Virtual Insanity - Designing Safer VR Spaces

UX Research and Design - School Project

TL;DR

Skills: Secondary Research, Ethnography, Recruitment, Survey Design, In-Depth Interviews, Qualitative Coding, Content Analysis, Usability Testing in VR, Persona Creation, Rapid Iteration, Prototyping

Top Lessons Learned:

My initial assumptions about who uses VR social platforms and why were challenged over the course of this project. It reinforced the importance of approaching projects with a questioning mindset and being open to revising initial beliefs.
This project involved gaining a comprehensive understanding of complicated issues and technology in a very short timeframe. I honed my ability to learn rapidly, prioritizing efficient research methods to effectively direct our design goals within the quarter.

Project Background

This project originated from a 10-week course during Fall 2022, part of the Master of Science in Human-Centered Design and Engineering program at the University of Washington. Guided by the "do no harm" principle and the stipulation to exclude website or app design, our team embarked on addressing harassment within VR environments. This area of interest was sparked by a news article detailing the problem of sexual harassment in Horizon Worlds.

My Role: Lead Researcher
Avani Babar: Designer
Kayda Norman: Designer
Kim Johnson: PM & Designer

With the surge in popularity of Virtual Reality (VR) spaces - both gaming and social - comes a pressing need to address the unique and unprecedented challenges of harassment within virtual spaces.

VR's immersive nature generates emotions akin to those felt in face-to-face scenarios; consequently, harassment in VR feels as real as in-person incidents, potentially leading to mental health struggles such as isolation, anxiety, and depression. Virtual spaces must be designed with these issues in mind so that people who enter them feel safe and secure.

This project aimed to explore some of the safety concerns that are present in virtual spaces and provide a solution that offers a richer and more inclusive experience.

Our initial design question was: How do we design a safe environment and harassment-free experience for people in VR spaces?

Research Overview

We had three main goals that directed my research methodology:

Develop an understanding of VR harassment issues.
Investigate user motivations for engaging in social VR platforms.
Identify pain points and gaps in VR experiences that contribute to harassment.

Developing an understanding of VR harassment issues

This project was very exploratory, so I started off by seeing what was already out there - what had people already studied, what were people saying, and what were they doing? We also continued to understand VR harassment through our other research methods (surveys and interviews) but this is where I began.
‍

Secondary Research

I started off by conducting secondary research - I searched for articles and papers on topics such as harassment in VR, building online communities, ethical considerations of VR, and embodiment.
This gave us insight into the larger issues and allowed us to start narrowing down our research questions and helped avoid redundant surveys or interview questions.

Investigating user motivations for engaging in social VR platforms

Before interacting with users I wanted to understand why people use VRChat. These motivations were also explored in the rest of my research (survey and interviews) but this is where I began.
‍

Content Analysis

Before moving forward with a survey or interviews, I searched for online forums that discussed two topics: why people use VRChat, and harassment in VRChat. I put the responses in spreadsheets and conducted a quick couple rounds of qualitative coding to see the major themes that emerged.
Several major themes emerged. Firstly, people used VRChat to either make friends or hang out with friends from IRL. Secondly, users felt that VRChat allowed them to be their true selves or who they want to be, whether that be having the ability to customize their appearance, or the anonymity and freedom to be more outgoing and social than they normally would. Other themes such as it's simply entertaining, or allowed them to express creativity through world and avatar building, also emerged.

Identifying pain points and gaps in VR experiences that contribute to harassment

Once I had a general idea of the types of harassment people faced and their motivations, I started off by using rapid ethnography to get a closer look. From my findings we then created a survey, followed by a round of in-depth interviews.
‍

Rapid Ethnography

I rented a VR headset and joined VRChat as a user to see what the experience was like understand behavior, and see what the environments and social norms were like.
This provided us with a clearer understanding of the platform and users, and helped us find discrepancies between what they had or had not pointed out to us.
Harassment was indeed widespread. Additionally, it was difficult to report and block other users, as the UI was overly complicated. New users spoke to other users for help on how to use features they were unfamiliar with, and some experienced users were more likely to help guide new members than others.

Survey

I created a survey in Qualtrics, and I recruited participants from two discord channels and five subreddits, from which we got 91 responses.
We decided a survey would be one of the best methods we could use to reach users across the country with varying experiences with VR, and this method gave us a way to gather information quickly with no extra cost.
A large majority of our participants had either been harassed (76%) or seen someone harassed (78%), and 73% of respondents said either they or someone else intervened in those cases to help the victim.
Respondents had mixed ideas on how to address harassment - 33% believed nothing could be done, 14% wanted better age restrictions (as they thought kids were the primary source), and 39% wanted better safety features or moderation.

Interviews

From our survey I was able to recruit 5 interview participants, and I conducted 5 one-on-one semi-structured in-depth interviews with them. The interviews were recorded, transcribed, and coded.
From these interviews I was able to gain more nuanced information about the types of harassment seen and experienced in VRChat, how they dealt with it, why people use VRChat, and what other pain points they encountered.

Research Findings & Personas

After the research phase was complete, our main insights were:

New users were overwhelmed by the complicated UI, navigation, and extensive safety features, and are not onboarded properly. VRChat has very little onboarding, and many users learned how to block or mute through the need to perform the actions repeatedly.
"It's like going to a bar with a daycare" - Long-term and expert users see children as the main source of harassment problems in VR. However, they did cite many other types of harassment, including verbal (personal insults, slurs, sexualized language, etc), physical (blocking the way, groping, etc), and environmental (being loud, drawing sexual imagery, etc).
"Yeah I really hate the new UI that came out 3 months ago there's like, 13 tabs. It's too much" - The customizable and free nature of VRChat is its strength, but also its weakness as it is overly complicated, and can lead to the perpetuation of harassment.
VRChat users primarily use VRChat to meet new people or hang out with friends they already know, with secondary motivations being entertainment, exploration of worlds, and self-expression.

