Handbook of Psychology of Investigative Interviewing: Current Developments and Future Directions

by Ray Bull, Tim Valentine, Dr Tom Williamson

  have said in previous accounts, while truth - tellers will attempt to reconstruct

  Truthfulness in Witnesses’ and Suspects’ Reports

  291

  an event they actually experienced. Because deceptive suspects possess the

  common - sense understanding that fewer details are easier to remember than

  multiple details, they will make their statements as short as possible in order

  to avoid contradictions and inconsistencies over repeated interviews. Thus, our

  fi rst hypothesis was that deceptive suspects would give shorter statements and

  omit more details than truthful suspects across repeated interviews.

  The second major interest of this research was the investigation of the

  impact of deception on recall time and duration estimates, a focus of interest

  that has yet to be empirically explored by researchers. An individual ’ s ability

  to provide an accurate account of both where he or she was and the times or

  durations of these events has practical forensic implications (Loftus, et al .,

  1987 ; Jackson, Michon & Melchior, 1993 ; Yarmey, 2000 ; Pedersen & Wright,

  2002 ). If a suspect provides alibi evidence of a timeframe of a critical event

  that overestimates or underestimates the actual duration or confl icts with evi-

  dence from witnesses ’ duration estimates of the same event, police can use the

  temporal statement as an investigative tool. Theoretical and empirical investi-

  gation of duration estimations date back to the nineteenth century, with

  Vierordt ’ s (1868) discovery that short intervals tend to be overestimated and

  longer ones underestimated. Consistent with Vierordt

  ’ s Law, forensically

  related research has shown that witnesses tend on average to overestimate the

  duration of relatively short events lasting a few minutes or less (e.g., Loftus

  et al ., 1987 ; Yarmey & Yarmey, 1997 ) and underestimate relatively long events

  lasting 20 minutes and more (Yarmey, 1990 ).

  People perceive and recall the duration of events subjectively and their

  estimations of time are infl uenced by cognitive, affective and subjective prefer-

  ence (Carmichael, 1997 ). Deceptive suspects have different purposes, goals

  and agendas from truthful suspects. That is, deceptive suspects probably wish

  to distort their time reports and distance themselves from the period in which

  the crime occurred. Deceptive participants in this study were told to be cred-

  ible in their reports. As a consequence they would probably make duration

  estimations constrained somewhat by reality rather than construct fantasy - type

  fabrications (Johnson & Raye, 1981 ; Johnson & Sherman, 1990 ). Whereas

  the research literature would support the prediction that truthful participants

  would overestimate the duration of their involvement in short - term, mundane

  activities (see Yarmey,

  2000 ), it was uncertain whether deceptive suspects

  would overestimate or underestimate the duration of both routine activities

  and the critical incident. It also was uncertain whether deceptive participants

  would differ from truthful participants in estimating the specifi c time in which

  the critical incident began.

  Research p articipants and p rocedures

  Nineteen men and women were randomly assigned to a truthful condition and

  17 to a deceptive condition (overall

  M age

  = 27.61 years,

  SD = 10.34).

  All participants (14 men and 22 women) were city - dwellers and worked in

  292

  Handbook of Psychology of Investigative Interviewing

  different occupations (business, homemakers, educators, college students, and

  so forth). The experiment used a 2 (Group: Deceptive, Truthful) × 3 (Repeated

  interviews: one - day, one - week, two - week retention interval) mixed - group fac-

  torial design.

  Two female psychology students acted as both observers and as mock police

  offi cers for the experiment. The observer accompanied the participant for a

  two - hour period during which time notes were taken of each change in an

  event or activity. Observations were done outside of normal working hours,

  but the participant was free to do any type of activity. Common activities

  included eating, household chores, and so forth. The observer did not interact

  with the participant during this period, except to announce that they were

  ‘ going for a walk

  ’ . The procedure involved taking on

  - the - spot records of

  behaviour as it occurred for two continuous hours. This record was then

  compared to the participant ’ s later report of his or her activities in an attempt

  to establish the accuracy or distortion of time - delayed recall. The observer

  recorded all activities initiated by the participant and the amount of time

  engaged in each activity in chronological order. Timing of activities was

  rounded off and recorded to the nearest minute. The level of detail for an

  activity unit was to be such that someone reading the description could visual-

  ize the events as they occurred. An activity unit of behaviour was defi ned as a

  natural and meaningful action with a start and ending and which lasted

  approximately one minute or longer in duration.

