Memon, A. and Vartoukian, R. (1996) The Effects of Repeated Questioning on Young Children's Eyewitness Testimony British Journal of Psychology, 87, 403-415

The Effects of Repeated Questioning on Young Children's Eyewitness Testimony.

Amina Memon & Rita Vartoukian
University of Southampton

Dr. Amina Memon (until 8/1998):
School of Human Development
University of Texas at Dallas,
Richardson, Texas, U.S.A., 75083-0688
email: amemon@utdallas.edu

After 8/1998:

Department of Psychology
University of Southampton
Southampton
S017 1BJ

Abstract

The aim of the present study was to explore the conditions under which repeated questions would influence memory performance. Children of five and seven years of age witnessed a staged-event and were then individually interviewed with a free recall test and closed and open-form questions, some of which were repeated in the interview. Some children were warned that questions may be repeated. The older children were more accurate on both open and closed question forms than the younger children. In both groups recall improved upon second questioning with open questions whereas accuracy of responses deteriorated somewhat upon repetition of closed questions. On the basis of this data it is concluded that if closed questions are repeated in a witness interview it may lead the witness to incorrectly assume that their first response was incorrect; however the findings support the use of repeated questioning as a probe for more information to open-ended questions.

Keywords: eyewitness, repetition, warning, questions, interview

The Effects of Repeated Questioning on Young Children's Eyewitness Testimony

Children can be questioned repeatedly about an event for a number of different reasons, the most obvious one being to elicit more detailed responses. This is important in everyday remembering (Fivush and Hammond, 1990) as well as in child witness interviews (Poole and White, 1991; Poole and White, 1993). It is well established that children provide less accurate and complete responses to specific questions as compared to adults (List, 1986, Cassel and Bjorklund, 1993) and this has raised concerns about the reliability of responses in real life investigations (Raskin and Yuille, 1989). The present paper addresses some empirical questions pertaining to the effects of repeating questions in child witness interviews.

Repeated testing has been associated with both positive and negative effects on memory reporting. Laboratory studies of memory have documented two main positive effects of repeat testing using a variety of stimulus materials from words to pictures. The first is reminiscence, or the recall of material that did not appear in an earlier test (Richardson, 1985). When this `new' information exceeds the amount of information that is forgotten, the `hypermnesia' effect has occurred (Erdelyi and Stein, 1981). A second advantage of repeated testing is the possibility that repeated testing will `inoculate' against forgetting (Warren and Lane, 1995). Both hypermnesia and inoculation effects presumably occur because repetition increases trace strength for recalled items, for example, by "redintegration," which is the refurbishing of traces due to feature activation (Brainerd, Reyna, Howe and Kingma, 1990; Brainerd and Reyna, 1990; Brainerd and Ornstein, 1991).

While the basic memory literature suggests repeated testing can have beneficial effects, there is a concern amongst legal professionals as to the effects of repeatedly interviewing an eyewitness, particularly a child witness. For example, a juror may wonder why a fact was not included in an earlier testimony and may think it was planted in 'witness' memory (see Myers, 1993; Memon and Bartlett, in press; for recent reviews of the implications of repeat questioning of child witnesses). The perceived credibility of testimony may also be influenced by consistency in responses to repeated questions (Drew, 1990) despite the evidence that consistency and accuracy are unrelated (Fisher and Cutler, 1992; Poole and White, 1993). Most importantly, it has been assumed that answers to repeated questions will be less accurate than original answers because witnesses react to the social pressure of repeated questions by offering speculative responses (Moston, 1990).

Several studies have explored hypermnesia for details in an eyewitness context. The first of these found that the number of correct details recalled about a taped violent event increases with repeated recall attempts (there was a relatively small increase in errors). The conclusion from this study was that witnesses will not necessarily provide a complete account on their first attempt at recall and later additions may not be less accurate or contaminated by information interviewers provide (Scrivner and Safer, 1988).

Poole and White (1995) arrive at a similar conclusion following a comprehensive review of studies that have examined the performance of child and adult witnesses across repeated questions. They found little evidence that repetition degraded accuracy when witnesses were asked open-ended questions. The only exception was a study which showed that compared to adults and older children, four-year olds were more likely to shift answers to repeated yes/no questions (Poole and White, 1991).

