War is forcing many people to flee from their hometowns, not knowing what to expect and whether they will be able to survive and return. For humans in countries at peace, it can be hard to imagine what it is like to stand in the shoes of a refugee and get the full scope of the impact it has on a refugee’s life. Personal relevance is low due to a lack of personal experience and a limited impact on one’s personal life (Kuzmičová & Bálint, 2019). Narratives allow readers to imagine the perspective of war victims by telling the victims’ stories from their point of view. Interactive digital narratives (IDNs) transform readers of a narrative into players permitting them to make choices on behalf of the war victim, encouraging perspective-taking even more (Green & Jenkins, 2014; Peng et al., 2010; Scholl et al., 2022).

A compelling example of such an IDN is the newsgame I am Mosul (2019) (www.iamosul.nl) created by journalists Frederick Mansell and Laurens Samsom to raise awareness about the devastating effects of war on the local citizens. I am Mosul increases personal relevance by bringing the war close to home, enabling a geolocalized experience (López-Arcos et al., 2017): the player gets to choose their own Dutch hometown and also has the agency to make choices on behalf of the main character on how to survive. As a newsgame, I am Mosul is at the intersection of narrative games and journalism (Bogost et al., 2010).

The goal of changing someone's perspective on the world through an experience is in line with the process of transformative learning (Mezirow, 2000, 2012). This process involves different phases, from being confronted with a disorienting dilemma, to a critical assessment of one’s own assumptions, to trying out and implementing new actions based on a new perspective (Mezirow, 2012, p. 86), ultimately transforming learners' mindsets (frames of reference) to be more open and inclusive.

Narratives have the potential to enable transformative learning by presenting their readers with such a disorienting dilemma making them reconsider their assumptions about, e.g., war and refugees, allowing a deeper or maybe even different understanding (Hoggan & Cranton, 2015). IDNs – such as I am Mosul - could increase transformative learning even more (Roth, 2019). The agency provided by IDNs enables learners to be actively involved in the learning process and construct their perception of reality. This way of active learning (Bonwell & Eison, 1991) through IDNs has been shown to stimulate transformative learning more than passive traditional narratives(Scholl et al., 2022)

Agency in interactive narratives can affect transformative learning through narrative engagement. Narrative engagement covers the constructs of identification and transportation (Bilandzic & Busselle, 2017). Players may be more prone to identify with the character when they can make decisions for this character, with a discernible effect on the narrative. This allows players to experience the events ‘through the eyes’ of this character and actively adopt the character’s goals (Green & Jenkins, 2014; Peng et al., 2010). This heightened identification with the character can facilitate the player’s transportation into the storyworld: “absorbed into the story through the position and role of the character with whom one identifies” (Cohen et al., 2015, p.240; Hand & Varan, 2008). As a result, players may internalize the perspectives of characters even more than in a traditional narrative (Hand & Varan, 2009), paving the way for transformative learning.

Personal relevance may be an additional factor enhancing transformative learning, via intrinsic motivation and identification. As for intrinsic motivation, personal relevance can be seen as an important antecedent. Personally relevant issues are more likely to be processed carefully than personally irrelevant issues (Ajzen et al., 1996, p. 44). As for identification, there are studies finding no effects of personal relevance on narrative engagement (e.g., identification) (Quintero Johnson et al., 2013; Soto-Sanfiel & Igartua, 2015). However, other studies do indicate that it may be easier to identify with the main character of a personally relevant narrative than with one of a personally irrelevant narrative (de Graaf et al., 2012; Hoeken et al., 2016; van Krieken et al., 2017). In I am Mosul, players experience a war in their own city, mixing a real setting (own city) with fictitious events (war in this city), which creates a so-called geolocalized experience (López-Arcos et al., 2017 ). By bringing the unfamiliar experience of war to a familiar city, the geolocalization in I am Mosul arguably makes the experience more personally relevant for the player, enabling identification with the main character.

In this paper, we investigate the effect of narrative agency on transformative learning. We also look into potential mediation effects of identification with the main character and transportation into the storyworld and investigate the potential moderating effect of personal relevance as well.

We formulated the following research question:

RQ: How does agency in a narrative newsgame affect transformative learning, and how is this effect mediated by identification and transportation and moderated by personal relevance?

To answer our research question, we ran an experiment comparing two versions of the narrative newsgame I am Mosul: (a) the original version with a higher level of agency and (b) an adapted version with a low level of agency. Furthermore, the newsgame either took place in a personally relevant or irrelevant city.

1. Theoretical framework

1.1 Interactive digital narratives

Interactive digital narratives (IDNs) allow users to actively participate in the narrative. They can take many forms, such as video games, mobile apps and websites, and often involve decision-making, branching narratives, and multiple endings that allow users to shape the direction and outcome of the story (Crawford, 2013; Koenitz, 2023; Murray, 2017). According to Roth and Koenitz 2016, IDNs afford “dramatic agency for interactors, and the ability to intentionally influence salient aspects (character development, sequencing, outcome etc.) of a narrative”. Murray (2017) describes this agency as “the satisfying power to take meaningful action and see the results of our decisions and choices” (p.159). Her use of the term agency is closely related to the concept of effectance in psychological research, the sense of active participation by intentionally influencing a game or storyworld (Klimmt & Hartmann, 2012). High agency can create a more engaging experience, as users feel more invested in the story and its outcome (Roth & Koenitz, 2016). The browser-based hypertext interface of I am Mosul allows players to exert agency over the storyline they follow by clicking on different options presented in the text (Green & Jenkins, 2014). The underlying branching structure allows them to experience the journalistic story in different ways depending on their choices. Combining journalism and games, I am Mosul falls under the header of newsgames.

