Questions asked during investigative
interviewing, their relationship with
obtained detail and the effect of short
training
Masterthesis Forensic Clinical Psychology
Femke Brankaert (10559388)
Supervisor UvA: B. Verschuere
Supervisor University of Portsmouth: A. Vrij
Date: September 2014
Abstract
This research examined which types of questions are typically asked by police officers during interviews (1), how much information they elicit (2) and if training has an effect on questions
asked (3). The types of questions distinguished are open-ended, probing,
encourager/acknowledgement (which we label “appropriate questions”), and closed, forced choice, leading, echo, opinion/statement and multiple at once questions (which we label “inappropriate” questions). The 50 police officers were not instructed to use a particular standard interview protocol when interviewing participants (half truthful, half deceptive) about witnessing a secret meeting or trip to a pub while a laptop had been stolen. Some police officers had a one-day training about best practices for investigative interviewing and interviewed someone before and after they had training. The data showed that police officers asked more inappropriate questions than appropriate ones but this was affected by the training in a positive way, as police officers used less inappropriate and more appropriate questions after the training than before. Furthermore, closed questions were used more frequently than every other type of question. Open-ended questions were only used less frequently than probing and closed questions, which is not in line with previous research. The amount of detail obtained was affected by the type of questions asked. Appropriate questions had a significant positive relationship with obtained detail, while inappropriate questions had a negative relationship which was not found significant. The implications of these findings and the limitations of the study are discussed. Most important limitation was the lack of ecological validity in this research. Despite this, we found interesting findings in our study, which may have important consequences for the practical work of police officers. More extensive research should look at the reason why police officers tend to use more inappropriate questions than appropriate questions and should look closely to the effect of training and after how long training effects expire.
Introduction
Interviewing witnesses and suspects is a common and important part of the work of police officers in any investigation (Snook, Luther, Quinlan & Milne, 2012). A vital part to investigative interviewing is asking questions. Asking the right questions could lead to getting the information needed and help the police to uncover what took place at the scene of the crime (Shawyer, Milne & Bull, 2009). For this reason police officers should ask questions that lead to the most amount of detail, so that they have more evidence and information to help aid the investigation. Interestingly, more information usually leads to a more accurate decision about the truthfulness of the interviewee (Vrij, 2008; Vrij, Granhag & Porter, 2010). Unfortunately, research has shown that police officers often struggle to ask the right questions and sometimes are not even aware of the types of questions that would prove most useful during an investigative interview (Soukara, Bull, Vrij, Turner & Cherryman, 2009). Evidence from court cases from around the world have revealed police officers to be less competent in their questioning techniques than would be expected (Bull & Milne, 2004).
One possible explanation for the fact that police officers do not use best practice known to date could be that police training for investigative interviewing is relatively new. In fact research into this area only began in the early nineties, so the training of police officers based on academic research is a fairly new concept (Griffiths & Milne, 2006). Before that, police officers were only taught how to interrogate someone by learning on the job. A new recruit would learn by observing a more experienced officer as at the time there standard training did not exist (Griffiths & Milne, 2006). In addition, the collaboration between academics and the police has often proved challenging, although it has improved over the last couple of decades (Griffith & Milne, 2006). As with every practical field, it normally takes some time for new research findings to be recognised and incorporated into police trainings. As a result police officers are not trained according to the newest findings of research and still use old or outdated techniques. Therefore, although police officers are working to the best of their abilities they are not utilising the expertise or research that is currently being developed and are therefore at a disadvantage when it comes to developing their interview skills.
Current research has already shown that asking open-ended questions (questions mostly beginning with “Tell”, “Describe” or “Explain”) or probing questions (questions beginning with one of the five WH words “who”, “where”, “when”, “which” or “why” or
beginning with “how”) are the most productive. It does not matter what the interviewee said or did before or what the officer already knows about the interviewee, interviewees give more information when asked open-ended or probing questions (Griffiths & Milne, 2006; Milne & Bull, 1999, 2006; Myklebust & Bjorklund, 2006; Shepherd, 2007; Sternberg, Lamb, Orbarch, Esplin, Mitchell, 2001). Open-ended questions also tend to lead to the free recall of events, which results in more details being recalled (Snook et al., 2012). Dunning and Stern (1992) found that witnesses tend to recall more information and more accurate and complete information when asked open-ended questions in interviews. In addition, the elicited information from open-ended questions is more accurate and gives more detailed information making it more valuable for the investigation (Fisher, Falkner, Trevisan & McCauley, 2000), possible because interviewees are allowed to collect their thoughts in their own way, instead of being distracted by the interviewer asking other directed questions (Brewer & Williamson, 2005). Finally, open-ended questions give the interviewee time to think, which will lead to more elaborate retrieval of memory (Brewer & Williamson, 2005).
