Contemporaneous notes and juror decision-making: Do reminiscence and expert testimony play a role?

Abstract

Eyewitnesses are often advised to write down their recollection of the event soon after it occurs. However, they sometimes omit details in the written note and recall them later, a phenomenon known as reminiscence. This research examined the impact of testimonial reminiscence and expert testimony on mock-jurors’ decision-making. Undergraduate students (N = 222) read a fictitious trial transcript, in which an eyewitness had written a contemporaneous note about the event. When testifying, the eyewitness either was consistent with the note or exhibited reminiscence. The transcript also included an expert witness testifying about eyewitness memory before the eyewitness testified, after the eyewitness, or not at all. Neither reminiscence nor expert testimony influenced juror verdicts; however, the reminiscent eyewitness was perceived to be significantly less credible than the consistent eyewitness. Expert testimony, regardless of when it was delivered, helped mitigate this by enhancing juror ratings of the reminiscent eyewitness’s credibility.

Eyewitness testimony is a crucial form of evidence in legal settings (Nayak & Khajuria, 2019), and is highly influential in juror decision-making (Loftus, 1975; Schweitzer & Nuñez, 2018). Given the pivotal role that eyewitness memory plays, researchers have attempted to find strategies to better preserve eyewitness memory. One promising avenue has been to ask witnesses to write down their experiences immediately after the critical event, while memory is still optimal (Gabbert et al., 2009; Horry et al., 2021). Having witnesses engage in immediate written recall helps them to preserve a detailed and accurate account of the incident (Hope et al., 2014; Krix et al., 2016). Empirical studies have consistently found that individuals who engage in high-quality, immediate written recall tend to exhibit more accurate and detailed memories for an incident immediately and after a delay than those who do not record their experiences right away (e.g. Gabbert et al., 2009; Horry et al., 2021; Paterson et al., 2015). Furthermore, there is some evidence that immediate recall can protect memory against misleading post-event information (Gabbert et al., 2012).

Immediate written recall accounts are referred to as contemporaneous notes in legal settings (Deck & Paterson, 2021; Horsman, 2021; Paterson et al., 2018), and their use is becoming increasingly popular in forensic settings (Hope et al., 2011; Horry et al., 2021). The state of New South Wales in Australia has permitted contemporaneous notes as court evidence, and witnesses can read them at trial to refresh their memories (Evidence Act, 1995, ss 32 and 34). With both empirical support and practical feasibility, encouraging eyewitnesses to create contemporaneous notes and use them as supporting evidence in court appears to be a promising strategy to enhance the quality of witness testimonies (Paterson et al., 2018).

However, despite the abundant evidence in support of their memory benefits, little attention has been given to the effects of contemporaneous notes within the legal setting. Specifically, minimal research has examined how the presence of contemporaneous notes influences juror decision-making. Given the empirically supported benefits of contemporaneous notes in improving eyewitness memory and the crucial influence eyewitness evidence has upon jurors’ decision-making, further investigation into the effects of contemporaneous notes on judicial outcomes is warranted. Hence, the current study extended this line of research by investigating the effects of contemporaneous notes in the legal decision process.

Juror perceptions of eyewitness memory and testimonial inconsistencies

Most people understand some basic properties of memory, such as its decline in performance over time (Oeberst, 2012), and they can apply such an understanding when evaluating eyewitness testimony (Lindsay, 1994). However, research has also shown that jurors lack correct knowledge about other, more specific factors affecting eyewitness memory accuracy (e.g. Akhtar et al., 2018; Bornstein et al., 2008; Kassin & Barndollar, 1992; Wise & Safer, 2010). Specifically, Benton et al. (2006) found that jurors disagreed with the experts on 87% of the common eyewitness issues they surveyed. Critically, this lack of knowledge influences how jurors evaluate eyewitness credibility. Their judgements are consistently found to be affected by misconceptions that do not possess an evidence-based relationship with eyewitness accuracy, and they are insensitive to factors that do (e.g. Berman & Cutler, 1996; Deck & Paterson, 2021; Penrod & Cutler, 1999; Wise & Safer, 2004).

One of the most common misconceptions is the belief that testimonial inconsistency is a robust predictor of eyewitness inaccuracy. Jury simulation studies have repeatedly found that mock-jurors perceive a prosecution eyewitness who gives inconsistent testimonies as less credible and accurate and the defendant as less guilty (e.g. Berman & Cutler, 1996; Brewer & Hupfeld, 2004; Brewer et al., 1999; Deck & Paterson, 2021; Fraser et al., 2022). Judges and attorneys also hold this misconception. Specifically, many courtroom instructions direct jurors to attend to the consistency of witnesses, whilst litigators often capitalise on witness inconsistencies to discredit them in court (Fisher et al., 2009). However, studies have shown that inconsistency in one testimonial dimension is not necessarily predictive of accuracy in other dimensions, or the overall witness accuracy (e.g. Brewer et al., 1999; Brock et al., 1999; Gilbert & Fisher, 2006; Odinot et al., 2013; Smeets et al., 2004). Additionally, there are various types of inconsistencies (Brewer & Burke, 2002; Erdelyi, 2010), and not all are correlated with inaccuracy. For example, while direct contradictions logically indicate that one of the eyewitness’s statements must be incorrect (Oeberst, 2012) and are indicative of inaccuracy (Gilbert & Fisher, 2006), other types of inconsistencies, such as reminiscence, can be fairly accurate (e.g. Fisher et al., 2009).

Reminiscence occurs when witnesses recall some information later, but not at an earlier timepoint (Payne, 1987). This may occur in the context of contemporaneous notes, when witnesses later recall details (e.g. at trial) that they had not included in their earlier brief notes (Hope et al., 2014). Evidence has suggested that reminiscence is not a predictor of overall eyewitness inaccuracy (Fisher et al., 2009), and reminiscent items themselves have actually been found to be fairly accurate (Brock et al., 1999; Gilbert & Fisher, 2006; Hope et al., 2014; Odinot et al., 2013; Oeberst, 2012). For example, Gilbert and Fisher (2006) measured participants’ recall of a videotaped mock crime at two timepoints: immediately after viewing the video and then after a 48-hour delay. They found that 98% of the participants gave at least two reminiscent details. On average, participants exhibited 8.19 reminiscent details, and the mean accuracy rate for those reminiscent details was 87% (significantly higher than 35% for contradictory details). Additionally, the number of reminiscent details exhibited by each participant was not significantly correlated with their overall accuracy; however, the number of contradictions was negatively correlated with their overall accuracy. Similar results were found in Oeberst (2012). Participants were assessed on their recall of a series of pictures at two time points, and it was found that every participant recalled at least one reminiscent item. The mean accuracy rate for the reminiscent items was 83.9%, which did not differ significantly from that of the consistent items (92.6%). Despite this, participants significantly underestimated the accuracy of those reminiscent items.