1.2 Newsgames

Newsgames are a specific form of IDN and can be seen as a unique genre of video games that aim to incorporate journalistic principles into their design and gameplay (Bogost et al., 2010; Bogost, 2020;García-Avilés et al., 2022 ; Lin & Wu, 2020; Plewe & Fürsich, 2018). They are created to provide players with a better understanding of complex events or situations. They can provide context and perspective that is difficult to achieve through other mediums such as written articles or news broadcasts, thus also engaging audiences who may feel disconnected to certain topics or who might not be interested in traditional news media. According to Bogost et al. (2010; 2020), the term “newsgame” suggests any intersection of journalism and gaming; “newsworthy topics in game form”.

Newsgames can take a variety of forms, but they typically involve the player taking on the role of a character in a real-life scenario, such as a journalist covering a breaking news story or  a refugee trying to survive in a war-torn country. Bogost et al. (2010) further mention “current event games”, a type of easily accessible newsgames which run in the browser and give insight into a current topic. I am Mosul can be seen as a relatively complex narrative-driven form of such a current event game.

Newsgames, such as I am Mosul, aim to provide players with a deeper understanding of the events and issues they are experiencing by immersing them in the situation and allowing them to make decisions that can impact the outcome of the game. This offers a unique and engaging way to explore complex issues and events (García-Avilés et al., 2022; Navarro-Remesal & Zapata, 2019), potentially enhancing or even transforming our understanding.

2. Transformative learning through IDNs

According to Murray (2017), IDNs allow players to transform themselves into someone else for the duration of the experience, letting them experience new perspectives through enactment rather than merely witnessing (pp. 212-214). This enactment can be seen as a precursor for personal transformation, potentially altering player’s attitudes, beliefs, and behavior in real life. In I am Mosul, the player enacts a refugee, which enables the exploration of and experimentation with different actions and consequences. When this exploration and experimentation yields a deeply personal and meaningful insight with the potential to change attitudes and behavior, the experience can be qualified as transformative (Roth, 2019), falling under the scope of Mezirow’s Transformative Learning Theory (Mezirow, 2000, 2012).

Transformative learning refers to “the process by which we transform our taken-for-granted frames of reference (meaning perspectives, habits of mind, mindsets) to make them more inclusive, discriminating, open, emotionally capable of change, and reflective so that they may generate beliefs and opinions that will prove truer or justified to guide action” (Mezirow, 2012, p. 76). Learners are confronted with new information that requires them to evaluate their past understanding of a topic, shifting their worldview through critical reflection. In contrast to just acquiring new knowledge, learners can find new meaning in their understanding and potentially in their lives. The process of transformative learning can thus involve a fundamental change in learners’ perception, making space for new insights examined from a new perspective. To arrive at this fundamental change, the learner goes through several phases starting with a transformative spark that Mezirow (2000) calls the "disorienting dilemma". He associates this dilemma with major life crises or transitions, challenging existing beliefs. A disorienting dilemma is an unexpected incident, a surprising outcome in our lives, which upon examination can reveal that we had misconceptions about reality.

Transformative learning occurs as a result of an impactful experience, with an IDN experience as a promising example (Roth, 2019). I am Mosul, for instance, contains several discomforting and disorienting dilemmas by creating the fiction of a war in the player’s hometown. When players have the agency to act on behalf of their character in the warzone, disorienting dilemmas could feel like their own (Chen & Martin, 2015; Scholl et al., 2022). Agency encourages or even forces players to reflect when making choices as these choices impact the narrative progression and have consequences, such as an exploding mine when searching for food. Being surprised and overwhelmed in I am Mosul can help players to understand the personal impact and the difficult decisions that people in a war situation have to make. This could trigger critical reflection and, ultimately, transformative learning. We therefore hypothesize a positive effect of agency on transformative learning:

H1: High agency enhances transformative learning compared to low agency.

Narrative agency, narrative engagement and transformative learning. Both identification with characters and transportation into the storyworld are key constructs of narrative engagement (Bilandzic & Busselle, 2017). Identification is about taking a character’s perspective, “a mechanism through which audience members experience reception and interpretation of the text from the inside, as if the events were happening to them” (Cohen, 2001, p.245). Transportation comprises engagement with the narrative as a whole. When readers are cognitively and emotionally involved in the unfolding of narrative events, they are transported into the storyworld (Bilandzic & Busselle, 2017; Green et al., 2004). Identification can be seen as a facilitator of transportation: “absorbed into the story through the position and role of the character with whom one identifies” ( Cohen et al., 2015, p.240; Bilandzic & Busselle, 2017). By taking on the role of a character, readers may be more cognitively and emotionally involved in the narrative. Therefore, an increased level of identification arguably yields a higher level of transportation.

Previous studies have pointed at various antecedents of identification and transportation (e.g., Cohen & Tal-Or, 2017; de Graaf et al., 2012; van Krieken et al., 2017). We reason that one such antecedent is narrative agency, particularly when it is operationalized as choices a player is invited to make on behalf of the character (cf. Green & Jenkins, 2014). I am Mosul (2019) has a branching structure providing players with such agency. The player chooses on behalf of the main character to, e.g., flee or stay, and is confronted with the consequences of either staying or fleeing. This agency can encourage identification with characters (Dillman Carpentier et al., 2015; Green & Jenkins, 2014; Hand & Varan, 2007; 2008; 2009; Peng et al., 2010). The offered agency in I am Mosul can be seen as an invitation to ‘step into the shoes’ of the main character, to accept “a role in the narrative setting” (Roth & Koenitz, 2016). Readers become players who actively adopt the goals of characters by making decisions for them. Moreover, players can better empathize with characters, because they experience the events that characters face directly instead of vicariously (Rigby & Ryan, 2016). Narrative agency is able to trigger transportation as well (Hand & Varan, 2007; Zhou & Kim, 2022). The unfolding of events is affected by choices of the player and as a result, players become invested in the outcomes of their choices and the course of the narrative as a whole (“the consequences of those events are felt more deeply” (Hand & Varan, 2008, p. 13), transporting them more (Murray, 2017).

