Empathy in action: cultivating altruism through immersive game experiences

Introduction: Immersive digital games are increasingly explored as tools for fostering empathy and prosocial behavior, yet limited research examines how these effects unfold in virtual reality (VR). This mixed-methods research examined Empathy in Action, a narrative-driven VR game developed for this study that engages players in assisting a distressed child through both physical tasks and moral decision-making.

Methods: Sixty-four adult participants completed pre- and post-surveys measuring empathy, altruism, immersion, and gameplay effects, followed by semi-structured interviews.

Results: Results revealed a significant increase in altruistic attitudes but a significant decrease in self-reported empathy, suggesting that altruism can be reinforced through immersive decision-making even when emotional resonance tapers post-play. Regression analyses indicated that perceived gameplay effect was the strongest predictor of altruistic outcomes, surpassing demographic or gaming background variables. Qualitative reflections supported these findings, with participants linking in-game helping behaviors to real-life relationships and envisioning applications in education, therapy, and rehabilitation.

Discussion: These results complicate assumptions that empathy and altruism rise in tandem, underscoring the role of reinforcement, moral agency, and reflective engagement in prosocial outcomes. The study advances theory by decoupling empathy and altruism within immersive contexts and highlights the broader potential of VR games as tools for cultivating moral action and social change.

1 Introduction

Digital games have increasingly been recognized as tools for promoting social learning and moral engagement. Prosocial games, which emphasize cooperation, helping, and moral decision-making, have been shown to foster empathy, altruism, and teamwork (Gentile et al., 2009; Greitemeyer and Osswald, 2010). Immersive games, in particular, have been integral to studying social and psychological interactions. In this current study, the term immersive games is used to describe digital games whose design features aim to generate a robust sense of presence and foster deeper involvement within a virtual world, resulting in detachment from physical reality (Han et al., 2024; Jennett et al., 2008; Pine and Gilmore, 2001). For example, games such as Game Company (2012), developed by thatgamecompany, encourage nonverbal cooperation (du Bois, 2022); Upper One Games (2014), created by Upper One Games and published by E-Line Media, fosters understanding while reinforcing the importance of community through cultural exploration (Evren, 2025); and LLC (2013), developed by 3909 LLC, challenges players with moral dilemmas that balance loyalty and empathy (Cabellos et al., 2022).

A common thread across these examples is that they are desktop-based games. While desktop prosocial games have demonstrated positive effects on helping behaviors, extended reality (XR) gaming environments such as augmented reality (AR) and virtual reality (VR)—which predominantly use first-person perspectives but can also include third-person formats—offer heightened emotional engagement and embodied interaction that are expected to produce stronger and potentially longer-lasting effects (Gorisse et al., 2017). Further studies suggest that immersion amplifies presence, empathy, and moral engagement (Baños et al., 2004; Fraser et al., 2012). Building on these insights, the current study investigates whether these heightened immersive affordances translate into meaningful prosocial outcomes.

Given this shift toward more immersive formats, the present research examines whether engaging in helping behaviors in an immersive VR game can translate into increased empathy, altruism, and moral decision-making beyond the game environment. While prior studies on desktop or screen-based games show that prosocial content promotes cooperation and helping behaviors (Gentile et al., 2009; Greitemeyer and Osswald, 2010), immersive VR offers deeper emotional and cognitive engagement that may emulate prosocial scenarios that encourage players to translate in-game behaviors into real-world actions (Gorisse et al., 2017).

Beyond empathy and altruism, this study examines whether immersive experiences can also strengthen interpersonal skills, such as perspective-taking, conflict resolution, and moral reasoning. These outcomes carry potential implications for designing educational and therapeutic games, as well as for applications in training programs and conflict resolution.

2 Research context

Research consistently shows that prosocial video games can positively influence real-world behaviors. Gentile et al. (2009) found that prosocial gameplay predicted higher cooperation across cultures, while Greitemeyer and Osswald (2010) showed that prosocial games increased the likelihood of players helping others in real-life situations.

Immersive games are particularly promising for deepening moral engagement. Narrative-driven experiences such as Journey and Papers, Please highlight how unique storytelling can foster empathy, moral reflection, and perspective-taking (Cabellos et al., 2022; du Bois, 2022). Prior AR and VR studies further amplify these effects, demonstrating that increased presence and embodiment heightened emotional responses (Rosenberg et al., 2013), first-person VR enhanced moral engagement (Gorisse et al., 2017), and immersive formats elicited stronger empathetic reactions that may lead to lasting behavioral change (Baños et al., 2004; Fraser et al., 2012). The empathy–altruism model (Batson et al., 2015) further supports this trajectory, positing that empathic concern for someone in need produces an altruistic motivation to help. XR studies (e.g., Herrera et al., 2018) build on this by showing how VR-induced empathy through vivid perspective-taking can translate into altruistic outcomes, such as increased willingness to donate after immersive simulations.

This current study extends these findings by testing whether helping behaviors within an immersive environment can both trigger immediate emotional responses and encourage sustained prosocial attitudes, thereby advancing our understanding of how immersive games can influence empathy and altruism.

2.1 Theoretical frameworks

This research is grounded in three theoretical frameworks: mental models theory (Johnson-Laird, 1983), social cognitive theory (SCT; Bandura, 1986), and prosocial behavior theory (Batson and Shaw, 1991). Together, these frameworks provide a lens for understanding how immersive narratives can shape players’ cognitive representations, reinforce behaviors through social learning, and motivate altruistic or prosocial outcomes.

2.1.1 Mental models theory

Mental models theory (Johnson-Laird, 1983) suggests that people create internal representations—or mental models—of how things work in the world. These models help people reason through problems, make predictions, and decide how to act in specific situations. They do not merely encompass factual knowledge. Rather, these models are dynamic simulations based on experience, allowing individuals to mentally “try out” actions and outcomes before committing to them in real life.

Immersive technologies like VR can provide particularly powerful contexts for mental model development because they allow users to interact with complex systems in dynamic, experiential ways. Rather than simply reading about or observing a process, individuals can actively test hypotheses, manipulate variables, and immediately see the outcomes of their actions in VR. For example, Slezaka et al. (2023) found that participants who actively constructed and manipulated virtual systems developed stronger and more accurate causal mental models than those who only observed or interacted with prebuilt scenarios. This suggests that active engagement in immersive environments supports a deeper understanding of cause–effect relationships. Similarly, Vogt et al. (2021) demonstrated that giving participants elaboration prompts—such as questions that encouraged them to predict outcomes or explain concepts—before entering a VR environment improved the accuracy and coherence of their structural mental models. In other words, encouraging reflection in advance helped individuals better organize and connect information during the immersive task, showing how thoughtful design strategies can shape the quality of mental models formed in XR environments.

In the context of prosocial gameplay, immersive games that require moral decision-making can promote helping behaviors, shape mental models aligned with prosocial values, and support real-world behavior change. For example, in Papers, Please, players act as border agents under authoritarian rule, facing the challenge of balancing official duties with the need to financially sustain their family. Cabellos et al. (2022) analyzed over 1,500 player reviews of the game and found that while most responses were oriented toward pragmatic success in gameplay, a subset of players reflected on deeper moral dimensions. Although fairness was less frequently acknowledged despite its salience in the narrative, concerns for family and immigrants, as well as the authority of the state, were more commonly acknowledged (Cabellos et al., 2022). In this context, mental models help players simulate reality so that they can more deeply reflect on complex social, emotional, and/or practical decisions. The findings also suggest that with additional scaffolding—such as explicit prompts or guided reflection—games like Papers, Please could more strongly activate epistemic goals and reinforce moral learning (Cabellos et al., 2022), further shaping prosocial mental models.

2.1.2 Social cognitive theory (SCT)

To complement mental models theory, social cognitive theory (SCT; Bandura, 1986) explains how people learn behaviors through observation, imitation, and reinforcement, especially in social contexts. It emphasizes that people do not learn solely through direct experience; they also learn by watching others and mentally rehearsing outcomes. As a result, SCT suggests that people are more likely to repeat behaviors they have seen modeled, found rewarding, and believed they can perform. Immersive games are ideal for reinforcing these altruistic actions by providing realistic feedback and immediate rewards for prosocial behavior.

Evidence from immersive environments supports these claims. For example, Mousavi et al. (2023) found that VR safety training improved university students’ knowledge and self-efficacy, with self-efficacy emerging as a key predictor of learning gains. Their findings highlight how immersive training environments can reinforce confidence and performance, thereby enhancing learning outcomes. Similarly, VR-based social cognition training for children with autism improved emotion recognition and social attribution by enabling participants to rehearse interactions with real-time feedback, demonstrating how modeled behavior and reinforcement in VR can translate to real life (Didehbani et al., 2016). Teacher trainees exposed to VR classroom simulations also reported higher self-efficacy, consistent with SCT’s principle that mastery experiences and observational learning strengthen confidence and motivation (Nissim and Weissblueth, 2017). Literature reviews further highlight that VR and AR in collaborative learning spaces naturally align with SCT, as they foster observational learning and social reinforcement—allowing individuals to observe and validate each other’s behaviors in immersive, socially interactive environments (Scavarelli et al., 2021).

