1. Introduction

Decision-making plays a crucial role in everyday life, encompassing trivial choices such as selecting a route to school and significant ones like career or marital decisions that can sometimes profoundly impact others. Within the legal context, juror decision-making (JDM) is particularly significant since determining the innocence or guilt of a suspect has major implications on their future. An alibi, defined as evidence proving a suspect’s presence elsewhere during a crime [1], is vital in this process. As noted by Levine and Miller [2], whether or not it is accurate, a strong alibi offers potentially conclusive evidence regarding the suspect’s ability to commit the crime, greatly impacting JDM and increasing the acquittal rate in jury verdicts. The evaluation of alibis within the framework of JDM is essential for ensuring justice. The importance of research in this area in preventing wrongful convictions or acquittals propelled the initiation of the present study.

While physical alibis like videotapes and receipts generally provide clear and reliable evidence, human alibis, particularly eyewitness testimony, are often less accurate [3]. In the context of criminal investigations, the discovery of inconsistencies or deliberately misleading information can discredit an alibi, rendering it ineffective in substantiating a suspect’s whereabouts at the time of the crime. Loftus [4] had participants read summaries of criminal trials featuring either reliable or discredited prosecution eyewitnesses and found that discrediting an eyewitness had a negligible impact on the subjects’ likelihood to convict, demonstrating a tendency among jurors to place undue trust in eyewitness testimony, regardless of its credibility. However, subsequent research has consistently observed the discrediting effect, which occurs when discredited eyewitness evidence results in a higher frequency of innocent verdicts compared to cases with no eyewitness at all. Weinberg and Baron [5] showed to 156 undergraduate students criminal trial scenarios containing three types of eyewitnesses: a trustworthy eyewitness, a discredited eyewitness, and no eyewitness. Findings inferred the propensity to ignore eyewitness evidence after it was discredited, irrespective of the discreditation’s nature or target, indicating that eyewitness accounts had a diminished effect on verdicts.

Similarly, Kennedy and Haygood [6] presented summaries of a robbery trial to 147 university students, involving the same conditions as Weinberg and Baron [5]. Their findings revealed a significant discrediting effect, where the presence of a discredited eyewitness led to the lowest guilt ratings, the presence of no eyewitness led to higher guilt ratings, and the presence of credible eyewitnesses led to the highest guilt ratings. This underscores the detrimental impact that discrediting an eyewitness has on their credibility and their influence on JDM. One perspective posits that eyewitness evidence should be excluded from consideration by the jury if it is proven false. However, this strategy might inadvertently backfire on the prosecution, as jurors could view this as an attempt to mislead the jury, leading to a distrust of the prosecution’s evidence. This might cause jurors to sympathize with the defendant and even persuade them to hold a belief in the defendant’s innocence. Given the reaction to discredited eyewitnesses, it stands to reason that jurors might similarly respond to discredited alibis by suspecting deliberate deception and a concealment of truth by the suspect, potentially leading to a “backfire effect”. In this scenario, jurors could perceive the defendant as guiltier with a discredited alibi than with no alibi, indicating an expectation for discredited alibis to increase the likelihood of guilty verdicts, mirroring the effect seen with discredited eyewitness testimony.

The phenomenon of the discrediting effect on alibis has prompted research into its influential factors. Olson and Wells [3] observed that alibis supported by physical evidence (e.g., ticket stubs) were more convincing than those backed by personal testimony (e.g., from a boyfriend). Culhane and Hosch [7] noted the credibility of alibis from impartial individuals (e.g., a neighbour) over those from someone with a personal connection to the defendant (e.g., a wife); Lindsay . [8] further confirmed that non-relative alibi witnesses decrease conviction rates more than alibis from relatives.