Narrative engagement can have a positive effect on transformative learning. Firstly, readers who identify with a character are more likely to feel empathy toward a character and will be more inclined to try and understand the views and attitudes of a character (de Graaf et al., 2012; Hoeken & Fikkers, 2014). Perspective-taking (Jarvis, 2012) and empathy (Mezirow, 2012) are not only relevant for identification, but are also key factors in transformative learning. By adopting a character's perspective, readers may develop empathy for that character, leading to an increase in shared understanding. As a result, readers may reflect on and adjust their own frame of reference, ultimately promoting transformative learning. Secondly, when being transported, readers concentrate all their cognitive processes on understanding the narrative. This heightened focus enables readers to process relevant learning content more deeply and increases their receptiveness to diverse perspectives (Green et al., 2004). This can facilitate transformative learning (Zhou & Kim, 2022).

Due to agency and narrative engagement, players may feel more personally responsible for the unfolding events, which may convert into a change of perspective and attitudes in the real world. We therefore expect that identification and transportation serve as mediators for the effect of agency on transformative learning leading to Hypothesis 2:

H2: Identification and transportation mediate the effect of agency on transformative learning: High agency leads to more identification than low agency, which subsequently leads to more transportation and more transformative learning.

Personal relevance and transformative learning. Personally relevant issues are of intrinsic importance, have personal meaning or have significant consequences for people's own lives (Kuzmičová & Bálint, 2019; Petty & Cacioppo, 1979, p.1916). Previous studies manipulated personal relevance for example through cultural proximity (Soto-Sanfiel & Igartua, 2015), topic familiarity (Quintero Johnson et al., 2013) or geographic proximity (Petty & Cacioppo, 1979). Although war is intrinsically important, it is often happening at geographically distant locations, with low perceived personal relevance as a result. To seemingly proximize the events, geolocalization can be used, increasing perceived personal relevance. Geolocalized games are related to AR games as they are “developing in both virtual and real scenarios” (López-Arcos et al., 2017, p.2); “the real world has been integrated within the interactive story” (p.4). I am Mosul is a clear example of a geolocalized game. In I am Mosul, players experience a war in their own city instead of in Mosul, mixing a real setting (own city) with fictitious events (war in this city). This creates a geolocalized experience, bringing the unfamiliar experience of war to a familiar city. The geolocalization in I am Mosul arguably makes the experience more personally relevant for the player, enhancing identification with the main character.

We expect that identification and transformative learning are highest when people get to make choices for the main character in a personally relevant city and lowest when choices are absent and the city is personally irrelevant. Hypothesis 3 and 4 are as follows:

H3: A higher degree of personal relevance in the narrative has a stronger effect of agency on transformative learning than a lower degree of personal relevance.

H4: Personal relevance moderates the effect of agency on identification: A higher degree of personal relevance in the narrative has a stronger effect of agency on identification with the character in the narrative than a lower degree of personal relevance.

This study received ethical clearance from the Research Ethics and Data Management Committee of Tilburg School of Humanities and Digital Sciences (REDC # 20201078). The survey items and dataset of this study can be found here: https://doi.org/10.34894/DTSZML.

4. Design

A 2 x 2 between-subjects design was used to investigate whether the agency in I am Mosul (IV: low vs. high) positively affects transformative learning (DV), and whether this effect is mediated by identification and transportation (MED 1 and 2). The moderator personal relevance (MOD 1: low vs. high personal relevance) was included to investigate whether this factor moderated the effect of agency on identification (MED 1) and transformative learning. The conceptual model can be found in Figure 1. Participants were randomly assigned to one of the four conditions (1. high agency / high personal relevance, 2. high agency / low personal relevance, 3. low agency / high personal relevance, 4. low agency / low personal relevance).

5. Participants

Participants were acquired through the Human Subject Pool of the Tilburg School of Humanities and Digital Sciences (n = 83) and the researchers’ personal networks (n = 132). Participants recruited through the Human Subject Pool received 0.5 credits of compensation for their participation. Participants recruited through the researchers’ personal networks received no compensation. The sample originally consisted of 230 participants. Participants who were familiar with I am Mosul and had played the game before (n = 8) or participants who had more than one wrong answer in the control questions about the narrative (n = 7) were excluded from the analyses, ending up with a final sample of 215 participants. The participants’ age ranged from 16 to 67 (M = 26.2, SD = 10.6) and consisted of 156 females (72.6 %) and 59 males (27.4 %). Most participants were highly educated (n = 174, 80.9 %)^[1]. As I am Mosul was presented in Dutch, only participants were recruited who were fluent in Dutch.

6. Material

This study focused on the newsgame I am Mosul (2019) (www.iamosul.nl) created by journalists Frederick Mansell and Laurens Samsom and published by the Dutch public broadcaster BNNVARA. This newsgame gives players the agency to choose a city, push a “reality check” button^[2] (comparing the events to what happened in Mosul) and – foremost – make choices on behalf of the main character. These choices concern actions (e.g., “You have to flee. Which three items would you bring along?”) with local effectance (e.g., not choosing to bring money disables you grabbing a taxi while fleeing). The newsgame uses a foldback structure, with main events remaining the same (e.g., having to flee from your hometown, returning to your hometown) and choices converging back to a main path. To test our hypotheses, we created a low-agency version of I am Mosul. In this version, created in collaboration with the journalists, the participants were only able to choose a city and use the “reality check” button. All choices on behalf of the main character were removed, yielding a fully linear story structure. For instance, instead of asking to bring along three things, the items were already selected. Both versions contained the same reality checks and main events.

As for personal relevance, the narrative took place in a city with either high or low personal relevance, manipulated through the instructions. In both high personal relevance conditions, participants were given the instruction to choose a city they felt close to. In the high-agency condition, they could choose between 35 Dutch cities. In the low-agency condition, the choice was between seven Dutch cities. In the high-agency / low personal relevance condition, participants received the instruction to choose a city they did not feel close to. In the low-agency / low personal relevance condition, participants had no choice and experienced the newsgame in the remote Dutch city Den Helder.

