Dead End: The Influence of the PC on Player Emotional State and Actions A Thesis Submitted to the Faculty of Drexel University by Ian C. Woskey in partial fulfillment of the requirements for the degree of Master of Digital Media June 2013

© Copyright 2013 Ian C. Woskey All Rights Reserved.

ii

DEDICATIONS

For all the late nights, and friends I’ve lost. For the few that believed in me, when I could not. For the path that lies ahead of me, and the future I see.

iii

ACKNOWLEDGEMENTS

My parents, without whom I would not be where I am or who I am today.

Bobby Speck, my ever dedicated friend and partner in crime, who was always there for me during the darkest of days.

My dedicated thesis committee, Jichen Zhu, James Malazita, and Garth DeAngelis. Without your insights and wisdom this project would not have come together as well as it has.

RenderHeads, who generously donated the software used to record Dead End gameplay for testing.

iv

TABLE OF CONTENTS

Contents LIST OF TABLES ................................................................................................................................. vi LIST OF FIGURES ............................................................................................................................... vii ABSTRACT......................................................................................................................................... viii 1. KEY TERMS ...................................................................................................................................... 1 2. INTRODUCTION & THESIS OVERVIEW ........................................................................................ 2 3. LITERATURE REVEIW .................................................................................................................... 5 3.1. THE PLAYER, THE PC, AND EMOTIONS ................................................................................ 5 3.2. PRESSENCE, DIFFICULTY, AND SURVIVAL HORRROR ...................................................... 9 4. METHOD ......................................................................................................................................... 12 4.1 GAME AND TEST STRUCTURE .............................................................................................. 12 4.2 THE PC ...................................................................................................................................... 16 4.3 THE ENEMY: SKINLESS ......................................................................................................... 19 4.4 BALANCING............................................................................................................................. 20 4.5 TESTING ................................................................................................................................... 20 5. DATA EVALUATION ..................................................................................................................... 22 5.1 GROUNDED THEORY .............................................................................................................. 24 5.2 QUALITATIVE DATA ............................................................................................................... 25 5.3 PC CHARACTERIZATION ........................................................................................................ 27 5.4 NARRATION EVAULUATION ................................................................................................. 31 5.6 QUANTITATIVE DATA ............................................................................................................ 35 5.7 DATA COMPARISON ............................................................................................................... 38 6. CONCLUSION ................................................................................................................................. 40

v

7. LIST OF REFERENCES ................................................................................................................... 43 8. APPENDIX ....................................................................................................................................... 46 8.1 POSITIVE AND NEGATIVE POST VIDEO ANSWERS ...................................................... 46

vi

LIST OF TABLES Table 1: Categorization Parameters ....................................................................................................... 33 Table 2: Averaged Coded Metrics .......................................................................................................... 37

vii

LIST OF FIGURES Figure 1: The game start prompt and view............................................................................................. 14 Figure 2: A diagram showing the different sections of the maze. ........................................................... 15 Figure 3: Left, fearful PC, right, confident PC.......................................................................................... 16 Figure 4: Top, confident walk animation. Bottom, fearful walk animation. ............................................. 17 Figure 5: The Skinless ............................................................................................................................ 19

viii

ABSTRACT Dead End: The Influence of the PC on Player Emotional State and Actions Ian C. Woskey Jichen Zhu, Ph.D.

This project endeavors to examine the effect of externalized emotional display of the player character (PC) on the emotional state of the player. It examines the possibility of instilling fear or confidence in the player by expressing emotional cues through PC animations. A large body of work exists examining this relationship on a broad scale, including the effect of presence and embodiment on player identification. In the literature review, a series of evidence is outlined that suggests the PC can have an emotional impact on the player—specifically when the player is feeling a sense of presence—and that emotions affect action. Presence in this context is the degree to which a player feels embodied and immersed in the game world. An action horror game called Dead End was developed to test these hypotheses. The mechanics and setup of the game will be outlined and the process for testing will also be explained. The playtesting was evaluated using digital subjectivities and ethnographic human factors. The quantitative results, based on metrics coded from participant play sessions, show no statistically significant differences in performance or play style. Qualitative results were generated using unstructured participant interviews. The interview results demonstrate that there were noticeable differences in how participants perceived the PC and the game.

ix

1

1. KEY TERMS Player The individual actively playing the game. Player Character (PC) The virtual entity through which the player acts in the game world. Presence The players sense of “being there” when playing game. Ambient difficulty Game difficulty not based on current situation, but the anticipation of future or imminent difficult encounters.

2

2. INTRODUCTION & THESIS OVERVIEW

The player character (PC) is the player’s vehicle for all interactions within the game world [1]. This entity mediates every interaction between the player and the game. It “…operates as a tool that extends the player’s ability to realise [sic] affordances within the game-world” [1]. This project endeavors to apply the strength of this relationship to see if desired emotional states can be elicited from the player. A game called Dead End was constructed to test this connection. In this game, the PC has a set of two animations: fearful and confident. The effect of these separate PC emotional states on the player was examined through both digital subjectivities and ethnographic human factors in user test studies. In the literature review, the similarities between how individuals process personal emotional response and perceived external emotional response will be shown [2]. Studies that show similarities between how individuals perceive the emotions of both real and artificial agents will also be examined [3]. This theory comes full circle with an explanation of the relationship between player and PC as both an extension of self and a vehicle whose restrictions control intentions. This will reveal the PC as both an external and internal regulator, acting on behalf of the player and influencing how the player chooses to act [1]. Without providing a definition for specific emotions, the PC was animated to show either fear or confidence during the participant playthrough. User interviews were conducted post-trial in an attempt to evaluate positive or negative feelings caused by these animations. The terms “fear” and “confidence” are used to define animations based on the goal of expression in animating. Though these animations may instill fear or confidence, whether or not an animation will elicit the specific correlated emotion is a part of what was tested, not an assertion made. The hypotheses

3

proposed are that animations displaying negative (fear) may instill a similar negative emotional response in the player, and conversely animations displaying positive (confident) emotions may instill a similar negative emotional response. Hypothesis 1: The externalized emotional state of the player character during the course of play will impact the emotional state of the player. Hypothesis 2: This emotional response will produce an effect on player performance and actions. Differing animation states based on game character emotional state and other factors is not a foreign concept to video games. The games of the Resident Evil series communicate the effects of taking damage both on a HUD health bar, and in the way the PC moves. As a PC in this series takes more damage they will begin to hold themselves as if clutching a wound, and in some cases move slower [4]. In the Uncharted series, there are sequences that illustrate the player sustaining injury through a story sequence. The PC will communicate this injury to the player during gameplay by staggering and clutching wounds [5]. In the Black and White game series, the player is put in charge of a mythical beast that it must care for and train. To train this creature the player is given the ability to punish or encourage it. Based on the player’s use of this mechanic the creature will give positive or negative emotional feedback in the form of corresponding animations [6]. In The Sims series, the player is put in charge of a number of semi-autonomous agents he or she creates [7]. The entirety of communication between these agents and the player is handled through body language, as they speak in a language of jargon called Simlish [8]. While these animation shifts are meant to effect the player, our research has found no prior studies that have attempted to codify and document the impact these game elements have on the player. It is possible that this technique is much more impactful on the

4

player than the simple nature of these animations shifts would suggest. If the concepts these hypotheses will endeavor to study are validated by positive results, and the nature of the playerPC relationship is as strong as suggested, there are constructive implications to the future of PC design. It may be possible for game designers to instill desired emotional states into the player by animating the PC to express these emotions. This would provide a significant functional and artistic impact on game design. For the purposes of testing these hypotheses, a third person action horror game called Dead End was developed. Dead End takes place in a dark maze populated with creatures that will try to kill the PC. A more specific dissection of the gameplay can be found in the methodology section. The PC was designed with two differing animations sets: fearful and confident. These different animations are purely aesthetic and have no direct impact on gameplay. Participants were split evenly between a fearful animation and confident animation group. We tested to see if this passive gameplay element could elicit an active effect on player action. As a pilot study, this test was not intended to arrive at a conclusive result, but to create a test result-based sketch of the possible correlations between PC behavior and player behavior. The results of this test were determined through a combination of gameplay metrics recording specific player actions, and an interview with each participant. These results are split between the qualitative analysis and our quantitative analysis. Our qualitative analysis showed that there was a strong correlation between descriptions of in-game actions and test group. Our qualitative results however showed no statistically significant difference between the two groups, or even between individual participants in some cases.