Researchers agree that closed questions lead to less information and will lead people to answer the questions in an unreliable way (Myklebust & Bjorklund, 2006; Oxburgh, Ost & Cherryman, 2012). Interviewees will in fact answer with only a “yes” or “no” response, which gives no opportunity to talk freely or obtain any additional information. People even give “yes” or “no” answers, if they actually do not know the answer (Waterman, Blades & Spencer, 2001), which could lead to incorrect answers being given.
Research also revealed that leading questions are inappropriate to use during investigative interviewing (Griffiths & Milne, 2006), mostly because leading questions create a high level of suggestibility. Leading questions communicate expectations and premises to an interviewee (Baxter, Charles, Martin & McGroarty, 2013). People tend to give desirable answers in high stake situations such as an investigative interview, because they do not want to contradict the interviewer (Gudjonsson & Clark, 1986). This could well lead to unreliable information. Another problem with leading questions is that the often misleading information embedded in these questions can be incorporated into a person’s memory and could eventually lead to false recall in later stages of the interview process (Gudjonsson & Clark, 1986; Loftus, 1979).
Following Griffiths and Milne’s (2006) categorisation of question types in forensic investigative interviewing, other frequently used questions are forced choice questions, echo
questions, encourager or acknowledgement questions, multiple questions at once and opinion or statement questions. Forced questions force the interviewee to choose between two or more options given by the interviewer (Snook et al., 2012). Echo questions are usually used to express disbelief or confirm information already given by the interviewee. An echo question repeats the information given by the interviewee but is phrased as a question (Snook et al., 2012). A multiple at once question means that the interviewer asks multiple questions at the same time without giving the interviewee time to respond between the questions (Snook et al., 2012). In this case the interviewee usually only answers the last asked question or chooses only one question to answer. Opinion/statement questions are expressions of opinion or put a statement to the interviewee (Snook et al., 2012). Table 1 gives a schematic overview of the different types of questions with some examples for each type of question. Oxburgh et al. (2012), used this categorisation of questions and found that open-ended, probing and encourager/acknowledgement questions (so called appropriate questions) related to obtaining more information. They also found that closed, leading, multiple at once, forced choice, echo and opinion/statement questions (so called inappropriate questions) resulted in less information being obtained in investigative interviews with suspects of sexual offences. The categorisation of inappropriate and appropriate questions is based on the effect of the question type on amount of detail.
The above suggests that according to researchers and practitioners in the field of forensic interviewing best practice would be to ask open-ended questions and maintain this style throughout the interview. Asking as many open-ended questions as possible is important as it allows the interviewer to get as much information as possible (Oxburgh et al., 2012; Poole & Lamb, 1998; Read, Powell, Kebell & Milne, 2009). Leading, closed and other inappropriate questions should be avoided where possible, because they lead to a lesser amount of information and possibly result in unreliable or untruthful answers (Baxter et al., 2013; Oxburgh, et al., 2012; Snook et al., 2012; Waterman et al., 2001). Despite these clear findings and suggestions, research has shown that asking open-ended questions is not common in investigative interviewing in the forensic setting. Instead, police officers tend to use closed, forced choice, multiple at once and other inappropriate questions (Baldwin, 1993; Clarke & Milne, 2001; Davies, Westcott & Horan, 2000; Myklebust & Bjorklund, 2006; Smith, Powell & Lum, 2009; Snook et al., 2012; Snook & Keating, 2011). In fact, police officers ask too much inappropriate questions, which leads to less valuable information obtained during interviews (Oxburgh et al., 2012). Asking information-gathering questions,
like open-ended questions, should also be used more because these are known to maximize the chance of detecting liars (Vrij et al., 2010). This means that real-life investigative interviewing is not based on the best practice known to date, which is a real loss for police departments, witnesses, suspects and the whole of society (Smith, et al., 2009; Snook et al.,
2012). Important decisions are based on the investigative interviews of police officers every
day and people’s lives can be affected by it in a major way. It is also concerning that the investigators may not have the ability to obtain complete and accurate information during forensic interviews (Snook et al., 2012). On the positive side, research has revealed that giving police officers training on questioning techniques has a positive effect on the questions asked by them and suggests that police officers can learn new interview techniques (McGurk, Carr & McGurk, 1993; Walsh & Milne, 2008). However, even trained interviewers struggle to implement the best practices that they learned (Soukara, Bull, Vrij, Turner & Cherryman, 2009).
