Annotated study list
The “What we read” column tells you whether our summary is based on the full paper or only its abstract, so you know how far to trust the detail.
| Study | Sample | Finding | Main limit | What we read |
|---|---|---|---|---|
| Anderson, Klofstad, Mayew & Venkatachalam (2014), PLOS ONE | 7 women + 7 men recorded; 800 listeners | Fry voices rated less competent, educated, trustworthy, attractive and hireable; women penalised more, by women listeners most | Fry was imitated on request | Full text |
| Alexander & Irons (2022), JASA meeting abstract | Natural vs elicited fry | Elicited fry can differ acoustically from natural fry | Meeting abstract; not a full-length paper | Abstract |
| Chao & Bursten (2020/2021), Hypatia | Philosophical analysis | Dismissing speech because of fry limits women’s communicative autonomy | Argument, not data | Abstract |
| Devaraj et al. (2023), Applied Sciences | 13 listeners; synthesized vowels | Fry perception is driven mainly by low F0; open quotient also matters; fry can be healthy or a disorder sign | Small panel; synthetic stimuli | Abstract and introduction |
| Uusitalo et al. (2022), Journal of Voice | 200 Finnish samples (95 men, 105 women) | Women’s creak rose from the 1990s to the 2010s; men’s change not significant; no age link | One language | Abstract |
| Pillot-Loiseau et al. (2019), Anglophonia | 9 French–English tandem pairs | English speakers creaked more; French speakers creaked more in English | Pilot; 9 pairs | Full text |
| Journal of Voice (2024), Dutch male speech | 30 Dutch men | Creak on about 0–5% of syllables; large individual differences | Men only | Abstract |
| Systematic review of creak prevalence (2020), PLOS ONE | Published prevalence studies | Methods too varied to pool into one English-wide rate | We read the abstract only | Abstract |
| Brown & Sonderegger (2025), Journal of Phonetics | 49 Canadian English–French bilingual public figures | Men and older speakers were creakier by multiple acoustic indicators | One bilingual public-figure sample; author summary and publication metadata | Author summary and metadata |
| Brown & Clayards (2026), Laboratory Phonology | 40 Canadian English listeners; matched-guise experiment | Strong voice-quality and moderate pitch effects; no overall face-gender effect, with a small interaction | Manipulated speech and face stimuli; not everyday prevalence | Full text |
| National Center for Voice and Speech (2026) | Expert commentary | No specific pathology tied to fry; not recommended for speech for clarity reasons | Commentary, not original data | Full text |
| A Brief History of Vocal Fry (2025), ASHA Perspectives | Historical review | Traces attitudes toward fry back to the 1930s | We read the abstract only | Abstract |
| Stewart et al. (2024), Journal of Voice | — | Listed for completeness; we could see only the title, so we do not summarise it | Not yet read | Title only |
Studies about perception
The 2014 PLOS ONE study is the anchor of the social debate, and its method is the anchor of the critique. Its speakers produced fry on request, and the 2022 JASA paper on natural versus elicited fry suggests performed fry may carry extra cues. The 2023 Applied Sciences study adds a technical point: listeners decide something is fry mostly from low pitch. Together they suggest that what people react to is partly the pitch and partly the stereotype. The philosophical side, Chao and Bursten in Hypatia, asks what it means when listeners use a voice quality as a reason to stop listening. More on the social perception page.
Studies about how often people fry
Prevalence studies disagree mostly because they measure different groups in different ways, which is exactly what the 2020 PLOS ONE review concluded. The Finnish, French and Dutch studies each cover one language and one setting. Brown and Sonderegger’s peer-reviewed 2025 paper analyses 49 bilingual public figures and finds more creak in men and older speakers in that sample. Brown and Clayards’ separate peer-reviewed 2026 paper studies perception; it should not be counted as a second prevalence sample. The full comparison is on men vs women.
Studies about harm
This is the thinnest part of the literature. The ASHA Leader’s 2016 feature notes that a 2011 study often cited for harm only suggested it might occur. The NCVS’s 2026 commentary finds no pathology specific to fry. Mayo Clinic’s warning about nodules is explicitly theoretical. Our health page sets the clinical positions side by side.
Research on synthetic creak
A separate strand comes from speech technology. A KTH team published neural speech synthesis with controllable creak at ICPhS 2023 and the CreakVC voice-conversion system at Interspeech 2024. They matter for why AI voices creak.
People you will see quoted
- Jeanne Brown (McGill University) co-authored the 2025 production and 2026 perception papers and presented related work at ASA in 2026.
- Lisa Davidson (New York University) is a linguist quoted in New Scientist’s coverage of that study, as an outside commentator.
- Penny Eckert (linguist) was NPR’s main voice in the 2015 “are we policing young women’s voices?” coverage.
If you prefer listening, The Vocal Fries is a linguistics podcast about language attitudes; despite the name, it covers far more than fry.