5

3. LITERATURE REVEIW 3.1. THE PLAYER, THE PC, AND EMOTIONS It is necessary to establish, on at least a general basis, that video games are capable of influencing the emotions of the player. Bernhaupt et al. conducted a study to determine the effectiveness of emotional interfaces on the emotional state of users. Players were charged with growing a virtual potted flower based on how positive their facial expressions were. It was found that their game “…directly influenc[ed the]...emotional states of participants” [9]. Though the gameplay aspects of this game were minimal, consisting only of smiling at a camera and watching the results, this process elicited positive emotional shifts in players. This study shows that even simplistic game design elements can elicit emotional reactions from the player [9]. Going forward, we will show that humanoid agents can elicit even more powerful emotional feedback in games. In Lankoski’s examination of player engagement with the PC [2], he explains the nature of affective mimicry. The basis of this phenomena is that an individual “involuntarily and automatically mimics another person’s expressed affects… people tend to smile and feel pleasure when they see another person smiling (or, conversely, expe-rience [sic] affects relating to pain if they see someone getting hurt)” [2]. Moody et al., in a study of affective mimicry, found that this phenomenon extends to body movements. Though this point was nonconclusive, Moody suggests that “[b]ody mimicry may be more evident when there is more consistent motivation to focus on another’s specific body movements (as when there are task demands beyond simply watching” [10]. The PC, as the centerpiece of every in-game interaction [1], represents a body whose movements, based on notions of embodiment [11], are at the center of the player’s attention. But the PC is an artificial entity, not a person. For

6

these concepts to apply to our hypotheses, it is necessary to establish that a player will evaluate the artificial PC in the same way he or she would evaluate a real person. A study released in 2012 shows that “emotional body language displayed by an agent or a human is interpreted in a similar way in terms of recognition” [3]. Participant in this study were asked to identify emotions demonstrated by a different array of artificial characters and a real person. Identification of emotions was consistent across these different actors. People passively engage in affective mimicry, shifting emotional states toward the emotional states of other perceived individuals [10]. People also evaluate the emotional expressions of artificial agents in the same way they evaluate other people [3]. From these two studies there is evidence to suggest that people engage in affective mimicry with artificial agents. It based on this logic that the hypotheses were developed. The PC is not, however, merely an individual artificial agent but a humanoid vehicle embodied and controlled by the player [1]. There is evidence to show that the perceptual relationship between player and PC is stronger than the perceptual relationship between two people, or a person and an artificial humanoid. When looked at in the context of affective mimicry, the PC represents an externally visualized perception of self. This is not to say that notions of self are confused by the PC. The PC embodies the player in the game world, as an extension of self. The PC represents the player’s only link to interact with the game. Presence is, “the sense of experiencing mediated space as unmediated” [1]. Presence is the player’s sense of being in the game world. When the player feels presence, when he or she feels in the game world, the natural place for the player to exist in the game is the PC. In this context the player’s sense of presence is the degree to which they identify the PC in a sense of self, “identification with one's character may be understood as the player entering a state where

7

he or she has an experience of 'being' the character” [12]. This relationship has obvious implications to how the PC’s animations could influence the player. The player, experiencing the game through the PC and viewing the PC as an extension of self, may perceive autonomous attributes of the PC within this extended notion of self. We have shown that these animations will be viewed through the phenomena of affective mimicry, instilling feelings of motion and emotion into the player [10]. However, current notions of affective mimicry are based on viewing an external autonomous entity, the player-PC relationship is based on viewing an external dependent entity. This positions this relationship, and its connection to affective mimicry, more akin to the player looking into a mirror, than at an autonomous entity. While in an informal manner the significance of this difference can be inferred, sadly there are no studies examining the effects of affective mimicry while participants look into a mirror. Further evidence concerning the relationship between player and PC shows the affect that the PC has on the player’s thought processes while playing the game. The player character is the agent through which virtually every in-game action is acted, “player character is both the input and the output language - it is through the character the user adds input to the system, and through the character that the user perceives the output calculated by the system" [12]. The player evaluates how to interact with the world based on feedback received from the PC. What the player can accomplish in game is limited by the functionalities and mechanics that make up the palette of possible PC action, “…the intentions of the player are mediated by the restrictions implicit with the …[PC]…, over time these restrictions …become embodied in the intentions and actions of the player” [1]. As the player becomes accustomed to the mechanics and actions that the PC is capable of, the way the player thinks about how he or she will act in the game is effected, embodying the restrictions of the PC [1]. In this way, how fast the PC

8

can move limits how quickly the player will expect to arrive at a destination, and how far the PC can jump will become how far the player expects to jump. The restriction of the PC’s pseudo-physical actions become the expected restrictions on the player’s embodied perception of their surrogate capabilities [1]. When this concept is examined in context with the implications of affective mimicry on the relationship between player and PC, we see that it is possible that this relational influence on how players form intentions within the game world may extend past functionality to form. We can draw the following question from this possible correlation: if the player’s intentions to use mechanics are formed by the mechanics themselves, are there similar implications to the aesthetics and emotional state of the PC and their influence on the player? In this way the emotional state of the PC may have a restrictive or additive impact on the player’s emotional state. We have so far outlined our reasoning behind why it may be possible to instill emotions in the player. In this section we will explain why emotions are important, and the reasoning behind the hypothesis that affecting emotions will affect player actions. Emotional response is an important part of modern gaming. “It can be argued that one of the main reasons for playing games is to achieve an emotional reaction of the player. To be surprised, happy, angry or anxious – to perceive different emotional states – is one of the main reasons to play games” [9]. Emotions are a powerful affecter on the way people experience media, and the actions they take. As video games are situated as both a perceived and acted upon media, emotions represent an important factor in game design. There are many studies that document the effect of emotions on decision making [13]. Berhaupt asserts “… emotions influence people’s attitude towards their current and next action and there is evidence that they play an essential role in rational decision making, perception, learning, and other cognitive functions” [9].

9

Cacioppo, in his paper documenting specific studies on emotions and general trends in emotional testing goals and techniques, asserts that while emotions are not the primitive action disrupting force that some assume, they do affect action. While he submits to the notion that extreme emotion can adversely affect decision making, he outlines emotion as a necessary grounding force for decision making; in fact emotion positively guides our decision making process [13]. To illustrate this concept, Cacioppo uses the example of Elliot, a man who developed a brain tumor that left him unable to feel emotion. Testing proved that none of Elliot’s other mental faculties had been affected, but “…Elliot began behaving irrationally… the lack of emotional guidance rendered decision making a dangerous game of roulette” [13]. Cacioppo asserts the importance of emotion in generating grounded contextually based decision making. If emotions play as strong of a role in decision making as Cacioppo suggests, then it is possible that affecting emotions will affect decision making games.