The first aim of the present study is to explore which types of questions police officers ask during interviews between police officers and suspects or witnesses. The prediction is that, police officers will ask significantly more inappropriate questions than appropriate ones, echoing the findings of previous research (Oxburgh et al., 2012; Snook et al., 2012) (Hypothesis 1). A further prediction is that, police-officers use open-ended questions the least and closed questions the most in relation to the other types of questions (Myklebust & Bjorklund, 2006; Oxburgh et al., 2012; Snook et al., 2012) (Hypothesis 2 and 3). Hypothesis 1 to 3 are to replicate the findings of previous research, with the difference of standardized circumstances and using no specific interview technique. The second aim of this study was to get a closer look at the relationship between the amount of detail obtained and the questions asked in an investigative interview. The expectation is that the more appropriate questions (e.g. open-ended, probing and encourager/acknowledgement questions) asked, the more relevant and detailed information the participants will disclose (Hypothesis 4). In addition, it is predicted, that the more inappropriate questions (e.g. forced choice, leading, closed, echo, multiple at once and opinion/statement questions) asked, the less detailed information interviewees will disclose, echoing previous research (Oxburgh et al., 2012) (Hypothesis 5). Finally, we will take a look at the difference between the type of questions asked by the police officers before they received a one-day training session on detecting deception techniques and after this training. The training focuses on questions officers should ask in suspect interviews that encourage suspects to disclose more information. Important in this
context is that during the training it was repeatedly emphasised that asking open-ended questions in particular is the best way to obtain as much accurate information as possible from truthful suspects. For this reason, the expectation is that the police officers will ask more appropriate and less inappropriate questions during the interview after the one-day training session compared with before (Hypothesis 6 and 7).
Method
Participants
Following ethical approval from the FBI the participants in this study were approached. Two types of participants were used. The interviewers were police officers working in the UK. The interviewees were students and staff from the University of Portsmouth.
(a)Interviewers
The interviewers were contacted with information about the study via email and volunteered to take part. There was no reward of any kind for the participating police officers. Eventually 50 police officers took part, of whom 58.8% were male. Their average age was M = 43.6 years (SD = 9.5) and their age ranged from 27 to 58 years. All police officers received a one-day workshop about the best practices in questioning techniques, lie detection and underlying research about these subjects and then interviewed a mock suspect or witness, and 24 of the 50 police officers also interviewed a mock suspect or witness before the training.
(b)Interviewees
Interviewees were recruited via advertisements on the University of Portsmouth website and earned £15 for participating. The only exclusion criteria was that people could not have taken part before in a similar deception study. Eventually 74 interviewees took part in the study, of whom 32.4 % were male. The average age of the interviewees was M = 29.9 years (SD = 12.5) and ranged from 18 to 61 years.
Procedure
The materials used were the transcribed interviews between the police-officers and interviewees. Before taking part in the study, both police-officers and interviewed participants signed an informed consent form. In this informed consent form the conditions of the study and the procedure were explained. For both interviewers and interviewees participation was confidential and their names were not used or connected to the results in any way. Besides 6
this, participants were informed that (I) they could withdraw at any time during the experiment, without any consequences, (II) that the data would be saved for a maximum of 5 years, (III) that it would be possible to obtain the results of the study and (IV) that they would receive a debriefing form. In the informed consent form the procedure and aim of the study were explained. Participants were given the opportunity to ask questions about the experiment. After signing the informed consent form, the interviewers were given a short introduction about the case. They were given no particular assignment of using a standard protocol, procedure or technique. They could ask the questions that they wanted to, and there was no limit on the number of questions they could ask. Their only assignment was to obtain as much information as possible that was relevant to the case. Some police officers received training between two interviews about questioning techniques and lie detection. The interviewers were told to terminate the interview after approximately 30 minutes.
For the interviewees there were four different conditions. They were put in one of the two scenarios. The first scenario was watching a short video about a secret meeting in which three people planned to put a spy device in a secret location. In the second scenario the interviewees were taken to a pub for fifteen minutes (truth condition) or only for one minute (lie condition) which was close to the research lab with one of the research assistants. In the pub scenario they were told after that a theft had taken place in the last fifteen minutes in the research lab, making them a suspect. After one of the two conditions they were told to lie or tell the truth to the officer when interviewed. Lying in the secret meeting scenario meant providing a mixture of truthful and deceptive information about the meeting. Lying in the pub scenario meant pretending to have stayed in the pub for fifteen minutes. The two scenarios are both relevant in real-life for investigative interviewing.
Coding of interview transcripts
(a)Type of questions
The types of questions asked by the interviewer were categorised according to Griffiths and Milne’s (2006) category system, as explained in the introduction. Broadly there are two categories: appropriate and inappropriate questions. These categories can be divided into more specific question types. The appropriate questions consist of open-ended questions, probing questions and encourager/acknowledgement questions. The inappropriate questions consist of closed questions, forced choice questions, leading questions, multiple questions at once, opinion/statement questions and echo questions. Table 1 provides a schematic overview
of these types of questions with examples for each type. Every question was allocated to one of these categories (no double coding took place) and for each interview the total number of every type of question was counted. This categorisation of question types has been chosen as it appeared frequently in the forensic investigative interviewing literature (Griffiths & Milne, 2006; Oxburgh et al., 2012; Snook & Keating, 2010; Snook et al., 2012). The interrater reliability for question coding was good for all the question types (.85 < ICC < .99).
