . . 0.453 CF4 . 0.562 . . 0.457 CF5 . 0.599 . . 0.429 CS1 . . 0.656 . 0.373 CS2 . . 0.951 . 0.244 CS3 . . 0.742 . 0.384

Chi-squared Test Value df p Model 84.935 62 0.028

Additional fit indices RMSEA 90% CI TLI BIC Model 0.056 [0.017 - 0.075] 0.964 -224.892

Appendices 139

Scree Plot

Development Procedure Step 1: Scale Collation IMI (Ryan): 1 (not at all true) - 4 (somewhat true) - 7 (very true) BPNSFS (Chen): 1 (completely untrue) - 5 (completely true) Need Satisfaction Scale (La Guardia): 7 points, no accompanying text provided BMPN (Sheldon): 1 (no agreement) - 5 (much agreement) NSFS (Longo): 1 (strongly disagree) - 7 (strongly agree) IBQ (Rocchi): 1 (do not agree at all) - 7 (completely agree) PNTS (Bartholemew): 1 (strongly disagree) 7 - (strongly agree) Step 2: First Draft Competence Frustration: 1) I seriously doubted whether I was playing the game well* 2) I felt disappointed with my performance in the game 3) I felt insecure about my abilities in the game 4) I felt like a failure because of the mistakes I made 5) I felt unable to do well in the game** 6) I did things that made me feel incompetent*** 7) There were situations where I felt incapable*** 8) At times, I was given feedback that made me feel incompetent*** 9) There were situations in the game that made me feel inadequate*** 10) I doubted if I could properly complete tasks in the game 11) I occasionally felt incapable of succeeding in the game 12) I sometimes felt unable to overcome challenges in the game 13) I often felt inadequate or incompetent *would anyone think this at the time? (or maybe doubt just seems ambivalent) **strongly worded ***seems empirical/descriptive ("I did x") rather than psychological ("I felt x") Autonomy Frustration: 1) I felt like I had to do most things* 2) I felt forced to do many things I wouldn't choose to do

140 Appendices

3) My actions in the game felt like obligations 4) I had a lot of unnecessary pressures* 5) I felt like I had to do things against my will 6) I felt prevented from making choices in the game*** 7) I felt pushed to play in certain ways 8) I felt obliged to follow decisions made for me 9) I felt prevented from choosing how I completed tasks in the game*** 10) I felt forced to follow directions regarding what to do 11) I felt pressured to play the game in a particular way 12) I felt pressured to adopt certain behaviours 13) I felt like my choices had been limited** 14) I felt controlled and pressured to play in certain ways *wording is strange **seems empirical/descriptive ("I did x") rather than psychological ("I felt x") ***may not be important (or it may be an expected part of the game) Step 3: Second Draft Competence Frustration: 1) I felt disappointed with my performance in the game 2) I felt insecure about my abilities in the game 3) I felt like a failure because of the mistakes I made 4) I doubted if I could properly complete tasks in the game 5) I occasionally felt incapable of succeeding in the game 6) I sometimes felt unable to overcome challenges in the game 7) I felt inadequate or incompetent Autonomy Frustration: 1) I felt forced to do things I wouldn't choose to do 2) My actions in the game felt like obligations 3) I felt pushed to play in certain ways 4) I felt obliged to follow decisions made for me 5) I felt forced to follow directions regarding what to do 6) I felt pressured to play the game in a particular way 7) I felt pressured to adopt certain behaviours 8) I felt controlled and pressured to play in certain ways

Appendices 141

Appendix C Study 2 Materials Descriptive Statistics Age WeeklyPlay FirstPlayed 0 1 0 1 0 1 Valid 74 74 74 74 74 74 Missing 0 0 0 0 0 0 Mean 23.22 22.49 16.31 18.53 6.973 6.824 Std. Deviation 5.604 5.367 12.73 17.01 2.952 3.067 Minimum 17.00 17.00 0.000 0.000 2.000 1.000 Maximum 45.00 51.00 54.00 96.00 21.00 16.00 0 = control condition, 1 = experimental condition Gender Control condition Experimental condition Gender Count % Gender Count % F 14 18.9 F 16 21.6 M 60 81.1 M 57 77.0 Undisclosed 0 0.0 Undisclosed 1 1.4 Total 74 100.0 Total 74 100.0