Additionally, the impact of the alibi provider’s age on alibi credibility and JDM has gained attention due to the prevalence of child witnesses in legal proceedings [9]. Dahl and Price [10] conducted a mock police investigation with alibi witnesses varying by age (a 6-year-old child versus a 25-year-old adult) and relationship to the suspect (son or neighbour). Results indicated that alibis from children not only diminished perceived guilt more effectively, but also led to higher credibility ratings compared to alibisfrom adult witnesses. Similarly, Gabora et al. [11] had mock jurors watch a video of a simulated child sexual abuse trial and reach an unanimous verdict. The trial involved either a 13- or 17-year-old complainant. In this instance, the complainant delivered their own eyewitness testimony, accusing the suspect of the crime. The results emphasized that jurors were more likely to convict with the younger complainant, who was also considered more credible than the older one.

Past research has typically classified alibi providers by age as either children (under 18) or adults (over 18), without delving into how discredited alibis from children at varying developmental stages affect JDM, despite the known variation in cognitive abilities by age. It is notable that a key cognitive development relevant to deceit, the theory of mind (ToM), which involves understanding others’ mental states, is undeveloped in most 3-year-olds but is stable in those aged 4 years and above [12]. This developmental milestone is crucial in the capacity to lie, as it involves representing and manipulating the mental states of oneself and others. Proficiency in lying, as Talwar and Lee [13] suggest, evolves and improves throughout childhood. Additionally, Ross et al. [14] argued that assessments of child witness credibility are largely influenced by their perceived honesty, with children generally viewed as more honest than adults due to a presumed lack of motive or capacity to lie. Given this perspective, it is plausible to anticipate that an alibi from a 3-year-old would be considered more credible than one from an older child, such as a 15-year-old, due to assumptions about their inability to deliberately deceive and would lead to lower perceived guilt.

In summary, prior research has separately examined the effects of discrediting an alibi and the age of the alibi provider on juror perceptions. This study seeks to integrate these aspects by exploring how the age of a child giving a discredited alibi influences jurors’ perceptions of guilt in a theft scenario. This study addresses this question by presenting to the participants a fictional theft case with three subsequent scenarios: no alibi, a discredited alibi from a 3-year-old, and a discredited alibi from a 15-year-old. Participants were then asked to assess the suspect’s guilt. The “no alibi” scenario served as a control to establish a baseline for perceived guilt, enabling the analysis to focus on the impact of the alibi provider’s age on juror perceptions. This study proposed the hypothesis that the participant-assigned guilt predictions would vary by condition: lowest for the scenario with no alibi, higher when a discredited alibi is given by a 3-year-old, and the highest when provided by a 15-year-old.

2. Study methods and design

This study employed a 1x3 between-subjects design. The independent variable was the type of alibi, with 3 levels: no alibi (NA), discredited alibi provided by a 3-year-old (DA3), and discredited alibi provided by a 15-year-old (DA15). The dependent variable was the participants’ perceived probability of the suspect’s guilt, measured on a scale from 0-100, after being presented with one of the alibi types. The between-subjects design was chosen to avoid influencing participants’ perceptions of increased guilt, which might occur if they were sequentially presented with two discredited alibis.

2.1. Participants

Prior to enrolling participants, the necessary sample size for the current study’s design was estimated by a G*Power power analysis [15]. Assuming a medium effect size (Cohen’s d = 0.5), a 0.05 alpha level, and a 0.80 targeted power, a sample size of about 42 individuals was enough to ensure that the research had sufficient power. The study recruited 49 participants (Mage = 21.61, SD = 8.10) through opportunity sampling by sending the experiment link to the researchers’ connections. Consequently, a diverse group encompassing different genders, races, and ages was collated. This sample size surpassed the minimum threshold assessed by the study’s power analysis, confirming its adequacy for testing this study’s hypothesis with strong statistical validity. Participants with high English proficiency were selected for this study due to its reliance on specific legal terminology, including terms like 'alibi’ and 'discredited’, which are integral to understanding and evaluating suspect guilt perceptions, making linguistic competence essential for accurate interpretation and response.

The participants were randomly assigned to one of three conditions: no alibi, 3-year-old discredited alibi, or 15-year-old discredited alibi with 13, 19, and 17 participants respectively. Demographically, the sample included 33 females, 15 males, and 1 identifying as other, while the racial composition split was 31 Asian, 14 Caucasian/White, 1 Hispanic/Latino, and 3 others. This distribution indicated a potential bias in the sample towards female and Asian demographics. All participants provided informed consent before the experiment, which had received approval from the ethics committee.