(b) Amount of relevant detail obtained
The answers given by the interviewees were coded for the presence of detail. Total number of details were coded, based on visual details, auditory details, action details, spatial details, and temporal details. Details were only counted if they were relevant to the investigation. The pub scenario had four categories which were seen as relevant to the core event; the time before and after the pub visit, the time in the pub, detail about the research assistant and details about other individuals in the study. For the secret meeting scenario there were also categories which were coded as relevant; details about the accepted and rejected sites for planting the device, details about the people in the secret meeting, details about the place where the secret meeting had taken place and details about the device. Each detail has only been coded once. Eventually there will be a total number of details per interview. The interrater reliability for the detail coding was good (ICC = .95).
Data analyses
Before carrying out the data analyses needed to test the Hypotheses, the conditions which have to be met for these data analyses, had to be examined. Since the total number of questions asked varied substantially, the proportions of every question type has been
calculated (number of questions per type of question/total number of questions). For the first Hypothesis, the mean proportions of appropriate and inappropriate question type per
interview were calculated. A dependent t-test was carried out to compare these means proportions. A 0.05 significance level has been used.
Since not all the data needed for Hypotheses two and three were normally distributed, a non-parametric test was used. The Wilcoxon Signed-Rank Test has been used to compare the mean proportions of the different types of questions. For these Hypotheses a mean proportion for every type of question was calculated. For the second Hypotheses the mean of the proportion of open-ended questions was compared with all other means, using a 0.05
significance level for each test. For the third Hypotheses the mean of the proportion of closed questions was compared with the other means, also using a significance level of 0.05 for each comparison.
For Hypotheses four and five a multiple regression analyses with the obtained details as the dependent variable and the appropriate and inappropriate questions as the predictor variables was carried out. Every police officer had one total score for amount of detail and a score for the proportional amount of appropriate questions and inappropriate questions asked. This analysis revealed how much of the variance in the details is accounted for by the type of question. First of all, the model had to be significant (significant F-test with alpha level of 0.05) and the model had to be relevant (R square). After that, to support the hypotheses, the b-parameter of the appropriate questions had to be positive and significant in the analyses and the parameter of the inappropriate questions had to be negative and significant. With the b-parameters, an alpha level of 0.05 was used. Finally, the betas indicated the relative importance of the variables.
For Hypotheses 6 and 7 paired sample t-tests were used. For these Hypotheses only the police officers who also interviewed someone before they received the one-day training (N = 24) were used. The mean proportions of asking appropriate and inappropriate question types before and after the one-day training were calculated. For Hypothesis 6 the mean proportions of asking appropriate questions after and before the training were compared. For Hypothesis 7 the mean proportions of asking inappropriate questions after and before the training were compared.
Table 1
Examples of question types used by interviewing officers.
Group of questions Type of question Description Exemplar questions
appropriate questions open-ended questions These questions encourage interviewees ‘Tell me about what you saw?’ to provide answers from free recall
memory. They allow for a wide range of
‘Can you tell me in as much detail as you can, about the meeting?’
responses, and typically start with “tell”, ‘Can you describe the man?’
“explain,” or “describe.” ‘Are you able to explain what happened?’ probing questions Questions that begin with one of the five ‘What about the ladies?’
WH-words (who, why, what, when, ‘Who had the device?’ where) or begin with “how”. ‘Where was that?’ encouragements/ Questions or statements which ‘Ah right, thank you.’ acknowledgements encourage people to go on with talking ‘You are doing a good job.’
or thank them for the effort. ‘Yeah, carry on.’
inappropriate questions closed questions Questions that can be answered only by ‘Did you get the impression they knew each other?’ “yes” or “no” and ask about new ‘Were you carrying anything?’
information. ‘Ok. So you walk into the building together?’ forced choice questions Questions that force the interviewee to ‘And erm, digital or analogue?’
choose between two or more options ‘Did you see the man or just, just the location?’ given by the interviewer. ‘Henry or George?'
leading questions This type of question suggests an ‘He’s left his orange there, hasn’t he?’ answer to the interviewee. The desired
answer is already in the question.
multiple questions at once Two or more questions are asked at the ‘So there, there are three white buttons? On a, on a same time. circular.. Is, is it like a ball?’
‘What happened? What did you see on the pictures then? What pictures did you see?’
echo questions The information of the interviewee is Participant: ‘And he is not there.’ repeated in a question form by the Officer: ‘And he's not there?’
interviewer. Usually to confirm or to Participant: ‘No I didn’t really take notice of that.’ express disbelieve. Officer: ‘Didn’t take any notice of that?’
opinion/statement This type of question or statement gives ‘You would remember that, I think.’
The interviewee information or an ‘It troubles me so I am asking you again.’
opinion from the interviewer.
Results
General results
The average length of the interview was 28.44 minutes (SD = 7.95) and ranged from 12.07 minutes to 59.07 minutes. The average number of questions asked per interview was 130.16 (SD = 53.57) and ranged from 44 to 268.
