The role of repeated interviewing in children's responses to cross-examination-style questioning

Abstract

The negative effect of cross-examination-style questioning on children's accuracy is likely to be due to the complex and credibility-challenging questions that characterize the interview. Given that cross-examination occurs after at least one prior interview, however, it is equally possible that repeated interviewing per se impairs children's accuracy, and that the questions asked have little bearing on children's responses. To examine this issue, 5- and 6-year-old children (n= 82) and 9- and 10-year-old children (n= 103) took part in a surprise event and were then interviewed using an analogue of direct examination. Either 1 week or 6 months later, half of the children were re-interviewed with an analogue of cross-examination designed to challenge their direct examination responses. Remaining children were re-interviewed with the same questions that were asked during direct examination. Children's accuracy decreased following their second interview, irrespective of age or delay; however, delay particularly impacted younger children's second interview performance. Children's accuracy was most impaired following a cross-examination-style interview. Overall, cross-examination-style questioning appears to be particularly detrimental to obtaining accurate event reports from children.

Background

Although considerable efforts have been made to modify legal procedures for children who negotiate the adversarial legal system, reforms have typically focused on children's direct evidence (American Professional Society on the Abuse of Children, 1990; 1997; Home Office, 1992; 2000). In contrast, reform of the cross-examination process, during which the opposing lawyer challenges the child's testimony, has lagged (Gupta, 1994; Hamlyn, Phelps, Turtle, & Sattar, 2004; Pipe & Henaghan, 1996). There is good reason to be concerned about this issue; laboratory research has repeatedly shown that children's accuracy decreases markedly when they are interviewed with an analogue of cross-examination (Righarts, Zajac, Jack, & Hayne, 2011; Zajac & Hayne, 2003; 2006; Zajac, Jury, & O’Neill, 2009). Although this effect is robust, there are unanswered questions about the precise mechanism(s) that account for it. In this study, we attempted to further clarify the factors that contribute to children's poor cross-examination performance.

Cross-examination research

Despite long-voiced concerns about the suitability of the cross-examination process for children (Brennan & Brennan, 1988; Eastwood & Patton, 2002; Zajac, Gross, & Hayne, 2003; Zajac & Hayne, 2003, 2006; Zajac et al., 2009), empirical research into cross-examination is a relatively new phenomenon. Numerous court transcript studies have now raised concern about the types of questions that cross-examining lawyers ask children (i.e., leading, complex, confusing, and credibility-challenging questions; Brennan & Brennan, 1988; Davies, Henderson, & Seymour, 1997; Davies & Seymour, 1998). Few studies have examined children's responses to these questions, but those that have suggest that cross-examination poses real problems for children. Zajac et al.'s (2003) analysis of 5- to 13-year-old children's courtroom testimony, for example, showed that children engaged in both helpful strategies (e.g., seeking or giving clarification) and potentially unhelpful strategies (e.g., complying with leading questions) when attempting to answer direct examination and cross-examination questions. Children's responses to cross-examination, however, were particularly concerning, with over 75% of children changing at least one aspect of their earlier testimony during this process.

Of course, changes to testimony can render a child's evidence more accurate or less accurate, and court transcript data do not allow researchers to distinguish between these possibilities. Consequently, researchers have recently taken cross-examination questioning into the laboratory. In Zajac and Hayne's (2003) seminal study, 5- and 6-year-old children visited the police station and then completed two interviews about their experiences. The first interview, direct examination, took place 6 weeks after the event and comprised open and closed questions about the visit. The second interview, conducted 8 months later, was developed using court transcripts and designed to mimic cross-examination. Its chief aim was to talk children out of their direct examination responses, irrespective of accuracy. Mirroring findings from transcript studies, the majority of children (85%) changed at least one of their original responses during the cross-examination interview. These changes were detrimental to children's overall accuracy. Subsequent studies have replicated these findings (Righarts et al., 2011; Zajac & Hayne, 2006; Zajac et al., 2009).

Explanations for the cross-examination effect

Why does cross-examination reduce the accuracy of children's event reports? Based on existing research, we can effectively rule out several explanations. First, we know that children undergoing cross-examination do not change their responses because they were originally mistaken. On the contrary, children's direct examination reports are generally highly accurate and, during cross-examination, children often make changes to responses that were originally correct (Righarts et al., 2011; Zajac & Hayne, 2003, 2006; Zajac et al., 2009).

Another possible explanation for children's poor cross-examination performance is that they do not remember the target event details. Significant delays are common in forensic settings (Eastwood & Patton, 2002; Goodman et al., 1992, Martone, Jaudes, & Cavins, 1996), and there are well-documented developmental differences in rates of forgetting (Brainerd, Kingma, & Howe, 1985; Brainerd & Reyna, 1995) and resistance to suggestion (Pezdek & Roe, 1995) over time. Poor cross-examination performance might therefore be explained by forgetting that occurs during the long delays between the target event and cross-examination. Righarts et al. (2011), however, discounted this possibility by manipulating the delay between the event and subsequent cross-examination interview (1 week or 8 months). Cross-examination performance did not differ as a function of delay; even children who were interviewed very soon after the event showed a decrease in accuracy during cross-examination. Furthermore, when the yes/no questions from the direct examination interview were repeated 1 week after the cross-examination interview, children's accuracy was high and did not differ from their direct examination interview performance.

The findings described above suggest that children's responses to cross-examination questioning are, at least largely, socially motivated. Further support for this notion comes from individual difference studies showing that children who are more self-confident (Vrij & Bush, 2000), more assertive (Muir-Broaddus, King, Downey, & Petersen, 1998), and have higher self-esteem (Howie & Dowd, 1996) are better able to resist misleading questions. These characteristics also buffer children from the negative effects of cross-examination questions (Zajac et al., 2009).

Given mounting evidence suggesting that children's responding reflects compliance rather than memory failure, it is most likely that children's poor cross-examination performance can be attributed to the specific types of questions asked. We know that children's reports are most consistent (Lamb & Fauchier, 2001; Orbach & Lamb, 2001) and most accurate (Cassel, Roebers, & Bjorklund, 1996; Dent & Stephenson, 1979; Poole & Lindsay, 1995; Quas & Schaaf, 2002) when elicited with open-ended invitations (e.g., “Tell me everything you remember”). In stark contrast, the leading, complex, and credibility challenging questions that characterize cross-examination (Brennan & Brennan, 1988; Davies et al., 1997; Zajac et al., 2003) contravene almost every guideline established for maximizing children's accuracy (e.g., Home Office, 2011).

