You ask and do not receive,because you ask wrongly

Abstract

Investigative interviewing can be thought of as the interaction between two clusters of factors: questions and techniques. Analyses of question types and interview techniques are often treated discretely, where one is the focal point at the expense of the other, or they are conceptually indistinguishable. To explore these relationships further, the current study examined a sample of interviews using the Griffiths Question Map and a taxonomy of interview methods. In addition to presenting maps as example interviews, the article explores the associations between interview clusters and suspect cooperation. Results indicate that greater use of appropriate questions was associated with greater use of rapport-building techniques and greater suspect cooperation, and a reduction in the use of accusatorial interrogation tactics, which themselves were associated with suspect resistance. Implications for research and practice include understanding the interaction between these—and potentially other—clusters.

Keywords

Investigative interviewing question types interrogation techniques;suspect cooperation

You ask and do not receive, because you ask wrongly

Imagine this scenario¹: Three heavily armed men enter a bank in downtown Los Angeles with the intention of taking an unusually large amount of money that was to be on location that day. As planned, they escape the bank with US $12 million, only to be met by a surveilling police force waiting for them outside. A massive shootout incurs on the streets in broad daylight, with dozens of casualties including one of the robbers and the getaway driver who were killed by police. The rampage continues for a short time until the other two robbers are apprehended and brought into custody. Interviews of the suspects commenced at the same time, as investigators sought to elicit admissions to the crime and its planning, determine if they received logistical support from other individuals, and gain intelligence on possible future scores. Each suspect was informed of his rights, and although neither invoked those rights, the suspects were uncooperative with detectives.

The first interviewer, Lt Hanna, says to the first suspect, Neil, “Did you intend to kill several police officers when you exited the bank?”. Meanwhile, in the next interviewing room, Sgt Drucker framed his question to Chris as, “Please tell me your reasons for being at the bank this afternoon”. Alternatively, Lt Hanna could have emphasized the zero-sum relationship between himself and Neil in the context of the interview and implied that whoever talks first will be treated more fairly by the police and prosecution. In the other interview, Sgt Drucker could have instead expressed to Chris that he has a job to do and holds no person animus against him while offering him a cigarette. With these possible approaches in mind, two fundamental questions arise for the research on and practice of investigative interviewing: which of these methods is more or less likely to turn an uncooperative suspect into a cooperative one? And which methods are best suited to produce admissions and information?

For the purposes of this inquiry, the term “methods” is used here to refer to the vast number and type of approaches interviewers may take in order to secure cooperation or information. Of interest are those methods that the interviewer herself can manipulate and employ, as opposed to other elements of the interview that are beyond her immediate control such as offender and case characteristics (Beauregard et al., 2010; Moston et al., 1992) and the physical context in which the interview takes place (Kelly et al., 2019). Thus, we focus on the words and actions of the interviewer, specifically the two approaches, or clusters, described above—the type of question asked (e.g., “What were you doing at that location?”), and the technique employed (e.g., depersonalizing the situation, prisoner’s dilemma or telling one suspect that the other is already cooperating). Arguably, these two clusters of methods are among the most researched, and in some studies one is quite clearly the focus at the expense of the other (Kelly et al., 2016); elsewhere, interviewer verbiage and actions have been treated equally, presumably without consideration of the conceptual distinctions (Soukara et al., 2009) or an explicit differentiation between them (Pearse and Gudjonsson, 1999). Moreover, scholarly focus on accusatorial interrogation such as the Reid Technique (Inbau et al., 2013) tends to focus on the techniques employed rather than the type of question asked by investigators.

To advance our understanding of what is effective in achieving an interviewer’s goals, the position of this article and its unique contribution to the literature is that we must understand that both questions are asked and techniques employed during every interview, and that these seemingly discrete elements ought to be considered in tandem. Although the conclusion of this piece is that successful interviews—and the study of what makes them successful—should consider the complex interplay between questions and techniques, we purposefully begin with the matter of investigative questioning. After all, the biblical quote that precedes the title passage of this paper is, “You do not have if you do not ask” (James 4: 2–3, Revised Standard Version).

Question types

In the broadest sense possible, research on question types has examined open- versus closed-ended questions. Although universally accepted definitions of open and closed questions are difficult to find (see Oxburgh et al., 2010, for an excellent discussion of the various typologies of questions), open questions are generally considered those in which a suspect may give a free-narrative account of the events under investigation and closed questions considerably constrain the interviewee’s ability to share information as they elicit brief and undetailed responses. Examples of open questions include free-narrative style imperatives (“tell” and “explain”) and sometimes the “5WH” (who, what, when, where, why, and how) interrogatives that are classified as open as well. Closed questions, by contrast, include some versions of yes/no, leading, forced choice, or multiple questions.

Perhaps a better question type dichotomy would be appropriate versus inappropriate (Milne and Bull, 1999) or productive versus unproductive (Griffiths and Milne, 2006; see also expected versus unexpected questions in the deception detection literature, Vrij et al., 2009). For instance, Milne and Bull argued that the 5WH questions should be considered closed because they elicit a limited response from the interviewee as opposed to the free reign offered by open-ended questions. They nevertheless considered such questions as appropriate and worthy of employing. Likewise, Griffiths and Milne differentiated the 5WH questions from the open category, classifying them as probing instead and treating them as productive.

Despite an apparent lack of precision in their definitions, these and other researchers have overwhelmingly reported that skilled interviewers should ask more open, appropriate, or productive questions (see also: Clarke and Milne, 2001; Griffiths et al., 2011; O’Mahoney et al., 2012; Read et al., 2009) which are more effective on a range of outcomes than closed, inappropriate, or unproductive ones (Oxburgh et al., 2012; Phillips et al., 2012; Walsh and Milne, 2008). These associations notwithstanding, research has demonstrated that interviewers in general still tend to ask closed questions at a higher rate than open ones (Baldwin, 1993; Oxburgh et al., 2012; Snook et al., 2012).

