Embedded Files
The Role We Play:
Table of Contents
Welcome!
This page summarizes the findings and methods of my Master's Thesis. The content here is designed to offer a clear and accessible overview of the work. For full citations, detailed analyses, and original source material, please refer to the complete thesis:
Overview
This research project examines how gender identity, personality traits, and self-efficacy influence the roles players choose in three major multiplayer titles: Overwatch 2, League of Legends, and Destiny 2. Across 234 participants, I combined validated psychological scales with game-specific behavior questions to understand why players gravitate toward certain classes and what this means for game design, community dynamics, and player engagement.
Background
Research Goals
Determine whether gender predicts preference for tank, damage, or support roles
Evaluate whether Big Five personality traits correlate with player role choice
Measure how self-efficacy varies across games with differing complexity and customization
Understand why players choose specific classes through qualitative motivations
Why These Games?
I chose Overwatch 2, League of Legends, and Destiny 2 not only for their popularity, but also because they share several key features relevant to my study, while also differing in ways that make comparison meaningful.
Role-based team dynamics
Whether you're queuing as a tank, support, or damage dealer, team success depends on players fulfilling distinct, interdependent roles.
This allows me to study how role selection aligns with gender identity and self-efficacy.
Live-service monetization model
These systems are designed to keep players engaged over time, making them ideal for examining how confidence and identity affect long-term play.
Games can be accessed for free
This makes them accessible to a wider and more diverse audience, which is important for capturing a broader sample in my study.
However, each game varies with its core gameplay loops and session objectives
League of Legends focuses on structured lane-based PvP matches,
Overwatch 2 focuses on fast-paced PvP objective control, and
Destiny 2 focuses on cooperative PvE raids.
These differences let me explore how different mechanical systems and social environments influence player behavior and role dynamics.
Methodology
Measurements
I primarily used 2 different metrics to get a better understanding of the players: Big Five Personality (IPIP-50) and Self-Efficacy (NGSE). Additionally, I collected demographic information, including age (over 18), race, gender, games played (Overwatch 2, League of Legends, and Destiny 2), and preferred roles in those games (tank, damage, or support).
Overall Survey Structure
The study was conducted using a multi-stage survey built in Qualtrics, designed to adapt dynamically to each participant’s experience across Overwatch 2, League of Legends, and Destiny 2. After collecting demographic information, the survey branched into game-specific sections based on which titles each participant had played. Within each branch, respondents completed validated psychological measures, including the IPIP Big Five Personality scale and the New General Self-Efficacy Scale, followed by qualitative statements about their preferred roles, motivation for role selection, and gameplay behaviors. This structure ensured that participants only answered relevant questions while maintaining consistency across all three games, resulting in clean, comparable data for analysis.
Distrubution
The survey was shared through approved posts on game-specific subreddits and official Discord servers, following direct coordination with community moderators. To supplement recruitment and ensure an adequate sample size, I also distributed the survey through a small number of personal and enthusiast Discord communities.
Data Processing
Combining Gender
Almost 80% of participants responded “man.”
Analysis requires 5+ members for each gender option, so some gender options were combined.
Classifying Customization
In order to sort the games by their customizability, I broke down how players are able to customize their gameplay to fit scenarios.
Overall, League of Legends has the least amount of options to deal with enemies.
Though there are many items players can buy, items must be relevant for the character's unique abilities and/or stats and helpful for the circumstances they are playing in.
Players must very slowly save up their gold throughout the match to buy items.
Counterplay is very slow and does not provide as much change to a player’s kit or playstyle.
Overwatch 2 is in the middle.
Players swap characters within their role (if they chose role queue) or to any character in the game (in open queue).
A complete change in kit constitutes a medium level of customization.
Lastly, Destiny 2 has the most options for dealing with its content(enemy encounters).
Players can change abilities and subclasses at any time.
Players can change their weapons (3 slots with 9 options each) or change their armor (4 slots with 9 options each).
