Stuttering and anxiety responses in virtual audiences closely correspond to those in live audiences
How this was rated
Within-subjects comparison in adults who stutter (n = 10) with very strong correlations between virtual and live conditions. Downgraded because of small sample size and single speaking task. This is a foundational validation study, influential in the field; confidence in the core finding has grown as later work has replicated the pattern.
Ratings use a simplified four-tier scheme (High, Moderate, Low, Very Low) informed by the GRADE working group. Learn more about how studies are rated.
A foundational study in ten adults who stutter showing that stuttering frequency during a challenging virtual audience speech correlated at Spearman rho = 0.99 with stuttering during a live audience speech, and that anticipatory apprehension and confidence measured before the virtual condition correlated strongly with the same measures before the live condition (rho = 0.82 and 0.88 respectively). The neutral virtual audience also correlated with the live condition, but less strongly (rho = 0.82 for stuttering frequency).
The evidence suggests that a well-designed virtual audience can evoke communicative responses closely matching those in a live audience, supporting VR as an ecologically valid rehearsal context for adults who stutter. This is one small-sample study and findings need to be interpreted alongside later work.
Key findings
- Stuttering frequency (percent syllables stuttered) during the challenging virtual audience condition correlated at Spearman rho = 0.99 with the live audience condition (p < .001); the neutral virtual audience also correlated significantly with the live condition at rho = 0.82 (p = .004)
- Pre-speech communication apprehension (PRCA-24) correlated at rho = 0.82 (p < .004) between virtual and live conditions
- Pre-speech speaker confidence (PRCS) correlated at rho = 0.88 (p < .001) between virtual and live conditions
- Mean stuttering frequencies were similar across conditions: live audience 2.73% syllables stuttered, neutral virtual audience 2.55%, challenging virtual audience 2.86%
- Participants' qualitative comments described feeling genuinely present in the virtual environment and experiencing physical anxiety responses consistent with real audiences
Background
One of the most persistent challenges in speech therapy is bridging the gap between the therapy room and everyday life. In-clinic practice often feels safe and manageable, but the real world introduces unpredictability, social pressure, and high-stakes situations that are difficult to replicate in a controlled setting.
Brundage and Hancock set out to test whether virtual reality could fill this gap - specifically, whether virtual audiences produce the same kinds of communicative responses (emotional, behavioral, and cognitive) as a real audience in adults who stutter.
What the researchers did
Ten adults who stutter each delivered three five-minute extemporaneous speeches over two days: one to a live audience of eight to ten people and two to virtual audiences delivered through a head-mounted display (one virtual audience with neutral behavior, the other with challenging behavior such as inattention and falling asleep). The order of speeches was counterbalanced. Speech topics were drawn at random from a stack of cards immediately before each speech, with no preparation time. Stuttering frequency was measured as percent syllables stuttered. Communication apprehension was measured with the Personal Report of Communication Apprehension-24 (PRCA-24) and confidence with the Personal Report of Confidence as a Speaker (PRCS), both administered before each condition.
What they found
Spearman rank-order correlations between virtual and live conditions were strong across all three measures. Stuttering frequency during the challenging virtual audience correlated almost perfectly with stuttering during the live audience speech (rho = 0.99, p < .001). The neutral virtual audience also correlated significantly with the live audience condition, though less strongly (rho = 0.82, p = .004). Communication apprehension scores (PRCA-24) measured before entering VR correlated at rho = 0.82 (p < .004) with scores before the live speech. Confidence scores (PRCS) correlated at rho = 0.88 (p < .001). Mean stuttering frequency was similar across the three conditions (live 2.73%, neutral virtual 2.55%, challenging virtual 2.86%).
Participants’ qualitative comments described feeling genuinely present in the virtual environment. They reported experiencing physical symptoms of anxiety, behaving as they would in front of a real audience, and finding the challenging audience particularly realistic. Several participants noted that the neutral audience felt unexpectedly unsettling because the avatars stared without blinking - highlighting that audience behavior, rather than audience size alone, drives communicative responses.
Why this matters
This study established a critical foundation: well-designed virtual speaking situations can elicit genuine communicative responses, not simulated ones. The near-perfect correlation between VR and live stuttering frequency means that clinicians can trust that work done in virtual environments reflects how someone would respond in the real world.
This has significant implications for the generalization phase of therapy, where clinicians need realistic practice opportunities but face logistical and ethical constraints in arranging real-world scenarios. VR offers a way to create graded, repeatable, and controllable speaking situations that feel genuinely challenging - without the social consequences of a difficult real-world experience.
Limitations
The sample was small (ten participants) and limited to adults with mild-to-moderate stuttering severity; correspondence between virtual and live settings may differ in people with more severe stuttering or in adolescents (a population the authors note was not tested). Only one speaking context - extemporaneous public speaking - was examined, so the ecological validity of other scenarios (conversations, phone calls, job interviews) was not directly tested. The virtual audiences contained approximately twice as many “people” as the live audience, although the authors note this is unlikely to fully explain the findings. The study established that VR elicits responses correlated with real-life equivalents but did not test whether VR-based practice leads to improved real-world outcomes. The correlation coefficients reported are measures of linear relationship between conditions, not equivalence; mean stuttering frequencies and questionnaire scores differed somewhat across conditions even where correlations were strong.
Implications for practice
Because communicative responses in VR closely match those in live settings, clinicians can use virtual audiences as ecologically valid practice environments. VR offers controlled, repeatable, and graded speaking situations that support the generalization phase of therapy, where real-world practice opportunities are often limited.
Implications for research
Follow-up work would benefit from larger samples, a wider age range, non-English speaking populations, and tasks beyond formal presentations. Long-term transfer from virtual rehearsal to day-to-day speaking remains an open question.
Where this connects to Therapy withVR
The study above is independent research and does not endorse any product. The notes below are commentary from withVR on how the themes in this research relate to features of Therapy withVR. The research findings are not claims about Therapy withVR.
Auditorium Environment
This study used a virtual audience - Therapy withVR's Auditorium seats 197 with up to 6 interactive front-row avatars, replicating the public speaking scenario used in this research.
Real-Time Clinician Control
The correlation between VR and real-world responses means clinicians can use Therapy withVR's real-time controls to create genuinely challenging practice situations from their laptop.
Eye Contact Toggle
Avatars can be set to follow the speaker with their gaze - adding the social pressure that this study showed produces authentic communicative responses.
Cite this study
If you reference this study in your work, the canonical citation formats are:
@article{brundage2015,
author = {Brundage, S. B. and Hancock, A. B.},
title = {Real Enough: Using Virtual Public Speaking Environments to Evoke Feelings and Behaviors Targeted in Stuttering Assessment and Treatment},
journal = {American Journal of Speech-Language Pathology},
year = {2015},
doi = {10.1044/2014_AJSLP-14-0087},
url = {https://withvr.app/evidence/studies/brundage-2015}
}TY - JOUR
AU - Brundage, S. B.
AU - Hancock, A. B.
TI - Real Enough: Using Virtual Public Speaking Environments to Evoke Feelings and Behaviors Targeted in Stuttering Assessment and Treatment
JO - American Journal of Speech-Language Pathology
PY - 2015
DO - 10.1044/2014_AJSLP-14-0087
UR - https://withvr.app/evidence/studies/brundage-2015
ER - Know of research that should be in this hub? If a relevant peer-reviewed study is not listed here, send the reference to hello@withvr.app. The hub is kept up to date as the literature grows.
Funding & independence
No withVR BV involvement in funding, study design, or authorship. Summary prepared independently by withVR using the published paper.