The mean percentage of each question type is showed in Figure 1. On average, 8.7% of questions asked in an interview were open-ended questions (SD = 5.5), 22.4% of questions
were probing questions (SD = 8.5) and 8.2% of the questions were
encourager/acknowledgement questions. Consequently, on average total amount of appropriate questions asked by the police officers during the interviews was 39.3% (SD = 11.3) Closed questions accounted on average for 46.0% (SD = 10.1) of the questions asked during an interview. Forced choice questions composed on average 1.6 % (SD = 2.7) of questions asked during an interview. Leading questions accounted on average for 1.5% (SD = 1.5), multiple at once questions composed on average 1.9% (SD = 1.8), echo questions composed on average 5.0% (SD = 3.7) and opinion/statement questions composed on average 4.7% (SD = 3,9) of questions asked during the interviews, respectively. On average, total amount of inappropriate questions asked by the police officers during interviewing was 60.7%. Closed questions were asked five times as much as open-ended questions on average (proportion closed questions/proportion open-ended questions = 5.29).
Question types
As predicted, police officers used on average significantly a greater proportion of inappropriate questions (M = .607, SD = .113) than appropriate questions (M = .393, SD = .113) during the investigative interviews (t(73) = -8.137, p < 0.01, r = 0.69).
Comparing the mean proportion of the open-ended questions with all the other types of questions, open-ended questions are only used less than closed questions and probing questions (Table 3). Unexpectedly, open-ended questions were significantly more used than forced, leading, multiple at once, echo and opinion/statement questions (Table 3). There was no significant difference between the use of encourager/acknowledgement questions and open-ended questions (Table 3). All these significant differences had medium to large effect sizes (-.29 < r < -.61) (Cohen, 1988) (Table 3).
When comparing the mean proportion of the closed questions with all the other types of questions, the results show that closed questions are used significantly more than any other type of question during the investigative interviews (Table 4). All of the significant differences had large effect sizes (-.58 < r < -.61) (Cohen, 1988) (Table 4).
Figure 1. Mean percentage (%) of type of questions asked per interview.
Impact of question types on obtained details
Table 2 shows the results of the multiple regression analysis with obtained details as the dependent variable and inappropriate and appropriate questions as predictor variables. Although the model is significant (F = 3.539, p < 0.05), only 9,2 % of the variability in the outcome is accounted for by the predictors. As predicted, the appropriate questions had a positive relationship with the amount of detail obtained, while the inappropriate questions had a negative relationship with the amount of detail obtained (Table 2). Although the
0 5 10 15 20 25 30 35 40 45 50 O pe n-ende d P robi ng E nc our age r/ A cknow le dge m ent C lo se d F or ce d c hoi ce L ea di ng M ul ti pl e a t onc e E ch o O pi ni on/ S ta te m e nt
Appropriate questions Inappropriate questions
M ean p er cen tag e (% ) 13
relationships were as expected, solely the appropriate questions had a significant relationship with the obtained details and can be seen as a predictor for obtained detail (Table 2).
Table 2
Results of the multiple regression analyses with total detail obtained as dependent variable.
B SE B β Constant 110.52 28.16 Appropriate questions 1.58 0.60 .36* Inappropriate questions -0.39 0.32 -.17 * p < 0.05 Note: R2 = .092 Training effect
Following the predictions, police officers used a significant greater proportion of appropriate questions during the interviews after the training (M = .412, SD = .123) than before the training (M = .358, SD = .090, t(23) = 3.334, p < 0.01, r = 0.24). They further used a significant greater proportion of inappropriate questions before training (M = .642, SD = .090) than after the training (M = .588, SD = .123, t(23) = -.334, p < 0.01, r = 0.24). In Figure 2 the changes in proportions are made visible.
Figure 2. Mean percentage (%) of inappropriate and inappropriate questions asked before and
after the one-day training. 0 10 20 30 40 50 60 70 80 90 100
Before training After training
M ean p er cen tag e (% ) Appropriate questions Inappropriate questions 14
Table 3
Wilcoxon Signed-Ranks Test results, comparisons with open-ended questions
* Significant with p < 0.001
a. proportion probing questions < proportion open-ended questions b. proportion probing questions > proportion open-ended questions c. proportion probing questions = proportion open-ended questions d. total number of participants
Variables Signed ranks N Mean ranks Sum of ranks z p r
Proportion probing questions – proportion open-ended questions Positive 5a 11.40 57 -7.054 .000* -.58 Negative 67b 38.37 2571 Ties 2c Total 74d Proportion encourager/acknowledgement questions – proportion open-ended questions
Positive 40 36.16 1446.5 -.965 .334 -
Negative 31 35.79 1109.5
Ties 3
Total 74
Proportion closed questions – proportion open-ended questions
Positive 1 2.00 2 -7.464 .000* -.61
Negative 73 37.99 2773
Ties 0
Total 74
Proportion forced choice questions – proportion open-ended questions
Positive 70 37.37 2616 -7.306 .000* -.60
Negative 2 6.00 12
Ties 2
Total 74
Proportion leading questions – proportion open-ended questions
Positive 68 39.88 2712 -7.135 .000* -.59
Negative 6 10.50 63
Ties 0
Total 74
Proportion multiple at once questions – proportion open-ended questions
Positive 68 39.00 2652 -7.155 .000* -.59
Negative 5 9.80 49
