Learning to design through character identication using a gamication learning app: Effects of learning quests on students’ design behaviours and performance

Character identication occurs when game players adopt the identities of game characters to different degrees. Character identication can enhance identity development and learning of young adolescents but how higher education students in professional programs develop character identication using a gamication learning mobile app and use it to support their learning of professional skills and development of professional identities have not been examined. This study describes the design of a character-based gamied learning quest to support the learning of design skills. Through 13 learning quests, 4 3 tertiary design and media students used the professional media roles and ‘power’ attributes of their chosen characters towards the design of an innovative media product. The study investigates the developed character identication for their professional. Content analysis of students’ video blog reections indicated that more than half of them were able to establish deep identication with their chosen characters. Students who had deep identication enacted their character by actively ideating, prototyping and testing design ideas whereas those who had shallow or no identication largely focused on ideation. Regression analysis showed that deep character identication during design conceptualization predicted students’ design project scores. How the study ndings can inform the practical applications of character-based gamication in higher education professional learning programs are discussed. through active identication that have resultant behaviours in real-life. The results show that the use of character-based learning quests has been effective in motivating some form of character identication among majority of the students and this is in line with studies of game players’ identication with game characters The study results also reveal that the level of students’ character identication may not remain constant and there was slightly less than


Introduction
Character identi cation is a strategy used to create immersive experiences in digital video games and massively multiplayer online games (MMOGs) (Burgess & Jones, 2020a;Sierra Rativa et al., 2020). Just as one can develop one's character by imitating the characteristics of people that one identi es with, one can also imagine themselves as "becoming" media characters by temporarily giving up consciousness of their own identities (Cohen, 2001). The more that people exchange their perspectives of identity with the media characters, the more intense the level of character identi cation (Cohen, 2001). The use of character identi cation is seen in gami cation for education (Jackson, 2016;Oxford Analytica, 2016) that adopt characters or avatars as the game design element (Deterding et al., 2011) to characterize and personalize student identities (Martin & Tyler, 2017). Cohen (2006) proposes that identi cation can bring about educational bene ts such as better concentration and retention of the messages that characters bring. More importantly, character identi cation can help students to develop the positive attributes of the characters that they aspire to be. During gameplay, characters are agencies that support mental rapprochement between players and characters (Burgess & Jones, 2020;Klimmt et al., 2009). Students can possibly develop con dence, motivation, and direction as they identify with game characters (Hoffner & Buchanan, 2005) and this helps them to develop and transfer different knowledge and skills from games to reality (Barnett & Coulson, 2010;Erikson, 1968;Mazalek et al., 2013). However, in the education, character identi cation is the least common gaming mechanism and preferences as compared to badges, points and leaderboards (Gachkova et al., 2020;Jackson, 2016;Majuri et al., 2018). Identi cation involves psychological matching (Bandura, 1986). MMOG studies reveal that game players have different kinds of character identi cation (Burgess & Jones, 2020b). Li et al. (2013) assessed player-avatar identi cation by the feelings during play, absorption during play, positive attitudes towards avatars and perceived importance to identity. It was found that primary and secondary students can develop conciseness of identity through gameplay. Other studies described different levels of identi cation. At a shallow level, players view themselves as being separate from their characters, having similarity identi cation, meaning, similar experiences, feelings and perspectives as their character (Hoffner & Buchanan, 2005) or wishful identi cation where they desire to match characters by reducing their psychological self-discrepancy (Bessière et al., 2007;Klimmt et al., 2009). At a deeper level, players unconsciously adapt their behaviour to characters to the point described as having embodied presence with characters (Van Looy et al., 2012). Some level of similarity or wishful identi cation can help game players to be more satis ed with their characters, have more enjoyable game experiences and better game performance , Klimmt et al., 2009. The more that game players perceived embodied presence, the more they internalized emotions and empathized with the messages delivered by the subject matter of the game (Van Looy et al., 2012). Oksanen et al. (2013) called attention to the potential effects of character identi cation for learning. In an interactive educational game for learning programming, character identi cation improved the quality of programs created by students as well as their intrinsic motivation and self-e cacy for programming (Kao & Harrell, 2018). In an interactive serious game for civics and moral education that is set in the aftermath of the 2010 Haiti earthquake, high school students who have embodied identi cation with their playing character had higher levels of global empathy and interest in the subject matter (Bachen et al., 2012(Bachen et al., , 2016. Character identi cation can help students who are adolescents and young adults to explore and develop their identities as they play different roles with game characters (Li et al., 2013;Van Looy et al., 2012). As higher education institutions confront the need to produce industry ready graduates (Brennan et al., 2014), character identi cation can perhaps be used as another way of helping students to establish their professional identities, talents and potential before they enter the industry. Nevertheless, this is still an area needing more examination as serious games used in business and engineering courses have more often been evaluated in terms of learning outcomes (Bulander, 2010) but these have shown mixed ndings (Bakhuys Roozeboom et al., 2017). The need for strong professional identity is critical for students who are entering creative industries such as design and media because their professionalism is highly intertwined with their ability to demonstrate unique creative voices that are driven from strong artistic identities (Lena & Lindemann, 2014). How character identi cation in uences the learning and professional identities of higher education students still need to be further investigated.
In view of the extant research gaps, this study examines the character identi cation that 43 Design and Media students at a higher education arts institution developed through playing character-based learning quests using a gami cation learning app. The following research questions are examined to develop implications for using character-based gami cation for higher education professional learning: (1) What kinds of character identi cation do design and media students adopt after playing characterbased learning quests?
(2) Do students with different levels of character identi cation manifest different design behaviours?
(3) How is design performance in uenced by character identi cation?