In an attempt to advance the utility of question type analyses, Griffiths (2008) developed the Griffiths Question Map (GQM) as a visual demonstration of how productive an interview was. Categorizing question types as productive or unproductive, the three productive questions are open (“Tell me about…”), probing (the who, what, when, where, why, and how questions), and appropriate closed (those used to clarify and confirm information). The unproductive questions are inappropriate closed (yes/no questions that do not refer to information already disclosed by the suspect), leading (the desired answer is presumed in the question itself), multiple (asking about several topics in a single question), forced choice (not allowing for an alternative response), and finally, questions that are actually statements of the investigator’s opinion. For any given interview, each question is sequentially plotted on the x-axis of a graph and the question’s relative productiveness located on the y-axis (the productive questions at the top). In better interviews, the GQM would have concentrations of questions plotted in the top third of the graph, with relatively fewer unproductive questions in the lower portion of the graph, and vice versa in less successful ones.

Despite its obvious utility to interview training (Griffiths, 2008) and assessment (Dodier and Denault, 2018; Kopše, 2017), research on the GQM in the published literature has been somewhat limited. Walsh and Bull (2015), for instance, used the question type categories put forth by Griffiths and Milne (2006) and presented example maps, and they found that skilled interviews were associated with more productive questioning that produced more complete accounts from suspects. More recent research has sought to build upon the novel idea to graphically show the progression of questions by also mapping interviewee responses (Farrugia et al., 2019; Waterhouse et al., 2019) and other interviewer behaviors (e.g., interruptions, empathy; Farrugia et al., 2019).

Techniques

Whereas the majority of research on question types emanates from Europe and the United Kingdom (UK) specifically, Snook and colleagues (2012) observed that “[s]tudies of suspect interviewing practices in North America…have focused largely on the types of interrogation tactics and techniques that are used to elicit confessions from offenders…” (p. 1330; emphasis in original). Like the type of question posed, a technique is seemingly difficult to define despite being the object of much empirical review. In some instances, techniques have been used as the umbrella term for all actions taken by the interviewer, including questions asked, although in other studies the term “tactics” has been used in this manner (Bull and Soukara, 2010). It is unclear whether the use of the word “tactic” was purposeful or not, but in no study has there been a discussion regarding the distinctions between these descriptors intended or assumed to be describing what one individual does in order to obtain information or admissions or cooperation from another individual.

One possible definition of a technique is any action on the part of the interviewer that is intended to create conditions favorable for the elicitation of information and/or admissions, including turning an uncooperative interviewee into a cooperative one. The problem with this definition, however, is that it could reasonably include question types. That skilled interviewers will use open-ended and productive questions is also a means of creating favorable conditions in the sense that it is empirically demonstrated that these methods will benefit the interviewer in the end.

What sets techniques apart, then, and justifies their discussion as a distinct cluster of methods is that we know interviewers and interrogators, especially in the USA, have more tools from which to choose whilst attempting to elicit information from uncooperative interviewees that are conceptually distinct from question types (Miller et al., 2018). As such, we can identify techniques as being verbal if not exclusively lingual; they can be the physical actions or simply the presence of more than just an interviewer and suspect; they can be embodied in inanimate objects or the space in which the interview is conducted (Dawson et al., 2015); and techniques are often—by design—undetectable by the suspect or an untrained observer. This is not to say that all techniques are coercive (or abusive in the physical sense), but they all, in their own unique way, attempt to alter the conditions of the interview in such a manner that benefits the interviewer and have nothing to do with the grammatical structure of an utterance. A brief review of key studies reveals the importance of techniques in the present schema with question type.

In the seminal study on American police interrogation, Leo (1996) found that among the 25 techniques observed, the minimization tactics of appeals to the suspect’s self-interest and conscience, and the use of moral justifications were most commonly used and associated with providing incriminating evidence. Recent research using interrogation recordings partially confirmed the high rate of minimization use among American investigators (Kelly et al., 2019). Feld (2006, 2013), on the other hand, found minimization techniques were less frequently applied to juvenile suspects; instead, Feld reported techniques related to maximization were more likely found in his sample. Finally, in the third major study to examine accusatorial interrogation, King and Snook (2009) reported high rates of techniques exemplified by the Reid method (Inbau et al., 2013), such as the use of theming and the alternative question, and direct accusations of guilt, were employed in their sample of Canadian interrogation recordings.

In a thorough examination of known techniques, Kelly et al. (2013) developed a taxonomy of interrogation methods by identifying nearly 70 unique techniques that they sorted into one of six conceptual domains, arguing that these six constructs are more useful than analyzing scores of specific techniques. First, the “rapport and relationship-building” domain seeks to establish a working relationship between investigator and subject, with example techniques in this domain including meeting the subject’s basic needs (e.g., offers of food, water), finding common ground, and expressing empathy. Next, “emotion provocation” is akin to theming (Senese, 2008) or minimization (Kelly et al., 2019), and included techniques such as appealing to the subject’s self-interest, appealing to his/her conscience, offering rationalizations, and flattery. “Context manipulation” is altering the physical or temporal space in which the interview occurs, and the “confrontation/competition” domain includes accusatorial techniques such as expressing frustration, and explicitly or implicitly making threats of punishment for non-cooperation. Next, appealing to the subject’s sense of cooperation and exchanges between interviewer and interviewee are included in the “collaboration” domain. Finally, showing or alluding to information from sources such as criminal records, witness or victim statements, surveillance footage, or forensic tests, represent the “presentation of evidence” domain.