Zero-order Correlations (Weekly Play, Age of First Play Experience) Kendall's Tau Correlations τ p τ p WeeklyPlay - FirstPlayed -0.123 0.039 FirstPlayed - IntEnj1 0.008 0.890 WeeklyPlay - IntEnj1 -0.034 0.549 FirstPlayed - CS1 0.080 0.190 WeeklyPlay - CS1 0.012 0.832 FirstPlayed - AS1 0.088 0.147 WeeklyPlay - AS1 -0.052 0.378 FirstPlayed - CF1 0.015 0.809 WeeklyPlay - CF1 0.044 0.463 FirstPlayed - AF1 -0.009 0.882 WeeklyPlay - AF1 0.100 0.081 FirstPlayed - Happy1 -0.006 0.916 WeeklyPlay - Happy1 -0.049 0.408 FirstPlayed - Calm1 -0.052 0.392 WeeklyPlay - Calm1 0.048 0.411 FirstPlayed - Anger1 -0.004 0.953 WeeklyPlay - Anger1 0.061 0.332 FirstPlayed - Sad1 0.005 0.940 WeeklyPlay - Sad1 -0.052 0.431 FirstPlayed - Tense1 0.082 0.202 WeeklyPlay - Tense1 < 0.001 0.992 FirstPlayed - V1 -0.051 0.398 WeeklyPlay - V1 0.035 0.544 FirstPlayed - Imm1 -0.042 0.488 WeeklyPlay - Imm1 -0.015 0.804 FirstPlayed - IntEnj2 -0.055 0.359 WeeklyPlay - IntEnj2 -0.003 0.952 FirstPlayed - CS2 -0.068 0.265 WeeklyPlay - CS2 0.056 0.337 FirstPlayed - AS2 -0.061 0.327 WeeklyPlay - AS2 0.117 0.051 FirstPlayed - CF2 -0.005 0.933 WeeklyPlay - CF2 0.024 0.690 FirstPlayed - AF2 0.003 0.962 WeeklyPlay - AF2 < 0.001 0.999 FirstPlayed - Happy2 -0.095 0.122 WeeklyPlay - Happy2 0.010 0.866 FirstPlayed - Calm2 -0.113 0.056 WeeklyPlay - Calm2 0.052 0.380 FirstPlayed - Anger2 0.062 0.345 WeeklyPlay - Anger2 0.068 0.280 FirstPlayed - Sad2 0.029 0.671

142 Appendices

WeeklyPlay - Sad2 -0.120 0.070 FirstPlayed - Tense2 0.044 0.489 WeeklyPlay - Tense2 -0.060 0.329 FirstPlayed - V2 -0.042 0.485 WeeklyPlay - V2 -0.036 0.528 FirstPlayed - Imm2 -0.032 0.610 WeeklyPlay - Imm2 0.056 0.347

Videogame Selection Criteria “Autonomy connotes an inner endorsement of one's actions, the sense that they emanate from oneself and are one's own. Autonomous action is thus chosen, but we use the term choice not as a cognitive concept, referring to decisions among behavioral options (e.g., Brehm & Brehm, 1981), but rather as an organismic concept anchored in the sense of a fuller, more integrated functioning. The more autonomous the behavior, the more it is endorsed by the whole self and is experienced as action for which one is responsible. [...] When autonomous, people experience themselves as initiators of their own behavior; they select desired outcomes and choose how to achieve them. Regulation through choice is characterized by flexibility and the absence of pressure. By contrast, being controlled is characterized by greater rigidity and the experience of having to do what one is doing. There is intention, but lacking is a true sense of choice. When controlled, people are, in the words of deCharms (1968), "pawns" to desired outcomes, even though they intend to achieve those outcomes.” (Deci1987; emphasis added) “Game designs also differ in the autonomy afforded within the game, such as the degree of choice one has over the sequence of actions, or the tasks and goals undertaken. Specifically, we expect autonomy to be enhanced by game designs that provide considerable flexibility over movement and strategies, choice over tasks and goals, and those where rewards are structured so as to provide feedback rather than to control the player’s behavior. In the gaming industry, this is reflected in the more recent movement towards a “procedural” structure, which is championed by a leading game developer, Will Wright, and is defined by a game’s ability to respond dynamically to an individual’s choices without constraining or anticipating them.” (Ryan2006; emphasis added)

