© 2001
James Alexander Tanford
Indiana University School of Law
211 S Indiana Av
Bloomington IN 47405


The interviewing process is not as simple as asking witnesses to tell you everything they know about an event. Many factors influence the information flow, distorting and limiting it. The witnesses themselves will be of all types and come from all levels of society. Some will be helpful; some will not want to become involved. Witnesses will have differing abilities to remember and articulate what they observed. Some will be honest and some will lie. You will see witnesses who are obnoxious, neurotic, dull, stupid, nervous, and who speak in accents you cannot understand. They cannot all be treated the same way. As an interviewer, you must learn the techniques for maximizing the completeness and reliability of information sought from so many different kinds of people.

Your own personality, attitude, and manner of asking questions can have a profound impact on the information you receive. Interviewing is a dynamic process that depends upon interaction between two persons. Your input, both conscious and subconscious, can have as much to do with the facts elicited as the input from the witness. If you communicate, either verbally or nonverbally, that some facts are unimportant, that you want to change topics, or that you do not believe what the witness is telling you, the witness is likely to stop sending information. Studies show that even the particular words you use, the way you phrase a question, or the sequence in which you ask questions can alter the way in which a witness remembers an event.


If the person being interviewed is your client, the attorney-client privilege will shield most of what the client tells you from forced disclosure. The attorney-client privilege does not extend to nonparty witnesses. Statements made by witnesses may be confidential, but they are not privileged and must be disclosed under proper circumstances. Thus, if requested, a formal statement made and signed by a witness must be turned over to your opponent during discovery. However, work-product reflecting your thought processes is not subject to discovery. It is common to draw a line between a verbatim record or signed statement and your own notes about what is important in the witness's testimony or how the witness fits into your case. Many lawyers do not take formal statements from favorable witnesses; they write summaries instead in order to evade discovery.

The leading case is Hickman v. Taylor, 329 U.S. 495 (1947), in which the Supreme Court exempted unwritten statements from discovery and required a showing of necessity before written witness statements became discoverable. The rule is different for expert witnesses. Discovery of the contents of their opinions generally is permitted, whether reduced to writing or not, under Fed. R. Civ. P. 26(b)(4). In federal criminal cases, the Jencks Act, 18 U.S.C.A. 3500, makes written statements by prosecution witnesses discoverable only after the witness testifies.


The effectiveness of an interview depends on a multitude of factors. Some things that affect the flow of information are beyond your control, including the quality of the witness's perception and recollection, differences or similarities in socioeconomic status or in the language used between yourself and the witness, unconscious emotional reactions to the other person, and etiquette barriers. You do have control over some factors that appear to have an effect on interview success --- the time and place of the interview, the number of interruptions, and the scope and wording of questions. Finally, personal factors may influence the accuracy of the interview. In this category fall your own biases and attitudes, your ability to listen, and your awareness of nonverbal communication. Before we discuss the strategy and tactics of interviewing, you should read the following materials so that you will understand the dynamics of interviewing.


Witnesses to an event do not remember it either accurately or completely. The data collected by psychologists testing witness recall is staggering in its implications. In most experiments, subjects remember as many things inaccurately as they do accurately, and they forget more than they remember. In one of the classic experiments in this area, a professor staged a simple assault in front of his class and then asked his students to describe what had happened. The student who made fewest mistakes nevertheless gave twenty-six percent erroneous statements; the least accurate witness was wrong eighty percent of the time. Students put words into the mouths of people who had been silent spectators, attributed actions to the participants that had not been done, and completely forgot essential parts of the event. See H. Munsterberg, On the Witness Stand 49-51 (1908).

What causes such distortions in memory? The answer is found in the nature of the memory process. Psychologists divide memory into three states: acquisition, retention, and retrieval of information. Events at any of these stages can be the cause of memory failure. The information simply may not have been perceived in the first place because the witness was not looking, the lighting was bad, the event did not seem important, or many events took place in a very short time. Alternatively, the event initially may have been perceived incorrectly because the witness was under stress or because the witness interpreted the event consistently with his or her expectations and stereotypes.

Even if information was perceived correctly, memory can become distorted during the retention stage. All witnesses forget some information as time passes. The information that is retained can be altered when new information is superimposed on the old. Thus, a witness may "remember" details learned from newspaper accounts or conversations with other witnesses.