In contrast, several studies have found degradations in accuracy across repeated questions (See Moston, 1990 for a review). In a study conducted by Moston (1987), children aged 6 and 10 years were interviewed about a staged event. Of the eight questions that were repeated, four were neutral and four were misleading. There were significantly more correct responses to the four neutral questions on first questioning as compared to second questioning. Moston concludes that interviewers should be prepared to accept that a child's first answer is probably the best they can give. There are a number of reasons to suggest that this conclusion is premature. Firstly it was based upon responses to four questions only, each of which was a closed question. Children perform less well with these types of questions (Dent and Stephenson, 1979, Davies, Tarrant and Flin, 1989, Lamb, Hershkowitz, Sternberg, Esplin, Hovav, Manor, and Yudilvitch, in press). Secondly, there is an indication in the Moston study that the younger children (six-year olds) were more likely to change their responses in the face of repeated questioning than the older (10-year old) children.

In response to these discrepant conclusions, Poole and White (1993) argued that the effects of question repetition will depend on witnesses' interpretation of the task. Open-ended questions are repeated in daily interactions (e.g., what did you say you did at school today?) and children as well as adults respond to such requests by repeating their original story. Therefore, repeating open-ended questions appears to be an innocuous procedure and one that may lead witnesses to recall a useful piece of additional information. A question that is repeated may be misinterpreted as a desire for additional or alternative information (Memon, Cronin, Eaves and Bull, 1993). This may increase inaccuracies, particularly if the question is misleading (Bruck, Barr, Ceci and Francouer, in press; Leichtman and Ceci, in press). There are many examples of this in the developmental literature (Rose and Blank , 1974; Siegal, Walters and Dinwiddy, 1988).

Dent and Stephenson (1979) and Bull (1992) have suggested that one reason why children (and adults) report erroneous information in recall tests is because social convention stresses giving an answer to a question even where respondents are really not sure of the answer. It is predicted that the following variables will reduce the effect of this social convention on children's responses: (a) telling children that they can answer a question with 'I don't know' or warning them that they may be tricked; (b) telling children that a repeat question does not mean that the previous answer was wrong.

Moston (1987) examined the effect on responses of telling children that they can say 'I don't know' and found that those given this instruction were marginally more likely to use it. Saywitz, Moan and Lamplear (1991) found that a similar warning had quite a different effect. In their study 7-year-olds participated in a number of activities which they were subsequently interviewed about. Before being tested, one group of children received training in strategies they could use to resist misleading questions which included the use of `I don't know.' The results showed the trained group produced 26% fewer errors in response to misleading questions. Contrary to expectations, the trained group also gave fewer correct responses as they tended to respond `I don't know' more frequently than the control group!

The effect of warning children that they may be tricked has also been examined by Warren Hulse-Trotter and Tubbs (1991). The children in this study (aged 6 and 11 years) were read a story and then asked questions about it, including misleading questions. The children were then told they had not performed well and to answer all the questions they were asked again. Younger children tended to shift their answers to non-leading questions following this negative feedback, suggesting that age may predict warning effectiveness.

There is one further variable that may mediate the impact of question repetition and that is question salience. The salience of a question has been defined as its `centrality' to the event and this variable has occasionally been examined in eyewitness studies on emotion and memory. Christianson (1992) suggests that central and peripheral information can be distinguished in terms of how well they are remembered within a scenario. One hypothesis in this area is that central details tend to be connected with the source of emotional arousal and our attentional resources will be directed towards them (as in a focus on a weapon) leaving limited resources for information that is spatially peripheral or unrelated to the source of arousal (Ellis and Ashbrook, 1988). An alternative explanation for the enhanced recall of central details is post-stimulus elaboration (e.g. more reference to, and opinions about, central actions and characters in an event; Christianson, Loftus, Hoffman and Loftus, 1991). The effects of centrality of detail or question salience on children's recall are inconclusive. In some studies children show superior recall of central details (Davies et al, 1989) however, other studies have found no differences (Warren and Lane, 1995; King and Yuille, 1987). King and Yuille (1987), for example, report that many peripheral features such as footwear were as salient for children (8-13 year olds) as were central details.