Ties 1
Total 74
Proportion echo questions – proportion open-ended questions
Positive 48 41.63 1998 -3.560 .000* -.29
Negative 25 28.12 703
Ties 1
Total 74
Proportion opinion/Statement questions – proportion open-ended questions
Positive 52 38.53 2003.5 -5.075 .000* -.42
Negative 16 21.41 342.5
Ties 6
Total 74
Table 4
Wilcoxon Signed-Ranks Test results, comparisons with closed questions
* Significant with p < 0.001
a. proportion probing questions < proportion closed questions b. proportion probing questions > proportion closed questions c. proportion probing questions = proportion closed questions d. total number of participants
Variables Signed ranks N Mean ranks Sum of ranks z p r
Proportion probing questions – proportion closed questions
Positive 66a 38.11 2515 -7.088 .000* -.58 Negative 5b 8.20 41 Ties 3c Total 74d Proportion encourager/acknowledgement questions – proportion closed questions
Positive 74 37.50 2775 -7.475 .000* -.61
Negative 0 0.00 0
Ties 0
Total 74
Proportion forced choice questions – proportion closed questions
Positive 73 37.00 2701 -7.425 .000* -.61
Negative 0 0.00 0
Ties 1
Total 74
Proportion leading questions – proportion closed questions
Positive 74 37.50 2775 -7.475 .000* -.61
Negative 0 0.00 0
Ties 0
Total 74
Proportion multiple at once questions – proportion closed questions
Positive 74 37.50 2775 -7.475 .000* -.61
Negative 0 0.00 0
Ties 0
Total 74
Proportion echo questions – proportion closed questions
Positive 74 37.50 2775 -7.475 .000* -.61
Negative 0 0.00 0
Ties 0
Total 74
Proportion opinion/statement questions – proportion closed questions
Positive 74 37.50 2775 -7.475 .000* -.61
Negative 0 0.00 0
Ties 0
Total 74
Discussion
The aim of this research study was to examine what type of questions are asked by police officers during investigative interviewing and how these questions affect the amount of detail that is obtained. In addition, this study examined the effect of training on the questions asked during interviewing.
Do police officers ask more inappropriate than appropriate questions?
Following the expectations, police officers asked more inappropriate questions than appropriate questions during the interviews. These results were in line with previous research (Oxburgh et al., 2012; Snook et al., 2012). Although we expected, based on previous research (Baldwin, 1993; Clarke & Milne, 2001; Davies et al., 2000; Myklebust & Bjorklund, 2006; Smith et al., 2009; Snook et al., 2012; Snook & Keating, 2011), that open-ended questions would have been asked less than every other type of question, this was not the case. Open-ended questions were only used less than closed questions and probing questions. For the other types of questions open-ended questions were used more or there was no difference (encourager/acknowledgement questions). Closed questions were asked most frequently compared to all the other type of questions, which is in line with previous research (Baldwin, 1993; Clarke & Milne, 2001; Davies et al., 2000; Myklebust & Bjorklund, 2006; Smith et al., 2009; Snook et al., 2012; Snook & Keating, 2011).
The fact that open-ended questions were asked more than most type of questions is not in line with the research in this area (Baldwin, 1993; Clarke & Milne, 2001; Davies, Westcott & Horan, 2000; Myklebust & Bjorklund, 2006; Smith, Powell & Lum, 2009; Snook et al., 2012; Snook & Keating, 2011), which may be explained by the fact that we used different inclusion criteria for open-ended questions (Oxburgh et al., 2012; Myklebust & Bjorklund, 2006; Snook et al., 2012). Questions as ‘Can you describe the man?’ and ‘Are you able to explain what happened?’ are strictly closed questions when we look at the format, but interviewees will never answer this question with “no” or “yes”. In fact, interviewees answer these questions as if they are open-ended questions and start to describe or explain what is asked, which is the reason why these questions have been treated as open-ended questions in this research. Another explanation for the fact that open-ended questions were used more than expected from the existing research, could be that all the data were used for this hypothesis. 50 out of 74 interviews were conducted after the interviewers had training and only 24 were conducted before. Most research to date has been with real-life investigative interviews
without training beforehand. As the training had an effect on asking appropriate and inappropriate questions, it would be reasonable that the open-ended questions were also affected as a group on their own by the training.
Do appropriate questions lead to more detail?
The analyses about the relationship between the obtained details and the questions asked show that appropriate questions are positively related to the amount of detail that will be obtained. Inappropriate questions are negatively related to the details obtained but not in a significant way. This last finding is not in line with the existing research (Oxburgh et al., 2012). As a whole, the multiple regression model was significant. If the model also will be relevant for practical use, is questionable though.
The fact that we did find a significant predictive model but not as relevant as hoped for the obtained details could be explained by the fact that only the total amount of detail and total questions asked were used during this research. In this way it was not possible to look at the effect of each question individually. Other research which found significant differences in detail obtained did look at each question individually (Oxburgh et al., 2012). For this reason it is difficult to make firm statements about the exact relationship between the obtained details and the questions asked in this research.