The way in which information is retrieved from memory also appears to affect recall. Even information perceived and retained accurately can become distorted by the method of questioning. First, recall is affected by whether a witness is asked specific questions or is asked only to give a narrative description. Narrative reports are more accurate but less complete; answers to interrogatories cover more of the events but contain many errors. In one experiment, subjects who described an event in narrative form were ninety-one percent accurate, but recalled only twenty-one percent of the details; subjects who answered a series of specific questions were only fifty-six percent accurate, but recalled (correctly or incorrectly) seventy-five percent of the details. Lipton, On the Psychology of Eyewitness Testimony, 62 J. Applied Psychology 90 (1977). Second, the wording of questions is crucial because small changes in word choice can have dramatic results. In a famous experiment by Elizabeth Loftus, subjects were shown a film of a car accident and then questioned about it. Those subjects who were asked how fast the cars were going when they "smashed" gave estimates averaging 40.8 miles per hour. When asked how fast the cars were going when they "hit," the subjects' answers averaged 34.0 miles per hour. Those questioned about how fast the cars were going when they "contacted" gave answers averaging 30.8 miles per hour. Loftus & Palmer, Reconstruction of Automobile Destruction: An Example of the Interaction Between Language and Memory, 13 J. Verbal Learning & Verbal Behavior 585 (1974).

1. The problem of poor memory may not be as bad as the Munsterberg experiment suggests. See Marshall, Marquis & Oskamp, Effects of Kind of Question and Atmosphere of Interrogation on Accuracy and Completeness of Testimony, 84 Harv. L. Rev. 1620 (1971) (experiment in which witnesses were ninety-eight percent accurate for the most important matters).
2. The psychology of memory is summarized in Elizabeth Loftus, Eyewitness Testimony (1979); .L. Taylor, Eyewitness Identification (1982); A. Yarmey, The Psychology of Eyewitness Testimony (1979).


Interviewer bias refers to the process by which the interviewer influences the interview -- it has nothing to do with prejudice. We are all biased -- we have political and social preferences and beliefs. When these preferences and beliefs intrude into the interview, they are likely to produce erroneous information.

The simplest example is our reaction to a client who suggests that her husband was abusing their daughter. Allegations of abuse are common in divorce situations. Some are true; some false. If the interviewer is sympathetic toward child abuse allegations, this receptiveness to the client's allegation may cause the client to provide extensive details about the incidents. If the interviewer is skeptical about the allegation, the opposite may occur -- the client becomes reluctant to provide any details. In both cases, the interviewer's bias has influenced the information transmitted, regardless of whether the underlying allegation is true or false.

Bias also may occur at a more subtle level. The behavior of the interviewer when asking questions and recording answers affects the flow of information. Consider the following two ways to phrase the same question:

Q1. Has your ex-husband been paying his child support?
Q2. I don't suppose your ex-husband has been paying his child support, has he?

Follow-up questions are even more likely to introduce bias. Consider:

Q: Has your ex-husband been paying his child support?
A: He's missed the last few payments.
Follow-up Q1: But generally he's been paying?
Follow-up Q2: So he hasn't been keeping up on his payments?

The client will probably agree with either characterization, although the implications are quite different. Without realizing it, you have let your preconceptions about whether fathers generally do or generally do not pay their child support bias the interview.

Bias is also introduced by your reaction to the client's answers. What ends up in your memory may not be what the client told you. You may simply fail to "hear" some things the witness says, especially things that run counter to your own attitudes or are contrary to what you expected. Or, you may "round off" an answer, and store the characterization in your memory. For example, if your client tells you her ex-husband has missed the last three support payments, you may remember her telling you that the husband doesn't make his support payments -- although he may have had a perfect record up until the last three. Finally, simple body language may bias the interview. Your act of jotting down an answer or not jotting it down may cause the witness to believe the subject is important or unimportant, causing him or her to expand on or stop talking about the topic.


It is self-evident that you can conduct a successful interview only if you hear what is being said. Effective listening, for some reason, is not an easy skill to master. There may be two reasons for this: 1) Most of us prefer to talk rather than listen; and 2) We are able to listen three or four times as fast as the other person talks. The average person speaks at 125 words per minute and can listen at 500 words per minute. There is a danger that the "gap" may be filled by thinking of similar situations, leaping ahead and anticipating what will come next in the story, thinking about the law, or wondering what you will have for dinner.

