Designing Accessible Classroom Communities on Discord

Secondary and higher education teachers have been recipe-swapping tips, strategies and postmortem classroom analyses online way before COVID gained precedence in our collective imaginary, but the pandemic has created a veritable linkstorm of teaching advice available through personal blogs, research portals and collective activism groups.

Not all of this advice was good advice.

And while surely everyone’s hearts were in the right place while authoring articles like this, the deluge of teaching advice available to crisis instruction teachers is a bit of a minefield of classism and accidental ableism: so many of these strategies assume physical abilities (the ability to sit for long hours, standard mouse manipulation), technological availabilities (cost of computing tech, availability of dongles and periphery tech) and affective resilience (the wellness to do online learning day in and day out) look mostly the same for everyone in the North American context. Notwithstanding the fact that classrooms are becoming increasingly internationalized and the assumption of single-timezone or even single-state education is likely misguided in pandemic learning, we also must pull apart the mythology that digital and physical access looks mostly the same for everyone in the room. These issues have been taken up in earnest over the past two years in the critical disability community, as well as critical digital media. What I haven’t seen too much of is critical spacemaking for madness – or psychosocial disability – students experiencing symptoms of long-term mental illness or short-term mental disability. They can’t “turn that off” for the classroom context, but in our efforts to become physically digi-accessible we might be leaving behind students whose flavour of madness (a term I use lovingly, as a mad educator myself) does not easily comply with rigid LMS systems, constantly changing syllabi or entirely asynchronous instructional methods.

I honour that mad stories and LXP (lived experiences) are very seldom shared in teaching contexts – apart from “how to deal” with people like me in the room – and I’m not partial to counterstorying with their voices unless I am directly collaborating with them. Noting this, I’ll use myself as an example to walk you through some of the vagaries of accommodating madness in a classroom setting, and then I’ll show you how I translated that into fully-online environments.

My flavour of madness is schizophrenia. Most teachers do not know what accommodations would be feasibly useful for psychosocial disabilities, and precious few are in active conversation about how to accommodate serious mental illness (often associated as illnesses with psychotic features). There is plenty of discourse on the non-relationship between violence and serious mental illness, so I’m taking for granted – in the space I’m sharing right now with you – that this isn’t your primary concern when thinking about how to best accommodate my needs.

Keeping in mind that every bodymind is necessarily different and experiences different needs and strengths (something I have liked to call “their kaleidoscopic mind” in the past), I’ll tell you how schizophrenia presents for me in the classroom:

• I have a very hard time with audio-only interaction. When I’m hearing things, it’s like two audio tracks playing simultaneously – and often, the track in my head is quite a bit louder than the person speaking across the room from me. In an in-class environment, sometimes the FM microphone system helps, but the “blurring” of voices is not accommodated well by FM systems.
• Somewhat similarly, this dual-track forces me to hyper-focus: when closed-captions are available to me, I find I can either read the captioning or process the speaker’s gestures, but not both. This causes me to miss information: for this reason I prefer transcripts whenever they are available (they’re usually not), while listening to the audio-track in the background. Most transcripts will indicate relevant physical actions or environmental interactions that are necessary to build context.
• I have unpredictable productivity hours – this is very difficult to accommodate in an in-class environment. Psychotic illness often feature manic periods, followed by days of total disengagement from the world. I have a very difficult time “presenting as normative” during a disengagement period, which can worry or startle other learners or draw attention to myself in a stigmatizing way. Online learning offers an interesting hybrid solution toward people experiencing “cycling”, as this disengagement period is relatively concealed from the classroom environment. Crip time is, of course, a popular solution to this conundrum.

You’ll notice that the menu-standard accommodations offered through your school’s Accessibility Office – solutions like “time 1.5x or time 2x”, “quiet environment”, “alternate format resources”, “interpreter”, “closed-captioning” – would help very little for bodyminds like mine. This refracts itself across the spectrum of psychosocial disability, rendering the burden of accommodation back on the student themselves, not the instructor or institution. I have had to create my own build-ins and solutions to suit my studies, and this holds true for students with bipolar disorder, major depressive disorder, autism spectrum diagnosis – the list goes on.

How does Discord intercept this dynamic?