Together, these results suggest that reminiscence is, in fact, a common phenomenon during repeated recalls, and that reminiscent items can be fairly accurate, and in some instances even comparable to their consistent counterparts (Oeberst, 2012). Additionally, there is clear distinction between reminiscence and contradiction, with only the latter being shown to indicate inaccuracy. Despite this, many jury instructions (e.g. The United States Department of Justice, 2015), legislations (e.g. Evidence Act, 1995, ss.38) and even empirical studies (e.g. Smeets et al., 2004; van Giezen et al., 2005) still conflate reminiscence and contradiction, presenting ‘eyewitness inconsistency’ in a single, undifferentiated way. This perhaps has contributed to the public’s lack of understanding about reminiscence. That is, individuals seem to be unaware of the true characteristics of reminiscence and significantly underestimate its accuracy. This misconception also seems to be apparent in forensic settings, where the credibility of eyewitnesses is significantly compromised if they exhibit reminiscence over time (Berman & Cutler, 1996; Deck & Paterson, 2021). If jurors can come to understand reminiscence and make decisions based on scientific evidence, then reminiscence should not necessarily be seen as an indication of eyewitness untrustworthiness.

To our knowledge, only two studies have examined mock-jurors’ perceptions about reminiscent eyewitnesses. In the first study, Berman and Cutler (1996) found that participant jurors exposed to a reminiscent eyewitness (who added additional information on the stand that was not mentioned during the pre-trial interview) rated the eyewitness as less credible than those exposed to a consistent eyewitness. They also rated the accused as less guilty, and were less likely to convict. This study, however, examined the effect of reminiscence when the differences were between investigative interviews and in-trial statements. Given reminiscence is highly likely to occur after witnesses have created contemporaneous notes (Hope et al., 2014), it is important to establish whether its detrimental effect on jurors’ perceptions also exists when differences arise between witnesses’ contemporaneous notes and in-trial statements. In Deck and Paterson (2021), mock-jurors read a fictitious trial transcript in which the primary eyewitness either composed a contemporaneous note immediately after witnessing a car accident, or did not compose a contemporaneous note. When there were contemporaneous notes present, the eyewitness’s in-trial statement either was consistent with their note or contained a critical detail that was not mentioned in their initial note (reminiscence). Like Berman and Cutler (1996), Deck and Paterson (2021) also found that reminiscent eyewitnesses were perceived to be significantly less credible than consistent eyewitnesses. They even found the reminiscent eyewitness to be considered less credible than an eyewitness who did not create a note. Taken together, evidence suggests that mock-jurors seem to regard reminiscent witnesses as less credible than consistent witnesses, despite empirical studies suggesting otherwise.

These findings regarding mock-jurors’ misconceptions about reminiscent eyewitnesses reinforce the idea that jurors do not always make judgements based on scientific findings regarding eyewitness memory (Wise & Safer, 2004). This is concerning given the influence that juror decisions have on legal outcomes. Jurors should act impartially and base their decisions on factors found to be diagnostic of eyewitness quality (Cutler & Penrod, 1995), while in reality they seem to lack necessary scientific knowledge and are often swayed by misconceptions. This could potentially lead to miscarriage of justice. Therefore, there is a need for researchers to investigate strategies to educate jurors, aiding optimal juror decision-making and upholding fair judicial outcomes (e.g. Neal et al., 2012).

Expert testimony to increase juror knowledge about eyewitness factors

To address this issue, many scholars advocate presenting expert testimony in court to educate jurors about issues surrounding eyewitness evidence (e.g. Benton et al., 2006; Kassin & Barndollar, 1992; Penrod & Cutler, 1999; Wise et al., 2007). Ideally, expert testimonies should provide jurors with the scientific knowledge necessary to evaluate the evidence at hand, setting aside their preconceived misconceptions and hopefully ensuring fair judicial outcomes (Shermer et al., 2011)

In recent decades, expert testimony has started to play an increasingly significant role in judicial decision-making (Greene & Gordon, 2016). Researchers have found expert testimonies to be effective in increasing mock-jurors’ awareness and sensitivity to factors indicative of eyewitness accuracy. Cutler et al. (1989) had mock-jurors view a videotaped trial in which eyewitness testimony was the central evidence. Half of the participants were exposed to an expert testimony explaining critical, case-relevant eyewitness factors, while the other half were not. Their results suggested that, compared to those who did not hear an expert, mock-jurors who heard the expert placed more weight on factors scientifically diagnostic of eyewitness accuracy (i.e. witnessing conditions), and less weight to factors that were commonly mistaken as indicators (i.e. witness confidence). Similar results were found in Wells (1986).

More recently, evidence has emerged supporting the ability of expert witnesses to correct common misconceptions that could potentially threaten the validity of juror decision-making (Coyle & Thomson, 2014; Deck & Paterson, 2022; Goodman-Delahunty et al., 2021; Ryan & Westera, 2018; Stobbs & Kebbell, 2003). For instance, a series of studies by Goodman-Delahunty et al. (2010, 2011, 2021) focused on the use of expert testimony in reducing misconceptions about child sexual abuse victims. Jurors often hold incorrect assumptions about the victims’ experience and indiscriminately discredit anyone who violates their stereotyped expectations. Such misunderstandings are associated with lower perceived child victim/witness credibility, higher perceived defendant credibility and fewer convictions (Cossins et al., 2009). Goodman-Delahunty et al. found that having an expert clarifying potential misconceptions significantly decreased jurors’ misunderstanding, and such growth in knowledge was associated with increased perceptions of victim credibility and likelihood to convict. Similarly, expert testimony has been found to effectively ameliorate jurors’ uncorroborated, negative perceptions of various types of witness, including repeated-event witnesses (Deck & Paterson, 2022) and witnesses with intellectual disabilities (Stobbs & Kebbell, 2003).

Research suggests that jurors lack knowledge about many factors affecting eyewitness memory (e.g. Akhtar et al., 2018; Marchand et al., 2020; Marr et al., 2021). Additionally, judges may also have inaccurate beliefs about eyewitness memory (Bjørndal et al., 2021; Bond-Fraser et al., 2019). These findings suggest that expert testimony on such issues falls outside the scope of the public’s common knowledge (Leippe, 1995), highlighting the need and the importance of admitting expert testimony in court (Benton et al., 2006).

Taken together, studies have found abundant support for the beneficial effect of expert testimony in ameliorating jurors’ misconceptions and increasing their reliance on scientific findings when making decisions. This, combined with its practical feasibility, renders expert testimony a promising tool to facilitate evidence-based juror decision-making. Therefore, expert testimony might also be helpful in rectifying jurors’ misconceptions about reminiscence and educating them about the benefits of contemporaneous notes. To our knowledge, no previous study has investigated this issue.

The timing of expert testimony

The effect of expert testimony on jurors’ judgements may also be influenced by when an expert takes the stand (Leippe et al., 2004). An eyewitness expert could testify before or after the eyewitness, and there are competing theories supporting each approach.