To test this hypothesis, however, we need to consider one additional factor: repeated questioning. Research suggests that repeated questioning can lead children to change their previous responses, resulting in reduced accuracy (Howie, Sheehan, Mojarrad, & Wrzesinska, 2004; Memon & Vartoukian, 1996). Given that cross-examination, by definition, occurs after at least one interview has taken place, we need to know whether children would alter their reports during cross-examination irrespective of the types of questions asked.

Repeated questioning and children's reports

The effect of repeated questioning on children's reports has traditionally been studied in two distinct contexts: repetition within an interview and repetition across interviews. Although most studies show children's response accuracy decreases when questions are repeated within an interview (e.g., Howie et al., 2004; Memon & Vartoukian, 1996; Siegal, Waters, & Dinwiddy, 1988), more relevant to cross-examination is the effect of repeated questioning across interview sessions.

Theorists suggest that re-presenting test material during a delay period can improve recall (Erdelyi & Stein, 1981; Roediger & Payne, 1982; Tulving, 1967). Multiple interviews, which commonly occur in forensic settings (Goodman et al., 1992; Tedesco & Schnell, 1987), may therefore be expected to result in improved memory performance. Repeated interviews, however, can facilitate or impair children's recall, depending on the type of questions asked (Gordon, Baker-Ward, & Ornstein, 2001; Poole & White, 1995). Children interviewed multiple times with open-ended questions tend to show greater net recall of correct information than those interviewed fewer times, with no increase in errors (Baker-Ward, Hess, & Flannagan, 1990; Leichtman & Ceci, 1995; Poole & White, 1991; Principe, Ornstein, Baker-Ward, & Gordon, 2000). When specific questions are repeated across interviews, however, interviewees tend to recall more correct information, but their error rate increases (Brainerd & Reyna, 1996, Experiment 1; Dent & Stephenson, 1979). Finally, there appears to be little doubt that children's accuracy across repeated interviews is most severely affected when those interviews involve misinformation (Ceci, Huffman, Smith, & Loftus, 1994; Gobbo, 2000; Leichtman & Ceci, 1995; Poole & Lindsay, 1995; Quas & Schaaf, 2002). It therefore appears that multiple interviews per se do not account for inaccuracies in children's reports; rather, errors can be attributed to inappropriate questioning during those interviews (Poole & White, 1995).

The present study

It is well established that cross-examination diminishes the accuracy of children's event reports, but the reason for this has not been empirically verified. This study is the first to investigate whether it is the questions that comprise cross-examination that are problematic, or if children's report accuracy would decrease following a second interview irrespective of the types of questions they are asked. This issue has marked implications for the criminal justice system, as it will allow us to ascertain whether changing the basic structure of cross-examination could facilitate children's accuracy.

After a surprise memory event, children were interviewed using an analogue of direct examination. Half of the children were then re-interviewed with an analogue of cross-examination, while remaining children were re-interviewed with the direct examination questions. Given that repeating specific questions across interviews can lead to a decrease in children's accuracy (Brainerd & Reyna, 1996; Dent & Stephenson, 1979), we expected accuracy to be compromised in the second interview, regardless of its format, but that children interviewed with cross-examination questions would be less accurate than those re-interviewed with direct examination questions.

We also explored two further variables. First, we manipulated the delay between the first and second interviews. We know that children's recall accuracy (Ornstein et al., 2006) and resistance to suggestion (Pezdek & Roe, 1995) decreases over time. Given that several months often lapse between the reporting of a crime and trial (Eastwood & Patton, 2002; Goodman et al., 1992; Martone et al., 1996; Westcott & Page, 2002), delay is a crucial factor to consider. While many countries allow children's direct examination to be recorded at time of disclosure and played in court, children must then wait several months before cross-examination occurs at trial (e.g., Gupta, 1994; Hamlyn et al., 2004; Pipe & Henaghan, 1996). Consistent with Righarts et al. (2011), we expected that delay would not affect cross-examination accuracy, but that children in the repeated direct examination condition would be less accurate after a long delay, relative to a short delay.

Second, we manipulated child age. Legislative accommodations for vulnerable witnesses (e.g., the use of CCTV and screens) have not only led to more children being called as witnesses, but also to younger children taking the stand (Wilson & Davies, 1999). Understanding how age affects children's performance is therefore crucial. As children get older, they provide more detailed reports of autobiographical events (Fivush, 2002; Gordon et al., 2001; Quas, Goodman, Ghetti, & Redlich, 2000), are more accurate over longer delay periods (Peterson, Moores, & White, 2001), and are more able to resist suggestion (Goodman, Quas, Batterman-Faunce, Riddlesberger, & Kuhn, 1994; Goodman & Reed, 1986; Pezdek & Roe, 1995). Although older children still have considerable difficulty with cross-examination questioning, they do tend to outperform their younger counterparts (Zajac & Hayne, 2006). We therefore hypothesized that younger children would be less accurate than older children following a second interview (irrespective of format) and a long delay.

Method

Participants

Five- and 6-year-old and 9- and 10-year-old children were recruited from eight schools in the Otago and Southland regions of New Zealand. Two participants did not complete the study because they were ill on the days of their interviews (attrition rate = 1%), leaving a final sample of 185 children (5- and 6-years-old: n= 82, M= 5.79 years, SD= .48; 9- and 10-years-old: n= 103, M= 10.04 years, SD= .65). The sample comprised approximately equal numbers of boys (n= 95) and girls (n= 90). Children were predominantly of New Zealand-European descent, and all had written parental consent to participate. Experimenters obtained each child's assent before each interview and children were able to discontinue an interview at any time.

Experimental procedure

Memory event

In class groups, children took part in a surprise visit to their local police station. During their visit, children took part in several activities. First, they got their thumbprints and ‘mugshot’ photos taken. They then visited unique parts of the station (e.g., watch house, jail cell, gymnasium, communications room). They also saw a police car with its siren sounding and lights flashing. Finally, the male police officer who led the tour turned the children into ‘honorary police officers’ for the day by spinning each child around by his or her shoulders. This final event was designed to mimic a situation in which children experience bodily touch by a male authority figure. The police officer told the children to keep this event a secret.