To play the game in English, open it in Chrome, right-click anywhere on the page and choose “Translate to English”.

7. Measures

All variables were measured through seven-point Likert scales. Transformative learning (DV) contained eleven items based on Roth’s (2016) eudaimonic appraisal scale and King’s (2009) Learning Activities Survey (LAS) (e.g., “I had an experience that caused me to question the way I normally act”, “I thought about acting in a different way from my usual beliefs and roles”) (α = 0.87). Perceived autonomy, serving as manipulation check, had three items (Roth, 2016) (e.g., “I had the impression that I was able to make many different events happen in the story”) (α = 0.69). Identification (MED 1) was measured through ten items (de Graaf et al., 2012) (e.g., “In my imagination I felt like I was the main character”) (α = .93). Transportation (MED 2) consisted of ten items, adopted from de Graaf et al. (2012) and adapted to fit the current material (e.g, “When I was experiencing the story, I no longer thought about other things on my mind”) (α = . 93). Perceived personal relevance was measured with the following item: “I feel connected to the city in which the story was set". Lastly, six true/false control questions were asked to check whether participants had not played I am Mosul before, and whether they had paid attention to the narrative (e.g., the false statement “The main character was shot.”).

8. Procedure

Due to COVID-19 regulations, the experiment ran completely online, with the research leaders being present for questions through Zoom for the participants via the university’s Human Subjects Pool. After providing informed consent, participants entered their age, gender and educational level. Then, participants were randomly assigned to one of the four conditions and were instructed that the newsgame should be experienced on a computer with sound on. The instruction that followed differed for each condition, as described in the Material paragraph. After this, participants played the newsgame I am Mosul in the version of their condition. Participants then filled in the items of transformative learning, identification, transportation, perceived autonomy and perceived personal relevance, followed by the attention check questions. Afterwards, participants were thanked for their time and effort and received a short debriefing explaining the objective of the study. The full experiment took approximately 15 - 25 minutes.

9. Data analysis

Firstly, people who had played I am Mosul before (n = 8) and people who made more than one mistake in the attention check questions (n = 7) were removed from the data. This made for a net total of 215 participants, who were initially assigned to either a personally relevant or personally irrelevant condition as described above. However, the perception of the participants was leading in whether or not the city was perceived as personally relevant. When participants in the personally relevant conditions indicated that they did not feel connected to the city (including neutral scores), they were moved to the personally irrelevant conditions (n = 16). When participants in the personally irrelevant conditions indicated that they did feel connected to the city, they were moved to the personally relevant condition (n = 29).

To test H1, we conducted a one-way ANOVA with agency as independent variable and transformative learning as dependent variable. To test H2 – H4, we conducted a moderated mediation analysis using Hayes’ (2022) PROCESS macro with agency as independent variable, transformative learning as dependent variable, identification (MED 1) and transportation (MED 2) as serial mediators, and personal relevance as moderator. Moderated mediation was established by means of Hayes’ (2022) index of moderated mediation, which represents the difference between the conditional indirect effects. As our conceptual model (see Figure 1) is not among Hayes’ (2022) preprogrammed models, we made a custom model using the B and W matrices in Table 1 and 2 (Hayes, 2022). These matrices specify which antecedent variables affect which consequent variables (B matrix), and which of these effects are estimated as linearly moderated (W matrix).

Table 1. B matrix specifying whether (1) or not (0) antecedent variables send an effect to consequent variables

	X	M1	M2
M1	1	0	0
M2	0	1	0
Y	1	1	1

Note. X = Agency, M1 = Identification, M2 = Transportation, Y = Transformative Learning

Table 2. W matrix specifying whether (1) or not (0) a path in the mediation model represented in the B matrix is linearly moderated by variable W

	X	M1	M2
M1	1	0	0
M2	0	0	0
Y	1	0	0

Note. X = Agency, M1 = Identification, M2 = Transportation, Y = Transformative Learning, W = Personal Relevance

Results

Table 3. Descriptive statistics

	Personal relevance	Agency	M	SD	n
Transformative learning	Low personal relevance	Low agency	4.12	1.02	47
		High agency	4.42	0.85	43
		Total	4.26	0.95	90
	High personal relevance	Low agency	4.46	0.87	59
		High agency	4.83	0.85	66
		Total	4.65	0.88	125
	Total	Low agency	4.31	0.95	106
		High agency	4.67	0.87	109
	Total		4.49	0.92	215
Identification	Low personal relevance	Low agency	4.02	1.24	47
		High agency	4.82	1.20	43
		Total	4.40	1.28	90
	High personal relevance	Low agency	4.63	1.16	59
		High agency	5.20	0.97	66
		Total	4.93	1.10	125
	Total	Low agency	4.36	1.23	106
		High agency	5.05	1.08	109
	Total		4.71	1.20	215
Transportation	Low personal relevance	Low agency	4.14	1.20	47
		High agency	4.89	0.96	43
		Total	4.50	1.15	90
	High personal relevance	Low agency	4.88	0.99	59
		High agency	5.01	1.13	66
		Total	4.95	1.06	125
	Total	Low agency	4.55	1.14	106
		High agency	4.96	1.06	109
	Total		4.76	1.12	215

Note. All consequent variables were measured on 7-point Likert scales. Higher scores indicate more transformative learning, identification, etc.

First, a one-way ANOVA showed that perceived autonomy was indeed higher for high agency (M = 4.84, SD = .86) than for low agency (M = 2.75, SD = 1.05) (F(1,213) = 255.085, p < .001), meaning that our manipulation was successful. To test Hypothesis 1, stating that high agency enhances transformative learning compared to low agency, a one-way ANOVA was done with agency as independent variable and transformative learning as dependent variable. Hypothesis 1 was supported: Transformative learning was higher for high agency (M = 4.67, SD = .87) than low agency (M = 4.31, SD = .95) (F(1,213) = 8.270, p = .004).