Based on the Story Model of juror decision-making (Pennington & Hastie, 1992), jurors organise the evidence they hear in trial to form their own story about the crime, which ultimately determines their verdict. Thus, specialised eyewitness knowledge (e.g. about the diagnostic cues of eyewitness accuracy) presented before the eyewitness testifies would better assist jurors in building a story framework to organise and interpret the evidence they hear subsequently (i.e. the eyewitness testimony). In contrast, when eyewitnesses finish testifying, jurors might have already formed their own story about the crime, and those forged beliefs may be hard to disconfirm (Lord et al., 1979). This would mean that expert testimony presented after the eyewitness may be less effective in influencing jurors’ already-formed judgements. Conversely, when an expert testifies after the eyewitness (and other evidence) they might have the benefit of recency (Leippe et al., 2004). That is, hearing the expert testimony later in the trial may mean that jurors are more likely to remember the expert advice than other evidence when considering their final verdict, especially if the case and evidence are tedious and complicated (Horowitz et al., 2001). Also, as individuals’ existing evaluations could be retroactively impacted by new information (Snyder & Uranowitz, 1978), jurors may retrospectively apply the knowledge to evaluate the eyewitness evidence when the expert’s knowledge is fresh in their mind before deliberation. Hence, since both approaches seem to have theoretical support, it is unknown which would yield a more beneficial outcome in practice.

Empirical studies have had mixed findings, supporting both theories. Brekke and Borgida (1988) found that when expert testimony was presented early in the trial and linked specifically to the case at hand, it was able to improve mock-jurors’ perceptions of the victim and increase guilty verdicts, compared to when it was presented after the victim had testified. Similarly, Schuller and Cripps (1998) found that when an expert for the defence testified prior to the defendant, mock-jurors were more likely to believe the defendant, and they gave a more lenient verdict than when the expert presented after the defendant. Conversely, Leippe et al. (2004) found expert testimony about the fallibility of eyewitness memory significantly lowered mock-jurors’ belief in the eyewitness and their ratings of guilt only when it was presented after the eyewitness testimony and accompanied by a judge’s reminder. Contrary to the above studies, Leippe and colleagues’ results suggest that the recency effect may have increased the saliency of the expert testimony. Recently, an additional study investigating the timing of expert testimony failed to find any significant differences (Terrance et al., 2014).

Given the competing theories and empirical evidence, we believe the optimal timing of expert testimony certainly warrants further investigation. We thus aimed to compare the effect of pre-eyewitness expert testimony to post-eyewitness testimony in mitigating mock-jurors’ misconceptions about a reminiscent eyewitness.

The present study

As reminiscence occurs frequently after witnesses compose their contemporaneous notes (Hope et al., 2014), it is important to increase our understanding of how a reminiscent eyewitness would be perceived by jurors. Therefore, the first aim of our study was to replicate the adverse effect of reminiscence on jurors’ perceptions of eyewitness credibility, as found in Deck and Paterson (2021). A replication would reinforce the misconception that jurors hold about reminiscent eyewitnesses, further supporting the notion that jurors do not make decisions about eyewitness memory based on scientific findings (Wise & Safer, 2004). The second aim of our study was to examine whether presenting expert testimony in court could ameliorate jurors’ uncorroborated misconceptions about reminiscence and promote evidence-based juror decision-making. Furthermore, as the timing of expert testimony may influence the persuasive effect of experts (Leippe et al., 2004), our study is the first to compare the effect of an expert testifying before and after the eyewitness in educating jurors about reminiscence.

In the current study participants played the role of a juror and read a trial transcript containing either a reminiscent or consistent prosecution eyewitness. In some versions of the transcripts there was an expert witness who testified for the prosecution before the eyewitness, whereas in other versions they testified afterwards, or not at all. Participants gave ratings of eyewitness credibility, defendant guilt and their verdict.

As reminiscence consistently lowers mock-jurors’ perceptions of eyewitness credibility (Berman & Cutler, 1996; Deck & Paterson, 2021), we hypothesised that a reminiscent (prosecution) eyewitness would be perceived by mock-jurors as less credible than a consistent eyewitness, and the defendant would be perceived as less guilty (Hypothesis 1, H1).

Given that expert testimonies are effective in enhancing jurors’ understanding about eyewitness issues (e.g. Benton et al., 2006; Kassin & Barndollar, 1992; Penrod & Cutler, 1999; Wise et al., 2007), we expected that expert testimony would be efficacious in educating jurors about the benefits of contemporaneous notes and reducing their misconceptions about reminiscence. That is, we hypothesised that when there is an expert witness, mock-jurors would give higher ratings of eyewitness credibility and defendant guilt than when there is no expert (H2).

Additionally, we expected a significant interaction between eyewitness type and expert testimony, such that the lower ratings of eyewitness credibility and defendant guilt for a reminiscent eyewitness compared to a consistent eyewitness would be significantly less pronounced when there is expert testimony (H3). Since the expert is expected to educate jurors about the science behind reminiscence, expert testimony should reduce the difference between their belief in reminiscent and consistent eyewitnesses.

Considering the competing theories and evidence regarding the optimal timing of expert testimony, we have two competing hypotheses: if presenting the expert testimony before the eyewitness leads to higher ratings of eyewitness credibility and defendant guilt than when it is presented afterwards, this would support the Story Model (Pennington & Hastie, 1992) and favour pre-eyewitness expert testimony (H4a). Alternatively, if presenting the expert testimony after the eyewitness leads to higher ratings of eyewitness credibility and defendant guilt, it would support the recency benefits (Leippe et al., 2004) and favour post-eyewitness expert testimony (H4b).

Subsequently, we have two competing hypotheses for the interaction between eyewitness type and expert timing. If pre-eyewitness expert testimony is superior at influencing jurors’ judgements, the lower ratings of eyewitness credibility and guilt for the reminiscent eyewitness than for the consistent eyewitness would be significantly less pronounced when the expert testifies before the eyewitness than after (H5a). If, instead, post-eyewitness expert testimony prevails, the lower ratings of witness credibility and guilt for the reminiscent eyewitness than for the consistent eyewitness would be less pronounced when the expert testifies after the eyewitness than before (H5b).

Method

Design

A 2 (eyewitness type: consistent, reminiscent) × 3 (expert testimony: present before eyewitness testifying, present after eyewitness testifying, absent) between-subjects design was employed. Participants were randomly assigned to read a fictitious trial transcript, in which either an eyewitness provided an in-trial statement that was consistent with their contemporaneous note, or an eyewitness later added reminiscent details to their in-trial statement that were not in their contemporaneous note. The transcript also included an expert witness who testified before the eyewitness took the stand, an expert witness who testified after the eyewitness, or no expert witness at all.

The primary dependent variables of interests were participants’ overall perceived eyewitness credibility, judgements of guilt and verdict. We additionally measured participants’ perceived credibility of the eyewitness’s contemporaneous note and in-trial statement, participants’ beliefs in the memory benefits of contemporaneous notes and the accuracy of reminiscent items.