Results
Hypotheses at a Glance
Hypothesis 1 - Fully Supported
My first hypothesis was that across all three games, non-males are more likely to choose the support role
I used a Pearson's Chi-squared test on each of the games, comparing binary gender and support role choice
The result for each of these Chi-squared tests was statistically significant for all three games, and supported the hypothesis
Gender (Binary) and Role (Binary) Chi-squared Analysis Probability Across Games
Hypothesis 2 - NOT Fully Supported
My second hypothesis was that across all three games, males exhibit higher self-efficacy compared to non-males
I used a one-way ANOVA test on each of the games, comparing binary gender and self-efficacy
The result for two of these ANOVA tests (Overwatch 2 and League of Legends) was not statistically significant and did not support the hypothesis
However, the result for the ANOVA test on Destiny 2 was statistically significant and supported the hypothesis
Gender (Binary) and Self-Efficacy ANOVA Probability Across Games
Hypothesis 3 - Fully Supported
My third hypothesis was that players will exhibit higher levels of self-efficacy when playing games with more choices
I used a one-way ANOVA test to compare customization and self-efficacy
Whereas in previous games, I was calculating tests within games, this test calculated the effect of the game choice itself
The result of this ANOVA test was statistically significant and supported the hypothesis
Average Self-Efficacy of Players by Game Choice
Hypothesis 4 - NOT Fully Supported
My fourth and final hypothesis was that across all three games, players with high agreeableness and low extraversion select the support class
I used a chi-squared test on each of the games comparing high agreeableness and low extraversion to a support role preference
The result for each of these Chi-squared tests was not statistically significant for all three games, and did not support the hypothesis
Role Preference (Binary) and High Agreeableness and Low Extraversion Chi-squared Analysis Probability Across Games
Discussion
As expected, non-male players were more likely to choose the support role.
Previous research suggested 3 main reasons:
Social expectations
Stereotype Reinforcement
Confidence
To investigate this, I analyzed responses to qualitative statements I asked at the end of each game-based section.
Social Expectations
First, I looked at the qualitative agreement responses to the statement related to social expectations:
“I chose this role because my teammates wanted me to play this role.”The highest (or tied for highest) means reported were for the support role across all three games.
This suggests support players are more willing to play roles based on their teammates' preferences over their own.
Qualitative Statement Agreement Means Across All Three Games (7-Point Likert Scale) for the Phrase: “I chose this role because my teammates wanted me to play this role.”
Stereotype Reinforcement
Next, I looked at the qualitative agreement responses to the statement related to stereotype reinforcement:
“I chose this role because my teammates are less critical of my performance when I play this role.”Initially, it may seem that male players generally feel more pressure, with their higher agreement scores for 2 out of the 3 games.
However, we should remember that Destiny 2 was the only game for which self-efficacy had a statistically significant connection to gender (Hypothesis 3).
In Destiny 2, non-male players had a higher agreement mean than male players.
Non-male players having lower self-efficacy (at a statistically significant rate) and feeling generally like teammates are less critical when they play support may indicate how players subconsciously reinforce the stereotype of non-male players selecting support.
Qualitative Statement Agreement Means Across All Three Games (7-Point Likert Scale) for the Phrase: “I chose this role because my teammates are less critical of my performance when I play this role.”
Confidence
To investigate confidence, I looked at the qualitative agreement responses in Destiny 2 related to confidence:
“I chose this role because I am skilled at this role.”I only looked at Destiny 2 because it was the only game with statistically significant differences in self-efficacy between male and non-male players.
Across all three roles, male players had the highest agreement means.
This qualitative data highlights how non-male players may feel less skilled overall, and choose support.
This may also be related to the notion that support characters require less mechanical skill in most games.
Additionally, this is consistent with the hypothesis that male players have higher SE overall.
Qualitative Statement Agreement Means Across Roles and Genders for Destiny 2 (7-Point Likert Scale) for the Phrase: “I chose this role because I am skilled at this role.”
Destiny 2 Stands out
Hypothesis 2
Destiny 2 was the only game that had a statistically significant relationship between SE and gender
Hypothesis 3
Destiny 2 had a much higher self-efficacy mean than the other two games
Why Does Destiny 2 Stand Out?
Prior Research shows:
1. Destiny 2 requires a high time investment
2. Destiny 2 has high customization
Qualitative Results show:
3. Role Satisfaction
4. PvE is more cooperative
1. Time Investment
Research consistently shows that, on average, men tend to play video games more than women.
This difference in playtime ratio is especially meaningful when analyzing high time-investment activities.
Destiny 2 end-game content requires a significant time investment to perform well because players need to:
Complete specific pre-requisite missions to unlock exotic items.
Complete missions in general to raise their weapon and armor levels.
Earn in-game money to unlock optimal sets of character abilities.
2. High Customization
Another reason Destiny 2 stands out from the other games in my study is because of how deeply it emphasizes kit customization and build flexibility.
Previous research has linked high self-efficacy with a preference for having more choices.
In Destiny 2, players have a lot of customization for their gameplay:
3 weapon slots, each with 9 choices in inventory per slot.
4 armor pieces, again with 9 options in inventory for each.
And unlike many other games, Destiny 2 lets players change their subclass and equipment at any time, even mid-activity.
This high level of control appeals strongly to individuals with higher self-efficacy, those who believe in their ability to optimize, experiment, and succeed under complex conditions.