3.2. PRESSENCE, DIFFICULTY, AND SURVIVAL HORRROR

Presence is a common goal in video game design (Ravaja et al., 2004). Presence represents the degree to which a player feels immersed or embodied within a mediated experience, in this case within a game world. By definition, presence has a direct effect on a player’s connection with the PC as an agent that blurs the line between artificial and self [12]. Not every genre of game seeks presence, but many rely heavily on it. Games represent an artificial or mediated experience. To create a game that can impact the player in a way that is more significant than objectively experiencing it as a game, it is necessary to create a sense of presence [1]. This connection between presence and the relationship between player and PC

10

leads to the conclusion that the effect of the PC’s emotional state on the player’s emotional state, if any, will be limited by the degree to which the player achieves presence while playing a game. The purpose of the animation shifts in this game are not meant to communicate game state information to the player, as found in the shifting animations in games like Resident Evil. In Resident Evil, where animation shifts are affected only by remaining health, the shifting animation is as effective as the player’s ability to perceive the shift [4]. We are not suggesting that presence will affect the player’s ability to perceive the change in animation state. This project will not test the player’s ability to evaluate the event of a shift in animation, but the effect animations may have on the player’s decision making concerning in game actions. We believe this effect is influenced by presence. There is evidence that games like Tetris do elicit a certain degree of presence in the player [14]. The question this project explored was not related to the kind of immersion achieved in this kind of game. These tests were performed within a 3rd person, charactercentric, realistic, combat-oriented game. The results generated by this project, and any conclusions drawn from those results, are primarily relevant within similar games. It will be possible to evaluate games like The Legend of Zelda or Dark Souls based on this project’s findings, but we will not suggest the impact of our findings on games within genres that do not possess similar qualities (such as Tetris). The theories outlined concerning affective mimicry are predicated by the player’s sense of presence. Without a sense of presence, the player is more likely to view the PC purely as a vehicle through which he or she plays the game. In a study testing level of presence inside virtual reality environments that respond to the player’s physical movements, it was found that “the more present a user feels in our stressful environment, the more physiological reaction the

11

user will exhibit” [15] . In this test users were positioned near a large virtual hole within the virtual environment. The more real and “present” this danger seemed, the more impact this virtual object was able to elicit from their users. The results of a study testing the effects of genre, realism, and difficulty on presence in participants showed that level of visual realism and difficulty enhance the level of presence in the player. Participants were tasked with playing games across multiple genres and difficulty levels. This study revealed that “[a] higher difficulty level of the game taxes the cognitive resources, thereby diminishing attention paid to cues signaling that the game environment is not real” [14]. This information influenced formative design decisions during the creation process of Dead End. The action horror genre was chosen for Dead End based on a combination of features inherent in the genre that elicit presence. Presence and consequence are necessary for a mediated horror experience. At a basic level the genre relies on the premise that, without putting the player in any true physical danger, fear can be instilled in the player [16]. This artificial apprehension creates a sense of consequence separate from concrete game mechanics. Presence is necessary to achieve meaningful consequence [1]. Consequence, and the difficulty that affects consequence, is what causes the player to fear. The positive correlation between game difficulty and presence shows that the difficulty inherent in the horror genre could be a powerful force for creating presence. Perron’s description of the genre notes that horror games both require and demand a sense of presence [16]. As presence is a necessary factor in the proposed game project, using a genre that requires and creates presence is a natural choice. Survival horror as a genre is characterized by a near-constant sense of apprehension and stress. Games like Resident Evil, Silent Hill, Amnesia: The Dark Descent, and many others

12

in the genre are constructed of sporadic and unexpected encounters. These encounters can take the player by surprise, and traditionally put the PC into life or death (win or lose) situations. The techniques within these games are generally used to make players uneasy, even in situations that may not contain explicitly dangerous content. This creates constant apprehension in the player as one of these encounters could lie around any corner, and in any seemingly-innocuous room.

The specific aesthetic and gameplay techniques used to

accomplish this vary from game to game, but one characteristic remains the same: ambient difficulty/stress. Survival horror games can turn the act of simply walking down a hall into an encounter characterized by the same tension as a more interactive encounter [16]. While no specific gameplay feature is present in this hall, based on the player’s expectations, previous experience in game, and knowledge of the genre, the player is actively affected by the fear of what could occur. “When we become aware of a danger, we are knocked off-balance. We start to anticipate what could happen next and to begin to dread that the threat might materialize itself” [16]. The difficulty and stress inherent in ambient difficulty is a powerful source of presence.

4. METHOD 4.1 GAME AND TEST STRUCTURE A short 3rd person action horror game called Dead End was constructed to test our hypotheses. This game was designed to facilitate examination of the influence of PC externalized emotion on player emotional state, perception, and performance. . A custom game

13

was developed to ensure complete control over game mechanics and testing parameters. A 40 person user test study was conducted in which each participant played through this game. The PC in Dead End was designed with two separate animations for each action: fearful and confident. Each participant was tasked with taking control of this PC to navigate and survive within the game environment. The game has three phases: a start screen, a brief tutorial, and the core game. The start menu allows the test administrator to enter the participant’s number, choose either Fear group (participants given fearful animation sets) or Confidence group (participants given confident animation sets), and start the tutorial phase. The tutorial phase was included to ensure that each participant understood how to control the basic gameplay mechanics before beginning the actual game. The player is tasked with attacking a cube to open a door, sprinting down a hallway to open a door, and crouching in front of the last door to open it. After completing the tutorial, the player is dropped into the actual game. The player is presented with a short prompt: “Find the Bonfire at the center of the maze. You have 15 minutes. Hint: look around this room”. Heeding this prompt, shown in figure 1, the player is intended to notice a nearby tower that denotes the end of the maze. The game ends when a participant reaches this tower or after 15 minutes has passed.

14

Figure 1: The opening Prompt.

The maze is populated with 47 zombie-like creatures internally referred to as “the skinless.” Upon seeing or hearing the PC, these enemies path towards him and attack. The enemies, the maze (shown in figure 2), and the death mechanic in this game, represent the primary gameplay difficulties. Our sound design and aesthetics represented a more abstract or passive difficulty within the game based on the concepts of ambient difficulty outlined in the literature review.

15

Figure 2: A diagram showing the different sections of the maze.

16

4.2 THE PC

Figure 3: Left, fearful PC, right, confident PC

The PC was created with two separate animations for each in-game action: fearful and confident. The parameters with which these animations were developed focused on incorporating appropriate body language expressing either aggression and confidence or timidity and fear in the PC. An example comparing both expressive versions of the run animation can be seen in figure 4. Creating affective expression was accomplished using an iterative process. Versions of these animations were viewed and evaluated candidly by third parties which consisted of primarily digital media students and other media specialists. Based on feedback and correct or incorrect evaluation of expression, animations were tweaked until a final product was reached. Though expression was an important goal for our testing, ensuring

17

no functional difference between the two PCs was equally as important. Movement speed and the physical properties of attacking were both determined entirely by the animations. It was necessary to ensure that the motions tied to these functionalities were consistent between both groups. This was accomplished by standardizing specific parts of the PC’s movements between both versions.

Figure 4: Top, confident walk animation. Bottom, fearful walk animation.

The PC has two core game mechanics: the variable movement types that make up the stealth system, and combat. The combat mechanic can be simplified down to two basic functionalities: positioning and attacking. The PC is controlled using the “W”, “A”, “S”, and

18

“D” keys, the standard movement control system for personal computer games. The PC can be turned using the mouse. This control scheme was chosen to allow players to gain positioning advantage over enemies. Players can attack using the left mouse button. This causes the PC to swing his pickaxe in a horizontal arch in front of him from right to left. If this button is pressed twice in quick succession the PC will attack a second time from left to right. The stealth mechanic in Dead End is based on 3 different movement types: walking, running, and crouching. Enemies can perceive the player through direct visual contact or through auditory contact with the player. Auditory contact is generated based on a variable radius around the PC. The faster the player is moving, the larger this radius becomes. An enemy within this radius will achieve auditory contact with the player. The intention was to create two opposed movement types: one that favored a more confident play style, giving the player a boost in speed at a cost of increased danger, and one that favored more fearful play, giving players added safety at the cost of slower speed. The use of this mechanic was monitored to see if any trends developed.

