Eyewitness testimony by individuals who stammer: Evidence, experience and perceived credibility · Maras K, Park S, Grafton P et al.

GENERAL DISCUSSION

In this first empirical examination of eyewitness testimony and stammering, we examined the evidential quality (Study 1a), lived experience (Study 1b) and perceived credibility (Studies 2 and 3) of witnesses who stammer. Across three studies, four key findings emerged. First, while PWS recalled a similar amount of information (and as accurately) as cognitive‐ability matched comparison participants across the interviews overall, they recalled fewer correct details during the initial free recall phase of the interview. This difference between groups dissipated in the cued recall phase. Second, PWS reported experiencing the interview as a cyclical interplay between communicative pressure, anxiety over how they would be perceived and their stammer severity. They attempted to navigate this by prioritizing either fluency (at the expense of testimony content) or content (often speaking through their stammer to relay their message). Third, in line with these lived experiences, PWS were perceived by mock jurors as less confident, but equally as capable and more likeable and trustworthy than comparison witnesses. Fourth, informing jurors that a witness has a stammer, and providing them with information about stammering, further increased ratings of their likeability, trustworthiness and capability, but had no effect on how confident the witness appeared.

It is positive that, despite the increased cognitive load that PWS reported experiencing due to anxiety, speech monitoring and concerns over others' perceptions, their overall eyewitness recall was as complete and accurate as witnesses who do not stammer. However, the fact that PWS recalled less information in the initial free recall phase of the interview suggests that the additional cognitive pressures of providing an account without external scaffolding diminished the amount of detail that they were willing or able to report (Boraxbekk et al., 2015; Chevroulet et al., 2021). In contrast, and in line with findings with other groups with communication difficulties such as autism (see Norris et al., 2020; Norris & Maras, 2021), there was no difference between groups during cued recall.

Consistent with this experimental finding, follow‐up surveys and interviews with PWS indicated that they navigated the challenges and demands of providing an account by either engaging in avoidance strategies (e.g. circumventing particular words or phrases they know they would find difficult to verbalize without stammering), or prioritizing testimony content and speaking through their stammer. Notably, individuals who employed avoidance strategies feared this hindered the clarity of their testimony, reflecting the cognitive struggle between achieving fluency and conveying intended messages (Bricker‐Katz et al., 2013; Iimura et al., 2016), paralleling the experiences of bilingual speakers providing testimony in their non‐native language (Akehurst et al., 2018; Alm et al., 2019; Hu & Naka, 2023).

An interesting avenue for future research will be to examine eyewitness recall by PWS in the context of Koriat and Goldsmith's (1996) monitoring and control framework, and in particular their strategic regulation of grain size in memory reporting (Goldsmith et al., 2002; Weber & Brewer, 2008). Within this framework, individuals balance competing goals of accuracy and informativeness by making a series of reporting decisions, such as whether to volunteer a detail (and if so, at what level of granularity) or withhold the detail altogether (Goldsmith et al., 2002). In Study 1, PWS reported needing to prioritize either fluency or accuracy, suggesting that the additional cognitive burden associated with providing an account without external cues resulted in fewer correct details reported in the free recall phase. However, based on the framework of Koriat and Goldsmith (1996), it is possible that PWS simply adopted a more cautious reporting strategy by raising their reporting threshold, choosing in their free recall to communicate only details they were highly confident in. Future studies should incorporate a confidence measure to examine the confidence–accuracy relation at the statement level.

Given that PWS appear to find unstructured recall more challenging, adapted interview techniques for witnesses who stammer may be appropriate. For instance, Maras et al. (2020) developed a Witness‐Aimed First Account (WAFA) interviewing technique for autistic witnesses – a population who experience difficulties in recalling episodic events in free recall. Unlike typical free recall, WAFA encourages individuals to self‐segment the event they witnessed, which is later displayed on Post‐It notes or similar as a reminder of the structure of the event (Maras et al., 2020). WAFA has been shown to yield higher accuracy and completeness of eyewitness accounts by both autistic and non‐autistic adults, and may be a fruitful line of investigation for PWS.

Stammering appeared to influence mock juror perceptions of witness credibility in a more heterogenous manner than anticipated. Specifically, PWS were rated as more likeable and trustworthy, equally as capable, but less confident than comparison witnesses. Whereas previous research found that other groups with communication difficulties (e.g. developmental language disorder, intellectual disability and autism) are often perceived more negatively across credibility domains (Henry et al., 2011; Lim et al., 2022; Spaulding & Blewitt, 2024; but see Maras et al., 2019), stammering seems to differentially influence aspects of witness credibility. This somewhat counterintuitive finding that reduced perceived confidence of PWS was accompanied by increased trustworthiness and likeability challenges the conventional view in forensic psychology that confidence is an important positive determinant of perceived witness credibility (e.g. Slane & Dodson, 2022). It does align, however, with Andrus' (2012) theoretical framework, which posits that the legal system constructs credibility through the appearance of diminished discursive agency. Andrus outlines how spontaneous, emotionally unfiltered, less polished speech is often treated as more trustworthy because of its lack of artifice: The speaker is perceived to be unable to fabricate a story, and their account appears more authentic. Accordingly, it may be the case that PWS are seen as trustworthy primarily due to perceptions of diminished discursive agency.