Research into the domain structure supports its utility. Studies have used them to describe the relative rate at which interviewers use the domains (Redlich et al., 2014; Russano et al., 2014) and in a multinational survey of interviewing practices (Miller et al., 2018). Uniformly, these studies’ participants have endorsed the “rapport and relationship-building” domain and eschewed the accusatorial methods contained across the “confrontation/competition” and “emotion provocation” domains. Observational research largely supports these self-reported studies that rapport tactics are emphasized more than the accusatorial ones (Kelly et al., 2015). Furthermore, “rapport and relationship-building” led to increases in suspect cooperation, whereas “confrontation/competition” and “presentation of evidence” reduced cooperation (Kelly et al., 2016).

The importance of rapport in investigative interviews is underscored by the Observing Rapport-Based Interpersonal Techniques (ORBIT) framework developed by Laurence Alison and colleagues (2013). Across a large (and growing) sample of law enforcement interviews with convicted terrorists, the research has consistently demonstrated that rapport-based interviewing techniques increase interviewee engagement and yield (Alison et al., 2013) and decrease counterinterrogation measures (Alison et al., 2014). Additional recent analyses suggest that rapport-based interviews reduce reactance, or a type of resistance produced by the pressure put forth by the interviewer to talk, cooperate, etc. (Surmon-Böhr et al., 2020). As such, it is critical to examine non-confrontational, non-judgmental interviewing styles for their influence on interviewee behavior.

Research questions and hypotheses

With few exceptions, the research on interviewing methods has examined the question type and technique clusters independent of one another. Where a single study has data on each, the analyses have not considered how questions relate to techniques, instead presenting them as equal things interviewers do (Soukara et al., 2009; Walsh and Bull, 2012). In this study, we sought to clarify the two clusters of investigative interviewing research by explicitly distinguishing between question types and interview techniques and to explore the associations between them. Using a sample of recorded suspect interviews from a large American police department, we sought to answer the following research questions:

What is the relationship between question type and interview technique?

What is the relationship between question type and suspect cooperation, controlling for interview techniques?

The first is a novel area of exploration in investigative interviewing research, as no previous study, to our knowledge, has purposefully analyzed the associations between questions asked and techniques employed with suspects. As reviewed above, the research is quite conclusive on two counts—first, appropriate or productive questions lead to better outcomes than inappropriate or unproductive questions; second, successful interviews are those that focus more on rapport-building and less on accusatorial interrogation techniques. It stands to reason, then, that there may be associations between questions and techniques, and two hypotheses for the current study come from this logical deduction:

H_1a: The proportion of appropriate-to-inappropriate questions will be positively related to the use of rapport-building techniques.

H_1b: Conversely, the proportion of appropriate-to-inappropriate questions will be negatively related to the use of accusatorial interrogation techniques.

Put differently, when investigators ask more appropriate questions than inappropriate ones, they will tend to emphasize rapport-building and eschew confrontational or psychologically manipulative methods. To examine this, we computed a single variable that measured the relative use of appropriate and inappropriate questions within a 5-minute interval of time (described in detail in the next section).

Next, the study builds upon a previous one reviewed above (Kelly et al., 2016) that examined the influence four interrogation domains (Kelly et al., 2013) had on suspect cooperation. The second research question, then, explicitly addresses the association between the type of question asked and how the suspect responds during an interview, independent of the domain influences on cooperation. We will employ both correlational and regression analyses to test the following hypothesis:

H₂: The proportion of appropriate-to-inappropriate questions will be positively related to suspect cooperation.

In addition to addressing these research questions and hypotheses, we also present two example GQMs from the current sample—one in which the suspect made self-incriminating admissions, and one in which the suspect denied involvement throughout. To our knowledge, no such maps have been published using American interviews, and using the appropriate-to-inappropriate measure, we selected these maps to illustrate the utility of that variable in the context of contributing to the research on the GQM.

Method

Sample

The data for the current study of question types and techniques come from 15 interviews provided by the Robbery-Homicide Division of the Los Angeles Police Department (RHD-LAPD), eight of which were video-recorded; the remaining seven were audio only. Thirteen of the suspects were men, nine were Latino/Latina, three African American, two were white and for one suspect race or ethnicity was unknown, 12 were murder investigations, and the average suspect age was 32.93 years (SD = 15.08). All interviews had two investigators present, and the mean interview length was 57 minutes and 16 seconds (SD = 0:28:52; range 0:14:15 to 1:37:35). We note here that the 15 interviews come from a larger sample of audio and video recordings (see Kelly et al., 2015) that were subsequently transcribed and used for measuring question type use, the procedures of which are described in the next section.

Procedure and measures

Whereas earlier research (Feld, 2006; Leo, 1996; Soukara et al., 2009) measured variables as being absent or present at any point during an interrogation, to more precisely capture the dynamics between investigator and suspect, we measured question types, domains, and suspect responses in 5-minute intervals (see Bull and Soukara, 2010; Pearse and Gudjonsson, 1999). Independent coders watched or listened to the recordings, pausing after each 5-minute interval to measure the domains and responses; a separate coder responsible for measuring question types used the transcribed version of the interrogations that were marked according to the interval system. As such, the 5-minute intervals represent the primary unit of analysis in this study (N=168).

