The Scharff-Technique: Eliciting Intelligence From Human Sources

Law and Human Behavior 2014, Vol. 38, No. 5, 478 – 489 © 2014 American Psychological Association 0147-7307/14/$12.00 DOI: 10.1037/lhb0000085 The Scharff-Technique: Eliciting Intelligence From Human Sources Simon Oleszkiewicz Pär Anders Granhag University of Gothenburg, Sweden University of Gothenburg, Sweden, and Norwegian Police University College Sebastian Cancino Montecinos This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. University of Gothenburg, Sweden This study is on how to elicit intelligence from human sources. We compared the efficacy of two human intelligence gathering techniques: the Scharff-technique (conceptualized as four different tactics) and the Direct Approach (a combination of open and direct questions). Participants (N ⫽ 60) were asked to take on the role of “sources” and were given information about a planned terrorist attack. They were to reveal part of this information in an upcoming interview. Critically, the participants were instructed to strike a balance between not revealing too much or too little information. As predicted, the participants revealed significantly more, and more precise, new information when interviewed with the Scharff-technique (vs. the Direct Approach). Furthermore, and as predicted, the participants in the Scharff condition underestimated how much new information they revealed whereas the participants in the Direct Approach overestimated how much new information they revealed. The study provides rather strong support for the Scharff-technique as an effective human intelligence gathering technique. Keywords: Scharff-technique, Direct Approach, human intelligence gathering, information elicitation interaction has a broad interest with regards to the type of information to be gathered, including information related to the past, present, and future (Evans et al., 2010; Redlich, 2007). Information elicitation is a specific form of human intelligence, in which the goal is to gather information in such a manner that the source remains unaware of the true effect of the exchange (Justice et al., 2010). The legal-psychology literature has very little to offer with respect to techniques aimed at eliciting intelligence from human sources. This is remarkable considering the prominent role of human intelligence collection and the resurgent interest in collecting intelligence in the period following 9/11. For example, only a handful of science-based psychological initiatives have had any impact on policies or practices of U.S. national security agencies from WWII until today (Brandon, 2011). Critically, it is likely that this trend also holds true for the rest of the world. Thus, although operational experience and the need for information tailored to the needs of national security interests have given rise to a wide array of human intelligence techniques, these have rarely been subjected to any scientific evaluation (Dujmovic, 2005; Varouhakis, 2013). The present study set out to remedy this by examining two techniques aimed at eliciting intelligence from human sources. Human intelligence is best described as the gathering of information by means of an interaction between two or more individuals (Justice, Bhatt, Brandon, & Kleinman, 2010). The general purpose of human intelligence gathering is to collect intelligence in order to improve national security (Evans, Meissner, Brandon, Russano, & Kleinman, 2010). To this end, the human intelligence Hanns Scharff: A Master at Perspective-Taking Hanns Joachim Scharff (1907–1992) was a very successful interrogator working for the German Luftwaffe during WWII (Shoemaker, 2008; Toliver, 1997). After the war, his interrogation techniques gained recognition due to his unconventional, humane, and noncoercive approach (Granhag, 2010). By observing the interrogations conducted at the Luftwaffe’s Intelligence and Evaluation Centre Auswertestelle West, Scharff discovered that prisoners adopted different strategies to resist the conventional interrogation techniques used (Toliver, 1997; Scharff, 1950). Hence, Scharff started to take the perspective of the prisoners (mainly American and British fighter pilots) in order to learn how to counteract their counterinterrogation strategies (Granhag, 2010). In essence, Scharff developed a human intelligence gathering technique that rested upon taking the perspective of the source (prisoner). Perspective taking is the cognitive capacity to consider the world from another’s viewpoint, which allows an individual to anticipate other people’s behavior and reactions (Galinsky, Maddux, Gilin, & White, 2008). Importantly, the ability to take the perspective of others is predictive of success in negotiations (Galinsky et al., 2008), and of importance for interrogators (e.g., Justice et al., 2010; Soufan, 2011). The 10⫹ years research pro- This article was published Online First April 7, 2014. Simon Oleszkiewicz, Department of Psychology, University of Gothenburg, Sweden; Pär Anders Granhag, University of Gothenburg, Sweden, and Norwegian Police University College; Sebastian Cancino Montecinos, University of Gothenburg, Sweden. Correspondence concerning this article should be addressed to Simon Oleszkiewicz, Department of Psychology, University of Gothenburg, P. O. Box 500, 405 30 Gothenburg, Sweden. E-mail: simon.oleszkiewicz@ psy.gu.se 478 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. ELICITING INTELLIGENCE gram on the Strategic Use of Evidence (SUE) technique provides ample evidence that the understanding of suspects’ counterinterrogation strategies (e.g., Granhag & Hartwig, 2008; Granhag, Mac Giolla, Strömwall, & Rangmar, 2012) can be translated into effective interview tactics (e.g., Granhag & Hartwig, in press; Granhag, Strömwall, Willén, & Hartwig, 