Finally, there is evidence from several studies conducted with children between the ages of 6 and 11 years that the type of information being recalled is critical. It has been consistently shown that descriptions about persons are less accurately reported as compared to details about actions (Dent and Stephenson, 1979; Davies et al, 1989; Cassel and Bjorklund, 1993, Memon, Wark, Koehnken and Bull, 1995; Milne, Bull, Koehnken and Memon, 1995). In the current study, the event details were categorised as descriptors and actions in order to compare the performance of younger and older children.

The aim of the present study was to explore the conditions under which repeated questions would influence memory performance. If the explanation that children actively change their answers because they do not understand that repetition is a check for consistency is correct, then specific questions will be more affected by repeated questioning than open-ended questions. This is because the former have a limited number of possible answers while open-ended questions are more commonly repeated in normal conversation as probes for more information (Poole and White, 1993). A pre-interview warning may, however, reduce degradations in accuracy to repeated questions. To our knowledge there are no published studies which report such a detailed analysis of question repetition effects.

Method

Participants

Fourty-four children from a local Primary School, ranging in age from 5 to 8 years, participated in the investigation. Twenty-four of the children were from Year 1 (mean age = five years and 11 months; 10 females and 14 males). Twenty children were from Year 3 (mean age = 7 years and 4 months; 11 females and 9 males).

The Witnessed Event

The event began with Actor 1 (male) entering the classroom with his flute and being introduced by the teacher as a student who had come to play a tune to the children. During the performance Actor 2 (female) knocked on the door and stormed in carrying a bag and umbrella and started arguing with Actor 1 about why the latter had not waited at the bus-stop for her. Actor 2 then decided she needed a drink of orange juice and something to eat. After further disagreements, Actor 2 recited a poem while Actor 1 played the flute. They then left. The event lasted about 4.5 minutes and was video-recorded.

Selection of Central versus Peripheral Details

Eight students from Southampton University viewed a recording of the staged event to be used in the main study and wrote down a list of as many things that they could remember about the film, including actions and descriptions. Items listed by at least four subjects were taken to be the core items and those mentioned by less than four subjects were used as the peripheral items, including items not mentioned by all of them but known by experimenters to be on the video. Ten of each of these categories were selected to make up the interview questions (see Appendix 1).

Interview Procedure

About 5 minutes after the actors left, the children were taken individually to a room to be questioned by a female interviewer. The interview (which was audio-taped) began with a free recall (i.e. "tell me as much as you can remember about what happened in class this morning."). Each child was then asked some questions about the event. It was explained to them that if they did not know an answer to a question, it did not matter and that they should not worry if they could not remember something, but that they should not make up any answers. Half the children were given the additional warning that some questions may be repeated. Each child was asked a total of 30 questions about the event, 10 of which were direct repetitions. Each interview lasted between 6 to 10 minutes.

Data Coding

Free recall narratives and responses to open-ended questions were scored for the number of accurate and inaccurate syntactic units (SUs) in each response (similar to the technique used by Poole and White, 1991). Examples of SUs are words describing a person (he/she), an action (drank) and an object (orange juice). In other words: 'she drank orange juice,' includes 3 SUs. This scoring technique was found to be more useful than one based on complete propositions because children are particularly fragmentary in giving their responses. Judgements about accuracy were based on a detailed description of the event taken from the filmed event, and a set of rules for scoring particularly difficult information. Information repeated within the free recall session was only counted once, and where statements were retracted, e.g. 'he wore blue...no black trousers', the second statement was taken as the final one. Also, only the individual SUs that were inaccurate were coded as incorrect e.g. the statement 'she ate some crisps', would be coded as 2 accurate SUs (she did have some crisps) and 1 inaccurate SU (she did not eat them). Occasionally the pronoun 'he' or 'she' was used inappropriately when correctly describing actions performed, so in these cases, only the pronoun was scored as incorrect.