19

4.3 THE ENEMY: SKINLESS

Figure 5: The Skinless.

When not aware of the player enemy agents will remain stationary, looping an idle animation. There are two ways for the enemy artificial intelligence (AI) to become aware of the PC: direct visual contact or auditory contact. Visual contact is created using a raycast which constantly points toward the center of the PC’s chest. This raycast returns the tag of every object it touches. When this tag returns “Player” the enemy will begin moving toward the PC. Aural perception is based on a diameter set around the player. This diameter becomes wider based on the speed the player is moving. When an enemy enters this diameter, it becomes aware of the PC’s position and begins moving toward it.

20

4.4 BALANCING Our design philosophy in every aspect of balancing the game mechanics of Dead End was to create a system in which participants could achieve varying degrees of success. The game was tested using participants who self-identified as familiar with or experienced playing video games. It was necessary to accommodate many variable skill levels within the desired group, imposing a restriction. These two concepts needed to work in harmony to create an accessible experience in which small differences in player action could create noticeable differences in performance. The movement system allowed players to choose between an aggressive, neutral, or timid option. This relationship bleeds into the combat system, giving players the option to progress stealthily and avoid combat, or engage in combat aggressively. There were many variables that needed to be balanced to create the desired experience: damage and health ratios (combat), enemy auditory radius (stealth), and enemy positioning within the maze (environment). Each of these factors came together to create the experience desired. 4.5 TESTING Test participants were contacted in three ways: a flyer posted around the URBN Center of Drexel University, direct contact recruiting within the URBN Center labs, and word of mouth. Upon entering the testing room, participants were first tasked with filling out a small survey before playing the game. Each participant was asked to note their gender, age, and college major. The survey consisted of five questions. Participants were tasked with ranking their familiarity with video games, their skill at playing video games, and their familiarity with the survival horror genre, each on a scale from one to five. The parameters of this study were based on using a pool of participants who were familiar with video games, and preferably avid game players. These questions were used to ensure each participant’s viability as a test subject

21

and later as a reference for our findings within individual play tests. Participants were asked to list games recently played that included melee combat. These answers were used to make sure that participants were familiar with the standard mechanics of close quarters combat. It was necessary to ensure participants were familiar with orienting a PC in relation to enemies and assessing the parameters of a real time combat system to connect with attacks. This ensured that any trends in attack accuracy would be less likely to be attributable to variable player skill. The final section of the survey explained the test procedure, other information required by IRB parameters, listed the controls used in the game, and asked participants to circle “yes” or “no” based on his or her understanding of the controls. After completing the survey, participants were taken into a room set up with a computer and speakers. The lights were turned off, and the door was closed to create a proper atmosphere for playing a horror game. Participants would remain in this room for the duration of the playtest, which at most would take fifteen minutes plus the time necessary to complete the tutorial. Four separate files were produced from the game engine each play session: a text file coding variables related to player action, a simple printout of the final coded values for convenience, and a video screen capture showing the entirety of the main game play through. After each play session, participants were tasked with narrating his or her game experience while watching the video of his or her play through. Each participant was given the same prompt to structure narration. The test administrator asked participants to narrate his or her experience, describe what was going through his or her head, why he or she made certain decisions, and what strategy, if any, he or she used. Additional comments made by the test giver were limited to reinforcing the starting prompt and encouraging participants to further elaborate on given statements. The third phase of the interview contained five standard

22

questions. These questions related to the participants evaluation of the PC, his or her evaluation of personal skill within the game, and their sense of tension/stress and fear within the game. Quantitative data used to evaluate participant play sessions was comprised of 12 individual metrics coded from gameplay. Seven metrics were related to combat: attack presses, PC attacks, PC attack hits, PC attack misses, successful enemy attacks, PC deaths, and enemies killed. Four metrics were related to time spent using each movement type: time spent stationary, time spent walking, time spent sneaking, and time spent running. The final metric logged total play time. The purpose of these metrics was to evaluate our second hypothesis, that differing animation states would create an impact on player performance and action. The purpose of collecting these metrics was to create a data based impression of player performance. This data was used to compare participant performance and action in the game between Fear Group and Confident Group.

5. DATA EVALUATION Participant play sessions were evaluated using two methods: qualitative and quantitative. The qualitative data collection and evaluation leveraged ethnographic human factors. Human factors centers around the examination of the interactions between humans and systems. The goal of this field is to better facilitate and understand these interactions in practical applications. Human factors function using a,”…holistic, human-centered approach

23

work systems design…” [17]. This discipline was chosen for our evaluations based on its practical approach and the ability to evaluate the relationship between participants and the game system. Human factors evaluations were conducted in the form of unstructured participant interviews. Interviews were conducted while each participant and the test giver viewed a video of that participants play session. This video was generated in-engine from each play session. The purpose of these interviews was to gather information related to the first hypothesis, attempting to examine the emotional state and perceptions of participants. The quantitative data collection and evaluation used methods common to digital subjectivities, coding metrics directly from gameplay. This technique was developed based on similar techniques used to evaluate gameplay [18]. The statistics coded were directly related to the use of the PC. These metrics were intended to articulate differences in play style and performance. This data was used to examine our second hypothesis, that any shifts in player emotional state would create performance and play style differences between our two test groups. The results of our quantitative study showed that within game mechanics used and metrics coded there was no statistically significant differences between the two groups. Conversely, the qualitative evaluations showed there was a significant difference in the way participants perceived his or her game experience. Similar to the quantitative data, there was a degree of variance within each testing group, but the results showed a strong trend toward negative comments and descriptions of in-game actions in Fear Group, and a less pronounced trend toward positive comments and descriptions within Confident Group. This data seems to show a positive correlation based on our first hypothesis, the externalized PC emotion will

24

effect player state and perception of the game. However the results also suggests that it is possible our second hypothesis is incorrect, that emotional or perceptual shifts caused by externalized PC emotional state may not cause a shift in player action. 5.1 GROUNDED THEORY Data collected from participant interviews was evaluated using grounded theory. Grounded theory is a method of data evaluation that precedes theory formation in process. Testing is conducted and data collected, then this data is evaluated and from the data theories are developed. This method does not strive to find absolute answers but to create a conceptualization of examined circumstances. Grounded theory data collection, related to human factors, primarily concerns itself with eliciting candid descriptions [19]. This test endeavored to examine participants’ decision making and perception. Based on grounded theory we developed our questions to elicit candid descriptions of participants’ experience. Though this project was designed based on hypotheses, grounded theory fits our data evaluation needs. The hypotheses were generated based on evaluating broad connections between player and PC. We wanted to analyze the impact that externalized PC emotions had on the Player. Though we hypothesized that externalized PC emotion would have an effect on the player we did not know exactly what this effect would be or how it would present itself. Our hypotheses outlined no specific form the interaction between player and PC would take. To evaluate these connections, and any trends therein, grounded theory was applied. Our first step in developing theory was data collection. The first step in developing this method was to broadly observe the data looking for possible trends. Based on these trends, systems were developed for evaluating and categorizing the video narration and standardized question responses.

