Game Studies

“I Harbour Strong Feelings for Tali Despite Her Being a Fictional Character”: Investigating Videogame Players’ Emotional Attachments to Non-Player Characters

Abstract

This study investigated players’ emotional attachment to two non-player characters, Tali and Garrus, from BioWare’s Mass Effect videogame series. Two forum threads, created soon after the release by BioWare of Mass Effect 2 in January 2010 dedicated to these two characters, were downloaded and analyzed using thematic analysis. The results found that players had developed intense emotional attachments to the characters, but the emotional attachment for the female character of Tali was expressed from the players’ point of view, while the emotional attachment for the male character of Garrus was expressed from the player character of Shepard’s point of view. These emotional attachments influenced how players engaged with the game mechanics of Mass Effect 2 (BioWare, 2010), for example players customized their player character’s armour to match that of Tali or Garrus. As previous research into videogame characters has focused on playable characters, this article advances knowledge by considering the relationship between players and non-player characters in videogames.

Keywords: videogame characters, non-player characters, Mass Effect, emotional attachment, qualitative research, player-centric

Introduction

Despite appearing frequently in videogames regardless of genre, non-player characters (NPCs), characters the player does not directly control, have received far less research focus than player-characters (PCs) (Daviault, 2012; Egenfeldt-Nielsen, Smith, & Tosca, 2008). Furthermore, when videogame characters are discussed, it is often in passing in explicitly narratological-positioned studies of videogames (Schröter & Thon, 2014), or the discussion focuses less on the characters in the context of the videogame, and more on them as a conduit of behaviour. For example, videogame characters are often analyzed to assess how female videogame characters are depicted in terms of visibility and sexualisation (Downs & Smith, 2010; Lynch, Tompkins, van Driel, & Fritz, 2016), or the effects on aggression and violence of playing as a violent character (Lin, 2013).  There remains a research gap in the Game Studies discipline regarding players’ emotional attachments to NPCs (Bopp, Müller, Aeshbach, Opwis, & Mekler, 2019; Mallon & Lynch, 2014), which is important for videogame developers because it is recognised that emotional engagement of players due to game elements, including characters, is critical for games to succeed (Lankoski, 2012; Bopp et al., 2019). The emotional attachments to characters in part is responsible for the success of Nintendo, Square and BioWare’s game series (Jørgensen, 2010; Wesley & Barczak, 2010). However, investigation into the phenomenon of emotional attachment to videogame characters in the Game Studies discipline is still emerging. This article examines players’ responses to two NPCs from a narrative videogame, a game in which the story plays a significant role (Egenfeldt-Nielsen et al., 2008), to add to the emerging body of research examining the emotional attachments that can form between players and NPCs.

Theoretical Background

Player Characters

Player characters (PCs), sometimes referred to as avatars, have thus far been the central preoccupation of academic research into attachment to videogame characters, rather than NPCs. Banks and Bowman (2016) investigated the emotional connection between players and their PCs in the context of an online multiplayer and Huang and Yeh (2016) found when examining a serious game--one designed for purposes beyond entertainment--that the more players became attached to their PCs, the greater the players perceived their mastery of the game. Lewis, Weber, and Bowman (2008) developed a scale for measuring character attachment using five dimensions focusing on PCs and defining character attachment as the psychological merging of a player’s and character’s minds. Bowman, Schultheiss, and Schumann (2012), using a similar definition involving “mind melding” as Lewis et al. (2008), found that character attachment can be useful to understand the motivations of both pro- and anti-social videogame players. A more recent study by Bopp et al. (2019), in which participants were free to discuss PCs or NPCs, found seven distinct forms of emotional attachment that demonstrate the range of relationships a player can have with a videogame character.

Another strand of research that has explored how players relate to videogame PCs has focused on the concept of identification, which explores how audiences lose awareness of themselves and imagine their sense of self as a character (Cohen, 2001; Klimmt, Hefner, & Vorderer, 2009). Identification has been a prominent focus of media research for decades, with attention turning to videogame characters more recently (Van Looey, Courtois, De Vochy, & De Marez, 2012). However, identification has focused on PCs in massive multiplayer online games (Bessière, Seay, & Kiesler, 2007; Teng, 2017; Van Looey et al., 2012; Wu & Hsu, 2018). Therefore, to date there has been little research to understand the emotional attachment between videogame players and NPCs.

Non-Player Characters

The lack of research into the emotional attachments prompted by NPCs is an important research gap to address because videogames use NPCs to build emotional attachments with their players. Many game genres utilize NPCs, which may help or hinder the player, to make the game interesting, immersive, and challenging (Daviault, 2012). Despite this, there is limited research into NPCs outside of the body of literature devoted to their design (Daviault, 2012). Coulson, Barnett, Ferguson, and Gould (2012) investigated the factors that may affect how players become attracted to videogame characters using some NPCs with the sufficient depth and characterisation that players could have extensive and meaningful interactions with from Dragon Age Origins (BioWare, 2009) and found that players can develop real and meaningful emotions, and even love, for these characters. Investigating players’ responses to NPCs in Half-Life (Valve Corporation, 2004), Daviault (2012) found players perceived the NPCs as friends and that they added emotion and enjoyment to the game, rather than considering them part of the game mechanics. Overall, 96 percent of Daviault’s (2012) participants stated that they considered NPCs as playing a role in the story. Mallon and Lynch (2014) investigated ways that game creators could make players feel involved and engaged with videogame characters, focusing on both PCs and NPCs, by analysing comments players diarized about their feelings during play. Their results found two criteria for when players care for videogame characters: 1) the character’s responsiveness and 2) whether the character stimulated deep and meaningful relations. It was found players enjoyed developing relationships between their PC and an NPC, and when deep relationships were established, particularly when the relationships were of a romantic nature, players expressed wonder and high levels of engagement. Waern (2015) focused on romantic relationships in videogames and found that the four romantic options for PCs in Dragon Age Origins prompted emotional attachment. Furthermore, it was found that players experienced a ‘bleed effect’ similar to that of non-computerised role play, and players shared emotions with their PC similar to that of identification.