In games like Never Alone, players cooperate with companions to overcome challenges, fortifying moral agency and cooperation (Evren, 2025). Never Alone is a puzzle platformer developed in collaboration with the Iñupiat people, where players guide a young girl named Nuna and her arctic fox companion through the tundra to stop an eternal blizzard. The game blends cooperative gameplay with the reward of short cultural insight videos, offering both storytelling and education about Alaskan Native traditions. Research further shows that Never Alone functions as more than entertainment, preserving and transmitting Iñupiat culture through authentic narratives and collaborative design (Evren, 2025). This alignment between gameplay and cultural storytelling provides players with modeled examples of cooperation, care, and cultural respect, which are social behaviors that SCT predicts can be internalized and reproduced in real life.

Building on this idea, Starks (2014) introduced cognitive behavioral game design (CBGD), a framework that combines SCT with multiple intelligences theory and practical game design strategies. The purpose of CBGD is to guide the creation of serious games that both educate and motivate players toward positive behavior change. Drawing from Bandura’s (1986) work, Starks emphasizes five SCT elements—knowledge, goals, outcome expectations, facilitation/social support (termed “encouragement” in CBGD; Starks, 2014, p. 3), and barriers—as critical to designing games that both teach and motivate. For example, self-efficacy is strengthened when players observe modeled behaviors, overcome barriers, and experience mastery in-game, which increases the likelihood of repeating those behaviors outside of gameplay (Starks, 2014).

Taken together, these examples demonstrate how SCT can be applied not only to explain why players feel more capable and motivated after cooperative play but also to guide intentional design choices that foster lasting behavioral change in immersive contexts.

2.1.3 Prosocial behavior theory

Building on SCT’s focus on learning and reinforcement, prosocial behavior theory shifts attention to the underlying emotional and motivational processes that drive altruistic action. Prosocial behavior theory emphasizes empathy, perspective-taking, and moral elevation as key drivers of altruistic behavior (Batson and Shaw, 1991). It explores the conditions that drive people to engage in actions that benefit others, such as helping, comforting, donating, or cooperating.

These mechanisms are particularly visible in games that prioritize connection over competition, where immersive design elements draw players into emotionally resonant interactions. For example, Journey fosters empathy by allowing players to assist strangers, creating shared experiences and moral elevation (du Bois, 2022). Journey is a wordless adventure game where players guide a robed traveler on a pilgrimage to a distant mountain, encountering ruins of a fallen civilization and cooperating with anonymous players along the way. Its design emphasizes peaceful, nonviolent interaction with multiplayer mechanics that encourage players to support each other through small acts of kindness—such as recharging one another’s cloaks or guiding each other through obstacles. This absence of competition and reliance on nonverbal communication highlights interconnectivity, understanding, and collaboration as the core of the experience (du Bois, 2022). Moreover, the game’s symbolic narrative portrays life’s struggles, mortality, and eventual renewal. Overall, Journey demonstrates how narrative and design can foster empathy and perspective-taking, reinforcing prosocial behavior theory’s focus on the emotional and cognitive processes that inspire helping behaviors.

Beyond traditional games, VR research has further demonstrated how immersive environments can activate prosocial mechanisms. Rosenberg et al. (2013) found that embodying the superpower of flight in VR led participants to engage in immediate helping behaviors afterward, suggesting that immersive role-play can prime superhero-related concepts and encourage altruistic action. Building on this, Herrera et al. (2018) showed that VR perspective-taking produced more positive, longer-lasting attitudes toward the homeless and greater willingness to support them (e.g., petition signing), highlighting VR’s potential to extend empathic engagement into sustained prosocial outcomes. Peck et al. (2013) demonstrated that embodying a Black avatar in VR significantly reduced implicit racial bias, indicating that immersive perspective-taking can reshape intergroup attitudes and potentially foster more prosocial orientations. Finally, Ahn et al. (2014) reported that participants who cut down a virtual tree in VR later engaged in more conservation behaviors—such as using fewer napkins and reporting greater recycling behaviors over time—demonstrating that embodied perspective-taking can translate moral concern into concrete, real-world prosocial actions. Collectively, these studies show that immersive design can move beyond empathy and perspective-taking alone, producing observable prosocial behaviors that directly support prosocial behavior theory.

The present study builds on this work by conceptualizing altruism as attitudinal and intention-based prosocial motivation—that is, self-reported willingness to help and prosocial orientations—rather than observed behavioral outcomes. More broadly, this trajectory of prosocial game research also extends to intervention studies. Saleme et al. (2020) conducted a systematic review of 11 interventions designed to promote prosocial skills and behaviors in children and adolescents. Results were mixed but promising: several studies reported increases in empathy, social skills, and positive bystander behavior, while others showed more modest effects. Importantly, observed prosocial behaviors—such as cooperation and adaptive social interaction—were most likely to improve when games incorporated opportunities for practice and feedback in real-life scenarios. This aligns with prosocial behavior theory, which emphasizes that empathy and perspective-taking must translate into actual helping behaviors to generate meaningful change. By highlighting both the promise and the limitations of current interventions, Saleme et al. (2020) underscore the need for games that deliberately integrate prosocial design elements, supporting the theory’s claim that empathy-driven play can foster altruism beyond the screen.

2.1.4 Research questions and hypothesis

These frameworks provide a comprehensive view of how helping-oriented activities in immersive games can influence empathy and altruism. Mental models theory explains how players’ cognitive structures evolve through moral decisions; SCT highlights how positive social behaviors are reinforced; and prosocial behavior theory focuses on the emotional drivers of helping behaviors. This integration explores both behavioral changes and underlying mechanisms, showing how immersive environments foster altruistic engagement within and beyond the gaming world. Through these frameworks, this study addresses the following research questions (RQs):

• RQ1: How do prosocial behaviors in immersive game environments influence players’ level of empathy and sense of altruism?

• RQ2: What are the effects of engaging in helping activities within immersive games on prosocial attitudes and moral decision-making?

Drawing on these contexts and prior works (e.g., Ahn et al., 2014; Batson et al., 2015; Herrera et al., 2018; Rosenberg et al., 2013), which posit that empathic concern motivates altruistic action, the researchers hypothesize that participants will exhibit a statistically significant increase in empathy and altruism after engaging in the immersive VR game designed for this study.

3 Materials and methods

3.1 Game design: empathy in action

Contributing to prior works demonstrating that immersive games foster deeper moral engagement through unique narratives (Cabellos et al., 2022; du Bois, 2022), and building on studies that observed stronger empathetic responses in immersive XR formats (Baños et al., 2004; Gorisse et al., 2017), the researchers and game development collaborators used the Unity game engine to create an immersive experience titled Empathy in Action. Visual stills from the final game environment are shown in Figure 1.

Figure 1

Figure 1. Visual stills from the Empathy in Action VR game environment. These scenes from the completed VR environment show the settings where players accompany characters in the game through narrative-driven tasks. (A) An aerial view of Unity Springs, the fictional community where the game takes place. (B) The main entrance of the Unity Springs community center where the game begins. (C) A street view of the neighborhood. (D) The community park entrance where players encounter a character in need. (E) An aerial view of the lake where players complete both physical and moral decision-making tasks. (F) An aerial view of the playground where players search for clues.

3.1.1 Mechanics

Best described as a narrative-driven VR game, Empathy in Action was designed with a strong emphasis on emotional engagement and understanding others. It combines the interactivity and structure of games with the immersive storytelling of virtual experiences. The game aims to examine how immersive environments influence empathy, altruistic decision-making, and a player’s sense of agency. To achieve this, Empathy in Action balances physical tasks with moral decision-making, allowing players to experience relatable scenarios and reflect on the outcomes of their actions. Research shows that both physical and non-physical acts of helping can produce meaningful prosocial outcomes, highlighting the importance of moral and emotional heroism (Peña and Chen, 2017; Rosenberg et al., 2013). Furthermore, by emphasizing social heroism—moral dilemmas involving empathy and justice rather than physical danger—this study takes a unique approach, expanding on previous research focused primarily on physical rescue scenarios (Gillath et al., 2008).

The game was deliberately designed with an on-rails (guided, restricted-path) narrative structure, where player movement and choices were directed to keep the story moving forward and allow completion of the experience within approximately 15 min. This design reflected consideration for study participants’ time while also guaranteeing exposure to all tasks and narrative elements in the game.

3.1.1.1 Narrative-driven VR game

Empathy in Action takes place in the fictional community of Unity Springs. Upon entering the immersive space, players are introduced to Sage, an NPC (non-player character) community guide whose name—meaning “wise”—reflects the character’s role. At the beginning of the game, Sage provides a short tutorial to help players acclimate to the controls. Players are then tasked with helping another NPC, a young boy named Alden (meaning “old friend”), reunite with his lost dog, Buddy (a straightforward symbol of companionship). Throughout the experience, players travel alongside Alden, completing both physical tasks (e.g., searching for clues) and moral decision-making tasks (e.g., choosing whether to comfort Alden through a breathing exercise or by discussing a plan to find Buddy), as illustrated in Figure 2. Alden also shares his emotional backstory with Buddy, and Sage returns at the game’s conclusion to reflect on the importance of empathy, kindness, and helping.

