Prior work has argued for interactive storytelling’s power in terms of providing therapeutic benefits Starks et al. (2016), Dias et al. (2018) and enabling learning experiences through educational games Weiß and Müller, (2008), Danilicheva et al. (2009), Melcer et al. (2015), Nguyen et al. (2018), Camingue et al. (2020). Specifically, narrative/storytelling is an important element that can be incorporated into educational games in order to maintain and increase students’ motivation Dickey, (2006), Rowe et al. (2011), Padilla-Zea et al. (2014), with some suggesting that integration of a good story into an educational game will determine its success or failure Göbel et al. (2009). Interactive storytelling has been incorporated into a number of educational games focusing on topics such as history Christopoulos et al. (2011), Song et al. (2012), STEM Danilicheva et al. (2009), Rowe et al. (2011), Weng et al. (2011), Zhang et al. (2019), and bullying Aylett et al. (2005), Watson et al. (2007). However, the majority of research on educational interactive storytelling games has focused on adaptivity Göbel and Mehm, (2013), Kickmeier-Rust et al. (2008), interactivity Song et al. (2012), Zhang et al. (2019), emergent narrative Aylett et al. (2005), player and knowledge modeling Magerko, (2007), Rowe and Lester, (2010), narrative planning and generation Riedl et al. (2008), Hodhod et al. (2011), Wang et al. (2016), Zook et al. (2012), and the game creation process itself Spierling, (2008), Christopoulos et al. (2011), Diez and Melcer (2020). As a result, there is comparatively little work evaluating the impact of an interactive storytelling approach on learning outcomes. This is especially the case for topics such as RCR with ethically complex concepts that require a variety of perspectives, which our previous work sought to address Grasse et al. (2022).

To better understand how interactive narrative game design relates to teaching efficacy, this paper extends our previous work by using SDT to explore players’ motivations for enjoying our RCR training game, Academical, as well as educational interactive narrative games more broadly. Our attempt to explicate this relationship ultimately serves to demonstrate the complexity of tensions between various intrinsic motivations (i.e., ARC factors) and educational interactive narrative game design choices that we intended to facilitate acquisition of learning outcomes. Based on the success of our prior research demonstrating the efficacy of Academical for training RCR learning outcomes Grasse et al. (2022), we posit that our insights from this work can be generalized beyond the scope of research ethics to more broadly inform the design of motivation within interactive digital narratives representing complexity.

Intrinsic motivations, such as those described by Self-Determination Theory, are thought to satisfy basic psychological needs and have been demonstrated to be potent drivers of enjoyment and engagement Tamborini et al. (2010), Peng et al. (2012), Oliver et al. (2016). The following descriptions broadly summarize each factor’s definition (Ryan and Deci, 2017, pgs. 513–17):

1) Autonomy: a sense of control or volition. The very nature of virtual environments removes some real-world constraints and opens up choices that are often unavailable in everyday realities (e.g., through customization); allows players to choose activities and roles from an increasingly large menu; and provides opportunities for action.

SDT also highlights a fourth motivational construct—presence or immersion—that is distinct from the above ARC factors because it is not a source of psychological need satisfaction. Ryan and Deci summarize presence/immersion as “people’s sense that they are, psychologically speaking, within the game world, as opposed to experiencing themselves as agents outside the game, manipulating controls or characters” and clarify that “people perceive and respond to events within a medium as if the medium were not there” (Ryan and Deci, 2017, p.520). Additionally, they state that the capacity for presence/immersion involves “the flow of psychological satisfactions that keep players fluidly and fully engaged within the game world” Ryan et al. (2006). Presence/immersion can certainly arise from conditions such as the realism of a story’s setting and/or characters Smith, (1994), Green and Jenkins, (2014), but these feelings of engagement (i.e., narrative, emotional and physical presence) can be enhanced and better predicted by “ongoing opportunities for meaningful, need-satisfying activities” Ryan et al. (2006).

