We in UCD and product development are very good at participant safety. Consent. Distress protocols. Safeguarding routes. A list of support organisations at the bottom of the discussion guide, actually listening to people during sessions to know when to intervene.
All of it matters.
Then the session ends, and a calendar invite tells you the next one starts in ten minutes.
What it actually looks like
I've sat with someone a few weeks after they lost a person they loved. They weren't there to talk about grief. They were there to talk about the service they'd been forced to use because of it. So I got both, in the same breath, with a discussion guide on my lap and my scratty little notebook with my chicken scratch notes.
I've spoken to parents terrified that if something changed with their benefits, they wouldn't be able to feed their kids. Not “concerned”. Terrified. Doing the maths out loud while I sat there with a prompt about “ease of completing the form”.
I've spoken to people convinced they were about to lose their jobs, trying to work out whether the thing I was researching would replace them.
I’ve gone down incredibly dark tales of addicts - learning about their gambling, substance issues and how they’re worried that they’ll succumb again.
And I've spoken to victims of modern slavery. One of those conversations was about a local street I’ve been down thousands of times. Shops, like real shops, proper shops with brand names. Bus stops. And groups of women were held there. Whilst people are just walking past, thinking about what to have for lunch, or moaning that they’ve missed the bus.
I’ll never look at that particular street the same way.
Then after that research session, straight into the next one. Resetting. Sat in front of another group of people and asked a stranger how they found the sign-in process.
Nobody planned for that bit. Nobody ever does.
We are not fucking recording devices
Research loves to pretend the researcher is neutral equipment. Questions in, insight out, no residue.
That’s bollocks.
We're people. It doesn't stay in the transcript. It comes home with you. It turns up at 11pm, or in the shower, or while you're stood in the kitchen staring at a pan of pasta.
One heavy session is manageable. Five in a day is something else. Five in a day, then a findings workshop, then “how's the report coming along?” isn't resilience. It's a working pattern, and we've designed it badly.
It's not just me
I'd love to say I'm an outlier and a special little snowflake. I'm not. (I mean I am, but that’s not for discussion right now.)
I've heard stories from other researchers, and they're grim. People stacking up the worst subjects imaginable with nothing behind them. People making a safeguarding call alone and going straight into a stand-up. People finding out the “support” on offer was a generic helpline number and a paragraph in a handbook or a squishy people manager going “awww, you ok?”
As UCD people we often moan and share horror stories about bad jobs, obnoxious teams or having to justify the job - but this is different, it’s about human support on hard topics, which are compounded.
The common thread is how threadbare it all is. Not weak support. Barely any.
A discipline that won't shut up about empathy, with next to nothing built to sustain the people doing the empathising.
"We're not therapists" is true, and it's not an answer
Yes, we're not therapists. Yes, boundaries matter. I'm not arguing for researchers to rescue participants or fall apart mid-session. I do not need an emotional support animal, although my laptop becomes the most interesting place to sit for my cat just as I’m about to start a remote session.
A boundary is not the same as being unaffected. Running a hard interview professionally and being fine afterwards are two completely different things, and treating them as one means nobody says anything until it's too late.
And the better you are, the worse it gets. Good researchers build trust fast. They sit with silence. People say, “I've never told anyone this before.” That's the skill working.
It also means you're the one holding it.
Do not tell me to TAKE A FUCKING WALK
When this comes up, the answer is always individual. Look after yourself. Take a walk. Have a bath. Have a drink tonight.
If your organisation has booked seven heavy interviews in a day, given you no support, and wants the playback by Friday, no amount of self-care fixes that.
“Resilience” has become a polite way of saying: “we've built a shit system, so get better at coping with it.”
This is a design problem. We design things for a living. So let's design this properly.
What I'd actually change
Put it in the research plan. If the subject is bereavement, abuse, exploitation, debt, or illness - say so up front. Plan for the researcher as well as the participant.
Work with the researcher to get their cadence and understand their way of working. If a researcher feels like it’s going to be heavy and they need to space it, do that. Real gaps. Not "grab a coffee while the next link loads". Time to step away, decompress, and write up. If a researcher feels like they just want to get it all done as quickly as possible, let them - but then allow the decompression and build it in.
Pair up on the hard stuff. One leads, one supports, both debrief. Nobody should be holding a disclosure and a safeguarding decision alone. Not just observers, on truly hard topics, buddy up with other researchers, or failing that, other people to prompt and get th answers to the difficult questions.
Schedule the debrief. “Shout if you need me” isn't a process. It’s an empty message.
Let people say no. "I'm not the right person for this session" is a valid professional judgement, not a confession of weakness.
Know the escalation route before you need it. Distressed participant. Distressed researcher. Disclosure of harm. Decide who to call in advance, not while sitting in the mess.
None of this is expensive. None of it is dramatic. It's mostly diary management and someone being willing to treat it as part of the job. I keep banging on about how people aren’t “resources”, this is at the core of it, treat your team as people.
Which makes it even more infuriating that we don't and drones to move on a jira board.
What I don't know
I don't have a perfect model. I've seen good practice in places and absolutely nothing in others, and I want to know what's out there.
If your team does this well, for the love of god, tell me how. If it does it badly, tell me that too and I can see if I can help in any way.
Because we spend a lot of time asking whether the participant is okay. Good. Keep asking.
It just can't be the only question.
Who's looking after the person doing the listening?