What Is a Frankenbite? The Ethics of Editing Interview Footage

A producer scrubs eight weeks of confessionals hunting for four words that will make a rivalry "pop" for the promo cut. We've seen that search end two ways: an honest line that was already there, or a sentence nobody on camera actually said. A frankenbite is a soundbite stitched together from fragments of different statements, often said at different moments, to produce a sentence the speaker never actually spoke. The term fuses Frankenstein and soundbite for a reason: you're grafting parts from separate bodies of speech onto one frame, and the seams aren't supposed to show. The uncomfortable part isn't that it's possible. It's that it's easy, fast, and in some production cultures, routine.

Origin of the term

Where "frankenbite" came from, and why editors never stopped using it

The word surfaced as shop-floor slang among reality TV editors sometime in the 1980s and 1990s. There's no clean coinage story and no single person who gets credit for it. What made it stick is that it names the technique with uncomfortable accuracy: take a word from here, a syllable from there, stitch them into a line the subject never delivered in that order, in that breath, or at all.

It belongs to a different category than a stringout or a radio cut. Those terms describe how you arrange real selects into a sequence. A frankenbite manufactures a select that doesn't exist in the raw footage. That distinction, arranging truth versus fabricating it, is where the ethics get genuinely hard, and where more than one editorial career has gotten uncomfortable fast.

How it's built

What are the actual mechanics of manufacturing a line from raw footage?

Reality production is the clearest case study because the raw material is so abundant. A contestant asked how they feel about a rival in week two, week five, and week eight hands the edit suite three separate pools of answers to that same question. Multiply that across a cast and a season, and an editor is sitting on hundreds of near-duplicate utterances, any of which can be recombined.

From those pools, an editor pulls individual words or short phrases and resequences them. A pause lifted from one take gets dropped between two words from a different one. Stumbles and filler vanish. The result can run three seconds, sound completely natural on playback, and say something the subject never meant to say.

Picture a concrete, hypothetical version of it: a cast member, asked about a rival in one session, hedges, saying they don't think the person is lying, though they could be wrong. Weeks later, in a different session, the same cast member calls someone else a liar about an unrelated matter. Lift just the accusation from that second clip, place it after a reaction shot from the first scene, and you've built a sentence that never existed as a sentence. Nothing was invented at the word level. Everything was invented at the meaning level.

That's the spectrum worth holding onto: at one end, cleanup, trimming "um," restarts, and dead air so a real answer reads cleanly. At the other, fabrication, assembling a confession, accusation, or character beat the subject would dispute on sight. Most of the argument about frankenbites lives somewhere in that middle distance, not at either extreme.

Documented cases

What happened when the seams actually showed?

The Bachelor franchise is shorthand for frankenbiting because the technique is visible to anyone paying attention. Contestant Blake Moynes became a widely cited example when a promo reassembled his words to imply a statement he hadn't made; once the full episode aired, the gap between promo and reality was obvious. Former contestant Jamie Otis, from season 17, put it bluntly: "I never, ever, ever said what it looks like I am saying on the show."

The more consequential case is the 2016 documentary Under the Gun, produced by Katie Couric. After an interview with members of the Virginia Citizens Defense League about background checks, director Stephanie Soechtig inserted a long pause following Couric's question. In the actual interview, the members had answered immediately. The edited pause made them look stumped. The Virginia Citizens Defense League filed a $12 million defamation lawsuit over it. A federal judge dismissed the case in 2017, finding the group had not shown actual malice, but the editorial manipulation itself was never seriously disputed, and Couric later acknowledged asking the director about the pause and being told it was added for effect. It remains one of the most cited frankenbite cases in documentary, not reality TV.

For the academic version of this argument, Manfred W. Becker's Creating Reality in Factual Television: The Frankenbite and Other Fakes (Routledge, 2021) draws on interviews with dozens of documentary and reality editors across North America and Europe. Becker is a documentary filmmaker and associate professor at York University's School of the Arts, Media, Performance & Design, which is why his findings land differently than pure media-studies theory: he talked to the people doing the cutting.

The ethical line

When does tightening a speaker's words turn into manufacturing their meaning?

Editors have always compressed interviews. A 45 minute sit-down becomes a two minute segment, which means the vast majority of what was said never airs. Nobody argues against cutting filler, repetition, or a tangent the subject wandered into and back out of on their own.

The line isn't duration, it's meaning. Honest documentary interview editing takes what a subject actually meant and finds the sharpest true version of it already sitting in the footage. A frankenbite pulls fragments out of their original context and assembles a new meaning that wasn't there. The voice on screen is genuinely theirs. The statement is not.