Does training help police officers to ask more appropriate questions and less inappropriate questions?
Finally, the one-day training covering lie detection, best practices in questioning techniques and underlying research had a positive effect on the questions asked by the police officers, as they asked more appropriate questions and less inappropriate questions after the training. This latter finding suggests that a more extensive collaboration between researchers and police trainers would be beneficial. Additional trainings for police officers could be helpful to improve their competence of investigative interviewing, by refreshing or extending their knowledge.
It should be noted however that there was no control group for hypothesis 6 and 7. For this reason, we cannot make the assumption that the change in questions asked after the one-day training is necessarily the consequence of this one-day training. It may be that the differences are due to other reasons such as coincidence or repeated measurements. Further research should rule out this possibility. Furthermore should follow-up research look at the 18
effect after some time. It is valuable to know if the training effect is maintained or rules out after a short period of time.
Limitations
The present study had some limitations. First of all, the interviewees used in this study were all students or staff members of the University of Portsmouth. This means there were no real offenders involved during this research study, which makes it a less realistic scenario. In addition, the fact that both interviewers and interviewees knew they were participating in a study instead of a real investigative interview has probably changed their behaviours in certain ways. The stakes were most likely less in this experiment than in real-life scenarios and the police officers may have been acting milder because they knew the interviewees were not real offenders but students or members of staff. Finally, a major limitation for this research is the fact that people were asked about a movie or scene they had seen or experienced a couple of minutes before. Normally people will see an event or crime in real-life and it will take at least a couple of hours before they are interrogated by the police at the police office, which makes their memory less accurate (Dunning & Stern, 1992). All these factors make this scenario artificial and less applicable to real-life scenarios.
Future research
Future research could focus on the effect of training with a larger sample size, to see if this finding will be replicated. In addition, it would be interesting to look at the effect of more extensive training than the one-day training used in this study. Since only one day of training already can make a difference, it could be possible that a more extensive and longer training would have even greater impact on the investigative interviewing of the police officers. In the best case, these data would be conducted with real investigative interviews to keep the research as realistic as possible. It would also be valuable to take a close look at how long the training effect stays intact before it expires. It may be necessary to repeat training sessions and feedback throughout the careers of police officers to keep their knowledge and practices in the best condition (Lamb, Sternberg, Orbach, Esplin & Mitchel, 2002; Lamb, Sternberg, Orbach, Hershkowitz, Horowitz & Esplin, 2002). Research also should have a close look at the details obtained per type of question individually, so that there will be more clarity about the effect of questions separately in terms of obtaining detail. Finally, research should focus on the question why police officers tend to use so many closed questions, even though they know it is better to ask open-ended questions. Research repeatedly found that police officers 19
use far more closed questions than any other type of question, but there is no research to our knowledge about the reasons behind this (Baldwin, 1993; Clarke & Milne, 2001; Davies et al., 2000; Myklebust & Bjorklund, 2006; Smith et al., 2009; Snook et al., 2012; Snook & Keating, 2011). It could be possible that asking closed questions is mentally less difficult, because it is easier to form a closed question or because you get a less extensive answer which gives police officers the chance to breath for a second and not pay as much attention needed when given an extensive answer. Knowing the cause, could be the beginning of finding a solution.
Conclusions
In conclusion, this research study replicated the finding that appropriate questions are asked less frequently during interviewing than inappropriate questions. Closed questions are by far the most used type of question during investigative interviewing. Asking appropriate questions has a positive relationship with obtaining detail, while inappropriate questions do not have a significant relationship with obtained detail. Despite this, the predicting value of the whole model is significant. Finally we found that refreshing the memory of police officers by a one-day training has an impact on the questions they ask in a positive way. They asked more appropriate and less inappropriate questions after than before the training, which offers a hopeful prospect for the future.
References
Baldwin, J. (1993). Police interviewing techniques: establishing truth or proof? The British
Journal of Criminology, 33(3), 325-352.
Baxter, J., Charles, K., Martin, M., & McGroarty A. (2013). The relative influence of leading questions and negative feedback on response change on the Gudjonsson Suggestibility Scale (2): implications for forensic interviewing. Psychology, Crime & Law, 19(3), 277-285.
Brewer, N., & Williams, K.D. (2005). Psychology and law: an empirical perspective. New York: The Guilford Press.
Bull, R., & Milne, B. (2004). Attempts to improve the police interviewing of suspects. United States: Springer.
Clarke, C., & Milne, R. (2001). A national evaluation of the PEACE investigative
interviewing course. Police Research Award Scheme Report no: PRAS/149.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences. New York: Academic Press.
Davies, G.M., Westcott, H.L., & Horan, N. (2000). The impact of questioning style on
the content of investigative interviews with suspected child sexual abuse victims.
Psychology, Crime & Law, 6(2), 81-97.