PENS Autonomy items: [redacted to comply with Copyright law] Aesthetics/Experience: relaxation looseness accessibility flexibility clarity Dynamics/Behaviours: Perceiving multiple approaches to play (and their likely outcomes) Independently considering, selecting, and enacting preferred goals and strategies Adapting to controls without meaningful effort Mechanics/Elements: Absence of time pressure Informational (rather than controlling) feedback Clear communication of goals and potential strategies

Appendices 143

Simple and immediately responsive controls Initial Tangram Game Brief Initialisation: Running the program first displays a menu with a title ("Tangrams") and two buttons: one labelled "Start", placed directly underneath the title, and another labelled "Exit", placed in the bottom-right hand corner of the window. No music plays in this scene. Clicking "Start" sends users to the "Tangram" scene; clicking "Exit" immediately closes the program (i.e., without asking "are you sure"). "Tangram" scene: Users are given a short introduction to the controls (e.g., similar to the Examples below). On-screen instructions will be accompanied by spoken ones expressing similar information. (This script is yet to be recorded.) Each tangram is formed by correctly placing a set of seven shapes into a larger outline. These shapes should initially be placed outside the bounds of the tangram outline (e.g., in a separate section). Users can rotate shapes in 45 degree increments. Each shape can only be used once in forming the tangram, and each tangram requires all shapes to be used. Multiple solutions are sometimes possible. When a tangram is complete, each smaller shape should (over a short period of time) become the same colour to give the impression of a unified whole. Following this, another tangram should be automatically provided; if the user has solved all available tangrams, the first tangram should be presented again (excluding the introductory tasks). Tangrams should be presented in the same order every time the program is run. The program terminates ten minutes after the "Start" button is clicked in the main menu. At this time, users are notified (by the scripted audio track) that the program will shortly close. Controls are disabled after this verbal notification has fully played. The entire screen slowly becomes the same colour, starting from the edges of the screen and moving in. Audio volume is reduced at a similar rate. When the entire screen is the same colour, the program terminates. Controls: Pieces are moved by clicking and dragging with the mouse. Pieces are rotated by pressing Z or X (for right-handed players) or the left and right arrow keys (for left-handed players). The Windows key is disabled. Visual Style: The main window (where tangrams are solved) and the sidebar (where pieces are placed) can be different colours, but both should be relatively light and undistracting. Tangram pieces are distinctly coloured (keeping colourblindness in mind). The program always runs in fullscreen mode. Tangrams may be locked to a grid, or moved freely; whichever is easier to implement. If a grid is used, however, it should be invisible. Audio: An ambient audio track (to be provided) begins to play (and loops, if necessary) after the "Start" button is clicked. Spoken instructions play at each scene in the tutorial, and at regular intervals afterwards. T2 Zero-Order Correlation Table (Control Condition)

Competence Autonomy Anger Depression Tension Frustration Frustration

Anger —

Depression 0.454 —

144 Appendices

(< 0.001)

0.169 0.430 Tension — (0.074) (< 0.001)

Competence 0.288 0.501 0.330 — Frustration (0.002) (< 0.001) (< 0.001)

Autonomy 0.245 0.211 0.261 0.192 — Frustration (0.007) (0.025) (0.003) (0.024)

T2 Zero-Order Correlation Table (Experimental Condition)

Competence Autonomy Anger Depression Tension Frustration Frustration

Anger —

0.305 Depression — (0.004)

0.259 0.412 Tension — (0.008) (< 0.001)

Competence 0.209 0.318 0.327 — Frustration (0.026) (0.001) (< 0.001)