Figure 2

Figure 2. Screenshot of moral decision-making in the Empathy in Action narrative VR game. This figure illustrates a dialogue choice where players select between different options to support an upset Alden, the little boy who lost his dog.

Through Empathy in Action, participants actively embody a community volunteer, a job formally assigned by Sage during the introductory tutorial. In this role, participants complete tasks such as comforting a distressed child, resolving interpersonal conflict, and making moral decisions under emotional pressure. Moreover, the game’s structure is intentionally grounded in the theoretical frameworks guiding this study.

First, because the game simulates emotionally and ethically complex situations in a realistic and embodied way, it enables players to form new or strengthened mental models (Johnson-Laird, 1983) of what it means to be an empathetic, morally engaged person. These mental models can influence future real-world behavior by making altruistic responses feel more familiar, practiced, and satisfying.

Second, social cognitive theory (SCT; Bandura, 1986) explains why players may not only feel good during the game but also leave with a heightened sense of efficacy and motivation to help others in real life due to having already succeeded in a simulated social situation. Unlike passive storytelling, Empathy in Action actively involves players in observing, practicing, and receiving gratitude for altruistic behavior.

Lastly, the game emphasizes key psychological processes underlying prosocial action, including empathetic concern, perspective-taking, moral elevation, and internalized values such as moral identity or a sense of duty. Prosocial behavior theory (Batson and Shaw, 1991) helps account for both the emotional engagement during gameplay and the potential long-term effects, such as sustained motivation to help others even in the absence of external rewards.

3.2 Mixed methods

A mixed-methods approach was employed to explore RQ1 and RQ2. Participants were recruited through digital flyers and word-of-mouth outreach within a local university community. The study sample included 64 adult volunteers (ages 18+) residing in the U.S., chosen for practical reasons such as ease of recruitment, proximity, and consistent cultural context. This sample size, determined by G*Power analysis for a medium effect size (Cohen’s d = 0.5), aligns with previous research showing moderate effects of immersive prosocial games on behavior (Greitemeyer and Osswald, 2010). Studies on XR-based interventions tend to report medium effect sizes for positive learning and attitudinal outcomes, making a sample of 64 participants both statistically and logistically appropriate (Chang et al., 2022; Garzón and Acevedo, 2019; Garzón et al., 2019; Garzón et al., 2019; Li et al., 2023).

Participants completed a pre-survey that collected demographic information and assessed baseline empathy, altruism, and gaming behaviors, primarily using previously validated scales such as the Interpersonal Reactivity Index (IRI; Davis, 1980) and the Prosocial Tendencies Measure–Revised (PTM-R; Carlo and Randall, 2002). They then engaged in Empathy in Action, the custom-designed VR game described earlier in Section 3.1.1.1, to fulfill a series of helping scenarios.

After gameplay, participants completed a post-survey to measure changes in empathy and altruism, followed by a semi-structured interview exploring their moral reflections, decision-making processes, emotional responses, and gameplay effects and mechanics. After completing the entire study, which generally lasted 30–45 min and did not exceed 1 h, participants received $40 in cash as compensation for their time.

This pre–post structure is necessary given that the research questions and hypothesis focus on changes in empathy and altruism following immersive gameplay. Consistent with prior VR empathy studies (Ahn et al., 2014; Herrera et al., 2018; Rosenberg et al., 2013), the pre-survey assessed baseline tendencies, while the post-survey captured state-level emotional and altruistic responses specific to the game narrative. To minimize familiarity or repetition bias, different items were used for the pre- and post-measures, and the post-survey focused on the characteristics and situations encountered during gameplay.

3.2.1 Pre-survey measures

Prior to gameplay, participants completed a pre-survey designed to establish baseline measures of empathy and prosocial tendencies. Three validated scales, adapted for the context of immersive gameplay, were used to capture constructs central to the study’s theoretical frameworks.

3.2.1.1 Interpersonal reactivity index (IRI)

An adapted 6-item version of the Interpersonal Reactivity Index (IRI; Davis, 1980) was used to assess cognitive and affective empathy. The scale included items from the fantasy, perspective-taking, and empathic concern subscales, each measured on a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). The adapted IRI demonstrated acceptable internal consistency (Cronbach’s α = 0.744).

3.2.1.2 Basic empathy scale (BES)

A 5-item version of the Basic Empathy Scale (BES; Jolliffe and Farrington, 2006) was included to measure participants’ self-reported empathic tendencies. Items were selected prior to data collection based on conceptual relevance to the study. While the adapted scale demonstrated low internal consistency overall (α = 0.571), the researchers noted that two items reflected cognitive empathy (i.e., understanding another’s perspective) and two reflected affective empathy (i.e., sharing another’s emotions) (Davis, 1983). However, these groupings were not formal subscales and both showed low reliability (cognitive α = 0.330; affective α = 0.048). The adapted BES composite score was not used, but the individual items were analyzed separately to capture exploratory dimensions between cognitive and affective empathy, thereby offering greater depth to the interpretation of empathy scores.

3.2.1.3 Prosocial tendencies measure–revised (PTM-R)

An adapted 15-item version of the Prosocial Tendencies Measure–Revised (PTM-R; Carlo and Randall, 2002) was used to assess prosocial tendencies across multiple domains, including public, anonymous, emotional, compliant, and crisis-oriented helping. Two items from the Public Helping subscale were removed post hoc due to negative item-total correlations. The final 13-item scale demonstrated acceptable internal consistency (α = 0.716).

3.2.2 Post-survey measures

After gameplay, participants completed a post-survey assessing empathy, altruism, perceived gameplay effects, immersive experience, and perceived impact. These measures were designed to capture immediate outcomes of the immersive intervention and to complement the baseline data from the pre-survey.

3.2.2.1 Empathy levels post-gameplay

Nine items developed for this study assessed players’ emotional connection, perspective-taking, and affective reactions toward the central character, Alden. Sample items included, “While playing the game, I imagined how I would feel if the events were happening to me or someone I care about,” and “When Alden was upset about losing Buddy, I felt touched by his situation.” Internal consistency for the scale was high (Cronbach’s α = 0.863).

3.2.2.2 Altruism levels post-gameplay

Seven items measured altruistic motivation and self-reported helping behavior following gameplay. Items assessed both in-game behavior and real-world intention (e.g., “Since playing the game, I feel more inclined to offer help even when no one is watching.“). Internal consistency was high (α = 0.860).

3.2.2.3 Gameplay effects: measuring change

Eight items assessed players’ perceived changes in empathy, motivation to help, and reflection on physical and moral decision-making during gameplay. This scale demonstrated excellent internal consistency (α = 0.901).

3.2.2.4 Immersive experience questionnaire (IEQ–Adapted)

Six items from the Immersive Experience Questionnaire (IEQ–Adapted; Jennett et al., 2008) assessed the degree of immersive presence experienced during gameplay. Sample items included, “While playing the game, I felt like I was actually present in the game’s world,” and “I felt emotionally involved in the game’s storyline.” The adapted IEQ demonstrated acceptable reliability (α = 0.832). In line with this study’s usage of the term immersive game, these items focused on players’ subjective sense of presence and emotional involvement.

3.2.2.5 Perceived impact of game

Two items developed for this study assessed players’ perceptions of the educational and emotional impact of immersive gameplay (e.g., “I believe virtual reality (VR) games like this one can help people become more empathetic.“). This brief scale demonstrated excellent internal consistency (α = 0.912).

3.2.3 Semi-structured interviews

To complement quantitative findings from the pre- and post-surveys, researchers conducted brief semi-structured interviews with participants shortly after gameplay. Due to the interviews’ concise and targeted nature, a descriptive content analysis was employed to summarize participant reflections in a structured, theory-informed way (Guest et al., 2011; Sandelowski, 2000). The codebook (see Table 1) was developed a priori, drawing on the study’s theoretical framework and survey instruments, which reduced ambiguity in interpretation and supported coding consistency.

Table 1

Table 1. Interview analysis codebook. This table displays the codes, descriptions, example indicators, and linked theoretical constructs used to guide the descriptive content analysis of participants’ post-game interview responses. The codebook reflects both predefined categories (derived from prosocial behavior theory, social cognitive theory, and mental models theory) and emergent themes identified during analysis.

The coding framework was grounded in the study’s conceptual foundations—mental models theory, social cognitive theory, and prosocial behavior theory—and aligned with constructs measured across instruments (e.g., empathy, decision-making, and emotional engagement). This approach enabled researchers to efficiently organize responses around predefined research questions and variables, rather than relying on more interpretive or inductive strategies typically suited for narrative-rich data.

Select illustrative quotes were also incorporated to capture perspectives that might not have been fully conveyed through surveys. This hybrid strategy supported both the identification of response patterns and the discovery of meaningful participant insights, enriching understanding of the emotional and cognitive effects of the gameplay experience. The analytic approach was well-matched to the study’s mixed-methods design, providing grounded findings without overextending interpretive claims.