According to Hypothesis 2, identification and transportation mediate the effect of agency on transformative learning. Our PROCESS analysis indeed demonstrates a significant indirect effect, both with identification and transportation as serial mediators (low personal relevance: b = 0.14, 95% CI = [0.04, 0.26] | high personal relevance: b = 0.10, 95% CI = [0.03, 0.18]), and with identification as a single mediator (low personal relevance: b = 0.25, 95% CI = [0.07, 0.47] | high personal relevance: b = 0.18, 95% CI = [0.05, 0.35]).

Hypothesis 3 was not supported: there was no direct effect of agency on transformative learning, neither for low personal relevance (b = -0.15, 95% CI = [-0.44, 0.14]) nor for high personal relevance (b = 0.16, 95% CI = [-0.09, 0.40]). Hypothesis 4 stated that personal relevance moderates the indirect effect of agency on transformative learning via identification (in isolation or followed by transportation). Our results did not support this hypothesis either. The index of moderated mediation was not significant, neither for identification and transportation as serial mediators (b = -0.04, 95% CI = [-0.16, 0.06]) nor for identification as a single mediator (b = -0.07, 95% CI = [-0.28, 0.13]). We did find a direct effect of personal relevance on identification: identification was higher for high personal relevance than for low personal relevance (b = 0.61, 95% CI = [0.17, 1.04]). An overview of the results of our moderated mediation analysis can be found in Table 4.

Table 4. Bootstrapped moderated mediation analyses for the effect of agency on transformative learning through identification and transportation as a function of personal relevance

	95% Confidence Interval
	b	Lower	Upper
Direct effects
Low personal relevance
Agency -> Transformative learning	-0.15	-0.44	0.14
High personal relevance
Agency →Transformative learning	0.16	-0.09	0.40
Unconditional
Agency → Identification	0.80*	0.33	1.27
Identification → Transportation	0.65*	0.56	0.74
Identification → Transformative learning	0.31*	0.20	0.42
Transportation → Transformative learning	0.22*	0.10	0.34
Indirect effects
Low personal relevance
Identification	0.25*	0.07	0.47
Identification → Transportation	0.14*	0.04	0.26
High personal relevance
Identification	0.18*	0.05	0.35
Identification → Transportation	0.10*	0.03	0.18
Index moderated mediation
Identification	-0.07	-0.28	0.13
Identification → Transportation	-0.04	-0.16	0.06

Note. Standard errors of the bootstrapped estimates are displayed in parentheses; 5,000 bootstrapped samples.

* = significant; CI does not contain 0.

9. Discussion

This study investigated the effect of agency on transformative learning using the narrative newsgame I am Mosul (2019). We checked for mediation effects of identification and transportation and also investigated whether personal relevance served as a moderator. We found a positive effect of agency on transformative learning, confirming Hypothesis 1.The narrative engagement constructs identification and transportation served as mediator, supporting Hypothesis 2. Higher agency indeed led to more identification than low agency, which subsequently led to more transportation and to more transformative learning. Identification also served as a mediator independent from transportation. We did not find a moderating effect of personal relevance, neither for transformative learning in general (H3) nor for identification (H4). However, personal relevance did have a direct effect on identification: the players identified more with the main character when the game was located in a personally relevant city than in a personally irrelevant one.

One could argue that I am Mosul is too simple. While the newsgame allows players to make decisions, they still have to choose from the given– mostly binary–options, and they always end up back in their city because of the game’s foldback structure. Hence, the given options might not cover what players would like to do if they actually had been in the imagined situation; perceived completeness is low (Crawford, 2013). However, the limited autonomy in the gameplay is in line with the meaning of the given war situation in which refugees have limited autonomy, which results in so-called ludonarrative harmony (Roth et al., 2018). Plus, we did find a positive effect for this relatively simple newsgame. Below, we discuss our suggestions for future research to further enhance the effectiveness of a newsgame such as I am Mosul. 

A different issue to discuss, is the way in which the news in the newsgame is delivered. In I am Mosul, players must click on the “reality check” button to see how the game events relate to the actual events in Mosul. In addition, at the end of the game, a summary of the events in Mosul is given. However, during data collection, we noticed that most participants did not click on the “reality check” button. Ignoring this button can be seen as an indication of the transportation into the storyworld that our participants experienced. Clicking on the button would have meant stepping out of the storyworld, lowering transportation (Bilandzic & Busselle, 2017; Green et al., 2004), with a negative effect on transformative learning (Green et al., 2004; Zhou & Kim, 2022). We can also question how detrimental ignoring the button is for delivering Mosul’s story. The story in the game is similar to Mosul’s story in many important aspects: an IS-like group of armed fanatics takes over a city, and many people have to flee. After a while, they can return to their city only to find it in ruins. Furthermore, “Mosul” is prominent in the game’s name. All in all, we hold the opinion that the game succeeds in striking the balance between communicating the news, engaging the audience and transforming the audience’s frame of reference about refugees in a war situation.

9.1 Suggestions for future research

Although personal relevance did have a direct effect on identification in this study, it did not serve as a moderator: the effects of agency on identification and transformative learning did not differ between experiencing the game in a personally relevant or a personally irrelevant city. This null result is in line with Soto-Sanfiel and Igartua (2015). For a follow-up study, it would be interesting to increase the sample, to be able to do a more in-depth analysis of the effect of personal relevance (PR). Within the current study, participants were already allocated to the low PR condition when they were neutral about their connection to the city. In a similar vein, participants were already allocated to the high PR condition when they felt somewhat of a connection with the city. The suggested follow-up study would allow for a focus solely on the participants with a strong (dis)connection with the city in the story. In this study, you could also investigate the antecedents for feeling a (dis)connection with the city. Does the city size matter for instance or does it matter whether someone was born there or how many years and in which life stage they have lived there? Furthermore, we could question how unrelated the irrelevant Dutch cities actually were to participants. Within this small country, these were still nearby cities to which participants could easily relate. Other studies found positive effects of nearby compared to distant locations (Petty & Cacioppo, 1979) or compared to more general descriptions of locations (local versus global frames: Degeling & Koolen, 2022). A suggestion for future research would therefore be to compare one’s own location with a location further away (in this case: Mosul) or with an undefined location (applying a global frame).