BioWare’s Characters

The videogame developer BioWare’s videogame characters have attracted their own small body of research as a result of their characters being recognised as well-developed with whom players can develop emotional attachments (Jørgensen, 2010). These attachments are facilitated by motion capture of body movements, facial expressions, and hours of fully voiced dialogue that the player can choose to engage with to better understand each squadmate’s traits, opinions, and personality (Jørgensen, 2010; Ward, 2008). As previously noted, Coulson et al. (2012) used the characters from BioWare’s Dragon Age series of games in their research regarding player attraction to videogame characters as did Waern (2015) when exploring ‘bleed’ and videogame romantic relationships. Toma (2015) analyzed the depiction and representation of the Dragon Age character Flemeth focusing on her role as an older and female character, while Pelurson (2018) focused on the depiction and reception of Dorian, an LGBTQ character from the same series. Zekany (2016) investigated the portrayal of romantic relationships between humans and non-humans and the reception of this portrayal by players, this time using the Mass Effect games. The BioWare videogame series, Mass Effect and Dragon Age, were the two most mentioned games by participants in Bopp et al.’s (2019) investigation of emotional attachment and videogame characters.

The findings of the research into BioWare’s characters appears analogous to the concept of parasocial relationships, which have long been explored in media research (Cohen, 2003). In contrast to identification, which explores how the audience and a character merge identities, parasocial relationships involve audiences viewing characters as separate entities (Bopp et al., 2019) that results in them feeling that they have a special connection and attachment with them (Eyal & Cohen, 2006; Sanderson, 2009). These relationships have been found to be comparable to relationships with real people (Cohen, 2003). However, research into the emotional attachments that videogame players form with NPCs is still emerging (Bopp et al., 2019), and further insight is needed to more fully understand these attachments.

Prior research into BioWare characters that has explored audience attachments and relationships with these characters has also focused on the design of the character in-game and their representation (Pelurson, 2018; Zekany, 2016). Other studies have required players to play a game purely for the purpose of the study (Daviault, 2012; Mallon & Lynch, 2014), and to prompt participants to reflect and write about the gameplay (Bopp et al., 2019; Mallon & Lynch, 2014), rather than examining organic and unprompted reactions. To provide new knowledge and continue to address the research gap concerning NPCs and emotional attachment, this research analyzes players’ posts from BioWare’s official forums for two of Mass Effect’s most popular NPC squadmates to understand players’ emotional responses to these characters: the male Garrus, a member of the fictional turian race, and the female Tali, a member of the fictional quarian race. Emotional attachment is conceptualised using Bowlby’s (1979) long-established explanation of attachment as an ‘emotion-laden bond between a person and a specific target’. This more flexible definition is used by this study because NPCs are being explored, and so the prior definition of the merging of a character’s and player’s minds is not appropriate. Bowlby’s (1979) often-cited definition of attachment also aligns with the recent description of character attachment proposed by Bopp et al., (2019), which is the ‘sense of liking, connection and closeness a player feels to any in-game character’, and was designed to be applicable to PCs or NPCs. This research also takes place over a longer time than Bopp et al. (2019), addressing one of their calls for future research, and provides a more in-depth analysis of a dataset than Waern (2015), so building on and expanding these studies. It also addresses Mallon and Lynch’s (2014) call for more research focusing on the characterisation of NPCs.

Methodology

Mass Effect Series Background

A brief summary of the gameplay of the Mass Effect series follows to provide context for those unfamiliar with the series. BioWare’s Mass Effect series is best described as a science-fiction shooter and role-player game hybrid (Carvalho, 2015). The original Mass Effect trilogy was released from 2007-2012 with the first game set in 2183 by which time humanity has made contact with several alien races and integrated into galactic society (Zakowski, 2014). A fourth game was later released in 2017 and is mostly set 600 years after the events of the original trilogy. This research is only concerned with the original trilogy as these are the three games in which Tali and Garrus appear. The PC of the trilogy is named Shepard, an elite military operative with customisable abilities and playable as either male or female, who is tasked with saving the galaxy from an enemy called the Reapers. In each game, Shepard is accompanied by squadmates, NPCs that join the player in the combat and act autonomously but can also be given limited instructions. Approximately 100 hours are required to play all three of the games in the original Mass Effect trilogy (Jørgensen, 2010). Players can select from a variety of dialogue options in each NPC conversation throughout the series, and choose between different actions, plot points or speech.

Research Design and Data Collection

The characters of Tali and Garrus were selected as the research focus due to the Mass Effect series’ ability to create well-developed characters with high-quality voice acting and models (Jørgensen, 2010). Both characters were the only NPCs to accompany the PC of Shepard through all three videogames, allowing the player to get to know them. They are also potential romantic partners for the PC in the second Mass Effect game; a male PC of Shepard can romance Tali, and a female PC of Shepard can romance Garrus. As such, they are examples of the characters Mallon and Lynch (2014) note are particularly appreciated by videogame players: characters who can form relationships with the PC of Shepard--which may be romantic in nature--and by extension the player themselves. This makes them appropriate to explore emotional attachments with NPCs.