Participants

Undergraduate students were invited to take part in the study in exchange for course credit. Participants were deemed eligible if they were over 18 years old and fluent in English. The age criterion was set as it is the minimum legal age of jurors in Australia, while the language requirement was to make sure participants could sufficiently comprehend the written transcript they were required to read as part of the experimental task. We did not collect data on additional jury eligibility criteria such as Australian citizenship or criminal record status. However, previous research has shown that jury-eligible community members and undergraduate students are similarly influenced by testimonial inconsistencies when evaluating eyewitness credibility and defendant guilt (Berman & Cutler, 1996). Moreover, a meta-analysis of 53 jury simulation studies by Bornstein et al. (2017) found only minimal differences between student and community samples in trial-related decision-making. Therefore, the use of a student sample in the present study is unlikely to have meaningfully impacted the results.

A total of 375 students completed the study. Of these, 50 failed the attention check question, and 103 failed one or more of the three manipulation questions. All of these individuals were excluded from the analyses because they either did not pay attention while completing the questionnaire or failed to understand or pay attention to critical manipulated details of the study. This left a final sample of 222 participants (48 males, 172 females, 2 undisclosed, M_age = 20.54, SD = 3.725). Participants self-identified as Asian (50.45%), Caucasian (33.78%), Middle Eastern (3.60%), African (0.45%), mixed cultural backgrounds (9.46%) or other (2.25%).

The final sample size was deemed adequate, based on a power analysis. Previous research investigating jurors’ perceptions of reminiscent witnesses and the efficacy of expert testimony have typically found medium effect sizes (e.g. Deck & Paterson, 2021; Terrance et al., 2014). A power analysis using G-power revealed that a total sample size of 202 would be required to detect a medium effect size for a 2 × 3 between-subjects analysis of variance (ANOVA; with interactions) with 90% power and alpha .05.

Materials

The study was pre-registered, and all materials can be found on the Open Science Framework via the following link: https://osf.io/g5hr3/?view_only=ee534cc007fc44abad170cfe4c19b7a7

Transcript

The study used a fictitious trial transcript of R vs Olivia Verco, adapted from the stimulus originally used in Paterson et al. (2013). In the trial, Ms Verco (the defendant) was accused of dangerous driving occasioning grievous bodily harm. The defendant was alleged to have been using her mobile phone while driving, resulting in her car crashing into another and causing the driver of the impacted vehicle to fracture her ribs and suffer a concussion. The transcript described the entire trial proceeding, including opening statements from both the prosecution and defence, testimonies and examinations of witnesses, closing statements and the judge’s concluding instructions to the jury. Six slightly different versions of the transcript were created in accordance with the six experimental conditions. Contents of the transcripts were the same except for sections relating to the experimental manipulations.

The first experimental manipulation concerned the testimony of the eyewitness. In all transcripts, one eyewitness (Mr Barns) testified for the prosecution, claiming that he witnessed the accident as it happened and the defendant’s driving behaviour at that time. The eyewitness recorded what he saw in his diary immediately after the accident (contemporaneous note) and read aloud this note on the stand. The law in New South Wales, Australia permits witnesses to read aloud their contemporaneous notes in trial to refresh their memories (ss 32 and 34 Evidence Act, 1995). After reading out his note, the eyewitness was then subjected to direct and cross-examinations. In the consistent eyewitness condition, the eyewitness’s account was consistent upon further questioning. That is, the content of his in-trial statement exactly matched what he initially stated in his contemporaneous note. On both occasions he claimed that he witnessed the defendant using her phone whilst the accident happened. Alternatively, in the reminiscent condition, the eyewitness added a reminiscent detail when being questioned in trial that was not recorded in his contemporaneous note. That is, he claimed that he witnessed the defendant on her phone only when being questioned in trial, and this detail was not initially recorded in his contemporaneous note.

The second experimental manipulation was the presence of expert testimony. In the two expert conditions, Professor Kennedy, an expert on memory, testified for the prosecution. The expert stated their credentials, gave testimony on the memory benefits of contemporaneous notes (Horry et al., 2021) and gave evidence that reminiscence was not an indicator of memory inaccuracy (Oeberst, 2012). The two expert conditions differed in when the expert testified during the trial. In the ‘expert present before’ condition, the expert took the stand right before the eyewitness testified, while in the ‘expert present after’ condition, the expert testified right after the eyewitness. In the expert absent condition, no expert witness appeared throughout the entire trial.

Measures

Verdict and guilt rating

For ecological validity, all participants were asked to indicate whether they believed the defendant was guilty or not guilty. Additionally, for a more sensitive measure, all participants were also asked to rate the defendant’s likelihood of guilt on a 7-point Likert scale item (1 = not guilty, 7 = guilty).

Perceived eyewitness credibility

All participants were asked to rate the credibility of the eyewitness on nine questions using a 7-point Likert scale (1 = not at all, 7 = completely). The questions were adapted from Connolly et al. (2008), who based their scale on Ross et al.’s (2003) two-factor model of credibility. This scale has been used in multiple studies assessing perceived eyewitness credibility (e.g. Deck & Paterson, 2021, 2022). The nine questions assessed perceived eyewitness honesty (through perceived truthfulness, honesty and likelihood of fabrication), perceived eyewitness cognitive competence (through perceived eyewitness intelligence, memory accuracy and understanding of the event), and overall eyewitness credibility (through perceived credibility, likeability and believability of the eyewitness). The item measuring the perceived likelihood of eyewitness fabricating their testimony was reversed coded. The nine items were shown to exhibit high internal consistency both in previous research (Deck & Paterson, 2021) and in the current study (Cronbach’s α = .895), therefore they were averaged to form a single score representing perceived eyewitness credibility (Cooper et al., 2014; Nunnally & Bernstein, 1994).

Judgements of the credibility of contemporaneous note

All participants were asked to rate the extent to which they believed the eyewitness’s contemporaneous note to be credible using a 7-point Likert scale (1 = not at all credible, 7 = completely credible).

Judgements of the credibility of in-trial statements

All participants were asked to rate the extent to which they believed the eyewitness’s in-trial statement to be credible on a 7-point Likert scale (1 = not at all credible, 7 = completely credible).

Belief about the memory benefits of contemporaneous notes

All participants were asked to rate the extent to which they believed the eyewitness’s contemporaneous note helped to improve his memory about the car accident on a 7-point Likert scale (1 = not at all beneficial, 7 = completely beneficial).

Belief about the accuracy of reminiscent items

Only participants in the reminiscent conditions were asked to rate the extent to which they believed the eyewitness’s reminiscence was accurate on a 7-point Likert scale (1 = not at all accurate, 7 = completely accurate).

Attention check

To ensure participants were paying attention while answering the questions, one attention check question was included in the questionnaire. Participants were asked to select ‘not credible at all’ as the correct response to the question. Participants who failed this attention check were excluded from the analyses.

Manipulation checks

To ensure that participants were paying attention while reading the trial transcript, they were asked three questions concerning the main manipulations. The questions were: (a) In what format did Mr. Barns (the eyewitness) record what he saw immediately after the car crash? (b) Did Mr. Barns mention any additional details in trial that were not included in his contemporaneous note? (c) Did an expert on memory testify during the trial? Participants who did not exhibit correct knowledge about the above questions were excluded from the final analyses, as this signalled a fundamental misunderstanding about the critical experimental manipulations.