2.2. Materials

Data collection was conducted online using the Gorilla Experiment Builder [16]. The experiment utilized textual descriptions of a theft crime and corresponding alibis across three conditions (NA, DA3, DA15) as stimuli. The text placement and font size of the background story was standardized across all conditions, which read: “One night, an antique vase was stolen from the local museum. The main suspect was Alex, a collector of rare items. The museum found the vase missing the morning after a special event. Security footage showed someone like Alex near the museum that night but it wasn’t clear. Alex’s exact location at the time of the theft was unknown, making the case a puzzle for the police”. This standardization aimed to maintain consistency and mitigate any response bias that might arise from variations in the presentation of the crime scenario. The narrative deliberately omits a definitive conclusion regarding Alex’s culpability, thereby creating an ambiguous context. This ambiguity is designed to encourage participants to rely solely on their personal judgment in assessing Alex’s guilt, which ensures that responses are based on the individual perceptions of the participants, thereby enhancing the fairness and objectivity of their assessments. The NA condition is established as a baseline for comparing the suspect’s perceived guilt across the groups, which stated: “When questioned, Alex claims to have been home alone on the night of the theft. No one, not even the neighbors, can confirm their whereabouts, making the police skeptical due to the lack of an alibi and the security footage.”

In the DA3 and DA15 conditions, the textual descriptions were kept identical, differing only in the name and age of the alibi provider. This ensured that any variance in participant responses is attributed to the age-related context of the alibi rather than textual differences, thus increasing the current study’s validity of measurement. In the DA3 condition, the description stated: “Alex states they were at home during the theft and were seen by Blake, their 3-year-old neighbor. When Blake is questioned, they initially confirmed seeing Alex at home but later show signs of uncertainty and contradicted their own statements. The police decided to discredit this alibi due to the lack of conclusive evidence.” The DA15 condition presented the same narrative, but the alibi was provided by Taylor, Alex’s 15-year-old neighbor.

The identity of the alibi provider in both conditions was purposefully set as the suspect’s neighbour due to their relative neutrality and low propensity to fabricate an alibi for the suspect. This decision was informed by the findings of Culhane and Hosch [7], who demonstrated that an alibi witness without a personal relationship to the suspect significantly impacts jury decisions. Their research indicated that the probability of conviction was reduced by half when the alibi was provided by a neighbour rather than a girlfriend, or in the absence of an alibi witness. Additionally, given the hypothesis of this study that the suspect’s perceived guilt would be higher in the conditions with discredited alibis compared to the baseline, the character of a neighbour was selected instead of any relatives to accentuate the increase in perceived guilt and make it easily observable in the study’s outcomes.

All stimuli were crafted to focus on clarity and simplicity, avoiding extraneous details irrelevant to the core investigation. The background was carefully constructed to avoid any definitive evidence about the suspect’s guilt, therefore facilitating straightforward comprehension of the scenario and allow participants to make fair and unbiased decisions. Gender-neutral names (Alex, Blake, Taylor) were deliberately selected for the suspect and alibi providers to circumvent stereotypes associated with certain genders being more predisposed to criminal behaviour, which could potentially influence participant judgments regarding the suspect’s guilt.

An attention check was presented following the background story and the alibi in each condition. The question asked, “In the story about the theft of the antique vase, what does Alex claim to have been doing on the night of the theft?” with the response options: “Attending a late-night party”, “Staying at home”, “Visiting a local museum”, and “Traveling out of town”. In all three conditions of the study, participants received consistent information stating that Alex, the suspect, claimed to have been staying at home. Consequently, the attention check was uniform across all participants. This check was designed to allow a standard measure of the participants’ engagement with the study’s details. Participants who selected the correct response, “Staying at home,” were considered to have passed the attention check and were thus included in the subsequent data analysis. This task enhanced the internal validity of the experiment since passing this check indicated the participants’ thorough understanding of the narrative.

