Stimulating Psychological Attachments in Narrative Games

Characterization in Games

The role of characterization can vary, and scholars of traditional film and literary studies have long debated the role of characters. In the narratology literature, an argument exists between psychological versus apsychological narrative (Bremond, 1970; Bruner, 1986; Nelson, 1996; O’Connor, 1969; Propp, 1928). At the extremes, games treat characters as simply functions for plot, and psychological essences are avoided.

The views of the Formalists and (some) structuralists resemble Aristotle’s in a striking way. They too argue that characters are products of plots, that their status is “functional,” that they are, in short, participants or actants rather than personages, that it is erroneous to consider them as real beings. Narrative theory, they say, must avoid psychological essences; aspects of character can only be “functions.” They wish to analyse only what characters do in a story, not what they are—that is “are” by some outside psychological or moral measure. Further, they maintain that the “spheres of action” in which a character moves are “comparatively small in number, typical and classable.” (Chatman, 1978, p. 111)

Chatman (1978) explores both sides of this debate in some depth, before giving his own conclusion that the issue of priority or dominance of character or plot is not meaningful. However, we found the issue of character versus plot-oriented design to be meaningful to narrative games and directly relevant to players’ experiences. In some games, characters are simply functions of what they do—for example, a stereotypical first-person shooter where the relationships between players’ avatars and non-player characters are entirely focused on what players do, in this case, combat. We found in player’s critiques that such relationships do not fully model the diversity of human experience, and that meaningful relationships, born out of game play, strengthen player’s engagement levels with a game.

As computer games have developed, the role of characterization has evolved. Characterization and character relationships become more complex in an interactive medium, partly because the number of relationships multiplies. We need to consider three kinds of relationships, between the following:

Moreover, the first kind of relationship, between Player and Avatar, deconstructs into two further kinds of relationship. First, a relationship between the player and the type of avatar that is already pre-defined and engages in behavior independent of what the player does (this usually happens in cut scenes, where the game takes control away from the player). Second, a relationship between the player and the type of avatar that only behaves as the player dictates (with few or no cut scenes). In this article, we look at the first and second types of relationships described above, but focus mainly on the relationship between player and non-player character.

Exploring Engagement

What constitutes engagement needs further examination. Benyon, Turner, and Turner (2005) describe engagement as being

concerned with all the qualities of an experience that really pull people in—whether this is a sense of immersion that one feels when reading a good book, or a challenge one feels when playing a good game, or the fascinating unfolding of a radio drama. (p. 61)

We presented our own definition of engagement at the beginning of this article, but because alternative terms exist for similar or overlapping concepts, it is worthwhile to explore them briefly here. Examples of alternative terms are interest, involvement, immersion (Mania & Chalmers, 2001; Singer & Witmer, 1999; Whitton, 2011), presence (Slater & Wilbur, 1995; Witmer & Singer, 1998), flow (Csikszentmihalyi, 1992), and psychological absorption (Brockmyer et al., 2009).

Researchers often use the term immersion to describe and quantify players’ levels of engagement in the game world. “Immersion is the idea of absorption in and concentration on an activity, losing track of time and loss of self-consciousness” (Whitton, 2011, p. 604). Similarly, Singer and Witmer (1999) describe immersion as “the psychological state characterized by perceiving oneself to be enveloped by, included in, and interacting with an environment that contains a continuous stream of stimuli and experiences” (p. 566). Mania and Chalmers (2001) propose an alternative viewpoint, in their view of technological immersion “as a quantifiable description of technology, mainly determined by the extent to which visual displays are (1) inclusive, (2) extensive, (3) surrounding, and (4) vivid” (p. 248). It is likely that most players experience some form of immersion when playing games, where their focus of attention is engulfed by the game world, although still remaining acutely aware of their surroundings outside the game world.

We perceive presence as

the subjective experience of being in one place or environment, even when one is physically situated in another ... a normal awareness phenomenon that requires directed attention and is based in the interaction between sensory stimulation, environmental factors that encourage involvement and enable immersion, and internal tendencies to become involved. (Witmer & Singer, 1998, p. 225)

Defined also by Slater and Wilbur (1995), as a “state of consciousness” (p. 604), our preferred description of presence is as the psychological feeling when a player becomes immersed in another world.