Direct examination interview

One to 4 days following the police station event (M= 2.45 days, SD= 1.60), each child was individually interviewed about the visit by an unfamiliar female experimenter. This direct examination interview (see Zajac & Hayne, 2003, 2006) was designed to mimic the videotaped evidential interview that children in many jurisdictions undergo after making a sexual abuse allegation. The interviews were conducted in a small room at the child's school, and were audiotaped and videotaped. After a brief rapport-building period, the experimenter began the interview by saying:

I heard that yesterday [on Monday] you did something really special – you got to visit the police station! Well, I couldn't go, so I want you to tell me what happened. Tell me everything that happened when you went to the police station.

The experimenter used minimal encouragers (e.g., ‘Mmm-hmm’; ‘wow!’; head nodding) and further open-ended invitations (e.g., ‘Can you tell me anything else about the police station?’) to encourage the child to report all that he or she could remember. When a child indicated that he or she could not remember anything else, the experimenter asked a series of four specific questions. Two of these questions concerned true events, and two concerned events that did not occur at the police station (i.e., false events). The introductory statement made by the experimenter and the questions appear below, with false events indicated by an asterisk.

Now I’m going to ask you some more questions. You might have already told me the answers, but if you have, just tell me again, okay?

• 1
  Did you get your photo taken?
• 2
  Did you get to try on handcuffs?*
• 3
  Did you see a lady come in and say that her child's bike had been stolen?*
• 4
  Did the policeman spin you around by your shoulders?

If the child responded affirmatively to any of the questions, the experimenter said, ‘Tell me more about that’. Again, the experimenter used minimal encouragers until the child indicated that there was nothing more to report. The child was then thanked and sent back to class.

Second interview

Following the direct examination interview, all children were interviewed a second time (see Table 1). Children were randomly assigned to either the short delay condition (approximately 1 week after the police station visit; M= 8.04 days, SD= 1.93) or the long delay condition (approximately 6 months after the visit; M= 6.11 months, SD= .17). The format of the second interview was also manipulated. Within each delay condition, children either received a cross-examination interview (cross-examination condition) or were asked the same yes/no questions that they had been asked in the direct examination interview (repeated direct examination condition).

Table 1. Participant ns as a function of age group, delay, and second interview format

Age group	Participants (n)				Total
	Repeated direct examination		Cross-examination
	Short delay	Long delay	Short delay	Long delay
5- and 6-year-olds	20	20	21	21	82
9- and 10-year-olds	24	27	25	27	103

All interviews were conducted by an unfamiliar female experimenter in a spare room at the child's school. In most cases, this second interview was carried out in the same room as the first interview. Occasionally, however, a different room was used due to space constraints at the school. All interviews were videotaped. At the conclusion of the interview, each child chose a sticker before returning to his or her classroom.

Cross-examination interview Our cross-examination interview was developed by Zajac and Hayne (2003, 2006). Its aim was to talk the child out of his or her direct examination responses, irrespective of accuracy. Experimenters adopted a professional, non-aggressive tone during this interview, which began with the following statement:

A little while ago, a lady came and asked you questions about the police station. Do you remember that? I’m going to show you a video of you talking to that lady. You need to watch and listen really closely to it, and then I’ll ask you some more questions. Okay?

To mimic court procedure in countries allowing pre-recorded direct evidence (e.g., New Zealand, Australia, and England), the child then watched a videotape of his or her direct examination interview. The experimenter sat with the child during this process, but remained silent unless redirecting the child's attention back to the screen. After the video, the experimenter said, ‘Okay, now I have some more questions to ask you’, and then proceeded to ask three introductory questions:

• 1
  Can you tell me how old you are now?
• 2
  Do you know when your birthday is?
• 3
  Can you tell me where you live?

Each child was then asked 10 questions for each of the four direct examination interview topics (i.e., photo, handcuffs, stolen bike, spinning around). The aim of each set of 10 cross-examination questions was for the child to change his or her direct examination responses. For example, if during the direct examination interview the child had stated that he or she was spun around, the aim of cross-examination was for the child to say that this did not occur. If an outright change did not occur, a secondary aim of cross-examination questioning was for the child to concede that his or her original answer might have been wrong (i.e., that the interviewer's suggestion was a possibility). Each question set followed the outline below (see Appendix A for four sets of sample cross-examination questions):

• Question 1: The first question clarified the response that the child gave in his or her direct examination interview (e.g., ‘You said in the video that the policeman spun you around, didn't you?’).

• Question 2–7: These questions were indicative of the type and proportions of questions that lawyers use while cross-examining a witness (e.g., leading, ambiguous, complex, and irrelevant).

• Question 8: This question challenged the child's certainty about the topic in question (e.g., ‘Are you sure the policeman spun you around?’).

• Question 9: This leading question expressed disbelief at the child's response and provided an alternative scenario (e.g., ‘I think maybe your friends got spun around, but you didn’t. That's what really happened isn't it?’). Some reasons for disbelief were obtained from actual transcripts of cross-examination of child witnesses (Zajac et al., 2003); others were things that parents and adults commonly say when challenging children. Four of the reasons for disbelief aimed to talk the child out of saying that an event had occurred, and four of the reasons aimed to talk the child into saying that an event had occurred (see Table 2).

• Question 10: This leading question was only asked if the child did not acquiesce to Question 9. It was designed to get the child to admit that the reason for disbelief given in Question 9 was a possibility (e.g., ‘That might have happened though, mightn't it?’).

Table 2. Experimenter's reasons for disbelief aimed at talking a child into or out of the occurrence of an event

Experimenter's aim	Reason for disbelief	Wording
Talk child into saying that an event	Lack of attention	‘I think it did happen but you weren't paying attention’
occurred	Forgetting	‘I think it did happen but you’ve forgotten it’
	Inability to remember	‘I think it did happen, but you can't quite remember it’
	Appeal to authority	‘But if your teacher said that it did happen she’d be right wouldn't she?’
Talk child out of saying that an event	Suggestion from others	‘I think that someone told you to say that’
occurred	Mistaken participation	‘I think it happened to your friend and not you’
	Purposeful fabrication	‘I think you’re making that up’
	Appeal to authority	‘But if your teacher said that it didn’t happen she’d be right wouldn't she?’