Listening involves not only hearing the words used by a witness, but also "hearing" the nonverbal communication. It has been estimated that at least sixty percent of all important messages are passed nonverbally. Social scientists have found that information is communicated by body movement and positioning (kinesics); variations in vocal qualities such as pitch, intensity, and rapidity (paralinguistics); spacing and distancing (proxemics); and by gaze and eye contact, head nods, gestures, touch, and choice and arrangement of artifacts such as clothing and furnishings.

Some common types of nonverbal cues are:

1. The too-vigorous and vociferous denial. Often, but not always, a witness who "doth protest too much" is hiding something. My favorite example was a tenant who had a pet cat in violation of the lease. On one occasion, the landlord was coming over to check on some damage to the house. The tenant removed all traces of the cat and carefully instructed his three children not to say anything about the cat. When the landlord arrived, the youngest child ran up and announced, "We don't have a cat."

2. Lapsing into silence. When a witness makes a tentative statement about something and then is silent. This silence may mean that the witness is unsure whether to proceed, or has a hidden agenda.

3. What is omitted. If a witness who omits key facts, they are probably important and negative. For example, a prospective client once called me with a question about the Federal Emergency Family and Medical Leave Act. His wife had been fired, he told me, for taking family leave to accompany him to Minnesota. He was disabled, and he had a doctor's note that she had to accompany him because he couldn't travel by himself. Conspicuously missing from this narrative was any mention of why they went to Minnesota or for how long. It turned out that he wanted to go fishing, and they were gone for a month.

4. Depersonalizations. A husband who refers to his wife as "She" instead of Ellen is probably engaging in a psychological distancing maneuver to avoid dealing with emotional tensions.

5. Pacing. Speech which is halting or hesitant may indicate that the client is still trying to sort things out in his or her mind or is uncomfortable.

6. Voice timbre and inflection. These are too numerous to mention. A wide variety of emotions are reflected in the tone of our voice -- love, affection, anger, hatred, sadness, excitement, worry, jitteriness, doubt, resignation, invitation, etc.

7. Physical gestures. The most commonly known form of "body-language" is the open/closed distinction made popular in the 1970s pop-psychology literature. Folded arms, clenched fists, crossed legs, and leaning back indicated distrust, irritation, embarrassment, or anger. Open arms, uncrossed legs, and leaning forward indicated friendliness, openness, honesty, and truthfulness. There is some danger, however, that people from different cultures use gestures differently. Body language, like other language, is culture-specific. A classic example of the danger of miscommunication occurred during the first visit of the Premier of the Soviet Union to the United States. As Nikita Khruschev stood in the rear of an open car, waving to the crowds, he occasionally clasped his hands over his head, moving them up and down in much the same way that a boxer will do after a victorious fight. The TV news commentators were all upset about this gesture as typical Russian braggadocio. However, it turns out that in the gestural language of many Eastern European countries, that clasped hands motion signified friendship.

8. Physiological Responses to Stress. Sweating, blanching, blushing, or in extreme cases, vomiting or sudden need to use the bathroom indicate stress, anxiety, or shame.

9. Facial Expressions. Obviously, many emotions are expressed through changed facial expressions. But be on the lookout also for situations when the witness may hide his or her facial expressions. The witness may avert his or her face, or suddenly adopt a wooden "poker face," suggesting that the witness wishes to avoid whatever would have been revealed had he or she kept looking.

NOTE: For more on nonverbal communication, see Harrison, Nonverbal Communication, in Handbook of Communication 93-115 (Pool & Schrams, ed. 1973); Watson, The Lawyer in the Interviewing and Counseling Process 48-58, 60-65 (1976).


An inhibitor is a barrier or obstacle to communication that should be avoided, circumvented, or removed from the witness's mind. A facilitator, on the other hand, is not merely the absence of an inhibiting barrier but is a positive force motivating the witness to communicate.