BioWare’s official Mass Effect forums, hosted on the BioWare Social Network, were selected as the research site because they attracted a large number of posters, who were likely to be dedicated Mass Effect players. Furthermore, the most loyal and devoted audiences of videogames are most likely to take part in online discussions and have opinions regarding the characters and narrative of the games (Burgess & Jones, in press. a). Before the forums on the BioWare Social Network were closed and deleted in 2016, threads devoted to discussing a particular character were common. Two threads devoted to the characters of Tali and Garrus were selected for analysis to investigate if conversations in the threads and responses to these characters changed over time. Player posts were analysed, rather than the game itself, as this research was player-centric and focused on players’ reactions and responses. The initial threads for each character, termed version one (V1), were shut down by the BioWare forums, but players started a second thread, termed version two (V2), for each character to continue their conversations. These threads were started immediately following the release of Mass Effect 2 (BioWare, 2010) and were selected because the extensive, and heated, discussion regarding the ending (Jennings, 2016) of Mass Effect 3 (BioWare, 2012) after its release would be expected to influence the discussion about each character. Details about each thread are presented in Table 1. Each thread was downloaded using a custom-built python script.

Table 1: Details of Analyzed Threads

Data Analysis and Validity

The methodological approach adopted to explore these threads involved qualitative analysis, which is appropriate because it is particularly useful for investigating online communities, such as forums (Kozinets, 2002).  Due to the size of the dataset (number of posts), it was deemed to contain sufficiently rich information and so be appropriate for qualitive analysis (Arvidsson & Caliandro, 2016; Scholz & Smith, 2019). Thematic analysis, following the guidelines and methods established by Braun and Clarke (2006), were used to explore the major themes arising from the analysis of the sub-arenas. Thematic analysis was selected because it is a flexible method that can be applied across a range of theoretical approaches (Braun & Clarke, 2006), which is appropriate considering the multi-disciplinary nature of Game Studies research (Hjorth, 2011). Thematic analysis can be used to comprehend how people feel, think, and behave in specific contexts (Guest, MacQueen, & Namey, 2012), which makes it an appropriate method to investigate players’ emotional attachments and responses to videogame characters. A theoretical, rather than an inductive approach (Braun & Clarke, 2006) was undertaken to provide a nuanced account of specific themes related to emotional attachment, instead of a description of all the data as thematic analysis’s flexibility can become a limitation if the research is not sufficiently focused (Braun & Clarke, 2006).

Automatic analysis of data using the analytical software program Leximancer was combined with a traditional human interpretative approach to assist with the analysis of the large amount of data (Arvidsson & Caliandro, 2016). Leximancer can be used to identify high-level concepts and extract their semantic and relational information from large volumes of text (Leximancer Pty Ltd, 2017). To do so, a text duplicate sensitivity setting was used to remove the phrases repeated on every single post or web page, for example ‘This topic is locked’, and to account for posters quoting each other. Compound concepts were equated when they were names that had not been consistently capitalized; for example, ‘Tali’ and ‘tali’. Several stop words were also added to Leximancer’s lexicon due to their frequent usage: ‘man’, ‘course’, ‘need’, ‘guy’, ‘guys’, ‘bad’, ‘work’, and ‘lol’. Each theme and every line of text that Leximancer identified was scrutinised by the researchers using contextual knowledge gained from immersion resulting from the researchers’ active story interpreter roles because a purely computer analysis of qualitative data can be less accurate and constrain the meaning of the data being analysed (Kozinets, Scaraboto, & Parmentier, 2018). This scrutiny resulted in themes being renamed and reorganised to account for context, sarcasm and community specific slang, jokes and references (Scholz & Smith, 2019) that Leximancer could not take into account. The themes and concepts generated by Leximancer were analyzed by starting with both of the first threads, and then moving on to both of the second threads when no new themes could be identified, or existing themes were not further refined.

The analysis was informed by an adaptation of netnography (Reid & Duffy, 2018) known as non-participatory netnography (Scholz & Smith, 2019), which does not involve interacting with research participants and community members and was necessary due to the threads that were analysed being locked. Netnography is a respected, flexible, and useful analysis approach to provide a record of a specific phenomenon and behaviour in social reality with no single correct way to conduct it and with multiple adaptations and advancements (Kozinets et al., 2018; Reid & Duffy, 2018). The netnographic role of active story interpreter was undertaken by the researchers (Reid & Duffy, 2018). This interpretivist role involves immersion in the data and its context to comprehend and decode cultural insights into the phenomenon being investigated to ensure a valid analysis (Reid & Duffy, 2018). Such an approach allows for more accurate representations of the feelings expressed by the participants and the research context under investigation (Mittell, 2015). The specific immersion for this research involved playing all three games of the Mass Effect original trilogy, reading other character threads on the forums, viewing player created art and stories about Tali and Garrus, and reading news articles and commentaries about the Mass Effect series. A codebook was derived by thematic analysis including definitions of each theme and example quotes. Peer debriefing in the form of weekly meetings were used to validate the thematic analysis and the codes derived from it (Creswell & Miller, 2000) that include definitions of each theme and example quotes. During the coding process, a number of codes and themes were identified. Only those that directly relate to the article’s focus on emotional attachment to the NPCs are listed in the Codebook.

Results

The themes and sub-themes that occurred in each individual thread evolved and changed between the first and second iterations of each character’s thread. In all four threads, the results reveal that players had strong emotional attachments to the two NPCs, but there were differences in the ways in which these emotional attachments were expressed. Players wrote about their emotional attachment and feelings in the Tali thread, but in the Garrus thread the focus was on the relationship and attachment between Garrus and the PC of Shepard. The results also revealed that the emotional attachments that players felt were strong enough to influence how they engaged with Mass Effect 2’s (BioWare, 2010) game mechanics. The Codebook follows in Table 2.

Table 2: Codebook

Analysis

The following section presents the themes that emerged from the data analysis. Sub-headings are used for clarity.