Study participants and module context
The study participants were 43 students majoring in design and media at a higher education arts institution in Singapore. Following ethics approval by the institution's research evaluation panel, students' informed consent was sought for study participation. All the students gave consent for study participation and there were 19 (44%) males and 24 (56%) females. The students were attending the Alternative Applications course during the August semester of 2019. The course duration is 15 weeks, allocated one lesson of four hours per week. It is a project-based course where students need to draw upon the ruins of the ancient Ayutthaya kingdom in Thailand, a UNESCO world heritage site, as creative stimuli to develop an innovative group-based design and media solution. A total of 12 teams with three to four students per team was formed. For this study, the learning tasks were designed with character-based gami cation and presented to students in the form of learning quests that were driven by a meaningful storyline (Giakalaras, 2016) that is related to saving the Ayutthaya kingdom. The course activities were designed to rst guide students to form their own team after understanding their own and peers' characters. The activities also support students to research, conceptualize, and develop their team projects. Students were free to express their creativity with different screen-based mediums, supported by the design thinking process to produce an alternative creation from Ayutthaya-inspired themes. The ve steps of design thinking were used as a process to stimulate creative and alternative ways of thinking (Owen, 2006;Plattner, 2013, Plattner et al., 2014. In this process, design idea is rst developed by having Empathy for the consumers of their design products prior to De ning their design problem, Ideating, Prototyping, and Testing their design ideas (Carlgren, 2013). In the study context, students learnt how to apply this design thinking to Ayutthaya as they must create a design and media product inspired by Ayutthaya. Thus, students need to develop empathy for Ayutthaya's history, the Ayutthaya kingdom and their citizens as a starting point of design.