25

5.2 QUALITATIVE DATA After each video narration was complete participants were asked a series of standardized questions. Two of these questions tasked participants with describing the PC. The first question asked participants to describe how he or she felt about the PC’s capabilities as a character to survive within the environment. The second question tasked participants to describe how they felt about the PC as a character. It was understood the character had no explicit backstory, but participants were asked to describe any impressions that had formed during play. These two questions were analyzed to see if participants’ perceptions of the PC and the game were in line with the intended goals of expression. Using a grounded theory method categorizations were developed after data was collected, based on trends noted in the data. The responses were evaluated and split into three broad categories: positive, negative, or critique. Positive answers contained descriptions of the PC’s character or capabilities as confident, successful, or competent. Conversely negative answers contained descriptions of the PC’s character or capabilities as timid, unsuccessful, or incompetent. Comments that addressed environmental factors in a similarly positive or negative manner were categorized as such. Answers categorized as critique offered analysis of the game system with no relation to the PC that could be categorized as positive or negative. Certain responses contained answers that were either neutral or a mix of positive and negative comments. An attempt was made to consolidate these kinds of responses into their own categories. However these comments were not consistent enough to consolidate into a fourth category, and therefore were not included in this evaluation. Only polarized comments that easily fit into the positive or negative were categorized as such. The majority of answers fit a categorization of either positive, negative, or critique.

26

The following answer given by a Fear Group participant was categorized as a critique, “his regular walk was weird, but fine”. This answer primarily commented on perceived quality in animation. There is both weak positive and negative language, but neither is strong enough to give an impression of the character other than a functional evaluation of his aesthetics. Another critique statement given by a Confident Group participants follows, “I was confused about his manacle.” This comment offered no positive or negative evaluation of the PC, instead articulating an issue with an aesthetic choice. All answers categorized as critique offered answers evaluating the game system, not a feeling or perception concerning the PC, his expression, or the environment. Answers that offered more telling information, based on a participants perceptions of the PC’s capabilities, character, or given scenario were categorized as either positive or negative. The following negative statement from a Fear Group participant offers an evaluation of the game as a whole, “[The PC’s capabilities were]…ok. The game conveyed a sense of complete hopelessness”. As you can see, this quote describes a strong negative feeling towards the game environment as a whole, and a relatively neutral evaluation of the player. The participant’s strong negative evaluation of the scenario in relationship to personal experience led this statement to be categorized as negative. The following positive statement from a Confident Group participant addresses a character oriented evaluation of the PC, “He looks the part; he looks like a zombie killer. He has been through a lot”. This quote expresses that the PC seemed appropriate in the situation, and he was skilled in handling that situation. This positions the PC positively in relationship to the given scenario and was categorized as positive.

27

Certain answers make it obvious that the emotional expression of the PC was lost on some participants. However, based on the ratio of positive to negative answers within each group it is suggestive that this was not the norm, and in fact the majority of participants had the intended impression of the PC. Within Confident Group, 12 answers were categorized as positive, 6 as negative and 19 as critiques. Within Fear Group, 5 answers were categorized as positive, 19 as negative, and 10 as critique. There is a tendency toward stronger positive responses from Confident Group and weaker positive responses from Fear group. Similarly, the negative responses tend to be more poignant from Fear Group and less poignant from Confident Group, generally based on a perceived limp in the Confident Group PC. Based on Confident Group responses concerning a perceived limp in the PC, it is possible aesthetic nonverbal communication issues affected the ratio of positive to negative responses in Confident Group. Without this issue, it is possible that this ratio would have been a closer inverse of Fear Group. There are two things these results suggest. The first is that, if only in a broad sense of negative or positive impression, our participants’ perception of the PC was more in line with what was intended than not. It also shows that in many cases there was a diverging opinion of the PC’s capabilities within the game, even though both versions of the PC were functionally identical and only aesthetically different. This is in contrast to our quantitative results which suggest that performance was relatively consistent across our testing population. 5.3 PC CHARACTERIZATION

In Dead End, no explicit backstory or characterization was created for the PC other than his aesthetics, two separate animations sets, and the environment-based horror situation.

28

Participants were left on their own to fill in whatever backstory or role he or she desired for the PC. Based on Hefner’s description of the importance of role to player identification, slightly in contrast to our results, this scenario created the best environment for our testing [20]. Hefner explains that while playing a game a player must be willing to fill given roles to properly identify with a character. He also asserts that without a given role players will be less likely to identify with a PC [20]. In our test results, without a strictly implied role, many participants reported experiencing high levels of immersion while playing Dead End. Many participants were able to articulate personality and situational factors they constructed concerning the PC. Some participants articulated backstories they had come up with to drive their motivation to play. One Confident Group participant explained: “I like him, I like his struggle. I like the design choices, his design gives me enough story. He just needs to get out and I can get behind that.” One Fear Group participant explained: “He seems like he’s not trained for this, he’s just a normal dude. Like a Saw [movie] victim”. Another Fear Group participant describes his backstory, “in reality I’m just some guy dropped into hell or on a bad trip. He got injured before the game started, he is having a rough time.”

Hefner suggests some sort of

characterization or backstory is necessary for a player to become immersed and self-aware through a PC. Our results however, seem to suggest players are not incapable of providing their own characterization and backstory, given minimal cues. This may not have been the case without the strong visual aesthetics and horror environment of Dead End. Players need to be willing to fill the role of a PC to perceive that character through a sense of self. It is explained that if a player objects to the role they are playing, this identification may be mitigated as a result [20]. In our game system, when a participant created

29

his or her own characterization and backstory with no predetermined role, design decisions were less likely to mitigate PC identification. However, the majority of participants created a negative role they would not normally be willing to fill.

Within Fear Group, the

characterizations of the PC presented negative evaluations of the PC and the given situation that would not normally be considered a desirable role. Within the results of the two standardized questions concerning the PC, we can see participants’ evaluations of the role presented within the game. Within negative answers from Fear Group, we see a consistent use of certain similar terms (“terrorized,” “weak,” “feeble,” and “vulnerable”) and evaluations of the PC’s mortality within the given scenario. Confident Group negative answers invariably used the words “limp”, “limping”, or “injured”. This can be attributed to a technical issue with the way animations were implemented in the game system that created a perceived limp. An examination of negative answers reveals that while Fear Group comments characterize the PC in a negative light based on his relationship to the given scenario, Confident Group participants do not relate the PC’s perceived injury to the scene. The answers given by Fear Group offer a connection between the PC and the horror situation while Confident Group does not explicitly reference the given scenario. These results seem to show it is possible PC emotional state affected more than the participants’ perception of the PC. The evaluation of the PC’s place within the given scenario differed between each group, even within answers that were categorized as negative. PC characterization was not seen across the entire participant population. Out of 40 participants, 3 explicitly complained about the lack of story, and 8 others did not offer any suggestions of a possible backstory. Of these 11 participants, 7 either reported having a very goal completion oriented strategy, or were designated as such based on review of his or her

30

play through video. Of these 11 participants, 8 critiqued the PC’s design or animation as inappropriate or distracting. 9 out of these 11 participants were from Confident Group. 29 out of 40 participants offered some kind of insight into his or her perceptions of the PC’s role. These responses range from simplistic, such as this answer from Fear Group “he seemed almost cowardly”, to more complex, such as this answer from C Group “he’s wearing a collar, he’s trying to escape. The improvised weapon made me feel like a survivor, his animations were fine”. Jan Van Looy et al. examine the causes of personal identification in Massively Multiplayer Online Role-Playing Games (MMORPGs). This study’s findings suggest that in MMORPGS, players that strongly identify with their avatar are more likely to generate personal backstories for this PC [21]. If this consistency in MMORPGs holds true in our game, it suggests that these participants were identifying with the PC, a necessary factor for effective perception and possible embodiment of PC animation [1]. However the results of our questioning also seem to show that the given emotional expression, specifically the fearful PC, elicited these backstories from participants. There was significantly more negatively categorized characterization of the PC in Fear Group than negative or positive categorizations of the PC in Confident Group. The majority of the most poignant statements concerning the PC are from Fear Group negative answers. This result suggests that Fear Group participants had more consistent identification with the PC, and that this relationship was stronger. This may be partly due to the horror scenario of the game. Participants seem to be willing to create and inhabit a more in-depth characterization of the PC when that characterization is in line with the given scenario. In this case, the most appropriate and acceptable characterization and role seems to be a fearful role, even though this would

31

normally be an undesirable role. Within Confident Group, we see less depth in both positive and negative characterizations of the PC. This suggests that the character’s animations being unaligned with the tone of the scenario mitigated participants desire or ability to create characterization with as much consistency and depth as Fear Group. When perceived as intended, confident animations may have mitigated Confident Group’s ability or desire to create characterizations of the player, even though it seems this was the preferred role to fill. In fact the only negative characterizations from Confident Group specifically concerned an unintended evaluation of the confident animation in a negative light. Our data suggests the animation state of the PC, without giving an explicit role, impacted the kinds of roles participants created for the PC. The results suggest that character animation and aesthetics is enough for players to either construct their own role, or possibly represents a strong enough character feature to constitute a given role without other explicit characterization.

