Guidelines on Investigatory Interviewing of Children: What is the Consensus in the Scientific Community?

Hollida Wakefield^*

American Journal of Forensic Psychology, 24(3), 57-74 (2006)

Research over the last several years dramatically demonstrates how child witnesses are susceptible to misleading information given to them in leading and suggestive interviews.  As a result of this research, professionals now agree on the basic ways children must be interviewed in order to get accurate, uncontaminated, forensically useful information.  Interviewers must avoid preexisting preconceptions about what happened and encourage children to tell about relevant events in their own words and interviews should be taped.  But despite this consensus, field interviewers often fail to use proper techniques.  Instead, they readily slip into interviewing behaviors that risk compromising the reliability of the information obtained.

Researchers began studying children as witnesses along with their susceptibility to misleading information following several highly publicized sexual abuse cases in the 1980s.  There have now been hundreds of such studies.  The general conclusion from this research is that although children are capable of providing accurate, reliable, and forensically useful information, they are vulnerable to suggestion.  Leading, suggestive, or coercive questioning can not only result in a child making inaccurate statements, it can cause the child to develop a subjectively real memory for an event that never happened. (1, 2, 3).  It isn't only preschool children who are suggestible — older children (as well as adults) are also vulnerable to the suggestive techniques used in the studies with preschool children (4).  But with proper interviewing techniques, suggestibility can be avoided and even young children can provide accurate information.

Forensic psychologists consider suggestibility of child witnesses to be reliable and valid enough to be testified about in court (5).  Evidence concerning the way children should be interviewed in forensic situations is also now sufficiently evolved so that recommended techniques can be used in interviewing children in custody evaluations (6).  A 2004 article by Bruck and Ceci (4) briefly summarizes the research in this area.
Professionals have now developed guidelines on how to conduct investigations and interviews in order to get accurate and reliable information (2, 3, 7-28).  These recommendations are based on the scientific research that demonstrates how faulty interview techniques can produce false and inaccurate accounts.  These guidelines show a strong consensus as to how interviews and investigations should be conducted.

I tabulated recommendations for 24 articles and books; these are summarized in Table I.  When two articles described recommendations of one group (e.g., the National Institute of Child Health and Human Development [NICDH] interview protocol) I only counted it once.  The column titled "eight selected articles" in Table I refers to guidelines for various professional groups along with the protocol for the widely used Stepwise Interview; the other articles are by professionals writing about conducting competent forensic interviews of children.

Adequacy of forensic interviews must be examined in light of all the times the child has been questioned about the alleged abuse from the moment of the initial disclosure.  This includes casual conversations with parents and other relatives as well as formal interviews by police, social workers, and therapists.  This is critical to the case evaluation since informal questioning can affect the responses of children in later, formal interviews.  Even nonsuggestive, open-ended questions don't guarantee accuracy when child witnesses have been exposed to prior misinformation (29).  If the child has been given misinformation in earlier formal and informal interviews, it may later become impossible to obtain an uncontaminated account.  This must be considered in terms of later interviews since the child's memory for the original event may have been altered.
 

GENERALLY ACCEPTED GUIDELINES

1.  Avoid bias; explore alternative hypotheses or explanations

The most important thing for a child interviewer to do to obtain a reliable statement from the child is to have no preconceived belief as to what happened.  The approach should be one of hypothesis-testing.  Unfortunately, many interviewers try to get the child to say things which confirm what they already think happened.  The importance of avoiding bias and taking a hypothesis-testing approach is basic and is specifically addressed by most of the articles that discuss interviewing guidelines.  For example, Ceci and Bruck (2) note that "Interviewer bias influences the entire architecture of interviews and is revealed through a number of different component features that are highly suggestive" (p. 80).  If the interviewer has a preconceived belief about what happened, he or she is likely to ask questions and get answers that confirm this belief.  A number of classic studies in social psychology demonstrate the powerful effect of preconceived beliefs on information an interviewer or experimenter gets (30, 31, 32).  Several recent studies show the effects of interviewer bias on the accuracy of statements made by children in interviews (2, 33, 34) as well as in other situations
(35).

To avoid biasing the interview, the interviewer must explore alternative hypotheses.  One is that the abuse occurred as alleged.  But there are other possibilities.  In general, alternative hypotheses often include the following (these are not exhaustive, but are offered as examples):

2.  Videotape (or at least audiotape) all investigatory interviews

There is a strong consensus that forensic interviews of child witnesses should be videotaped, or at least audiotaped.  Only electronic recording can ensure an accurate record of the interview.  Without a tape, there is no way to know just what was said by the interviewer to elicit a response from the child.  There is no way to know just what the child said. There is no way to determine whether the child's statements are the result of a leading, coercive, and contaminating interview rather than the child's account from his or her own memory and personal knowledge.  There are no good reasons for not taping an investigatory interview of a child witness and many compelling reasons for doing so (36-41).