After the interview, the experimenter ascertained how the child was feeling, and made a comment such as, ‘Some of those questions were pretty tricky, you answered them really well.’

Repeated direct examination interview.  The repeated direct examination interview provided an alternative format for the second interview. The initial part of this interview was the same as the cross-examination interview – the experimenter introduced the interview, each child watched his or her direct examination interview, and then the interviewer asked the three introductory questions. After this point, however, children in this condition were merely asked the same yes/no questions that they were asked during their direct examination interview.

Coding

Direct examination

Children received 2 points for each correct response to the four yes/no questions asked during the direct examination interview. No points were awarded for an incorrect response. Given that children in forensic settings are often encouraged to say I don't know rather than give an incorrect response (Morgan Libeau, Woodham, & Rickard, 2003), one point was awarded if the child indicated that he or she was uncertain as to whether an event had taken place (e.g., I don't know or I can't remember). Response scores to the questions were added to give a direct examination score ranging from 0 to 8 (or 0 to 4 when true and false events were considered separately). Two experimenters independently scored 25% of the direct examination interviews. An inter-observer agreement value of 100% was obtained.

Cross-examination

Coding of children's cross-examination responses was based on that of Zajac and colleagues (Zajac & Hayne, 2003; 2006; Zajac et al., 2009). Points were added or deducted from children's direct examination scores if children changed their original response. Two points were added or deducted if the child acquiesced with the experimenter's alternative suggestion (i.e., if the child agreed with the experimenter's suggestion in Question 9). Points were added if acquiescence was the correct response and deducted if acquiescence was incorrect. For children who did not acquiesce completely, but admitted that the alternative suggestion was possible (i.e., agreed with the suggestion in Question 10), 1 point was added or deducted from their score for that question. For children who did not change a given response during cross-examination, no points were added or deducted. Examples of each of these scoring scenarios appear in Appendix B. Children's cross-examination accuracy scores could range between 0 and 8. Two experimenters independently coded 25% of the cross-examination interviews. An inter-observer agreement value of 95% was obtained. Disagreements were resolved through discussion.

Repeated direct examination

The repeated direct examination interviews were submitted to the same method of scoring as the cross-examination interviews. Points were added or deducted from children's direct examination scores if children changed their original response. Two points were added if the yes response was correct (e.g., ‘Did you get your photo taken?’), but the points were deducted if yes was an incorrect response (e.g., ‘Did you get to try on handcuffs?’). If a child responded I don't know or I can't remember, 1 point was added or deducted in a similar fashion. If a child did not change his or her direct examination response, no points were added or deducted. Points were combined to give an accuracy score ranging from 0 to 8. Two experimenters independently coded 25% of the repeated direct examination interviews. An inter-observer agreement value of 100% was obtained.

Results

This study was a 2 (Age: 5- and 6-year-olds, 9- and 10-year-olds) × 2 (Delay: short delay, long delay) × 2 (Second Interview Format: repeated direct examination, cross-examination) × 2 (Interview Phase: direct examination, second interview) design, with repeated measures over interview phase.

Accuracy scores

Direct examination interview

Recall that children were assigned direct examination accuracy scores based on their responses to the four yes/no questions. Total accuracy scores could range from 0 to 8. Overall, both 5- and 6-year-old and 9- and 10-year-old children provided largely accurate direct examination reports (M= 7.42, SE= .08). The majority of younger (62%) and older children (86%) were completely accurate. The lowest direct examination score was 4; this was achieved by 10% of younger children and 2% of older children. Recall that children could express uncertainty about whether an event had occurred (i.e., respond I don't know). The frequency of these responses was very low. For younger children, n= 4 (1.2%), direct examination responses were classified this way. For older children this figure was n= 3 (0.7%).

We then examined children's responses to the true and false event questions separately. A 2 (Age) × 2 (Event Type: true, false) ANOVA, with repeated measures over event type, revealed two significant main effects. First, children's reports about the false events (i.e., trying on handcuffs; seeing a lady reporting her child's bike as stolen; M= 3.87; SE= .04) were more accurate than their reports about the true events (i.e., getting their photo taken, being spun around by the police officer; M= 3.55; SE= .07), F(1, 183) = 16.21, p < .0001, f= .30 (see Table 3).¹ A 2 (True Event: photo, spin) × 2 (Response Accuracy: accurate, inaccurate) chi-square contingency test revealed that responses to the questions about getting a photo taken (90%) and being spun around (88%) were similarly accurate, χ² (1, N= 370) = .45, p= .50, w= .03, power = .99. Likewise, a 2 (False Event: handcuffs, bike) × 2 (Response Accuracy: accurate, inaccurate) chi-square contingency test revealed that responses to the questions about trying on handcuffs (98%) and the stolen bike (97%) were similarly accurate, χ² (1, N= 370) = .41, p= .52, w= .03, power = .99. Second, regardless of event type, older children's accounts of the police station visit were more accurate than those of the younger children, F(1, 183) = 13.01, p= .0004, f= .27 (see Table 3). The Event Type × Age interaction was not significant, F(1, 183) = 1.72, p= .19, f= .10, power = .93.

Table 3. Mean response accuracy (standard error) for questions about true and false events as a function of age

Age group	Event type		Total
	True	False
5- and 6-year-olds	3.33 (.12)	3.77 (.09)	7.10 (.15)
9- and 10-year-olds	3.73 (.08)	3.95 (.03)	7.68 (.09)
Overall	3.55 (.07)	3.87 (.04)	7.42 (.08)

Second interview

As in the direct examination interview, accuracy scores in the second interview could range from 0 to 8. Our next step was to examine how the second interview impacted on children's accuracy levels, and to assess the contributions of age and delay to children's performance. To do this, children's accuracy scores were submitted to a 2 (Age) × 2 (Delay) × 2 (Second Interview Format) × 2 (Interview Phase) ANOVA, with repeated measures over interview phase.