I did a twitter thread about compassionate instruction using Discord that this post is based on, and if you’re looking for the short & sweet answer, I can gladly direct you there. When the pandemic forced us online, auto-generated closed captions and flex time around assignment submission became popular accommodation mainstays: two features that Discord doesn’t offer directly, however I would argue that the wealth of features it offers for psychosocial disability make a compelling argument for its uptake in more higher education online pedagogies:

This is a screen capture of the core Discord server interface, featuring three panels: the left-hand panel denotes everything you can interact with in the server (including text channels, audio channels, and guides). The center panel is the active window, which displays channel content (based on what you’ve selected in the left-hand panel) and asynchronous interactions between all server users. The right-hand panel displays all current server members, which can be divided in many ways (into pre-defined teams, into self-assigned teams or roles, into power or ranked roles, into mini-communities, etc.). The example captures I’ll be using here are from a first year undergraduate writing course, a traditional 4-month course that teaches “academic writing” strategies and methodologies.

What’s important to notice here is that Discord greatly favours text interaction, which is perfect for a variety of disabilities (including ADHD, b/Blindness, long-term injury and TBI) but also perfect for psychosocial disabilities insofar as I have by-default opted out of highly overstimulating “video-first” interfaces like Zoom and MS Teams. Those endemically popular teaching tools greatly favour visual interaction, with prominence given to video feeds and visual content-sharing – with chatbox interaction added as more of a low-benefit afterthought, rather than a central community feature. Because of my hyperfocus and aural hallucinations, I am much more easily able to traverse platforms that favour text-first interaction (and also provides the possibility of disengaging and reengaging depending on my cycling phase – this is important! Sync video leaves out those who cannot be available consistently in abled timeframes).

These text channels are stable and asynchronous: this means that unlike Zoom or Google Meet, any text entered into any channel on the server does not delete after synchronous calls end. This provides a lot of possibilities for combining sync/asynchronous interactions (and continuing productive conversations!), as well as rendering old-tech, panoptical “discussion boards” available on Blackboard, Brightspace and Canvas relatively obsolete.

But simultaneously, I recognize that fully-non-visual teaching is probably not in our near future. Here is a screen capture of a synchronous interaction with the students during a Monday morning writing class:

You’ll see the interfacing has changed a little here, which can take a minute to get used to (I have written onboarding documents that are free to use!). The left-hand panel now displays a pop out panel (which can be dragged anywhere on your monitor) depicting myself streaming via webcam, and my screen share with my students (“showing what’s on your monitor >> to their monitors”). The center panel has remained the classroom anchor, live-updating the “chat” window as students join the server and write messages to each other to be seen both synchronously and asynchronously, while the right-hand panel remains stable as a repository of server members (and their respective roles, by default colour-coded but also delineated clearly by “role markers” rendered in plaintext).

I use the synchronous classes to run through powerpoint presentations with my students – a higher ed classroom mainstay – and sharing my own video for personality, as I’ve found the students’ morale improves if they are able to see me lecture. That said, there is no black box awkwardness (endemic to platforms like Zoom and MS Teams) when students choose not to share video. I have a policy of never requiring video because we don’t know what students’ home situations are and whether they are comfortable making their private environments public displays – I do not think we should impose public presentability on spaces meant to be private. They can “opt in” to seeing my camera and/or my slides (and it is possible to see my slides without seeing my camera feed!), and I leave that interactionism up to them.

The screen share feature is a direct copy of your monitor – so unlike Google Meet (which requires you to select 1 monitor) or Bongo (which requires you to select 1 application), I can cascade through a powerpoint presentation, a Word document, a PDF rubric and photo libraries of my cats all in the same lecture without any disorganization or recapturing. Similarly, students can choose how large your screen-share feed is (depending on their physical or technological needs, including bandwidth restrictions) and re-render your video feed transmission without your permission (if I stream at 750fps, students can choose to watch it in 250fps etc.). This is important for students who do not have easy access to more expensive technologies like a good quality laptop: these lectures are completely watchable on smartphones, tablets, even a smart fridge if you wanted to bother installing Discord as an active app on it.