Game creators aim to design games to immerse players and engage them in the experience, to not disrupt them from their flow of activity. This “satisfying, exhilarating feeling of creative accomplishment and heightened functioning” (Csikszentmihalyi, 1975, p. xiii) is known as flow. Csikszentmihalyi (1992) describes flow theory as

the state in which people are so involved in an activity that nothing else seems to matter; the experience itself is so enjoyable that people will do it even at great cost, for the sheer sake of doing it. (p. 4)

Human beings, Csikszentmihalyi argues, are activated completely when we feel fully alive, full of potential and purpose, achieving the greatest form of happiness available: intense, optimistic engagement with the world around us (McGonigal, 2011, p. 36). Indeed, in later work, Moneta and Csikszentmihalyi (1999) argue that

the quality of daily subjective experience is a function of two variables (1) the skills which the person s/he feels they have in relation to the activity, and (2) the perceived challenges of the activity. (p. 606)

Both variables are key to a player’s level of engagement when playing games.

To conclude our review of the definition of engagement, the psychological and emotional connections players make to a game are critical to their enjoyment. Psychological absorption is the term used to describe total attention and full committed engagement in the present experience (Irwin, 1999).

In contrast to immersion and presence, and in common with flow, being in a state of psychological absorption induces an altered state of consciousness. (Brockmyer et al., 2009, p. 625)

From a gaming perspective, players may become detached from the world around them, when transfixed in this altered state. The psychological absorption and emotional connection that players can experience when playing games is a factor that game creators aspire to invoke.

Therefore, terms such as interest, immersion, presence, flow, or psychological absorption during game play are applicable in their contribution in defining engagement. They are part of engagement and help to explain what we mean by it. In our study, we use a player’s engagement, or an absence of it, as an index to point to causal factors, both in people’s consciousness and in the games, which produce or reduce appeal. Engagement is the end effect of other factors or conditions. In our study, we did not attempt to deconstruct this end effect, unless game players spontaneously offered such information. Our engagement index is thus relatively simple and inclusive. We denote as markers of engagement (a) any expression of positive or negative affect in relation to game play, (b) the retention of interest to continue playing, or (c) a stated desire to repeat the experience.

Finally, although we focus on characterization in this article, computer games are highly complex entities comprised of many elements. Technological advancements have led to high definition graphics, beautifully engineered soundtracks, powerful game engines, and exquisitely designed game controls that have brought very appealing qualities to today’s games. These factors all contribute significantly to a player’s overall enjoyment of a computer game.

Research Design

Many elements affect players’ enjoyment of games, for example, graphics, sound, animation, environment, or difficulty level, among others. In this article, we focus specifically on characterization. Our study aimed to extract the reasons why players feel involved with and engage with game characters.

Data Collection

Over the course of two autumn periods (2011 and 2012), two separate groups of players consisting of 48 participants in total investigated the factors that actively engaged them in game play. Participants were asked to choose two computer games within the same genres to play, which could be online or off-line games, followed by a number of exercises. First, stopping intermittently, participants recorded their reactions, impressions, feelings, and reflections in a stream of consciousness diary form. We asked them to describe overall reactions as well as minute-by-minute reactions. Second, participants examined the transcripts of their writings to see what seemed to be generating positive and negative reactions. Third, they extracted from their diaries (a) the factors that affected their engagement with the selected games, and (b) any assessment criteria that they might draw from their game experiences. We had told them that the latter should, hypothetically, aid game designers. Finally, the players presented their findings during a debriefing session to their class members and their lecturer, outlining their findings from their game experiences. Participants filled in their diaries on an individual basis, though in the second year, we asked the students to collaborate in pairs to produce their evaluation criteria. Factors relating to a game’s graphical and aesthetical qualities, physical controls and interface design, audio, animation, or the performance of the game engine were discussed; however, this article is solely focused on the extraction of assessment criteria that pertain to the aspect of characterization in computer games.

Data Analysis

We took an inductive approach when analyzing the data, noting specific observations before proceeding to a likely generalization. The development of conclusions regarding which issues significantly affect players enjoyment was data-driven rather than theory-driven, given that conclusions were drawn from player’s diaries, and their analysis of those diaries in which they were asked to determine what issues were of importance to them, structured by their own language, categorizations, perceived associations and meanings.