A significant Age × Delay × Interview Phase interaction emerged, F(1, 177) = 3.97, p= .048, f= .15, and was explored by conducting two separate 2 (Age) × 2 (Interview Phase) ANOVAs for the short and long delay conditions, each with repeated measures over interview phase. At the short delay (see Figure 1, left side), children obtained higher accuracy scores in the direct examination interview (M= 7.54, SE= .11) than during the second interview (M= 6.20, SE= .22), F(1, 88) = 42.72, p < .0001, f= .70. The main effect of age approached significance, F(1, 88) = 3.11, p= .08, f= .19, power = .65, with a trend for older children (M= 7.09, SE= .18) to provide more accurate responses than younger children (M= 6.61, SE= .20), regardless of interview phase. The Interview Phase × Age interaction was not significant, F(1, 88) = .28, p= .60, f= .06, power = .65.

Figure 1. Mean response accuracy (+1 SE) at each Interview Phase (direct examination, second interview) as a function of Age (5- and 6-year-olds, 9- and 10-year-olds) and Delay (short delay, long delay).

At the long delay (see Figure 1, right side), significant main effects of interview phase and age were qualified by a significant Interview Phase × Age interaction, F(1, 93) = 7.43, p= .008, f= .28. Accuracy was lower in the second interview than in the direct examination interview, both for younger children, t(40) = 5.88, p < .0001, d= .92, and older children, t(53) = 4.09, p= .0001, d= .56. Older children, however, were more accurate than younger children during the direct examination interview, t(93) =−3.33, p= .001, d= .69, and the second interview, t(93) =−4.76, p < .0001, d= .99.

The four-way ANOVA also revealed a significant Interview Phase × Second Interview Format interaction, F(1, 177) = 79.43, p < .0001, f= .67 (see Figure 2). Children's accuracy during the direct examination interview did not differ as a function of subsequent group membership (i.e., the format of the second interview; repeated direct examination, M= 7.51, SE= .11; cross-examination, M= 7.34, SE= .12), t(183) =−1.00, p= .32, d= .15, power = .96. This was expected given that children were randomly assigned to the second interview conditions. In contrast, there was a significant effect of interview format on children's accuracy during the second interview. Children's accuracy diminished following a second interview, whether children received a repeated direct examination interview (M= 7.19, SE= .15), t(90) = 2.75, p= .007, d= .29, or a cross-examination interview (M= 4.65, SE= .22), t(93) = 11.10, p < .0001, d= 1.14. Accuracy during the second interview, however, was considerably lower when children were cross-examined, t(183) =−9.34, p < .0001, d= 1.37.

Figure 2. Mean response accuracy (+1 SE) at each Interview Phase (direct examination, second interview) as a function of Second Interview Format (repeated direct examination, cross-examination).

Response changes

The finding that children's accuracy decreased during the second interview tells us that children made changes to their direct examination reports. The final phase of our analysis was to examine these changes. Children were considered to have changed a direct examination report if, in the second interview, they received a score of 2 (direct contradiction) or 1 (partial contradiction) for a given question.

The majority of 5- and 6-year-olds (66%) and 42% of 9- and 10-year-olds made at least one change to their earlier responses. All four original responses were changed by 24% of younger children and 10% of older children. A 2 (Age) × 2 (Delay) × 2 (Second Interview Format) between-subjects ANOVA revealed that children in the cross-examination condition made nearly 10 times as many changes (M= 2.42, SE= .15) as children who received a repeated direct examination interview (M= .26, SE= .06), F(1, 177) = 203.28, p < .0001, f= 1.07.

The three-way between-subjects ANOVA also revealed main effects of age and delay, qualified by a significant Age × Delay interaction, F(1, 177) = 7.89, p= .006, f= .21. At the short delay, older (M= 1.04, SE= .21) and younger children (M= 1.24, SE= .24) made similar numbers of changes, t(88) = .65, p= .52, d= .14, power = .75. At the long delay, however, younger children (M= 2.20, SE= .24) made more changes than older children (M= 1.11, SE= .20), t(93) = 3.47, p= .001, d= .72. Delay did not influence the number of changes that older children made, t(101) =−.24, p= .81, d= .05, power = .80, but younger children made more changes after a long delay than after a short delay, t(80) =−2.81, p= .006, d= .44.

Because children's response changes could be directed towards or away from the truth, our next step was to examine the changes as a function of original response accuracy. The proportion of correct direct examination responses changed was examined in a 2 (Age) × 2 (Delay) × 2 (Second Interview Format) between-subjects ANOVA. As expected, given that children in the cross-examination condition changed nearly 10 times as many of their direct examination responses as children in the repeated direct examination condition, a main effect of second interview format was revealed. Children in the cross-examination condition changed a greater proportion of their correct responses (M= .60, SE= .04) than children in the repeated direct examination condition (M= .06, SE= .02), F(1, 177) = 201.69, p < .0001, f= 1.07. Significant main effects of age and delay were qualified by an Age x Delay interaction, F(1, 177) = 10.07, p= .002, f= .24. At the short delay, younger (M= .30, SE= .06) and older (M= .28, SE= .05) children changed a similar proportion of correct responses, t(88) = .27, p= .79, d= .06, power = .75. At the long delay, however, younger children (M= .53, SE= .06) changed a greater proportion of correct responses than older children (M= .26, SE= .05), t(93) = 3.50, p= .001, d= .73. The proportion of correct responses changed by older children did not differ as a function of delay, t(101) = .23, p= .82, d= .05, power = .80; younger children changed a greater proportion of correct responses after a long delay than after a short delay, t(80) =−2.72, p= .008, d= .42.

Of course, children could also change incorrect direct examination responses during the second interview. The proportion of incorrect direct examination responses that children changed was therefore submitted to a similar ANOVA. Children in the cross-examination condition changed a greater proportion of incorrect responses (M= .77, SE= .08) than children in the repeated direct examination condition (M= .26, SE= .10), F(1, 37) = 10.89, p= .002, f= .54. No other significant main effects nor interactions were revealed (non-significant Fs ≤ 3.49, ps = .07–.86, fs = .03–.31, power = .11–.36), although there was a trend for younger children (M= .65, SE= .09) to change a greater proportion of incorrect responses than older children (M= .36, SE= .13), F(1, 37) = 3.49, p= .07, f= .31, power = .33. In conducting these analyses, we were only able to consider children who had made errors in their direct examination reports (n= 45); as such, power values were low.