Demographic questionnaire

Participants were asked to answer demographic questions about their age, gender and cultural background.

Procedure

All participants were provided a link to complete the study online. After giving informed consent, they were instructed to read the trial transcript while imagining themselves as a juror sitting through the trial. Participants were informed that they would later be asked questions about the trial. They were told that they could take notes while reading as real jurors are allowed to do so in NSW (NSW Law Reform Commission, 2012); however, they would not be able to refer back to the full transcript once they finished reading. After reading the transcript, participants were asked to complete the measures. Upon completion of the study, participants were fully debriefed.

Results

Main analysis

Dichotomous verdict decision

A total of 76 of the participants gave guilty verdicts, while the remaining 146 gave non-guilty verdicts (see Table 1 for dichotomous verdict decisions across conditions). A binary logistic regression analysis was conducted to examine the probability of conviction for mock-jurors. The probability of conviction was predicted through five categorical predictors: witness type (consistent vs. reminiscent), Expert Contrast Term 1 (expert presence: absent vs. present), Expert Contrast Term 2 (the timing of expert: before vs. after), the interaction between expert presence and witness type, and the interaction between timing and witness type. Witness type was dummy coded using SPSS’s Indicator contrast, with ‘consistent’ as the reference category. Expert condition (three levels: absent, present before, present after) was represented using two orthogonal custom contrasts: Contrast 1 compared expert absent (1) to the average of expert-present conditions (−.5, −.5); Contrast 2 compared expert present before (1) to present after (−1), with expert absent coded as 0.

Table 1.

Mock jurors’ verdict decisions across conditions.

Expert testimony	Consistent eyewitness	Reminiscent eyewitness
Absent	10 (23)	8 (28)
Present before	13 (23)	13 (24)
Present after	19 (20)	13 (28)

Note: Frequency of a guilty verdict for each condition is presented outside the parentheses. Frequency of a non-guilty verdict for each condition is presented in parentheses.

The omnibus model for the logistic regression was not statistically significant, ${\chi }^2\left(5,N=222\right)=6.355,p=.273\text{,}$ Cox and Snell $R^2=.028\text{,}$ Nagelkerke $R^2=.039.$ The model was able to predict mock-jurors’ verdict decision 65.8% of the time. The result of a Hosmer and Lemeshow test suggested that the model was a good fit for the data, ${\chi }^2\left(4,N=222\right)=0.000, p=1.000\text{.}$ See Table 2 for the coefficients for all predictors in the model. As reported in Table 2, none of the predictors significantly predicted the mock-jurors’ verdict, all ps > .176.

Table 2.

Predictor coefficients for logistic regression predicting dichotomous verdict.

	b	SE (b)	p	Exp(B) [95% CI]
Constant	0.878	0.209	<.001	2.405
Witness type	−0.393	0.290	.176	0.675[0.382, 1.193]
Expert Contrast 1: absent vs. present	0.563	0.467	.229	1.755[0.702, 4.388]
Expert Contrast 2: present before vs. present after	−0.154	0.481	.749	0.857[0.334, 2.200]
Interaction Contrast 1: Expert Contrast 1 × Witness Type	−0.041	0.646	.950	0.960[0.271, 3.408]
Interaction Contrast 2: Expert Contrast 2 × Witness Type	0.673	0.674	.318	1.961[0.523, 7.348]

Note: CI = confidence interval. Witness type was coded using SPSS Indicator contrast with ‘consistent’ as the reference group. Expert condition (absent, present-before, present-after) was contrast-coded using two custom orthogonal contrasts: Contrast 1 compared expert absent (1) with the average of the two expert-present conditions (−.5, −.5), and Contrast 2 compared expert present before (1) with expert present after (−1), with expert absent coded as 0.

Guilt rating

A 2 × 3 ANOVA was conducted to investigate the effects of eyewitness type (consistent, reminiscent) and expert testimony (absent, present before eyewitness, present after) on guilt ratings. Means and standard deviations of guilt rating are reported in Table 3. Planned and orthogonal contrasts with a decision-wise error rate (α = .05) were conducted to address our main hypotheses. However, no statistically significant results were found, all ps > .06. See Figure 1 for a visual illustration of mock-jurors’ guilt rating across conditions.

Table 3.

Means and standard deviations of guilt rating.

Expert testimony	Consistent eyewitness	Reminiscent eyewitness
Absent	4.15 (1.30)	3.56 (1.21)
Present before	3.92 (1.40)	4.16 (1.26)
Present after	4.44 (1.50)	3.85 (1.30)

Note: Standard deviations are presented in parentheses.

Figure 1.

Mock-jurors’ guilt rating. Note: Mean perceptions of the defendant’s guilt are shown for mock-jurors who were exposed to either a consistent or a reminiscent eyewitness. Mock-jurors did not hear any expert testimony, heard the expert testimony before the eyewitness testifying, or heard the expert testimony after the eyewitness testifying. Error bars show ±1 SEM.

Perceived eyewitness credibility

The same 2 × 3 ANOVA and contrasts were conducted to examine the impact of eyewitness type and expert testimony on mock-jurors’ perceptions towards an eyewitness, using perceived eyewitness credibility as the dependent variable. Averaged over different expert manipulations, a reminiscent eyewitness (M = 4.43) was perceived by mock-jurors to be significantly less credible than a consistent eyewitness (M = 4.77), F(1, 216) = 8.009, $p<.01\text{,}$ ${\mathrm{\eta }}_p^2=.036\text{.}$ However, this main effect is qualified by a significant interaction between witness type and expert testimony, F(1, 216) = 3.991, $p<.05\text{,}$ ${\mathrm{\eta }}_p^2=.018$ (see Figure 2). To further examine this significant interaction, a simple main effect contrast analysis was used to examine the effect of expert presence on mock-jurors’ perceptions for the reminiscent and consistent eyewitnesses separately. It was found that the presence of expert testimony (averaged across timing) was significant in increasing mock-jurors’ perceived credibility for a reminiscent eyewitness, F(1, 216) = 14.373, $p<.001\text{,}$ ${\mathrm{\eta }}_p^2=.062$, but not for a consistent eyewitness, F(1, 216) = 0.777, $p=.379\text{,}$ ${\mathrm{\eta }}_p^2=.004\text{.}$

Figure 2.

Mock-jurors’ perceived eyewitness credibility. Note: Mean perceptions of the credibility of the consistent or reminiscent eyewitness are shown for mock-jurors who did not hear any expert testimony, heard the expert testimony before the eyewitness testifying, or heard the expert testimony after the eyewitness testifying. Error bars show ±1 SEM.