Design of character-based gami ed learning quests
The study was conducted during the rst seven weeks of the semester where students conducted background research and developed their design conceptualization. The ctional narrative of the learning quests was about a group of newly chosen Treasure Makers who were on a mission to uncover the old treasures of Ayutthaya for inspiration to create new treasures (media-based applications) that could be used to save Ayutthaya from being demolished for urbanization. This was to impart the message about the universal value of conserving the historic city of Ayutthaya. Each student role-played a Treasure Maker with a character of their choice, made decisions and experienced the story through the character's eyes (Cohen & Tal-Or, 2017;Marache-Francisco & Brangier, 2013). The quests unfolded according to a storyline embedded with hints and strategies to support problem-solving for the quest challenges. Each learning quest had gami ed learning content and tasks (Kingsley & Grabner-Hagen, 2015) where students earned experience points and achievement badges to level up their character's status rank throughout the course. The quests were created using an existing commercial gami cation learning mobile app (See Figure 1). It is suitable for the study as it allows educators to create their own quest video and students to submit their quest challenge video back for review and gain experience points.
Students chose their characters during the rst lesson by playing the quest "Who Are You?" The 20 characters used in the quest were created and validated against the Personality-related Position Requirements Form (PPRF) (Raymark et al., 1997). These are Explorer, Mingler, Protector, Intellectual, Independent, Nurturer, Fighter, Communicator, Strategist, Idealist, Associate, O cial, Celebrity, Champion, Creator, Controller, Investigator, Leader, Scholar and Founder. To foster professional identi cation, each character was further related to professional roles in the design and media profession such as illustrators, animators, media producer, manager, and writer. For example, the Creator character was related to the roles of artist, writer, mobile app developer, product designer, technologist etc. It also comes with a list of "powers" that describe its personality traits and abilities, such as optimistic, perfectionist, spontaneous, curiosity, and idealistic etc. (See Figure 2).
After choosing two characters, students also chose two roles and three "powers" from each of the two characters to represent themselves in their team (See Figure 3). A debrief was conducted with students after the quest to discuss how they could draw upon the roles and "powers" of their characters for their team project. After the debrief, the students continued to play another quest "Find Your Comrade" where they were to "speed date" to nd their own teammates by introducing their roles and "Powers". They were guided in the quest to form a team of different yet complementary roles and abilities.
Students then played two other learning quests designed to help them develop and pitch their individual design ideas to their team members. The quest "Are You Taya?" guided students to research on the historical background of Ayutthaya based on one out of the four themes of their choice: cultural and lifestyle, architecture and landscape, economic and trade, and technology and innovation. This quest was narrated by a ctional non-player character, named "Taya", a local Ayutthaya citizen. The students were led by "Taya" to research deeper into one theme. After completing the quest, students had to do a blog post on the course blogsite to present their ndings to their team. The second quest "Tuk Tuk" helped students to prepare a creative concept for a new treasure for saving Ayutthaya that is developed from the Ayutthaya values of pride, enterprise and adaptability. This treasure seeks to use inspiration from the old world to develop a new screen-based creation that could bene t today's generation. The students were again guided by the "Taya" character. After completing the quest, students had to integrate their concepts as a team in order to pitch their nal concept to the lecturers. To stimulate thinking about design processes, students made two individual re ections as video blogs (vlogs) with a duration requirement of within ve to nine minutes. In vlog 1, they re ected upon their processes for design research and how they planned to contribute to their team project. In vlog 2, they re ected upon how they developed design concepts to support their team project. The vlogs were graded as part of their individual weekly progress marks.
Students made project submissions during week 7 which constituted their mid-term assessment. The weightage of the assessment includes 60% based on team deliverables which include team project proposal and team concept poster artwork, and, 40% based on individual deliverables which consists of individual presentation slide deck as well as creative process journal and vlogs. The students spent 28 on-course hours to work on their project and majority of them spent additional off-course hours to complete the project.

Data collection and analysis
The following data sources were used to answer the research questions.

