Misinformation is one society’s most pressing issues, spreading division and chaos across the globe. Videogames have become one of the more promising mediums for misinformation interventions, teaching players common falsehood indicators and helping them develop their discernment abilities. This thesis details the design, playtesting, and evaluation of a digital game intervention intended to improve players’ misinformation discernment abilities. Through a design narrative, it first describes the design goals and development of an initial iteration of DoomScroll, a modified version of the popular social deception game Among Us. It then reports on the playtesting of DoomScroll and how player feedback was used to refine and improve the mod’s gameplay. It concludes by detailing a randomized control trial conducted to more rigorously test DoomScroll’s efficacy as a misinformation intervention tool when compared to a control condition. Player feedback regarding its socialization and replayability suggests that DoomScroll did meet its design goals by improving player motivation and increasing the likelihood of repeated use. However, further analyses showed that the mod did not produce any statistically significant improvement in players’ discernment abilities. The qualitative findings were more positive, suggesting an uneven distribution of process-driven learning did occur. This study contributes insights to improve misinformation game design and into the potential of modding as a medium for educational game development.

As esports has gained a growing audience, parents and educators have raised concerns about deviant behaviors associated with it (tilt). The present study investigated players’ stress levels and emotions around in-game stress events by means of physiological measurements (heart rate [HR], heart rate variability [HRV], and facial expressions) to find out how in-game stress events affect player’s stress levels according to different groups of players and different event contexts. While groups of young players (N = 119) aged 14-25 played three matches of League of Legends, their in-game logs, video recordings of gameplay, in-game communication, and physiological data (HR, HRV, and facial expressions) were collected. Linear mixed-effects models were fitted to explore the effects of different player groups and in-game contexts related to the stressful events during the gameplay. The higher game skill level and later in-game phase were positively associated with earlier recognition and anticipation of upcoming events before they occur. Winning a match, later in-game phase, Enemy and Myself events were positively associated with a higher increase of HR during a stress event. Older age and Team and Teammate events were positively associated with a smaller decrease in HR after stress events. Lower game skill level, older age, and later in-game phase are associated with a bigger drop of HRV during stress events, which implies higher stress levels. Overall, the contribution of this study lies in exploring the in-game variables that impact player tilt employing physiological measures of players’ emotions.

Failure is a crucial part of a game experience. Players expect to be challenged to a level that invites failure, showing lower engagement if they don’t fail at all. While an important aspect of the game experience, the ways video games frame failure and how players react while playing remains understudied. Borrowing from literature in psychology, this dissertation uses mastery orientation, a measure of how individuals react to failure, to develop a behavioral measure which observes how 56 individuals reacted to failure while playing notoriously challenging video game, Cuphead for two weeks.

Results validate this measure, showing that those who reported higher mastery orientation scores also show more mastery-oriented behaviors in-game, show more mastery-oriented behaviors sequentially, and are less likely to abandon a level before completing it. No change in mastery orientation towards game contexts was initially found after playing Cuphead for two weeks, although closer inspection shows a leveling out of mastery orientation scores as those who started on the lower end of the scale increased while those who started in the higher end of the scale decreased. This is shown in women significantly increasing in their mastery orientation scores after playing Cuphead for two weeks while men significantly decreased. Visualizing these effects suggest that some women began the study underconfident in their ability to persist through failure while some men began the study overconfident, with mastery orientation scores leveling out through the course of the study. Finally, distribution of gameplay behaviors shows that players persist for longer when further into the game, suggesting that the more time and effort the player expends, the more likely they are to continue to persist through failure. This sheds some light on how players react to failure depending on where in the game they currently are.

This study opens a new perspective for how researchers and game developers can understand the behaviors players take when they encounter failure, develops a new methodology to gauge mastery orientation, and begins to show where and why in a game players begin to give up. The implications of these results and areas for further investigation are discussed.

Much of team communication is a trained process for teams in high pressure environments, like the highly regimented checklists that structure aircraft takeoff and landing. Elite esports teams use communication to maintain tight coordination in the face of chaotic and stressful stimuli, closely resembling behavior of high performing teams in domains seemingly is similar from competitive video games. This 2 x 2 factorial design explores differences between in-game communications in League of Legends teams that vary in both experience as a team and expertise with the game. The findings describe how content, style, and amount of communication differ between novice and expert teams; whether those differences relate to experience with the game, experience as a team, or both; and whether differences in communication relate to team performance under pressure. This work demonstrates methods that can be built on to explore whether these same communication patterns and solutions can effectively transfer across domains in the interest of training for safety and performance in higher consequence contexts.

The rise of deepfake technology—AI-generated videos that convincingly depict fabricated events—poses growing challenges for digital media discernment, public trust, and memory. While technical detection methods have advanced, little empirical research has analyzed how deepfakes influence human cognition and memory. This dissertation empirically assesses whether deepfake videos can implant false memories in Generation Z adults and whether active debriefing or media literacy interventions mitigate these effects.

Drawing from literature in cognitive science, false memory, and misinformation studies, this research employs an experimental design involving over 1,600 participants aged 18–26. Participants were exposed to a combination of real videos and deepfake videos featuring celebrity figures. Later, they were assessed for memory retention, deepfake detection, and confidence in discernment. Results show that deepfakes can produce false memories, especially when participants are unaware that they are viewing falsified videos. However, active debriefing (informing participants post-exposure about the presence of deepfakes) significantly reduced the retention of these false memories.

Contrary to expectations, prior knowledge of the featured celebrities, familiarity with deepfake technology, and frequency of social media use did not improve detection accuracy or false memory resilience. Participants frequently relied on intuitive audiovisual cues (e.g., lip-sync mismatches and unnatural eye movements) over factual reasoning, and often expressed high confidence in inaccurate judgments. Emotional responses to deepfakes were also pronounced. Many participants reported anxiety, mistrust, and moral concern about the implications of deepfake technology and the irresponsible use of generative AI tools.