3. Role Satisfaction
Additionally, Destiny 2 players appeared to be more satisfied with their roles overall when compared to other games.
They showed consistently higher mean agreement scores compared to other games with the phrases:
"I like the positioning involved in this role."
"I enjoy the play-style of this role."
Qualitative Statement Agreement Means Across All Three Games (7-Point Likert Scale) for the Phrases: “I like the positioning involved in this role,” and "I enjoy the play-style of this role."
4. PvE is more cooperative
Destiny 2 is primarily PvE (Player vs. Environment), which means that players normally fight against computer-controlled enemies rather than other players.
Destiny 2 players showed consistently lower mean agreement scores with the phrases:
"My teammates wanted me to play this role."
"My teammates are less critical of my performance when I play this role."
Lower mean agreement scores indicate that teammate pressure overall may be lower in Destiny 2 than in the other titles.
Qualitative Statement Agreement Means Across All Three Games (7-Point Likert Scale) for the Phrases: “My teammates wanted me to play this role,” and "My teammates are less critical of my performance when I play this role."
Limitations and Future Research
Time Played
In the survey, time was measured categorically rather than continuously.
Categorical playtime meant that I could not establish a clear linear association.
Destiny 2 players have the highest categorical average, but that may be misleading since the highest option is 100+ hours.
I recommend that future researchers allow players to manually enter how much time they have spent playing the game.
Skill
League of Legends and Overwatch 2 both have ranks to measure skill, but Destiny 2 does not.
Additionally, Destiny 2 PvE has no enemy team to compare yourself against.
Progress/contribution in team fights is more accurately measured by light level.
Light level is just the average level of a player’s equipment + a seasonal item.
Overall, this makes it hard to establish a player's skill level.
For future research, I recommend either:
Focus exclusively on ranked OR non-ranked games.
Include a variety of games that have AND don’t have ranks.
Either way, it would be a good idea to record ranks to see if there is any association for competitive games.
Participation
Many of the posts on Reddit needed moderator approval.
Posts were often taken down immediately by moderation bots.
Waiting for human moderator approval meant that posts did not show up in the “newest posts” for a while after they were posted.
This may have led to lower engagement.
The Discord channel in which the survey was posted was also an important factor
Posting the survey link in a general chat meant it was quickly buried by other chats and not seen by many possible participants.
Posting the survey in announcement channels was not allowed by many servers, but when it was, it tended to yield more responses
Support was the most chosen role by far, and people may have preferred the support role in one game because they chose it in the others
Conclusions and Advice for Industry
1. Emphasize the support role to attract non-male audiences
Research (including this study) consistently shows that non-males tend to prefer support roles at a significantly higher rate than male players.
By highlighting the importance of support characters, showcasing their impact on team success, and even glorifying their contributions in trailers, advertisements, and community engagement efforts, developers can create a more inclusive and appealing experience for non-male audiences.
This approach not only broadens the game's appeal but also fosters a community where support players feel valued and recognized, ultimately encouraging long-term engagement and retention.
Furthermore, marketing efforts that frame support roles as strategic, skillful, and essential can help challenge outdated stereotypes and elevate their status within gaming culture, making them more attractive to a diverse player base.
2. Allow more customization and counterplay to increase playtime
If the goal is to increase playtime (and, by extension, revenue), offering players more customization options is a highly effective strategy.
Prior research has related higher self-efficacy with higher playtime, and more playtime means more revenue.
This study showed that players were more satisfied with their role when they could customize their positioning and playstyle more.
This implies that when players can tailor their experience to fit their preferences or enemies, they are more likely to stay engaged for longer periods.
Whether it’s through unique character builds or adaptable gameplay mechanics, giving players meaningful choices could enhance immersion and investment.
The more a game caters to diverse playstyles and self-expression, the more likely players are to return, increasing both retention and overall playtime.
3. The Holy Trinity is still effective for describing roles in combat
Despite the evolution of game mechanics and player roles, the Holy Trinity of tank, support, and damage remains an effective way to categorize and understand player preferences.
This foundational framework persists because it aligns with core gameplay dynamics
Tanks provide durability and frontline control,
Supports enable team survival and strategy, and
Damage dealers focus on offense and elimination.
These roles not only create clear and intuitive team structures but also cater to different player motivations
whether they enjoy leading engagements,
assisting teammates, or
maximizing their individual impact.
Even as games introduce hybrid roles and more nuanced mechanics, the Holy Trinity remains a reliable approach to balancing gameplay, designing engaging team interactions, and assigning roles in a way that resonates with both new and experienced players.
Thank you for reading!
For more information about the research, please feel free to check out the full paper linked at the top of this page, or email me through the link at the bottom of the page!
Google Sites
Report abuse