Research question 1 -character identi cation
Research question 1 was answered through content analysis (Weber, 1990) of the vlogs that students submitted. Out of the 43 students, 39 of them submitted vlog 1 while 35 of them submitted vlog 2. Thus, a total of 39 vlog 1 and 35 vlog 2 were collected. Each vlog constituted a unit of analysis. The videos were watched repeatedly and students' utterances as well as critical gestures (e.g., pointing a speci c part of their artwork) were transcribed into text. Van Looy et al.'s (2012) framework was used as a starting point for coding, but constant comparison (Creswell, 1998) revealed the need for adaptation. Some students expressed no speci c identi cation whereas those with embodied identi cation also shared ideas for enhancing embodiment of desirable character attributes. A second rater coded a random selection of 40 vlogs (54.05%) and a Cohen's kappa of 0.84 established the reliability of coding through which the coding protocol with three categories was established (See Table 1). The frequency distributions of students expressing different categories of character identi cation for vlogs 1 and 2 were analysed to answer the research question. Research question 2 -character identi cation and design behaviour Content analysis (Weber, 1990) of the vlog transcripts was used to answer research question 2. The transcribed text for each vlog was broken down into messages in the form of unique ideas as a unit of analysis (Daniel & Harland, 2017). A total of 100 units of analyses were derived. Each unit of analysis was coded with the ve steps of design thinking (Plattner, 2013;Plattner et al., 2014) to understand the creative processes that students used to research and create their design concepts. Two new categories, Plan and Re ect emerged from the analysis as students also revealed what they did before and after applying the design thinking processes. A second rater coded a third of total units that were selected randomly to derive a Cohen's kappa of 0.84. Adequate reliability was established, and the coding protocol was set up (see Table 2). The units of analyses were organized according to students' character identi cation that was established in research question 1. Thematic analysis (Braun & Clark, 2006) was used to identify the common features of students' design behaviour by their character identi cation. Research question 3 -character identi cation and design performance Each students' project submission was graded with the ve dimensions from the course assessment rubrics -(1) creativity and innovation (2) application of knowledge, skills, research, and experience (3) amount of effort (4) execution competency, and (5) uni cation of cohesiveness, coherence, and relatedness. Stepwise regression was used to answer the research question with project grade as the dependent variable. The three levels of character identi cation that emerged from research question 1 for each vlog were converted to dummy variables and used as independent variables.

Results
Research question 1 -students' character identi cation Table 3 shows students' character identi cation that were expressed through each vlog. Vlog 1 In vlog 1, students re ected about how they carried out background research for their design. Table 4 shows that about 23% of the students did not mention character identi cation and they made general comments about their research ndings. For example, Student 1 reported that, "There are the eight temples that one has to visit in Ayutthaya." Slightly more than a quarter of the students had shallow identi cation. These students used the roles and 'powers' that they desired to draw out interesting insights about Ayutthaya, as Student 2 shared: The role I used is Social Media Trendsetter. [power] where I found a humorous side story of the King Naresuan that I believe not many know i.e., King Naresuan's father was an elephant. I thought that it would be quite funny as a headline in my blog post...

Vlog 2
In vlog 2, students re ected on how they and their team developed their design concept. Slightly more than one third of the students did not draw upon character identi cation. These students described their design concept through random items or subjects, one of their interests, one real-life problem to solve and one of the four themes of Ayutthaya. Students had to associate them together to form a new creative concept. Student 4 explained: We chose [the] oating wet market in Ayutthaya for our treasure. The problem is environmental pollution. The theme is lifestyle and culture, our interest is travelling. Random word is "NASA" (National Aeronautics and Space Administration). We combined all of them to derive the [idea that the] littering problem in Ayutthaya can be solved by the data generated from the space satellite where we can see the littering in the river of Ayutthaya. We [will] share these data with the local and tourists in Ayutthaya so that they aware about the littering.
About 14% of the students expressed shallow identi cation. These students articulated more details and had clearer vision about their design concept as compared to those without identi cation, as shown in Student 5's re ection: We want to do an interactive animated storyboard website to spread awareness around the world so that they know about the ooding problem in Ayutthaya... Many do not know how bad it is and how it [has] affected the historical temples. My role in the group is a conceptual artist, and I am the one to do most of the drawing for the website. It is going to be amazing as the story will let people realize and spread awareness to save Ayutthaya.
Slightly more than half of the class had deep identi cation with their character. These students emphasized how their roles and 'powers' helped them to contribute to their teams. Student 6 described: Being a journalist and writer, I contribute through these roles by attending the group discussion. During the discussion, I always write key points. By doing so, it helps me to catch up with the discussion, further improve and develop the points so that the concept can become more innovative. The 'powers' that I used during the discussion were 'positive' and 'optimistic' because in the process of doing the creative proposal, I felt that our group seemed to be discouraged. We haven't gotten the perfect idea about the innovation that we want to create... so by using the powers that I have, I encouraged my team members and myself to persevere and continue thinking of [the] innovation that we can create by doing what we can. Eventually, our group had our own creative idea which made me feel relieved and proud…

