The accuracy, reliability of memory and credibility of the eyewitness testimonies (EWTs) of children is the focus of many heated debates. There are many issues concerning child EWT: children may be more affected by the interviewer and the techniques used to acquire the testimony and developmental differences between children from different age groups may affect their ability to resist suggestibility (Goodman, 2006) and thus, their statements may be deemed less reliable by jurors. However, if children are not taken seriously as witnesses then many criminals may remain unpunished, especially when a child is the only witness to a crime or when the child is the victim of the crime itself. In this essay I will discuss the conditions that affect the accuracy and, often more importantly, the perceived accuracy of child EWT. I will also investigate how research in EWT could inform real life interview practices to improve the reliability of child eyewitnesses. For the purposes of this essay I will define a child from the age of 3 up to early adolescence. First, I will explore the importance of accuracy, competency and credibility in EWT. Second, I will look at the factors and conditions that affect memory recall and testimonial accuracy in children. I will then discuss the interview techniques that have the ability to either enhance or hinder the reliability of testimony in children and finally, I will show how current research can introduce improvements to increase the dependability of child EWT.
From the perspective of justice the most important aspect of an EWT is its accuracy, but this is often hard to measure, especially if there is only one witness. A study by Brewer and co-workers (1999) investigated the factors that contribute to perceived testimony accuracy. They suggest that inconsistent statements of witnesses from one interview to the next lead potential jurors to believe that the testimony is inaccurate. The results of the study imply that when a testimony is inconsistent in some details that are related to specific dimensions of the testimony it does not affect the accuracy of other dimensions or the general memory reliability of that particular incident (Brewer et al, 1999). Brewer and colleagues used adult participants, most of whom had some inconsistencies in their interviews; therefore, it is natural to assume that if the participants had been children they would also have contradicted themselves with similar results.
The exact relationship between consistency and accuracy is not very clear and little research has been done to gain a better understanding of the issue (Brewer et al, 1999). Since accuracy is difficult to assess directly, it is often inferred by the credibility and competency of a witness. These are two distinct qualities that imply whether the witness is being truthful or able to recall important details of a particular situation (Perry and Wrightsman, 1991), so both are highly sought after in an eyewitness. The general perception of child EWT is that young children are honest but not necessarily competent eyewitnesses (Perry and Wrightsman, 1991). This has serious implications for jurors' decisions in court cases because the perceived credibility and competency of a child witness may determine whether their statements are considered reliable and accurate. The following paragraphs will discuss studies that have investigated the factors that affect the perceived accuracy of child EWT.
McCauley and Parker (2001) examined if victim's age, nature of case or juror's gender had an impact on the credibility of EWT and how this affected the decisions of a mock jury. The study used transcripts based on the testimonies of girls aged 6 or 13 that had been the victims of robbery or sexual assault. The researchers found that the age of the child and the gender of the jury members affected the perceived honesty and memory reliability of the EWTs. In general, the 13-year olds were perceived to be more honest and to have a better memory than the 6-year olds; however, there was a trend for female jurors' to believe that both had equally good memory and honesty. Furthermore, the nature of the case significantly influenced the perceived accuracy and credibility of the children: victims of sexual assault were thought to have better memory of the incident and to be more credible witnesses than victims of robbery. The results also implied that credibility influenced the final verdict whereas the age of the witness did not: increased witness credibility was expressed in higher conviction rates, especially in the sexual assault cases. This study illustrates the complexity of assessing the reliability and accuracy of child EWT: the victim's age, nature of the case and gender of the jury all have an impact on the credibility and perceived accuracy of the EWT.
In addition to the study of McCauley and Parker (2001), Wright and collaborators (2010) explored the perceived memory accuracy and competency in children aged 3 to 18. Adult participants had to rate the reliability of the witness memory and their honesty for two different scenarios where a child was a bystander. The results suggest that adults believe that the memories of older children are more reliable, with the fastest increase in perceived reliability and honesty with age in children up to the age of 6. The perceptions of children older than 6 were mixed and influenced by participants' gender: female participants believed that children's EWT was equally reliable from the age of 6 up to early adolescence while males continued to think that age increased EWT dependability (Wright et al, 2010). This study indicates that children are less likely to be perceived as credible and competent witnesses before the age of 6, however for older children it is very much dependent on the maturity and characteristics of the child witness in each individual case. Similarly to McCauley and Parker (2001), for the cases with EWT from older children the current study suggests a difference between the perceptions of male and female jurors on child EWT reliability and that this factor is more dominant in determining the perceived reliability of a child EWT than age. This result is important when considering how juries are formed for cases involving child EWT: male dominated juries are less likely to believe the testimonies of younger children while female dominated juries may accept child EWT too readily; both potentially leading to unjust verdicts.