The players therefore carried out some of the data analysis before the researchers carried out follow-on analysis using content analysis as the primary technique. We deemed content analysis the most suitable method, as it is an effective way to describe the content of a body of messages (Wimmer & Dominick, 1987). Defined by Holsti (1969) as “any technique for making inferences by objectively and systematically identifying specified characteristics of messages” (p. 14), content analysis allows researchers to reduce large amounts of text systematically into fewer content categories based on an explicit set of coding rules. It is defined by Lasswell (1948) as the study of “Who (says) What (to) Whom (in) What Channel (with) What Effect” (p. 37). We selected issues relevant to the players’ experiences of characterization in narrative games from the body of content and sorted these into categories before we could make inferences based on player input.

As the sample size was relatively small (48 players), some limitations arose on how the researchers could conduct content analysis. For example, in cases where players brought up a niche issue relating to characterization, they made relatively few comments on the issue raised. In such cases, the researchers reviewed similar data from a forum titled, Top 10 Gaming Moments (www.exisle.net), which included player’s critiques of characterization. When insufficient detail was evident in the research data, this forum was examined to see what it had to say about the particular issue and see how, or indeed if, it aligned itself with any category and supported or negated an emerging idea. Essentially, the forum acted as a methodological evaluation tool, either supporting or criticizing the (niche) issue raised by the players.

Furthermore, the first of the criteria discussed below, titled responsiveness, extends on earlier research by one of this article’s authors, into factors that engage players in adventure and role-play games.

The research design did not deliberately account for environmental factors that may affect engagement and immersion. Nevertheless, unless students were working too close to the submission deadline, the games were played in a relaxed environment, in the players’ home or in a friend’s home, in participants’ own time—which closely simulates their usual game-play environments. Moreover, we allowed players to choose games they had previously played. However, pausing to fill in the diary did change the usual game-play conditions. Although we consider that this should not impact heavily on responses such as general likeability, it is likely to have affected immersion to some degree.

Findings

Outlined below are two criteria focusing on what makes dynamic, engaging, and interactive characterization in narrative games, presented as assessment criteria, for use during formative evaluation of an emerging design. Evaluation criteria for use early in a game’s design and before developers expend substantive effort and resources, bring the evaluation process closer to the production process, which is advantageous for producers, given the high cost of game development. We envisage that this study could assist game creators in developing heterogeneous and interactive characters and a more engaging game.

Criterion 2: Stimulating Deep and Meaningful Relations

Imagine a game where you have a team of four non-player characters (NPCs) to help you in combat. Take one NPC; you fight alongside them, but you know nothing about them except their first name. They provide the same help in the team as other members. They do not stand out as markedly different from the other NPCs in the team. Whether they are the person minding your back, or not—it does not matter. If they die—it does not matter. Another NPC team member will step in. This scenario describes what we define as a shallow relationship. The NPC has functional utility for the player, but if an alternative NPC fulfills their function, the player feels no regret. The player does not deeply care about them. The game character is not meaningful to the player outside of its function. In contrast, a deep relationship is one in which the player cares for the characters, which can include antagonistic relationships. If the player feels antagonism or anger, they can still care.

Is it necessary that the players care? Is it important to form deep relationships with NPCs? Our answer is—no, it is not, but when players do, high levels of engagement and wonder at the experience are reported. Moreover, given the sometimes hundreds of NPCs in a game, it would not be possible to form deep relationships with them all.

So how does one build deep, as distinct from shallow, player to non-player character (NPC) relationships? How can we build meaningful associations, where the player does care? We propose a set of techniques, which we now present in summary. We have detailed and examined them in the sections that follow (see Table 1).

OPEN IN VIEWER

Table 1.

Stimulating Deep and Meaningful Relations: Techniques That Help Build Relationships Between Player and NPC.

Criterion 2.1	Whether the relationship with a particular NPC is different from other relationships
Criterion 2.2	Whether the interaction takes place over multiple channels, as distinct from one channel
Criterion 2.3	Whether the character is round or flat
Criterion 2.4	Whether the players’ relationships with NPCs are capable of change resulting from the players’ actions and behaviors
Criterion 2.5	Whether relationship-forming is graduated
Criterion 2.6	Design interactions to evoke psychological essences and emotions

Note. NPC = non-player character.