According to Ryan and Deci, “different types or genres of video games afford distinct profiles of the psychological needs that they can satisfy” Ryan and Deci, (2017). Thus, it is firmly understood that the ARC definitions described by SDT are not relevant to every game. For instance, SDT (and its affiliated Player Experience of Need Satisfaction survey, i.e., PENS) does not officially define how relatedness can contribute to intrinsic motivations to enjoy and/or feel engaged with single-player games (e.g., interactive narratives like Academical). The authors of SDT specifically describe single-player role-playing games (RPGs) as satisfying autonomy and competence, whereas similar games with multiplayer options (e.g., massively multiplayer online RPGs, or MMORPGs) further satisfy the need for relatedness through supportive interactions with other players (Ryan and Deci, 2017, p. 518). However, the theory’s designers admit (and some studies indicate) that players can experience relatedness through interactions with non-playable characters (NPCs; i.e., interactive characters that are not controlled by humans), particularly via “contingent responsiveness” Rigby and Ryan, (2011), Ryan and Deci, (2017), Dechering and Bakkes, (2018). For example, an NPC could foster a sense of relatedness with the player by providing them with aid in the form of an action or advice. Further bolstering this claim, the Ubisoft Perceived Experience Questionnaire (UPEQ), a similar validated game-specific survey based on SDT, amends the concepts covered by the PENS survey to explicitly include assessment of relatedness to NPCs via 3 of its 9 relatedness questions Azadvar and Canossa, (2018). Thus, it is possible for a single-player game (particularly one with a strong narrative and developed characters) to support enjoyment and/or engagement via each of the ARC factors described by SDT, including relatedness.

While the UPEQ survey is able to assess relatedness within single-player narrative game content, other game and media researchers have suggested that the definition of relatedness within this context (i.e., contingent responsiveness) can be further refined and extended Smith, (1994), Tyack and Wyeth, (2017), Adachi et al. (2018), Dechering and Bakkes, (2018). More specifically, this body of work suggests that, under certain conditions, it is possible for players to feel relatedness with the main character of a story (playable or not) in addition to supporting characters or NPCs. Drawing from this body of work, the following section describes the relatedness criteria we included in the final version of the rubric we used to code players’ qualitative answers for the presence of ARC factors.

Relatedness in games is typically interpreted in terms of the player’s relatedness to other players, particularly in a multiplayer context Ryan and Deci, (2017). Here, however, we interpret relatedness primarily in terms of the player’s relatedness to fictional characters in general, including both the player characters that they control and the non-player characters with whom they interact. In this regard we follow both Tyack and Wyeth, (2017), who extend relatedness to include parasocial relationships Rubin and Perse, (1987), Schramm and Hartmann, (2008), Kavli, (2012) with non-player characters, and Bopp et al. (2019), who address player feelings of attachment to both player and non-player characters.

The idea that spectators of fiction understand fictional characters by modeling them as other people is well-supported by the modern cognitive-narratological understanding of character (Schneider, 2001, p. 608), as is the idea that readers can therefore experience a sense of relatedness to fictional characters. Smith, (1994) proposes a structure of sympathy consisting of three distinct levels of “imaginative engagement” with fictional characters: recognition, in which spectators are presented with “legible and consistent” information about characters that lead them to model these characters as internally coherent others; alignment, in which spectators are given access to a character’s subjectivity to facilitate perspective-taking; and allegiance, in which spectators are led to “root for” (or against) characters on the basis of moral evaluation. The first of these three levels is largely connected to the spectator’s sense that the characters are believable, while the latter two levels increasingly address affective relationships with characters of the sorts that we usually discuss under the label of “identification”. However, all three levels involve the application to fictional characters of cognitive functions (such as perspective-taking) that are normally applied to other people.

Because relatedness centers on affective relationships, we code for relatedness any player response that contains evidence of the presence (or conspicuous absence) of either of the following constructs drawn from Smith’s “structure of sympathy”:

1) Alignment: Players being given privileged access to a character’s subjectivity, including their emotions and inner thoughts, and consequently taking the character’s perspective to some extent.

However, we do not code for relatedness (and instead code for presence/immersion) when only recognition is present. This is because players can recognize characters without identifying or sympathizing with them further.

3) Recognition: Players suggesting that characters are psychologically believable, that their actions seem to stem from a “legible and consistent” mental agency Smith, (1994).

Academical is a choice-based interactive narrative game that provides players the opportunity to role-play as various unique characters immersed within a collection of ethical dilemmas that are common in research. The game comprises fundamental ethical topics that inform all aspects of the research process and highlights how this process can be complicated by many factors such as power dynamics and marginalized identities. Each playable scenario in Academical centers on a conversation between two stakeholders in the RCR issue at hand, one of whom is controlled by the player—in the sense that they select dialogue options for that character (see Figure 1). By virtue of these choices, the player will ultimately reach one of several possible endings, a subset of which represent successful navigation of the situation. Upon reaching a good ending for the first character, the player then unlocks the other interlocutor and replays the scenario from that person’s viewpoint. In turn, reaching a good ending for the second character in a given scenario unlocks the next scenario/RCR topic.