The latter concept is supported by the lack of Confident Group

characterizations compared to Fear Group, and the lower number of positive characterizations within the entire testing population. This also suggests that, in contrast to observed character behavior within some blockbuster hits in the genre, confident animations in a horror experience may actually mitigate a player’s willingness or ability to identify with the PC. As established in the literature review, survival horror is a genre that only achieves its purpose when players are immersed and identifying with the PC. This result is in direct contrast to the current direction of a large portion of the genre.

5.4 NARRATION EVAULUATION

32

The contents of our unstructured play through narrations were analyzed based on the findings in our evaluation of participant perception of the PC. These narrations were split into separate statements for analysis. The division of these statements was based on descriptions of a particular moment, action, thought, or observation. Each statement was evaluated as positive, negative, critique or other and a tally was generated of these results. Unlike the standardized PC questions, the majority of narration statements are non-characterizing, mostly pertaining to statements describing participants’ strategies and feelings during play. There was a wider variety of responses within narrations compared to the standardized PC questions. The method for categorizing statements as positive or negative needed to be broadened for narrative results. There was increased use of negative language and cautious tactics within the Fear Group and an increased use of positive language and aggressive or confident tactics in Confident Group. Based on these observations, a system was created for categorizing statements within participants’ narration as either positive or negative.

Many of these

parameters were generated based on consistencies we saw within each group. For example, a trend toward creative strategizing was noted, specifically in relation to navigation, within Confident Group. These kinds of behaviors were categorized as positive. Within Fear Group, a trend appeared towards participants describing actions taken to maintain the PC’s health and to keep him alive. Statements describing this kind of behavior were seen in both groups, but were seen more frequently in Fear Group and were categorized as negative. Some parameters were generated based on their intrinsic relationship with either fearful or confident play. Statements about being cautious and relying on stealth were seen in both groups, but categorized as a negative or fearful action. Similarly a number of participants in both groups

33

described having a “gung-ho” attitude towards playing the game. These participants would start off running straight into the maze with reckless abandon. Statements using this kind of language and describing this kind of activity were categorized as positive or confident. The full list of Categorization Parameters can be seen in Table 1. Table 1: Categorization Parameters

Categorization parameters Negative

Checking around corners Moving slowly Cautious descriptors on actions Strategies developed to maintain health Specific mentions of being afraid, “freaked out” or panicked Revaluations of strategy towards caution Complaints about the setting Avoiding confrontation with multiple enemies Positive Starting without caution Direct strategies, related to forward movement without caution Exploratory strategies Not avoiding confrontation with multiples Not caring about game consequences Strategies related to navigation Expressing being comfortable with combat or other game features Expressing being calm about combat mechanics Stating he or she felt there was no need to crouch Going in without thinking about strategy Being confident or not scared

For each instance of the listed statements, a tally of positive and negative responses between participants was created. Within Confident Group there were 16 negative statements and 29 positive statements. Within Fear Group there were 50 negative statements and 15 positive statements. While every participant did not fall in line with this trend, our results

34

suggest that the majority did.

This statistically significant difference suggests that our

participants may have perceived his or her game experience differently, strategized differently or at least communicated his or her experience differently. This is a significant find specifically in its relation to the negative results of our quantitative data. The narration results show significant differences in how participants evaluated their play experience. The only variable element between the games these two groups played was the passive aesthetic aspect of the PC’s animation sets. The qualitative results seem to suggest that even with this difference in perception, PC animation state elicited no documented effect on player performance or play style. This would suggest that the correlation between PC state and player action may not be as strong as originally postulated or that the shift in perception suggests. Part of the formative theory was that a strong thematic environment would enhance player immersion. Design choices were based on research showing that immersion was an important aspect of player-PC identification. It seems possible that certain means of creating immersion can hinder the effect of the PC on the player based on similar performance statistics between both groups. The extreme aesthetic elements of the survival horror genre create a sense of presence and embody the player within the game world. These features may overshadow or decrease the importance of the medium through which the player becomes embodied in the game world. Additionally, certain aspects of gameplay may not have functioned as desired. For example, many participants had unexpected difficulties with the maze. Based on our results there is a strong indication that PC emotional state can drive a player to spontaneously form impressions of and backstories for a character. Additionally, PC emotion seems to filter participants’ interpretations of a game, their performance, and the

35

difficulties faced, without creating any real impact in these areas. This is a powerful tool for game developers to consider in any genre. These results suggest that this passive game element can create an active effect on the player, which could allow for crafting game “content” by directing the player’s own creativity. Recently there has been a rise in popularity of sandbox titles that leverage user generated content such as Minecraft, Little Big Planet, and Terraria. This emerging genre shows the trend in design philosophy toward creating systems for users to create their own experiences instead of explicit content generation. It is possible that the player’s creativity can be leveraged for more than physical content generation, but the generation of story, characterization, atmosphere and other abstract aesthetic elements. This may be accomplished using subtle manipulations of emotional states in the player to enforce desired emotions. Instead of requiring complicated systems to facilitate player creativity this tool merely leverages the power of a player’s mind to generate perceptual game content. As previously outlined, The PC in Dead End has no explicit characterization. Our test results show that in many cases the PC’s expressive animations were enough to shift and define participant’s perception of the PC as a character as well as their perception of the game as a whole. Using a similar blank canvas approach to the PC, it may be possible for other games to shape complex characterizations and perceptions of game events, merely using emotive PC animations. This has the potential to create experiences that represent more than the sum of their parts. It allows players to freely project and customize their personal story, but gives developers the tools to guide this experience. 5.6 QUANTITATIVE DATA

36

The metrics coded from our 40 participant test study revealed no statistically significant differences in the results between our two test groups, which can be seen in Table 2. Many of the statistics between these two groups were nearly identical. Accuracy only differed by 1%, with 41% accuracy in Fear Group and 42% in Confident Group. Between all participants, the lowest accuracy was 28% and the highest was 58%, with 36 out of 40 participants falling between 30% and 50% accuracy. Other statistics varied more between individual participants, but the average of each statistic compared between groups showed no statistically significant difference. The consistency seen between accuracy in all participants was in direct opposition to expectations.

Accuracy was expected to be a useful statistic in evaluating player

performance. It was expected for performance to vary to a much larger degree between individual participants regardless of any correlations between action and the PC’s expression. This result seems to suggest that while emotional state in PC animation seemed to influence perception, language choice, and descriptions of the game, it had little or no effect on performance, choices within the game, or play style within this game system. Trends seen in our quantitative analysis suggest players perceived the PC’s capabilities differently. The most telling example of this is Fear Group’s evaluation of the Confident Groups PC. After each interview was complete participants where shown the other version of the PC and asked to give their impressions. The majority of Fear Group participants evaluated the run speed of the confident PC as faster than the PC they had used. Similar trends can be seen in the question targeting participants’ perception of the PC’s capabilities. This shows that participants’ perception of the PC and his capabilities were grounded by the animation state. If a player’s perception of the PC’s capabilities are filtered by externalized emotions it stands to reason that his or her choices when controlling this PC may be similarly affected.