Criterion 2.1: Whether the relationship with a particular NPC is different from other relationships

Designing heterogeneous, varied characters, which differ from each other in appearance and behavior, helps to generate a feeling that the relationship with each may be different. It means that the NPCs cannot be easily interchanged. This is shown in the following quotations:

In Botanicula the user becomes emotionally attached to the characters who travel as a party. Each has their own strengths and weakness and you tend to love each one for their own quirky animations and abilities. In Botanicula your friends are constantly with you, it gives a joyful feeling ... (Group C)

Both Machinarium and Botanicula have different approaches to characters, that enable the player to empathize with them. In Botanicula, for instance, attention to detail, the colors, shapes and behaviors of every creature and the environment, create a very believable world. (Group C)

Criterion 2.2: Whether the interaction takes place over multiple channels, as distinct from one channel

The next technique arises out of opposition to the scenario painted at the beginning of this section. In that scenario, the interactions between the player and the non-player characters occur along one channel or dimension. This is a helper channel. We argue that interactions along additional channels to the one that serves the NPC’s function in the game will go further to form a relationship.

Interaction channels other than help offered or received, might, for example, include humor exchanged, going through a harrowing experience with the NPC, or having the background of the NPC, outside of their function in the game, become known. It can include fleshing out the NPC, seeing different sides of them, understanding how they think and act, learning about their philosophies and attitudes, and becoming aware of their emotional makeup through their responses to their surroundings, to others with whom they interact, and to events that occur, and through a history of such interactions.

In HALO, conflict and conversation is used quite well. Mostly using conversation, the player is given a fairly good idea of the main characters personalities; this is effective because it makes their deaths seem more tragic rather than just another event within the game. In HALO: REACH THE SPARTANS are seen on a more personally level, unlike other games. (Player 5)

Not all games supply diverse interaction channels. For example, first-person shooters may base interactions on combat only. Although such interactions are based on help, and trust, they will not fully model the breath, diversity, and heterogeneity of human interactions.

Criterion 2.3: Whether the character is round or flat

Narrative studies in film and literature suggest the next technique. A round character can surprise. They are non-stereotypical. They do more things than the plot requires them to do. They are not predictable, showing only a few traits, that is, always happy, always sad, always evil, or always good. Some of the traits may conflict, that is, the shy schoolteacher who goes parachute diving. They may change. They intrigue, because players can discover more.

The following comment from Buck Hammerstein (2011) shows the effect of the game facilitating surprise discovery, when a player behaves in unexpected ways:

The moments where a gamer explores the game world outside of the direct goals and is rewarded by clever world responses by developers who have imagined that we would perhaps do something out of the box is always precious.

In contrast, in traditional narratology, a flat character is one who is predictable, who only does things the plot requires them to do, who shows a limited number of (expected) traits, and who does not change. We suggest that these distinctions apply to game characters as much as to characters in film or literature. Earlier, we noted that asking what makes players feel engaged with game characters is not the same question as asking what makes game characters interesting or engaging. However, a natural overlap exists, as answers to the second question will contribute to the first question, as players are more likely to engage with interesting characters. The degree to which characters are round and complex depends on how important they are. Some only need to be “seen at a distance, like strangers or acquaintances” (Best et al., 1995).

Criterion 2.4: Whether the players’ relationships with NPCs are capable of change resulting from the players’ actions and behaviors

The relationship between player and NPC should be capable of change, of being transformed, born out of the players’ interaction with the character. The following example shows that if you treat the NPCs well, they will trust you:

Did you ever see a film called “The Thing”? It’s a very good film with Kurt Russell. He doesn’t know who to trust and the game is the sequel of that. It’s about what happens after and the people around you are reacting to your actions, because they don’t know who to trust either, so if you act irrationally like shooting your gun off or pointing your gun at them for no real reason, they will start to get agitated and annoyed, until finally they just destroy you completely and kill you. So they are acting directly to how you are acting. So if you give them weapons, they will trust you more. (Player quote reported by Mallon, 2008, Section 4, pp. 1-15)

This technique clearly replicates the responsiveness proposition above—The relationship between the player and NPC changes as a result of the player’s choices and behaviors.

Criterion 2.5: Whether relationship-forming is graduated

A significant issue that effects player responses to game-play relationships is whether the relationship is presented as already formed, or built up gradually, through the game play.