Following the question type categories of Griffiths and Milne (2006), we coded for three appropriate questions—open, probing, and appropriate closed—and five inappropriate questions—inappropriate closed, forced choice, leading, multiple questions, and opinions or statements made by investigators. Each question type was counted in the 5-minute interval, and we calculated appropriate and inappropriate categories by summing the individual question types. These question categories are instructive insofar as they are descriptive, but in order to analyze the influence of questions and their relationships to both interrogation domains and suspect responses, we created a single variable. Using the equation

(Appropriate Questions – Inappropriate Questions)/(Appropriate Questions + Inappropriate Questions)

resulted in our primary question type variable, the Appropriate Question Differential (AQD). The AQD represents the relative difference between appropriate and inappropriate questions asked during the 5-minute interval, and it ranges from −1 to +1. A score of +1 is interpreted as all questions asked in the interval were appropriate; 0 is either an equal number of appropriate and inappropriate questions were asked by investigators, or no questions were posed at all; finally, −1 represents the instance when all questions were inappropriate.

To measure the methods used in the recordings, we operationalized four of the six domains of the interrogation taxonomy described above (Kelly et al., 2013)—rapport and relationship-building, emotion provocation, confrontation/competition, and presentation of evidence.² Each was measured on a three-point emphasis scale (0 = none, 1 = moderate, 2 = major-exclusive), and in any 5-minute interval, multiple domains could be observed.

Our primary outcome measure was a composite variable of two separate measures of information the suspect offered that was coded as cooperative and resistant. Cooperative information was operationalized as offering non-incriminating information that may or may not have been related to the crime, self-incriminating information, including admissions, and alibis. Resistant information, on the other hand, was coded when the suspects offered denials, claims of poor memory, silence, or invocation of their right to silence. As with the domain emphases, these variables were coded on a three-point scale (0 = not present, 1 = somewhat present, 2 = strongly present). Despite the 5-minute interval coding procedure allowing for suspects to demonstrate both cooperative and resistant information within the same interval, the two variables were strongly correlated with one another (r = −.91, p < .001) indicating that the suspect was either cooperative or resistant during the interval. We created a single five-point cooperation scale that served as the dependent variable in this study: 1 = strong resistance; 2 = weak resistance; 3 = neutral; 4 = weak cooperation; 5 = strong cooperation. In this sample, the mean suspect cooperation score was 3.43 (SD = 1.16) indicating that, on average, suspects were generally neutral in response to investigators. For more information on the domain emphasis measures and suspect cooperation, see Kelly et al. (2016).

Results

The descriptive statistics for the specific question types, appropriate and inappropriate question categories, the computed AQD variable, and interrogation domain variables are presented in Table 1. The three most commonly asked questions per 5-minute interval were appropriate closed, inappropriate closed, and probing. Summing the variables in the appropriate and inappropriate categories, we found that, on average, investigators asked more than eight appropriate questions (8.39) and fewer than six inappropriate questions (5.66) in each interval, yielding an overall AQD score of 0.24. The pattern of results for the domains in this sample was the same as reported by Kelly et al. (2015), with rapport and relationship-building emphasized the most, followed by presentation of evidence and emotion provocation, with confrontation/competition being the least emphasized.

Table 1.

Descriptive statistics of question types, categories and AQD, and domains.

Question type	Mean (SD)	Min	Max
Appropriate	8.39 (7.15)	0	40
Open	0.17 (0.66)	0	5
Probing	4.01 (3.69)	0	20
Appropriate closed	4.21 (4.55)	0	22
Inappropriate	5.66 (4.79)	0	20
Inappropriate closed	4.21 (4.12)	0	16
Forced choice	0.14 (0.47)	0	3
Leading	0.63 (0.96)	0	5
Multiple	0.46 (1.01)	0	6
Statements	0.22 (0.82)	0	8
Appropriate Question Differential	0.24 (0.51)	−1	1
Domains
Rapport and relationship-building	0.86 (0.77)	0	2
Emotion provocation	0.59 (0.73)	0	2
Confrontation/competition	0.40 (0.58)	0	2
Presentation of evidence	0.81 (0.75)	0	2

The relationships between question categories, the AQD, and the four domains in Table 2 support Hypotheses 1a and 1b that appropriate questioning strategies are associated with rapport-building tactics and less so with accusatorial-style interrogation. Specifically, where the proportion of questions was more appropriate, investigators were significantly more likely to emphasize the rapport and relationship-building domain, and the AQD was negatively related to the emotion provocation, confrontation/competition, and presentation of evidence domains. Finally, although we discuss this in greater detail in the next section, we note here the positive and moderately strong correlation between appropriate and inappropriate categories.

We found support for Hypothesis 2, that appropriate questions would be associated with suspect cooperation, in bivariate correlations that both appropriate questions and higher scores on the AQD used were significantly associated with higher levels of cooperation. Also, of the four interrogation domains, only rapport and relationship-building was significantly associated with higher suspect cooperation, whereas the other three domains were significantly related to lower cooperation.

Table 2.

Pearson correlations and significance levels between question categories, domain emphases, and suspect cooperation.

	Inappropriate	AQD	RRB	EP	CC	PE	Cooperation
Appropriate	0.43 (<.001)	0.22 (.01)	0.29 (<.001)	−0.19 (.02)	−0.26 (.002)	−0.34 (<.001)	0.28 (.001)
Inappropriate	1	−0.55 (<.001)	0.05 (.54)	−0.05 (.53)	0.01 (.94)	−0.01 (.94)	−0.02 (.80)
AQD		1	0.18 (.03)	−0.16 (.05)	−0.17 (.04)	−0.22 (.01)	0.24 (.004)
RRB			1	−0.17 (.03)	−0.23 (.003)	−0.37 (<.001)	0.42 (< .001)
EP				1	0.17 (.03)	0.08 (.30)	−0.31 (< .001)
CC					1	0.30 (<.001)	−0.40 (< .001)
PE						1	−0.43 (< .001)

Note: AQD: appropriate question differential; RRB: rapport and relationship-building; EP: emotion provocation; CC: confrontation/competition; PE: presentation of evidence.