The 64 semi-structured interviews were divided between two researchers for qualitative coding. Given the exploratory scope and manageable dataset, a single-coder approach was applied, consistent with recommendations for small-scale qualitative studies (Braun and Clarke, 2021; Busetto et al., 2020). While multiple coders can strengthen inter-rater reliability in large-scale projects, the goals of this study centered on identifying illustrative quotes, reinforcing survey trends, and highlighting reflections relevant for future game design and research. Accordingly, theoretical alignment and consistency were prioritized over inter-rater quantification.

4 Results

4.1 Prototype usability testing

Prior to launching the official study, researchers conducted a prototype usability test with 14 participants at a public game night event. Usability data were collected from 13 participants in the pilot phase (a group distinct from the official study sample, but representing 20.31% of its size, N = 64) through semi-structured post-game interviews. The purpose of this pilot phase was to evaluate the functionality and playability of an early iteration of Empathy in Action and gather initial feedback on mechanics, immersion, and narrative flow.

Despite being a bare-bones version of the game, participants responded positively overall. On average, players rated their interest in playing the full version of the game at 3.77 out of 5. Interview feedback highlighted several strengths, including the engaging “lost dog” scenario and the game’s potential to foster empathy through embodied decision-making. At the same time, participants noted areas for improvement, such as navigation mechanics (preferring the option of controllers over hands-free teleportation), enhancing NPC emotional expressions, and incorporating both audio and text dialogue to improve immersion.

Findings from this preliminary usability phase informed revisions to the final version of the game, including enriched environmental detail, clearer guidance from the community guide NPC, and greater accessibility considerations.

4.2 Descriptive statistics

Among the official sample (N = 64), 33 participants identified as female, 25 as male, and six as non-binary. In terms of generational demographics, one participant identified as Generation X, 13 as Millennial, and 50 as members of Generation Z. Participants represented a range of racial and ethnic backgrounds: 68.75% identified as White or Caucasian, 21.88% as Latino or Hispanic, 17.19% as Asian, 4.69% as Black or African American, 4.69% as Other or Unknown, 3.13% as Native Hawaiian or Pacific Islander, and 1.56% as Native American. This distribution ensured representation across multiple identity groups, reflecting an assorted participant pool.

Gaming habits varied throughout the group: 41 participants reported playing video games frequently or very frequently (i.e., at least once a week), 10 identified as occasional players (one to three times per month), eight as rare players (less than once per month), and five indicated that they never play video games.

Experience with VR was also diverse. Only two participants reported using VR devices regularly (at least once per week), and nine were occasional users (one to three times per month). The majority (N = 37) had only used VR once or twice prior to participating, and 16 had never experienced VR before the study. Prior to gameplay, only 10 participants reported feeling comfortable or very comfortable navigating VR or XR environments. This mix of seasoned gamers and VR novices provided a broad spectrum of perspectives on the immersive gameplay experience.

Descriptive statistics were calculated for all scale variables. On a 5-point Likert scale where 1 = strongly agree and 5 = strongly disagree, lower scores indicate greater agreement with statements reflecting empathy, altruism, and related constructs. Participants reported relatively high baseline levels of empathy (M = 1.76, SD = 0.36) and moderate-to-high levels of altruism (M = 2.46, SD = 0.41). Post-game scores indicated a shift in both constructs, with post-game empathy increasing numerically to M = 2.17 (SD = 0.67) and post-game altruism decreasing to M = 2.11 (SD = 0.66). These results suggest that participants agreed more strongly with altruism-related statements after the game, but agreed less strongly with empathy-related statements post-game.

4.3 Group comparison analysis

To explore whether demographic and gaming experience variables were associated with differences in post-game responses, a series of one-way ANOVAs were conducted across multiple domains of the post-survey, an approach that is appropriate for comparing mean differences in continuous outcomes across categorical groups. These domains included self-reported empathy, altruism, perceived gameplay effects, immersive presence, and perceived impact of the game. Although only empathy, perceived gameplay effects, and perceived impact yielded significant differences, the analyses considered the full range of post-game measures.

4.3.1 Education level

Participants’ post-game empathy scores significantly differed by highest level of education completed, F (4, 59) = 2.64, p = 0.043. Post hoc comparisons using Tukey’s HSD (Honestly Significant Difference) indicated that participants who completed trade or vocational school (N = 3, M = 1.48, SD = 0.21) reported significantly greater agreement with empathy-related statements compared to those with a bachelor’s degree (N = 14, M = 2.47, SD = 0.98). The effect size was moderate (η² = 0.152). Given the uneven subgroup sizes and exploratory nature of the analysis, these findings should be interpreted cautiously. Rather than reflecting a systematic trend across education levels, the result represents a single significant pairwise comparison.

Similarly, perceived gameplay effects differed significantly across education levels, F (4, 59) = 2.79, p = 0.034. Participants with a bachelor’s degree (N = 14) reported the highest gameplay effect scores (M = 2.71, SD = 0.81), while those with vocational training (N = 3, M = 2.00, SD = 0.0) and master’s degrees (N = 13, M = 1.54, SD = 0.69) reported lower levels. The effect size was moderate (η² = 0.159). As with the empathy findings, these differences arise from exploratory post hoc comparisons and should not be interpreted as a consistent ordinal pattern across educational attainment. Variability may be influenced by small and imbalanced subgroup sizes, and future research with larger samples is needed to examine these patterns more robustly.

4.3.2 Simulation game frequency

There was also a significant group difference in perceived impact of the game based on how often participants played simulation games, F (4, 54) = 2.59, p = 0.047. Non-players (N = 18, M = 2.25, SD = 0.84) perceived the game as having more impact than those who played simulation games occasionally (N = 10, M = 1.45, SD = 0.52), who reported the lowest impact scores overall. This may suggest that players with less exposure to simulation-based mechanics found the experience more novel, emotionally engaging, or instructive. The effect size was moderate (η² = 0.161).

4.4 Paired samples t-tests

To examine whether changes in empathy and altruism from pre-to post-gameplay were statistically significant, paired-samples t-tests were conducted. For empathy, scores significantly increased numerically from pre- (M = 1.76, SD = 0.36) to post-game (M = 2.17, SD = 0.67), t (63) = −5.01, p < 0.001, indicating a statistically significant decrease in self-reported empathy. The effect size was medium-to-large (Cohen’s d = 0.66). For altruism, scores significantly decreased from pre- (M = 2.46, SD = 0.41) to post-game (M = 2.11, SD = 0.66), t (63) = 4.10, p < 0.001, indicating a statistically significant increase in altruistic attitudes. The effect size was medium (Cohen’s d = 0.51). Average participant scores on empathy and altruism before and after gameplay are shown in Figure 3.

Figure 3

Figure 3. Pre-game vs. post-game scores for empathy and altruism. This bar graph shows average participant scores on empathy and altruism before and after playing the immersive VR game. Lower scores indicate stronger agreement with statements reflecting empathy and altruism. After gameplay, average empathy scores increased slightly (indicating less agreement with empathy-related statements), while altruism scores decreased (indicating stronger agreement with altruism-related statements). Error bars represent standard deviations. These results suggest that while participants became slightly less likely to describe themselves as emotionally empathetic, they were more likely to endorse helping behaviors after playing the game.

These results further highlight a divergence between empathy and altruism: while participants reported feeling less emotionally resonant after gameplay, they simultaneously expressed stronger prosocial orientations. This pattern suggests that the game may have encouraged participants to reflect on helping as a moral or behavioral imperative, even as their self-perceived emotional resonance diminished.

The semi-structured interview results reiterate these findings. One participant related the game’s lost dog scenario to the pet-populated neighborhood they live in, saying they could apply their in-game helping behaviors to real-life: “It may not be my animal—I don’t have an emotional tie to this [situation]—but I could appreciate that if I were in this situation, I would want somebody else to help me.” Reflections like this one reinforce the survey results, suggesting that the game may have left participants with a heightened sense of altruistic motivation, even if it was not an emotionally resonant interaction.

4.5 Exploratory cognitive vs. affective empathy comparison

To explore the nature of participants’ empathic responses more closely, two cognitive empathy items (e.g., “I can often understand how people are feeling even before they tell me.“) and two affective empathy items (e.g., “After being with a friend who is sad, I usually feel sad myself.“) from the adapted Basic Empathy Scale (BES; Jolliffe and Farrington, 2006) were grouped into subscales. Cognitive empathy reflects the ability to recognize and understand another person’s emotions, whereas affective empathy reflects the extent to which one shares or feels those emotions (Davis, 1983; Jolliffe and Farrington, 2006). Although the BES composite was excluded due to low internal consistency, and both the cognitive and affective empathy subscales showed similarly low reliability, the researchers retained the individual items because they provided important distinctions between cognitive and affective empathy and added nuance to the overall empathy scores. A paired-samples t-test revealed that participants reported significantly greater agreement with cognitive empathy items (M = 1.03, SD = 0.31) than with affective empathy items (M = 1.16, SD = 0.43), t (63) = 2.52, p = 0.014, with a medium effect size (Cohen’s d = 0.31). The two subscales were moderately correlated (r = 0.35, p = 0.005).