From our user research data, we created personas together of our primary users, secondary users, and we also created an anti-persona. My teammate Avani designed these personas below:

New Design Question & Solution

After synthesizing our data, we decided to focus on new users through the persona Newbie Mushroom. Unlike experienced users, newbies found it challenging to navigate VRChat's safety features. Expert users were also likely to have used other VR platforms or have previous gaming experience, causing them to adjust to the controls more quickly and be more accustomed to harassment behaviors.

We reframed our design question to “How can we equip new users with the skills to navigate VRChat and make new connections within 1 week (or ~4 hours of use) of joining the platform?”

Along with our design question we decided on the following design goals:

Make meeting new people with shared interests easier.
Simplify onboarding for new users.
Help new users navigate the VRChat spaces.
Inform new users of the VRChat rules, culture, and social norms.
Provide users the skills to deal with unwanted interruptions.

We individually sketched out many ideas for design solutions, and after seeing some overlap in our individual sketches, we decided on three main themes:

A “shared interests” feature so that users may meet other people with similar interests. When users approach each other, an icon will appear over their head if they have shared interests, which will help foster quicker connections.
A VR “buddy” system, where new users can be paired with new or experienced users so they are not alone when entering VRChat for the first time. Users will be matched with buddies with similar interests.
An onboarding experience for new users that provides a tutorial for controls, rules and safety, exploring worlds, and finding avatars. This tutorial also teaches users how to add friends, mute, and block other players. Currently, VRChat doesn’t have an onboarding tutorial.

_{Click to expand the images below}

User Flows, Scenarios & Testing

Created from our design ideas and refined sketches, we created scenarios and user flows that were used as the basis for our first prototype. We prioritized the new user onboarding experience, as we felt that teaching users how to keep themselves safe was the most important aspect, followed by the "shared interests" feature, and then the VR "buddy" system. We all participated in prototyping, and split up the work so each person had something to prototype.

_{Click to expand the images below}

Once we had medium fidelity prototypes completed, we began user testing. We conducted four usability testing sessions (each member conducted a session) with our onboarding solution; we were able to load our prototype into a virtual desktop so that users could test the prototype by using a VR headset and controls. We then had them go into VRChat and gave them tasks to complete to see if we had properly taught them how to navigate different worlds, mute/block/report users, and make friends.

We split up the tests between two iterations: the first was a low-fidelity prototype which we tested on two users, then made some rapid iterations and upgraded to a medium-fidelity prototype which was again tested on two users. Each member moderated and conducted user testing sessions, and was responsible for a portion of the prototyping.

Usability Testing

The most crucial change after the first round of testing was reducing the amount of information in the tutorial, and shortening it overall. Participants felt they were overloaded with information and it took too long to complete.

The second round of usability testing focused more on the interaction rather than the content - for example, participants didn't know they should be pressing "next," and there were some confusing graphics.

_{Low Fidelity Prototype}

_{Medium Fidelity Prototype}

With the feedback from our four user testing sessions and a cross-class critique, we iterated one more time, finalizing our information architecture and making a high-fidelity prototype.

Design Solution

HERE is our clickable high fidelity prototype, which we completed in Figma.

Our design solution is an onboarding tutorial for new users using VRChat. This tutorial covers 6 sections: how to use the controllers, how to use the launch pad, where to find avatars, how to update your profile, how to explore worlds, and interactions such as adding friends and blocking other players.

This tutorial is designed to give new users a safe space to practice interactions such as muting and blocking, without the stress of needing to escape a real harassment situation. This tutorial was built to demonstrate the user interface of VRChat in its current state.

The second feature of our design solution is our shared interests feature. This is a section of the user's profile that allows them to add their interests. In addition, when a player is in proximity to other users that share similar interests, an icon appears near their usernames to let users know they have something in common.

The third feature of our design solution is the VR Buddy System. Many users learned how to use the block and mute features by asking other users what to do, and experienced users also reported intervening in harassment situations by teaching new users how to use these features. The VR Buddy System pairs new users with another user so they can explore VRChat together from the start.

Project Demo

Reflections

Limitations: This project had several limitations. First, we had to work within the current UI of VRChat which is very complicated and less than ideal. We were also very constricted on time, which limited many aspects of the project we would have liked to expand on, such as recruiting more diverse participants, conducting more user testing, polishing our designs, and potentially reworking the UI entirely. We were also limited to what software we could use and the experience we had as a team with the software we had available to us.

Lessons Learned: This project was a great learning experience for me in having more empathy for users, particularly a user group I may not completely understand. When I first began this project I believe I had some preconceived notions of what the users of VRChat were like based on stereotypes or the "loudest voice in the room" I had seen online. However after interviewing many users and using VRChat myself, I gained a new understanding of what motivated people to use VRChat, and realized that in the future I would need to be even more careful of confronting any ideas I had about users before starting the research process.

What I Would Do Differently: If i could do this project again I would spend more time trying to recruit a more diverse group of participants, in ethnicity, gender, and sexuality. I would also focus more on our personas throughout the design process in order to inform our design decisions and save the time we spent reworking our first prototype.