We aim to extend our research as well by investigating transformative learning with interactive narratives with a more impactful disorienting dilemma, this dilemma being key to a transformative experience (Mezirow, 2000). I am Mosul is artificial in the sense that it does not confront players with a disorienting dilemma facing life-threatening consequences. Instead, the game finds ways of saving the player (deus ex machina), for instance when players want to search for food in an abandoned car, they are warned of mines and stopped. Plus, players will always end up alive and back in the war-torn but freed city. Newsgames such as I am Mosul could include more dramatical consequences to convey the meaning of war: a matter of life and death. Many don’t survive, one wrong action or being at the wrong place can end lives. This is in line with what Ryan (2011) calls ontological participation in a digital narrative text: “the life of the avatar is at stake: every run of the program creates a new life story for the avatar and a new history for the fictional world” (p.45). Players need to find strategies to survive, making the experience including the disorienting dilemma more real and potentially more transforming. Note that such narrative newsgames can be shocking and may give rise to feelings of anxiety and subsequent avoidance of the game (Hunt & Shehryar, 2011). It may not be a coincidence that, of the currently published newsgames (Gómez-García & De La Hera Conde-Pumpido, 2023), none make use of Ryan’s (2011) ontological participation. Excessively punitive consequences for suboptimal decisions may result in heightened frustration, premature disengagement, or psychological reactance. Calibrating the appropriate degree of discomfort is therefore critical in facilitating a genuinely transformative experience. Given the variability in individual thresholds for discomfort, a worthy focus in future research would be the development of a more adaptive model of Interactive Digital Narrative (IDN) design. Dynamically attuning to the interactor’s responses and capacities would be key in order to maintain engagement and support reflective transformation, also for other pressing topics such as climate change.

10. Conclusion

Overall, the results of this study largely substantiate our assumptions and demonstrate the potential for transformative learning of interactive digital narratives such as the newsgame I am Mosul (2019). We found that I am Mosul enhances transformative learning compared to a low-agency version of it. Having the agency to choose on behalf of the main character does invite a player to identify with this main character, seeing the world from a different perspective (Dillman Carpentier et al., 2015; Green & Jenkins, 2014; Hand & Varan, 2007; 2008; 2009; Peng et al., 2010). This higher identification, in turn, enables a transformative learning experience, directly and through transportation into the storyworld (Green & Jenkins, 2014; Hand & Varan, 2007; Zhou & Kim, 2022). Personal relevance did not moderate the effects of agency but it did directly affect identification: people identified more with the main character when they played the game in a personally relevant city than in a personally irrelevant city.

Journalists Frederick Mansell and Laurens Samsom created I am Mosul to raise awareness about the devastating effects of war on local citizens. They let their players step into the shoes of a citizen faced with war in the player’s own city. While keeping the agency limited and staying away from overly shocking content (which may lead to anxiety and avoidance of the game), they were able to deliver a newsgame that transformed their players’ frame of reference about the effects of war. For the design of newsgames, this demonstrates the potential of relatively simple role-playing interactive narratives as well as of geolocalization to engage audiences with news that they may feel disconnected to otherwise. 

While we agree with Bogost’s (2020) assertion that newsgames have yet to fully realize their potential, this study on I am Mosul demonstrates the promise of newsgames in their ability to inform and engage. This makes them a valuable addition to the world of journalism and interactive media.

Notes

1. We checked whether education level affected the results. This was not the case. ↩︎
2. During data collection in the lab, we noticed that most participants did not click on this “reality check” button. ↩︎

Acknowledgements

We thank journalists Frederick Mansell and Laurens Samsom for granting us permission to use their newsgame I am Mosul and giving us access to their source material. We also thank Dieuwertje Schipper who participated in this study for her Master’s thesis at Communication and Information Sciences at Tilburg University.  

Funding details

N/A

Declaration of interest

The authors report there are no competing interests to declare.

Data availability statement

The data and survey of this study are openly available on Dataverse at https://doi.org/10.34894/DTSZML.