Tali

The two threads devoted to discussion about the female NPC of Tali contained multiple explicit examples of players posting that they loved the character and explaining why (Player Feelings themes). For example, one player noted: “I love Tali”, and agreed with another who said she was: “[...] a character that you can actually fall in love with” (Tali V2, Player Feelings). Another stated: “[...] the attraction to Tali has far, far more to do with her character (intelligent, sweet, shy, quirky, selfless, courageous, honest)” (Tali V1, Player Feelings). The Tali threads also contained discussion about the development of her personality and character through the first two Mass Effect games. For example: “her character development and your characters interactions with her in ME2 [Mass Effect 2] are awesome and very well done” (Tali V1, Personality/ Character). The comments posted about Tali’s character suggest that players developed strong, emotional attachments to her based on what they perceived was her personality.  The strength of these attachments supports Mallon and Lynch’s (2014) findings that characters, who can be the focus of a romantic relationship, are greatly appreciated by players, which is evident by the passionate, emotional reactions that Tali prompted from posters in the Tali threads. One player wrote that Tali made them feel that they “wanted to hold her and tell her everything would be okay” (Tali V2, Player Feelings), while another called her the “sweetest, cutest and most adorable character in mass effect” (Tali V2, Player Feelings).  The attachments expressed would also seem to classify Tali as both a ‘crush’ and ‘trusted close friend’ according to the categories of emotional attachment to characters developed by Bopp et al. (2019).  Players expressed their belief that Tali was courageous, they could depend on her and that she would fight alongside them, while also noting their infatuation with her. This combination of friendship along with romantic feelings could account for the intensity of the attachments expressed by players. These players are experiencing two different types of attachment at the same time.

Interestingly, when players were commentating on their love for Tali, the comments were all written from their viewpoint, not that of their PC, the character of Shepard. Players discussed their emotional attachments to Tali, not how they perceived their PC of Shepard and Tali felt about each other. Although discussion about the relationship between the PC of Shepard and Tali eventuated in the second Tali thread, there remained a focus on how players felt about her. In this way the emotional attachments took the form of the bleed affect suggested by Waern’s (2015) research. Players thus appeared to be experiencing identification with their PC and merging their PC’s (Shepard’s) feelings for Tali with their own feelings. Perhaps as a result, the intensity of emotional attachments expressed towards Tali does not appear to completely align with the definition of attachment provided by Bopp et al. (201), which was a ‘sense of liking, connection and closeness’ felt towards a character. Liking would appear to not rise to the level of devotion expressed by the posters, indicating a definition that allows for levels of intensity of feeling is required.

The ability of players to continue their romantic relationship with Tali in the next game in the series, Mass Effect 3 (BioWare, 2012), was also an important topic of discussion and concern in the Tali threads. Players considered her return, and the return of the romantic relationship between her and Shepard, mandatory. For example, one player wrote: “If she lives, she MUST remain a squad member, and if you romanced her, she MUST remained romanced in ME3 [Mass Effect 3]” (Tali V1, Hope), while another stated: “Tali being in ME3 as a full squadmate and a continuing possible romance is an absolute must” (Tali V1, Hope). Many of the posters’ desired inclusions in Mass Effect 3 (BioWare, 2012) beyond Tali’s return involved furthering the romance with her.  One player echoed multiple other posts writing: “I want to bloody marry her at the end of ME3 and build a house for her and live with her on her home world darn it” (Tali V1, Hope). Players felt so strongly about the romance with Tali that if BioWare did not include it, they cautioned there would be consequences. For example, “[...] it would be a huge mistake from BioWare not to include Tali in ME3 both as a squadmate and a love interest” (Tali V2, Hope), and another stated they would “never forgive” BioWare if that happened (Tali V1, Hope).

Among the stated consequences of Tali not appearing as a squad mate were players warning they would hate the game, not buy the game, and even jokingly “[...] throwing eggs at the bioware office if Tali isn't a LI [love interest] and squadmate in ME3 [...]” (Tali V1, Hope). The discussion around the status of Tali and her romance in Mass Effect 3 (BioWare, 2012) was a significant concern in both Tali threads. Players posted that they had initiated a romance with Tali and wanted this romance to continue and develop further in Mass Effect 3. However, despite these requests, Mass Effect 3’s (BioWare, 2012) executive producer, Casey Hudson, surprisingly stated that there was discussion about removing Tali as a squadmate, but several of the game’s writers argued against this (Matulef, 2013). The existence of this discussion indicates that the emotional attachments players develop with NPCs were not understood or recognized by their creators and game developers. This is surprising given the importance of emotion to the success of videogames (Lankoski, 2012). Videogame players form expectations of what will happen to the plot and characters in continuations and can react negatively if these are contradicted (Burgess & Jones, in press b). Furthermore, the emotional attachments players can form for NPCs are also important for developers to understand the expectations of the players for videogame continuations.

Garrus

The two threads devoted to the male NPC of Garrus also demonstrated strong emotional responses about him by players, but these manifested in a different way compared to the two Tali threads. Although there was a small discussion that aligned with the Tali thread, with posters expressing their appreciation and positivity for the unfolding process of developing a romantic relationship with him through the writing and cut-scenes, there was significant criticism of some aspects of the romance in the first Garrus thread. For example, that the final cut-scene was too short, and more physical intimacy and emotion should be added to the dialogue. One player wrote: “It is not much for us Garrus fans to ask for a scene that shows there is at least some degree of love there” (Garrus, V1, Romance Mechanics).  The critique of the romantic relationship between Shepard and Garrus continued in the second Garrus thread but to a lesser extent, as it appeared partially exhausted in the first Garrus thread. Although Mallon and Lynch (2014) found that players particularly appreciated forming romantic relationships with NPCs, there was a divergence between the Tali and Garrus threads regarding satisfaction with the romantic relationship players could develop. Like Tali, players’ emotional attachments to Garrus fell under both the ‘crush’ and ‘trusted close friend’ categories (Bopp et al., 2019), but players felt that Shepard’s romantic relationship with Garrus did not reflect the close friendship between the characters. The NPC of Tali reveals she had romantic feelings for the PC of Shepard before the events of Mass Effect 2 (BioWare, 2010), while the NPC of Garrus appears to not have considered a romantic relationship with the PC of Shepard before the player initiates the dialogue options. One player noted they were “[...] jealous of the Tali fans to be honest, since she basically tells Shepard that she had been crushing on him all throughout ME1 and just never thought he'd be interested in her” (Garrus V1, Tali). Another player explicitly compared the Garrus romance to other potential romance options and found it lacking: “[...] they gave us the least closure for our relationship and the least-eventful lovescene [sic] (though most emotional)” (Garrus V1, Desire for Intimacy). Therefore, those players, who chose to form a romantic relationship between their female PC of Shepard and the male NPC of Garrus, appreciated and enjoyed it but did not perceive it as being equal to the ones players could form between the male PC of Shepard and the female NPC of Tali, or some of the other NPCs, who were also potential romance options. Players were actively comparing Shepard’s romance with Garrus to that of other romantic interests for Shepard.