Shifts in character identi cation
From the 34 students who submitted both vlogs, it was found that the level of character identi cation remained consistent for slightly more than half of them (n=18, 52.9%). Seven students (20.6%) shifted towards deeper levels of identi cation during vlog 2 (from none to shallow or deep; or from shallow to deep). The rest used shallower forms of identi cation (from deep to shallow or none; or from shallow to none). Analysis of the vlogs showed that the congruence between the design task and roles selected by students may in uence their character identi cation. For example, Student 7 chose the roles of Designer, Artist and Musician and the 'powers' of 'creative' and 'original'. She did not apply these roles to the research design task of vlog 1 but was able to nd shallow identi cation for the design conceptualization task of vlog 2. Students who maintained some form of character identi cation were exible and imaginative in manoeuvring the attributes of their character. For example, Student 5 switched from the role of Writer to Artist in vlog 2 as this was more applicable for design conceptualization.

Research question 2 -students' design behaviour with respect to character identi cation
Design behaviour for vlog 1 Analysis of results from table 4 shows that students with different kinds of character identi cation approached the task of researching their design problem differently. A total of 140 occurrences in all vlogs were identi ed. The nine students who did not mention their characters in vlog 1 tend to jump right into Ideation where they extracted literal ideas from their research. For example, Student 8 shared that: Ayutthaya is the ancient capital in Siamese Kingdom. It was strategically located on an island surrounded by three rivers connecting the city to the sea, so it was a good trading point for the traders.
[The treasure found is] the hand-weave fabric for trade... The ten students who had shallow identi cation made more attempts to Empathize, De ne, and Ideate. For example, Student 9 shared: I decided to blog on the symbology and features of the Ayutthaya temples. I read up on chedi (a Buddhist stupa) and why it is bell-shaped? (De ne). It was interesting. It is so old and so much history behind it…. It is important to us because this is our past (Empathy). I found out that chedi is inspired by Hindu, I mean from India. I am interested in reading up architecture as it explains the culture and traditions, [for example], temples as a record of who and what they are, and how they are before (Ideate).
In comparison, the students who had deep character identi cation had a more holistic design approach.
For example, Student 10 explained the 'powers' that helped her to move from Ideation to the stages of Prototype and Test. As she designed the blog post of her research presentation, she drew upon the 'power' of 'sensitive' in her role as Project Manager. She prototyped her solution by being sensitive "towards how readers are reading my blog post." She prototyped her blog design with "clear subtitles and pictures to engage them to make reading more enjoyable." She also tested her prototype by drawing inspiration from the power of "pursing goals." The achievement of her "goal in mind to save Ayutthaya" became her way of testing if she has made a "good blog" that could pitch her ideas well to her team.
Some students with deep character identi cation also explained how they drew upon their characters to Plan their design processes with their team. For example, Student 11 said, "The 'powers' are leading team, organizing... I hope to be useful to the team and organize things for the team to get better. ... I tried to manage my research on what to do rst."