Were incorrect responses more likely to be changed than correct ones? To answer this question, separate 2 (Direct Examination Response: inaccurate, accurate) x 2 (Second Interview Response: change, no change) chi-square contingency tests were conducted for younger and older children, as a function of delay and second interview format. Given that jurors may look unfavourably on witnesses who lack confidence (Brennan & Brennan, 1988; Brewer & Burke, 2002; Park, 2003), I don't know responses during the direct examination interview were considered to be inaccurate for the purpose of these analyses. Recall that there were very few of these types of responses for both 5- and 6-year-olds (n= 4, 1.2%) and 9- and 10-year-olds (n= 3, 0.7%). For younger children undergoing cross-examination, accurate responses were just as likely as inaccurate responses to be changed, regardless of whether the delay was short (χ² (1, N= 84) = .58, Fisher's exact p= .50, w= .08, power = .78) or long (χ² (1, N= 84) = .70, Fisher's exact p= .68, w= .09, power = .78). Similar findings were observed for the repeated direct examination interview (short delay, χ² (1, N= 80) = 1.20, Fisher's exact p= .33, w= .12, power = .76; long delay, χ² (1, N= 80) = .25, Fisher's exact p= .69, w= .06, power = .76). Older children showed the same pattern of findings for the cross-examination interview (short delay, χ² (1, N= 100) = .001, Fisher's exact p= 1.00, w= .003, power = .85; long delay, χ² (1, N= 108) = .09, Fisher's exact p= 1.00, w= .03, power = .85), and the repeated direct examination interview after a long delay, χ² (1, N= 108) = .03, Fisher's exact p= 1.00, w= .02, power = .87. None of the older children changed their responses when re-interviewed with the direct examination questions at the short delay. Given that children in some jurisdictions are actively encouraged to say I don't know when they are uncertain of an answer (e.g., Morgan Libeau et al., 2003), the data were reanalysed with I don't know responses coded as accurate. No changes in the direction of findings were observed. Taken together, these findings suggest that children's accurate responses were just as susceptible to change following a second interview as inaccurate responses. Neither of the second interview formats appeared to differentially target inaccurate over accurate responses.

In summary, children performed well in the direct examination interview, but made numerous changes to their original responses during the second interview, to the detriment of their accuracy. While age and delay played a role in the number of response changes that children made and their overall accuracy, the format of the second interview exerted the greatest effect. The cross-examination interview was considerably more detrimental to children's reports than the repeated direct examination interview.

Discussion

Every year, thousands of young witnesses around the world give evidence in court (New Zealand Ministry of Justice, 2010; Plotnikoff & Woolfson, 2009). Given the enormity of legal outcomes for victims and alleged offenders alike, obtaining accurate testimony from witnesses is crucial. Unfortunately, the questioning style used during cross-examination has repeatedly been shown to be detrimental to children's accuracy (Zajac & Hayne, 2003, 2006; Zajac et al., 2009). We investigated two possible explanations for this phenomenon: questioning style and repeated interviewing. To do this, children's accuracy during cross-examination was compared to that of a second group of children who were interviewed with closed questions only. In this way, we could examine whether children's accuracy decreased following a second interview, irrespective of the sorts of questions they were asked.

Direct examination performance

Children's responses to the direct examination questions were generally accurate. Consistent with other research, however, older children were more accurate than younger children (e.g., Bjorklund, Bjorklund, Douglas-Brown, & Cassel, 1998; Brady, Poole, Warren, & Jones, 1999; Memon & Vartoukian, 1996). Considerable changes in children's cognitive abilities across development are likely to account for these age differences. These changes include an increase in storage capacity (Schneider & Pressley, 1989), more efficient use of encoding and retrieval strategies (Bjorklund, Coyle, & Gaultney, 1992; Kobasigawa, 1974; Schneider & Pressley, 1989), faster information processing (Dempster, 1981; Schneider & Pressley, 1989), and a greater knowledge base (Bjorklund, 1987); all of these factors are likely to facilitate children's memory for personally experienced events.

Regardless of age, children answered questions about false events more accurately than those about true events, although the effect size was relatively small. Unfortunately, we could not counterbalance true and false events in a memory event run by an external agency, so it is possible that the nature of our event made errors of omission more likely than errors of commission. For example, getting a photo taken may not have been perceived as salient, offering a potential explanation for lower rates of recall (Howe, 1997; Howe, Courage, & Peterson, 1994; Peterson, 1999; Peterson et al., 2001). Furthermore, children were told not to tell anyone about being spun around by the police officer, which may also have resulted in fewer children disclosing this event in subsequent interviews (Bussey, Lee, & Grimbeek, 1993; Wilson, Powell, Raju, & Romeo, 2004). Alternatively, it may be that children's rejection of the interviewer's questions about false events was due to their using memorability-based strategies. That is, children may have perceived that the false events (e.g., trying on handcuffs) would have been memorable had they occurred. Consequently, when unable to recall event details, children would easily be able to infer that the events did not happen (Ghetti, 2008; Ghetti & Alexander, 2004).

Second interview performance: The effect of interview format

Children's accuracy decreased in the second interview, regardless of format. That is, even when children were merely asked the direct examination questions again in their second interview, there was a small decrease in accuracy. This finding differs from those of Quas and colleagues (e.g., Quas, Malloy, Melinder, Goodman, & D’Mello, 2007; Quas & Schaaf, 2002; see Goodman & Quas, 2008, for a review), which suggest that repeating specific questions about true and false aspects of an experienced event across interviews does not decrease the accuracy of children's reports. While we cannot be certain why our findings differed, it may be that social factors played a role. Unlike Quas and colleagues’ participants, children in our study watched a video of their first interview before being questioned a second time. As a result, it is possible that children were confused as to why the interviewer was asking the same questions again, and therefore interpreted the repeated questions as a signal to change their previous responses (Donaldson, 1982; Hartwig & Wilson, 2002; Lyons, 2002; Rose & Blank, 1974; Siegal et al., 1988).