The free recall protocols were coded for (i) correct details (classified into central, peripheral; action and descriptors) (ii) incorrect details and (iii) accuracy: The proportion of correct information as a function of total reported information. Responses to closed questions were scored as 'correct' or `incorrect' with an appropriate code assigned. (Less than 5% of the responses fell into the `don't know' category and they were not included in the analysis). In three cases there were missing data in cells as a result of non-responses to questions. These cases were not included in the final analysis of data.

The responses of eight subjects (18% of the sample) were scored by two coders to assess intercoder reliability for the two dependent measures: Accurate information and inaccurate information. They were found to have 95% and 96% agreement respectively.

Results

The results section is organised as follows. The free recall data are presented first followed by the question data. The question data presents an analysis of total correct responses to open questions and a comparison of accuracy rates (proportion correct) to open and closed questions.

Free Recall

On average the 5-year-olds recalled only half as much correct information as the 7-year-olds (means = 14.50 and 33.5 respectively, t (25) =-3.07, p<.01). However, there were no significant differences in accuracy rates: 94.1% of the five-year-olds total recall was accurate compared to 97.5% for the seven-year-olds.

The recalled details were then subdivided and analysed according to whether they comprised action details or descriptive information (objects/persons). As is typically found in studies of children's memory for staged events, correct recall consisted of more action statements both for 5-year olds (82.4% action details vs. 17.6% descriptive details; t (24) = 4.23, p<.01) and 7-year-olds (79.3% action details vs. 20.6% descriptive details; t (20)= 4.84, p<.01). In terms of proportion accurate, there were no significant differences in accuracy for action versus descriptive information (means = 95.1% and 94.0% respectively). This is interesting as it suggests that although children report fewer descriptive details their memory for descriptions is just as accurate as their memory for actions.

Question data

Analyses of specific questions focused on open and closed questions that were presented twice during the interview, with other questions only serving as filler items. It is not the aim of this study to show differences in the amount of information recalled across open and closed questions, nor to compare accuracy rates across these two question forms, because the amount and accuracy of information could be due solely to the difficulty of specific questions we selected. However, what is of interest is how children's responses to these question forms vary as a function of repetition and warning.

Open ended questions

(i) Total correct information

The total number of accurate SUs were entered into an Analysis of Variance (ANOVA) with age (5 or 7 years) and warning condition (warned and unwarned) as between subject factors, and question salience (central vs. peripheral), information type (action vs. object/person descriptor), and trial (first or second presentation of the question) as within-subject variables[1]. As expected the 7-year-olds reported more details than the 5-year-olds (means = 4.04 and 2.57 respectively, F(1,40) = 7.10, p <.02) and subjects reported more on the second trial than on the first, means = 3.43 and 3.04 respectively, (F (1, 40) = 5.42, p <.05). There was also a main effect of question salience with more accurate recall of central items, means = 4.72 and 1.76, F (1,40)= 72.75, p<.01). Finally, there was a question salience by question type interaction, (F(1,40)= 14.97, p <.01) which suggests that central objects are more likely to be correctly recalled as compared to central actions (means = 5.69 and 3.81 respectively) but that the opposite is the case for peripheral objects and actions (means = 1.16 and 2.35 respectively).

(ii) Accuracy (proportion correct)

Table 1 showing mean proportion correct for open questions



5 YEARS

7 YEARS


1st
2nd
1st
2nd
Central
warned
.71
.72
.66
.70

unwarned
.66
.70
.85
.92
Peripheral
warned
.25
.31
.32
.57

unwarned
.37
.32
.71
.79

Accuracy rates on first and second presentations were compared by entering the proportion of total SUs that were accurate into an ANOVA with age and warning condition as between subject factors and question salience, information type and trial as within subject factors. There was a significant effect of age; the 7-year-olds were more accurate than the 5-year-olds, (F (1,38) = 10.01, p <.01) as shown in Table 1[2]. There was also a significant effect of question salience (F(1,38) = 27.98, p <.01) which showed that central details were more accurately recalled than peripheral details. There was a significant interaction between question salience and question type (F (1,38) = 4.86, p<.05). Simple main effects tests showed no difference between action and object questions that were central (means =.69 and .78, F (1,38) = 1.27, p>.05), but the children were more accurate on peripheral questions about actions than objects (means = .52 and .37, F (1,38) = 3.93, p=0.05).