This study was approved by the Institutional Review Board at the University of California, Santa Cruz (a Tier 1 research university). All participants were recruited as a convenience sample from an undergraduate course offered through UCSC’s engineering department. Two weeks before the conclusion of the course, participants were informed of the study through email and offered extra credit toward their class grade in exchange for completing all parts of the study. Participants were told that the purpose of the study was to test the efficacy of a new RCR training program.

This study was conducted entirely online. Participants accessed the game and post-game survey using the same methods as prior studies on this RCR training tool Grasse et al. (2022)—through their preferred web browser on their personal computers and without any supervision beyond automated data collection. Two of the nine possible scenarios were selected for students to play through (i.e., peer review and authorship). Participants were instructed to play through each character at least once in each scenario—equating a minimum of 4 total playthroughs (2 characters/playthroughs per module)—before completing the post-survey. After playing Academical, participants responded to short answer questions asking them to describe which aspects of the game they found most and least enjoyable.

Three of the authors independently coded all of the qualitative answers (154 total, 77 answers for each enjoyment question) to determine whether their content described feelings of autonomy, relatedness and/or competence. Participants’ answers could be categorized as including more than one factor (e.g., autonomy and competence). The initial rubric used for coding was based on definitions described in SDT-based game literature Ryan et al. (2006), Przybylski et al. (2010), Dechering and Bakkes, (2018), Azadvar and Canossa, (2018). After using this rubric, the ratings for autonomy and competence showed high agreement (percent agreement = 98.7 and 97.4%, respectively). However, there was still disagreement on what answers qualified as relatedness—particularly for responses that involved enjoyment of “perspective-taking” (a prominent feature of interactive narratives, especially Academical). Using both game- and media-based SDT literature Smith, (1994), Tyack and Wyeth, (2017), Dechering and Bakkes, (2018), the rubric was subsequently refined to clarify how perspective-taking could support feelings of relatedness with any of the game’s characters. The coders then separately re-evaluated the responses for relatedness content, resulting in high agreement (percent agreement = 98.7%). Though presence/immersion is not categorized as an ARC factor, it is a substantial component of the PENS survey and indicated as an expression of a player being motivated to engage with game content. Therefore, we also report presence/immersion results (percent agreement = 100%) by showing the quantity of responses that described recognition of either the characters or setting as being familiar, real or believable Smith, (1994), Ryan et al. (2006); Green and Jenkins (2014). Since most literature discusses need satisfaction, the “most enjoy” rubric served as the reference point for coding the “least enjoy” answers (e.g., a lack of choice was regarded as low autonomy). A summary of coding themes and representative responses can be found in Tables 1 and 2. Table 1 shows the final rubric used for coding SDT-based need satisfaction (most enjoy), while Table 2 shows the final rubric used to identify SDT-based need frustration (least enjoy). Responses that received two or more coders’ votes were included in the motivational factor tallies provided in Table 3, whereas responses with only one vote were excluded.

A total of 99 undergraduate students registered for this study through a participant recruitment website hosted by the university. Within this pool, 77 successfully participated by completing the study requirements. Nine of these participants reported that they had received prior RCR training. Altogether, we analyzed data from 51 males, 23 females and 3 non-binary individuals. The average participant age was 20.7 ± 2.2 years (median = 20, range = 18–29), which is a typical age for university students who are starting to engage in research and consider applying to graduate school. In response to the Likert survey question, participants on average indicated some interest in pursuing a research career (mean = 3.74, SD = 1.7, range = 1–7).

Participants reported pursuing the following undergraduate degrees: Arts and Design: Games and Playable Media (n = 29), Computer Science: Computer Game Design (16), Cognitive Science (16), Computer Science (9), Technology and Information Management (4), Psychology (2), Economics (1), Electrical Engineering (1), Film and Digital Media (1), Physics (1), Politics (1), and Sociology (1). Two participants reported that they had not yet declared a major, while seven reported pursuing two majors. Design-based (30) and engineering (30) degrees were the most common, followed by science (20) and economics/politics (2).

In the following sections, we quantify and summarize the reasons participants gave for their most and least enjoyable aspects of the game. Notably, some participants provided multiple distinct reasons within a response, even for the same ARC factor. In Table 3, we simply tally the number of responses that contain each of the SDT motivational factors. However, the proceeding qualitative breakdown of these answers further delineates the presence of multiple distinct mentions of the same factor, which explains why the sum of these cases may exceed the total number of responses that contain that factor.

