Discussion

Impact on Feedback Quality (RQ1)

The first research question asked: "Does a game-based learning intervention increase the quality of feedback provided during Computer Science peer code review?" Following the intervention, students' feedback became more detailed and specific compared to earlier rounds of traditional PCR, suggesting a potential relationship between the GBL context and increased feedback quality. This aligns with previous research showing that peer review quality improves when students are more invested in the process and understand its purpose (Huisman, Saab, Van Driel, & Van Den Broek, 2018; Indriasari, Luxton-Reilly, & Denny, 2020).

A key design element of the intervention was the integration of academic performance into the logic of the game. Students' feedback scores influenced the number of yellow action cards they received, which shaped their strategic options during play. This structural connection reflects the concept of meaningful play, in which actions must have visible and integrated consequences within the system (Salen & Zimmerman, 2003). Rather than functioning as an isolated academic task, peer feedback became a meaningful in-game action that affected future possibilities. This aligns with prior studies emphasizing that feedback tasks are more effective when they are situated in authentic, high-stakes contexts (Brown, Walia, Radermacher, Singh, & Narasareddygari, 2020).

Many students described this relationship as motivating. One wrote, "I wrote way more, and spent more time looking at the code, and debugged it" (Participant 12). Another shared, "It gave me an incentive to put more effort and details into my peer feedback than before" (Participant 7). Others echoed similar sentiments:

• "Knowing that you get more yellow action cards definitely prompted me to giving more in-depth peer feedback" (Participant 9).
• "It simply motivated me to look at their code further and give more insightful feedback" (Participant 13).
• "It improved my motivation to write better comments and improved feedback" (Participant 33).

These responses support research suggesting that well-designed game environments can enhance student motivation by making learning tasks more personally meaningful (Papastergiou, 2009; Proulx, Romero, & Arnab, 2017). While this pattern was common, it was not universal. One student noted, "Not really, I just maybe a slightly more detailed version of what I would usually write" (Participant 5). Another explained, "Knowing that the cards I would get for the card game was impacted by my peer review assessment, it didn't really impact the way I graded people or in the game" (Participant 20). These responses reflect the variability in how students responded to the intervention, shaped by individual preferences, perceptions of fairness, or prior experiences with peer review (Falchikov, 2013). Taken together, the findings suggest that embedding peer feedback into the mechanics of a game environment can encourage increased effort and depth in peer review for many students.

Motivational Outcomes (RQ2)

In addition to changes in peer feedback quality, the study also asked: "Does the game-based learning intervention influence students' perceived competence, autonomy, and relatedness, as conceptualized by Self-Determination Theory?" This examined how the game-based intervention may have influenced students' motivation to participate in peer review. Drawing on SDT, the following sections explore how students experienced the core psychological needs of autonomy, competence, and relatedness during the intervention.

Autonomy

The observed increase in autonomy suggests that the game may have supported a more self-directed and engaging peer review experience. However, this finding should be interpreted with caution due to inconsistencies in survey wording. Most post-test items referenced the game explicitly (e.g., "The card game made me feel more in control of the peer feedback I gave"), while the pre-test items used neutral phrasing (e.g., "I feel in control of the peer feedback I provide"). This shift may have influenced how students interpreted the questions, potentially measuring not just their sense of control, but their perception of how much the game changed that control. In the case of the autonomy scale, Q6 scores declined post-intervention, suggesting either that students felt less in control because of the game, or that they felt no additional autonomy and rated the game accordingly. Due to this ambiguity, Q6 was excluded from the autonomy scale, as noted in the Methodology chapter.

Despite this issue, the game design included features aligned with autonomy-supportive principles. According to SDT, autonomy is enhanced when learners experience meaningful choice and self-direction (Deci & Ryan, 1994). Students had flexibility in how they structured their feedback and when to evaluate, aligning with game design principles like meaningful play (Salen & Zimmerman, 2003). Prior research supports the motivational value of player agency and strategic freedom in learning games (Indriasari, Denny, Lottridge, & Luxton-Reilly, 2023; Papastergiou, 2009). Future iterations should standardize survey items across time points and consider adding new game mechanics that enhance autonomy more explicitly, such as choosing focus areas for feedback, or unlocking branching outcomes based on their peer review decisions.

Competence

Although feedback quality improved significantly, students did not report a corresponding increase in perceived competence. One possible explanation is that students' self-perceptions, or lack thereof, did not align with their actual performance. While the game encouraged more detailed peer reviews, it may not have provided the kind of feedback or reinforcement that helps learners recognize skill development. Research suggests that competence is best supported when learners receive clear indicators of success and opportunities for guided improvement (Bandura, 2012; Hattie & Timperley, 2007).