References

1. Ajzen, I., Brown, T. C., & Rosenthal, L. H. (1996). Information Bias in Contingent Valuation: Effects of Personal Relevance, Quality of Information, and Motivational Orientation. Journal of Environmental Economics and Management, 30(1), 43–57. https://doi.org/10.1006/jeem.1996.0004
2. Bilandzic, H., & Busselle, R. (2017). Chapter 1. Beyond Metaphors and Traditions: Exploring the Conceptual Boundaries of Narrative Engagement. In F. Hakemulder, M. M. Kuijpers, E. S. Tan, K. Bálint, & M. M. Doicaru (Eds.), Linguistic Approaches to Literature (Vol. 27, pp. 11–27). John Benjamins Publishing Company. https://doi.org/10.1075/lal.27.02bil
3. Bogost, I. (2020). Curiosity Journalism, or the First Decades of Newsgames. Convergence: The International Journal of Research into New Media Technologies, 26(3), 572–576. https://doi.org/10.1177/1354856520918075
4. Bogost, I., Ferrari, S., & Schweizer, B. (2010). Newsgames: Journalism at Play. MIT Press.
5. Bonwell, C. C., & Eison, J. A. (1991). Active Learning: Creating Excitement in the Classroom. School of Education and Human Development, George Washington University.
6. Chen, J. C., & Martin, A. R. (2015). Role-Play Simulations as a Transformative Methodology in Environmental Education. Journal of Transformative Education, 13(1), 85–102. https://doi.org/10.1177/1541344614560196
7. Cohen, J. (2001). Defining Identification: A Theoretical Look at the Identification of Audiences with Media Characters. Mass Communication and Society, 4(3), 245–264. https://doi.org/10.1207/S15327825MCS0403_01
8. Cohen, J., & Tal-Or, N. (2017). Antecedents of Identification: Character, Text, and Audiences. In F. Hakemulder, M. M. Kuijpers, E. S. Tan, K. Bálint, & M. M. Doicaru (Eds.), Linguistic Approaches to Literature (Vol. 27, pp. 133–153). John Benjamins Publishing Company. https://doi.org/10.1075/lal.27.08coh
9. Cohen, J., Tal-Or, N., & Mazor-Tregerman, M. (2015). The Tempering Effect of Transportation: Exploring the Effects of Transportation and Identification During Exposure to Controversial Two-Sided Narratives: The Tempering Effect of Transportation. Journal of Communication, 65(2), 237–258. https://doi.org/10.1111/jcom.12144
10. Crawford, C. (2013). Chris Crawford on Interactive Storytelling (Second edition). New Riders.
11. de Graaf, A., Hoeken, H., Sanders, J., & Beentjes, J. W. J. (2012). Identification as a Mechanism of Narrative Persuasion. Communication Research, 39(6), 802–823. https://doi.org/10.1177/0093650211408594
12. Degeling, D., & Koolen, R. (2022). Communicating Climate Change to a Local but Diverse Audience: On the Positive Impact of Locality Framing. Environmental Communication, 16(2), 243–261. https://doi.org/10.1080/17524032.2021.1998177
13. Dillman Carpentier, F. R., Rogers, R. P., & Barnard, L. (2015). Eliciting Behavior from Interactive Narratives: Isolating the Role of Agency in Connecting with and Modeling Characters. Journal of Broadcasting & Electronic Media, 59(1), 76–93. https://doi.org/10.1080/08838151.2014.998222
14. García-Avilés, J. A., Ferrer-Conill, R., & García-Ortega, A. (2022). Gamification and Newsgames as Narrative Innovations in Journalism. In J. Vázquez-Herrero, A. Silva-Rodríguez, M.-C. Negreira-Rey, C. Toural-Bran, & X. López-García (Eds.), Total Journalism (Vol. 97, pp. 53–67). Springer International Publishing. https://doi.org/10.1007/978-3-030-88028-6_5
15. Gómez-García, S., & De La Hera Conde-Pumpido, T. (2023). Newsgames: The Use of Digital Games by Mass-Media Outlets to Convey Journalistic Messages. Games and Culture, 18(4), 449–474. https://doi.org/10.1177/15554120221105461
16. Green, M. C., Brock, T. C., & Kaufman, G. F. (2004). Understanding Media Enjoyment: The Role of Transportation into Narrative Worlds. Communication Theory, 14(4), 311–327. https://doi.org/10.1111/j.1468-2885.2004.tb00317.x
17. Green, M. C., & Jenkins, K. M. (2014). Interactive Narratives: Processes and Outcomes in User-Directed Stories: Interactive Narratives. Journal of Communication, 64(3), 479–500. https://doi.org/10.1111/jcom.12093
18. Hand, S., & Varan, D. (2007). Exploring the Effects of Interactivity in Television Drama. In P. Cesar, K. Chorianopoulos, & J. F. Jensen (Eds.), Interactive TV: a Shared Experience (Vol. 4471, pp. 57–65). Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-540-72559-6_7
19. Hand, S., & Varan, D. (2008). Interactive Narratives: Exploring the Links between Empathy, Interactivity and Structure. In M. Tscheligi, M. Obrist, & A. Lugmayr (Eds.), Changing Television Environments (Vol. 5066, pp. 11–19). Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-540-69478-6_2
20. Hand, S., & Varan, D. (2009). Interactive Stories and the Audience: Why Empathy is Important. Computers in Entertainment, 7(3), 1–14. https://doi.org/10.1145/1594943.1594951
21. Hayes, A. F. (2022). Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-based Approach (Third edition). The Guilford Press.
22. Hoeken, H., & Fikkers, K. M. (2014). Issue-relevant Thinking and Identification as Mechanisms of Narrative Persuasion. Poetics, 44, 84–99. https://doi.org/10.1016/j.poetic.2014.05.001
23. Hoeken, H., Kolthoff, M., & Sanders, J. (2016). Story Perspective and Character Similarity as Drivers of Identification and Narrative Persuasion: Perspective, Similarity, and Identification. Human Communication Research, 42(2), 292–311. https://doi.org/10.1111/hcre.12076
24. Hoggan, C., & Cranton, P. (2015). Promoting Transformative Learning Through Reading Fiction. Journal of Transformative Education, 13(1), 6–25. https://doi.org/10.1177/1541344614561864
25. Hunt, D. M., & Shehryar, O. (2011). Integrating Terror Management Theory into Fear Appeal Research. Social and Personality Psychology Compass, 5(6), 372–382. https://doi.org/10.1111/j.1751-9004.2011.00354.x