Let’s take a look at a more recent server screen capture:

You’ll notice a number of channels on the left-hand server panel. I’ve created a central “#classroom-live” channel where sync/asynchronous lectures and post-lecture conversations happen, but the students have a number of other ways to interface and build community within the space of the “ENGL109 Classroom”. Some of these channels are full-server channels (meaning everyone registered can see them): #homework-help is an asynchronous Q&A that takes away the “awkwardness” of having to e-mail me questions or visit synchronous office hours to ask one two-minute question. Students can ask each other homework questions, or they can summon me directly to reply. This also acts as a ready-repository of ALL questions asked – we know that students tend to ask the same questions! So this saves you a lot of recursive e-mail answering time, or posting long “announcement” posts containing Q&A questions you’ve been fielding all on your own. As semester has progressed, students have felt more comfortable answering each other’s questions rather than always waiting for me – I see this as a big win for community building within the Discord server.

There are also 5 “secret” channels toward the center of the left-hand panel: “#rats-private”, “#tigers-private”, “#lemmings-private”, “#gators-private” and “#polar-bears-private”. These are student-selected Animal Kingdom team names, based on teams they were randomly assigned in Week 1 of this semester. There are 5 students per team (and the right-hand server user panel denotes by colour and by column which online users belong to which team), and they are given entirely private spaces to do team activities with each other and interact in classroom space. These channels are “greyed out” because I have them set to muted: this means I can see their conversation, but I do not get messages/pings about their use of their team community window unless they specifically summon me with “@sarah currie (Instructor)”. Students are told very early on that I am not monitoring that space – something we can’t emulate the same way on Brightspace or Blackboard forums – and they can synchronously or asynchronously leave messages for each other or use that window as a group chat to build camaraderie and inner community. I think we should be trusting students to more non-supervised spaces within classroom architecture, but if there is a problem reported or they need me I can be invited into their space. Being “invited” sends a much different message than “panoptically watching all team interactions via server forums”. It builds trust, and it builds team spirit as the “personalities” of each chat develop over the course of the semester.

Similarly, all teams have their own A/V channels (denoted with the “sound icon” toward the bottom of the left-panel server lists. They can use those private team channels for synchronous video-audio calls (though audio is by default prefaced on Discord, rather than visually punishing you for not wanting to use video), homework workshops, synchronous team activities, or asynchronous chatting and team interactionism. Similarly to the text channels, when we are outside of synchronous class contexts (where I sometimes use these rooms as pseudo-”breakout rooms” and will come chat with each team privately), I won’t join their A/V servers outside of synchronous classes: those are their workspaces to use however they wish, even if they’re not working on content from my course.

Continual, Dynamic Community Building

It’s important to reiterate the power of servers designed to create community first, instead of monitor interactionism as a primary premise. Online, I’ve been calling this “community-first versus policing-first” architecture: much of MS Teams, Blackboard and Brightspace design spaces to be monitored and tightly controlled via panoptical strategies of dataveillance taking, lack of private space, and attendance-based completion checking. On Discord, the paradigm is reversed: I am invited into spaces and the classroom is mostly their architecture to use how they please. Students use the “#off-topic” forum to share pictures of their pets and boost morale, while I use channels like “#announcements” to keep the course on track and send gentle reminders about upcoming work due, upcoming meetings or suggestions for teamwork elements. I frequently reinforce and remind students how to use the platform itself, and continually reinforce that this is mostly their space and I’m just a facilitator within it to make sure things don’t go haywire. Team chats are relatively active during class workweeks and easily allow for crip time (because of the text-first reliance), as all of these channels premise asynchronicity as perfectly acceptable interaction – this is opposite of Zoom or MS Teams, who clearly treat asynchronicity as more of an “afterthought” or “catch-up metric” for those who do not have the ability to be available during synchronous timeslots or when everyone else works on abled time.

The always-availability makes any work style acceptable (forums like Brightspace will “flag” students who work “very early in the morning” or “very late at night” as risk-students), and the server itself is only 80mb. This is a stark contrast to RAM-heavy platforms like Teams, which has become infamous among pandemic architecture for shutting down or freezing lower-cost computers, laptops and cell phones. Discord is free and does not require Enterprise licensing from your institution, it does not have overlays for phone interaction (overlays and retrofitting is very passe in the disability community), and it does not matter where or from what timezone you’re interacting with content. This way, students not on “normative time” aren’t afterthoughts – they have equal interactionism with course components and anytime is a good time for interfacing with the ENGL109 Classroom server.