Although repeated interviewing appeared to play a small role in children's second interview performance, children's event reports were most adversely affected by the cross-examination interview. Children who were cross-examined made more changes to their direct examination responses, changed a greater proportion of their correct responses, and obtained lower overall accuracy scores than those who were merely asked the direct examination questions again. These findings support the notion that the nature of the interview, rather than the fact that children are interviewed more than once, exerts the primary influence on children's accuracy. Because children in the cross-examination interview experienced a more prolonged interview, is it possible that they simply had greater opportunity to change their initial responses than those in the repeated direct examination condition? We cannot refute this possibility, as the number of questions asked was not standardized across the two interview formats. While we acknowledge this as a limitation of our study, we do not believe this discrepancy to be the primary explanation for findings, given recent repeated questioning research suggesting that the relation between greater opportunity to change responses and decreasing accuracy is not necessarily linear (Krahenbuhl, Blades, & Eiser, 2009). A more likely explanation is that children's performance is most vulnerable when a second interview comprises leading, linguistically complex, and confrontational questions; the very questions that are commonplace during cross-examination.

What makes cross-examination questions so difficult for children to answer accurately? We propose three main reasons. First, cross-examination represents a highly unique conversational situation for any witness, but perhaps especially for children, whose knowledge of conversational rules develops over time and with experience (Lamb & Brown, 2006; Walker, 1993). Children typically assume that adults are genuine conversation partners (Snyder & Lindstedt, 1995); their understanding that someone might be intentionally deceptive in conversation does not develop until later in middle childhood (Demorest, Meyer, Phelps, Gardner, & Winner, 1984). Consequently, children may fail to understand that they may be purposefully misled during cross-examination.

Second, because adults often know the answers to the questions that they ask children (Lamb & Brown, 2006), the highly suggestive nature of many cross-examination questions may lead children to assume that the interviewer is knowledgeable about the event in question (Snyder & Lindstedt, 1995), especially given lawyers’ apparent authority (Ackerman, 1983; Ceci, Ross, & Toglia, 1987; Lepore & Sesco, 1994; Thompson, Clarke-Stewart, & Lepore, 1997; Tobey & Goodman, 1992). Children's ability to understand what the cross-examining lawyer could possibly know about the alleged event, and how that information was acquired, is unlikely to be fully developed (Perner & Wimmer, 1985; Sodian & Wimmer, 1987; Wimmer, Hogrefe, & Perner, 1988).

Finally, the frequent use of leading questions during cross-examination creates an atmosphere in which a particular response is expected. Children are highly likely to answer these types of questions, often incorrectly, even when told that it is okay to say I don't know (Cassel et al., 1996). Child witnesses rarely express uncertainty or ask for clarification if they do not understand a question (Zajac et al., 2003), and children often speculate and provide responses to questions when they could not possibly know the answer (Poole & White, 1993), or when questions are nonsensical (Hughes & Grieve, 1980; Waterman, Blades, & Spencer, 2000). When children feel compelled to answer every question, their responses may seriously undermine their accuracy.

Although the focus of this study has been children, it is likely that other vulnerable witnesses are also susceptible to the negative effects of cross-examination questioning. Several studies, for example, have shown that adult rape complainants face lengthy, linguistically challenging, and coercive cross-examinations (Brereton, 1997; Danet & Bogoch, 1980; Kebbell, Deprez, & Wagstaff, 2003). Furthermore, like children, adult sexual assault complainants undergoing cross-examination make a significant number of changes to their earlier testimony (Zajac & Cannan, 2009). Recent research using a student sample suggests that the accuracy of adults’ reports also suffers under cross-examination questioning (Valentine & Maras, 2011).

Second interview performance: The effects of age and delay

Age and delay interacted to affect performance in the second interview, irrespective of interview format. Specifically, at the long delay, but not the short delay, younger children made more changes to their earlier reports, changed a greater proportion of correct responses, and displayed lower overall accuracy levels than older children. These findings are consistent with research showing that young children's recollections decay faster than those of older children (Baker-Ward, Gordon, Ornstein, Larus, & Clubb, 1993; Brainerd & Reyna, 1995; Flin, Boon, Knox, & Bull, 1992; Ornstein, Gordon, & Larus, 1992).

One theory that might account for these findings is the fuzzy trace theory of cognitive development (Brainerd & Reyna, 1990), which proposes that encoded information is stored on a fuzzy-to-verbatim continuum. Fuzzy traces are vague representations of encoded material, which degrade relatively slowly; verbatim traces are exact representations of encoded material, which degrade relatively quickly. Individuals are assumed to have a preference for processing fuzzy traces because this requires less effort. Younger children, however, appear to rely more heavily on verbatim traces (Brainerd & Reyna, 1990), perhaps accounting for their poorer performance at the long delay.

Practical implications

Our findings suggest that reform of the cross-examination process may be necessary to facilitate the accuracy of children's testimony. Such reform, however, would be highly controversial, with some lawyers (Davies et al., 1997) and judges (Cashmore & Bussey, 1996), indicating that they would resist reform initiatives. A central tenet of the adversarial system is the right to challenge a witness's testimony, and cross-examination has traditionally been promoted as an effective means of testing the veracity of a witness's report (Cashmore & Bussey, 1996; Eichelbaum, 1989).

If we were to change the format of cross-examination, what type of interview would be most effective? To stop children from changing correct responses under cross-examination, we know that we should ask open-ended questions (Baker-Ward et al., 1990; Leichtman & Ceci, 1995; Poole & White, 1991; Principe et al., 2000; Quas & Schaaf, 2002) and avoid suggestion (Ceci et al., 1994; Leichtman & Ceci, 1995; Gobbo, 2000; Poole & Lindsay, 1995; Quas & Schaaf, 2002) and possibly even yes/no questions (Brady et al., 1999; Dent & Stephenson, 1979; Donaldson, 1982; Memon & Vartoukian, 1996; Poole & White, 1991; Siegal et al., 1988, but see Principe et al., 2000; Quas & Schaaf, 2002). It is important to acknowledge, however, that children's primary accounts of their past experiences will not always be accurate. In this sense, the ideal cross-examination protocol would maximize favourable changes (i.e., changes directed towards the truth) while minimizing unfavourable changes (i.e., changes directed away from the truth). At present, no such interview protocol exists (Howie et al., 2004). Unfortunately, the questions that are problematic for children to answer correctly are the same questions that are most likely to effectively challenge credibility (Davies et al., 1997; Davies & Seymour, 1998). Recall that in the present study, even repeating very simple closed questions resulted in a decrease in accuracy.