In contrast to the Tali discussions, the first Garrus thread discussions were based primarily on the emotional responses of players to the relationship they perceived had developed between their PC of Shepard and the NPC of Garrus, rather than players’ attachment to Garrus themselves. The theme, Player Feelings, garnered fewer posts in the Garrus threads compared to the Tali threads. Instead, players’ discussions focused more on the perceived nature of the relationship between Shepard and Garrus. For example: “Garrus wouldn't be getting so emotional and nervous over something like that, and as was already stated, he doesn't have a fetish for humans - Shepard is special to him for other reasons” (Garrus V1, Relationship). Another noted: “I think it's very telling that when Shepard initiates the romance, Garrus says yes with barely a thought” (Garrus V1, Relationship). The posters in the Garrus threads viewed their Shepard as sufficiently developed to theorize and describe their romantic relationship with another character. They discussed ‘Shepard’ and what Shepard thought or felt much more than what they felt shown by them writing in third person, rather than first person. Rather than experiencing bleed, identification, or the definition of character attachment by Lewis et al. (2008) involving “mind melding” as the Tali posters and devotees appeared to, the players posting in the Garrus threads appeared to feel a different kind of emotional attachment or impact. This attachment aligns more with Bowlby’s (1979) attachment definition of ‘an emotion-laden target-specific bond’, as well as resembling a parasocial relationship (Eyal & Cohen, 2006; Sanderson, 2009). Research into user preference for PCs, using the Second Life virtual world as the research context, found that 82% of players selected a PC the same gender as themselves (Rymaszewski et al., 2007). BioWare’s own statistics also indicated that 82% of players selected the male Shepard (Makuch, 2013). Since Tali is only available as a romantic option for a male Shepard, and Garrus for a female Shepard, that suggests there may be a difference in how different genders of players feel and express attachment for characters.

Attachment Motivated Behaviour

The explicit comparison between the NPCs of Garrus and Tali by posters to other romanceable characters, and discussion about these other romances, was encompassed by the Other Characters and Other Romances themes. The Tali threads, when not discussing Tali, tended to discuss the other characters favourably, but in general terms, rather than focusing on the other romantic options. For example: “It is a testament to BioWare's ability to create believable characters that we can relate to as though they were living, breathing people” (Tali V1, Other Characters). The discussion about other romantic options was not in-depth, but when present posters did not compare them favourably to Tali. The posters in the Tali threads were satisfied that the romantic relationship they had developed was rewarding, and they held positive views towards the other NPCs that populated the game, even if they were viewed as not being satisfactory romantic options. They did not feel the need to discuss or explore the other romantic interests in depth as they were secure that their relationship with Tali was fulfilling in contrast to the posters in the Garrus thread, who discussed and critiqued the other romantic options, instead of the other NPCs more generally.

Due to players’ emotional attachments to Tali and Garrus, posters compiled information and used it to create theories for new topics to discuss to avoid becoming repetitive as the discussions in the original threads began to repeat and converge. An example is the racial differences between the human Shepard, the turian Garrus and the quarian Tali, which appeared to have been motivated by the romantic relationship Shepard could develop with Garrus or Tali, since most of the differences in species focused on speculated biological or cultural compatibility. For example: “I am sure she feels that Quarian children are necessary but Shepard has made her feel something no one ever has and she doesn't seem like the type to take things lightly” (Tali V2, Quarian Culture and Biology). In the second Garrus thread, the romantic relationship between Shepard and Garrus also caused some speculation: “My theory: The poor guy probably thought human culture was similar enough to turian culture that he could just go about seducing her like he would a turian [...]” (Garrus V2, Turian Culture and Biology). These conversations eventually evolved to include some speculation as to the culture and biology of both races, rather than focusing on their implications for the romances. For example: “[...] turians are very poor at retaining heat (as a result of their warm environment), so that's why you rarely see turians out of armor [...] even on climate-controlled space stations they're completely covered, to the point of wearing gloves” (Garrus V2, Turian Culture and Biology). The second Tali thread also moved on to extensive discussion about the war between the quarians and a robotic race, the Geth, who occupied the quarian home world. Several points of view were offered on the chances of peace and the outcomes of a military engagement (Returning Home). The threads’ discussions continued to be heavily motivated by emotional attachments to the NPCs of Tali and Garrus but matured with more analysis and interpretation occurring when shallower topics of discussion were exhausted and became repetitive. Motivated by their emotional attachments for the characters, players explored topics they felt were important to them.