Some students acknowledged that the game motivated them but did not necessarily make them feel more skilled. One student noted, "Yes, the card game influenced me to give more feedback but up to a certain degree. I wouldn't go too out of my way, but if it was an easy bug fix that I could provide, I would give them the feedback on it" (Participant 15). This suggests that while the game increased effort, it may not have reinforced growth in evaluative ability, a key aspect of perceived competence in SDT.

To address this, future iterations could incorporate scaffolding mechanisms such as worked examples, or formative peer review training. These supports not only help students give better feedback but also build evaluative confidence. Additionally, creating explicit opportunities for students to compare their feedback across game rounds, through guided reflection or pre-post exemplars, may help them recognize their progress and strengthen their sense of competence.

Relatedness

While relatedness did not show a statistically significant improvement, qualitative responses revealed mixed experiences. Some students found the game engaging and social, while others did not see a strong connection between the game and peer interactions. The competitive nature of the game may have contributed to this, as competition can sometimes emphasize individual performance over collaboration (Nicholson, 2015). One student noted, "The game made me a bit competitive, which affected the way I gave feedback to have a better chance of winning the card game" (Participant 19). As shown in Figure 5, while the competitive format aligned with most students' preferences, it may have been less effective for those who value collaboration. Still, the strong preference for playing games with others suggests that the social aspect of the card game was welcomed, even if the structure didn't consistently maintain a sense of peer connection.

Figure 5

Player Preferences

Some students may have perceived the feedback process as a social exchange, adjusting their tone or approach in response to the awareness of a peer audience. One student noted, "I did make sure the way I wrote the feedback was more professional," (Participant 28) suggesting that the intervention may have encouraged more thoughtful peer-to-peer communication.

These responses point to a need for a more balanced motivational structure. While competition can enhance motivation, research shows that environments emphasizing shared goals, peer support, and mutual respect are more likely to foster relatedness (Ardic & Tuzun, 2021; Powell & Kalina, 2009). Future iterations of the intervention could incorporate hybrid game mechanics, such as team-based challenges, collaborative missions, or peer mentoring roles. These additions may help ensure that students not only remain motivated but also feel a stronger sense of connection to one another through the feedback process.

Limitations and Future Directions

Sample and Contextual Limitations

While the study demonstrated statistically significant improvements in feedback quality and autonomy, several contextual factors limit the generalizability and sustainability of the findings. The intervention was conducted over two weeks in a single CS course, with a modest sample size ($N=42$ pre-test, $N=39$ post-test, $N=37$ for feedback quality). A short-term intervention is unlikely to produce lasting changes in motivation or feedback practices, even if early results are promising. Additionally, specific demographic data (e.g., age, sex, ethnicity) were not collected in order to reduce survey fatigue and focus on core constructs. However, this limits the ability to analyze subgroup trends or understand how diverse learner profiles may respond to the intervention. Future studies should examine whether similar outcomes are observed in larger, more varied populations, particularly in introductory courses, and consider incorporating demographic variables to support more nuanced analysis.

Furthermore, the novelty of the intervention may have influenced initial perceptions. Pre-test data indicated that most students had limited prior experience with GBL, with 57% having rarely participated in it and 24% never having done so. Prior research suggests that novelty effects can temporarily increase motivation (Papastergiou, 2009). Future studies should explore longer-term interventions or repeated gameplay sessions to assess whether motivation is sustained once the novelty wears off.

Additionally, the researcher also served as the instructor. While steps were taken to anonymize survey responses and ensure voluntary participation, it is possible that this dual role influenced student behaviour or motivation. Future implementations could be replicated in classrooms led by independent instructors to reduce the risk of researcher bias and further assess the generalizability of the results.

Methodological Considerations

One limitation of the study is the absence of a control group, which restricts the ability to attribute observed changes solely to the GBL intervention. Without comparing outcomes to a group using traditional peer feedback practices, causal inferences remain tentative. Future studies should include a control or comparison condition to isolate the specific effects of GBL on student motivation.

Another methodological consideration relates to the use of a Large Language Model (LLM) to classify peer feedback based on a predefined taxonomy. Although a subset of LLM-generated classifications was manually reviewed and found to align well with human-coded interpretations, the validation process was informal and not independently verified. While iterative refinements to the prompt and scoring logic improved consistency, the reliability of the automated analysis cannot be fully confirmed. Future research should incorporate more formal validation procedures, such as inter-rater reliability with multiple human coders, to increase confidence in the accuracy and reproducibility of automated feedback scoring (Cheng, Chen, Foung, Lam, & Tom, 2018).