The final statistical analysis explored the effects of the AQD and domain variables in a multiple ordered logistic regression model examining its association with suspect cooperation (Table 3). The proportional odds assumption of an ordered logistic regression model was not violated in this sample according to the Brant test of parallel regression (χ² = 23.25, df = 18, p = .18). Controlling for when in the interview it took place (interval number) and the interdependence of observations (i.e., intervals are more alike within interviews than between interviews) using the cluster option in Stata that produces robust standard errors, we found that the AQD was not significantly associated with cooperation, although rapport and relationship-building was associated with greater cooperation and emotion provocation and presentation of evidence was associated with lower cooperation. Given the significant association between the AQD and suspect cooperation in the bivariate analysis was rendered non-significant³ in the regression model, we found only partial support for Hypothesis 2.

Table 3.

Multiple ordered logistic regression predicting suspect cooperation.

	B	SE	p	95% CI
Appropriate Question Differential	0.37	0.51	.47	−0.63, 1.37
Rapport and relationship-building	0.98	0.21	< .001	0.57, 1.39
Emotion provocation	−0.64	0.21	.002	−1.05, −0.23
Confrontation/competition	−0.77	0.56	.17	−1.86, 0.33
Presentation of evidence	−1.09	0.30	< .001	−1.68, −0.50
Interval	0.09	0.04	.04	0.01, 0.18
Model N / Pseudo R²	141 / .19

Figures 1 and 2 are presented as examples of the Griffiths Question Map (GQM). These figures were purposely chosen because of their similar lengths and dissimilar outcomes, where the greater-than-average AQD was associated with suspect admissions and the lower-than-average AQD with a denial. Whereas the analyses to this point used the 5-minute interval as the unit of analysis, each dot on the map is an individual question per the GQM convention. Figure 1 is of an interrogation that lasted ∼ 95 minutes (19 intervals) and of a suspect who made admissions of guilt. Investigators asked a total of 179 questions, with an AQD of 0.50 which is greater than the sample average as a whole (0.23). Figure 2 is that of a suspect who denied guilt throughout the 90-minute (18 intervals) interrogation, with an AQD of 0.13. As is clear in the figure, investigators asked many more questions of this suspect, and its total of 309 questions is the most of any of the interrogations in the current sample. The implications of these question maps, including the role of the AQD in them, is discussed in the next section.

Figure 1.

GQM of interrogation that ended with a partial admission (AQD = 0.50; SD = 0.42).

Figure 2.

GQM of interrogation that ended in denial (AQD = 0.13; SD = 0.43).

Discussion

Let us return to the fictional bank robbery and interviews from the opening of this article. Lt Hanna could have posed his closed question to Neil, the first suspect as, “Did you intend to make some poor bastard’s wife a widow or were you just trying to take the score and things got out of hand?”. As phrased, this is a forced choice question (providing either/or options) and would be regarded as inappropriate following the definitions above. It is also properly considered as an interview technique, as it represents the alternative question of the Reid method (Inbau et al., 2013) and can be considered as offering a rationalization for the behavior, a form of minimization (Kelly et al., 2019).

Another example of the interaction between questions and techniques is Sgt Drucker could say to Chris, “Tell us what you were doing hanging out near the bank several times in the past two weeks”, current analyses framed interview while showing him still photographs taken from nearby closed-circuit cameras. Because there is evidence demonstrating that the suspect had been possibly surveilling the bank, the question is rooted in fact and provides the suspect an opportunity to offer his account of the situation to the authorities. There is neither an accusation nor opinion implicitly or explicitly stated in the question, it is not leading or closed in any manner, and by all measures would be considered a productive question combined with a technique of evidence presentation.

These examples highlight the importance of considering the relationships between two clusters of interview methods—question type and techniques—and their relative influences on suspect cooperation as explored in the current paper. As hypothesized, we found that appropriate questions were associated with higher use of rapport-based techniques and lower use of accusatorial interrogation techniques, and in bivariate analyses, appropriate questions were related to higher cooperation. These findings demonstrate that associations between questions and techniques exist, and although the relationship between questions and suspect cooperation becomes non-significant in the final ordered logistic regression model, the study overall makes several important contributions to the literature.

Implications for research and practice

Development of the AQD stands as a contribution in its own right, given that investigators tended to ask a lot of questions, and the AQD captures the relative rate of question types asked in a single measure. Indeed, investigators asked an average of ∼ 151 questions per interrogation. This number is reported here instead of above because it is not of direct import to the primary research questions and hypotheses, nor are we prepared to make a judgment about an ideal number of questions that should be asked. Instead, we cite it here to note that a lot of questions were asked during the sample interviews and to contextualize the significant, positive relationship between the number of appropriate and inappropriate questions asked (from Table 2). This result indicates that investigators in this sample simply asked a lot of questions with little or no discrimination between the types. This may not be a surprising revelation—as far as we are aware, the investigators had no proper instruction on questioning strategies—but future training and evaluations of same should look to reverse this correlation.

A negative correlation between appropriate and inappropriate questions would, by definition, lead to the desirable outcome of higher scores on the AQD. Although statistical associations between the AQD and whether or not a suspect confesses should be considered in future research, with only 15 interrogations in the sample, such an examination was outside the scope of the data for the current study. However, this variable advances the research on question types in investigative interviewing generally.