Several of the discussions about Tali and Garrus also revolved around how players’ emotional attachments for them affected how they used the game mechanics in Mass Effect 2 (BioWare, 2010), for example player customisation. One poster asked: “Anyone else here colored Shepard's armor to match Garrus?” (Garrus V1, Game Manipulation) and received several posts in agreement. A similar discussion occurred in the Tali threads, where players made their Shepard’s helmets match Tali’s, whose face is always hidden behind her helmet. For example: “Did anyone else choose to play their Shepherd with an enclosed helm just to be in fellowship with Tali?” (Tali V2, Game Manipulation). Another example was the preoccupation with trying to influence the composition of a cut-scene towards the end of the game to more prominently feature Garrus or Tali over the other NPC in the party. The cut-scene involves Shepard saving one of the two NPCs in the playable party from falling off a ledge, and then is later saved by that character. Players wanted to make sure it was Tali or Garrus who was saved, and then did the saving because of the emotional impact on the player. This question was asked and debated so frequently that one poster noted: “My friend and I have pooled our data, and which squadmate you save is still up in the air; at first I thought LI  [love interest] overrrode anyone else, but then he saved someone else while Tali was in his party” (Garrus V1, Game Manipulation). The selection of NPCs for certain tasks in the final mission of Mass Effect 2 (BioWare, 2010), where more than one character will perform optimally, was also influenced by players’ attachments. For example, a Tali poster explained: “I send Tali into the vents because the dialogue with Shep congratulating her success makes me happy” (Tali V2, Game Manipulation). Daviault (2012) found that players’ responses to NPCs in Half-Life were steeped in emotion, and they viewed characters as an element of the narrative/storyline, rather than as part of the game mechanics. However, it appears that some NPCs in the Mass Effect series could prompt an emotional attachment strong enough that players reflected on how they could use the game mechanics to further these emotional attachments for no tangible benefit.

Conclusions and Future Research

This research analyzed unsolicited forum posts to explore how players responded to and expressed their emotional attachment to two NPCs, Tali and Garrus, in the first two videogames in the Mass Effect trilogy to add to the body of knowledge exploring emotional attachments to NPCs. Players explicitly ascribed personality and depth of character to the two NPCs and proudly declared their love and attachment for them. Motivated by their emotional attachments, players engaged in theory and analysis regarding the biological and cultural aspects of the different races of Garrus and Tali. Players also spent time and effort attempting to manipulate the game due to these emotional attachments, which demonstrates the strength of their relationships with the NPCs.

 Several findings and areas for future research have also been identified. Emotional attachment to NPCs is still an emerging body of research in the Game Studies discipline, especially when compared to that exploring attachments to PCs, and this article builds on prior work by conducting research over a longer period of time and examining a greater quantity of data in-depth. The strength of the emotional attachments that players developed for the NPCs of Tali and Garrus supported Mallon and Lynch’s (2014) suggestion that characters, who players can build a romantic relationship with, are more appreciated. This research also found support for two of Bopp et al.’s (2019) categories of emotional attachment to characters, although it was found that each character fell under more than one category. The emotional attachments expressed by posters also appeared to be more intense than encompassed by the definition provided by Bopp et al. (2019), indicating emotional attachment to NPCs requires a definition that accounts for these powerful feelings or a scale to measure the degree of attachment, which future research could develop and validate. Future research could also analyze the videogames that have prompted emotional attachments, in conjunction with analyzing players’ reactions and responses to this attachment, to triangulate and understand what it is about certain characters, games, and stories that prompts attachment.

Players expressed intense emotional attachments and viewed the NPCs as characters possessing distinct personalities and they were romantic interests. It was found that most likely male players conflated themselves with the male PC of Shepard when describing their emotional attachment to, and love for Tali, while most likely female players discussed their emotional attachment to Garrus through the point of view of their female PC of Shepard. Consequently, the posts about Garrus focused on discussing and theorising about the relationship between him and female Shepard, while the Tali posts focused on the players’ relationship with Tali. The attachments expressed about Garrus appeared to be closer to parasocial relationships (Eyal & Cohen, 2006; Sanderson, 2009), and followed the definition of attachment provided by Bowlby, while the attachments expressed about Tali appeared to be an example of identification (Cohen, 2001; Klimmt et al., 2009), bleed (Waern, 2015), and mind melding (Lewis et al., 2008). Since identification has predominantly been explored in relation to MMOs, this research adds to knowledge of the concept, as well as knowledge about attachment to NPCs. Future research could attempt to build on this finding and explore and identify differences between how male and female players feel and express emotional attachment to NPCs.

Finally, despite the fact that the data analyzed in this study came from the official Mass Effect forums, the Mass Effect series developer, BioWare, did not appear to fully understand the emotional attachment players had developed with the Mass Effect 2 (BioWare, 2010) NPCs they had created.  Even with players extensive pleas in the BioWare forums posts for the return of Tali, BioWare considered not including her as a squadmate in Mass Effect 3 (BioWare, 2012). Players develop expectations based on their emotional attachments to the characters, world, and stories of videogames (Burgess & Jones, in press b). Ignoring players’ expectations for characters in continuations can be dangerous given the emotional attachment they can have for them, and emotion is vital for videogames to be successful (Lankoski, 2012). Analyzing the content that players create and share should be a priority for videogame developers with ambitions to create emotionally attractive games (Enevold & MacCallum-Stewart, 2015) to ensure that they are aware of how their players are responding to their characters, which is especially important if continuations for a videogame are intended.

Acknowledgements

The authors thank Dr Joanna McIntyre for her comments on an earlier draft of this manuscript and the anonymous reviewers, whose comments strengthened this article. The corresponding author wishes to acknowledge the Commonwealth of Australia for providing financial assistance under the Research Training Program scholarship.