A different approach to facilitating accuracy centres on preparing children for cross-examination. Teaching children to respond I don't know where appropriate or warning children about the difficult nature of a forensic interview (Beuscher & Roebers, 2005; Gee, Gregory, & Pipe, 1999; Memon & Vartoukian, 1996; Righarts & Zajac, 2011; Saywitz & Moan-Hardie, 1994) has met with minimal success. Greater promise has been observed with training interventions that teach children how to identify and respond to incomprehensible questions (Peters & Nunez, 1999) or that provide children with the opportunity to practice difficult questions and gain feedback about their responses (Gee et al., 1999; Righarts & Zajac, 2011). Designing effective preparation programmes to maximize children's accuracy remains an important area for future research.

Regardless of the questions asked or the degree of preparation conducted, our findings suggest that cross-examination should occur sooner rather than later, particularly for younger children. Current legislation in many countries allows for cross-examination to occur at the same time that children's direct evidence is pre-recorded (i.e., at the time that an allegation is made to police), but this rarely occurs (Gupta, 1994; New Zealand Ministry of Justice, 2010; Pipe & Henaghan, 1996). Reducing the delay to cross-examination may also confer benefits outside of improving accuracy, such as allowing victims earlier access to psychological support.

Concluding remarks

Our findings provide empirical support for the long-held hypothesis that questioning style is the most potent factor affecting children's cross-examination responses. Furthermore, for younger children, this effect appears to be compounded by longer delays to cross-examination. These results lend weight to the argument that policy makers need to address the difficulties inherent in the cross-examination interview. Any reform, however, must strike a delicate balance between accommodating vulnerable witnesses and the defendant's right to a fair trial.

Footnotes

• 1

  Cohen's (1988) measures of effect sizes have been included for all statistical analyses involving t-tests, ANOVAs, and chi-square tests (d, f, and w, respectively). Power values to detect medium effect sizes (d = .50, f = .25, w = .30) have been included for all non-significant findings. These power values indicate the probability of rejecting the null hypothesis should a medium sized (or larger) effect exist.

Acknowledgements

Funding for this study was provided by the Marsden Fund Council (from Government funding administered by the Royal Society of New Zealand) and by the University of Otago. The authors thank Bridget Irvine for her contribution to coding the data. The authors also acknowledge the invaluable assistance of Senior Constable Paul Howard, and the children who participated in this research, along with their parents and teachers.

Appendices

Appendix A

Four sets of sample cross-examination questions

Target activity 1 (photo) initially inaccurate child

• 1
  In the video, you said that you didn’t get your photo taken, didn’t you?
• 2
  Can you tell me why you think that?
• 3
  You got your fingerprints taken, didn’t you?
• 4
  Can you remember what you did straight after that?
• 5
  You lined up at some point, didn’t you?
• 6
  Did you do maths yesterday at school?
• 7
  Did you like going to the police station?
• 8
  Are you sure you didn’t get your photo taken?
• 9
  I think maybe you just forgot about getting your photo taken. That’s what happened, isn’t it?
• 10
  That might be what happened though, don’t you think?

Target activity 2 (handcuffs) initially inaccurate child

• 1
  You said on the video that you got to try on handcuffs, didn’t you?
• 2
  Who was there when that happened?
• 3
  The handcuffs would have been pretty heavy, were they?
• 4
  Did you have them on for a long time or a short time?
• 5
  Did they feel funny?
• 6
  Did someone drive you to school this morning?
• 7
  And the policeman put them on you?
• 8
  Are you sure you got to try on handcuffs?
• 9
  I think maybe your friends got to try on handcuffs, but you didn't. That's what really happened isn’t it?
• 10
  That might have happened though, mightn’t it?

Target activity 3 (bike) initially accurate child

• 1
  You said in the video that you didn’t see a lady come in about her child's bike, didn’t you?
• 2
  Can you tell me how you know that?
• 3
  You saw where she could have come in though, didn’t you?
• 4
  Was it right at the start?
• 5
  Had you been to the police station before?
• 6
  Can you tell me what you’ve been doing at school this morning?
• 7
  The police station was really big wasn’t it?
• 8
  Are you sure you didn’t see a lady come in about her child's bike?
• 9
  I think maybe you did see her, but you can’t quite remember it. That’t what happened isn't it?
• 10
  But that might be the case though, don’t you think?

Target activity 4 (spinning) initially accurate child

• 1
  You said in the video that the policeman spun you around, didn’t you?
• 2
  How do you know that?
• 3
  Was everyone else there?
• 4
  Tell me a bit about what happened when he spun you around.
• 5
  Did you get your fingerprints taken at the police station?
• 6
  And where did the policeman spin you around?
• 7
  Where was everyone standing?
• 8
  And are you sure the policeman spun you around?
• 9
  If your teacher told me that it didn’t happen, she would be right about that, wouldn’t she?
• 10
  But she might be right about that, mightn’t she?

Appendix B

Examples of 0-, 1-, and 2-point cross-examination interview responses

0-Point response (child does not acquiesce to suggestion)

• Interviewer: Are you sure that the policeman spun you around?

• Child: Yes.

• Interviewer: I think maybe you didn’t get spun around. I think someone told you to say that. That’s what really happened isn’t it?

• Child: No.

• Interviewer: That might be the case though, mightn't it?

• Child: No.

1-Point response (child admits that suggestion is possible)

• Interviewer: Are you sure you didn’t get to try on handcuffs?

• Child: Yes

• Interviewer: I think maybe you just forgot about getting to try on the handcuffs. That’s what happened, isn’t it?

• Child: No response.

• Interviewer: But that might be what happened though, don’t you think?

• Child: I don’t know.

2-Point response (child acquiesces with suggestion)

• Interviewer: Are you sure that you got your photo taken?

• Child: Yes.

• Interviewer: If your teacher told me that it didn’t happen, she would be right about that wouldn't she?

• Child: Yes.