1. Inhibitors

a. Competing Demands for Time.
The witness has other ways he or she would like to be spending this time. Children may need attention, the witness may need to get back to work, etc.

b. Ego Threat.
Witnesses tend to withhold or repress information which may threaten their self-esteem or is embarrassing.

c. Etiquette. Information perceived to be inappropriate may be withheld. This includes information that would be in poor taste or would show disrespect to the lawyer. Some people believe there are things adults don't discuss in front of children, men in front of women, married couples in front of unmarried people, students in front of teachers, etc. The witness may simply not want to offend you.

d. Trauma.
Emotional trauma is an unpleasant feeling associated with a disturbing experience. The unpleasant feeling will return if the disturbing experience is brought to the conscious level. The witness may not want to relive the original emotions associated with the experience, and so will have suppressed the experience itself. It is unwise for you to play "Young Therapist" at this point, and try to get the witness to remember the event in order to obtain a cathartic release from guilt feelings. It takes years of training to become a qualified therapist, and you may do your client more harm than good.

e. Forgetting. There is a natural fading of memory. The longer an event is in the past, the more likely the witness will have only a partial memory of it. Even some of the simplest and most obvious facts may have been forgotten.

f. Chronological Confusion. This term refers to the tendency to confuse the chronological order of two or more experiences. The witness may either be unsure which event happened first, or may remember the order incorrectly. Confusion also is likely when a witness is asked whether a condition that currently exists, also existed at an earlier time. This type of confusion is especially common when asking about relations between people, especially deteriorating ones.

g. Inferential Confusion. It is extremely common for witness to make errors of inductive and deductive logic. Induction errors are most common -- witnesses tend to make erroneous, exaggerated, and unfounded generalizations based on a few anecdotal concrete experiences. Deduction errors are less common and more difficult to detect. They occur when a witness is asked to give concrete examples to illustrate the interviewer's generalization.

h. Case threat. This applies to clients and others with a specific interest in the outcome of the litigation. If such a witness believes that revealing information will be harmful to the case, he or she may withhold it. In this manner, a defendant may deny that he or she had a motive for or was in the vicinity of a crime.

i. Role expectations. Clients and witnesses have certain preconceived ideas of what attorneys look like, how they talk and dress, what their offices will look like, and so on. These can be called role expectations. These expectations may be barriers in themselves, e.g., a client who expects the lawyer to define what is important and therefore waits to be asked specific questions. They also may inhibit communication in other ways. If a witness expects a lawyer to be a well-dressed, middle-aged, white male, and meets someone of different appearance, this failure to fulfill expectations can cause distrust and reluctance to communicate.

j. Perceived irrelevancy. This inhibitor is so obvious it often is overlooked. If the witness does not think information is relevant, he or she is not likely to offer it, and the witness may not give serious consideration to the answer even if asked.

k. Greater need. This inhibitor is similar to competing time demands. A witness may have a greater or more immediate need to discuss one issue than another, and may be unable to concentrate on topics the lawyer wishes to explore. A first offender in jail awaiting trial, having heard stories about homosexual attacks, may be so desperate to get out on reduced bail he simply is unable to talk about the facts of the offense. Similarly, a witness to a crime may believe her life in danger if she testifies, and this greater need will cause her to deny that she knows anything relevant.

l. Interruptions. If you interrupt a witness during an interview, this will inhibit the flow of information. If you interrupt an answer, you may communicate that you think the matter unimportant. If you permit the interview to be interrupted by telephone calls and secretaries, you may threaten the witness's ego by communicating that he or she is less important than your other activities.

2. Facilitators . Facilitators to communication are social-psychological products of the relationship between the information and the situation in which it is communicated. Some of them apply to any type of information and others only to more narrow categories of information. Using facilitators should maximize the flow of relevant information and maintain optimal interpersonal relations.

a. Fulfilling Expectations. The witness will tend to respond, consciously or unconsciously, to your expectations, both verbally and non-verbally expressed. By steadily and consistently communicating your expectations to the witness -- that she or he will cooperate fully and truthfully -- you will exert a subtle force affecting the witness's cooperation. Expectations also can act as an inhibitor. If the interviewer communicates that he or she does not expect the witness to be willing to discuss the case, the respondent may fulfill that expectation.

b. Recognition. A good interviewer takes advantage of every opportunity to give the witness sincere recognition for a job well done, or a sensible course of action taken. Did your client save all the letters from the collection agency? Excellent! That's just what a lawyer would have done.