When considering the nature of case sensitivity in the study by McCauley and Parker (2001) one should note that child EWTs related to sexual assault are more emotionally negative and stressful than many other crimes. The impact of such salient experience on memory is important for the assessment of their EWT accuracy. Shrimpton and colleagues (1998) investigated the effect of stress on children's memory in 4- to 12-year olds who either had a venipuncture  or who watched the procedure being demonstrated. This is an inconvenient procedure, which would cause mild stress in comparison to the stress experienced by children who are sexually assaulted; in laboratory conditions it is difficult and probably unethical to produce comparable stress levels. The study showed that the group who experienced distress recalled the event just as well as the control group; albeit the older children remembered more details and had fewer inaccuracies in their responses than the younger children (Shrimpton et al, 1998). They also found that the 4-5-year olds were more vulnerable to misleading questions but reducing the time delay between the event and the interview had a positive impact on the accuracy of responses, especially in younger children (Shrimpton et al, 1998). The study concludes that children are able to accurately recall stressful events that they have experienced regardless of their age and that older children remember more details in free recall. This is important for child EWT because it demonstrates that children can give accurate testimonies even if the witnessed event was stressful. The study also adds evidence that one should have confidence in younger children's reports that, although often not as detailed as older children's accounts, are nevertheless accurate.
While stressful events have been shown not to hinder children's recall ability (Shrimpington et al, 1998), other intervening events might have a negative impact. The effects of events that occur between the incident and the interview on memory recall are important for assessing the reliability and accuracy of child EWT in cases where the event took place a long time ago. Principe and co-workers (2000) investigated how different intervening events affected long term memory in young children aged 3-5 who underwent physical examination. The control group was interviewed straight after the appointment and then in 3 months time, whereas the 3 other groups received an intervening encounter in the form of another interview, a videotape of someone else's physical examination or a visit to doctor's office. The results showed that the group that had an intervening interview and the videotape group had better memory recall of the event in the final assessment than the control group or the return visit group (Principe et al, 2000). The researchers suggest that the first two conditions offered a full reminder of the original event whereas the exposure to the initial context was not a strong enough reminder of the examination (Principe et al, 2000). In sum, intervening events can have a great impact on a child's ability to remember the initial event. The opportunity to reassert the details of the initial experience can have a positive effect on memory recall; however accuracy can be hindered if children are faced with misguiding information.
In addition to occasions where children have the opportunity to discuss the event of interest, rumours and discussion with others may influence the accuracy of children's memories (Principe and Ceci, 2002). In a study with preschoolers, Principe and Ceci (2002) showed that rumours about an event influenced the reports of children who did not experience the incident themselves. This is important for EWT because children may incorporate things they did not witness by hearing the accounts of others. Thus, although using free recall is considered more accurate when dealing with children (Larsson and Lamb, 2009), it can produce erroneous testimonies (Ceci et al, 2007), especially if the child has been exposed to discussions about the incident. The current view is that young children are the most suggestible group, as well as the group most likely to be lead astray by misleading questions (Shrimpton et al, 1998),
Candel and co-workers (2007) investigated the effect of peer discussion on the accuracy of children's memory in slightly older children than Principe and Ceci: 6- and 12-year old groups watched alternative versions of a video clip either in pairs or alone. The results indicated that older children were better in remembering details and had higher accuracy, while in the pair condition they incorporated more supplementary information that they did not see in the video based on the discussion with their co-witness (Candel et al, 2007). As shown in previous examples, there seems to be a developmental difference in children's memory ability, which need not influence the overall reliability but rather the accuracy of detail. Both older and younger children's accuracy was negatively affected in the co-witness condition by the conformity effect but it was more pronounced amongst the older children. Therefore, suggestibility is not a problem with only the EWTs of the youngest witnesses but possibly even more so with early adolescents and more research should focus on the developmental differences of resistance to suggestibility of false information amongst older children. The result of this study is in contrast to the correlation of perceived accuracy with age found in the earlier studies and suggests that jurors are either not aware or over simplify their personal thoughts on the relationships between suggestibility, age and accuracy.
Apart from the factors that affect recall ability, the way an interview is carried out can have an impact on the accuracy of testimony. Multiple interviewing can have a positive affect on memory recall when it helps to strengthen the memories as demonstrated by Principe and others (2000). However, repeated questioning in a legal setting has been shown to be disruptive. Krähenbühl and colleagues (2009) interviewed 4-9 year old children after a week's delay about an event they had witnessed at school. The researchers based their questioning pattern on the police interviews carried out with children, which included numerous repetitions. The findings indicated that after the first repeat of a question, 25% of the responses to questions children were unsure of initially were different from the those in the original answer; thereafter the number or repetitions did not affect the answer (Krähenbühl et al, 2009). Also, children from all age groups changed their answers more often when they were asked questions that they could not answer based on the experienced event, even if they initially said that they did not know the answer. The implications of this study are that repetitive questioning in legal context could undermine the child's credibility as a witness when he or she gives inconsistent answers, and as Brewer and co-workers (1999) suggest, contradictive statements in one's testimony are perceived to indicate EWT inaccuracy. This emphasises the importance of making sure children are aware that not knowing the correct answer is alright and that they should not be forced into responding to a question that they do not know the answer to. Avoiding repetitive questioning may not always be possible but investigators should know that children might feel pressured into giving inaccurate statements.