Autonomy 0.088 0.072 -0.016 0.070 — Frustration (0.338) (0.460) (0.860) (0.417)

H3 Multiple Regression Analyses (Non-significant Need Satisfaction Predictors) Anger Model Summary F Model R R² Adjusted R² RMSE R² Change df1 df2 p Change 0 0.438 0.192 0.175 0.489 0.192 11.377 3 144 < .001 1 0.516 0.266 0.240 0.469 0.074 7.200 2 142 0.001 Note. Null model includes Anger1, CS2, AS2 ANOVA Sum of Mean Model df F p Squares Square 0 Regression 8.164 3 2.721 11.38 < .001 Residual 34.443 144 0.239 Total 42.606 147 1 Regression 11.335 5 2.267 10.29 < .001 Residual 31.272 142 0.220 Total 42.606 147 Note. Null model includes Anger1, CS2, AS2

Appendices 145

Coefficients Model Unstandardized Standard Error Standardized t p 0 (Intercept) 1.294 0.236 5.485 < .001 Anger1 0.368 0.079 0.347 4.629 < .001 CS2 -0.124 0.041 -0.289 -3.023 0.003 AS2 0.023 0.043 0.051 0.531 0.596 1 (Intercept) 0.708 0.345 2.051 0.042 Anger1 0.271 0.080 0.256 3.365 < .001 CS2 -0.046 0.045 -0.107 -1.032 0.304 AS2 0.015 0.048 0.034 0.316 0.752 CF2 0.157 0.044 0.308 3.578 < .001 AF2 0.038 0.047 0.077 0.801 0.424

Depressed Affect Model Summary F Model R R² Adjusted R² RMSE R² Change df1 df2 p Change 0 0.406 0.165 0.147 0.316 0.165 9.476 3 144 < .001 1 0.510 0.260 0.234 0.299 0.095 9.110 2 142 < .001 Note. Null model includes Sad1, CS2, AS2 ANOVA Sum of Mean Model df F p Squares Square 0 Regression 2.830 3 0.943 9.476 < .001 Residual 14.336 144 0.100 Total 17.167 147 1 Regression 4.461 5 0.892 9.970 < .001 Residual 12.706 142 0.089 Total 17.167 147 Note. Null model includes Sad1, CS2, AS2 Coefficients Model Unstandardized Standard Error Standardized t p 0 (Intercept) 1.403 0.165 8.514 < .001 Sad1 0.150 0.069 0.166 2.159 0.033 CS2 -0.099 0.027 -0.362 -3.712 < .001 AS2 0.006 0.028 0.021 0.213 0.832 1 (Intercept) 1.047 0.237 4.417 < .001 Sad1 0.086 0.068 0.095 1.256 0.211 CS2 -0.046 0.028 -0.169 -1.644 0.102 AS2 -0.006 0.031 -0.021 -0.191 0.849 CF2 0.113 0.028 0.349 4.084 < .001 AF2 0.013 0.030 0.043 0.448 0.655

Tension

146 Appendices

Model Summary F Model R R² Adjusted R² RMSE R² Change df1 df2 p Change 0 0.501 0.251 0.236 0.639 0.251 16.121 3 144 < .001 1 0.536 0.287 0.262 0.628 0.036 3.575 2 142 0.031 Note. Null model includes Tense1, CS2, AS2 ANOVA Sum of Mean Model df F p Squares Square 0 Regression 19.74 3 6.580 16.12 < .001 Residual 58.78 144 0.408 Total 78.52 147 1 Regression 22.56 5 4.511 11.45 < .001 Residual 55.96 142 0.394 Total 78.52 147 Note. Null model includes Tense1, CS2, AS2 Coefficients Model Unstandardized Standard Error Standardized t p 0 (Intercept) 1.322 0.301 4.390 < .001 Tense1 0.436 0.077 0.419 5.662 < .001 CS2 -0.138 0.055 -0.236 -2.499 0.014 AS2 0.042 0.056 0.070 0.755 0.451 1 (Intercept) 1.048 0.463 2.265 0.025 Tense1 0.385 0.078 0.370 4.908 < .001 CS2 -0.076 0.059 -0.131 -1.289 0.200 AS2 0.010 0.065 0.017 0.159 0.874 CF2 0.155 0.058 0.224 2.674 0.008 AF2 -0.025 0.063 -0.037 -0.396 0.693