This has implications for the language ideology embedded within UK and US law (such as Res Gestae evidence and the excited utterance exception to the hearsay law), which not only determines which kinds of speech are admissible in court but also shapes who is seen as a credible speaker (see Andrus, 2012; Coffey, 2023). There is a general tendency of the justice system to be unable to co‐consider agency and vulnerability; discursive agency and trustworthiness are seen as mutually exclusive. That is, a speaker is either agentic and reflective and thus perceived as less trustworthy, or emotional and impulsive and thus seen as more trustworthy. By legally privileging utterances made in emotional states, the system effectively credits speakers when they are presumed to not be in control of what they say (i.e. their statements were an excited utterance). This tension highlights the limited binary structure embedded in the legal system (Andrus, 2012), raising questions around why concepts such as confidence are viewed as essential to credibility, and why this might not always be accurate or serve all populations equally (see also Brown et al., 2017).

It may be pertinent that stammering is a more readily observable characteristic to an outside observer than that associated with other disorders such as autism or developmental language disorder. Indeed, many participants in Study 2 commented on witnesses' stammers. As such, PWS may be perceived as vulnerable from the outset, and likeable and trustworthy as a result. This is in line with research showing that individuals perceived as vulnerable are usually viewed as being higher in warmth (Fiske et al., 2007). The perception of vulnerability may also evoke empathy towards witnesses who stammer due to their noticeable struggle with speech, which jurors might have equated with honesty and reliability (Green & Brock, 2000). Although this has implications for witnesses whose stammering is more covert (i.e. through the deployment of masking, copying or avoidance strategies) and who therefore appear fluent, even though they may not be recalling everything that they would otherwise wish to. Equally, witnesses' stammers – and the para‐ and non‐verbal behaviours associated with managing them – appear to negatively influence juror perceptions of their confidence. This finding aligns with previous research suggesting that nervous and tense body language negatively affects juror perceptions of witnesses' confidence (Iudici et al., 2023; Wessel et al., 2006) and that stammering is a sign of nervousness or lack of confidence (Amick et al., 2017; Boyle, 2017; Sibanda & Mothapo, 2024).

The finding that information provision increased juror credibility ratings of testimony in Study 3 is consistent with previous research with other groups with communication disorders such as developmental language disorder (Hobson et al., 2022; Horsham et al., 2025) and autism (Crane et al., 2018; Maras et al., 2018, 2019). They are also consistent with Amick et al. (2017), who found that increased familiarity with stammering led to reduced social stigma and more positive social perceptions of people who stammer. A likely explanation is that improving knowledge of stammering enables jurors to attribute atypical behaviours – both stereotypical ones (e.g. hesitations and elongations of syllables) as well as lesser known characteristics (e.g. facial grimacing, rapid blinking and strategically navigating syntax) – to the condition, rather than as indicative of a less competent and credible witness (see Kelley, 1973). However, jurors appear to be over‐compensating for the stammer in their decision‐making process, given that in Study 2 (where jurors were not provided with prior information about stammering) PWS witnesses were already rated as more likeable and trustworthy than non‐stammering witnesses; a difference that was amplified with information provision in Study 3. Similar findings have been reported for autistic witnesses, whereby the provision of autism diagnosis information for jurors resulted in them rating autistic witnesses as more credible than their non‐autistic comparisons, despite no differences in the completeness or accuracy of testimony per se (Maras et al., 2019). It may be pertinent, however, that in Study 2, there was no difference between PWS and comparison witnesses in ratings of their speech quality, yet speech quality predicted credibility. It is possible that the free recall component (which was shown to mock jurors) enabled PWS to employ successful masking strategies (albeit at the expense of the completeness of their accounts), resulting in them being perceived as more credible than they would otherwise.

The only aspect of credibility that the provision of information did not affect was perceived witness' confidence. This might be due to ingrained societal biases that equate fluency and smooth speech with confidence (Guyer et al., 2021), or it might simply reflect a genuine lack of confidence on the part of the witness – a supposition supported by the qualitative follow‐up survey and interviews with PWS in Study 1b. It is worth noting, however, that jurors rated how confident the witness appeared generally, rather than how confident they appeared specifically in terms of the reliability of their account. This is an important point for future research to address, particularly in light of previous evidence that fears over misperceptions create a dissonance between PWS's internal confidence in their ability to provide testimony and the anticipated external misperceptions of listeners (Boyle, 2018).