c. Altruistic Appeals. This can work to overcome etiquette or self-esteem barriers in some cases. It may be easier for a victim of sexual abuse to talk about it, if she thinks she will thereby help other abuse victims.

d. Sympathetic Understanding. Perhaps the simplest facilitator -- we all respond well to a (genuine) sympathetic response from someone else. Offering a sympathetic, non-judgmental ear to someone who needs to be understood, will often trigger a flood of information.

e. Changing the Frame of Reference. If you cannot get at information from one direction, change the frame of reference. Most of an interview is conducted from the witness's point of view. When this runs into case threats and other inhibitors, you may be able to circumvent the problem by changing the perspective. If a criminal defendant denies any involvement in the crime, switch to the cop's perspective -- ask, "What do you think the cops will say on the stand?" After your divorce client tells her story about her lying, cheating husband, ask her what she thinks he will say about her.

f. Ask "Why?" One of the problems with many interviews is that they are superficial. The information you get is general and oversimplified. Suppose you ask a sixth-grader how he is doing in school. He says he's doing okay, has a new teacher and is getting better grades. Sounds good, right? Try asking "Why are your grades improving?" You might hear, "Oh, she's real easy and doesn't make us do much homework."

NOTE: One of the best discussions of facilitators and inhibitors is found in Robert Gorden, Interviewing: Strategy, Techniques, and Tactics 104-33 (Rev. ed. 1975).



Broad, open-ended questions produce the most reliable information from the client or other person being interviewed. They permit the client to say what's on his or her mind, picking the topics and the direction of the interview. The client's memory is least affected by the interviewer's biases. For example:

What can we do for you?
What happened next?

However, open-ended questions are not adequate by themselves because they do not provide enough detail. You will therefore have to also make use of more directed questions.


In a directed question, you select the topic or subject, but the client determines what information is relevant. For example:

What happened after the social worker got there?
How is your daughter doing at school?

Like broad questions, this kind of directed question produces reliable information, but not enough detail. For eliciting factual details, you will need to use narrow questions.


Narrow questions probe for very specific pieces of information not supplied voluntarily by the client in response to broad or directed questions. The fact that the client did not supply the detail voluntarily may mean that the client is uncertain of it. When you force him or her to give you an answer, you will get the client's best "guess" as to what the detail is -- sometimes accurate and sometimes not accurate. The best way to maximize accuracy is to use an unrestricted narrow question.

1. Unrestricted Narrow Question. In an unrestricted narrow question, you select the general topic and the specific detail you want, but require the client to supply the information rather than answering "yes/no." This maximizes the witness's opportunity to tell you that he or she does not remember the detail, reducing the amount of false information. For example:

What was the social worker's name?
How much is your rent?

If important facts are still missing, you might try restricted narrow questions.

2. Restricted Narrow Question. A restricted narrow question not only directs the client to a topic and detail, but supplies a plausible answer to the client to jog his or her memory. The plausible answer must be presented in a nonjudgmental and nonleading way, of course. The problem with restricting the client's answer is that you implicitly force the client to choose one of your answers, even if they are all wrong. This increases the likelihood of false information entering the interview. For example:

Have you seen any signs of drug use?
Do you usually take her to the doctor when she's sick?

3. Leading Questions. The most restrictive question is the leading question in which you not only provide details but suggest to the client that these details are true. Leading questions should never be used, because they make it difficult for the client to disagree with you even if you are wrong, raising a high probability of false information. For example:

You want custody, don't you?
The social worker has been to your home before, hasn't she?

NOTE: See David Binder & Susan Price, Lawyers as Counsellors (1991).


Always begin with broad questions, and go from broad to directed to narrow questions -- like a funnel. For example:

[Start with broad questions:]

Q: Tell me about the visit from the social worker.
A: She said the school counselor had called her because Sally wasn't in school.
Q: What else did she say?
A: She wanted to know why Sally was missing school.
Q: What happened next?
A: I told her Sally was sick last week.
Q: What else happened?
A: That's about it.

[Go to directed questions:]

Q: How is your daughter doing in school?
A: Not too good. She's been sick a lot this year.
Q: What's the matter with her?
A: I don't know. She just feels bad.