  Approximately 40 minutes into the observation session the observer

  announced that they were leaving the residence (assuming they were indoors

  at this time) and going for walk. Once outside the observer placed an envelope

  on the ground as if someone had dropped it. The participant had earlier

  been instructed to pick up the envelope, read its contents and then put the

  ‘ found ’ materials in their pocket or purse. Participants in the truthful condi-

  tion found an envelope containing a one - page advertisement fl yer. This was

  accompanied by a note stating that the participant was to read the ad for

  information purposes only, but was not expected to memorize it. Participants

  in the deceptive condition picked up an envelope containing a bank note

  worth $25,000 payable on demand to the bearer. An accompanying note

  informed the participant that ‘ although the money does not belong to you,

  you will keep it rather than trying to locate its owner or hand it over to the

  police. From this point on you realize that you must protect yourself. No

  one, particularly the police, should know that you have kept the money. When

  you are questioned about the lost note you will be trying to fool the inter-

  viewer (a mock police offi cer) but will try to be credible and convincing in

  your reports ’ .

  After ‘ fi nding ’ the lost materials all participants were instructed to carry on

  their activities as they saw fi t during the rest of the observation period.

  Participants also were told that the observer and mock police offi cer did not

  know which group they were in and they were not to reveal this information.

  The participants were interviewed one day, one week and two weeks later. The

  Truthfulness in Witnesses’ and Suspects’ Reports

  293

  mock police offi cer told the participants that a witness had reported that they

  in the general vicinity of where the envelopes were lost and that the police

  wanted them to answer a few questions about their whereabout
s at the approx-

  imate time of the incident. The participants were asked to put in writing all

  the activities they could recall engaging in during the two - hour observation

  period and the amount of time spent performing each one, or the time it

  started and stopped. They were told to recall those types of activities lasting

  approximately one minute or longer, or any activity they considered to be

  distinct or important, regardless of its duration. In order to allow participants

  to determine their own order of free recall, they were told that the chronologi-

  cal order of their recall was not important. When participants fi nished their

  report they were asked if they remembered being told by the observer: ‘ We

  are going for a walk ’ . If the participants said ‘ yes ’ , they were asked to estimate

  the specifi c time of day that occurred.

  Results and d iscussion

  To summarize our fi ndings, the analyses of free recall scores showed that, in

  contrast to truthful suspects, the reports of deceptive suspects declined sig-

  nifi cantly from the initial interview compared to one and two weeks later.

  Signifi cant differences were found for: the total number of activities reported

  regardless of accuracy; the percentage correspondence of activities scores

  (that is, the agreement between the number of correct activities observed by

  the research assistant and reported by the participant converted into ratio

  scores); and the level of precision of details reported. Contrary to our expecta-

  tions, truthful suspects did not show an increase in the amount of correct

  information reported across interviews (hypermnesia) or new information

  recalled across interviews (reminiscence effects). Furthermore, no signifi cant

  difference was found between truth

  - tellers and deceptive suspects in the

  chronological structure of their free narrations of the two - hour observation

  period.

  No signifi cant differences were found between truthful suspects and decep-

  tive suspects in their duration estimations for seven categories entailing routine

  mundane activities. However, on the critical event involving a walk, a signifi -

  cant difference in duration estimations was obtained. Whereas truthful suspects

  on average overestimated this short - duration event, deceptive suspects were

  highly consistent in signifi cantly underestimating the duration of the critical

  event. These differences occurred in spite of the fact that no signifi cant differ-

  ences were found between the two groups in their estimation of the time when

  the walk event began.

  In line with our interpretation of Granhag & Str ö mwall ’ s (2000) ‘ repeat

  vs. reconstruct

  ’ hypothesis, this experiment showed that the number and

  percentage accuracy of activities and the number of details reported by

  deceptive suspects decreased across interviews. In order to remember what was

  294

  Handbook of Psychology of Investigative Interviewing

  previously said, a deceptive suspect probably would attempt to follow a remem-

  bered script. In subsequent interviews, deceptive criminal suspects are repeat-

  ing an originally deceptive account rather than trying to reconstruct an account

  that actually occurred. Thus, they will use fewer words and give fewer details

  with repeated interviews in an attempt to ensure that their recall will appear

  accurate and credible. Coarsely grained but credible responses, as opposed to

  more specifi c responses, would appear to be the preferred and wiser choice for

  deceptive suspects. In contrast, because memory is reconstructive, it would be

  expected that truthful reports will gain, lose and change information over time

  and, in an attempt to be more informative, will show more fi ne - grained

  responses (Loftus, 1979 ; Baddeley, 1990 ; Goldsmith, Koriat & Weinberg -

  Eliezer, 2002 ).