These results suggest that participants felt more confident in understanding how others feel (cognitive empathy) than in emotionally sharing those feelings themselves (affective empathy). This pattern offers insight into why altruism increased despite a decrease in empathy scores: participants appeared to rely on reflective, understanding-based processing rather than emotional mirroring during and after gameplay. Interview data support this interpretation: several participants described becoming more self-critical of their emotional resonance while still reporting motivation to help. For instance, one participant reflected, “I’m more of a reactionary person, and having to listen to Alden talk about the history of the dog, I just kind of wanted to go find the dog. I realized that’s something that I probably need to work on in my own personal life.” Further sentiments illustrate that participants engaged in cognitive empathy over affective empathy. For example, one participant noted that they did not directly share Alden’s emotions, however, they did indicate an intellectual understanding of Alden’s feelings and were motivated to help.

Others described a tapering of affective engagement during Sage’s final reflection on the importance of empathy and kindness—some zoning out, some noting it had little effect because empathy was already central to their identity, and others saying it still shaped how they think about connecting with others despite reduced attentiveness once Buddy had been returned to his owner, Alden. These reflections suggest that cognitive empathy and moral reflection persisted even as affective engagement declined, reinforcing the idea that altruistic motivation can be maintained without sustained emotional immersion.

This exploratory breakdown of empathy types offers a more detailed lens on participants’ emotional responses and helps contextualize how these patterns may relate to other aspects of the gameplay experience.

4.6 Pearson correlation analyses

The researchers conducted Pearson correlation analyses to examine whether one variable (e.g., empathy) tended to move together with another (e.g., altruism), essentially checking how strongly these variables were related after gameplay. Correlation analyses revealed several strong and statistically significant relationships among the post-game variables (see Figure 4). Participants who reported feeling more empathetic after the game also reported being more altruistic (r = 0.811, p < 0.001), suggesting that these two qualities often went hand in hand. In addition, participants who said the game had a meaningful effect on them personally—for example, by making them think or feel differently—also tended to report higher levels of both empathy (r = 0.835, p < 0.001) and altruism (r = 0.846, p < 0.001). During the semi-structured interviews, one participate noted: “I think [the game] definitely enhanced my view in the sense of when you see someone who needs help or is struggling. You want to do your best effort to bring that help to them, provide comfort, and help them achieve what they’re trying to do.” This reflection highlights how engaging in a narrative VR game can enhance a person’s emotional resonance and willingness to help those in need.

Figure 4

Figure 4. Correlation matrix of post-game measures. This heatmap displays Pearson correlations between post-game self-report measures. Darker shades represent stronger positive relationships. All correlations were statistically significant at the p < 0.01 level. In particular, post-game empathy and altruism were strongly correlated and both were closely linked with participants’ perceived gameplay effect and sense of immersion.

Relatedly, participants who felt more immersed in the virtual environment also reported higher empathy (r = 0.556, p < 0.001) and altruism (r = 0.568, p < 0.001). This means that those who felt truly “present” in the game were more likely to feel emotionally connected and motivated to help others. Additional interview data support these findings. When asked if they felt immersed in the game’s world and emotionally connected to the story, one participant replied, “I didn’t feel distracted or anything like that from outside things. It was just so real that my emotions just kind of took over.” A different participant said, “It’s hard to be bored or drawn away from the story when you’re actually experiencing it … I think I was actually able to feel connected and be inclined to think critically about Alden and his situation because I feel like I was really in that world with him.”

Finally, those who believed the game had a broader social impact—such as teaching empathy or encouraging prosocial behavior—were also more likely to report higher levels of empathy (r = 0.468, p < 0.001), altruism (r = 0.525, p < 0.001), immersion (r = 0.371, p = 0.003), and personal gameplay effect (r = 0.612, p < 0.001).

These correlations indicate that participants who reported stronger empathy and altruism after gameplay also tended to feel more immersed and personally affected by the experience. The strongest relationships were observed between empathy, altruism, and perceived gameplay effect, suggesting that participants who felt personally changed by the experience also felt more likely to adopt prosocial attitudes. In one of the interviews, a participant expressed that a prosocial scenario like the one depicted in the game “could happen to [them] tomorrow,” emphasizing the parallels between the game narrative and reality. In addition, if a similar situation were to occur in real life, the participant expressed confidence in their actions. Specifically, they noted, “Now I feel like I know exactly what I would do.”

4.7 Predictive modeling (regression analyses)

As part of the effort to understand what drives altruistic behavior in immersive contexts, three multiple linear regression models were tested to examine the predictive influence of demographic characteristics, gaming habits, and gameplay-related experiences.

The first model examined demographic variables, including age, gender, education level, political affiliation, and political view. This model was not statistically significant, F (5, 58) = 0.88, p = 0.502, and explained only 7% of the variance in post-game altruism (R² = 0.07). None of the demographic predictors were individually significant, suggesting that background characteristics did not meaningfully influence altruistic responses.

The second model assessed participants’ gaming behavior, including frequency of gameplay across multiple genres (e.g., simulation, narrative, VR) and VR comfortability. This model was also not statistically significant, F (9, 49) = 0.60, p = 0.791, R² = 0.099, indicating that participants’ gaming habits did not predict how altruistic they felt after playing the game.

In contrast, a third model using scale-based predictors—such as pre-game empathy, pre-game altruism, post-game empathy, gameplay effect, immersion, and perceived impact—was statistically significant and accounted for a substantial portion of the variance in altruism, F (6, 57) = 33.80, p < 0.001, R² = 0.781. Among these predictors, perceived gameplay effect emerged as the strongest and most reliable predictor of post-game altruism (β = 0.589, p < 0.001). Baseline empathy also significantly predicted altruism (β = 0.179, p = 0.031), while post-game empathy approached significance (β = 0.215, p = 0.096). The other scale-based predictors—baseline altruism (β = −0.036, p = 0.653), immersion (β = 0.104, p = 0.181), and perceived impact (β = 0.013, p = 0.871)—were not significant. These findings suggest that the extent to which participants felt personally impacted by the game was a key factor in predicting their post hoc altruistic responses—more so than prior gaming experience or demographic background. Figure 5 demonstrates the relative strength of each scale-based predictor in influencing post-game altruism.

Figure 5

Figure 5. Standardized regression coefficients for predicting post-game altruism. This horizontal bar graph displays standardized regression coefficients (β), which show the relative strength of predictors of post-game altruism. Higher beta values indicate stronger predictors. Perceived gameplay effect was the strongest and most statistically significant predictor of altruistic responses after gameplay (β = 0.589, p < 0.001). Pre-game empathy (β = 0.179, p = 0.031) also contributed meaningfully. Other variables, including post-game empathy, pre-game altruism, immersion, and perceived impact, were not statistically significant predictors. Darker bars indicate predictors that reached or approached statistical significance. This figure highlights which aspects of the immersive experience were most influential in shaping altruistic attitudes.

4.8 Additional insights from post-game interviews

During post-game interviews, participants reflected on how immersive experiences like Empathy in Action might extend beyond the game. A large portion (N = 42) believed that immersive helping could translate into real-world helping behaviors, and 34 indicated it could shape moral decision-making. For instance, one participant shared, “I think it’s one thing to say, ‘I’m going to [help],’ and it’s another thing to do it. So, I think getting practice in a game helps bridge that gap.” Furthermore, many participants (N = 51) explicitly linked in-game behaviors to real-life helping, while 39 connected their choices to personal relationships. Among those of the latter, one stated, “My roommate is going through a lot right now, and so this game kind of reminded me of ways I could be more [helpful to] her.”

Importantly, 35 participants described moral choices—such as comforting Alden or selecting dialogue options—as more impactful than physical tasks, underscoring the central role of narrative and emotional decision-making. For many, these reflections highlighted that immersive gameplay did not simply model prosocial behavior but encouraged them to consider how they might act in everyday interpersonal situations. This pattern is consistent with prior VR research showing that immersive practice can shape real-world behavior. For instance, Didehbani et al. (2016) found that VR-based social cognition training improved emotion recognition and social attribution in children with autism, demonstrating how rehearsing modeled interactions in virtual contexts can carry over into everyday social situations. In this current study, one participant reflected on immersive game environments as a safe space to rehearse and practice social interactions, saying, “It’s a way for the person playing the game to know what’s possible without real-world consequences. In the real world, when you decide to help or not help someone, you don't know what the outcome is going to be like. In a game, if you fail, you can just try again. And I think that gives a positive reinforcement to the people playing the game.”