H4 Multiple Regression Analyses (Non-Significant Interactions) Happiness Model 1 R2 = 0.488, R2(adj) = 0.473, ΔR2 = 0.488, ΔF(4, 143) = 34.022, p < 0.001 Measure B SE β sr2 p (intercept) -0.191 0.346 - - 0.583 Happiness (T1) 0.231 0.066 0.215 0.044 < 0.001 Competence 0.242 0.050 0.370 0.082 < 0.001 Autonomy 0.035 0.057 0.051 0.001 0.541 Immersion 0.270 0.055 0.343 0.085 < 0.001 Model 2 R2 = 0.489, R2(adj) = 0.471, ΔR2 = 0.001, ΔF(1, 142) = 0.399, p = 0.529 Measure B SE β sr2 p (intercept) -0.898 1.172 - - 0.445 Happiness (T1) 0.234 0.066 0.218 0.045 < 0.001

Appendices 147

Competence 0.406 0.265 0.623 0.008 0.128 Autonomy 0.026 0.059 0.038 0.001 0.664 Immersion 0.394 0.205 0.501 0.013 0.056 Competence * Immersion -0.027 0.042 -0.334 0.001 0.529

Calmness Model 1 R2 = 0.286, R2(adj) = 0.266, ΔR2 = 0.286, ΔF(4, 143) = 14.290, p < 0.001 Measure B SE β sr2 p (intercept) 1.115 0.426 - - 0.010 Calmness (T1) 0.415 0.077 0.402 0.147 < 0.001 Competence 0.205 0.065 0.292 0.050 0.002 Autonomy -0.055 0.072 -0.075 0.003 0.451 Immersion 0.013 0.070 0.015 < 0.001 0.854 Model 2 R2 = 0.286, R2(adj) = 0.260, ΔR2 = 0.000, ΔF(1, 142) = 0.008, p = 0.928 Measure B SE β sr2 p (intercept) 0.997 1.374 - - 0.469 Calmness (T1) 0.415 0.077 0.402 0.145 < 0.001 Competence 0.205 0.065 0.292 0.050 0.002 Autonomy -0.033 0.245 -0.046 < 0.001 0.891 Immersion 0.034 0.248 0.041 < 0.001 0.890 Autonomy * Immersion -0.004 0.043 -0.048 < 0.001 0.928

Model 1 R2 = 0.286, R2(adj) = 0.266, ΔR2 = 0.286, ΔF(4, 143) = 14.290, p < 0.001 Measure B SE β sr2 p (intercept) 1.115 0.426 - - 0.010 Calmness (T1) 0.415 0.077 0.402 0.147 < 0.001 Competence 0.205 0.065 0.292 0.050 0.002 Autonomy -0.055 0.072 -0.075 0.003 0.451 Immersion 0.013 0.070 0.015 < 0.001 0.854 Model 2 R2 = 0.288, R2(adj) = 0.263, ΔR2 = 0.002, ΔF(1, 142) = 0.477, p = 0.491 Measure B SE β sr2 p (intercept) 2.105 1.496 - - 0.162 Calmness (T1) 0.409 0.077 0.396 0.141 < 0.001 Competence -0.024 0.337 -0.034 < 0.001 0.943 Autonomy -0.042 0.075 -0.057 0.002 0.579 Immersion -0.160 0.261 -0.189 0.002 0.539 Competence * Immersion 0.037 0.054 0.433 0.002 0.491

