Heuristic
Do not make zero filled pauses the speaking goal
A cleaner transcript is not automatically a better conversation.
When it fits
- Every 'um' or 'uh' in a recording is counted as a failure and fluency training has become filler elimination.
When to avoid it
- Filled pauses vary by language, speaker and context. Habit-reversal methods can reduce them, but small experimental evidence shows that reducing disfluency alone need not improve overall public-speaking ratings.
Why it matters
Treat recurring filled pauses as one possible signal, not the definition of speaking quality. They can matter locally when they make an answer sound unready, uncertain or hard to follow, but recent spontaneous-speech experiments found no simple global penalty on the speaker's overall competence, trustworthiness or warmth. If fillers cluster at a recurring transition, train that transition. Then check the message, listener response and overall delivery—not just the count.
An example
If 'um' repeatedly appears before the recommendation, practice a brief silent pause followed by the recommendation. Do not spend the week removing harmless fillers from casual conversation.
Check your result
The targeted moment becomes easier to follow or sounds better prepared without making the speaker more rigid, rushed or scripted.
Keep this limit in mind
- Filled pauses vary by language, speaker and context. Habit-reversal methods can reduce them, but small experimental evidence shows that reducing disfluency alone need not improve overall public-speaking ratings.
Evidence and sources
In 2026 experiments using spontaneous speech, filled pauses did not reduce global ratings of speaker competence, trustworthiness or warmth, but they did affect some local judgments of readiness, certainty, trust and future interaction; effects also differed with perceived expertise.
Laboratory judgments do not define the value of filled pauses in every language, culture, task or relationship.
Does disfluency affect social judgments and decisions? Evidence from spontaneous speech · Abstract results
In a small multiple-baseline study, awareness training reduced public-speaking disfluencies, but expert ratings of overall speaking effectiveness did not improve and sometimes worsened after the disfluency-focused component.
Only two participants were studied; the finding is a caution against equating one metric with overall quality, not an estimate of typical effects.
The efficacy of remote video-based behavioral skills training and awareness training on public speaking performance · Abstract results