2013; Clemens, Granhag, & Strömwall, 2011). Scharff used perspective-taking to identify and play on the counterinterrogation strategies that his prisoners adopted in order to withstand the interrogation (Granhag, 2010; Granhag et al., 2013). Utilizing different sources on Scharff’s method (e.g., Toliver, 1997; Scharff, 1950), three such counterinterrogation strategies can be identified: (i) “I will not tell very much during the interrogation”; (ii) “I will try to figure out what they are after, and then make sure not to give them what they want”; and (iii) “It is meaningless to withhold or deny what they already know.” Having identified the most common counterinterrogation strategies used, Scharff formed his own tactics to those of the prisoners. In brief, perspective taking is at the very core of the Scharff-technique. The most recognized Scharff-tactic is that he convinced his prisoner that he already knew all the important information (i.e., he painted the “illusion of already knowing it all”). But there is much more to the Scharff-technique than this single tactic. Granhag (2010) derived four tactics from the Scharff framework, which will be employed for the present study (see also Granhag et al., 2013). The Scharff-Tactics The first tactic is the friendly approach. Scharff was known for his friendly and conversational approach and was knowledgeable in both American and British customs. The second tactic is not pressing for information. Instead of asking explicit questions, Scharff told long and detailed stories related in a fashion that offered his prisoners the opportunity— even encouragement—to add details or correct apparent errors. The third tactic is the illusion of knowing it all. Scharff often opened the interaction by stating that it was unlikely that the prisoner would be able to offer any information beyond what he already knew. Scharff then told a detailed story that made it utterly clear that he was very well informed on the topic. The fourth tactic is the Confirmation/disconfirmation tactic. Scharff worked with confirmation and disconfirmation in creative ways. For example, he would systematically present claims to the prisoner that he himself knew the correct answer to, but every once in a while he would make a claim for which he did not know the correct answer. Hence, Scharff could elicit information without revealing his information objectives even though his prisoners did not say very much. For an example illustrating the specifics of the confirmation tactic, consider a situation where there is intelligence pointing in two different directions; either there will be a terrorist attack in Washington or in New York City. Assume there is more reliable information pointing toward Washington than toward NYC. An interviewer who wants to elicit information from a source on the location for the attack has a number of different options. One alternative is to avoid asking an explicit question and instead make a claim. The interviewer can either claim what s/he considers to be the most probable alternative (“So, we know that Washington is the target!”) and note whether the source dis/ confirms. Or claim what s/he considers to be the less probable 479 alternative (“So, we know that New York is the target!”), and note whether the source dis/confirms. The advantage of using claims is that the source might be more willing to respond to these than to explicit questions as this would be a less active form of complicity. More specifically, from the perspective of the source, confirming a claim might be viewed as “I only confirmed what they already knew,” and disconfirming might be viewed as “I only told them they were off target.” Importantly, our conceptualization of the Scharff-technique is very different to the Scharff-technique as described in the U.S. Army Field Manual (FM 2–22.3). The Army Field Manual provides a sketchy description, essentially limiting the technique to the “illusion of knowing it all” tactic. In contrast, we argue that the Scharff-technique is a rather complex interview technique, consisting of several tactics which are related in intricate ways (see Granhag et al., 2013). Here it should also be emphasized that the aim of the current study is not to offer a full and final conceptualization of the Scharff-technique. Instead, the main aim is to conduct an empirical test of one set of tactics derived from the framework used by Hanns Scharff. The Direct Approach In 2009 President Obama formally established the U.S. Army Field Manual FM 2–22.3, Human Intelligence Collector Operations as the new gold standard for conducting intelligence interrogations (Executive Order 13491). The Army Field Manual instructs the interrogator to approach the source in a business-like manner, posing direct questions (the Direct Approach), and to stay with this approach for as long as the source answers these questions in a truthful manner. Only when the source avoids or refuses to answer the direct questions, or when the answers are perceived as untruthful, is the interrogator recommended to adopt an alternative approach. Essentially, the Direct Approach rests on a combination of open-ended and explicit questions, while avoiding the use of leading questions. In accordance to the Army Field Manual’s recommendation, the Direct Approach has proven to be one of the most commonly used techniques by both military (81%) and civilian (45%) interrogators (Redlich, Kelly, & Miller, 2011). Because of its widespread use, we consider the Direct Approach a relevant point of comparison for the current study. The First Test of the Scharff-Technique In the first empirical study of the comparative efficacy of intelligence gathering techniques, Granhag et al. (2013) introduced an experimental paradigm mirroring some central