Vitality

148 Appendices

Model 1 R2 = 0.470, R2(adj) = 0.455, ΔR2 = 0.470, ΔF(4, 143) = 31.694, p < 0.001 Measure B SE β sr2 p (intercept) -1.012 0.515 - - 0.052 Vitality (T1) 0.391 0.065 0.375 0.133 < 0.001 Competence 0.289 0.082 0.276 0.046 < 0.001 Autonomy 0.141 0.093 0.130 0.008 0.131 Immersion 0.256 0.091 0.202 0.030 0.005 Model 2 R2 = 0.474, R2(adj) = 0.455, ΔR2 = 0.004, ΔF(1, 142) = 0.966, p = 0.327 Measure B SE β sr2 p (intercept) -2.619 1.714 - - 0.129 Vitality (T1) 0.387 0.065 0.371 0.130 < 0.001 Competence 0.298 0.083 0.284 0.048 < 0.001 Autonomy 0.436 0.313 0.399 0.007 0.167 Immersion 0.555 0.318 0.439 0.011 0.083 Autonomy * Immersion -0.054 0.055 -0.449 0.004 0.327

Model 1 R2 = 0.470, R2(adj) = 0.455, ΔR2 = 0.470, ΔF(4, 143) = 31.694, p < 0.001 Measure B SE β sr2 p (intercept) -1.012 0.515 - - 0.052 Vitality (T1) 0.391 0.065 0.375 0.133 < 0.001 Competence 0.289 0.082 0.276 0.046 < 0.001 Autonomy 0.141 0.093 0.130 0.008 0.131 Immersion 0.256 0.091 0.202 0.030 0.005 Model 2 R2 = 0.470, R2(adj) = 0.451, ΔR2 = 0.000, ΔF(1, 142) = 0.019, p = 0.891 Measure B SE β sr2 p (intercept) -1.259 1.879 - - 0.504 Vitality (T1) 0.391 0.066 0.374 0.132 < 0.001 Competence 0.348 0.435 0.332 0.002 0.426 Autonomy 0.138 0.096 0.127 0.008 0.155 Immersion 0.300 0.336 0.237 0.003 0.373 Competence * Immersion -0.010 0.069 -0.074 < 0.001 0.891

Intrinsic Motivation Model 1 R2 = 0.554, R2(adj) = 0.541, ΔR2 = 0.554, ΔF(4, 143) = 44.357, p < 0.001 Measure B SE β sr2 p (intercept) 0.772 0.420 - - 0.068 Intrinsic Motivation (T1) 0.046 0.049 0.053 0.003 0.350 Competence 0.231 0.060 0.275 0.046 < 0.001

Appendices 149

Autonomy 0.151 0.068 0.173 0.015 0.029 Immersion 0.466 0.067 0.459 0.151 < 0.001 Model 2 R2 = 0.555, R2(adj) = 0.540, ΔR2 = 0.001, ΔF(1, 142) = 0.486, p = 0.487 Measure B SE β sr2 p (intercept) -0.165 1.408 - - 0.907 Intrinsic Motivation (T1) 0.048 0.049 0.055 0.003 0.330 Competence 0.450 0.319 0.536 0.006 0.161 Autonomy 0.138 0.071 0.158 0.012 0.052 Immersion 0.630 0.246 0.622 0.021 0.011 Competence * Immersion -0.036 0.051 -0.344 0.002 0.487

Experimental Script (Italicised sections were spoken aloud; plain text was recorded) Step 1: Set up the game Run Spore; select a new world Minimise the game Place the consent form, participant info sheet, and a pen on the desk Make sure your desk has both the experiment and interview scripts, a pen, and a notebook Step 2: Explain the study and receive informed consent Before we start -- Just so you know, for most of this experiment, you’ll be given instructions from a pre-recorded audio track. I’ll be in the room the whole time, but my actual involvement will be fairly minimal. Hi. Thanks for coming in today. This audio recording has been prepared to provide a consistent experience for all participants. It’s not very long, so please wait until the end before asking questions. Feel free to read the information sheet in front of you, but to summarise: Your participation will involve playing a tangram game for 10 minutes, then playing Spore for 30 minutes. Before and after you play Spore, you’ll fill out some surveys. The entire session should take around one hour. Please turn your phone to silent mode, then leave it on a desk behind you until the study is complete. If you need to use the bathroom, you should go now – there won’t be any opportunity to go until we’re finished. If you have any questions, please ask them now. Otherwise, sign and date the consent form, then return it to the researcher. Step 3: Begin the tangram game Control Condition Please direct your attention to the “tangrams” icon in the top-right corner of the desktop. If you’ve never played with tangrams before, that’s ok; the game will teach you the rules and controls, and you can feel free to play in whatever way makes sense to you. Please note that your ability to solve tangrams is not being assessed. The game will close automatically after ten minutes. Make sure the chair height is comfortable for you before starting the game. If you have any questions, please ask; otherwise, go ahead and start the game.