There was a warning by age interaction (F (1,40) = 4.41, p<.05). Simple main effects tests showed that the 5-year-olds were generally unaffected by the warning, (F (1,38) = 0.02, p>.05), but warning resulted in less accurate reports for the seven-year-olds (F (1,38) = 8.82, p<.01).

The effect of trial on accuracy of responses to open questions fell short of statistical significance (F (1,38) = 3.51, p=.06). The means however were in the predicted direction and suggested an increase in proportion accurate upon second testing (see Table 1).

Closed questions

Table 2 showing mean proportion correct for closed questions



5 YEARS

7 YEARS


1st
2nd
1st
2nd
Central
warned
.56
.53
.76
.67

unwarned
.56
.48
.76
.72
Peripheral
warned
.57
.54
.83
.64

unwarned
.52
.53
.68
.65

Since closed questions generally elicit one word responses, we were concerned only with accuracy rates and not the overall amount of information reported. The data concerning closed questions were analysed by entering the number of correct responses into an ANOVA with age and warning as between-subject factors and question salience, question type, and trial as within-subject factors. As for open-ended questions, there was a significant effect of age, with the 5-year olds reporting less accurately than the seven-year-olds (F (1,40) = 10.13, p <.01). While there was no effect of information type on accuracy of open questions, there was a main effect of information type on closed questions (F (1,40 ) = 5.97, p <.05). There responses to qyestiona about actions tended to be more accurate than responses to questions about descriptions (see Table 2). There was a significant interaction between information type and age (F (1,40) = 6.73, p<.05). Analysis of simple effects showed that there were small differences between younger and older children in the responses to questions about actions, but when it came to object/person descriptions, the younger children fared less well (means = .25 and .55, F (1,40) = 13.96, p <.01). This is consistent with the findings of previous researchers (Cassel and Bjorklund, 1993; Davies et al, 1989).

Again the effect of trial fell short of statistical significance (F (1,40) = 3.04, p=.08) with the means suggesting a fall in the accuracy of responses to closed questions upon repetition (see Table 2).

Discussion

Repetition of open questions appears to have no harmful effects; children increased the total correct information they provided without decreasing the accuracy of responses. There was a non-significant trend for an increase in accuracy of responses to open questions that were repeated. These results are consistent with previous work on question repetition (e.g. Poole and White, 1991) and studies that report a hypermnesia effect following repeated testing (Payne, 1987; Scrivner and Safer, 1988). In contrast, children tended to be less accurate in response to closed questions that were repeated.

It is clear that cognitive and social factors may interact in interesting ways to influence responses to repeated questions and there is evidence to support this interpretation in the present study. The increases in amount of correct recall with repetition of open questions is compatible with notions of retrieval practice (Roediger and Payne, 1982) and trace redintegration (Brainerd et al, 1990). While responses to closed questions that are repeated may reflect social pressures to produce an appropriate response as is sometimes found when children are presented with misleading suggestions (Bruck et al, in press). The findings from the present study may guide future research on the relative contribution of cognitive and social mechanisms which determine responses to repeated questions and the conditions under which the effects may be modified.

One of the limitations of the present study was that salience of questions was determined by adults viewing a video-tape of the event that children witnessed. A more effective procedure for selecting the core and peripheral items of an event needs to be developed. Question saliency could be looked at using scaled items of saliency instead of two extreme groups (Cassel and Bjorklund, 1993). An innovative procedure is currently being developed in order to compare eyewitness memory for schema consistent and inconsistent details (Milne et al, 1995; Wark, Memon, Koehnken and Bull, 1995). The procedure involves interviewing children of various ages about the to-be-remembered event prior to conducting a memory study (with other children). This is a useful way of becoming familiar with children's vocabulary and the salience of different types of information (as a function of age, experience and so on). An alternative procedure would be to define centrality in terms of memorability and use the witnesses free recall accounts to define an item as a central or peripheral one.