An unexpected theme the researchers noticed across several post-game interviews was that many participants (N = 28) reported that Sage’s introductory tutorial in the game not only gave them necessary context and background of the in-game world but also a “role,” “job,” “intention,” and/or “mindset.” For example, one participant said that the in-game setting was a “perfect opportunity to showcase the duties that are related to being a community volunteer.” Another participant described, “It almost felt like a job, and like I had to complete the job in order to keep going.” Another participant said Sage’s introduction gave them direction to be a “role model” and placed them “in this position [in which they] have to help people [because] it’s part of the job.” This may suggest a potential factor in why post-survey results showed an increase in altruistic attitudes, a decrease in empathy, and an agreement towards cognitive over affective empathy. While Sage’s introductory tutorial gave players an understanding of the world, the fact that Sage helped them embody the role of a volunteer possibly changed how people felt and approached Alden. They felt a responsibility or obligation to help others and solve problems, adopting a task-focused framing rather than a purely emotional one. These insights demonstrate the value of effective game design, narrative priming (e.g., tutorials), and the role of guiding characters (like Sage).

Overall, these qualitative insights reinforce the survey and correlation findings by showing that players not only recognized the potential for immersive prosocial experiences to shape moral attitudes but also envisioned how these lessons could extend into their personal and social lives.

4.9 Anticipated use cases

In addition to reflecting on empathy, altruism, and moral decision-making, participants identified potential applications for immersive prosocial games like Empathy in Action. Suggested use cases included students or classroom settings (N = 5), younger children (N = 13), adults (N = 3), and those with disabilities such as autism (N = 1). One participant suggested such games can serve as useful simulations where individuals, particularly children, are encouraged to help others: “It’s not every day that people encounter these types of things. So, for children, for example, I imagine this could be a good way of introducing them to how to help others.” Another participant, who is a teacher, also agreed that immersive prosocial games can allow kids to “test out real-life situations where they would have to build empathy and care about others.”

Others proposed therapeutic or rehabilitative applications, such as therapy (N = 2), conflict resolution training (N = 2), and rehabilitation programs for prisoners, violent offenders, or individuals in the justice system (N = 1). For example, one participant said, “I think it could be a good form of therapy to teach baseline lessons of how empathy should be shown in the real world.” These responses highlight that participants not only engaged with the game’s immediate emotional and moral dimensions but also envisioned concrete ways it could be scaled for broader educational and social purposes.

In addition, one participant, who works professionally in game development, further emphasized the novelty of this design space, stating, “I think that VR really does lend itself to empathy-based gaming, and I’m surprised this is one of very few attempts at an empathy-focused VR game I’ve seen.” This perspective underscores both the timeliness of Empathy in Action and the gap in industry offerings that participants identified.

5 Discussion

The results of this study provide a nuanced look at how immersive, prosocial game experiences influence empathy and altruism. In addressing RQ1 (How do prosocial behaviors in immersive game environments influence players’ level of empathy and sense of altruism?), quantitative findings revealed a significant increase in altruism but a decrease in self-reported empathy, suggesting a more complex relationship than originally anticipated. This partially supports the hypothesis, which predicted significant increases in both empathy and altruism after gameplay, but indicates that altruistic outcomes may emerge even when empathy does not rise in tandem.

5.1 Interpreting empathetic and altruistic gameplay effects through theory

The observed increase in altruism aligns with the core premise of social cognitive theory (Bandura, 1986), which posits that behaviors reinforced in emotionally engaging environments are more likely to be internalized. Empathy in Action presented players with direct opportunities to help a vulnerable character (Alden) through a combination of physical actions and moral decision-making. These moments likely served as reinforcement cues—making players feel competent and rewarded for their actions—which enhanced their motivation to help others. The observed increase in post-game agreement with altruistic statements supports this interpretation. Similar effects have been demonstrated in other immersive studies: Mousavi et al. (2023) found that VR safety training increased university students’ knowledge and self-efficacy, with self-efficacy predicting stronger learning gains, while Nissim and Weissblueth (2017) reported that teacher trainees exposed to VR classroom simulations experienced higher self-efficacy. Both findings support SCT’s emphasis on mastery experiences and observational learning as key drivers of confidence and motivation.

The current study’s findings also support prosocial behavior theory (Batson and Shaw, 1991), which emphasizes empathy, perspective-taking, and moral elevation as drivers of altruism. Although self-reported empathy declined, participants’ altruistic responses increased, suggesting that the game’s emotional arc may have created a moral or motivational shift, even if participants became more critical of their own emotional resonance. This outcome further informs RQ2 (What are the effects of engaging in helping activities within immersive games on prosocial attitudes and moral decision-making?), by suggesting that moral decision-making opportunities—combined with reinforcement—can sustain prosocial attitudes even in the absence of heightened empathy. Such a finding echoes prior works (Gentile et al., 2009; Greitemeyer and Osswald, 2010), which similarly demonstrate that immersive games can promote prosocial behaviors without requiring strong or lasting empathic engagement, provided that players feel empowered to act.

Conversely, the decrease in empathy contradicts expectations and may be due to several factors. According to mental models theory (Johnson-Laird, 1983), players construct internal representations of the scenarios they experience, which in this case included intense emotional and moral moments. However, after completing the game, participants may have shifted their focus from emotional immersion to cognitive reflection. This could explain why empathy scores decreased—not necessarily because players became less empathetic, but because their mental models shifted from emotional identification to critical self-assessment. This interpretation is further supported by exploratory findings showing participants expressed stronger cognitive empathy than affective empathy, suggesting they may have intellectually understood others’ emotions in the game without deeply feeling them—reinforcing the notion of reflective rather than emotional engagement post-play.

This complexity is also fortified by correlational analyses, which revealed that empathy and altruism were still strongly and significantly associated with one another, even though only altruism increased on average. In particular, both traits were closely linked with participants’ reported levels of immersion and perceived gameplay impact. Those who felt more immersed in the virtual environment—and those who believed the experience had a meaningful personal effect—were significantly more likely to report higher empathy and altruism. These findings reinforce the idea that emotional and moral outcomes in immersive gameplay are not isolated, but deeply interconnected, and may depend more on how the experience is perceived and internalized than on any single measured construct. Furthermore, feedback from participants—including those with gaming industry experience—underscored that empathy-focused VR remains an underdeveloped area. This observation strengthens the broader significance of the findings: immersive prosocial gameplay holds potential well beyond research settings, with room for growth in mainstream applications.

5.2 Subgroup differences in responses

Additional subgroup analyses shed light on how individual characteristics shaped participants’ responses. Participants’ post-game empathy levels significantly differed by education level, with those who completed trade or vocational programs reporting the highest agreement with empathy-related statements, and those with bachelor’s degrees reporting the lowest. This may suggest that individuals with applied or technical education backgrounds were more emotionally engaged by the experience. Similarly, perceived effects of the gameplay varied by education level, with bachelor’s degree holders reporting the strongest gameplay impact and others—especially those with vocational or graduate-level education—reporting lower levels. These differences may reflect distinct interpretive lenses or expectations for media engagement shaped by educational background.

Participants’ prior experience with simulation games also affected how they perceived the game’s impact. Specifically, those who never played simulation games reported significantly higher impact scores than those who played them occasionally. This finding suggests that novelty, or a lack of familiarity with similar game formats, may have heightened players’ sensitivity to the emotional and narrative aspects of the experience. Taken together, these subgroup effects highlight how background factors—such as education or gaming habits—can influence the extent to which immersive narratives evoke emotional or prosocial responses.

5.3 Predictors of altruistic outcomes

Regression analyses further clarified which factors most strongly predicted post-game altruism. While neither demographic characteristics nor gaming habits significantly predicted altruistic outcomes, participants’ gameplay-related experiences proved to be highly predictive. More specifically, perceived gameplay effect emerged as the strongest and most statistically significant predictor of post-game altruism, followed by pre-game empathy. These results reinforce the idea that players’ perceptions of the game’s emotional and moral resonance are more predictive of altruistic responses than demographic or behavioral factors. Interview data aligned with this pattern: many participants (N = 42) believed immersive helping could translate to real-world helping, while others (N = 34) reflected that the game shaped how they might approach moral dilemmas beyond play. Several (N = 39) explicitly connected in-game behaviors to personal relationships, further illustrating that self-reported gameplay impact was not just an abstract score, but a lived sense of transformation.

5.4 Decoupling empathy and altruism

Furthermore, the divergence between increased altruism and decreased empathy may reflect a shift in how emotional engagement unfolds during and after gameplay, an idea supported by previous research on immersive prosocial experiences. For instance, Gorisse et al. (2017) suggest that first-person VR can heighten emotional engagement, though such engagement may depend on narrative context and may diminish once the scenario concludes. Similarly, Rosenberg et al. (2013) found that embodying superheroes in VR encouraged participants to engage in immediate helping behaviors, though the study did not examine whether such effects were sustained over time. This suggests that immersive experiences may prompt short-term prosocial actions without necessarily fostering longer-term empathic engagement. In this current study, the emotional impact of the game may have peaked during gameplay, followed by a decrease in self-perceived empathy as players gained distance and perspective post-play.

