by Andy Field
We also investigated whether children who remembered more about an event were more confident in correcting the interviewer when they knew them to be wrong. Finally, since older children are generally found to have better memories for events, the effects of training might be expected to interact with age: without training, younger children should be more affected by interviewer’s distortions than older children, and any beneficial effects of training should be most evident for the younger children.
16.3 Method
* * *
[Here, we give an overview of the study, to help the reader avoid getting bogged down in the details provided in subsequent sections:]
Children were exposed to a staged event (a male stooge performing various actions in their classroom). Each child was interviewed individually three weeks later, to determine what they could recall of the event. Before the interview, half of the children were given training in resisting any distortions introduced into their account by the interviewer; the remainder were given no such training.
16.4 Design
* * *
[Here, we specify the independent variables and their levels. We have two independent variables, each with two levels. This is a wholly independent measures design: each child takes part in only one of the four experimental conditions:]
There were four conditions, corresponding to the permutations of two independent variables: age of child (‘younger’, aged 6–7 years, or ‘older’, aged 9–10 years) and training condition (trained to correct the interviewer’s distortions of their testimony, or not trained).
[We provide a brief mention of our dependent variables. We have two dependent variables in this study:]
Each child participated in only one condition, and provided two scores: one score for the number of times they corrected the interviewer, and one score for each piece of information correctly remembered.
Participants
[Our participants are a ‘special’ group, since they are all school children. Note, however, that we give all of the relevant information, but without getting bogged down in irrelevant detail:]
Participants were children at a local primary school. The ‘younger’ children used were 6–7 years old (mean age: 82.7 months; range: 78–89 months). The ‘older’ children were aged between 9–10 years (mean age 118.6 months, range 114–124 months). The parents provided informed consent but the children were naïve as to the nature of the experiment.
16.5 Procedure
The Staged Event
[Sufficient details are given here of how the experiment was performed to enable the reader to replicate it in all its important details:]
Children were exposed to a staged event, which served as the basis for the subsequent test of their recall. The event took place between 9:45 am and 10:00 am on the last day before the start of the school’s Christmas holiday. A confederate entered the classroom and shouted out the question ‘Where’s your teacher?’ He then asked about the lights and proceeded to turn them off and on twice. Each classroom went very dark, due to the poor weather conditions outside. This served to gain the children’s full attention. The confederate then walked around the classroom with a screwdriver, talking to himself, but loudly enough for all of the children to hear. He then bumped his leg on the table before asking the teacher for the time and finally leaving the classroom. The staged event was performed twice: once for the class of younger children, and again for the class of older children. The entire event lasted for approximately 75 seconds on both occasions.
The Interview Procedure
After the Christmas holiday period (approximately three weeks after the staged event), each class of children was randomly divided into two groups – those who were interviewed after being given training in correcting interviewer’s distortions of their testimony, and those who were interviewed without having received such training. Those children who had been allocated to the training groups were given training in how to notice and correct any distortions made by the interviewer. This was conducted prior to the interview. The child was told that they and the experimenter were going to play a game in which the child had to say ‘no that’s not what I said’ every time the interviewer did not repeat exactly what the child had just said. The child was given an example and then the opportunity to practice this during a conversation about Christmas. For example if the child had said they did not visit anyone, then the interviewer would repeat back that they had stayed in all the time.
Each child was interviewed individually by one of the experimenters (JG), in the teacher’s staff room. The interview lasted for about fifteen minutes. At the start of the interview, the experimenter asked a couple of questions to ascertain whether the child knew what it meant to tell the truth and to tell a lie, and the child was asked to tell the truth. The experimenter reinstated the context of the staged event the child had witnessed, by reminding the child of where they had been sitting and what they had been doing on that day. The child was then asked questions about six different aspects of the scenario: what the confederate looked like; what he was wearing; what he was carrying; what time of the day the incident occurred; the confederate’s actions; and what he said. After each of the six sections of the interview, the experimenter summed up what the child had just said, but with some distortion added to their responses. For example, if the child said the man had worn jeans, then the experimenter said the man was wearing work clothes. The experimenter made six distortions of this kind during the interview. All interviews were recorded on cassette tape, for subsequent analysis.
16.6 Results
* * *
[For clarity, the results are described separately for the two measures that were taken. It was decided that tables were preferable to graphs for showing the descriptive data in this particular case. Note that the descriptions of the study’s main findings are perfectly intelligible without reference to the tables – all the essential information for the reader to understand what was found is located here, in the text. Note that every time we present the results of a statistical test, we also explain to the reader what they mean. Note too that the description of the results is confined to a description of what was found – not what it means in terms of psychological theory. The interpretation of the results in this sense will follow later, in the Discussion.]