As for its applicability to the GQM, the AQD offers a global assessment of the quality of questioning that can be applied across a sample of interviews or as a measure of individual performance. This, then, supplements the common use of the GQM as a training tool that graphically shows an interviewer’s performance with a statistical calculation. Indeed, we can examine the available GQMs and calculate AQDs for them. The interviews classified as “skilled” by Griffiths in his PhD thesis (2008) had an AQD of 0.83, and the skilled interview in Walsh and Bull (2015) was 0.84. The interview of a child abuse suspect reported by Griffiths and Milne (2006), on the other hand, had an AQD of −0.14, and the “oppressive” interview example offered by Griffiths was −0.15. Put differently, the AQD represents a single score for an interviewer’s performance, and although we offer no thresholds herein, future research may seek to determine AQD ranges for the quality of questioning during investigative interviews.

Next, there is a much-needed conversation in the investigative interviewing literature about the theoretical links between the independent and dependent variables in our studies. Early research in the field focused on the outcome of confessions, notably false confessions, and much research in the USA still tends to focus on what causes both true and false confessions because of the weight this outcome is given in the legal system. As interviewing models change their focus from confession-seeking to information-gathering (e.g., PEACE), the research follows suit and grows beyond even just examining these two outcomes to examine other behaviors of or responses by interviewees (e.g., cooperation, resistance, engagement). What has not been debated or fully fleshed out are the theoretical links between x and y.

For example, this study suffers in this regard with the dependent variable of cooperation. Although higher scores on the AQD were significantly (but modestly) correlated with suspect cooperation, this relationship disappeared in the ordered logistic regression model that included the four domain measures. The non-relationship in the regression model may have been predicted if we had previously considered what questions are meant to elicit—namely, verbal responses that provide information. The resistance or cooperation of suspects may just not heavily depend on question type; instead, this outcome may be more theoretically relevant to domains that encompass the techniques investigators may take on the broadly defined scale of rapport-based to accusatorial approaches. Cooperation, then, may serve as a necessary-but-not-sufficient mediating variable when the outcome of interest is information yield. In short, rapport may be better suited to gain the cooperation of a suspect than asking a TED question, but simply finding common ground or employing mimicry will not directly elicit information. An investigator must ask a question to receive that information.

As our findings demonstrate, however, higher AQD scores were significantly associated with the rapport and relationship-building domain and negatively associated with the other three domains that are (or in the case of presentation of evidence, may be) accusatorial in nature. There are indications, then, that questions may take on meaning beyond the structural (literal) or pragmatic (implied) qualities of them, and similar to the previous paragraph, appropriate questions are also insufficient on their own for reaching an investigator’s goal.

Although the data presented here were an imperfect vehicle to test the complexity of investigative interviewing, we did find that investigators who asked appropriate questions were likely to employ rapport-based techniques and vice versa. Putting this finding together with the prior two paragraphs, this study points to understanding and considering interview methods and suspect responses as a complex sequence in which rapport-based techniques may make an uncooperative suspect cooperative, and appropriate questioning practices result in more and better information. This sequence is applicable—albeit in different ways—to both researchers and professional interviewers. For the former, future studies and statistical models should be developed to properly test it; for the latter, it can be a consideration for evaluating the progress of an interview and the interview methods employed throughout it.

Limitations

This study is based on a small sample of non-randomly selected suspect interviews provided by a single investigations unit from a single police department in the USA. It would be a brazen act to claim that the results presented above are at all representative of investigative interviewing, even within the unit that provided the recordings, let alone reflective of practice in other parts of the USA or beyond. Although we are confident in the results as they pertain to the sample data presented above, we must recognize that the arguments and findings presented herein ought to be tested with larger and more representative samples of investigative interviews.

A third cluster?

The current analyses framed interview methods as one of two clusters—question type and techniques—and defined them as being under the control of interviewers. Another area of research, however, points to the likelihood that a third cluster exists but went unobserved in the current study. Specifically, we had no measures of the attitudes or dispositions of the interviewers who asked the questions and employed the techniques. There is a good deal of evidence in the literature to state that the disposition of the interviewer is a coequal cluster of interview methods, beginning with the Holmberg and Christianson (2002) analysis of convicted offenders’ perceptions of being interviewed. In that study, the authors reported an array of attitudes interviewers could adopt, sorted into two general styles—humanity versus dominance. Unsurprisingly, humane approaches were preferred by the offenders who were more likely to confess than if the interviewer adopted a dominant stance (see Kebbel et al., 2008, 2010, for similar findings with sex offenders specifically).

A concept related to humanity and that arguably falls within the disposition cluster is that of empathy. Demonstrations of empathy are within the control of the interviewer, and they may take on the form of a technique as defined here (e.g., saying to the suspect, “I understand that this is difficult for you to talk about”). Oxburgh and Ost (2011: 181) argue that “empathy is not just about the police officer ‘showing’ empathy…[it] can be seen as a multi-dimensional phenomenon comparison both cognitive processes and emotional (or affective) capacities”. Put differently, empathy is demonstrated in many ways throughout an interview and can be understood as a general disposition.

The ORBIT framework (Alison et al., 2013) referred to above advanced the measurement of investigative interviewing in ways that speak directly to the current analyses. ORBIT rests on two pillars of what we refer to here as clusters. First, they adapted Leary and Coffey’s (1954) Interpersonal Behavior Circle model, which is the disposition of the interviewer (and the suspect to boot) that they defined as either adaptive or maladaptive. The second cluster captures the use of six principles of Motivational Interviewing (MI; Miller and Rollnick, 2009)—acceptance, autonomy, adaptation, empathy, and evocation—which generally fits the definition of technique given above. Note that empathy is included in this list of techniques which underscores the conceptual messiness that arises between research teams and across studies, but empathy could very well be both a technique and a disposition depending on how it is operationalized and measured. Of direct relevance is the finding that greater use of MI techniques are associated with higher interviewer adaptive scores, a between-cluster positive association that echoes in this study.