Based on quantitative responses and interview reflections, findings from this research suggest that emotional intensity peaked during gameplay and declined afterward, leaving participants more self-critical of their empathy yet still motivated to help. This pattern is consistent with the idea that empathic responding can follow multiple routes: affective resonance may taper as emotional fatigue or evaluative recalibration sets in, while cognitive understanding and moral reflection remain active. In line with critiques emphasizing the distinction between emotional resonance and principled prosocial motivation (Bloom, 2017; Zaki, 2014), our exploratory cognitive–affective comparison (Section 4.5) suggests that participants may have relied more on reflective, understanding-based processing than on sustained emotional mirroring after gameplay. Such comments reveal that players may become more self-critical of their emotional resonance while still reporting motivation to help, reiterating the quantitative pattern of decreased empathy but increased altruism.

Ultimately, these results suggest that empathy and altruism are not always linear or simultaneous. A person can feel compelled to help even if they do not feel deeply connected emotionally in that moment. These findings complicate the assumption that empathy and altruism always move in tandem, suggesting that the relationship between emotional connection and helping behavior may be mediated by how players interpret and internalize their experience. While foundational work in prosocial behavior theory has emphasized empathy as a driver of altruism (Batson and Shaw, 1991; Batson et al., 2015), alternative scholarship highlight that empathy does not consistently lead to helping, particularly when factors such as emotional overload, cognitive framing, or perceived efficacy come into play (Bloom, 2017; Zaki, 2014). The results also support the value of designing immersive experiences that reinforce moral agency and decision-making, not just emotional immersion.

5.5 Limitations

While this study offers valuable insights into the social and emotional effects of immersive prosocial gameplay, several limitations should be acknowledged. Identifying these boundaries not only clarifies the scope of the current findings but also highlights avenues for refining measurement and design in future research.

First, the adapted Basic Empathy Scale used in the pre-survey (Jolliffe and Farrington, 2006) showed poor internal consistency and was ultimately excluded from composite analyses, which limits the reliability of empathy measurement in the current study. Although individual items were analyzed to differentiate cognitive and affective empathy, this decision restricts comparability with prior studies that employed validated full-scale versions. Future studies may benefit from using complete versions of validated empathy scales or designing new instruments specifically suited to immersive contexts.

Second, all measures relied on self-report, which can introduce social desirability bias or response distortion, particularly in emotionally charged research contexts. Future research could address this by incorporating behavioral observations, physiological measures, or peer reports.

Additionally, the study captured only short-term effects; it is unclear whether increases in altruism would persist over time or translate into real-world behavior. Future work could employ longitudinal designs and follow-up assessments to evaluate the durability of these effects.

Moreover, the sample, while diverse, was limited to U.S.-based adults, which may reduce the generalizability of results across cultures and age groups. Future studies could expand to cross-cultural samples and younger populations to capture developmental and cultural variation.

In addition, although the target sample size (N = 64) was determined a priori via G*Power to detect medium-sized effects typical of XR intervention studies (e.g., Chang et al., 2022; Li et al., 2023), it remains a modest sample drawn from a single university community. As such, the study was underpowered to detect small effects or more complex interaction patterns, and the findings should be interpreted as preliminary.

Relatedly, for many participants, the use of a VR headset and immersive environment was relatively novel. Novelty and initial discomfort with the technology may have amplified immersion and self-reported impact for some participants while dampening it for others, which makes it difficult to fully disentangle the effects of the prosocial narrative from the effects of the medium itself. Future research with larger, multi-site samples and participants with a wider range of VR experience will be important for testing the robustness and generalizability of these patterns beyond the university setting.

Lastly, since the post-game measures were tailored to the specific game experience, some differences from the pre-game measures (which were more general) may reflect measurement variance (i.e., the degree of spread or dispersion in a set of measurements) as well as genuine change. Future work could strengthen comparability by designing alternative parallel pre- and post-measures that balance general constructs with context-specific relevance, in addition to introducing more decisions and consequences in the game to expand understandings of gameplay effects.

5.6 Significance and broader impacts

This study advances insights into how immersive media can foster prosocial behavior by demonstrating that altruistic intentions can increase even without parallel gains in empathy. This finding highlights a critical nuance: immersive games may shape moral action not only through emotional resonance but also by reinforcing decision-making agency, feedback, and behavioral modeling. Such insights expand existing theory by showing that empathy is not the sole driver of altruism, and that immersive experiences can cultivate a sense of responsibility and competence that motivates helping.

Beyond theoretical contributions, these results carry practical implications for the design of educational and therapeutic games. Developers and educators can leverage immersive environments to simulate moral dilemmas, reward helping behaviors, and strengthen players’ confidence in their capacity to contribute positively to others. Interview data reinforce this applied potential: participants explicitly envisioned use cases in schools, therapy, training, and rehabilitation, underscoring that immersive prosocial games can be designed with diverse audiences and contexts in mind.

Looking ahead, the study underscores the broader societal potential of immersive media as tools for social change, healing, and cultural learning. By situating players in narratives that model care, responsibility, and interconnection, immersive games can serve as powerful complements to traditional education, therapy, and advocacy—broadening the scope of how digital media can promote collective wellbeing. Overall, these findings suggest that immersive games like Empathy in Action represent not just an academic proof of concept but an entry point into a largely untapped design space and area of research. Industry perspectives reinforce this view, indicating that the field is only beginning to explore how immersive media can be leveraged for social good.

Data availability statement

The datasets presented in this article are not readily available because of ethical restrictions and the need to protect participant privacy. Requests for access to the raw, anonymized data were not approved by the institutional review board (IRB). Requests to access the datasets should be directed to c2xvcmVuem9AdW9yZWdvbi5lZHU=.

Ethics statement

The studies involving humans were approved by the Institutional Review Board (IRB) at the University of Oregon. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

Author contributions

SL: Conceptualization, Data curation, Formal Analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing. LO: Formal Analysis, Investigation, Writing – review and editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. Financial support for this study was provided by a grant from the Center for Science Communication Research in the School of Journalism and Communication at the University of Oregon.

Acknowledgements

This project was made possible by the contributions of members of the Oregon Reality (OR) Lab at the University of Oregon. The authors thank Dr. Daniel Pimentel, Assistant Professor of Immersive Psychology and Director of the OR Lab, for his support in game development and securing project funding. We also acknowledge Luke Walker, Ph.D. student in Communication and Media Studies, for his contributions to game development. Jason deParrie-Turner, graduate of the Immersive Media Communication master’s program, and Isaac Wu, current master’s student in the program, assisted with development during an early iteration of the VR game and contributed to data collection. We further thank Tai Le, graduate of the Immersive Media Communication and Strategic Communication master’s programs, for assistance with data collection.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Ahn, S. J. G., Bailenson, J. N., and Park, D. (2014). Short- and long-term effects of embodied experiences in immersive virtual environments on environmental locus of control and behavior. Comput. Hum. Behav. 39, 235–245. doi:10.1016/j.chb.2014.07.025

CrossRef Full Text | Google Scholar

Bandura, A. (1986). Social foundations of thought and action. Englewood Cliffs, NJ: Prentice-Hall.

Google Scholar

Baños, R. M., Botella, C., Alcañiz, M., Liaño, V., Guerrero, B., and Rey, B. (2004). Immersion and emotion: their impact on the sense of presence. Cyberpsychology and Behav. 7 (6), 734–741. doi:10.1089/cpb.2004.7.734

PubMed Abstract | CrossRef Full Text | Google Scholar

Batson, C. D., and Shaw, L. L. (1991). Evidence for altruism: toward a pluralism of prosocial motives. Psychol. Inq. 2 (2), 107–122. doi:10.1207/s15327965pli0202_1

CrossRef Full Text | Google Scholar

Batson, C. D., Lishner, D. A., and Stocks, E. L. (2015). “The empathy-altruism hypothesis,” in The Oxford handbook of prosocial behavior (Oxford University Press), 259–281.

Google Scholar

Bloom, P. (2017). Against empathy: the case for rational compassion. London: Random House.

Google Scholar

Braun, V., and Clarke, V. (2021). Thematic analysis: a practical guide. London: Sage Publications.

Google Scholar

Busetto, L., Wick, W., and Gumbinger, C. (2020). How to use and assess qualitative research methods. Neurological Res. Pract. 2 (1), 14. doi:10.1186/s42466-020-00059-z

PubMed Abstract | CrossRef Full Text | Google Scholar

Cabellos, B., Pozo, J. I., Marín-Rubio, K., and Sánchez, D. L. (2022). Do pro-social video games promote moral activity? an analysis of user reviews of papers. Educ. Inf. Technol. 27 (8), 11411–11442. doi:10.1007/s10639-022-11072-x

CrossRef Full Text | Google Scholar

Carlo, G., and Randall, B. A. (2002). The development of a measure of prosocial behaviors for late adolescents. J. Youth Adolesc. 31 (1), 31–44. doi:10.1023/a:1014033032440

CrossRef Full Text | Google Scholar

Chang, H. Y., Binali, T., Liang, J. C., Chiou, G. L., Cheng, K. H., Lee, S. W. Y., et al. (2022). Ten years of augmented reality in education: a meta-analysis of (quasi-) experimental studies to investigate the impact. Comput. and Educ. 191, 104641. doi:10.1016/j.compedu.2022.104641

CrossRef Full Text | Google Scholar

Davis, M. H. (1980). A multidimensional approach to individual differences in empathy. JSAS Catalog Sel. Documents Psychol. 10, 85.