Additionally, averaged over different witness types, mock-jurors who heard the expert testimony (M = 4.74) perceived the eyewitness to be significantly more credible than did those who did not hear it (M = 4.32), F(1, 216) = 10.671, $p<.01\text{,}$ ${\mathrm{\eta }}_p^2=.047\text{.}$ However, there was no significant difference in perceived eyewitness credibility between those who heard the expert before hearing the eyewitness (M = 4.73) and those who heard the expert after the eyewitness (M = 4.75), F(1, 216) = 0.013, $p=.909\text{,}$ ${\mathrm{\eta }}_p^2<.001\text{.}$ Furthermore, there was no significant interaction between witness type and the timing of expert testimony, F(1, 216) = 0.042, $p=.838\text{,}$ ${\mathrm{\eta }}_p^2<.001\text{.}$

Supplementary analyses

To further understand the effect of expert testimony on mock-jurors’ perceptions of the eyewitness, additional supplementary analyses were conducted, as described below.

Judgements of the credibility of contemporaneous note

A 2 × 3 ANOVA and contrasts were run to investigate the impact of eyewitness type and expert testimony on mock-jurors’ perceptions of the eyewitness’s contemporaneous note. Means and standard deviations are shown in Table 4. Averaged over different expert conditions, mock-jurors in the reminiscent conditions (M = 5.22) showed significantly higher belief in the credibility of the eyewitness’s contemporaneous note than those in the consistent conditions (M = 4.58), F(1, 216) = 13.342, $p<.001\text{,}$ ${\mathrm{\eta }}_p^2=.058\text{.}$ No other statistically significant results were observed, all ps > .08.

Table 4.

Means and standard deviations of perceived credibility of contemporaneous note.

Expert testimony	Consistent eyewitness	Reminiscent eyewitness
Absent	4.42 (1.20)	4.97 (1.54)
Present before	4.53 (1.18)	5.11 (1.27)
Present after	4.79 (1.45)	5.59 (1.12)

Note: Standard deviations are presented in parentheses.

Judgements of the credibility of in-trial statements

A 2 × 3 ANOVA and contrasts were conducted to investigate the impact of eyewitness type and expert testimony on mock-jurors’ perceptions of the eyewitness’s in-trial statement. Averaged across different expert manipulations, the in-trial statements of a reminiscent eyewitness (M = 3.94) were rated by mock-jurors to be significantly less credible than the statements of a consistent eyewitness (M = 4.60), F(1, 216) = 16.901, $p<.001\text{,}$ ${\mathrm{\eta }}_p^2=.073\text{.}$ However, there was a significant interaction between witness type and expert presence, F(1, 216) = 6.384, $p<.05\text{,}$ ${\mathrm{\eta }}_p^2=.029\text{.}$ Given this, a simple main effect contrast analysis was conducted to understand the effect of expert presence at each level of the witness type (reminiscent, consistent). The results suggested that the presence of expert testimony (averaged across timing) significantly increased mock-jurors’ belief in the credibility of the reminiscent eyewitness’s in-trial statement, F(1, 216) = 11.813, $p<.001\text{,}$ ${\mathrm{\eta }}_p^2=.052\text{;}$ but not their belief in the consistent eyewitness’s statement, F(1, 216) = 0.038, $p=.846\text{,}$ ${\mathrm{\eta }}_p^2<.001.$ A graphical illustration of the interaction was depicted in Figure 3.

Figure 3.

Mock-jurors’ judgements of the credibility of in-trial statements. Note: Mean judgements of the credibility of eyewitness’s (consistent or reminiscent) in-trial statement are shown for mock-jurors who did not hear any expert testimony, heard the expert testimony before the eyewitness testifying, or heard the expert testimony after the eyewitness testifying. Error bars show ±1 SEM.

Additionally, averaged across witness type, when there was an expert present (M = 4.40), mock-jurors on average exhibited significantly higher belief in the eyewitness’s in-trial statement than when there was no expert present (M = 4.01), F(1, 216) = 5.044, $p<.05\text{,}$ ${\mathrm{\eta }}_p^2=.023.$ However, there was no significant difference between the belief of mock-jurors who heard the expert before the eyewitness (M = 4.29) and that of those who heard the expert after the eyewitness (M = 4.51), F(1, 216) = 1.309, $p=.254\text{,}$ ${\mathrm{\eta }}_p^2=.006.$ This lack of difference again did not differ significantly as a function of witness type, F(1, 216) = 1.923, $p=.167\text{,}$ ${\mathrm{\eta }}_p^2=.009.$

Belief about the memory benefits of contemporaneous notes

A 2 × 3 ANOVA and contrasts analyses were conducted to assess effects of eyewitness type and expert testimony on participants’ belief about the memory benefits of contemporaneous notes. Means and standard deviations are reported in Table 5. Mock-jurors who were exposed to a consistent eyewitness (M = 5.36) reported significantly higher belief in the memory benefits of contemporaneous notes than those who were exposed to a reminiscent eyewitness (M = 4.99), F(1, 216) = 4.985, $p<.05\text{,}$ ${\mathrm{\eta }}_p^2=.023.$ No other results were statistically significant.

Table 5.

Means and standard deviations of belief in the benefits of notes.

Expert testimony	Consistent eyewitness	Reminiscent eyewitness
Absent	5.15 (1.12)	4.86 (1.25)
Present before	5.25 (1.00)	5.08 (1.30)
Present after	5.67 (1.24)	5.05 (1.22)

Note: Standard deviations are presented in parentheses.

Belief about the accuracy of reminiscent items

For the reminiscent condition only, a one-way ANOVA was conducted to investigate the effects of expert testimony (absent, present before, present after) on participants’ belief about the accuracy of reminiscent items. Means and standard deviations are reported in Table 6. There was a significant effect of expert testimony on mock-jurors’ beliefs about the accuracy of reminiscence, F(2, 110) = 3.636, $p<.05\text{,}$ ${\mathrm{\eta }}_p^2=.062.$ Contrast analyses revealed that mock-jurors who heard an expert testimony (averaged across timing) rated their belief (M = 3.84) in the accuracy of reminiscent items to be significantly higher than those who did not hear an expert testimony (M = 3.17), F(1, 110) = 6.990, $p<.01\text{,}$ ${\mathrm{\eta }}_p^2=.060.$ However, the mean belief did not differ between those who heard the expert before the eyewitness (M = 3.78) and those who heard the expert after the eyewitness (M = 3.90), F(1, 110) = 0.187, $p=.666\text{,}$ ${\mathrm{\eta }}_p^2=.002.$

Table 6.

Means and standard deviations of belief in the accuracy of reminiscence.

Expert testimony	Reminiscent eyewitness
Absent	3.17 (1.23)
Present before	3.78 (1.33)
Present after	3.90 (1.22)

Note: Standard deviations are presented in parentheses.

Discussion

The current study was the first to examine the effect of testimonial reminiscence and expert testimony on mock-jurors’ decision-making, specifically in the context of contemporaneous notes. Contrary to hypotheses, none of our experimental manipulations significantly influenced mock-jurors’ perceptions of the defendant’s guilt and their verdict. However, in accordance with previous research, we found evidence demonstrating the detrimental effect that testimonial reminiscence can have upon mock-jurors’ ratings of eyewitness credibility. Furthermore, our findings suggested that expert testimony could be a promising tool to facilitate informed juror decision-making, as it improved their perceptions of the credibility of a reminiscent eyewitness. Our results have the potential to inform legal practice and contribute to fair judicial outcomes. All main findings and their implications are discussed in turn.