Michael Rabiger's Directing the Documentary (Focal Press, first published 1987) describes what amounts to an inside out method: build structure entirely from what people actually said, which constrains the edit to the truth of the footage rather than a predetermined story you're hunting for proof of. A frankenbite runs that process backward, starting from the dramatic line you want and mining the footage for pieces that can be forced into saying it.

Errol Morris, director of The Thin Blue Line, has described refusing to cut from a transcript at all, telling Transom that he edits from the tapes themselves rather than the page. His objection runs parallel to the frankenbite problem: both are warnings against treating the page as the truth when the person on screen is the actual truth. Worth taking seriously even if your process is transcript first start to finish. Every selected line still needs a real listen, in context, before it locks.

Reality TV and documentary

How did a reality-TV technique become a documentary-ethics problem?

Reality television all but invites the frankenbite because the production model demands it. There's no script. There's rarely a clean second take of the exact line you need. What there is, across weeks of shooting, is a massive pool of raw material where the sentence you need might exist only in fragments, scattered across a dozen sessions.

Documentary carries a harder ethical burden. Its subjects are presented as real people speaking for themselves in a specific, real context, and the audience's trust depends on that being true. Paper editing as a discipline exists to build the best honest assembly the footage supports, not the most dramatically convenient one. Becker's research found frankenbiting shows up in documentary more often than editors tend to admit, frequently defended internally as mere cleanup, and that where the ethical line sits shifts with who the subject is and what's actually at stake. A frankenbite in a branded testimonial is one category of problem. One in a political documentary is a different animal entirely.

Common mistakes

Where do even careful editors cross the line without meaning to?

Nobody sets out to fabricate a quote. Most frankenbites happen gradually, through small decisions that each feel defensible in isolation. We'd flag these four as the places it usually starts.

• Trusting the transcript over the recording. A line reads clean on the page but sounds sarcastic, hesitant, or joking on playback. If you build the cut from text alone and never listen back to the assembled sequence, you'll miss tonal mismatches the words can't show you.
• Closing a gap without flagging it. Joining a sentence from minute four to one from minute thirty-one, inside the same interview, isn't automatically dishonest. Doing it without a mental or literal note that you did it is how the practice drifts into carelessness.
• Treating "it sounds natural" as the bar. A smooth edit and an honest edit aren't the same test. Smoothness is a production value. Honesty is a different, non-negotiable standard, and a bite can pass the first while failing the second.
• Cutting for the headline instead of the context. If a statement only makes sense with its setup or its caveat, the short, punchier version without that setup is a choice you're making, not something the footage forced on you.

The broader skill of pulling genuine, honest selects is covered in how to find the best soundbites and how to do a paper edit, step by step. Working only with what a subject actually said is a constraint, and constraints usually produce better stories than manufactured ones.

Transcript-linked editing

Does editing from a transcript make a frankenbite more or less likely?

Part of why frankenbites slip through on a traditional timeline is that trimming and dragging a clip severs it from where it came from. A few cuts later, its original position in the conversation is invisible unless you went looking for it.

A transcript-based editor keeps every word tied to its exact place in the source interview, which is the opposite problem to solve for. In ScriptCut, AI Select marks the strongest moments in a long recording while keeping each one attached to its original timecode, so selecting a line for your assembly shows you exactly where it sat in the conversation, what the subject said right before it, and what came right after. That doesn't stop anyone from building a frankenbite on purpose. It does make the joins visible instead of invisible, which is the first condition for catching a bad one before it ships, since you or a producer reviewing the cut can see the gap between two selected lines at a glance instead of hunting for it. The full mechanics of that kind of AI-assisted selection and story building are worth understanding before you decide whether to trust it on a sensitive subject.

The discipline is still on you. But a workflow where the transcript doubles as both the selection surface and the playback check, something also worth setting up if you're running a transcript studio workflow in Premiere Pro, makes context harder to ignore by accident. Fewer manufactured moments make it into the cut when the source is one click away from every selected line. We think that visibility, more than any policy memo, is what actually keeps a cutting room honest on a deadline.

If you're building an editorial process around this, the honest version costs you almost nothing in quality and buys you a defensible cut if anyone ever asks where a line came from. Here is how we'd start: begin every assembly in ScriptCut with the transcript visible, listen to every select in context before it locks, and treat "can you defend this join to the subject's face" as the actual bar, not "does it sound smooth."

Sources

• Manfred W. Becker, Creating Reality in Factual Television: The Frankenbite and Other Fakes, Routledge, 2021
• Variety: Katie Couric, Under the Gun Filmmakers Sued for $12 Million for Defamation
• realityblurred: How The Bachelor Distorted One Contestant's Reality
• Transom: Errol Morris, A Conversation
• Michael Rabiger, Directing the Documentary, Routledge