The DV was quantified using a slider task (as illustrated in Figure 1) that scaled from 0 to 100, measuring the participants’ perceived guilt of the suspect (0 = perception of complete innocence; 100 = strong belief in the suspect’s guilt). The participants simply needed to click and drag the slider to indicate their predictions. This task was chosen based on its ease of use, straightforward nature for participant comprehension, and its ability to yield clear, numerical results that facilitate subsequent analysis and calculations.

Note. The participants simply needed to click and drag the slider to indicate their predictions.

2.3. Procedure

All participants reviewed an information sheet with a concise overview of the research and gave their informed consent, then provided their demographic information including age, gender, race, and level of English proficiency. Following this, participants were randomly assigned to either the NA, DA3, or DA15 condition and then presented the background story, the condition-specific alibi, the attention check, and the slider task. Finally, participants were debriefed, which included expressing gratitude for their participation, soliciting their perceptions about the study’s objectives, and addressing any queries they might have had. To ensure the participants’ thorough comprehension of the materials and considered decisions, no time limits were imposed on any section of the study.

3. Result

All participants passed the attention check; hence, no data was eliminated. As depicted in Figure 2, the data demonstrates an elevation in the mean probability of guilt in the DA3 condition (M = 50.37%, SD= 18.49) relative to the NA condition (M = 42.75%, SD = 19.69). This trend of increased perceived guilt probability is further amplified in the DA15 condition (M = 66.67%, SD = 25.84). Conversely, a comparison between the DA3 and DA15 conditions illustrates a marked reduction in the mean probability of the suspect’s guilt in the former (M = 50.37%, SD = 18.49) as opposed to the latter (M = 66.67%, SD = 25.84). Overall, participants perceived the suspect as least guilty in the no alibi condition, followed by the 3-year-old discredited alibi condition and the 15-year-old discredited alibi condition (means indicated as the crossbars in the box plots in Figure 2). The top and the bottom line of the boxes are the 75% and the 25% quantile of each condition’s data. As indicated by the violin plot, the perceived guilt in the DA15 condition exhibited a wider distribution compared to the other groups. This graph was plotted using R studio.

Levene’s test was employed to ascertain the homogeneity of variance across conditions. The non-significant results, F(2, 48) = .95, p = .39, satisfied the assumption of equal variances. Additionally, a non-significant Shapiro-Wilk test (W = .97, p = .23) suggested a normal distribution of the perceived guilt data. Given these preliminary findings, a 1x3 ANOVA was conducted to explore differences among the perceived guilt. Results revealed a significant effect, F(2, 48) = 5.49, p = .006, ges = .17, indicating notable disparities in perceived guilt among conditions.

Three pairwise comparisons were conducted to further dissect these variances. The comparison evaluating the perceived probability of guilt between the DA15 and the NA condition exhibited a significant difference, t(48) = 3.04, p = .01, indicating the participants perceived the suspect as more guilty in the DA15 compared to the NA condition. The comparison assessing the difference in perceived guilt probability between the DA15 and the DA3 condition was also significant, t(48) = 2.38, p = .01, implying the participants perceived the suspect as more guilty in the DA15 compared to the DA3 condition. However, the comparison between the DA3 and the NA condition was non-significant, t(48) = .93, p = 1.0, suggesting no substantial effect on the perceived probability of guilt in these conditions.

4. Discussion

Outcomes of the current study offered partial corroboration for the hypothesis, delineating a gradation in guilt assessment from participants: minimal for the NA condition, intermediate for the DA3 condition, and maximal for the DA15 condition. The analytical comparison underscored a heightened perception of guilt associated with the discredited alibi from the 15-year-old, relative to both the 3-year-old’s discredited alibi and the absence of an alibi. These outcomes are consistent with prior studies that observed a pronounced discrediting effect, and with jurors deeming alibis from younger providers as more credible (e.g. Dahl & Price [10], Kennedy & Haygood [6]). Consequently, this study’s findings enrich the existing discourse on the discrediting effect, contributing to a deeper comprehension of JDM in contexts of discredited alibis across different ages of witnesses.