Contrary to the expectation that clarifying the purpose of repetition to young children would influence responses, there was no effect of warning. There could have been a number of reasons for this. Firstly, the nature of the warning was very general and only given once (prior to free recall). The warning may have had a greater impact had it been reiterated prior to the questioning phase. Secondly, the warning makes little sense to the children as they are used to answering repeated questions (See Durkin, 1986 for an excellent review). Finally, it is possible that the children did not fully understand the warning. A practice session or example may have helped (cf. Memon, Cronin, Eaves and Bull, in press). Data from a recently completed study throws some light on what may be going on. Lindsay, Gonzales and Eso (1995) measured memory in a standard suggestibility paradigm, except that for half of the subjects (5-year-olds, 9-year-olds and adults) the test instructions informed them that they had received misleading suggestions and asked them not to base their memory reports on anything mentioned in this phase because that information was wrong (opposition instructions). In terms of correct responses, suggestions reduced correct report for subjects given the standard instructions. The opposition instructions eliminated this effect in free recall and reduced it in cued recall. Importantly, the youngest children benefited from the warning just as much as adults when performance was measured in terms of correct recall. However, in terms of incorrect report of suggested details, the warning reduced but did not eliminate the suggestibility effect in 9-year-olds and adults, and had no effect at all on preschoolers' incorrect reports in free and cued recall. Lindsay et al suggest that when given the standard instructions, the younger children remembered two potential answers (seen and suggested) but they typically `kept mum' and did not volunteer either, unless they were given the opposition instructions, in which case they reported the seen detail. When the younger children in the standard instruction group reported a suggested detail, they really were not aware that they were doing so they thought they were reporting what they had seen. Hence the opposition instructions increased correct reports but did not decrease incorrect reports. Clearly this line of research is important and more of it is needed to understand the effects of pre-interview instructions.

Children can be reliable witnesses as long as adults use careful questioning (Bull, 1995). Instead of using a very general warning, perhaps children could benefit more from explicit statements which explain why questions may be repeated. Future research should look into the feasibility of devising warnings better suited to inducing resistance to repeated questions particularly misleading ones. At the same time it is important to develop effective means of communicating to children the intent and purpose of such interviewing. This is compatible with the approach recommended by the British Home Office Memorandum of Good Practice (1992). Finally, the impact of repeat questions when they are part of a particular interviewing technique (e.g the Cognitive Interview) needs to be assessed.

References

Baker-Ward, L., Hess, T.M., & Flannagan, D.A. (1990). The effects of involvement on children's memory for events. Cognitive Development,5, 55-69.

Bartlett, D. and Memon, A. (in press). The role of language in the courtroom. To appear in D. Carson & R. Bull, Handbook of Psychology in Legal Contexts.Chichester: Wiley.

Brainerd, C.J. & Ornstein, P.A. (1991). Children's memory for witnessed events: the developmental backdrop. In J. Doris (Ed.), The Suggestibility of Children's Recollections. American Psychological Association: Washington, D.C. (pp.10-20).

Brainerd, C.J. & Reyna, V.F. (1990). Gist is the grist: Fuzzy-trace theory and the new intuition. Developmental Review, 10, 3-47.

Brainerd, C.J., Reyna, V.F., Howe, M.L., & Kingma, J. (1990). The development of forgetting and reminiscence.Monographs of the Society for Research in Child Development,55, (3-4 Serial No. 222).

Bruck, M., Barr, R., Ceci, S.J. & Francouer, E. (in press). I hardly cried when I got my shot! Influencing children's memories about a visit to their paediatrician Child Development.

Bull, R. (1995). Innovative techniques for the questioning of child witnesses especially those who are young and those with learning disability. In M. Zaragoza, J.R. Graham, G.C.N. Hall, R. Hirschman & Y.S. Ben-Porath (Eds.),Memory and testimony in the child witnessThousand Oaks, CA: Sage.

Bull, R. (1992). Obtaining evidence expertly: The reliability of interviews with child witnesses. Expert Evidence: The International Digest of Human Behaviour, Science and Law, 1(1), 3-36.