This paper has aimed to extend existing conceptualizations for the relatedness factor of SDT so that it can be better understood and applied within the context of single player educational interactive narrative games. Relatedness traditionally requires a minimum of a dyadic relationship characterized by “contingent responsiveness” that allows a person to psychologically thrive, experience well-being, and support the expression of their “true self.” Indeed, Ryan et al. (2006) explicitly posit that “immersion players desire to … engage in role-play and ‘be part of the story’,” which implies that role-play is more closely associated with presence/immersion than relatedness. According to Tyack and Wyeth, (2017), “it remains improbable that relatedness needs can be fully satisfied without any form of direct human contact.” However, various studies (theirs included) have suggested that relatedness can at least be partially satisfied within the context of media and story/character-based games Smith, (1994), Oliver et al. (2016), Tyack and Wyeth (2017), Dechering and Bakkes, (2018), Adachi et al. (2018). This can occur, for instance, when a person forms a parasocial relationship with a celebrity or movie character or when they form a more interactive supportive relationship with an artificial character in a game.

Role-play is an activity common to games that explicitly engages a person’s capacity to experience the world from another’s perspective as well as to experience alignment and allegiance with fictional characters. Many people feel very motivated to engage in these kinds of activities and feelings because it is enjoyable (e.g., reading choose your own adventure books). Ultimately, we determined that previous research supported the idea that: People enjoy role-playing and connecting with fictional characters because it is intrinsically motivating (i.e., via relatedness). This theory indicates that the act of role-playing and/or connecting with fictional characters is able to satisfy the psychological need to support genuine self-expression and feelings of well-being. Green and Jenkins, (2014) explain that role-play, specifically within the context of interactive narratives, induces identification with a character. They further state that “narrative can affect self-perceptions, … temporarily broaden the reader’s perspective, … and expand the self or allow people to explore new possible selves.” The authors then emphasize that the act of perspective-taking “may translate into motivation for real-world behavior change.“ ¹ Interactive narratives are typically designed to support self-expression through the provision of choice on behalf of one or more characters (though our analysis indicates that Academical could be improved in this respect). Additionally, it is possible for educational interactive narratives to encourage feelings of well-being through attitude and behavioral change. Indeed, we previously found that our game is capable of significantly improving attitudes about RCR training Grasse et al. (2021), though we do not know if these changes in attitude were associated with the exploration of possible selves or increased feelings of well-being. Our current research evaluating the reasons why players felt that playing Academical was enjoyable provides some support to the theory that role-play induces feelings of relatedness. Of the 77 participants in our case study, many made a point to explain that they especially liked the perspective-taking (i.e., role-playing) experience that the game provided. That this experiential aspect was such a prominent explanation for the game’s enjoyment, combined with the fact that the majority of the other responses involved other intrinsic motivations, suggests that the role-playing nature of the game may have contributed to SDT-based need satisfaction. However, more research is required to confirm this hypothesis.

This case study had a variety of limitations, especially in its ability to confirm the relationship between role-play and relatedness. This issue could have been addressed by asking the participants to complete a validated SDT-based survey such as the PENS or UPEQ. Furthermore, though our analysis is based on rubrics informed by popular psychological theory (i.e., SDT), other literature suggests that our assessment of presence/immersion may be oversimplified McMahon, (2003), Mestre et al. (2006), Roth and Koenitz (2016). Moving forward, future studies will account for the differences and relationship between presence and immersion using more nuanced assessment tools, such as the Game Experience Questionnaire (GEQ) Johnson et al. (2018) ² or Roth’s measurement toolbox for evaluating the user experience of interactive digital narratives Roth, (2016). Next, we recognize that the participant pool (i.e., undergraduates pursuing mainly design and engineering degrees) is not an ideal target population for this study. For example, the number of negative responses that mentioned feeling unrelatable to the content of the stories could be lower for graduate students and/or undergraduates intent on pursuing a career in science. Additionally, the results we discussed were most relevant in the assessment of our game and therefore may not hold true for other narrative-based games. In particular, educational interactive narrative games are by design more relatable than those for entertainment, and so some of our observations may only be applicable for the former sub-genre. Finally, our data collection methods were designed to only capture information about what made the game most or least enjoyable. Therefore, we were not able to report the prevalence or subjective importance of every SDT-based factor for each participant.

While this study constitutes a valuable starting point, in future studies we want to dive more deeply into determining explanations for why players perceive some game design features to assist enjoyment and engagement more than others (e.g., by conducting semi-structured interviews rather than online surveys). We can use these results to better predict how much impact certain changes will have on improving the quality of the game experience for a wide player audience. Ultimately, an interesting line of research would involve manipulating features of the game that we expect would directly impact specific need satisfactions Tamborini et al. (2010), Peng et al. (2012). In this way, we would be able to confirm whether the act of role-playing supports intrinsic need satisfaction via relatedness.