Two measures were taken for each child:
(a) Number of items correctly recalled:
[First, we explain exactly what was scored.]
One point was given for each item correctly recalled, up to a maximum of 24 points.
[The descriptive statistics . . .]
Table 1 shows the mean number of items recalled correctly by children in each of the four groups.
Table 1. Memory scores as a function of age and training condition
[The inferential statistics . . .]
A two-way Analysis of Variance (ANOVA) with independent measures on both variables was conducted on these data. This revealed no significant main effect of training (F(1, 36) < 1): giving training to children to correct the interviewer’s distortions of their testimony did not affect the children’s ability to recall the events. There was a significant main effect of age (F(1, 36) = 7.96, p < .01, r = .43): older children recalled significantly more items than did younger children. There was no significant interaction between age and training (F(1, 36) < 1).
(b) Number of distortions corrected:
[Details of what was scored . . . ]
One point was given for each distortion that was corrected, up to a maximum of 6 points. Correction of a distortion was held to have occurred if the child said ‘no that’s not right’ or ‘no’, or if they repeated what they said originally. A sample of the replies was given to an independent rater who was naïve as to which condition the samples had come from. Cohen’s Kappa was 85% for the memory scores and 93% for the correction scores.
[Descriptive statistics . . . ]
Table 2 shows the distortion-correction scores for each group. It app
ears that, regardless of age, children in the training conditions corrected more distortions than children in the non-training conditions.
Table 2. Number of distortions corrected, as a function of age and training condition
[Inferential statistics . . . ]
A two-way independent measures ANOVA confirmed this impression: it revealed a significant main effect of training (F(1, 36) = 27.06, p < .001, r = .66), and no significant effect of age, either as a main effect or in interaction with training (F(1, 36) < 1 in both cases). Thus, the children who had received training in correcting the interviewer corrected significantly more distortions than did the untrained children, and this was true regardless of the children’s age.
(c) Relationship between memory score and the number of corrections made:
A Pearson’s correlation test was performed to determine whether the number of details that a child could remember correctly, was related to the number of interviewer’s distortions that they corrected. The results indicated that there was no significant correlation between these two measures (r (38) = .10, ns): children who were able to remember a larger number of the details of the staged event were no more or less likely to correct the interviewer than were children with poorer recall.
16.7 Discussion
* * *
[First, a summary of the principal results of this study:]
The main result of this study was that training a child to correct an interviewer who mis-reported their testimony significantly increased the child’s tendency to correct any distortions made by the interviewer in a subsequent interview. This was true for both younger and older children, despite age-differences in the total amount of information correctly recalled (age-differences that were consistent with those found by previous researchers: review in Ceci & Bruck, 1993). There was no evidence that training affected the amount of information that was successfully recalled by the children.
[Next, we compare our findings to those of previous researchers, and try to account for any discrepancies between them:]
At first glance our results do not appear to replicate the findings of Roberts and Lamb (1998), who found that the children did not correct two thirds of the distortions made. However, the children in that study were involved in a real child protection interview and as such were only given a short time to practice correcting the interviewer’s distortions of their testimony, as part of the rapport building phase of the interview. In contrast the children who participated in the present study were actively encouraged to practice correcting distortions made by the interviewer, for about 10 minutes. This implies that in order for such training to be effective, the child should be formally trained and given time to practice correcting distortions. However, further research is required to confirm whether this is so, and to determine the precise factors which determine whether or not training is effective.
The present study is consistent with the work of Gee, Gregory and Pipe (1999), who trained children to resist misleading questioning by an interviewer about a trip to a science centre. Gee et al. also found that the children in the training condition were less likely to comply with the interviewer when the interviewer gave the child misleading information. Taken together, these studies suggest that if the child is given permission to correct an interviewer, and time to practice doing so, they will challenge the authority of the interviewer and correct them if they provide inaccurate information – excessive suggestibility need not be an inevitable problem when interviewing children.