Google Scholar

Davis, M. H. (1983). Measuring individual differences in empathy: evidence for a multidimensional approach. J. Personality Soc. Psychol. 44 (1), 113–126. doi:10.1037/0022-3514.44.1.113

CrossRef Full Text | Google Scholar

Didehbani, N., Allen, T., Kandalaft, M., Krawczyk, D., and Chapman, S. (2016). Virtual reality social cognition training for children with high functioning autism. Comput. Hum. Behav. 62, 703–711. doi:10.1016/j.chb.2016.04.033

CrossRef Full Text | Google Scholar

du Bois, Z. (2022). Journey: visualising peace through gaming. Museum of Peace at the University of St Andrews. Available online at: https://peacemuseum.wp.st-andrews.ac.uk/2022/06/04/journey-visualising-peace-through-gaming/(Accessed September 13, 2025).

Google Scholar

Evren, F. B. (2025). Digital game narratives as living culture: never alone. Milli Folklor. 19 (146), 88–99. doi:10.58242/millifolklor.1429467

CrossRef Full Text | Google Scholar

Fraser, A. M., Padilla-Walker, L. M., Coyne, S. M., Nelson, L. J., and Stockdale, L. A. (2012). Associations between violent video gaming, empathic concern, and prosocial behavior toward strangers, friends, and family members. J. Youth Adolesc. 41 (5), 636–649. doi:10.1007/s10964-012-9742-2

PubMed Abstract | CrossRef Full Text | Google Scholar

Garzón, J., and Acevedo, J. (2019). Meta-analysis of the impact of augmented reality on students learning gains. Educ. Res. Rev. 27, 244–260. doi:10.1016/j.edurev.2019.04.001

CrossRef Full Text | Google Scholar

Garzón, J., Pavón, J., and Baldiris, S. (2019). Systematic review and meta-analysis of augmented reality in educational settings. Virtual Real. 23 (4), 447–459. doi:10.1007/s10055-019-00379-9

CrossRef Full Text | Google Scholar

Gentile, D. A., Anderson, C. A., Yukawa, S., Ihori, N., Saleem, M., Ming, L. K., et al. (2009). The effects of prosocial video games on prosocial behaviors: international evidence from correlational, longitudinal, and experimental studies. Personality Soc. Psychol. Bull. 35 (6), 752–763. doi:10.1177/0146167209333045

PubMed Abstract | CrossRef Full Text | Google Scholar

Gillath, O., McCall, C., Shaver, P. R., and Blascovich, J. (2008). What can virtual reality teach us about prosocial tendencies in real and virtual environments? Media Psychol. 11 (2), 259–282. doi:10.1080/15213260801906489

CrossRef Full Text | Google Scholar

Gorisse, G., Christmann, O., Amato, E. A., and Richir, S. (2017). First-and third-person perspectives in immersive virtual environments: presence and performance analysis of embodied users. Front. Robotics AI 4, 33. doi:10.3389/frobt.2017.00033

CrossRef Full Text | Google Scholar

Greitemeyer, T., and Osswald, S. (2010). Effects of prosocial video games on prosocial behavior. J. Personality Soc. Psychol. 98 (2), 211–221. doi:10.1037/a0016997

PubMed Abstract | CrossRef Full Text | Google Scholar

Guest, G., MacQueen, K. M., and Namey, E. E. (2011). Applied thematic analysis. Thousand Oaks, CA: Sage Publications.

Google Scholar

Han, D. I. D., Melissen, F., and Haggis-Burridge, M. (2024). Immersive experience framework: a Delphi approach. Behav. and Inf. Technol. 43 (4), 623–639. doi:10.1080/0144929x.2023.2183054

CrossRef Full Text | Google Scholar

Herrera, F., Bailenson, J., Weisz, E., Ogle, E., and Zaki, J. (2018). Building long-term empathy: a large-scale comparison of traditional and virtual reality perspective-taking. PLOS One 13 (10), e0204494. doi:10.1371/journal.pone.0204494

PubMed Abstract | CrossRef Full Text | Google Scholar

Jennett, C., Cox, A. L., Cairns, P., Dhoparee, S., Epps, A., Tijs, T., et al. (2008). Measuring and defining the experience of immersion in games. Int. J. Human-Computer Stud. 66 (9), 641–661. doi:10.1016/j.ijhcs.2008.04.004

CrossRef Full Text | Google Scholar

Johnson-Laird, P. N. (1983). Mental models. Harvard University Press.

Google Scholar

Jolliffe, D., and Farrington, D. P. (2006). Development and validation of the basic empathy scale. J. Adolesc. 29 (4), 589–611. doi:10.1016/j.adolescence.2005.08.010

PubMed Abstract | CrossRef Full Text | Google Scholar

Li, F., Wang, X., He, X., Cheng, L., and Wang, Y. (2023). How augmented reality affected academic achievement in K-12 education–a meta-analysis and thematic-analysis. Interact. Learn. Environ. 31 (9), 5582–5600. doi:10.1080/10494820.2021.2012810

CrossRef Full Text | Google Scholar

Mousavi, S. A., Powell, W., Louwerse, M. M., and Hendrickson, A. T. (2023). Behavior and self-efficacy modulate learning in virtual reality simulations for training: a structural equation modeling approach. Front. Virtual Real. 4, 1250823. doi:10.3389/frvir.2023.1250823

CrossRef Full Text | Google Scholar

Nissim, Y., and Weissblueth, E. (2017). Virtual reality (VR) as a source for self-efficacy in teacher training. Int. Educ. Stud. 10 (8), 52–59. doi:10.5539/ies.v10n8p52

CrossRef Full Text | Google Scholar

Peck, T. C., Seinfeld, S., Aglioti, S. M., and Slater, M. (2013). Putting yourself in the skin of a black avatar reduces implicit racial bias. Conscious. Cognition 22 (3), 779–787. doi:10.1016/j.concog.2013.04.016

PubMed Abstract | CrossRef Full Text | Google Scholar

Peña, J., and Chen, M. (2017). With great power comes great responsibility: superhero primes and expansive poses influence prosocial behavior after a motion-controlled game task. Comput. Hum. Behav. 76, 378–385. doi:10.1016/j.chb.2017.07.039

CrossRef Full Text | Google Scholar

Pine, J., and Gilmore, J. H. (2001). Welcome to the experience economy. Health Forum J. 44 (5), 10–16.

PubMed Abstract | Google Scholar

Pope, L. (2013). Papers, Please [video game], 3909. LLC.

Google Scholar

Rosenberg, R. S., Baughman, S. L., and Bailenson, J. N. (2013). Virtual superheroes: using superpowers in virtual reality to encourage prosocial behavior. PLOS One 8 (1), e55003. doi:10.1371/journal.pone.0055003

PubMed Abstract | CrossRef Full Text | Google Scholar

Saleme, P., Pang, B., Dietrich, T., and Parkinson, J. (2020). Prosocial digital games for youth: a systematic review of interventions. Comput. Hum. Behav. Rep. 2, 100039. doi:10.1016/j.chbr.2020.100039

CrossRef Full Text | Google Scholar

Sandelowski, M. (2000). Whatever happened to qualitative description? Res. Nurs. and Health 23 (4), 334–340. doi:10.1002/1098-240x(200008)23:4<334::aid-nur9>3.0.co;2-g

PubMed Abstract | CrossRef Full Text | Google Scholar

Scavarelli, A., Arya, A., and Teather, R. J. (2021). Virtual reality and augmented reality in social learning spaces: a literature review. Virtual Real. 25 (1), 257–277. doi:10.1007/s10055-020-00444-8

CrossRef Full Text | Google Scholar

Slezaka, R. J., Keren, N., Gilbert, S. B., Harvey, M. E., Ryan, S. J., and Wiley, A. J. (2023). Examining virtual reality as a platform for developing mental models of industrial systems. J. Comput. Assisted Learn. 39 (1), 113–124. doi:10.1111/jcal.12731

CrossRef Full Text | Google Scholar

Starks, K. (2014). Cognitive behavioral game design: a unified model for designing serious games. Front. Psychol. 5, 28. doi:10.3389/fpsyg.2014.00028

PubMed Abstract | CrossRef Full Text | Google Scholar

Thatgamecompany (2012). Journey [video game]. Sony Computer Entertainment.

Google Scholar

Upper One Games (2014). Never alone (Kisima Ingitchuna) [video game]. E-Line Media.

Google Scholar

Vogt, A., Babel, F., Hock, P., Baumann, M., and Seufert, T. (2021). Prompting in-depth learning in immersive virtual reality: impact of an elaboration prompt on developing a mental model. Comput. and Educ. 171, 104235. doi:10.1016/j.compedu.2021.104235

CrossRef Full Text | Google Scholar

Zaki, J. (2014). Empathy: a motivated account. Psychol. Bull. 140 (6), 1608–1647. doi:10.1037/a0037679

PubMed Abstract | CrossRef Full Text | Google Scholar