Design behaviour for vlog2
Analysis of results from table 5 shows that the general trend of students' design behaviours was largely consistent with those reported during vlog 1 when they re ected about their design conceptualization processes in vlog 2. However, there was more evidence of engagement in Re ection. No. of students who submitted vlogs = 35 As students worked on their design conceptualization, those who did not draw upon character identi cation still focused on Ideation but also made attempts to Empathize and De ne. Referring to the issue where many did not know Ayutthaya's traditional foods, Student 12 de ned their group's proposition for promoting Ayutthaya through a food recipe app for foreigners at Ayutthaya to "feel at home even at foreign country." The team empathized with the fact that foreigners would desire to "cook their home food and able to taste international food" while at Ayutthaya and their idea was for their app to "provide recipes from all round the world and sent the ingredients to their doorstep which they ordered from us." Some students with shallow character identi cation made attempts to Test and Prototype. For example, Student 7 re ected how she created the design sketches that were shared with her team-mates and the lecturers for feedback: I use my roles as Artist and Designer to make new sticker design. I also design tote bag. This is the sketch (showing her sketch of tote bag), this is the colour sketch (showing her colour sketch), and this is the nal design (show her nal sketch of used water bottles collage into a whale). This week, I used my 'power' of creativity to create the sticker and [tote bag] designs.
In comparison, Student 13 who had deep identi cation De ned why his team chose to produce a game app to spread knowledge about Ayutthaya through Empathizing with the need of people to have fun because, "for us, we want to bring card game and recipe together because the problem is that many do not know the treasures in Ayutthaya. You can have fun playing the card game and gain knowledge at the same time." Drawing upon his role as Product Designer with the 'powers' of 'focus', 'one-of-a-kind' and 'friendly', he took initiative to develop a prototype of the card game concept in Adobe XD that was tested with his team-mates. The positive reception from his team-mates a rmed his efforts to manifest his 'powers', "because I tried my best to being one-of-a-kind to bring the best of me to everyone around me." He also re ected on the positive effects of drawing upon his "power" where, "I really enjoy doing this kind of card game because it gives me a positive aura …" Research question 3 -students' design performance by character identi cation Stepwise regression found that only a deep level identi cation in vlog 2 signi cantly predicted project score which explained 42% of the model variance (See Table 6).

Discussion
These ndings reveal the following about using character identi cation in educational contexts: Character identi cation process Bandura (1969) asserts that students could pattern their characters' actions through active identi cation that have resultant behaviours in real-life. The results show that the use of character-based learning quests has been effective in motivating some form of character identi cation among majority of the students and this is in line with studies of game players' identi cation with game characters Klimmt et al., 2009;Yee et al., 2009;Yee & Bailenson, 2007). The study results also reveal that the level of students' character identi cation may not remain constant and there was slightly less than half of the students shifted their identi cation according to congruence between their character attributes and the design task. As compared to the results of character identi cation study from Bachen et al. (2016), this shows that the development of character identi cation in educational contexts cannot be assumed as the kinds of learning tasks and how students align themselves professionally with their characters can in uence the level of their character identi cation (Hoeken et al., 2016).

Character identi cation in uences design behaviour
Character identi cation can help students to translate character behaviours into professional practice (Barnett & Coulson, 2010). The study results show that students' professional behaviours could vary according to the extent of their character identi cation. Analysis of students' vlogs reveal that when they did not establish character identi cation, they focused on creating design ideas with little problematization, which is typical of novice designers (Lawson, 1997). With shallow identi cation, students were able to draw upon some desirable character attributes to support design problematization through applying Empathize and De ne of the design thinking process. Deep identi cation appears to encourage positive action as students sought to translate character attributions into action. Like good designers who spend more time developing artefacts and exploring ideas (Lawson, 1997), these students were more able to move their ideas forward by making Prototypes to Test their ideas as part of design thinking.

Character identi cation in uences design performance
Several studies have indicated that character identi cation enhanced learning performance (Bachen et al., 2016;Cohen, 2006;Kao & Harrell, 2018;Oksanen et al., 2013). However, the study results suggest that these relationships are not straightforward in the context of design learning. The regression analysis suggests that design performance is determined by how students engage character identi cation for different design tasks. It appears that character identi cation during design research does not predict design performance whereas only deep identi cation predicts design performance during design conceptualization. This could be because design conceptualization has more direct linkages with students' project deliverable whereas deep identi cation supports students to engage in more comprehensive design behaviours. The combination of these factors in turn improves their design performance. These ndings support and build upon existing evidence of the previous research on character identi cation whereby students must identify their own identity and personality with similar roles and attributes of the character to order achieve deep identi cation. Furthermore, this study a rmed that imagination (Kusuma et al., 2018;Bal, 2019) is needed for gami cation which leverages on game design elements to create engagement and imagination (Kapp, 2012). This can be hypothesized that imagination is needed to assume deep identi cation (Cohen, 2006).