37

In the literature review the functional feedback loop between player and PC that instructs how the player expects to interact with the game world is outlined. It is also shown that this relationship drives in game choices. As an example, how quickly the player expects to walk is driven by how fast the player walks. Testing results suggest externalized PC emotions can interact with this relationship to the extent that it effects the perception of run speed. From these concepts the conclusion can be drawn that a perceptual shift in this feedback loop can affect interpretation of in game actions, and affect their subsequent use. The results from the quantitative analysis do not illustrate these connections. Based on strong logic and results based correlation it seems possible that external factors normalized the quantitative results. Table 2: Averaged Coded Metrics

C group

F group

Attack Presses

232.2

304.7

PC attacks

96.25

122.7

attack/press

2.41

2.5

PC Hits Dealt

45

49.3

Accuracy

42%

41%

PC Hits Taken

102.7

125.2

PC Deaths

3.15

3.85

Enemies Killed

12

15

Stationary Time

113.8

108.3

Stationary ratio

14%

13%

38

Sneak Time

54.8

43

Sneak Ratio

7%

5%

Walk Time

233.3

267.9

Walk ratio

28%

32%

Run Time

343.9

384.1

Run Ratio

49%

48%

Total Time

762.8

832.7

5.7 DATA COMPARISON

The two-pronged approach to generating data from user tests fed directly into the attempt to gather information on two hypotheses. Participant interviews directly targeted the first hypothesis: externalized PC emotions would elicit an emotional or perceptual effect on the PC. The goal of these interviews was to extract as clear a view of a participant’s perception and feelings about the game as possible. Metrics coded from play sessions directly targeted the second hypothesis, that any emotional or perceptual shift would have an impact on a player’s performance or choices in game. Based on the results of this test, it seems that the documented differences in participant perception created no significant effect on the performance and play style metrics coded. To further evaluate these metrics, 10 participants’ results from each group were isolated. These participants, based on both analysis of standardized questions and narrative responses, had the most accurate interpretation of the PC and narration most in line with his or her animation group. The

39

average of coded metrics within these 20 participants also produced no statistically significant differences between Fear Group and Confident Group. In fact, the majority of shifts brought slightly dissonant metrics even closer to each other between the two groups. This suggests that participants’ play styles and performance may have been more related to how immersed they felt in the game. If this is true, then it is possible that participant play style was driven by the survival horror setting or some other consistent feature of the game. The use of varying animation states within Dead End was constructed based on creating an appropriate testing system. These animations were completely static within each group, and play sessions extremely short. Even so, the effect of these animations on participant perception, in the form of internal character development and evaluation of scenario, was pronounced. We have suggested that this difference in perception, based on our research, would logically affect play. Based on critique feedback, certain areas of the game that may have had a normalizing effect on coded metrics were identified. Participants complained about becoming lost within the maze environment of Dead End. Some participants spent more time than intended lost within the maze, becoming frustrated. The more frustrated participants became, the more likely he or she would begin running around the maze looking for an exit. In many of these cases, before becoming frustrated, participants had approached the game in a different way. This behavior was seem in both testing groups and could have had a normalizing effect on the averages of movement type use. The majority of participants approached combat using similar tactics. Most participants approached combat directly at first. Finding this strategy effective most participants never reevaluated or developed strategies. This was unintended, and possibly led to the strong similarities in accuracy across all participants.

40

6. CONCLUSION The testing results provide two interesting evaluations of the effect of emotion in PC animation states, one based on the quantitative data and one based on the qualitative data. Qualitative analysis examined our first hypothesis, that the externalized emotional state of the PC would have an impact on the player. interviews.

This was accomplished through participant

Quantitative analysis examined the second hypothesis, that shifts in player

emotional state and perception would have an effect on performance and action in the game. This was analyzed using qualitative results and revealed positive correlations between PC expression and participants’ description of their play through. Analysis of this data suggests that a participant’s perception of his or her experience was affected by the PC’s animation state. The method for evaluating this data was based on trends in strategy and language seen in either testing group using a grounded theory method. The quantitative data showed that there was no statistically significant difference between metrics coded from both groups related to performance and play style. The expectation was that the interaction between animation type and player action would be strong enough to create an effect on player action. This was not the case within the simulation. There were strong correlations in the qualitative analysis between animation state and participants’ evaluation of the PC’s capabilities. This result, in context with research concerning the relationship between player and PC, suggests that this perceptual shift can affect in game action. It is possible that certain aspects on the game may have played a more substantial role in forming player action than the PC. Participants relatively consistently had

41

issues navigating the maze, some never leaving the starting area. As participants became more frustrated, they became more likely to run around the map looking for new content. The combat system may have not had the depth necessary to encourage or even allow for varying styles of play. This seemed to produce extreme similarities in how participants approached Combat, which could account for not just consistencies between the two groups, but between all participants. Participants consistently complained about the speed of the crouch movement type. Thirty six out of forty participants cited run as his or her preferred movement type. This could have led to the consistency between both groups of high use of the run mechanic and low use of the crouch mechanic. If a future test shows that emotional state in PC elicits an effect on player action, there are important implications to game design theory. One of the most intriguing is the possibility of using aesthetics as a balancing tool in games. Similar to the director system in the Left 4 Dead series, the emotional state of the PC could be tweaked to either give players an encouraging boost, or give them an indication of when they may not be capable of dealing with a presented game scenario. This concept could be used to add a different level of difficulty to game scenarios, or give players a sense of empowerment when needed. The possibility that a player’s perception of in-game events can be filtered through the emotional state of the PC without affecting performance is also a powerful design tool. It is possible that emotional shifts in PC animation states can be used to instill a whole range of emotion within the player. We established earlier that games do leverage expression in animation to elicit emotional responses. Though PC expression is already used to influence players, there is a significant difference between creating a thoroughly understood design science and designing on a whim and hoping for the best. The correlations presented in the

42

findings represent the first step in the process of developing a new layer of interactive design for video games. To fully understand the relationship between player and PC, and this subject’s effect on game design, it is necessary to make further studies. Based on the data, we have shown that a game with mechanics polarized between a positive and negative emotional state, can use PC animations to drive perception of in-game events in a positive or negative direction. At this point it is not known whether or not the game needs to be constructed in this polarized fashion to elicit this kind of effect. It is also inconclusive whether or not externalized PC emotions illicit an effect on player performance or action. It is strongly recommended that the effects of externalized PC emotional states on player perception and performance be examined further.