[Go to unrestricted narrow questions:]

Q: Has she seen a doctor?
A: Yes, I took her to PromptCare.
Q: Do you remember about when that was?
A: Yeah, last month.
Q: What did the doctor tell you was wrong with her?
A: I don't know.
Q: What was the doctor's name?
A: I don't know.
Q: About how many days has Sally been sick in the last 2 months?
A: I don't know, about 2 or 3 days a week.

[Go to restricted narrow questions:]

Q: Did the doctor prescribe any medicine?
A: No.
Q: Did you tell the school Sally was sick?
A: Yeah, I sent a note last month.
Q: Did you talk to anyone from the school?
A: No.
Q: Did you send just one note or more than one?
A: Just the one.
Q: Did you ever take her back to see the doctor?
A: No.

NOTE: Recent social science experiments provide further guidance in choosing between open and narrow questions. In one study, witnesses saw a short film depicting a robbery. When asked by an open-ended question to narrate what they remembered, witnesses were ninety-one percent accurate in the details they recalled, but were able to only recall twenty-one percent of the details in the film. When asked only specific, narrow questions, they were able to recall many more details (seventy-five percent), but accuracy fell off to fifty-six percent. See Lipton, On the Psychology of Eyewitness Testimony, 62 J. Applied Psychology 90 (1977).See also Marquis, Marshall & Oskamp, Testimony Validity as a Function of Question Form, Atmosphere, and Item Difficulty, 2 J. Applied Soc. Psychology 167 (1972) ( accuracy decreases but completeness increases continually as question form becomes increasingly narrow). It is thus apparent that the use of narrow questions increases both the number of details remembered accurately and the number of details recalled incorrectly. If you use open questions first, you at least can be fairly confident of the accuracy of the information elicited. However, if you use narrow questions first and they elicit information only fifty-six percent accurate, you will not know which are the accurate details, and the net result will be less certainty even though you have more information. See generally E. Loftus, Eyewitness Testimony 93-94 (1979).



It is helpful to arrange the furniture in a manner that gives you and the client options on the degree of formality of the interview. The most formal arrangement occurs with the interviewer behind a desk and the client on the opposite side with the expanse of the desk between them. This arrangement fixes the distance between the two participants (typically 5-7 feet) and limits visual contact to the head, arms, and upper body. A less formal desk arrangement is the cross-corner arrangement where the interviewer and client interact across a small corner portion of the desk or table. This position allows greater flexibility in the closeness of the two participants (3-5 feet) and provides greater visual contact, particularly eye contact, between the participants, and reduces the "barrier" of the large desk top. It still results in a "protected space", which may make the client more comfortable, as the participants are separated by a portion of the desk.

Most legal interviews should be conducted with either the formal cross-desk or the somewhat less formal cross-corner design. Psychological research suggests clients prefer the cross-corner setting over the formal setting, while lawyers tend to prefer the cross-desk. A workable compromise uses two chairs, one at the corner and one across the desk, and allows the client to select which seat makes him or her more comfortable. This eliminates making assumptions about a client's preferences (which may be wrong). If you are comfortable with it, you might consider directing the client to the corner chair, or indicating that it is the client's choice. Clients unfamiliar with lawyers may assume the interview is supposed to be formal and will automatically select the most formal seating arrangement.

NOTE: See Fey and Goldberg, Legal Interviewing from a Psychological Perspective: an Attorney's Handbook, 14 Willamette L.J. 217 (1977):


Picking the time to interview a witness involves three separate issues: how much time should elapse between the event and the interview, what time during the day is best, and when one witness should be interviewed in relation to others. You may have no control over these factors. By the time you become involved in a case, days, weeks, or months may have passed already. A witness may be available only at a certain time. However, if you do have some room to maneuver, timing can be important.

Interviews should be conducted early in your investigation, as close to the event as possible. Memories fade with time and become clouded with new information. As time passes, chronological and inferential confusion increase. The closer to the event you conduct the interview, the more reliable will be the information. In one study, the ability of witnesses to correctly identify pictures they had been shown dropped from one hundred percent two hours after the event to around fifty percent accuracy four months after the event. However, the big drop in accuracy occurs in the first few days. You must act quickly.