Let’s talk more explicit accessibility while I have you here:

This screen capture displays the Accessibility Dashboard, which every server admin and server user has access to via a “gear symbol” control panel. Basing the encoding in HTML offers students dynamic text interfacing options, including bold and italicized text, server-created emojis and images, image-sharing, file-sharing, GIF-sharing and sticker sharing. This allows psychosocially disabled users to flair content, censor content (with the classic “black bar” reveal function), italicizing warning content, as well as the ability to communicate their thoughts and idea via mediums other than voice and text. Sometimes, an image (or a meme) really does communicate a thousand words better and faster, and visually-oriented students love making use of that feature. Students are also able to “@ping” specific students, specific teams, myself, or other forums (for quick linking across the server, but also linking each other to content they might like).

Discord does not natively offer auto-generated closed captioning (though the Scriptly Bot does offer autogenerated text to speech and screenwriting!), BUT it is worth knowing that autogenerated closed captioning is not actually ADA compliant as a classroom accommodation! These autogenerated CCs developed on Google Meet, Zoom and MS Teams only produce 60% accuracy, a full 20% below the minimum requirement: if a student in your class has asked for closed-captioning as an accommodation, autoCC is actually not good enough to be compliant with accessibility laws. Recording softwares like Screencast-O-Matic and Vidyard have very robust closed-captioning editing tools that you can use after lectures to fix recorded captions, to be able to post an ADA-compliant video for asynchronous viewers. I record all my live lectures using Screencast-O-Matic, it takes about an hour to edit and correct autogenerated CC content generated by the program, and then I post compliant CCed videos to my students’ LMS for anyone who missed the live lecture. This makes Discord JUST as CC-compliant as any other “canon” pandemic platform, if you are paying attention to accessibility laws.

For those of you within access community, Discord is also RIAA and WASG compliant, as well as optimized for RIAA (UK), JAWS, TTS, screenreaders ZoomText and Kurzweil. This is already more than “canon” platforms, but I can keep going. The accessibility dashboard above also denotes a number of psychosocial customizations: reduce noise, reduce motion, colour inversion, disable GIFs, and customizing saturation. This is great news for mad kin like me, who make use of saturation slide bars and text-first interfacing. This is great news for mad kin with ADHD and need motion reduction and GIF disablement; for chronic pain users who may need saturation or reverse-colourization for easement on the eyes; for Blind users who require TTS or Kurzweil interfacing; even for epileptic users who frequently make use of reduce motion and disable GIFs to improve their stabilizing experience. Most “accessibility vetting” DOES NOT take into account mental illness or psychosocial disability accommodations! This is something we are responsible for noticing, and intentionally picking programs that provide some of the current available accommodations for these students and instructors.

Many neurodiverse students rely on custom keybindings and lowtech settings to interface well with many extended hours of online learning – Discord handles that, Zoom does not. I know most pandemic instructors never asked for a crash course in online compatibility – but students didn’t ask for that crash course either. Responsibilizing them with creating and applying their own accommodations in our classroom environments is deflection of institutional responsibility to provide them safe learning environments. More so, it just makes us better, more compassionate instructors to be aware of all the problems and interface blockages in our community rooms – not just the most popular ones.

If you want to know more, I did a whole seminar on “Compassionate Pedagogy and Discord” for DHSI’s Pedagogy Special Series in 2021. I recognize and empathize with the exhaustion and the trepidation instructors have for introducing new tech to the “pandemic canon” – but the pandemic canon only took into account the most visible disablements and disabilities, which is to say not very many of them at all. We have so many students every term with psychosocial disability, mental illness or neurodiversities who were not accommodated pre-pandemic, and they continue to receive precious little accommodations 2 years into online learning, despite a wealth of solutions already available to them that we “refuse to utilize”.

Yes, pandemic instruction is hard. But we can build back better with open access servers that develop strong online community atmospheres. Psychosocial disability should be intentionally welcomed to our classrooms too, with community-first pedagogy: not policing-first.

Contributed by sarah madoka currie

29 January 2022