Even experienced interviewers are unable to accurately recall their specific, verbatim questions and the child's answers that are necessary for evaluating an interview (42).  This includes times when they take verbatim notes during the interview (38).  Reports based upon the recollections of interviewers are likely to be inaccurate and underestimate the degree to which they used closed and leading questions as opposed to open-ended prompts.  When there are no tapes of an interview, there is no way to know the extent to which a child's statements are in response to leading and suggestive questioning.  If it is impossible to videotape the interview (for example. a police officer must take a statement at the child's house), it can be easily audiotaped.

3.  Interview the child alone

The child should be interviewed alone unless he or she is too young to separate from the parent.  A parent or other supportive adult sitting in on the interview can either intentionally or inadvertently cue the child and contaminate the interview.  The only exception to this is when a very young child refuses to separate from a parent.  But this is not desirable and in such cases the parent should be cautioned not to participate in the interview or cue the child in any way.  Also, following the rapport phase of the interview, if the child seems comfortable, the interviewer can ask the child if the parent can leave and wait nearby.

There should also be only one interviewer.  To the extent that the child perceives pressure to say what she thinks the interviewer expects to hear, more than one interviewer will increase this perceived pressure.  Also, children are more likely to go along with what they believe an interviewer expects if the interviewer is identified as an authority figure (2).  I have seen tapes of interviews with as many as four or five adults present in the interview, including police officers in uniform wearing guns.  If it is considered necessary for a team to be jointly involved in the interview (such as a social worker from child protective services and a police officer), the team can discuss in advance what topics need to be addressed and then only one person interview the child.  The other person can observe the interview through a one-way glass mirror and there can be an opportunity to consult before the interview is over.
 

4.  Have a rapport building phase at the beginning

There should be a rapport building phase at the beginning of the interview.  One purpose of this part of the interview is to talk about neutral topics and help the child become more comfortable.  But it is also to encourage and teach the child to give information to the interviewer.  The interviewer should avoid asking a series of closed and forced choice questions during this phase of the interview.

Such questions tell the child that this is like school where there are right and wrong answers and the teacher knows the right answer and is testing the child to see if the child also knows.  Adults routinely test children by asking them questions to which the adult already knows the answer and children are not accustomed to being questioned by authoritative adults when only they have the information and the adult does not.

But in investigative interviews, the child is the source of novel information.  Therefore the interviewer must let the child know from the beginning that only he or she has the answers.  The interviewer must explain the child's role, motivate the child to give detailed and complete accounts of events they have experienced, emphasize the importance of telling only about true events that actually happened, and encourage the child to correct inaccurate statements made by the interviewers (43).  This is best accomplished by beginning the interview with open questions where the interviewer clearly does not have the information.
 

5.  Have a practice interview

During the rapport phase there should be one or more practice interviews where the child is asked open questions about neutral topics, such their last birthday party or the first day of school, and encouraged to give detailed narrative answers.  These practice interviews allow the interviewer to gauge the child's memory and ability to describe past events.  They also allow the child to practice giving information in response to open, nonleading questions.  Research indicates that interviewers get better information from children when they begin with such practice interviews (43).  Children who have the opportunity to practice giving lengthy narrative responses to open-ended questions in the rapport phase continue this behavior in the substantive part of the
interview.
 

6.  Provide ground rules

Young children have a tendency to try to answer any question an adult asks and may provide answers to unanswerable questions such as "Is milk bigger than water?" or "Is red heavier than yellow?" (44).  Therefore, child interviews should begin with ground rules that include telling the child the interviewer doesn't know the answers and that it is all right for the child to say "I don't know" or "I don't remember," and that the child should correct the interviewer if she says something wrong.  It helps if the interviewer practices the ground rules by asking an unanswerable question (e.g., "What is the name of my cat?") and praising the child when he or she says, "I don't know."  The interviewer can also deliberately get information wrong (e.g., "You said you have a younger sister and an older brother" when the child has two brothers) and then reinforce the child for correcting the interviewer.