Practical implications

Findings have several practical implications, not least in highlighting the need for individual and systemic accommodations for PWS in providing testimony. Participants expressed confidence that the right adjustments could mitigate the negative impact of testimony contexts, facilitating the clarity and accuracy of their testimony, but that this requires a multifaceted approach. In their post‐testimony reflections, PWS discussed the importance of individual accommodations, such as providing extra time, incorporating pauses and allowing appropriate preparation. These were deemed essential for alleviating stammer‐related anxiety and facilitating better speech management. As noted above, adapted interview techniques such as those developed for autistic witnesses may be beneficial (Maras et al., 2020); however, the current findings also underscore the importance of positive interpersonal dynamics between interviewers and interviewees, which extend beyond verbal communication. Indeed, when interviewers built rapport and demonstrated positive body language, patience and understanding, participants reported that they felt better able to provide testimony due to reduced fear of negative reactions (see also Connery et al., 2020). Some participants also suggested that stammer disclosure would be particularly beneficial to their ability to provide testimony by helping to set listener expectations and reducing the pressure they feel to employ these avoidance strategies (Young et al., 2022). However, other PWS are reluctant to report their condition for fear of discrimination (Butler, 2013). Yet the current findings highlight that their perceived likeability, capability and trustworthiness are significantly improved when their difficulties are disclosed. Future research should examine the effect of diagnosis disclosure by PWS not only on how credible they appear, but also on the testimony they subsequently provide in the knowledge that observers are aware of their stammer.

Participants called for increased awareness and education within the CJS to challenge stereotypes and mitigate prejudices that may impact listener perceptions. They also suggested that alternative mediums of testimony should be offered, such as video conferencing, written statements and pre‐recorded testimony at court, to enhance accessibility and reduce anxiety around listener perceptions (Constantino et al., 2022), and that familiarization with the judge and courtroom would be particularly beneficial at trial. These suggestions echo those suggested in the Equal Treatment Bench Book (Judicial College, 2024) and are in line with the Special Measures that are available for vulnerable witnesses under part two of the Youth Justice and Criminal Evidence Act 1999. Special Measures permit adaptations to how evidence is gathered and presented in criminal courts in England and Wales, including screening the witness from the accused, giving evidence by live link or in private, removal of wigs and gowns while the witness gives evidence, video‐recorded evidence‐in‐chief, video‐recorded cross‐examination and re‐examination, evidence given through an intermediary and the use of aids to communications. A judge may also order any non‐statutory ‘extra’ special measures, where appropriate (see Cooper & Mattison, 2017).

Findings also have wider implications for the participation of individuals with communication disorders in CJS proceedings. Considering the importance of juror perceptions of credibility for trial outcomes, calls for routine assessments to identify language difficulties amongst those entering the youth justice system (e.g. Bryan et al., 2015; Winstanley et al., 2021) could be extended to those providing evidence more broadly.

Limitations

It is important to note limitations of the current research. First, the PWS group in Study 1 was interviewed by an interviewer who already knew of their stammering status. Given that self‐monitoring often arises from the concerns about the perspectives of listeners (Jackson et al., 2021), this may have allowed the PWS witnesses to focus more effectively on the recall task. Future studies involving mock jurors could provide insights into how the presence of evaluators (e.g. police officers, jurors and judges) would impact the performance of PWS.

Second, the testimony interview was conducted virtually, and although the structure mirrored a real‐life interview, it would not have been as cognitively or emotionally demanding (Bagnall & Maras, 2025). Additionally, findings may reflect the views of a niche sample of PWS, given the high chance of selection bias, since the people who volunteered to participate were likely at a specific point with their stammer where they felt confident to do so. Resultantly, the findings may not reflect the views of the broader stammering population, given the diversity of stammering experiences (Tichenor et al., 2022).

Third, there was a baseline age and gender difference between the PWS and comparison groups in Study 1, with the PWS group being slightly older and proportionally more male, which previous evidence indicates may have an effect on eyewitness recall (e.g. Aizpurua et al., 2009; Areh, 2011; Maeder et al., 2012; West & Stone, 2013). It is worth noting, though, that the current PWS group was significantly younger than that of the older adults included in previous studies (Aizpurua et al., 2009; West & Stone, 2013). Groups in the current study were also well matched on cognitive ability. Nevertheless, it is important for future research to replicate these findings with more closely age‐ and gender‐matched groups.