Another reason for interviewing early is to talk with the witness before your opponent does. We have already seen that the way in which questions are asked can affect recall. If you wait until the other side has interviewed a witness, that person's memory may have been altered by the prior questioning.

Selecting a good time of day also can facilitate communication. You always should try to interview a witness at a time when he or she has the fewest competing time demands. Interview witnesses at their convenience, not yours.


To understand how an interview can be conducted most effectively, it is beneficial to divide it into stages. While there are many ways to conceptualize the progressive development of information during an interview, the following five-stage model reflects the collective advice of psychologists:

1. First Stage --- Mutual Evaluation . The first stage of an interview can be called mutual evaluation. This is the introductory stage, in which you meet the witness, break the ice, explain your purpose, and begin the process of building trust and rapport, and in which you and the witness get to know each other. This stage can be quite short, especially if you previously have met the witness. Some interviewers consciously keep it as short as possible, believing it better to get to the facts immediately. Psychologists generally advise that a few minutes devoted to small talk will help make the witness feel more comfortable and relaxed, will help build rapport, and will improve the quality of subsequent information.

The tone of the interview will be influenced by your first contact with the witness. Thus, how you greet the witness is important. If he or she comes to your office, you have a great deal of control. Unless you think it necessary to belittle the witness, you should not keep him or her waiting. If you greet the witness personally, you build up the witness's importance; if you have a secretary show the witness in while you are engaged in some other activity, such as signing a letter, you reinforce the status differences between you.

After you greet the witness, a minute or two of small talk will help break the ice. You may ask the witness if he or she would like a cup of coffee, offer to take the person's coat, discuss sports, or ask the witness about his or her background. It usually is appropriate to take this time to make sure the witness understands the ground rules of the interview --- how long it will take, its confidentiality, what its purposes are, how it will be structured, and whether it will be recorded in written or taped form. Explaining the purpose allows you to begin maximizing facilitators and minimizing communication inhibitors. You can tell the witness he or she is being interviewed because of his or her importance to the case (minimizing ego threat, providing recognition); that you are sorry to have to put him or her through another reliving of a traumatic incident (sympathetic understanding), and so on.

The first few minutes is also the best time to clear up any potential social problems, such as whether smoking is allowed and the names by which you will refer to each other. Discussing these preliminary matters has a benefit beyond clearing up potentially uncomfortable situations. It gives you a few minutes to evaluate the witness; to get used to the witness's speech patterns, vocabulary, accent, and level of nervousness. Having this background information will assist you in anticipating communication inhibitors and evaluating subsequent nonverbal communication cues.

2. Second Stage --- Witness's Overview. The next stage of the interview is to obtain the witness's recollection of the events in narrative form, without interrupting to ask for details. Remember that a witness's free narration, while not likely to be very complete, will be highly accurate. During this stage, your job is to listen, not to talk. You should encourage the witness to talk, facilitate communication, and be careful not to influence inadvertently what the witness thinks is important. It is probably a bad idea to take notes at this stage, since they distract you from listening and may subtly influence the witness's conception of what is important. This phase will provide you with a summary of the events from the witness's perspective and an outline of issues to pursue in detail later in the interview. Obviously, open-ended questions should be used at this stage.

If the witness is articulate, allow the witness to talk without interruption. This helps the witness relax, bolsters self-esteem, and facilitates communication.

Not all witnesses will spontaneously pour forth a long narrative of facts upon merely being asked to tell what happened. Some may be reluctant or shy; some may offer only general conclusions. You may have to take an active role encouraging communication and probing for more detailed information. At this stage, neutral or passive probe techniques should be used --- techniques that will stimulate the witness to talk without influencing the content of what is said. There are three types of neutral probes:

a. Silence. Silence in the middle of an interview is very awkward. There is a natural tendency to fill it in with more conversation. You must be careful not to immediately jump into gaps in conversation with questions. If you let the silence continue, the witness will probably start talking again. Care must be exercised not to overuse silence because it may embarrass a witness who cannot think of what to say next.

b. Noncommittal encouragement. This category includes remarks, gestures, and semi-verbal noises that indicate that you wish the witness to continue. Unlike silence, these probes let the witness know you are listening. This category includes head nods, expectant facial expressions, and brief remarks such as "uh huh," "really," "I see," and "is that so."

c. Neutral questions. The third kind of neutral probe is the nonsuggestive open question which, like asking the witness to tell what happened, starts the narrative again when it has stalled. This category includes questions such as "tell me more," "can you think of anything else," and "what else happened?" To be effective as neutral probes, these questions must not contain topic suggestions.

3. Third Stage --- Starting Point Determination. After the witness has completed his or her overview, you will want to probe more deeply for details. In order to minimize chronological confusion, this model assumes that you will go back over the witness's version of events in chronological order. To do this, you first must identify the point at which the chronology should begin. In some interviews, the starting point will be obvious from the witness's summary version of the events. For instance, if you are interviewing a police officer who responded to a radio call and went to the scene of an accident, the starting point almost certainly will be the radio message. You still must verify that this is the first relevant event, but that will be easy to do by a question such as, "Can you think of anything that happened before the radio call?"

In other situations, determining the chronological starting point may be more difficult. One inhibitor to communication is perceived irrelevancy. If a witness has rejected earlier events as irrelevant or inconsequential, that person may not mention them during the overview stage, and will be unlikely to volunteer them if asked whether anything relevant happened earlier. A witness to a traffic accident may think that the sight of a speeding car is the first relevant event. If you accept this, you may miss out on crucial events that happened earlier --- the witness may have heard the sound of brakes, seen a man on the corner who will turn out to be a valuable eyewitness, or overheard the owner of the car say that the brakes were bad the day before the accident.

The technique for ascertaining the first relevant event is to cause the witness to think about what happened before the main event. Since the witness already has told you what he or she perceives to be relevant, neutral open questions such as, "Tell me what happened earlier that day," may not elicit any new relevant information. At this stage, narrow questions that suggest particular topics will help the witness focus his or her memory. Common sense will help you guess what might have happened earlier so that you can ask about it. However, you must be careful to avoid leading questions. If you say to the witness, "Nothing else happened earlier, did it?," the witness will probably agree with you. It is more productive to ask the witness to focus on specific people, things, and events --- whether he or she had any previous contact with people mentioned, whether he or she had seen people or objects before the main event, or whether the witness knows anything about events logically related to the main one.

4. Fourth Stage --- Detailed Chronology. Once you have obtained an overview of the problem from the witness, and have ascertained the starting point, it is time to probe for details. In most cases, a straight step-by-step chronological order will maximize the completeness of the information. Some system of note taking usually is recommended during this phase, since particular important details, such as the names and addresses of other witnesses, are difficult to remember accurately. During this stage, you must take control of the tempo and the scope of detail of the interview, preventing the witness from jumping ahead, and probing for details. This is known as topic control.

This is when you use the "funnel" sequence of questions, from broad to directed to narrow. During this part of the interview, in which you are seeking more detail, two probes are commonly used:

a. Elaboration. A request for elaboration may take several forms. First, elaboration might imply a need for continuing the "story" or finishing the trend of thought. This would include such probes as, "... and then?" "Then what happened?" "What happened next?" Second, the elaboration might not imply a "moving on" with the story, but merely requests the respondent to say more about the topic at hand. For example: "Tell me more about that." "What else could you say about that?" "Is there anything you would like to add?"

b. Clarification. The clarification probe not only asks for more information on the topic under discussion, but it also specifies the kind of additional information that is needed. A request for clarification may take many specific forms of two general types. First, the interviewer might request that the respondent give a more detailed sequence of events, beginning at a certain point in the action described in the immediately preceding response. For example: "What happened right after you opened the door?" "Where were you just before the lights went out?" "What happened that night after you got home?" Second, the interviewer might probe for more detailed information on some specific aspect, rather than some particular period of time. This would include such probes as: "When did that happen?" "How did you find out about that?" "How did you feel when you saw the building collapse?" "Why do you suppose they did that?"

5. Fifth Stage -- Conclusion. The final stage is to draw the interview to a conclusion. A poorly handled farewell can destroy the delicate rapport you have built up during the interview. In general, you should do three things: 1) Summarize the main facts for the witness to verify. 2) Set an agenda for future meetings and obligations, being specific about any future contacts. 3) Immediately (after the client leaves) write up everything you learned in as much detail as possible.