Indeed, we acknowledge here that neither the current analyses nor the work of Alison et al. (2013, 2014) were the first to analyze the more than one cluster of interview methods. Oxburgh and colleagues (2012) examined both empathy and question type in their sample of sex offenders, Soukara et al. (2009) measured a variety of question types and techniques (see also Bull and Soukara, 2010), and Walsh and Bull (2012) analyzed attitudes (i.e., disposition) and tactics (i.e., techniques). These studies laid the groundwork for later examinations, such as the current one, to begin understanding the associations between clusters.

Conclusion

For the final time, recall the fictive interviews from above, and imagine Lt Hanna saying to Neil, “What do you say I buy you a cup of coffee?”. This is a closed question, of course, but this utterance is unlikely to be coded in a study strictly evaluating the type of question asked. Meeting the suspect’s basic needs by offering him something to drink, however, can be operationalized as a technique of rapport-building (Kelly et al., 2013). Moreover, offering the suspect coffee can demonstrate empathy for his state of deprivation while detained and being questioned or it can be considered an invaluable technique that facilitates rapport-building and contributing to a sense of reciprocity. In short, this question is a simplistic example of the basic premise of this paper: that questions, techniques, and dispositions should not be treated as discrete elements of investigative interviewing, as they mostly have in the empirical literature. Instead, researchers must consider the interactive effects of the three components, because they have each been demonstrated as critical to the effectiveness of investigative interviewing.

In conclusion, the primary takeaway of this inquiry is similar to the parable of the blind men and the elephant. In that story, six blind men each described a section of an elephant that they were touching. The discrete descriptions of a leg, the trunk, the tail, an ear, the stomach, and a tusk were incorrectly interpreted, but only when considered together could the various body parts accurately describe an elephant. As it pertains to investigative interviewing and interrogation methods, then, question type, techniques, and dispositions can be considered the various body parts of the larger whole that is an interpersonal dynamic between two individuals where one is in possession of information or knowledge that the other is trying to obtain.

Footnotes

Acknowledgements

An earlier version of this research was presented at the 2016 International Investigative Interviewing Research Group (iIIRG) in Buckinghamshire, United Kingdom. The authors wish to thank the following individuals: Susan Brandon and Christian Meissner, Captain William Hayes and Detectives Greg Stearns and Tim Marcia of the Robbery-Homicide Division of the Los Angeles Police Department for providing the recordings.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was funded by the High-Value Detainee Interrogation Group (HIG) contract awarded to subcontractors Christopher Kelly, Allison Redlich, and the University at Albany, State University of New York, through the University of Texas at El Paso. Statements of fact, opinion and analysis in the paper are those of the authors and do not reflect the official policy or position of the FBI or the U.S. Government. Additional support was provided by the Saint Joseph's University Summer Scholars Program.

Notes

References

Alison

Noone

Eltnib

Christiansen

(2013) Why tough tactics fail and rapport gets results: observing rapport-based interpersonal techniques (ORBIT) to generate useful information from terrorists. Psychology, Public Policy, and Law 19: 411–431.

Alison

Noone

Elntib

Waring

Christiansen

(2014) The efficacy of rapport based techniques for minimizing counterinterrogation tactics amongst a field sample of terrorists. Psychology, Public Policy, and Law 20: 421–430.

Baldwin

(1993) Police interview techniques: establishing truth or proof? The British Journal of Criminology 33(3): 325–352.

Beauregard

Deslauriers-Varin

St-Yves

(2010) Interactions between factors related to the decision of sex offenders to confess during police interrogation: a classification-tree approach. Sexual Abuse 22(3): 343–367.

Bull

Soukara

(2010) Four studies of what really happens in police interviews. In: Lassiter

Meissner

(eds) Police Interrogations and False Confessions: Current Research, Practice, and Policy Recommendations. Washington, DC: American Psychological Association, 81–95.

Clarke

Milne

(2001) A National Evaluation of the PEACE Investigative Interviewing Course. London: Home Office.

Dawson

Hartwig

Brimbal

(2015) Interviewing to elicit information: using priming to promote disclosure. Law and Human Behavior 39: 443–450.

Dodier

Denault

(2018) The Griffiths Question Map: a forensic tool for expert witnesses’ assessments of witnesses and victims’ statements. Journal of Forensic Sciences 63(1): 266–274.

Griffiths

(2008) An examination into the efficacy of police advanced investigative interview training. PhD thesis, University of Portsmouth.

10.

Griffiths

Milne

(2006) Will it all end in tiers? In: Williamson

(ed) Investigative Interviewing: Rights, Research, Regulation. Cullompton, UK: Willan, 167–189.

11.

Griffiths

Milne

Cherryman

(2011) A question of control? The formulation of suspect and witness interview question strategies by advanced interviewers. International Journal of Police Science & Management 13(3): 225–267.

12.

Farrugia

Oxburgh

Gabbert

(2019) Effective evaluation of forensic interviews: the Forensic Interview Trace (FIT). Investigative Interviewing: Research & Practice 10(1): 20–31.

13.

Feld

(2006) Police interrogation of juveniles: an empirical study of policy and practice. The Journal of Criminal Law and Criminology 97(1): 219–316.

14.

Feld

(2013) Kids, Cops, and Confessions: Inside the Interrogation Room. New York: New York University Press.

15.

Holmberg

Christianson

(2002) Murderers’ and sexual offenders’experiences of police interviews and their inclination to admit or deny crimes. Behavioral Sciences & the Law 20: 31–45.

16.

Inbau

Reid

Buckley

Jayne

(2013) Criminal Interrogation and Confessions. Burlington, MA: Jones and Bartlett Learning.

17.