References

Arvidsson, A., & Caliandro, A. (2016). Brand public. Journal of Consumer Research, 42(5),727-748. doi: https://doi.org/10.1093/jcr/ucv053

Banks, J., & Bowman, N. D. (2016). Avatars are (sometimes) people too: Linguistic indicators of parasocial and social ties in player-avatar relationships. New Media & Society, 18(7), 1257-1276. doi: https://doi.org/10.1177/1461444814554898

Bessière, K., Seay, A. F., & Kiesler, S. (2007). The ideal elf: Exploration in World of Warcraft. CyberPsychology & Behavior, 10(4), 530-535. doi: https://doi.org/10.1089/cpb.2007.9994

Bopp, J. A., Müller, L. J., Aeshbach, L. F., Opwis, K., & Mekler, E. D. (2019, October). Exploring emotional attachment to game characters. CHI PLAY '19: Proceedings of the Annual Symposium on Computer-Human Interaction in Play,313-324. doi:  https://doi.org/10.1145/3311350.3347169

Bowlby, J. (1979). The making and breaking of affectional bonds. London, UK: Tavistock Publications.

Bowman, N., Schultheiss, D., & Schumann, C. (2012). ‘I'm attached, and I'm a good guy/ gal!': How character attachment influences pro- and anti-social motivations to play massively multiplayer online role-playing games. Cyberpsychology, Behavior & Social Networking, 15(3), 169-174. doi: https://doi.org/10.1089/cyber.2011.0311

Burgess, J., & Jones, C. (in press a). Exploring the forced closure of a brand community that is also a participatory culture. European Journal of Marketing.

Burgess, J., & Jones, C. (in press b). “In a story-driven game, if the ending sucks it can ruin the whole franchise”: Evaluating a narrative brand’s crisis recovery. Leisure/Loisir.

Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77-101.doi: 10.1191/1478088706qp063oa

Carvalho, V. (2015). Leaving earth, preserving history: Uses of the future in the mass effect series. Games and Culture, 10(2), 127-147. doi: https://doi.org/10.1177/1555412014545085

Cohen, J. (2001). Defining identification: A theoretical look at the identification of audiences with media characters. Mass Communication and Society, 4(3), 245-264. doi: https://doi.org/10.1207/S15327825MCS0403_01

---- (2003). Parasocial breakups: Measuring individual differences in response to the dissolution of parasocial relationships. Mass Communication and Theory, 6(2), 191-202. doi: https://doi.org/10.1207/S15327825MCS0602_5

Coulson, M., Barnett, J., & Ferguson, C. (2012). Real feelings for virtual people: Emotional attachments and interpersonal attraction in video games. Psychology of Popular Media Culture, 1(3), 176-184. doi: http://psycnet.apa.org/doi/10.1037/a0028192

Creswell, J. W., & Miller, E. L. (2000). Determining validity in qualitative inquiry. Theory Into Practice, 39(3), 124-130. doi: https://doi.org/10.1207/s15430421tip3903_2

Daviault, C. (2012). Does game playing experience have an impact on the player-PNPC relationship? Bulletin of Science, Technology & Society, 32(6), 441-446. doi: https://doi.org/10.1177/0270467612469075           

Downs, E., & Smith, S. L. (2010).  Keeping abreast of hypersexuality: A video game character content analysis. Sex Roles, 62(11-12), 721-733. doi: https://doi.org/10.1007/s11199-009-9637-1

Egenfeldt-Nielsen, S., Smith, J. H., & Tosca, S. P. (2008). Understanding video games: The essential introduction. New York, NY: Routledge.

Enevold, J., & MacCullum-Stewart, E. (2015). Introduction. In J. Enevold, & E. MacCallum-Stewart (Eds.), Game Love, (pp. 1-21). Jefferson, NC: McFarland.

Eyal, K., & Cohen, J. (2006). When good friends say goodbye: A parasocial breakup study. Journal of Broadcasting & Electronic Media, 50(3), 502-523. doi: https://doi.org/10.1207/s15506878jobem5003_9

Guest, G., MacQueen K. M., & Namey, E. E. (2012). Applied thematic analysis. Thousand Oaks, CA: Sage.

Hjorth, L. (2011). Games and Gaming. London, UK: Berg.

Huang, L-Y., & Yeh, Y. (2016). Mediated enactive experience versus perceived mastery experience: An enhancing mechanism versus a mediator of character attachment and internal political efficacy in serious games. Computers in Human Behaviour, 55(Part B), 1085-1096. doi: https://doi.org/10.1016/j.chb.2015.10.029

Jennings, S. (2016). Co-creation and the distributed authorship of video games. In L. J. Jensen & K. D. Valentine (Eds.), Examining the evolution of gaming and its impact on social, cultural, and political perspectives (pp. 123-146). Hershey, PA: IGI Global.

Jørgensen, K. (2010). Game characters as narrative devices. A comparative analysis of Dragon Age: Origins and Mass Effect 2. Eludamos. Journal for Computer Game Culture, 4(2), 315-331. doi: http://www.eludamos.org/eludamos/index.php/eludamos/article/view/vol4no2-13/192

Klimmt, C., Hefner, D., & Vorderer, P. (2009). The video game experience as “true” identification: A theory of enjoyable alterations of players' self-perception. Communication Theory, 19(4), 351-373. doi: 10.1111/j.1468-2885.2009.01347.x

Kozinets, R. V. (2002). The field behind the screen: Using netnography for marketing research in online communities. Journal of Marketing Research, 39(1), 61-72. doi: https://doi.org/10.1509/jmkr.39.1.61.18935

----, Scaraboto, D., & Parmentier, M-A. (2018). Evolving netnography: How brand auto-netnography, a netnographic sensibility, and more-than-human netnography can transform your research. Journal of Marketing Management, 34(3-4), 231-242. doi: https://doi.org/10.1080/0267257X.2018.1446488

Lankoski, P. (2012). Computer games and emotion. In J. Sageng, H. Fossheim, & T. Mandt Larsen (Eds,), The Philosophy of Computer Games. Philosophy of Engineering and Technology, vol 7, (pp. 39-55). Dordrecht, Netherlands: Springer.