Relationships built up gradually assist in players buy-in as they go on a journey with the characters. Examples of ways in which relationships can form gradually were already given when discussing Criterion 2.2; for example, they can be built up upon help offered or received, humor exchanged, and so on. This idea is supported in the following critique of the game HEAVENLY SWORD:

I think that allowing the player to actually participate in several of the characters’ emotional journeys helped create that elusive player-to-character bond. Too often characters can become ... interchangeable and nothing more than talking parts of the level design. (Rhianna Pratchett, in an interview reported by Nutt, 2009)

Criterion 2.6: Design interactions to evoke psychological essences and emotions

Interactions can have little emotional content. For example, a player walks into a store, an NPC serves them and they walk out. The dialogue between the two is commonplace. Such an exchange invokes little emotion. Conversely, depending on the event, psychological essences or elements such as trust, suspicion, liking, dislike, admiration, fear, happiness, drama, tragedy, goodness, badness, politeness, and rudeness may be invoked. In Criterion 2.4 section, we saw players responding positively to opportunities to make moral or attitude choices (“if you give them weapons, they will trust you more”). Equally, they respond positively to making decisions concerning norms, values, and emotions, to being good or bad, rude, or polite. They like evidence that such choices are consequential. Players made favorable comments about opportunities to interact on a psychological level with NPCs. In the following quotes, the players clearly enjoy the emotions evoked:

GEARS OF WAR 3 offers so many twists and turns in the story it’s just amazing. Character relationships are so deep, that when one the main four die, it is so emotionally sad it brought tears to people’s eyes. Any Gears fan will admit it was one of the saddest moments in the game and I’m not afraid to admit that I shed a tear for Dom when he died. That’s what makes a good game though is it not? For a game to be so evocative something must have been done right. (GEARS OF WAR 3; Player 34)

I won’t lie. I teared up a little at his death. Mordin was an awesome character and I didn’t want to see him go. He went out a hero, but I wish he didn’t have to die at all. (MASS EFFECT 3; Forum thread started by Virgil Vox, September 23, 2012, titled Top 10 Gaming Moments)

Debriefing

It is important to analyze the role of debriefing in the methods and learning outcomes of the study. Game players may neither reflect on what they have played nor critically analyze the events, including the learning that takes place during the course of the game play. Indeed, Crookall (2010) argues that amid all the wonderful creations of games, some gamers seem to have “forgotten that the learning comes from the debriefing, not from the game” (p. 907). Debriefing played a key role in the study undertaken.

The exercise carried out by the students in this study was essentially a reflective and critical analytical process, intended to aid their learning. For instance, the participants documented in diary form, a journal of their thoughts and opinions while playing computer games. Following this process, players studied the transcripts of what they had written, and reflected on their experiences to identify the critical factors that enhanced or impeded their engagement with the games. We then asked them to extract any assessment criteria from their game experiences. As such, the exercise was a reflective process.

Players were also given time to reflect on the process, through the staged separation of the diary, distinguishing between the factors that made games engaging or disengaging, and the report containing their evaluation criteria. After submitting their written report, the participants presented their experiences in class to their fellow classmates and to their lecturer, an author of this article. The presentation of results evolved into a discussion where participants were able to “share their experiences in a frank, open and honest manner” (Pearson & Smith, 1986). To ensure a successful debriefing session took place, one that enhanced the learning experience, the facilitator (researcher) provided a “supportive climate” (Savoldelli, Naik, Hamstra, & Morgan, 2005) “where students feel valued, respected, and free to learn in a dignified environment” (Fanning & Gaba, 2007, p. 116). The discussion that ensued proved to be a fruitful exercise as players outlined their experiences and deliberated the factors, including the two criteria groupings discussed in this article, that lead to engaging games.

The debriefing process enhanced the players’ critical eyes in terms of looking at and assessing games. By the end of the debriefing stage, players were more capable of distinguishing between what engages them and what does not engage them. Players learned from their peers collaboratively during their presentations and during discussions following their presentations. Furthermore, the learning occurred through experiential practices. Moreover, the debriefing process provided a platform for all participants to air their game-play experiences, allowing players, the participants, the chance to further learn from the exercise.