Cassel, W.S., & Bjorklund, D.F. (in press). Developmental patterns of eyewitness memory and suggestibility: An ecologically based short-term longitudinal study. Law and Human Behaviour.

Christianson, S. (1992). Emotional stress and eyewitness memory: a critical review. Psychological Bulletin, 112 (2), 284-309.

Christianson, S., Loftus, E.F., Hoffman, H. & Loftus, G. R. (1991) Eye fixations and memory for emotional events. Journal of Experimental Psychology: Learning, Memory and Cognition. 17693-701.

Davies, G., Tarrant, A., & Flin, R. (1989). Close encounters of the witness kind: children's memory for a simulated health inspection.British Journal of Psychology, 80, 415-429.

Dent, H.R., & Stephenson, G.M. (1979). An experimental study of the effectiveness of different techniques of questioning child witnesses. British Journal of Social and Clinical Psychology,18, 41-51.

Drew, P. (1990). Strategies in the Contest Between Lawyer and Witness in Cross-Examination. In J. Levi and A. Walker, (Eds.) Language in the Judicial Process. New York: Plenum Press.

Durkin, K. (1986). Language and social cognition during the school years. In K. Durkin (Ed.), Language development in the school years, Kent: Croom Helm, (pp. 203-233).

Ellis, H.C. & Ashbrook, P.W. (1988). Resource allocation model and the effects of depressed mood states on memory. In K. Fiedler & J. Forgas (Eds.), Affect, Cognition and Social Behaviour. Toronto, Ontario, Canada: Hogreff.

Erdelyi, P. & Stein, J.B. (1981). Recognition hypermnesia: the growth of recognition memory (d') over time with repeated testing. Cognition,9, 23-33.

Fisher, R. P. & Cutler, B. L. (1992, September). The relation between consistency and accuracy of eyewitness testimony. Paper presented at the Third European Law and Psychology Conference, Oxford.

Fivush, R., & Hammond, N. (1990). Autobiographical memory across the preschool years: toward reconceptualising childhood amnesia. In R. Fivush and J. Hudson (Eds.), Knowing and Remembering in Young Children.New York: Cambridge University Press.

King, M. & Yuille, J. (1987). Suggestibility and the child witness. In S.J. Ceci, M. Toglia, & D.Ross (Eds.), Children's Eyewitness Memory. (pp.24-35). New York: Springer-Verlag

Lamb, M., Hershkowitz., I, Sternberg., K, Esplin., P, Hovav., M, Manor., T. & Yudilvitch, L. (in press). Effects of investigative utterance types on Israeli children's responses. International Journal of Behavioural Development

Leichtman, M.D. & Ceci, S.J. (in press). The effects of stereotypes and suggestions on preschoolers' reports. Developmental Psychology

Lindsay, D. S., Gonzales, V. & Eso, K. (1995). Aware and unaware uses of memories of postevent suggestions. In M. Zaragoza, J.R. Graham, G.C.N. Hall, R. Hirschman & Y.S. Ben-Porath (Eds.),Memory and testimony in the child witness (pp. 86-108).Thousand Oaks, CA: Sage.

List, J. A.(1986). Age and schematic differences in the reliability of eyewitness testimony. Developmental Psychology, 22, 50-57.

Marquis., K, Marshall, J., & Oskamp, S. (1972). Testimony validity as a function of question form, atmosphere and item difficulty. Journal of Applied Social Psychology,2,167-186.

Memon, A., Cronin, Ó., Eaves, R. and Bull, R. (in press). An empirical test of the mnemonic components of the Cognitive Interview. In G.M. Davies, S. Lloyd-Bostock, M. McMurran & C. Wilson (Eds.), Psychology and Law: Advances in Research. Berlin:De Gruyter.

Memon, A., Cronin, Ó., Eaves, R. & Bull, R. (1993). The cognitive interview and child witnesses. In G.M. Stephenson & N.K. Clark (Eds.), Children, Evidence and Procedure.Issues in Criminological & Legal Psychology. No. 20. British Psychological Society, Leicester, UK.

Memon, A., Wark, L., Koehnken, G. and Bull, R. (1995). Children's event memory: The effects of interview technique and repeated testing. Manuscript submitted for publication

Milne, R., Bull, R., Koehnken, G. and Memon, A. (1995). Are children suggestible with the cognitive interview? Manuscript submitted for publication.