43

7. LIST OF REFERENCES [1] P. Bayliss, "Beings in the Game-World: Characters, Avatars, and Players," in Paper presented at the proceedings of the 4th Australian Conference on Interactive Entertainment, Melbourne, Australia, 2007. [2] P. Lankoski, "Player Character Engagement in Computer Games," Games and Culture, vol. 6, no. 4, pp. 291-311. [3] A. Beck, B. Stevens and K. A. Bard, "Emotional Body Language Displayed by Artificial Agents," ACM Trans. interact. Intel. Syst.,, vol. 2, no. 1, pp. 1-29. [4] M. Mikami, Director, Resident Evil 4. [Film]. Capcom, 2005. [5] NaughtyDog, Director, Uncharted 2: Among Theives. [Film]. Sony Computer Entertainment, 2009. [6] P. Molyneux, Director, Black and White 2. [Film]. Lionhead Studio, 2005. [7] K. Jørgensen, ""I'm overburdened!" An Empiracle Study of the Player, the Avatar, and the Gameworld.," 2009. [8] T. S. Studios, Director, The Sims 3. [Film]. Electronic Arts, 2008. [9] R. Berhaupt, A. Boldt, T. Mirlacher, d. Wilfinger and M. Tschelgi, "Using Emotion in Games: Emotional Flowers," in International Conference on Advances in Computer Entertainment Technology, Salzburg, Australia, 2007. [10] E. J. Moody and D. N. Mcintosh, "Mimicry of Dynamic Emotional and Motor-Only Stimuli," Socila Psychological and Personality Science, vol. 2, no. 6, pp. 679-686, 2011. [11] P. Dourish, Where the Action is: The Foundation of Embodied Interaction, MIT Press, 2004. [12] M. Hitchens, A. Drachen and D. Richards, "An Investigation of Player to Player Character Identification via Persona Pronouns," in The Proceedings of the 8th Australian Conference on Interactive Entertainment: Playing the System, Aukland, New Zealand, 2012. [13] J. T. Cacioppo and W. L. Gardner, "Emotion.," Annual Reveiw of Psychology, vol. 50, pp. 191214, 1999.

44

[14] N. Rajava, M. Saminen, J. Holopainen, T. Saari and J. Laarni, "Emotional Responce Patterns and Sence of Pressence During Vidoe Games: Potential Criterion Variables for Game Design," in The Proceedings of the Third Nordic Confererence on Human Computer Interaction, Tampere, Finland, 2004. [15] M. Meehan, B. Insko, M. Whitton and F. P. Brooks, "Phsychological Measures of Presence in Stressful Virtual Environments," ACM Trans. Graph., vol. 21, no. 3, pp. 645-652, 2002. [16] B. Perron, "Coming to Play at Frightening Yourself: Welcome to the World of Horror Video Games," in The Aesthetics of Play, Bergen, Norway, 2005. [17] G. Salvendy, Handbook of Human Factors and Ergonomics, Hoboken, New Jersey: John Wiley & Sons, 2012. [18] M. V. Aponte and Natkin, Difficulty in Videogames: An Experimental Validation of a Formal Definition., Lisbon: Paper presented a tthe proceedings of the 8th international conference on advances in computer entertainment technology, 2011. [19] B. G. Glaser and L. A. Strauss, The Discovery of Grounded Theory: Strategies of Qualitative Research, Piscataway, New Jersey: Seventh Paperback Printing, 1967. [20] D. Hefner, C. Klimmt and P. Vorderer, Identification with the Player Character as Determinant of Video Game Enjoyment, Laxenburg: IFIP, 2007. [21] J. V. Looy, C. Courtois and M. D. Vocht, Player Identificaiton in Online Games: Validation of a Scale for Measuring Identificaiton in MMORPGs, vol. Fear and Games 2010, Leuven: ACM, 2010. [22] N. Ducheneaut, M.-H. ". Wen, N. Yee and G. Wadley, Body and Mind: A Study of Avatar Personalization in Three Virtual Worlds, vol. New Media Experiences 1, Bosnten, MA.: CHI, 2009. [23] L. Klastrup, Death Matters: Understanding Gameworld Experiences, vol. ACE 06, California, CA.: ACE, 2006. [24] D. Milam and M. S. E. Nasr, Analysis of Level Design 'Push & Pull' within 21 games, Monterey, CA.: ACM, 2010. [25] J. V. Looy, Games and Self-Imaging, a Comparative Media Perspective, Vols. Vol 3, No 1 (2009), Cambridge, MA.: Eludamos. Journal for Computer Game Culture, 2009. [26] K. Nakazawa, Director, Silent Hill 3. [Film]. Japan: Konami Computer Entertainment Tokyo.

45

[27] H. Miyazaki, Director, Dark Souls. [Film]. Japan: From Software, 2011. [28] S. A. Barab, M. Gresalfi and A. Ingram-Goble, "Transformative Play," Educational Researcher, vol. 39, no. 7, pp. 525-536, 2010. [29] M. D. Dickey, "Engaging by Design: How Engagement Stategies in popular Computer and Video Games Can Inform Instructional Design," Educational Technology and Developement, vol. 53, no. 2, pp. 67-83, 2005. [30] D. Milam and M. S. E. Nasr, "Analysis of Level Design 'Push & Pull' withing 21 games," ACM, Monerey, CA., 2010. [31] D. Punday, "Ivolvement, Interruption, and Inveitability: Melancholy as an Aesthetic Principle in Game Narratives," SubStance, vol. 33, no. 3, pp. 80-107, 2004. [32] B. Rolfe, C. M. Jones and H. Wallace, "Designing Dramatic Play: Story and Games Structure," in the 24th BCS Interaction Specialist Group Conference, Dundee, United Kingdoms, 2010. [33] N. Zagalo, A. Torres and V. Branco, "Emotional Spectrum Developed by Virtual Storytelling," in the Third International Conference on Virtual Storytelling: Using Virtual Reaity Technologies for Storytelling, Strasburg, France, 2005.

46

8. APPENDIX 8.1 POSITIVE AND NEGATIVE POST VIDEO ANSWERS POSITIVE •

“Potent” C group

•

“He looks like he has been through a lot, he has some experience” C group

•

“Seems fine” C group

•

“Survivor of a zombie apocalypse” C group

•

“I like him, I like his struggle. I like the design choices, his design gives me enough story. He just needs to get out and I can get behind that”. C group

•

“Easy enough, see a demon, hit it until it falls down” C group

•

“He looks the part, looks like a zombie killer, he has been through a lot” C group

•

“Hardcore badass, he could deal with punishment” C group

•

“He was cool looking, I had a sense that he should be there, vandal” C group

•

“pickaxe helps, he’s a regular guy” F group

•

“He’s wearing a collar, he’s trying to escape. The improvised weapon made me feel like a survivor, his animations were fine” F group

•

“pretty good” F group

•

“He seemed good at fighting them off” F group

•

“Fine, I could kill anything unless there were 3” F group

NEGATIVE •

“He seems a little off balance and clumsy, does not seem prepared” C group

47

•

“Not great, he would be better if he was a monster, I was scared” C group

•

“Post-apocalyptic, he’s ambling and unsteady, I am not really sure how I feel about him, he’s from some other universe” C group

•

“blood covered limping and hitting people” C group

•

“He was limping, and had some kind of weird walk. I saw this and thought I would only last a few hits.” C group

•

“He seems like he’s not trained for this, he’s just a normal dude. Like a saw (movie) victim” F group

•

“He could live or die, he’s been tortured, and this is his last chance. He is confused or scared” F group

•

“He’s wearing ratty jeans and a tee-shirt and a dog collar. Running away, breaking out.” F group

•

“He is as terrified as I am. His posture does that, you put your own emotions in but…” F group

•

“He is realistic, you can kill 3 but you will die” F group

•

“I thought the walking motion was weird at first, but because of the collar I can tell he has been tortured” F group

•

“My backstory for this game is that in reality I’m just some guy dropped into hell or on a bad trip. He got injured before the game started, he is having a rough time” F group

•

“He looks like someone that escaped from imprisonment, woefully underequipped. He is scared shitless and vulnerable” F group

•

“He seemed almost cowardly” F group

48

•

“It was hard to deal with enemies from multiple directions” F group

•

“He makes me feel uncomfortable. He is covering himself, he seems scared and gives me the same feeling” F group

•

“He shouldn’t be able to swing his pickaxe so fast. Walk cycle makes him seem like a dude stuck in maze” F group

•

“His movement was creepy and weird, I liked it, it enhanced my fear” F group

•

“He was relatively easily killed” F group

•

“He felt weak, he would die in three hits” F group

•

“He felt feeble, if I was him I would die” F group

•

“Not weak, but unprepared” F group