Discussion

A question asked in this research is whether it is necessary for players to have deep and meaningful relationships with (some) NPCs. Relationships based on action sequences and tied to interactions required by the plot, where characters are only functions of what they do, may be the only kind of relationships presented in a game. Relationships can be shallow, rather than deep, but as illustrated with players’ comments, players express wonder and high levels of engagement when a deep and meaningful, as distinct from a shallow, relationship occurs with game characters.

Bruner helps to explicate how we defined above shallow versus deep relationships. Bruner’s (1986) conception of narrative includes what he titles the “landscape of action” and the “landscape of consciousness.” The latter is the interpretation that humans put on actions and events in terms of goals, intentions, motivation, values, and affect. An idea emerging from Bruner’s definition is that an action sequence on its own provides simply the skeleton, but not the flesh of an event. Shallow relations invoke the landscape of action and to a much lesser extent the landscape of consciousness.

We need to consider what puts the “flesh” on the skeleton events. How do players begin to interpret actions and events in terms of goals, intentions, motivation, values, and affect? What makes them care or feel relationships form? Our two criteria above, namely, responsiveness of the characters and stimulating deep and meaningful relations go some way to achieving this, but they are not exhaustive. To develop this work, future work should focus solely on characterization, across a more extensive number of games, with a larger sample size of participants.

The first criteria, responsiveness, is significant in creating what Salen and Zimmerman (2004) describe as meaningful play, where the player feels that the environment responds to their input:

Meaningful play in a game emerges from the relationship between player action and system outcome; it is the process by which a player takes action within the designed system of a game and the system responds to the action. The meaning of an action in a game resides in the relationship between action and outcome. (p. 40)

Meaningful play results from the player having an effect:

The core of meaningful play lies in the relationship between action and outcome. As a player uses core mechanics to take action, outcomes accumulate. These outcomes take many forms: sensory feedback, strategic achievement, emotional gratification, social relationships, and so on. As a player advances through a game, it is crucial that the game provide meaningful play at every moment. (Salen & Zimmerman, 2004, p. 411)

As indicated above, many kinds of outcomes are possible, or ways in which computer agents can respond intelligently to the player input. Such outcomes make play meaningful, which correlates to our idea of responsiveness.

Both the responsiveness and the stimulating deep and meaningful relations criteria we propose are abstract constructs, at a somewhat general level, which we deconstructed into a series of operational specifications, examples, and testing procedures to show how they may be implemented. They encompass a variety of techniques located across multiple dimensions. Again, the set of techniques we offer are unlikely to be an exhaustive listing of the many parameters affecting those abstract constructs. The relationship between the players’ actions in the game and the characters’ subsequent responses have a profound effect on the levels of engagement with the game. Games with higher levels of responsiveness and deeper relationships correlate with reports of higher levels of engagement.

Conclusion

The development of new technologies has also led to an evolution of the role of characterization in a game narrative setting. Not all designers can easily share developments from particular game cases, unless they document the lessons learned. A goal of this study and the methods employed was to identify constituents meaningful to players’ experiences and assessments. Recording how players appraise narrative games meant that lessons from particular games successes and failures can be drawn, then documented and distributed as evaluation criteria to aid other game creators.

When designing characterization, a game designer should be able to find in the literature specific details or concrete examples showing how to incorporate design principles or build related performance indicators. We might expect literature on narrative theory in games, in particular, to provide guidance on characterization. One of the intended contributions of this research is its attention to the concrete. Design or formative evaluation guidance may be quite general, offering largely abstract, broadly defined, general principles that a product should meet. We not only offer general evaluation criteria, but also provide the type of detail that explains how to achieve them, in many instances.

Good characterization is an important factor in creating good game experiences. Players want to experience meaningful relationships with the game and its characters. They want to erode the boundary lines between man and machine, between player and character. The development of deep connections between the players and game characters is something they desire. Failure to create bonds between players and their avatar and between their avatar and NPCs builds the barrier between man and machine. Players enjoy becoming one with the main character, sensing their elation at defeating an enemy, empathizing when losing a loved one, recognizing trepidation when faced with adversity, or feeling their angst when taking on an arduous task or deliberating a decision, and so on. They enjoy experiencing wish fulfillment, as they seek to deepen relationships with NPCs. This study has shown that meaningful relationships are something that players relish in games. Born out of game play, meaningful relationships between players and characters strengthen engagement levels with a game. Players enjoy experiencing deep, emotional connections, even with a computer-generated character. Put simply, players want to care.