Examples of ground rules include:

7.  Ask open questions and encourage a free narrative from the child

The most reliable and forensically useful information from children is obtained by encouraging the child to give a free narrative of the alleged events and by asking a series of open, nonleading questions (e.g., who?, what?, when?) or asking the child to "tell me everything you remember about ..."  The research evidence is clear: freely recalled information is more likely to be accurate than information obtained in response to yes/no and forced choice questions.  Consequently, all of the articles discussing guidelines for child forensic interviews make this recommendation.  Even children as young as four can provide substantial amounts of forensically relevant information in response to free-recall prompts (45).  This means that interviewers do not have to rely on forced choice and yes/no questions even with preschoolers.

The substantive portion of the interview should be also introduced in as open a way as possible.  The NICDH investigative interview protocol gives detailed examples of how to progressively phrase such beginning questions (16) and how to continue the interview using open-ended prompts.  Some examples of how to use open-ended probes to introduce the topic of the interview include:

Whenever the child gives response that is on track, the interviewer should encourage a narrative response by asking, "Tell me everything you can remember about that."  When the child pauses, the interviewer should follow up with additional open-ended prompts such as, "And then what happened?," "Tell me more about that."  Such open questions should constitute as much of the questioning as possible.  Interviewers can ask the child to repeat something that wasn't clear or encourage the child to continue the narrative by repeating a phrase, but they should never interrupt the child to redirect the interview or to ask specific questions.  Only when it is clear that the child is not going to provide additional information in response to the open-ended prompts should the interviewer turn to specific questions.
 

8.  Pair specific questions with opened-ended prompts

After obtaining as much information as possible with open questions, interviewers may need to ask specific questions to address important areas that have not been mentioned by the child.  When this is necessary, it should be later in the interview; such questions should not be asked at the beginning.  But it is a common error for interviewers to ask specific questions rather than encouraging narrative responses (23, 46, 47).  When a more specific question must later be asked, it should be paired with an open question.  For example, if the child is asked if his clothes were on or off and says, "Off," the interviewer could then say, "Tell me everything about how they got off"  If the interviewer asks if anything happened in the bedroom and the child says, "Yes" the interviewer can then say, "Tell me everything that happened there."  The risk of getting inaccurate information from such closed questions can be minimized if they are paired with an open-ended prompt.
 

9.  Avoid pressure, coercion, suggestion through giving the child information, asking leading questions, and repeating questions

Although open-ended questions can be repeated without contaminating the child's statements, interviewers should avoid repeating specific, closed, and yes-no questions.  When children are asked the same question repeatedly, they can change their answers to conform to what they think the interviewer wants to hear (2, 3, 48).

Interviewers should never ask suggestive questions which provide information about allegations.  The general principle is that the interviewer shouldn't ask a question about something unless the child has already brought it up.  Obviously, pressure and coercion should never be used.  All the guidelines warn against this.  But in practice, many interviews are leading and suggestive (see 2 and 49 for transcripts of suggestive interviews).  Even with the attention paid to the importance of avoiding contaminating interviewing techniques, this remains a problem (23).  I regularly review videotapes that include closed, forced choice, and leading and suggestive questions with few open-ended prompts.
 

10.  Avoid play, fantasy, and imagining

The interviewer should avoid using such terms as "pretend" or "imagine" or engage in imaginative play as part of the interview.  False disclosures of abuse can sometimes occur in response to techniques involving fantasy, imagery, visualization and reenactment during play (24).  Guided imagery techniques can be particularly suggestive and can lead to the child confusing an imagined event for something that really happened.  Techniques such as having puppets talk to each other, as were used in the McMartin preschool case, should be avoided.

11.  Avoid reinforcing specific responses

Social reinforcement can have a powerful effect on behavior and interviewers should never selectively reinforce specific responses. Research shows that such reinforcement during interviews can readily elicit false allegations of wrongdoing from children (50, 51). Wood and Garven (25) note that several types of interviewer behavior are forms of selective reinforcement or punishment that can contaminate interviews, including:

Although it is important to create a warm and supportive environment, all such selective reinforcement of the child's responses must be carefully avoided.
 

CONTROVERSIAL TECHNIQUES

There is no research supporting the use of anatomical drawings where body parts are named, and these are not generally recommended as part of an interview protocol.  There is an indication that such drawings may decrease the reliability of the information obtained (52).  When used at the beginning of the interview, the anatomical drawings may communicate to the child that what is to be discussed is body parts and touching.  They may confuse very young children who don't understand that a drawing of a naked body has an abstract relationship to an actual person (53).  The drawings should not be necessary with older children.  I recently reviewed a taped interview in which the interviewer showed anatomical drawings to a 13-year-old girl of normal intelligence and then asked if she were a girl or a boy.  Since interviewers should encourage a child to perform at as a mature and effective level as possible, beginning the interview by asking a teenager such a question detracts from the serious purpose of the interview.