Lewis, M. L., Weber, R., & Bowman, N. D. (2008). They may be pixels, but they’re MY Pixels: Developing a metric of character attachment in role-playing video games. CyberPsychology and Behavior, 11(4), 515-518. doi: https://doi.org/10.1089/cpb.2007.0137

Leximancer Pty Ltd (2017). Leximancer white paper. Retrieved from https://info.leximancer.com/science/

Lin, J. (2013). Identification matters: A moderated mediation model of media interactivity, character identification, and video game violence on aggression. Journal of Communication, 63(4), 682-702. doi: https://doi.org/10.1111/jcom.12044

Lynch, T., Tompkins J. E., van Driel, I. I., & Fritz, N. (2016). Sexy, strong, and secondary: A content analysis of female characters in video games across 31 years. Journal of Communication, 66(4), 564-584. doi: https://doi.org/10.1111/jcom.12237

Makuch, E. (2013, March 23). Mass Effect 3 players prefer male Shepard. Gamespot. Retrieved from https://www.gamespot.com/articles/mass-effect-3-players-prefer-male-shepard/1100-6405842/

Mallon, B., & Lynch, R. (2014). Stimulating psychological attachments in narrative games. Engaging players with game characters. Simulation & Gaming, 45(4-5), 508-527. doi: https://doi.org/10.1177/1046878114553572

Matulef, J. (2013). Mass Effect 3 nearly didn't have Tali as a squadmate. Eurogamer, 23 March. Accessed 7 December 2017. http://www.eurogamer.net/articles/2013-03-23-mass-effect-3-nearly-didnt-have-tali-as-a-playable-character

Mittell, J. (2015). Complex tv: The poetics of contemporary television storytelling. New York, NY: New York University Press.

Pelurson, G. (2018). Mustaches, blood magic and interspecies sex: Navigating the non-heterosexuality of Dorian Pavus. Game Studies, 18(1). Retrieved from http://eprints.leedsbeckett.ac.uk/6021/1/Mustaches%2C%20Blood%20Magic%20and%20Interspecies%20Sex%20-%20Gaspard%20Pelurson.pdf

Reid, E., & Duffy, K. (2018). A netnographic sensibility: Developing the netnographic/social listening boundaries. Journal of Marketing Management, 34(3-4), 263-286. doi: https://doi.org/10.1080/0267257X.2018.1450282

Rymaszewski, M., Wagner, J. A., Wallace, M., Winters, C., Ondrejka, C., & Batstone Cunningham, B. (2007). Second life the official guide. New Jersey, NJ: Wiley.

Sanderson, J. (2009). ‘You are all loved so much’: Exploring relational maintenance within the context of parasocial relationships. Journal of Media Psychology, 21(4), 171-182. doi: 10.1027/1864-1105.21.4.171

Scholz, J., & Smith, A. N. (2019). Branding in the age of social media firestorms: How to create brand value by fighting back online. Journal of Marketing Management, 35(11-12), 1100-1134. doi: https://doi.org/10.1080/0267257X.2019.1620839

Schröter, F., & Thon, J-N. (2014). Video game characters: Theory and analysis. Diegesis Interdisciplinary E-Journal for Narrative Research, 3(1). Retrieved from http://elpub.bib.uni-wuppertal.de/servlets/DerivateServlet/Derivate-4064/dej14030103.pdf

Teng, C-I. (2017). Impact of avatar identification on online gamer loyalty: Perspectives of social identity and social capital theories. International Journal of Information Management, 37(6), 601-610. doi: https://doi.org/10.1016/j.ijinfomgt.2017.06.006

Toma, E. (2015). Powerful elderly characters in video games: Flemeth of Dragon Age. Journal of Comparative Research in Anthropology and Sociology, 6(2), 51-65. Retrieved from https://www.ceeol.com/search/article-detail?id=459106

Van Looey, J., Courtois, C., De Vocht, M., & De Marez, L. (2012). Player identification in online games: Validation of a scale for measuring identification in MMOGs. Media Psychology, 15(2), 197-221. doi: https://doi.org/10.1080/15213269.2012.674917

Waern, A. (2015). ‘I’m in love with someone that doesn’t exist!’ Bleed in the context of a computer game. In J. Enevold, & E. MacCallum-Stewart (Eds.), Game Love, (pp. 25-45). Jefferson, NC: McFarland.

Ward, N. (2008). Mass(ively) effect(ive): Emotional connections, choice, and humanity. Eludamos. Journal for Computer Game Culture, 2(2), 289-292. Retrieved from https://www.eludamos.org/index.php/eludamos/article/view/48/91

Wesley, D., & Barczak, G. (2010). Innovation and marketing in the video game industry. Boston, MA: Gower Publishing Limited.

Wu, S-U., & Hsu, C-P. (2018). Role of authenticity in massively multiplayer online role playing games (MMORPGs): Determinants of virtual item purchase intention. Journal of Business Research, 92, 242-249. doi: https://doi.org/10.1016/j.jbusres.2018.07.035

Zakowski, S. (2014). Time and temporality in the Mass Effect series. A narratological approach. Games and Culture, 9(1), 58-79. doi: https://doi.org/10.1177/1555412013512421

Zekany, E. (2016). “A horrible interspecies awkwardness thing”: (Non)human desire in the Mass Effect universe. Bulletin of Science, Technology & Society, 36(1), 67-77. doi: https://doi.org/10.1177/0270467615624565

Ludography

BioWare. (2009). Dragon Age: Origins. Redwood City, CA: Electronic Arts.

BioWare. (2007). Mass Effect. Redmond, WA: Microsoft Game Studios.

BioWare. (2010). Mass Effect 2. Redwood City, CA: Electronic Arts.

BioWare. (2012). Mass Effect 3. Redwood City, CA: Electronic Arts.