The explanation for these findings can be seen through two lenses: the perceived honesty and the ability of the alibi provider to lie. Young children are often seen as more truthful, lacking the intent or ability to deceive, in contrast to older individuals who are viewed as having more cognitive skill in accurately observing and recounting events. Consequently, if older individuals offer a discredited alibi, they are suspected of intentionally fabricating their testimony to benefit the suspect, rather than merely misremembering or misreporting details [14]. The inability of 3-year-olds to fully develop a theory of mind suggests a perceived lack of capacity for deception [12]; Gabora et al. [11] posit that during childhood, while children are considered to possess accurate memories, they are considered too young to have motives for false accusations, reflecting an absence of intent to deceive in younger children. These understandings may lead participants to view a 15-year-old providing a discredited alibi as intentionally deceptive, resulting in higher guilt ratings for the suspect in scenarios involving older children.

The comparison between the DA3 and the NA condition did not reveal any significant differences, indicating that the presence of a discredited alibi from a 3-year-old does not substantially alter the perceived likelihood of guilt compared to cases where no alibi is presented. This observation is consistent with the findings of Pozzulo et al. [17], which, with a large sample of 231 participants, failed to identify any significant interaction between the age of the defendant and the corroboration of their alibi. Fawcett and Winstanley [18] explored how the age and confidence of child alibi witnesses impact mock jurors’ decision-making, testing witnesses aged 8, 12, and 16 years with varying levels of confidence. While there was a general bias towards acquittal, the study found no notable correlation between the witnesses’ age or confidence and the jurors' decisions. Nevertheless, witnesses who displayed confidence were judged to be more credible and reliable than their less confident counterparts. The lack of significant difference in guilt assessments between the NA and DA3 condition could be attributed to a reduced scepticism towards child witnesses, in contrast to the suspicion often directed at adult witnesses providing an alibi.

A significant factor contributing to the validity of this study was the selection of gender-neutral names (Alex, Blake, Taylor) for the suspect and alibi providers, aimed at avoiding gender-related stereotypes about criminal propensity. Nevertheless, it is possible that these names might imply a specific nationality or racial background, potentially affecting participants’ perceptions of the suspect’s guilt. Another limitation of this study arises from the demographic composition of the study’s participants, who were primarily female and of Asian descent. Gabora et al. [11] observed that female jurors are more likely to convict the defendant and view the complainant as more credible compared to male jurors. Thus, the predominance of female participants in this study may have skewed the guilt assessments toward a ceiling effect. Additionally, to mitigate the potential for anchoring bias, the study strategically positioned the starting point of the slider (utilized to measure perceived guilt) at the midpoint of the scale. This approach was intended to minimize the influence of initial value setting on participants’ responses, thereby more accurately capturing the impact of the alibi provider’s age on guilt judgments.

5. Conclusion

This study offers several implications for legal practice. Its insights are useful in appeal processes or case re-evaluations where convictions depend significantly on child witness testimony. The judicial system may need to examine the influence of age-related biases on the credibility assessments of witnesses and their impact on trial outcomes. Additionally, legal professionals could be trained on recognizing and counteracting age-related biases. Legal protocols could be revised to ensure that the credibility of witnesses is judged impartially, free from age-related prejudices. Moreover, understanding the nuances in alibis provided by children of different ages can enhance the legal representation of minors in court, promoting a more equitable treatment of alibis across various age groups. To enhance our understanding, future studies could adopt a refined methodological framework, such as a 2x3 factorial ANOVA, which would incorporate an intermediary condition wherein an alibi is initially presented without being discredited, before its inaccuracy is disclosed to participants. Evaluating the perceptions of guilt pre- and post-disclosure would enable a precise analysis of the impact that discrediting an alibi has on JDM, facilitating the isolation of the discreditation effect and thereby elucidating the nuanced psychological processes engaged in by jurors when confronted with evidence of varying credibility.