  The fact that the mean scores for correspondence of activities at one day were

  highly similar for both truthful and deceptive participants deserves particular

  consideration. Both truthful and deceptive suspects reported approximately

  one - third of the total possible number of activities observed by the research

  assistant. The ‘ repeat vs. reconstruct ’ hypothesis would predict that truthful

  accounts should contain more details than deceptive accounts. So what can

  account for these fi ndings? As Granhag & Str ö mwall (2000) theorize in their

  own study, participants in the current study were asked to distort an original

  event rather than create a complete fabrication. Distortions may in general

  contain a greater number of details than pure fi ction, potentially negating the

  effects of deception on the number of details initially reported by participants.

  Furthermore, in this investigation both truthful and deceptive suspects oper-

  ated in the very familiar surroundings of their own home and nearby streets.

  The context of the events for both deceptive and truthful suspects would have

  been relatively similar in terms of visual, auditory and spatial details. Performing

  a mock crime in familiar surroundings would have been highly conducive

  for the construction of logical and well - rehearsed reports by the deceptive

  suspects. Such familiarity would facilitate credible and logical stories by both

  deceptive and truthful suspects.

  Additionally, our fi ndings support the basic premise of the hypothesis put

  forth by Undeutsch (1982) , which argues that fewer details are provided by

  deceptive individuals because it is relatively easier to recall fewer details later.

  Thus, in actual forensic settings in which a criminal suspect is interviewed

  repeatedly, the number of details and the length of particular accounts over

  time may be important clues to veracity. Note that these are not absolute

  quantities but relative ones in which the point of comparison is the individual

  him - or herself. Individuals vary greatly in the number of words and details

  that they use in general, and this is based on a variety of factors, including

  intelligence, educational level and language profi ciency. In an initial interview,

  it is thus diffi cult if not impossible to predict how many words and details

  might be used. But over time with repeated interviews one may see decreases

  in quantities and fi ne details, and these decreases can serve as important indica-

  tors of deception.

  Truthfulness in Witnesses’ and Suspects’ Reports

  295

  Interestingly, neither truthful nor deceptive suspects reported what they did

  during the walk, apart from stating that they were going for the walk. It is

  likely that no disclosures were given regarding activities undertaken during the

  walk because participants were repeatedly instructed by the research assistant

  not to reveal into which group they had been placed. These fi ndings suggest

  that our participants were highly compliant in following the experimental

  instructions. Furthermore, the methodology used in this investigation (written

  free responses without direct questioning from the mock police offi cer) prob-

  ably contributed to the lack of details reported by participants during the walk.

  Follow - up studies of the present investiga
tion should investigate the differ-

  ences between truthful and deceptive suspects as a function of type of interview

  and suspect responses.

  This study provides preliminary support for the equal ease with which

  both deceptive and truthful participants can effectively account for their time

  during an event or series of events. Thus, asking a criminal suspect when an

  event took place or how long that event lasted may not be a reliable clue to

  deception if the act of deception has no differential effect on one ’ s ability to

  estimate its duration. Recall of event duration is seen by legal experts to have

  potential signifi cance in the courtroom. In her article directed to defence

  lawyers, Steele (2004) advises that attention should be paid to ‘ Event Duration

  … for how long did the witness see the culprit? … the witness ’ estimate is

  almost certain to be wrong; generally, the witness will overestimate durations ’

  (p. 10). Although the witnesses referred to by Steele are victims of or

  bystanders to a crime, it is also true that guilty suspects are witnesses to

  their own involvement in a crime. In the present investigation truthful sus-

  pects and deceptive suspects did not differ in their duration estimations for

  routine, mundane events. However, deceptive participants in contrast to

  truthful participants signifi cantly and consistently underestimated the duration

  of ‘ the walk ’ (the point during the study when participants were asked to go

  out and fi nd either the bank note or advertisement). Note that the research

  literature on duration estimations for events lasting three minutes or less is

  highly reliable, that is, short - duration events are on average overestimated (see

  Yarmey, 2000 ). The implications of this fi nding are potentially considerable,

  particularly given the lack of signifi cant fi ndings for any of the other duration

  estimates. The key experimental manipulation of being assigned to either

  the truthful or deceptive condition is made most salient to participants when

  they are instructed to go out for the walk. Although they already know

  whether they will be behaving truthfully or deceptively, ‘ the walk ’ is the point

  at which these instructions actually become an event that the participants must

  act upon.

  The practical psycho - legal implications of the current study centre on the