150 Appendices

Experimental Condition Please direct your attention to the “tangrams” icon in the top-right corner of the desktop. If you’ve never played with tangrams before, that’s ok; the game will teach you the rules and controls, and your ability to solve tangrams is not being assessed. The game will close automatically after ten minutes. Make sure the chair height is comfortable for you before starting the game. If you have any questions, please ask; otherwise, go ahead and start the game. Step 4: Participant completes the first set of surveys: Now direct your attention to the three icons in the centre of the desktop. Double-click the one on the left, labelled “Survey 1”. [beat] Your participant number is written on a note attached to the bottom- left corner of the monitor. Be aware that the researcher is unable to answer any questions about the survey — you should interpret each question with your own judgement. Please begin the survey now. Step 5: Game play: It’s now time to start playing Spore. The game is already running; you can select it from the taskbar at the bottom of the screen to bring it up. If you’ve never played Spore, that’s ok; the game will teach you how to play itself. The researcher will be timing the session from behind a partition; just play through as you normally would, until you’re told to stop. Please put on the headphones now, and adjust them until you’re comfortable. If you have any questions, please ask; otherwise, let the researcher know that you are ready to begin. [30 minutes pass] Thirty minutes have passed since you began playing. Please use the game menus to exit Spore. [beat] You should now be looking at the desktop. Please double-click the centre link labelled “Survey 2”, and answer the survey questions until you reach the end. Step 6: Debriefing Ok, that’s the end of the experiment. Before I begin the debriefing, what did you think the aim of this experiment was? The purpose of this experiment is to determine the extent to which videogames can improve player wellbeing after a negative event. To this end, you played a tangram game with high or low levels of control. If you would like to be rewarded for your participation today, please open the third survey link on the desktop and follow the instructions. Do you have any questions? Thanks! I need to ask you not to discuss this experiment with other people, as it’s less effective when participants know how it works. Does that make sense? Step 7: Interview (optional) I’d now like to invite you to participate in a short interview about your play session, and your thoughts about games and wellbeing more generally. This interview will take roughly fifteen minutes. Thank you; have a great day. Interview Questions Autonomy Spore is a fairly open-ended game, but a lot of people also enjoy more linear games, where you're not given as much opportunity to customise a character, or decide how you

Appendices 151

want to approach situations. How do you feel when you play games that emphasise options, like Spore, compared to how you feel when you play more linear games? (This could lead into a discussion about how they like the experience being offered in a linear game, preference for the pick-and-choose nature of open-world/sandbox, rejection of the dichotomy so they can talk about The Witcher 3...) Tell me about a time where you stopped playing a game because you didn't want to do what was required to progress. Tell me about a time where you forced yourself to keep playing a game even though you didn't enjoy what you were doing to progress. Why did you do that? Tell me about a time where you were willing to act a certain way to progress, even without being given a choice. Tell me about a time where you played a game that offered multiple options, but none of them appealed to you. When are the decisions you make in games most interesting? Least interesting? Eudaimonic Wellbeing “Wellbeing”, or “wellness”, are becoming more popular terms in our culture – what does “wellbeing” mean to you? Self-Determination Theory – which is what this study is founded in – describes wellbeing as something more than just feeling happy: there are all sorts of things people do to feel good in the moment that aren’t particularly healthy in the long term. Instead, wellbeing is said to result from acting in ways that we reflectively value. Is playing videogames, in general, something you reflectively value? (Why/why not?) Are there any particular games that you feel this way about? (Tell me more...)

152 Appendices