The Hidden Weeks: Transcribing a Study by Hand

This is Talk Type, from the team at Pepys, where we turn talk into text.

Let's do the math nobody warns you about in grad school. You finished the interviews. The recordings are sitting in a folder. And now, before you can code a single theme or write a single line, you have to turn all that talk into text. That part is where the weeks go.

Start with one audio hour. The University of Bath's research-data guide puts hand transcription at four to seven hours of typing for every single hour of audio. Peer-reviewed work stretches it wider. Bailey, writing in Family Practice back in two thousand eight, found at least three hours per audio hour at the fast, low-detail end, climbing to ten hours when you're marking pauses, overlaps, and every nonverbal cue.

Now scale it up. How big is a study? Vasileiou and colleagues reviewed two hundred and fourteen published qualitative studies. The typical one ran somewhere between fifteen and thirty-one interviews. So picture a middle-of-the-road project. Thirty interviews, about an hour of audio each. Thirty hours of talk waiting to become text.

Here's where it gets heavy. Thirty audio hours times the Bath figure of four to seven hours of typing per hour. That's roughly one hundred and twenty to two hundred and ten person-hours. Not minutes. Hours. Somebody, probably you, sitting at a keyboard, pausing and rewinding, for the better part of two hundred hours.

Turn that into a calendar. The U.S. standard workweek is forty hours. Divide it out and a single thirty-interview study eats three to five full working weeks. Call it fifteen to twenty-six eight-hour days of nothing but typing. And that's the honest version, because nobody actually transcribes eight focused hours a day. Add the breaks, the corrections, the second pass, and the real calendar stretches longer.

Maybe you're thinking you'll just run fewer interviews. Fair. Saturation research says the themes often show up early. Hennink and colleagues found code saturation at around nine interviews. But here's the catch. Cutting interviews trims the interviewing. It doesn't touch the typing. Every recorded hour you keep is still an hour you have to transcribe by hand. The workload scales with the audio, not the ambition.

And outsourcing doesn't erase the labor. It just moves it. Someone still types every recorded minute. The person-hours don't vanish, they change hands. Which is exactly why transcription, not interviewing, is so often the longest phase of the whole project.

This is the before. Before you've read a word. Before you've noticed a pattern. Before the actual thinking begins. Weeks of pure mechanical labor standing between you and your own data.

That's the gap automatic transcription closes. Software drafts the whole thing in minutes, so instead of building a transcript from a blank page, you're correcting one that already exists. With Pepys there's no thirty-minute cap either. Long interviews get split and stitched back together, so a two-hour session isn't a problem, it's just a file.

The interviews were the study. The typing was never supposed to be.

That's this episode of Talk Type. The full write up, with the links and sources, is in the show notes. Pepys transcribes any file or link, any length, pay once, and we never train on your audio. Your first sixty minutes are free at pepys dot co. Thanks for listening, and we'll see you next time.