While the Intrinsic Motivation Inventory (IMI) is a validated tool for measuring perceived motivation, it relies on self-reported data. Prior research cautions that students' perceptions of their learning or motivation do not always align with actual learning outcomes or behavioural changes (Persky, Lee, & Schlesselman, 2020). Furthermore, while paired-sample t-tests would have been more appropriate for comparing pre- and post-intervention motivation scores, the anonymity of survey responses made it impossible to link individuals across time points. This necessitated the use of independent-sample t-tests, which are less sensitive to within-student changes. Future studies may consider implementing pseudonymous identifiers to enable matched analyses while maintaining participant confidentiality.

Game Design Implications

Although peer feedback was not an action students performed within the game, it played a central role in determining gameplay conditions. Students' prior feedback quality influenced how many yellow action cards they received at the start of each game session, directly affecting their strategic options and potential for success. This mechanic reflects the principles of meaningful play, in which player actions must be both discernible and consequential within the system (Salen & Zimmerman, 2003).

By embedding peer feedback into the logic of gameplay, the intervention went beyond superficial gamification and aligned more closely with GBL, where learning tasks are structurally integrated rather than layered on as external rewards (Deterding, Dixon, Khaled, & Nacke, 2011; Nicholson, 2015). While some definitions might describe aspects of this intervention as gamification, GBL is not mutually exclusive with gamification and in this case provides a more comprehensive framing (Al-Azawi, Al-Faliti, & Al-Blushi, 2016). The learning task was not merely incentivized, it became embedded in the game's structure and progression, consistent with GBL design principles.

While this integration supports meaningful play, the connection between peer review and in-game performance was only revealed after the first game. This delayed feedback may have limited players' ability to perceive how their actions shaped outcomes, one of the core conditions for meaningful play. Future iterations could make this relationship more visible during gameplay by incorporating mid-game PCR bonuses, dynamic feedback quality indicators, or real-time adjustments to player status. These additions would preserve the game's existing structure while enhancing the immediacy and clarity of the learning-gameplay connection.

Student feedback also revealed valuable insight to refine gameplay. While many enjoyed the competitive format, others noted underdeveloped mechanics or a lack of narrative context. Comments such as "There should be some lore to the game, like why are we collecting all the green cards?" (Participant 12) and "The trade [card] is almost not used" (Participant 14) point to ways the game could improve in thematic coherence and strategic balance. Enhancing underutilized mechanics and incorporating light narrative framing could boost immersion and perceived relevance. Additionally, integrating cooperative elements, such as team-based objectives or collaborative card effects, may better support students who are socially motivated, thereby reinforcing the SDT need for relatedness (Powell & Kalina, 2009).

Implications for Practice

This study reinforces the need to design peer learning environments that make evaluative tasks meaningful, socially relevant, and autonomy-supportive. As a CS educator, I plan to continue using game-based interventions to promote high-quality PCR, particularly by embedding learning goals into well-structured game systems. The intervention's success suggests that students are more likely to invest effort when feedback is clearly linked to in-game outcomes. However, student responses also highlight the importance of transparency, balance, and narrative coherence in game design. For example, some students expressed confusion about game mechanics or a desire for deeper thematic immersion, indicating that instructional games should be refined not only for pedagogical impact but also for clarity.

Drawing from both the literature and student feedback, I plan to refine future iterations of this intervention by:

• emphasizing the relationship between feedback quality and gameplay success during play;
• strengthening under-utilized mechanics (e.g., trade cards) and rebalancing elements to sustain strategic depth;
• adding light narrative framing to reinforce coherence and purpose (e.g., explaining why students are "collecting" resources);
• incorporating cooperative mechanics (e.g., team-based bonuses, peer mentoring cards) to support students who are more socially motivated and to foster relatedness.

These recommendations echo best practices identified earlier in the study. From a social-constructivist perspective, learning is most effective when it is shared, situated, and actively constructed through interaction. In peer review, this means creating structures that promote thoughtful critique and reflective dialogue, not only procedural compliance. Similarly, as outlined in SDT, students are more intrinsically motivated when they feel competent, autonomous, and connected. The design of the game should continue to support these needs, both through gameplay and the surrounding instructional context.

Finally, this study supports my broader aim to align peer feedback with constructive alignment principles, ensuring that learning outcomes, assessments, and activities reinforce one another. If peer feedback is to be valued as a professional skill, it must be assessed as such and designed in ways that connect classroom instructional strategies to authentic professional practices. Integrating these principles into my course design, especially in preparation for students' industry internships, will help bridge the gap between educational and workplace expectations.