Jurors’ misconception about testimonial reminiscence

As hypothesised (H1), we found that a reminiscent eyewitness was perceived by mock-jurors to be significantly less credible than a consistent eyewitness. This finding replicates the findings of Deck and Paterson (2021) and extends the findings of Berman and Cutler (1996) by showing that reminiscences that arise between an eyewitness’s contemporaneous note and in-trial statement can also adversely affect jurors’ judgements of the eyewitness. Our finding adds to evidence suggesting that jurors do not always make decisions that are consistent with scientific findings (e.g. Wise & Safer, 2004). Since jurors who lack the necessary knowledge about eyewitness memory could be swayed by their false intuitions, our results suggest the need to investigate feasible and effective mechanisms to foster jurors’ understanding of eyewitness memory (Shermer et al., 2011). Interestingly, although reminiscent eyewitnesses were perceived as less credible overall, their contemporaneous notes were rated as more credible than those of consistent eyewitnesses. This finding suggests that mock-jurors may place particular trust in initial, contemporaneously recorded accounts. In contrast, subsequent discrepancies – such as those found in reminiscent testimony – may be met with increased scepticism, potentially due to concerns about memory distortion or fabrication. This distinction highlights jurors’ nuanced evaluation of memory evidence and underscores the evidentiary weight they attribute to contemporaneous documentation, even when subsequent testimony raises credibility concerns.

The beneficial effect of expert testimony

As predicted (H2), we found that mock-jurors rated the credibility of eyewitnesses more highly when they heard an expert testimony about eyewitness memory than when there was no expert. This finding was consistent with the results of previous studies, which have revealed the beneficial effect that a witness-affirming expert can have upon mock-jurors’ perceptions of the witness (e.g. Deck & Paterson, 2022; Goodman-Delahunty et al., 2021). We also found a significant interaction between eyewitness type and expert presence (H3), such that the mock-jurors’ lower credibility ratings of a reminiscent eyewitness compared to a consistent eyewitness were significantly less pronounced among those who heard an expert than those who did not hear one. This significant interaction was driven by an increase in the mock-jurors’ credibility ratings of the reminiscent eyewitness when there was an expert present. Namely, although expert testimony improved mock-jurors’ mean ratings of eyewitness credibility (averaging together that of the reminiscent and consistent eyewitness), when examining its effect on the two types of eyewitnesses respectively, expert testimony was effective in enhancing mock-jurors’ perceptions of the reminiscent eyewitnesses but had no effect in improving their perceptions of the consistent eyewitness. Why was the expert only successful in affirming one but not the other?

The results of our supplementary analyses aided our understanding of this unexpected main finding. In our supplementary analyses, we examined the experts’ influence on jurors’ perceptions of the eyewitness’s contemporaneous note and in-trial statements separately. We also assessed mock-jurors’ explicit understandings of reminiscence and the memory-enhancing benefits of contemporaneous notes. We found that, averaging across the witness type, expert testimony successfully increased mock-jurors’ beliefs in the eyewitness’s in-trial statements, but not their beliefs in the contemporaneous note. Additionally, the expert only increased mock-jurors’ beliefs in a reminiscent eyewitness’s in-trial statements, but had no impact on their beliefs in a consistent eyewitness’s statements. This was presumably because the expert was capable of enhancing jurors’ conscious beliefs in the accuracy of reminiscent items but was unable to heighten their beliefs in the memory benefits of contemporaneous notes. Therefore, mock-jurors’ increased trust in a reminiscent in-trial statement seemed to stem from their fundamental increased understanding of the science behind reminiscence, which, consequently, led to their elevated perceptions of the reminiscent eyewitness. On the other hand, the expert failed to assert the memory-enhancing benefits of contemporaneous notes and hence was unable to increase jurors’ beliefs in a consistent in-trial statement and their perceptions of a consistent eyewitness. This finding is novel given our study is the first to demonstrate the effect of expert testimony in ameliorating jurors’ misconception about reminiscence. It is also unanticipated, as we expected our expert testimony (which contained research findings about both contemporaneous notes and reminiscence) to concurrently educate mock-jurors about the benefits of contemporaneous notes and mitigate their misunderstandings about reminiscence. However, it was only efficacious for the latter but not the former.

The most appropriate explanation might be that mock-jurors may already have some understanding about the memory benefits of contemporaneous notes. Although literature indicates apparent deficits in jurors’ eyewitness knowledge (e.g. Benton et al., 2006), some research shows that the media portrayals of factors influencing eyewitness memory have dramatically increased over the past two decades (Desmarais et al., 2008); as laypeople are having more exposure to eyewitness issues, their understanding of them may have been enriched. Indeed, a meta-analysis presents some evidence that the public’s knowledge of eyewitness issues has improved over time (Desmarais & Read, 2011). Specifically, evidence suggests that laypeople nowadays commonly do understand that memory for an incident is the most optimal immediately after the event and subsequently decays over time (e.g. Oeberst, 2012, 2015). Given the improvement in people’s eyewitness knowledge in general, jurors today may be able to intuitively understand the memory benefits of contemporaneous notes without the help of an expert. Being aware of the deterioration of memory over time, they could independently apply this knowledge when evaluating the credibility of an eyewitness who composed a contemporaneous note and is consistent with it at trial. Hence it is plausible that mock-jurors’ beliefs about a consistent eyewitness and their contemporaneous note are already favourable to start with, so the presence of an expert cannot further improve their perceptions. Conversely, jurors are not familiar with reminiscence; they consistently mistake it as indicative of memory inaccuracy (e.g. Berman & Cutler, 1996) and substantially underestimate the accuracy of reminiscent items (Oeberst, 2012, 2015). As some argue (e.g. Cutler et al., 1989; Greene & Gordon, 2016), expert testimony pertaining to issues that jurors are already familiar with has limited impact compared to when it is on areas that jurors understand less. Since mock-jurors possess little understanding about reminiscence, it makes sense that an expert testimony would be valuable in educating them and influencing their decisions.

Null effects for the timing of expert testimony

The timing of expert testimony had no significant impact on any of the dependent variables of interest (H4a, H4b). Specifically, there was no significant difference between the perceived eyewitness credibility, guilt rating and verdict of mock-jurors who heard the expert testimony before the eyewitness and of those who heard the expert testimony after the eyewitness. There was also no interaction between the timing of the expert testimony and the eyewitness type (reminiscent, consistent; H5a, H5b). Both approaches (expert before vs. after) had theoretical and empirical justifications, thus making it difficult to predict which would prevail under the current circumstance. Previous studies that have found support for one approach over the other (Brekke & Borgida, 1998; Leippe et al., 2004; Schuller & Cripps, 1998) differ drastically from the present study in terms of the methodologies employed and the purpose of the expert testimony, hence it is not surprising that our findings are not in accordance with what they have found.