Kebbell

Alison

Hurren

(2008) Sex offenders’ perceptions of the effectiveness and fairness of humanity, dominance, and displaying an understanding of cognitive distortions in police interviews: a vignette study. Psychology, Crime & Law 14(5): 435–449.

18.

Kebbell

Alison

Hurren

Mazerolle

(2010) How do sex offenders think the police should interview to elicit confessions from sex offenders? Psychology, Crime & Law 16(7): 567–584.

19.

Kelly

Dawson

Hartwig

(2019). Context manipulation in police interviews: a field experiment. Journal of Experimental Criminology. https://doi.org/10.1007/s11292-019-09389-8

20.

Kelly

Miller

Redlich

(2016) The dynamic nature of interrogation. Law and Human Behavior 40(3): 295–309.

21.

Kelly

Miller

Redlich

Kleinman

(2013) A taxonomy of interrogation methods. Psychology, Public Policy, and Law 19(2): 165–178.

22.

Kelly

Redlich

Miller

(2015) Examining the meso-level domains of the interrogation taxonomy. Psychology, Public Policy, and Law 21(2): 179–191.

23.

Kelly

Russano

Miller

Redlich

(2019) On the road (to admission): engaging suspects with minimization. Psychology, Public Policy, and Law 25(3): 166–180.

24.

King

Snook

(2009) Peering inside a Canadian interrogation room: an examination of the Reid Model of interrogation, influence tactics, and coercive strategies. Criminal Justice and Behavior 36: 674–694.

25.

Kopše

(2017) The use of the Griffiths Question Map for the analysis of interrogations proceeded by courts–a case study of lawyer’s interrogation technique. Journal of Criminal Justice and Security 3: 234–251.

26.

Leary

Coffey

(1954) The prediction of interpersonal behavior in group psychotherapy. Group Psychotherapy 7: 7–51.

27.

Leo

(1996) Inside the interrogation room. Journal of Criminal Law & Criminology 86(2): 266–303.

28.

Mann

(director) (1995) Heat [Film]. Regency Enterprises.

29.

Miller

Redlich

Kelly

(2018) Accusatorial and information-gathering interview and interrogation methods: a multi-country comparison. Psychology, Crime & Law 24(9): 935–956.

30.

Miller

Rollnick

(2009) Ten things that motivational interviewing is not. Behavioural and Cognitive Psychotherapy 37: 129–140.

31.

Milne

Bull

(1999) Investigative Interviewing: Psychology and Practice. Chichester, UK: Wiley.

32.

Moston

Stephenson

Williamson

(1992) The effects of case characteristics on suspect behaviour during police questioning. British Journal of Criminology 32(1): 23–40.

33.

O’Mahoney

Milne

Grant

(2012) To challenge, or not to challenge? Best practice when interviewing vulnerable suspects. Policing 6(3): 301–313.

34.

Oxburgh

Ost

(2011) The use and efficacy of empathy in police interviews with suspects of sexual offences. Journal of Investigative Psychology and Offender Profiling 8(2): 178–188.

35.

Oxburgh

Ost

Cherryman

(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.

36.

Oxburgh

Myklebust

Grant

(2010) The question of question types in police interviews: A review of the literature from a psychological and linguistic perspective. International Journal of Speech Language and the Law 17(1): 45–66.

37.

Pearse

Gudjonsson

(1999) Measuring influential police interviewing tactics: a factor analytic approach. Legal and Criminological Psychology 4(2): 221–238.

38.

Phillips

Oxburgh

Gavin

Myklebust

(2012) Investigative interviews with victims of child sexual abuse: the relationship between question type and investigation relevant information. Journal of Police and Criminal Psychology 27(1): 45–54.

39.

Read

Powell

Kebbell

Milne

(2009) Investigative interviewing of suspected sex offenders: a review of what constitutes best practice. International Journal of Police Science & Management 11(4): 442–459.

40.

Redlich

Kelly

Miller

(2014) The who, what, and why of human intelligence gathering: self-reported measures of interrogation methods. Applied Cognitive Psychology 28(6): 817–828.

41.

Russano

Narchet

Kleinman

Meissner

(2014) Structured interviews of experienced HUMINT interrogators. Applied Cognitive Psychology 28(6): 847–859.

42.

Senese

(2008) Anatomy of Interrogation Themes. Chicago, IL: John E. Reid and Associates.

43.

Snook

Luther

Quinlan

Milne

(2012) Let ‘em talk: a field study of police questioning practices of suspects and accused persons. Criminal Justice and Behavior 39(10): 1328–1339.

44.

Soukara

Bull

Vrij

Turner

Cherryman

(2009) What really happens in police interviews of suspects? Tactics and confessions. Psychology, Crime & Law 15(6): 493–506.

45.

Surmon-Böhr

Alison

Christiansen

Alison

(2020) The right to silence and the permission to talk: Motivational Interviewing and high-value detainees. American Psychologist 75(7): 1011–1021.

46.

Vrij

Leal

Granhag

Mann

Fisher

Hillman

Sperry

(2009) Outsmarting the liars: the benefit of asking unanticipated questions. Law and Human Behavior 33: 159–166.

47.

Walsh

Bull

(2012) How do interviewers attempt to overcome suspects’ denials? Psychiatry, Psychology and Law 19(2): 151–168.

48.

Walsh

Bull

(2015) Interviewing suspects: examining the association between skills, questioning, evidence disclosure, and interview outcomes. Psychology, Crime & Law 21(7): 661–680.

49.

Walsh

Milne

(2008) Keeping the PEACE? A study of investigative interviewing practices in the public sector. Legal and Criminological Psychology 13(1): 39–57.

50.

Waterhouse

Ridley

Bull

La Rooy

Wilcock

(2019) Mapping repeated interviews. Journal of Police and Criminal Psychology 34(4): 392–409.