[Now, we start to discuss problems and limitations of the present study. First, a limitation – the study doesn’t pinpoint the source of the effect that was found . . . ]
Although we have demonstrated that training increases children’s readiness to correct their interviewers, the source of this effect remains to be determined. One possibility is that it arises primarily from differences between the trained and untrained children in terms of the rapport that developed between them and the interviewer. Rapport has been demonstrated to have a significant influence on adult suggestibility to misleading questions (Bain & Baxter, 2000), and there is reason to believe that similar effects occur in children (review in Ceci & Bruck, 1993). In the present study, children in the training stage spent more time with the interviewer and had time to build up a rapport with her. Evidence for the importance of a good rapport between interviewer and child comes from a study by Goodman, Bottoms and Schwartz-Kenney (1991), who interviewed 3–4 and 5–7 year olds about a visit to a Health Clinic for inoculation. If the children were interviewed with a cool and detached interview style, the 5–7 year olds performed better than the 3–4 year olds. However, a warm, empathic and encouraging interview style improved performance for both age-groups and eliminated age-differences in susceptibility to leading questions.
An alternative possibility is that the effects in the present study arose from the fact that the trained children had explicitly been given permission to challenge the interviewer, and had practice in refusing to conform to the interviewer’s perceived intentions. The possibility that the untrained children conformed to the interviewer as an authority figure is in line with the results of Goodman et al.’s (1995) study, which found that children were less suggestible when interviewed by their mother than when they were interviewed by a stranger. Similar conclusions come from Ceci, Ross and Toglia’s (1987) study, which found that preschool children were less influenced by misleading information when it was given to them by a seven-year old child, than when it came from an adult.
[Suggestions for future research, to remedy the limitations of the present study, and extend the generality of its findings:]
In further research, it would be useful to attempt to identify the relative contributions to the trained children’s behaviour of the rapport-building and training components. These two factors could be separated out by including a condition in which the interviewer builds up a rapport with the child but does not train them to correct the distortions, or vice versa. In the present study, the same experimenter conducted both the training and the interviewing phases of the study, and had therefore built up a rapport with some of the children prior to the interview phase. This may have given the children in the training conditions more confidence in correcting the interviewer when she distorted what the child had said. Future research could usefully investigate whether the effects of training found in the present study persist when the trainer and interviewer are different people.
There are several other issues which also need to be explored. For ethical reasons, the current experiment used a neutral event. It would be desirable to investigate the effects of training when the event is of more consequence to the child, such as when a child witnesses a criminal act. Child witnesses to a criminal event have been found to be less suggestible than those witnessing a neutral event (Ochsner, Zaragoza & Mitchell, 1999). On the basis of this, one would therefore predict that similar effects to those in the present study would be found – namely that children would correct the interviewer – but that this effect would be even more marked. Ideally, however, this prediction should be tested with more ecologically-valid events.
Finally, in the present study, distortions of the child’s testimony occurred every time the interviewer summed up. In future research, ecological validity could be increased by ensuring that some of the testimony was repeated correctly and some parts were distorted, to see whether the child would continue to correct the interviewer. Additionally, in real life an interview would last longer than 10–15 minutes. As the interview progresses, it is possible that the child might not remember to correct the interviewer whenever they made a mistake, especially if the interviewer distorted only a few pieces of the information provided by the child. Consequently, future research should examine the present effects within the context of considerably longer interviews, to see if training continues to be as effective under these circumstances.
16.8 References for the Example
Bain, S.A. and Baxter, J.
S. (2000). Interrogative suggestibility: the role of interviewer behaviour. Legal and Criminological Psychology, 5, 123–133.
Bjorklund, D.F., Cassel, W.S., Bjorklund, B.R., Brown, R.D., Park, C.L., Ernst, K. and Owen, F.A. (2000). Social demand characteristics in children’s and adults’ eyewitness memory and suggestibility: the effect of different interviewers on free recall and recognition. Applied Cognitive Psychology, 14, 421–433.
Ceci, S.J. and Bruck, M. (1993). Suggestibility of the child witness: a historical review and synthesis. Psychological Bulletin, 113 (3), 403–439.
Ceci, S.J., Ross, D. and Toglia, M. (1987). Age differences in suggestibility: psychological implications. Journal of Experimental Psychology: General, 117, 38–49.
Dent, H. (1982). The effects of interviewing strategies on the results of interviews with child witnesses. In A. Trankell (Ed.) Reconstructing the past: The role of the psychologist in criminal trials. Netherlands: Kluwer.
Dent, H. and Flin, R. (Eds.) (1992). Children as Witnesses. Chichester: Wiley.
Dent, H. and 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.
Gee, S., Gregory, M. and Pipe, M.E. (1999). ‘What colour is your pet dinosaur?’ The impact of pre-interview training and question type on children’s answers. Legal and Criminological Psychology, 4, 111–128.
Goodman, G.S. and Reed, R.S. (1986). Age differences in eyewitness testimony. Law and Human Behaviour, 10, 317–332.