Sometimes not only how and what type of question is asked but also the context of the interview can affect EWT accuracy: children tend to be more inaccurate in their statements if the interview is carried out by an authority figure (Lowenstein et al, 2010). Lowenstein and others (2010) conducted a study using 9-10 year olds to investigate the effects of the presence of uniformed officials on the accuracy of children's eyewitness identification. Children were exposed to a staged crime and had to identify the offender from a photo line-up where the culprit was present or absent and with the procedure carried out by either uniformed or non-uniformed personnel. The results of the study indicate that children were faster and more accurate in their decisions when the target was present in the line-up. When the target was absent children in the non-uniformed group took longer on their decisions and were less confident in their conclusion yet more accurate than children in uniformed group. The uniform present group identified more possible targets and in the culprit-absent condition had a greater false-positive rate,. This can be explained by childrens' desire to conform to the demands placed on them and children may feel that they are expected to identify the target (Lowenstein et al, 2010). The presence of a uniformed officer is likely to pressure the child to identify someone from the line-up, which may lead to miscarriages of justice and reduce children's credibility as witnesses. Similar pressure need not only apply to line-up identification but also to EWT interviews when carried out by a uniformed authority figure. These results imply that the accuracy of children's EWT can highly depend on the interview context and this should carefully be considered when arranging the collection of a child EWT. It should be noted that there is a great variance in the practices of the UK constabularies concerning the wearing of uniform when dealing with child eyewitnesses (Lowenstein et al, 2010). A standardised practice of not wearing a uniform across the UK may prove helpful in ensuring unbiased child witness accounts and increase their accuracy.
There are several other interview techniques used with children that can also have a great impact on the accuracy of EWT, for example prop-assisted and verbal interviews that include methods like the cognitive interview, narrative elaboration and others (Melinder et al, 2010). The verbal interview is common in legal settings and forensic psychologists are more likely to use the prop-assisted technique. Melinder and colleagues (2010) have recently tested the effects of these techniques on children's recall ability: 4-year old children were interviewed about a medical examination either after 2 and 6 weeks or only after 6 weeks by an non-uniformed policeman or by a psychologist neither of which knew what details the examination had included. The researchers revealed that throughout verbal interviews children were asked more free recall questions and the discussion did not drift off from the event of interest. In the prop-assisted interview children made more commission errors, possibly because stimulating children's fantasy to help them talk about their experiences may lead the psychologists to ask more questions about imaginary things resulting in erroneous reports and also, children may get distracted by some more desirable toys. However, children were better able to correctly reject false suggestions using props but this effect was not prevalent in the second interview (Melinder et al, 2010). Thus, using verbal interview with children may help produce more accurate EWTs than prop-assisted interviews because the latter introduces more suggestibility into the reports.
As has been shown, there are many factors and conditions that affect the reliability and accuracy of child EWT. Age, intervening events and repeated interviews can have both positive and negative effects on the accuracy of EWT. From the previously discussed examples it has emerged that interviewer and the questioning methods themselves can have a great impact on the quality of received statements. To maximise the benefits of child EWTs the interviewers should take into account the developmental differences of children from different age groups i.e. young children have been shown to give less detailed testimonies, which is not an indicator of poor memory but could reflect linguistic abilities that develop with age (Larsson and Lamb, 2009). Thus, it is best to use open-ended questions that encourage free recall and minimise interviewer bias but as Ceci and others (2007) show, open-ended free recall is not always error free. As young children are more suggestible they are more vulnerable to leading questions, positive-negative enforcements and forced answer questions that will introduce errors into child EWTs (Larsson and Lamb, 2009) that can undermine their credibility as witnesses. To improve the credibility of child EWT the interviewee should be made aware of the significance of his/her testimony and that they are not expected to be able to answer every question. Also, it may improve the accuracy of received EWT if the interview is carried out by a non-uniformed officer to reduce the pressure that the child is under and if the repetition of questions is kept to a minimum. Since the reliability of child EWT is easily compromised by suggestive interview techniques and misleading information, especially in children under the age of 6, an independent monitor, trained in the practices discussed, should be present to ensure unbiased questioning. One should note however that it has been shown that not only the EWT's of children but also those of adults may be distorted by suggestive methods (Loftus, 2004) so this by no means is an issue solely associated with the young. In summary, there are many simple methods that can be implemented to increase child EWT accuracy and more work needs to be carried out in this complex matter that is still not yet fully understood.