Moston, S. (1990). How children interpret and respond to questions: Situational sources of suggestibility in eyewitness interviews. Social Behaviour, 5, 155-67.

Moston, S. (1987). The suggestibility of children in interview studies. First Language, 7, 67-78.

Myers, J.E.B. (1993). The competence of young children to testify in legal Proceedings. Behavioural Sciences & The Law, 11, 121-133.

Payne, D.G. (1987). Hypermnesia and reminiscence in recall: A historical and empirical review.Psychological Bulletin,101, 5-27.

Poole, D.A., & White, L.T. (1995). Tell me again and again: stability and change in repeated testimonies of children and adults. In M. Zaragoza, J.R. Graham, G.C.N. Hall, R. Hirschman & Y.S. Ben-Porath (Eds.),Memory and testimony in the child witness Thousand Oaks, CA: Sage.

Poole, D.A. & White, L.T (1993). Two years later: effects of question repetition and retention interval on the eyewitness testimony of children and adults, Developmental Psychology,29, 844-53.

Poole, D.A. & White, L.T. (1991). Effects of question repetition on the eyewitness testimony of children and adults. Developmental Psychology,27, 975-986.

Raskin, D. & Yuille, J.C. (1989). Problems of evaluating interviews of children in sexual abuse cases. In S. Ceci, D.F. Ross, & M.P. Toglia (Eds.), Perspectives on children's testimony, (pp. 184-207). New York: Springer-Verlag.

Richardson, J.T. (1985). The effects of retention tests upon memory: an historical view and an experimental analysis. Educational Psychology, 5, 85-114.

Roediger, H.L. & Payne, D. G. (1982). Hypermnesia: The role of repeated testing. Journal of Experimental Psychology: Learning, Memory & Cognition, 8, 66-72.

Rose, S.A., & Blank, M. (1974). The potency of context in children's cognition: an illustration through conservation.Child Development, 45, 499-502.

Saywitz, K.J., Moan, S.F., & Lamplear, V.S. (1991). The effect of preparation on children's resistance to misleading questions. Paper presented to American Psychological Association on Child Development Meeting, California.

Scrivner, E., & Safer, M.A. (1988). Eyewitnesses show hypermnesia for details about a violent event. Journal of Applied Psychology, 73, 371-377.

Siegal, M., Waters, L., & Dinwiddy, L. S. (1988). Misleading children: causal attributions for inconsistency under repeated questioning. Journal of Experimental Child Psychology,45, 438-456.

Wark, L., Memon, A., Koehnken, G., & Bull, R. (1994). Uses and abuses of schematic knowledge. Department of Psychology, University of Southampton, Working Paper, RWPS/95/1.

Warren, A. & Lane, P. (1995). Effects of timing and type of questioning on eyewitness accuracy and suggestibility. In M. Zaragoza, J.R. Graham, G.C.N. Hall, R. Hirschman & Y.S. Ben-Porath (Eds.),Memory and testimony in the child witness Thousand Oaks, CA: Sage.

Warren A., Hulse-Trotter, K., & Tubbs, E.C. (1991). Inducing resistance to suggestibility in children. Law and Human Behaviour15, 273-285.

Acknowledgements

I am grateful to Debra Poole, Ray Bull, Norman Freeman, Sarah Stevenage and anonymous reviewers for their comments on earlier drafts of this paper.

Footnotes

[1] The data on question type and question salience, these results must be interpreted with caution because the content of the various questions may be responsible for differences in the amount reported (Marquis, Marshall & Oskamp, 1972).

[2] The accuracy rates reported here are lower than those reported in the literature and there may be two reasons for this. First, there were a number of schema inconsistent details in the event such as the colour of the main character's trousers. Most children erroneously recalled these as being `black' when in fact they were `white'. Second, it was obvious that children had difficulty with peripheral details such as the colour of shoes, a detail which they probably did not encode in the first place. There were notable age differences. The 7-year-olds were more accurate and reported more information than 5-year-olds.