Dunning, D., & Stern, L.B. (1992). Examining the generality of eyewitness hypermnesia: a close look at time delay and question asked. Applied Cognitive Psychology, 6(7), 643 657.
Fisher, R.P., Falkner, K.L., Trevisan, M., & McCauley, M.R. (2000). Adapting the cognitive interview to enhance long-term (35 years) recall of physical activities. Journal of
Applied Psychology, 85(2), 180-189.
Griffiths, A., & Milne, R. (2006). Will it al end in tiers? Police interviews with suspects in Britain. In Williamson, T. (2006). Investigative Interviewing: Rights, research,
regulation. Cullompton: Willan Pubslishing.
Gudjonsson, G.H., & Clark, N. K. (1986). Suggestibility in police interrogation: a social psychological model. Social Behaviour, 1(2), 83-104.
Lamb, E., Sternberg, K.J., Orbach, Y., Esplin, P.W., & Mitchell, S. (2002). Is ongoing
feedback necessary to maintain the quality of investigative interviews with allegedly abused children? Applied Developmental Science, 6(1), 35-41.
Lamb, E., Sternberg, K.J., Orbach, Y., Hershkowitz, I., Horowitz, D., & Esplin, P.W. (2002). The effects of intensive training and ongoing supervision on the quality of investigative interviews with alleged sex abuse victims. Applied Developmental
Science, 6(3), 114-125.
Loftus, E.F. (1979). Reactions to blalantly contradictory information. Memory & Cognition,
7(5), 368-374.
McGurk, B.J., Carr, M.J., & McGurk, D. (1993). Investigative interviewing courses for police
officers: an evaluation. London: Home Office Police Research Group.
Milne, R., & Bull, R. (1999). Investigative Interviewing: Psychology and practice. Chichester: John Wiley & Sons.
Milne, R., & Bull, R. (2006). Interviewing victims of crime, including children and people with intellectual difficulties. In Kebbell, M.R., & Davies, G.M. (2006). Practical
psychology for forensic investigations. Chichester: John Wiley & Sons.
Myklebust, T., & Bjorklund, R.A. (2006). The effect of long-term training on police officers’ use of open and closed questions in field investigative interviews of children (FIIC).
International Journal of Investigative Psychology and Offender Profiling, 3(3),
165-181.
Oxburgh, G., Ost. J., & Cherryman, J. (2012). Police interviews with suspected child sex offenders: does use of empathy and question type influence the amount of
investigation relevant information obtained? Psychology, Crime & Law, 18(3), 259- 273.
Poole, D.A., & Lamb, M.E. (1998). Investigative interviews of children. Washington DC: American Psychological Association.
Read, J.M., Powell, M.B., Kebbell, M.R., & Milne, R. (2009). Investigative interviewing of suspected child offenders: a review of what constitutes best practice. International
Journal of Police Science & Management, 11(4), 442-459.
Shawyer, A., Milne, B., & Bull, R. (2009). Investigative interviewing in the UK. In Williamson, T., Milne, B., & Savage, S.P. (2013). International developments in
investigative interviewing. Cullompton: Willan Publishing.
Shepherd, E. (2007). Investigative interviewing: The conversation management approach. Oxford: Oxford University Press.
Smith, R.M., Powell, R.M., & Lum, J. (2009). The relationship between job status,
interviewing experience, gender and police officers’ adherence to open-ended questions. Legal and Criminological Psychology, 14(1), 51-63.
Snook, B., & Keating, K. (2011). A field study of adult witness interviewing practices in a Canadian police organization. Legal and Criminological Psychology, 16(1), 160- 172.
Snook, B., Luther, K., Quinlan, H., & Milne, R. (2012). LET ‘EM TALK! A field study of police questioning practices of suspects and accused persons. Criminal Justice and
Behavior, 39(10), 1328-1339.
Soukara, S., Bull, R., Vrij, A., Turner, M., & Cherryman, J. (2009). What really happens in police interviews of suspects? Tactics and confessions. Psychology, Crime & Law,
15(6), 493-506.
Sternberg, K.J., Lamb, M.E., Orbach, Y., Esplin, P.W., & Mitchell, S. (2001). Use of a structured investigative protocol enhances young children’s responses to free-recall prompts in the course of forensic interviews. Journal of Applied Psychology, 86(5), 997-1005.
Vrij, A. (2008). Detecting lies and deceit: pitfalls and opportunities. Chichester: John Wiley & Sons.
Vrij, A., Granhag, P.A., & Porter, S. (2010). Pitfalls and opportunities in nonverbal and verbal lie detection. Psychological science in the public interest, 11(3), 89-121. Walsh, D.W., & Milne, R. (2008). Keeping the PEACE? A study of investigative
interviewing practices in the public sector. Legal and Criminological Psychology,
13(1), 39-57.
Waterman, A.H., Blades, M., & Spencer, C. (2001). Interviewing children and adults: The effect of question format on the tendency to speculate. Applied Cognitive Psychology,
15(5), 521-531.