There are similar criticisms about discussions of good touch / bad touch.  Guidelines on how to conduct forensic interviews of children do not mention beginning the interview with good touch - bad touch discussions.  Wood, McClure, and Birch (26) observe that agencies continue to use the good touch / bad touch discussion for no particular reason other than they had been doing it for years.  I continue to see good touch / bad touch discussions in tapes I review.  I am unaware of any research supporting this procedure.  What it risks is telling the child from the beginning that the purpose of the interview is to talk about genital touching.

Most guidelines do not recommend using anatomical dolls.  The few that do caution how they are to be used (e.g., 7, 9, 10, 27, 28).  Yuille, et al., (28) note that they should be used only as a last resort and Carnes et al. (9) state that they should be used with "caution" and "only when absolutely needed."  But there are often problems with the way the dolls are used by practitioners in the field (54).  In addition, very young children cannot use dolls as symbols or representations for themselves, and make more errors when using the dolls (55).  Wolfner, Faust, and Dawes (56) critique the dolls and their failure to add incremental validity to the interview.  Many professionals oppose their use.  The conclusion, therefore, is that the dolls are controversial and not generally accepted in the scientific community (57, 58).  There is no evidence that they add to the completeness and accuracy of the information obtained and they are susceptible to increasing the suggestiveness of the interview.
 

INTERVIEWS IN THE FIELD

There is now a clear consensus in the professional community as to how children should be interviewed.  But this hasn't always translated to workers in the field.  For example, estimates of the frequency of the use of leading questions that introduce information to children vary from 13% to 60% with the majority in the 40% to 50% range (59).

In 1990 Underwager and Wakefield (49) reported on an analysis of 36 actual cases involving 150 interviews and 62 interviewers.  The interviewers didn't encourage free recall; instead they relied on closed questions, pressure, and suggestion and they appeared to be trying to substantiate abuse they had already concluded was real.

In 1996 Warren et al. (23) looked at 42 transcripts of sexual abuse interviews conducted by child protective services personnel and found that the interviewers failed to follow practices recommended by researchers on children's testimony.  The interviewers rarely conducted practice interviews, seldom provided ground rules, and failed to begin with open-ended questions, instead relying on specific, yes-no questions throughout.  They frequently introduced new material not previously disclosed by the children.  That same year Lamb et al. (60) reported on their examination of 22 audiotaped interviews from 12 field interviewers in Israel.  Most questions were directive rather than open-ended, and many were leading.

In 2004 Gilstrap (59) examined 80 interviews conducted by 41 field interviewers with 40 children ages 3 to 7 about staged events.  She compared the behavior of these real world interviewers to the types of questions studied in research settings.  She found that the field interviewers asked a substantial amount of leading questions (42%) and that approximately one-third of the leading questions introduced inaccurate information.  The field interviewers repeated questions 12% of the time and introduced novel information 18% of the time.

Saywitz and Geiselman (61) observe that interviewing guidelines designed to maximize the completeness of children's reports are not always based on the realities of work conditions on the front lines.  A problem for workers in the field is that although young children's spontaneous descriptions of past events are accurate, their descriptions are often too incomplete to be useful.  Important legal decisions cannot be made without more information.  Although field workers can gather additional information with more questions, their methods risk undermining the accuracy of the children's statements.  Saywitz and Geiselman have therefore developed approaches specifically geared to field workers to elicit more complete and consistent accounts from children.  These approaches, narrative elaboration, and cognitive interviewing, are based on research in their laboratories and appear to be a promising way for field interviewers to get more complete but accurate information from young children.  Researchers from the National Institute of Child Health and Human Development have also reported on field studies of their interview protocol (NICHD protocol) which demonstrate that with their protocol even young children can provide a substantial amount of forensically relevant and accurate information in response to free-recall prompts (12, 16, 45, 46, 47).
 

CONCLUSIONS

Research over the last several years dramatically demonstrates the importance of properly interviewing child witnesses. Interviewers with preexisting biases who ask leading, suggestive, questions risk confirming their beliefs and getting false information.  There is now a clear consensus in the scientific community about how children must be interviewed in order to get accurate, uncontaminated, forensically useful information.  Unfortunately, field interviewers aren't using these techniques.  Instead, they readily slip into undesirable behaviors that risk compromising the integrity of the interview and the reliability of the information the child gives them.

REFERENCES