A potential explanation for the null effects is that our experiment might not have created enough temporal distinction between the two expert testimony conditions. The recency benefits associated with post-eyewitness expert testimony are more likely to surface when the trial is long and complex (Horowitz et al., 2001). However, due to practical constraints, the entire trial transcript was purposefully restricted to an approximate 15–20-minute reading time, which does not reflect the length of an actual trial. Furthermore, there were no additional testimonies introduced between the eyewitness and the expert in our fictitious trial, while, in reality, other forms of evidence would often be presented. Consequently, the contrast between an expert testifying before or after the eyewitness in our experiment was minimal. Plausibly, if a significant difference between the effect of pre-eyewitness expert testimony and post-eyewitness expert testimony does exist in real trials, it might not have manifested under our experimental manipulation. Future research is advised to employ more realistic trials to investigate the potential difference.

The lack of effect upon guilt perceptions

We hypothesised that mock-jurors, if exposed to a reminiscent prosecution eyewitness, would perceive them to be less credible than a consistent eyewitness, and would subsequently rate the defendant to be less culpable. However, the guilt ratings of those who saw a reminiscent eyewitness did not differ from those who saw a consistent eyewitness, despite a significant difference between their perceptions of witness credibility. This result is in accordance with the findings of Deck and Paterson (2021), but inconsistent with Berman and Cutler (1996) who found that mock-jurors tend to perceive the defendant as less guilty and are less likely to convict them when exposed to a reminiscent prosecution eyewitness. Moreover, expert testimony affirming the prosecution eyewitness did not lead to increased perceptions of defendant guilt among mock-jurors, again contradicting our expectations. This result contradicts previous studies that have revealed the effect of expert testimony in influencing mock-jurors’ perception of guilt (e.g. Goodman-Delahunty et al., 2021; Stobbs & Kebbell, 2003). However, other studies also fail to establish significant effects of expert testimony upon guilt ratings (e.g. Ryan & Westera, 2018). Additionally, the interaction between eyewitness types and expert testimony did not have a noticeable impact on mock-jurors’ guilt ratings.

The absence of effect may be due to the overall prosecution evidence in our trial stimulus relative to the strict legal criteria for conviction. The Australian criminal justice system requires jurors to only convict if they are convinced that the defendant is guilty beyond a reasonable doubt. The prosecution thus carries a very high burden of proof, and the strength of their evidence is strongly correlated with convictions (Devine et al., 2016). On the other hand, jurors tend to exhibit a leniency bias (i.e. they tend to acquit the defendant) when they are not fully certain about the defendant’s guilt (MacCoun & Kerr, 1988). In our trial stimulus, we deliberately did not include incontestable evidence (e.g. video footage) in order to make the eyewitness testimony the focal point. Our manipulations of eyewitness evidence successfully influenced jurors’ perceptions of the eyewitness; however, it could be that the overall prosecution evidence was not strong enough to eliminate all reasonable doubts and satisfy the high standard for conviction.

Additionally, some literature indicates that mock-jurors’ perceptions of witness credibility are more likely to influence their guilt ratings when there is little additional case evidence (other evidence unrelated to the witness) to consider (Voogt et al., 2020). Unlike some studies (e.g. Stobbs & Kebbell, 2003, which observed changes in both participants’ guilt ratings and perceptions of witness credibility) that utilised vignettes containing only the experimentally manipulated parts of the trial, our study employed a more ecologically valid transcript depicting the entire trial with different forms of evidence favouring either the prosecution or the defence (e.g. environmental conditions, testimonies from other witnesses etc). Thus, mock-jurors’ perceptions of defendant guilt may not strictly align with their perception of eyewitness credibility, as other evidence (besides the experimentally manipulated evidence concerning eyewitness memory) may also contribute to their final verdict.

Limitations and future directions

A few limitations of this study must be addressed. Firstly, like most jury stimulation experiments, our study employed a convenient undergraduate student sample. While student samples are commonly used in jury research to examine the psychological mechanisms underlying juror decision-making (e.g. Bornstein, 1999; Bornstein et al., 2017), we recognise this limits the generalisability of the findings. Hence, replication with a more diverse, representative sample would ensure that our findings are generalisable. In addition, participants were recruited through a university research participation pool rather than a source analogous to official juror summons procedures. Secondly, unlike real trials, there was no group deliberation involved. Future research should compare jurors’ pre-deliberation and post-deliberation decisions to further our understanding of the impact of reminiscence and expert testimony on different stages of jury decision-making. Third, our use of a written trial transcript decreased the ecological validity of our study because it lacked certain visual and audio cues, such as a speaker’s tone, confidence and appearance, which could potentially influence the jurors’ judgements (e.g. Cutler et al., 1988). Furthermore, the trial transcript used in this study was fictitious, though informed by prior mock-juror research and constructed to reflect plausible legal scenarios. However, we acknowledge that it was not based directly on real case materials nor reviewed by legal professionals, which may affect how well the evidence mirrors what would be presented in an actual courtroom. Future research hence should consider inclusion of genuine case materials or trained legal professionals to ensure that evidence presented to mock-jurors would closely reflect the quality of evidence admissible in the courtroom (Wilmott et al., 2021). Lastly, although efforts were made to simulate a realistic courtroom environment through detailed trial materials, we acknowledge that the trial process was simplified and lacked the full procedural and emotional weight of a genuine trial. Future research could consider replicating our findings with field studies involving real judges and juries making decisions under a naturalistic setting. However, field studies come with a myriad of ethical, legal and logistical considerations, making them extremely difficult to conduct, and are very rare in the area (Bornstein, 2017).

Implications and conclusions

In recent years, witnesses have been advised to write a contemporaneous note to consolidate and enhance their memory of a critical incident (e.g. Paterson et al., 2018). However, our research suggests that such notes may inadvertently discredit an eyewitness who omits key details in the note but recalls them on a later occasion. It appears that jurors do not understand reminiscence, and they may falsely equate this common phenomenon with low eyewitness credibility, despite scientific evidence suggesting otherwise (e.g. Oeberst, 2012, 2015). However, promisingly, our research suggests that expert testimony (regardless of when it is presented in a trial) can increase jurors’ knowledge about eyewitness reminiscence. Notably, we found that participants who encountered expert testimony rated reminiscent witnesses as more credible than those who did not. This suggests that it is possible to use expert testimony on eyewitness memory to enhance jurors’ reliance on scientific knowledge when making decisions, consistent with past research (e.g. Cutler et al., 1989). Thus, expert testimony may be used as a corrective precaution against misconceptions about reminiscence, and should be considered by judges and attorneys when there are eyewitnesses who need to testify multiple times over the case.

Ethical standards

Declaration of conflicts of interest

Shuyu Lin has declared no conflicts of interest.

Helen M. Paterson has declared no conflicts of interest.

Ethical approval

All procedures performed were in accordance with the ethical standards of the University of Sydney’s Human Research Ethics Committee [2022/245] and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.

Informed consent

Informed consent was obtained from all individual participants included in the study.

Data availability statement

The data that support the findings of this study are openly available on the Open Science Framework at: https://osf.io/g5hr3/?view_only=ee534cc007fc44abad170cfe4c19b7a7