Implications for higher education
The ndings of this study suggest the following implications for the adoption of character identi cation as a strategy for higher education learning and application into educational technology.
Design characters to support professional identi cation Character identi cation can be used as a gami cation strategy to help students identify professional roles and identities. The study ndings suggest that in order to foster professional learning, game characters need to be designed in ways that are not only identi able to the personality traits of students (Hoffner & Buchanan, 2005) but also relatable to their professional roles (Bandura, 1986;Klimmt et al., 2009). These kinds of relatedness help students to develop interest in learning (Bachen et al., 2012) as well as deeper understanding and identi cation with their professional roles.
Create imagination with character attributes Cohen (2001) suggests that the intensity of character identi cation is highly dependent upon one's capacity for imagination. He de ned identi cation as an imaginative process through which one assumes the identity, goals, and perspective of a character (Cohen, 2006, p. 261). Through gami cation, the students can attach, project and experiment with the characters. Getting "into" the character did not appear challenging as slightly more than half of the students were able to develop deep identi cation even from the rst design task. For these students, character identi cation served as a creative stimulus for their design thinking. However, there is also another half of the students who were not able to connect their character attributes to the design tasks. These results suggest that students' capacity for imagination cannot be assumed. When designing lessons with character identi cation, strategies to help students customize and fantasize (Van Looy et al., 2012) ways to exploit characters in professional thinking need to be considered.
Discover students' identity and talent Talent is one's self-perceptions of one's goals, interests and talents in one's career (Ibarra, 1999). Ricoeur (1992) believes that character, personality and identity coincide. The study results show that character identi cation helped students to articulate and activate positive attributes that can be a way of helping them to discover their identity (Martin & Tyler, 2017), thus, forming their talents. The ability to form deep identi cation with characters gives students direction and con dence to enact character attributes through their design work. This in turn improved design performance as students did not learn and design blindly but were consciously rationalizing their design thinking. Character identi cation can be used to trigger talent awareness among higher education students, giving them direction to exhibit their talents in their learning projects. It can also help lecturers to understand students' pro les and aspirations and provide better individualized support.

Develop characters into gami cation educational technology
Character identi cation can be designed into character-based gami cation educational technology and learning environments such as mobile and desktop apps and Massive Open Online Courses (MOOC).
Students can easily identify and continue develop their professional roles and talents throughout their studies with these platforms. Lecturers can also easily access students' pro les to better guide students with their projects.

Limitations and Future Research
There are several limitations in this study that could constitute areas for future improvement and research. Firstly, the study examined students' one-time selection of roles and 'powers' of the characters at a point in time whereas their desired roles and characters may change over time (Bialystok, 2009).
While some insights on students' character identi cation shifts have been uncovered in the study, further study into the factors and reasons for identi cation shifts are needed. Secondly, the study examined students' re ections and self-described design processes. In future studies, student interviews could be carried out to further probe why they identi ed with characters and the character design elements that supported their identi cation. Thirdly, the study did not consider in team-based contexts even though the module did require teamwork among students to complete a project. Future studies could investigate the effects of teamwork on character identi cation as well as how game elements shape character identi cation. Finally, this study focused on design and media students. The applicability of the ndings to other elds of study in higher education can be further examined.

Conclusion
In this study, there is evidence that character identi cation can in uence the learning processes and learning performance of design and media students. Character identi cation seems to be an important yet overlooked component in gami cation design. The types, levels and shifts of identi cation are factors that need further consideration for better character identi cation experiences and effectiveness in educational gami cation.

Declarations Compliance with Ethical Standards
Disclosure of potential con icts of interest: