Witness Coaching by Prosecutors   > See full article in > Pace Law Faculty Publications.

((The below is copied from a news article and some errors were created by the software.))
Bennett L. Gershman*
Witness coaching has been described as the ”dark”l-some
have even called it “dirty secret of the U.S. adversary system.”
It is a practice, some claim, that more than anything else has given
trial lawyers a reputation as purveyors of falsehood^.^ Witnesses
are prepared by lawyers in private, no records are kept, and the
participants do not openly discuss the encounter. Of false or

misleading testimony results, the only persons who know about it
are the participants themsel~es.~ And the capacity of crossexamination
to expose improper coaching is extremely limited?
Given its controversial nature, one would expect the practice
and ethics of witness coaching to have attracted close scrutiny by
courts and commentators. Interestingly, however, the subject has
received relatively modest attenti~n.A~ handful of judicialg and
ethics opinionsI0 have discussed superficially the subject of witness

preparation and coaching. Practitioner manuals typically offer
general guidance on how to prepare witnesses,]’ and occasionally
address tactical and ethical issues involved in coaching.12 Scholarly
commentary has examined the ethical limits of witness
preparation, particularly by differentiating acceptable techniques
from improper techniques, which promote false or misleading
testimony.13 In addition, popular culture occasionally has
dramatized the subject.I4 However, despite a discrete body of
literature devoted to witness preparation generally, there has been
very little discussion by courts and commentators on witness
preparation and coaching by prosecutor^.’^
The absence of critical examination of witness coaching by
prosecutors is puzzling for two reasons. First, there is an
increasing concern among courts, lawmakers, and commentators
that some prosecutors use the adversary system not to serve truth,
but for self-serving purposes.16 According to thls view, the conduct

of some prosecutors in investigating cases and preparing witnesses
to give testimony is undertaken not to ascertain, present, and
protect the truth, but rather to manipulate the truth in order to
secure a conviction.
Second, there is an increasing concern-amply documented
by recent reports of wrongful convictions-that the criminal justice
system is seriously prone to error.” Critics contend that these
errors are attributable to defects in the adversarial trial process,
mostly from incompetent representation by defense lawyers and
trial errors by prosecutor^.^^ However, recent disclosures suggest
that the origin of many, perhaps most, of these miscarriages of
justice occurs before the cases actually reach the courtroom for

trial. Indeed, the inability of criminal trials to produce accurate
results may be attributable in many cases to techniques used by
prosecutors to prepare, shape, and polish the testimony of their
witnesses.
Absent any contemporaneous record of a prosecutor’s pretrial
interaction with witnesses,19 it is exceedingly difficult for observers
to investigate the preparation process to ascertain the extent to
which prosecutors or police may have improperly influenced
witnesses overtly, covertly, or even unwittingly to give false or
misleading testimony.20 Moreover, the difficulty of analyzing the
witness-preparation process is compounded by the failure of courts
and prosecutors to recognize or appreciate how cognitive factors
such as memory, languages, and suggestion can affect the
accuracy and truth of a witness’s testimony.
Given the secrecy surrounding the prosecutor’s preparation of
her witnesses and the inability to review the process meaningfully,
it is virtually impossible to ascertain whether and to what extent
witnesses have been coached by prosecutors and police to give
false or misleading testimony. Nevertheless, inferences can be
drawn from cases, commentary, and empirical evidence to
illuminate this murky process. First, it is indisputable that some
prosecutors coach witnesses with the deliberate objective of
promoting false or misleading testimony. Prosecutors do this
primarily to (1) eliminate inconsistencies between a witness’s
earlier statements and her present testimony, (2) avoid details that

might embarrass the witness and weaken her testimony, and (3)
conceal information that might reveal that the prosecutor has
suppressed evidence.
Additionally, prosecutors have the ability, consciously or
unconsciously, to strengthen the case by questions and suggestions
that cause the witness to fill gaps in memory, eliminate ambiguities
or contradictions, sharpen language, create emphasis, and alter
demeanor. Some witnesses, moreover, are vulnerable to
prosecutorial suggestions, or receptive to prosecutorial cues.
Indeed, the prosecutor’s interaction with the cooperating witness is
a paradigmatic example of the dangers and abuses of coaching.
Finally, the absence of any contemporaneous record of the
prosecutor’s preparation of witnesses encourages improper
coaching by hiding the process from meaningful oversight by
courts or defense counsel. Without some basis to believe that
coaching occurred, a court would not invoke prophylactic
safeguards to detect or prevent coaching. And absent any
documentation of the witness-preparation process, the defense
lawyer has no basis to challenge the witness’s testimony as the
product of improper conduct by the prosecution.
A. Incentives for Coaching
1. Eliminate Discrepancies
A prosecutor bent on obtaining a conviction may attempt to
eliminate any significant inconsistencies in a witness’s testimony.
Some prosecutors overtly influence their witness to alter
materially an earlier, inconsistent version. In Kyles v. Whitle~a, ~~
capital murder case, the prosecutor elicited testimony from Isaac
Smallwood, a key eyewitness who gave an extremely detailed
description of the killing. Smallwood claimed he saw Kyles
struggle with the victim, produce a small, black .32 caliber gun
from his right pocket, shoot the victim, and then drive off in the
victim’s LTD.25 The prosecutor argued to the jury: “Isaac
Smallwood, have you ever seen a better witness[?]”26
However, in a statement Smallwood made to the police
shortly after the killing, which the prosecutor did not disclose to
the defense, Smallwood gave a vastly different account of the
crime. He told the police that he did not see the actual killing, did

not see the assailant outside the victim’s vehicle, and saw the
assailant for the first time driving toward him in a Th~nderbird.~’
The U.S. Supreme Court reversed the conviction because the
prosecutor violated his constitutional obligation to disclose this
information to the defense.28 The majority opinion, however,
noted how Smallwood’s original story apparently had been
“adjust[ed]” by the prosecutor by the time of the Disclosure
of the earlier statements, the majority observed, not only would
have “destroyed confidence in Smallwood’s story,” but also would
have “rais[ed] a substantial implication that the prosecutor had
coached him to give it.”30
A major incentive for prosecutors to use cooperating
witnesses is to support an uncertain but consistent version of the
facts, rather than to confirm an inconsistent version of the facts
that may represent more of the truth.” As an example, one writer
describes a proffer session in which a cooperating witness has
identified several people as being present at a meeting to
distribute The witness fails to identify a particular
individual as being present. The prosecutor, however, firmly
believes from other evidence that this person was present at the
meeting. When asked specifically whether this person was present,

the witness repeats that he was not present. The witness is asked:
“Are you telling me that [he] was not there?” The witness now
knows what the prosecutor wants to hear. After a break, when the
witness is asked again, he now remembers that this individual was
there.
2. Avoid Embarrassing Details
Some prosecutors attempt to influence the testimony of
witnesses to conceal information that might embarrass the witness
and weaken the prosecutor’s case. As a consequence, the witness
may be coached to give testimony that may be literally true but
creates a false or misleading impression, while allowing the
prosecutor to avoid a charge of outright solicitation of perjury.
One method of evasion is to instruct the witness to claim a lack of
memory or knowledge about a question that may be embarrassing
or otherwise harmful.” A prosecutor might reinforce this subtle
message by reassuring the witness that while he has a duty to
answer every question truthfully, an answer such as “I don’t know”
or “I don’t remember” is a perfectly acceptable response when the
witness is not absolutely certain of the answer.34
Another strategy of concealment is to instruct the witness to
answer only the question asked, and not to volunteer
informati~n.~A~n example is Alcorta v. Texas,3Yn which the
Supreme Court reversed the defendant’s murder conviction for
stabbing his wife to death. The defendant admitted the killing but
claimed it occurred in a fit of passion after discovering his wife,

whom he had already suspected of marital infidelity, kissing one
Castelleja late at night in a parked car. Castelleja, the only
eyewitness to the killing, testified that his relationship with the
deceased had been nothing more than a casual friend~hip;~an d
that he had driven her home from work on the night she was
killed.
However, during pretrial preparation, the witness told the
prosecutor that he had had sexual intercourse with the defendant’s
wife on five or six occasions shortly before her death. This fact, if
known or believed by the defendant, would have provided the
defendant with a powerful motive for the killing. The prosecutor
advised the witness that he should not volunteer any information
about sexual intercourse but if specifically asked about it, to
answer truthf~lly.T~h~e prosecutor’s questions at trial, as reflected
in one significant colloquy, were obviously designed to allow the
witness to give literally truthful answers about his relationship with
the deceased while carefully avoiding the subject of his sexual
conduct with the de~eased.~’
3. Conceal Suppressed Evidence
A prosecutor may be motivated to engage in improper
coaching to prevent the revelation of material information that the
prosecutor did not disclose to the defense as required by due
process.”O Needless to say, a prosecutor who is predisposed to
violate his constitutional and ethical obligation to disclose
favorable evidence to a defendant is also capable of molding her
witness’s testimony to protect the nondisclosure from being

discovered.” Indeed, it is arguable that every case in which a
prosecutor deliberately conceals exculpatory evidence from the
defense may also involve an effort by the prosecutor to coach the
witness’s testimony to hide the suppression.
In Walker v. City of New Y~rkf,o~r e~x ample, a prosecutor
almost certainly coached a cooperating witness to give false
testimony to conceal from the defense information that would
have undermined the witness’s credibility. Walker describes a
prosecutor’s debriefing and preparation of a cooperating witness
in an investigation of the robbery of an armored truck and murder
of the truck driver. At the initial proffer session, the witness
identified two individuals as having participated in the crime. The
prosecutor subsequently learned, however, that one of these
alleged accomplices could not have committed the crime because
he was in prison on the date of the robbery. Undeterred, the
prosecutor elicited testimony from the cooperator in the grand
jury and at trial that did not mention a second accomplice. The
decision by the Second Circuit condemned the prosecutor’s failure
to disclose the inconsistency. The court did not discuss the reason
for the witness’s failure to mention the existence of a second
perpetrator, an omission that undoubtedly resulted from careful
coaching by the prosec~tor.~~
B. Cognitive Factors Facilitating Coaching
1. Memory
An extensive body of scientific literature holds that memory is
highly fallible, and the process of memory retrieval and
reconstruction extremely fragile.44 The perception of an event

typically does not leave a single, clear image; it is heavily
influenced by a variety of factors, including the manner in which
the memory is sought to be retrie~ed.M~~an y studies describe the
distorting effects of suggestive q~estioning.~W~h ereas witness
preparation certainly can assist a witness in remembering and
retrieving a truthful recollection, preparation also can distort a
witness’s underlying memory and produce a false rec~llection.~~
And because of the complex nature of memory, it may be difficult
for the witness himself to distinguish between a genuine
recollection of a previously unrecalled fact, and an imagined
recollection based on suggestions from the inter~iewer.~~
Many prosecutors do not appreciate the dangers associated
with retrieving a memory of an event.49 A prosecutor, through the
use of questions and suggestions has the ability to influence a
witness to remember facts and fill gaps that may be inaccurate, but
which the witness may come to believe is the truth.50 In addition,
because of the prosecutor’s unique status as the attorney for the
government, she ordinarily is viewed by the witness as a highly

knowledgeable official who can be trusted to use the facts
re~ponsibly.~’ Indeed, because of the prosecutor’s power and
prestige, the witness may try to conform his recollection of the
event to what the witness believes the prosecutor wants to hear.52
Experts and courts recognize that facts are slippery, and the
process of memory retrieval can be treachero~s.~”~ Justice
Stevens noted in Nix v. White~ide;~ “facts” often are highly
ambiguous and uncertain. To an appellate court after a case has
been tried and the evidence sifted by others,
a particular fact may be as clear and certain as a piece of
crystal or a small diamond. A trial lawyer, however, must
often deal with mixtures of clay and sand. Even a pebble
that seems clear enough at first glance may take on a
different hue in a handful of graveless
And to experts who have studied the psychology of human
memory, a witness’s recollection of an event is a delicate amalgam
that is retrieved, reconstructed, or imagined from this combination
of clay, sand, and gra~el.~”
The potential for witness coaching by prosecutors and police
to alter memory is powerfully illustrated in the Supreme Court’s
decision in Strickler v. Greene.57 In Strickler, a capital murder trial,
a key prosecution witness, Anne Stolzfus, initially told police that
she had only “muddled memories” about a kidnapping in a mall,
and could not identify the perpetrators, the victim, or the

automobile.~A~t trial, however, Stolzfus gave astonishingly
detailed testimony about the event.5y She gave a detailed
description of all three perpetrators? the victim:’ and even
remembered the license plate number of the van.62 Denying the
suggestion that she had learned these details from news reports,
she answered, “I have an exceptionally good memory.”63
‘The process by which Stolzfus’ memory improved so
remarkably, however, was revealed in a series of documents
prepared by a police detective, which were never disclosed to the

defense.@ These documents were based on interviews between
Detective Claytor and Stolzfus in which her memory continued to
expand over time because, she claimed, of “the associations that
[Detective Claytor] helped me make.”65 The Supreme Court
addressed whether the prosecutor violated his constitutional duty
by not disclosing these statements to the defense. The Court
found no violation.66 The Court never discussed the other
important question, i.e., whether the prosecutor and police
encouraged Stolzfus to reconstruct her memory to accord with the
government’s theory of
2. Language and Communication
Prosecutors may attempt to fill gaps by asking questions,
making statements, or displaying evidence that highlights specific
facts.” Indeed, several former prosecutors found it “disturbing”
and “dangerous” that so many prosecutors gave facts to
cooperating witnesses in the course of eliciting information from
them.69 Moreover, the language used by a prosecutor in
communicating those facts to the witness may significantly
influence the witness’s responses. Experts have shown that using
specific words in a question can distort a witness’s recollection or
recounting of an event.70 For example, asking a witness whether

he saw “a” car is much less suggestive than asking the witness
whether he saw “the” car.7′ Similarly, asking the witness whether a
person “smashed” another’s face may produce a decidedly
different response than asking the witness whether a person “hit”
the other person.72
Prosecutors have the ability by their choice of language to
telegraph to the witness specific facts that the prosecutor wants the
witness to say. For example, when a witness initially is uncertain
of the identity of persons who were present at a critical meeting, a
prosecutor could properly try to refresh the witness’s recollection
by asking him whether a named person was present.73 However, if
the witness continues to express doubt, a prosecutor who focuses
repeatedly on whether that person was present is ultimately going
to convey to the witness the prosecutor’s expectations and the
witness eventually will get the message and say it.74
Many prosecutors appear to be unaware of the extent to
which they express verbally or non-verbally a genuine interest in
certain facts, or communicate disappointment when the witness
does not know particular facts, and thereby tip off the witness to
what they want him to say.75 Some prosecutors are not subtle
about this type of communicative message. A prosecutor, for
example, might signal to a cooperating witness, either explicitly or
implicitly, that he is not helping himself by omitting certain details.
According to several former prosecutors, the witness “somehow
now for the first time . . . finds information that helps the
government .”76
Presuppositions or assumptions in questions also can create a
false recollection. For example, after being intensively questioned
by the prosecutor about whether the target of an investigation
displayed a gun, the witness might acknowledge that he
remembers a gun because the story has become implanted in his
mind as a fact, either because he heard about it from others rather
than observing it firsthand, or because the prosecutor strongly

suggested the fact.77 Asking a witness to retell an event over and
over may convince the witness that his story is true.78
There are several communicative techniques that interviewers
use to shape a witness’s testimony.79 Among the most common are
asking leading questions,80 showing a witness a document to
refresh her recollection,R1 informing a witness of what another
witness has said about the incident,x2 and giving the witness a
lecture on the consequences of saying one thing or another.83
C. Dangerous Witnesses
Some witnesses are especially vulnerable to coercive or
suggestive interviewing techniques. The most susceptible of these
witnesses are (1) children, because of their immaturity and
impressionability; (2) identification witnesses, because of the
inherent unreliability of eyewitness testimony; and (3) cooperating
witnesses, because of their enormous incentives to falsify or
embellish.
1. Children
Some witnesses are especially vulnerable to suggestive
interviewing techniques. A familiar and frequently cited example
is the testimony of young children in sexual abuse cases. Many
instances of wrongful convictions are attributable to the testimony
of child witnessesx4 Courts have increasingly scrutinized the
reliability of the testimony of young children for coercive or

suggestive pretrial interviewing techniq~es.~F~o r example, in
Idaho v. the Supreme Court found that a child’s
accusation of sexual abuse was based on suggestive and leading
questioning by an interrogator who had a preconceived idea of
what the child should be disclo~ing.~A~d ditionally, courts have
also focused on the absence of spontaneous recall, the bias of the
interviewer, the use of leading questions, multiple interviews,
incessant questioning, vilification of the defendant, ongoing
contact with peers and references to their statements, and the use
of threats, bribes, and cajoling.88 Courts have also noted the failure
to videotape or otherwise document the initial interview ses~ion.~”
2. Identification Witnesses
Identification witnesses are among the most unreliable
witnesse~.~A s noted above,gl these witnesses may adapt their
testimony to what they believe accords with the prosecutor’s

expectations. Prosecutors may also attempt to “adjust” the
testimony of these witnesses to strengthen the probative impact of
their identifi~ation.~T~h ese witnesses may add facts to their
memory that appear to the witness to be an accurate reproduction
of what the witness originally perceived. This process of “memory
adjustment” is exemplified by the testimony of Ann Stolzfus in
Strickler v. G1-eene,9a~n d Isaac Smallwood in Kyles v. Whitle~.~~
This adjustment often involves the testimony of identifying
witnesses that the prosecutor knows is factually weak or
unreliable, but is presented to the jury with an aura of certainty
and confiden~e.~~
The testimony of several eyewitnesses in the murder trial of
Randall Dale Adams, memorialized in the film documentary The
Thin Blue Line,” offers a dramatic commentary on the dangers of
testimony of identification witnesses. Three rebuttal witnesses-
Emily Miller, her husband R.L. Miller, and Michael Randalloffered
critical testimony identifying Adams as the killer.97 The

testimony was given confidently, with some bravado. However, as depicted in the film, these witnesses appeared to have given
contrived testimony that probably was the product of coaching by
the prose~utor.In~d~e ed, these witnesses’ subsequent narrations of
their accounts of the incident for the camera-a starkly revealing
portrait that captures their venality and deception-is a
devastating commentary of the artificiality of courtroom testimony
and how a prosecutor’s apparent coaching produced a terrible
miscarriage of justice.
3. Cooperating Witnesses
The cooperating witness is probably the most dangerous
prosecution witness of all. No other witness has such an
extraordinary incentive to lie. Furthermore, no other witness has
the capacity to manipulate, mislead, and deceive his investigative
and prosecutorial handlers. For the prosecutor, the cooperating
witness provides the most damaging evidence against a defendant,
is capable of lying convincingly, and typically is believed by the
jury.

The prosecutor’s pretrial coaching of cooperating witnesses is
vulnerable to many of the potential abuses noted above. The
cooperating witness is (1) easily manipulated by coercive and
suggestive interviewing techniques; (2) readily capable of giving
false and embellished testimony with the prosecutor’s knowledge,
acquiescence, indifference, or ignorance; (3) readily capable of
creating false impressions by omissions or memory alterations that
in the absence of any recordation or documentation eludes
disclosure and impeachment; and (4) able to present his testimony
to the jury in a truthful and convincing manner, which because of
the nature of the cooperation process is difficult to impeach
through cross-examination.I0O
A prosecutor has a powerful incentive to accept a
cooperator’s account uncritically.lOl Many prosecutors, if they are
candid, will admit that in some cases they really do not know
whether the cooperator is being truthful or dishonest. This is
particularly the case when a prosecutor lacks evidence to
corroborate the cooperator’s account. Moreover, some
prosecutors have a predetermined view of the facts of a case that
constrains their ability or willingness to assess the cooperator’s
credibility objectively. They may have a theory of the case that
they developed from other evidence or from reliance on the
opinion of the case agent. These prosecutors believe that theory
to be true, and to the extent that the cooperator’s version is
inconsistent with this theory, the prosecutor may conclude that the
cooperator is lying or withholding information.
Cooperators are manipulative, and some prosecutors can be
easily manipulated. Some prosecutors trust their cooperators too

much–one former prosecutor described the relationship as
“falling in love with your ratv-and this mindset skews the
prosecutor’s ability to evaluate the cooperator’s credibility
objectively. These prosecutors may neglect to probe the
cooperator’s story or background intensively to uncover
inconsistencies or outright lies. A recent illustration is United
States v. W~llachi,n~ w~h~ic h a key cooperating witness, Anthony
Guariglia, gave perjured testimony about his gambling habits that
the prosecutors could easily have checked but did not.Io3
Moreover, some prosecutors have a cramped view of their
ethical duty as ministers of justice. They believe that serving
justice means getting convictions and putting bad people in jail.lW
This mindset may be particularly noticeable with younger
prosecutors, whose experience is confined to administering the
federal sentencing guidelines. These prosecutors have been
described as “Guidelines babies.”lo5 They often exhibit a
“mechanistic” and “hardened” view of justice.Io6 They perceive
themselves as cops, and exude a “macho” persona wherein “[tlhey
don’t ask what’s the right thing to do. They just want the right
result.”107
Cooperators appreciate that their value depends on giving the
prosecutor “what they want to hear.”los This message is “drummed
into defendants at the MCC (Metropolitan Correction Center)
that you have got to have good information for the go~ernment.”‘~~
Many professional participants in federal criminal practice believe
that the Federal Sentencing Guidelines, particularly by their ability
to confer unprecedented and enormous rewards on cooperators
who provide law enforcement with “substantial assistance,” create

a powerful incentive for cooperators to exaggerate and falsify
information.’I0
Moreover, some cooperators may not even appreciate the
difference between truth and untruth. Cooperators frequently
come from alien environments of crime and deceit that produce a
mindset, according to some prosecutors, that “[tlruth equals what I
know or what I can be caught at.””‘ Cooperators do not share the
prosecutor’s “obsession with exact facts.”112T hey use language in
a loose, non-literal fashion that allows them to make false or
exaggerated assertions that they might believe to be true. They
might assume, for example, that if they have knowledge of certain
information, it is immaterial how they came to learn it, whether
through personal observation or based on what they may have
heard.ll”

Detecting coachmg is difficult for two reasons. First, there
typically is no verbatim record or other significant documentation
of the interview session. What occurs in this private encounter is
known only to the prosecutor and witness. Second, crossexamination
has a very limited capacity to expose improprieties in
the preparation process. Preventing coaching typically rests on the
integrity and sense of fairness of the individual prosecutor.
Preparation of witnesses, when done properly, is an essential
means of readying the prosecutor and witness for adversarial
testing. By working closely with the witness, the prosecutor is able
(1) to ascertain the truth fully, fairly, and objectively, (2) present
the truth, as she honestly ‘understands it, in an accurate, fair, and
effective manner, and (3) protect the truth from being discredited
and distorted by adversarial attack. These truth-serving goals,
while applicable to all lawyers, would seem to be most clearly
applicable to prosecutors who are unique among lawyers in their
ethical obligation to seek justice rather than merely gain a partisan
advantage.Il4
However, for those prosecutors who do not view their role as
ministers of justice, prophylactic procedures may be necessary to
protect the fact-finding process from improper influences. These
procedures might include: (1) a pretrial taint hearing to expose
witness contamination; (2) expert testimony on memory and
suggestive interviewing techniques; and (3) recording interview
sessions for in camera judicial inspection.
A. Documentation
Witness preparation is done in private. Since there are no
audio or video recordings of the interview process, there is
virtually no way of learning precisely what transpired during the
preparation session. Neither the. witness, the police, nor the
prosecutor will readily acknowledge improper coaching. Nor is it
the practice of prosecutors or police to prepare extensive written

or recorded evidence of the interview that might shed light on
whether the witness was coached.I1To be sure, documentation of
the preparation process, if available, might reveal overt attempts
to shape the witness’s testimony. However, documentation would
not expose subtle cues that might produce false or misleading
testimony that are attributable to a witness’s own “suggestibility,”
“confabulation,” and “memory hardening.”Il6
A prosecutor is legally required to disclose to the defense, for
impeachment purposes, pretrial statements that a government
witness made to the prosecutor, police, or other government
agent.”‘ However, the prosecutor is able to limit the availability of
such statements by a variety of tactics. First, neither the police nor
the prosecutor has any legal obligation to take notes.118 Some
agents as a matter of policy do not take notes specifically to avoid
creating contradicting evidence.lIy Some prosecutors do not

encourage note-taking, and occasionally even forbid government
agents from taking notes.Iz0 According to one former prosecutor,
“[t]hereYsa certain unconscious arrogation of power about it
Another former prosecutor stated: “[Tlhe office lore is don’t take
too many notes or figure out how to take notes so that they are
meaningful to you and no one else. You don’t want a complete set
of materials that you have to disclose.”lZ2 Prosecutors and their
agents typically do not prepare extensive notes, and when they do
take notes, they try to do it in a safe way that avoids discl~sure.’~~
Thus, notes of significant comments, contradictions, and
inconsistencies by a government witness are exempt from
disclosure unless the notes are “a substantially verbatim recital of
an oral statement made by said witness to an agent of the
government and recorded contemporaneously with their making
of such oral statement.”124 These notes may be withheld from the
defense even if they contain significant impeachment evidence
when it is shown that the notes are selections, summaries, or
interpretations by the government agent.125
Also, it is not uncommon for a government witness to be
interviewed by a prosecutor after the witness has been intensively
debriefed by the police. If the police do not take notes, or if they
do not disclose their notes to the prosecutor, the prosecutor may
never know what the witness initially told the police, whether the
witness’s initial account changed, or the extent to which the story
was shaped or polished by police during the initial interview
session. If the police employed the kinds of suggestive or coercive
techniques described above, and absent any available record to
document such conduct, it is unlikely that a prosecutor’s

subsequent probing could effectively recreate the circumstances to
demonstrate any improper influence on the witness’s subsequent
testimony. Moreover, as noted above, the witness herself may be
unaware of the subtle techniques that may have influenced her
testimony.Iz6
B. Cross- Examination
Cross-examination is assumed to be the most important
adversarial safeguard to discovering the truth.lz7 However, there is
no empirical basis for this assumption.128 In Ceders v. United
States,Iz9 the Supreme Court observed that skillful crossexamination
is a vital safeguard to uncovering improper
preparation and coaching of witnesses. The Court assumed that
the line between ethical pretrial preparation and unethical
coaching is easily defined,I3O and that interrogation of the witness
by opposing counsel could disclose improper influences. However,
given the subtle ways that a witness’s testimony can be
manipulated, it is highly unlikely that cross-examination will
disclose ~0aching.l~~
First, one of the cardinal precepts of cross-examination is to
avoid asking questions of which the examiner does not know the
answer.’32 Thus, lacking a factual basis to believe that a witness’s
memory has been manipulated, that an “I don’t remember” is false
or misleading, or that a failure to mention an incriminating fact is
the product of improper coaching, it is unlikely that a crossexaminer
would focus on the discrepancy, or be able to prepare an
effective impeaching strategy about something of which he is

ignorant.133 Moreover, even if a witness’s testimony has been
improperly influenced during the coaching session, the opposing
counsel would have no basis to believe that the witness’s clear and
convincing testimony is the product of an altered memory.
Indeed, as noted above, it may often be the case that the witness
herself is unaware of any improper infl~ence.’~~
Additionally, although it is commonly recognized that the
testimony of a cooperating witness is inherently suspect, and that
the process of preparing and coaching the cooperating witness can
impair the integrity of the truth-finding process, cross-examination
is made even more difficult when the cooperating witness has been
carefully coached to testify that any benefit is speculative,
uncertain, and contingent upon his giving truthful testimony in
accordance with his cooperation agreement.’35
C. Protocol for Witness Preparation
There is nothing wrong with a prosecutor assisting a witness
to give testimony truthfully and effectively. However, under their
obligation to serve justice, prosecutors should be able to regulate
their own conduct to insure that witnesses are not exposed to
suggestive questioning that may create false or misleading
testimony. Prosecutors should be trained and supervised in
interviewing protocols, the vulnerabilities of certain witnesses, and
the psychological literature relating to memory, language, and
comm~nication.’~~
The following is a protocol for witness preparation by
prosecutors. It is based largely on this writer’s experience as a
prosecutor. It differs in several important respects from advice
contained in practitioner-oriented publications typically addressed

to civil litigation attorneys.i37 In preparing a witness for testifying
at trial, a prosecutor should:
1. Demand Truth
Advise the witness to tell only the truth as he knows it, not
what he thinks he knows, or what someone else knows.’38 Advise
the witness not to embellish facts, or fill in gaps of memory, and
that if he does not remember something, to say
2. Be Objective
Evaluate the witness’s story objectively to determine its
accuracy and believability. Ask open-ended questions initially,
and use more specific questions after the witness has given a
complete account of the event.I4O Never put words in the witness’s
mouth or suggest answers.I4′ Know as much as possible about the
witness’s background and any interest the witness might have to
falsify, and probe these areas carefully without suggesting any
desired response.

Explain courtroom procedures, where the witness will sit, and
the order of q~esti0ning.lD~o~ not advise an identification witness
where the defendant will be sitting.'” Advise the witness to speak
in a loud, clear voice so the jurors can hear what is being said, and
to make eye contact with the jury when appr~priate.’A~d~v ise the
witness to sit straight, avoid distracting body language, and dress
appr0priate1y.l~~
4. Rehearse Direct Examination
Write out direct examination questions ahead of time and
rehearse specific questions with the witness.'” The witness should
be encouraged to use his own words whenever possible, and not to
use slang or offensive expressions.14′ Control the witness’s answers
to the extent of preventing the witness from giving long, rambling
narratives, and ensure that the witness does not violate rules of

evidence. If tangible evidence is being introduced, show the
evidence to the witness and go through the process of laying the
foundation. Have the witness do whatever demonstration he will
be asked to do in the courtroom.
5. Reconcile Inconsistencies
Cautiously try to reconcile the witness’s testimony with other
evidence. This may include prodding the witness’s recollection
with prior statements that the witness made, or referring to other
facts in the case, including the testimony of other witnesses.14s
Never suggest what a witness should say, or telegraph what the
witness should say, by revealing inconsistencies or weaknesses in
the witness’s te~tim0ny.I~R~ev iew the witness’s prior testimony
with the witness, but do not allow the witness to read any prior
testimony unless absolutely necessary.150
6. Prepare Cross-Examination
Prepare for cross-examination by going through all
discrepancies in the witness’s prior testimony or other
statements.I5l Try to ensure that the witness can correct or explain

each discrepancy. Go through a witness’s criminal background
and any interest the witness may have in testifying. Make sure that
the witness acknowledges prior convictions and bad acts. A
witness should be encouraged not to volunteer information, to
limit answers as much as possible to “yes” or “no,” not to guess or
embellish an answer, to make an effort to remember, and not to be
afraid to acknowledge a mistake.152T he witness should be told not
to look at the prosecutor during cross-examination, not to answer
a question when the prosecutor stands up to object, and to answer
in a positive tone if he is asked whether he spoke to the police or
prosecutor about the case. A prosecutor should do a mock crossexamination
with the witness to anticipate what defense counsel
will likely ask.153
D. Remedies
1. Pretrial Taint Hearing
A pretrial “taint” hearing should be required when there is
some basis to believe that a witness’s testimony has been
improperly influenced by suggestive or coercive interviewing
techniques. Such a hearing is not unusual. It has been authorized
in many instances in whlch police or prosecutorial conduct has
placed the integrity of the fact-finding process into question and
there is a need for the procedural protection of a pretrial hearing
to exclude from a potential prosecution the prejudicial effects of
tainted evidence. Thus, pretrial hearings have been employed to
determine the admissibility of in-court identification testimony
because of pretrial suggestivene~s,s~ta~t~em ents of children in
sexual abuse cases,Is5 hypnotically-recalled in-court testimony,lS6

breathalyzer evidence because of prior falsified police breathalyzer
reports,ls7 and evidence following police investigatory
misconduct .Isx
A pretrial taint hearing into the reliability of a witness’s
testimony based on pretrial suggestiveness should consider all
relevant circumstances, including any inconsistency between a
witness’s statements, the interest or motivation of a witness to
falsify, the presence or absence of corroboration, the nature of the
corroboration, the inherent believability of the statements, the
existence of any documentation of the debriefing and preparation
sessions, the rewards and other inducements to testify, the scope of
punishment to which a cooperating witness may be exposed in the
absence of cooperation, the manner and form of the questioning,
the number of interview sessions, and the person or persons
present when the statements were made.
Defense counsel, in order to obtain a pretrial hearing, would
have the burden initially of making some factual showing that a
witness has been subjected to improper conduct by the prosecution
that has caused the witness to adjust his testimony. In making its
determination, a court should consider whether, under all the
circumstances, the interview and preparation sessions give rise to a
substantial likelihood of false, inaccurate, or misleading testimony.
2. Expert Testimony
Given the capacity of the witness-preparation process to
produce false or distorted testimony, courts should allow experts
in cognitive psychology to testify how memory, language, and
communication can produce false, inaccurate, or misleading
testimony.Is’ Rule 702 of the Federal Rules of Evidence authorizes

the use of scientific testimony by an expert to assist the jury in
understanding the extent to which suggestive influences in the
debriefing and coaching of witnesses may produce false or
inaccurate testimony.160 AS now occurs with eyewitness
te~tirnony,’t~h~e expert could identify the factors that influence
perception and memory, the extent to which witnesses are
susceptible to suggestive influences, and how pretrial interviewing
techniques are likely to produce false, inaccurate, or embellished
te~tirn0ny.lC~~on trary to the belief of many jurors, an expert could
testify that there is no necessary correlation between a witness’s
confidence and the accuracy of her te~tim0ny.l~E~xp erts could
also counter a belief held by some jurors that witnesses have a
better memory for dramatic events.16″
3. Recording
To enable the defendant to challenge the veracity of the
witness effectively, and a jury to assess his credibility, all
interviews with potential trial witnesses should be electronically
recorded either by audio or videotaping. Videotaping would be
preferable to sound recording as it would depict the physical
interaction and body language of the participants. The use of such
a procedural safeguard is unusual, but hardly novel. Videotaping
of interview sessions with child witnesses is not
Moreover, videotaping has been used when it is important to
document whether the government used unfair tactics to produce
evidence, such as a defendant’s confession,16o6r for interrogations
conducted before and after hypnosis.I6′

Recording the interview session is essential to disclose the
presence or extent of the different types of suggestive influences
discussed above. Taping would reveal overt attempts to influence
the witness’s testimony by use of leading questions or other cues
that alert the witness to the expected answer. Whereas recording
of the sessions would not necessarily reveal whether a witness’s
story was a fabrication from the start, it might demonstrate
whether the witness embellished his story to please the
government or filled in details to make the story more complete or
persuasive, and the extent to which his story crystallized and
became more confident over several interview sessions.
It may be that courts should have the authority to conduct an
in camera inspection of the recording, and preclude the use of any
portions that contain embarrassing or sensitive material. To be
sure, a prosecutor could properly seek to preclude from disclosure
statements by a cooperating witness that might compromise an
ongoing investigation. A court also could limit the use of the
recorded interview session to those portions that reveal that a
witness is trying to please the interviewer, confabulating the story
by appearing to fill in details to make the story more coherent and
complete, or demonstrating memory “hardening” by appearing to
suddenly and confidently remember new details.
The preparation of witnesses for testimony at trial is a
necessary and acceptable part of the prosecutor’s function.
However, some prosecutors engage in conduct that goes beyond
permissible trial preparation. These prosecutors either overtly,
covertly or even inadvertently, cause witnesses to give testimony
that is false, inaccurate, or misleading. And given that witnesscoaching
is done in private, there is usually no evidence of
improper conduct. Therefore, the ability of cross-examination to
reveal such improper conduct is extremely limited.
Coaching typically is accomplished through memory
reconstruction, suggestions that improve testimony, and cues that
alter testimonial language. Some witnesses such as children,
identification witnesses, and cooperating witnesses are highly
susceptible to coaching. These witnesses are capable of adjusting
their testimony based on leading, suggestive, coercive or
intimidating questions or statements. Furthermore, the prestige

and power of the prosecutor enhances her ability to influence the
witness’s testimony improperly.
Given the potential of witness coaching to skew the factfinding
process, this Article offers several suggestions to expose
improper influences and prevent false or inaccurate testimony.
These include a pretrial taint hearing when there is some basis to
believe that a witness has been improperly influenced, expert
testimony to assist the jury in understanding the vulnerability of
memory and the dangers to accurate testimony from certain types
of interviewing techniques, and electronically recording witnesspreparation
sessions. Any or all of these recommendations, if
adopted, would protect the fact-finding from overzealous conduct.