26. Jarvis, C. (2012). Fiction and Film in Transformative Learning. In E. W. Taylor & P. Cranton, The Handbook of Transformative Learning: Theory, Research, and Practice (1st ed, pp. 486–502). Jossey-Bass a Wiley Imprint.
27. King, K. P. (2009). The Handbook of the Evolving Research of Transformative learning, Based on the Learning Activities Survey (10th Anniversary edition). IAP, Information Age Publishing, Inc.
28. Klimmt, C., & Hartmann, T. (2012). Effectance, Self-efficacy, and the Motivation to Play Video Games. In P. Vorderer, Playing Video Games: Motives, Responses, and Consequences (1st ed., pp. 133–145). Routledge. https://doi.org/10.4324/9780203873700
29. Koenitz, H. (2023). Understanding Interactive Digital Narrative: Immersive Expressions for a Complex Time (1st ed.). Routledge. https://doi.org/10.4324/9781003106425
30. Kuzmičová, A., & Bálint, K. (2019). Personal Relevance in Story Reading. Poetics Today, 40(3), 429–451. https://doi.org/10.1215/03335372-7558066
31. Lin, J.-H. T., & Wu, D.-Y. (2020). Newsgames for the Greater Good: The Effects of Graphic Realism and Geographic Proximity on Knowledge Acquisition and Willingness to Help. Journalism & Mass Communication Quarterly, 97(1), 30–51. https://doi.org/10.1177/1077699018820315
32. López-Arcos, J. R., Vela, F. L. G., Padilla-Zea, N., Paderewski, P., & Medina-Medina, N. (2017). Introducing an Interactive Story in a Geolocalized Experience. Proceedings of the XVIII International Conference on Human Computer Interaction, 1–7. https://doi.org/10.1145/3123818.3123850
33. Mezirow, J. (2000). Learning as Transformation: Critical Perspectives on a Theory in Progress (1st ed). Jossey-Bass.
34. Mezirow, J. (2012). Learning to Think Like an Adult. Core Concepts of Transformation Theory. In E. W. Taylor & P. Cranton, The handbook of transformative learning: Theory, research, and practice (1st ed). Jossey-Bass a Wiley Imprint.
35. Murray, J. H. (2017). Hamlet on the Holodeck: The Future of Narrative in Cyberspace (Updated edition). The MIT Press.
36. Navarro-Remesal, V., & Zapata, B. P. (2019). First-Person Refugee Games: Ludonarrative Strategies for Playing the Stories of Refugees and Asylum Seekers. In N. Zagalo, A. I. Veloso, L. Costa, & Ó. Mealha (Eds.), Videogame Sciences and Arts (Vol. 1164, pp. 3–17). Springer International Publishing. https://doi.org/10.1007/978-3-030-37983-4_1
37. Peng, W., Lee, M., & Heeter, C. (2010). The Effects of a Serious Game on Role-Taking and Willingness to Help. Journal of Communication, 60(4), 723–742. https://doi.org/10.1111/j.1460-2466.2010.01511.x
38. Petty, R. E., & Cacioppo, J. T. (1979). Issue Involvement Can Increase or Decrease Persuasion by Enhancing Message-Relevant Cognitive Responses. Journal of Personality and Social Psychology, 37(10), 1915–1926. https://doi.org/10.1037/0022-3514.37.10.1915
39. Plewe, C., & Fürsich, E. (2018). Are Newsgames Better Journalism? Empathy, Information and Representation in Games on Refugees and Migrants. Journalism Studies, 19(16), 2470–2487. https://doi.org/10.1080/1461670X.2017.1351884
40. Quintero Johnson, J. M., Harrison, K., & Quick, B. L. (2013). Understanding the Effectiveness of the Entertainment-Education Strategy: An Investigation of How Audience Involvement, Message Processing, and Message Design Influence Health Information Recall. Journal of Health Communication, 18(2), 160–178. https://doi.org/10.1080/10810730.2012.688244
41. Rigby, C. S., & Ryan, R. M. (2016). Time Well-spent? Motivation for Entertainment Media and Its Eudaimonic Aspects Through the Lens of Self-Determination Theory. In L. Reinecke & M. B. Oliver (Eds.), The Routledge Handbook of Media Use and Well-Being: International Perspectives on Theory and Research on Positive Media Effects (pp. 34–48). Routledge Taylor & Francis Group.
42. Roth, C. (2016). Experiencing Interactive Storytelling [Vrije Universiteit]. https://research.vu.nl/ws/portalfiles/portal/42167962/complete+dissertation.pdf
43. Roth, C. (2019). The ‘Angstfabriek’ Experience: Factoring Fear into Transformative Interactive Narrative Design. In R. E. Cardona-Rivera, A. Sullivan, & R. M. Young (Eds.), Interactive Storytelling (Vol. 11869, pp. 101–114). Springer International Publishing. https://doi.org/10.1007/978-3-030-33894-7_11
44. Roth, C., & Koenitz, H. (2016). Evaluating the User Experience of Interactive Digital Narrative. Proceedings of the 1st International Workshop on Multimedia Alternate Realities, 31–36. https://doi.org/10.1145/2983298.2983302
45. Roth, C., Van Nuenen, T., & Koenitz, H. (2018). Ludonarrative Hermeneutics: A Way Out and the Narrative Paradox. In R. Rouse, H. Koenitz, & M. Haahr (Eds.), Interactive Storytelling (Vol. 11318, pp. 93–106). Springer International Publishing. https://doi.org/10.1007/978-3-030-04028-4_7
46. Ryan, M.-L. (2011). The Interactive Onion: Layers of User Participation in Digital Narrative Texts. In R. Page & B. Thomas (Eds.), New Narratives: Stories and Storytelling in the Digital Age (pp. 35–62). UNP - Nebraska Paperback. https://doi.org/10.2307/j.ctt1df4h49
47. Scholl, J., Pandrea, M., & van Enschot, R. (2022). How to Help Your Depressed Friend? The Effects of Interactive Health Narratives on Cognitive and Transformative Learning. Frontiers in Communication, 7, 966944. https://doi.org/10.3389/fcomm.2022.966944
48. Soto-Sanfiel, M., & Igartua, J.-J. (2015). Cultural Proximity and Interactivity in the Processes of Narrative Reception. International Journal of Arts and Technology, 1–21.
49. van Krieken, K., Hoeken, H., & Sanders, J. (2017). Evoking and Measuring Identification with Narrative Characters – A Linguistic Cues Framework. Frontiers in Psychology, 8, 1190. https://doi.org/10.3389/fpsyg.2017.01190
50. Zhou, C., & Kim, S. (2022). Testing the Effect of an Interactive Narrative: The Mediating Role of Transportation and the Moderating Roles of Narrative Ending and Issue Involvement. Health Communication, 1–10. https://doi.org/10.1080/10410236.2022.2126693