features of a typical human intelligence interaction. The results showed that the intelligence gathering techniques examined (Open Questions, Specific Questions, and the Scharff-technique) elicited a similar amount of new information. As predicted, sources interviewed with the Scharff-technique incorrectly perceived to have revealed a lesser amount of information than sources interviewed with Specific Questions. Furthermore, as predicted the sources interviewed with the Scharff-technique found it more difficult to read the interviewer’s information objectives compared to the sources in the two control conditions. The unexpected finding that the Scharff-technique did not elicit more new information than the two control techniques was attributed to a rather disorganized imple- OLESZKIEWICZ, GRANHAG, AND CANCINO MONTECINOS 480 mentation of the “illusion of knowing it all” tactic and the “confirmation” tactic. For the current study, we will remedy this by properly establishing the “illusion of knowing it all” before putting the “confirmation” tactic into play. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. The Present Study The major aim is to examine the efficacy of the Scharfftechnique and the Direct Approach. To this end we will draw on an experimental set-up mirroring some main features of a typical human intelligence situation. Moreover, we will use a set of new measures for examining the efficacy of intelligence gathering techniques, introduced by Granhag et al. (2013). Most importantly, the present study will advance the previous work by Granhag and colleagues on several accounts. First, we will implement the central Scharff-tactics in a stepwise manner (i.e., establishing the “illusion of knowing it all” before introducing the “confirmation” tactic). Second, we will compare the Scharff-technique against the Direct Approach, which is a more relevant comparison. Third, we will introduce yet another measure of efficacy; the degree of precision of the new information elicited (see, e.g., Goldsmith, Koriat, & Weinberg-Eliezer, 2002). We consider this measure highly warranted as more precise intelligence is often more actionable. Fourth, we will include two postinterview checklists for the participants to mark (1) the specific pieces of information that they perceived to have revealed during the interview, and (2) the specific pieces of information that they believed was known to the interviewer prior to the interview. This will allow us to assess the extent to which each participant over- or underestimated the amount of new information revealed. This is a central measure of efficacy of information elicitation which could not be properly tapped in the previous study by Granhag et al. (2013). Fifth, we will—in addition to the total amount of new information elicited— also map how much each particular phase of the interview contributed to the total amount of new information. To this end we divided the interview into three phases (initial open-ended question, claims/specific questions, and the final open-ended question). interviewer. In contrast, for the Direct Approach the sources are free to report the information they have planned to tell, and this is expected to be a mix of new information and information already known by the interviewer. We predict that the Scharff-technique will outperform the Direct Approach in terms of new information elicited both for the first (Hypothesis 3a) and the second (Hypothesis 3b) phase of the interview. For the third phase, we did not expect any difference as the question asked is the same for both conditions. Subjective Perceptions of the Interview We predict that sources in the Scharff condition will perceive the interviewer to have more information prior to the interview compared to sources in the Direct Approach (Hypothesis 4a). Furthermore, we predict that the sources in the Scharff condition will have a less clear understanding of what information the interviewer was after (Hypothesis 4b). The basis for this prediction is that for the Scharff-technique no specific questions will be asked. Relating Objective and Subjective Measures We predict an interaction effect related to (a) the objective amount of new information revealed and (b) the perceived amount of new information revealed. We predict that the sources in the Scharff condition will perceive to have revealed significantly less new information than they objectively did, whereas we expect the sources in the Direct Approach condition to perceive that they have revealed more new information than they actually did (Hypothesis 5). The rationale behind this prediction is that we expect the Scharff tactics to result in the sources unknowingly revealing new information. In contrast, the sources in the Direct Approach are expected to report a mix of new and old information, but estimate that (almost) all revealed information will be new to the interviewer. Method New Information Elicited We predict that the sources in the Scharff condition will reveal more new information than the sources in the Direct Approach (Hypothesis 1), and we predict that the new information elicited will reach a higher degree of precision for the Scharff condition (Hypothesis 2). Both predictions are backed up by the same twofold reasoning. First, for sources faced with the Scharfftechnique, the interviewer will tell a “story” in order to create the “I already know it all” illusion. Hence, it will make no sense for the sources to repeat information that was part of the interviewer’s story (i.e., to tell information that the source knows the interviewer to already hold). In order to be perceived as helpful sources, they instead need to tell information beyond what the interviewer just told. If they do, the information they provide will be new to the interviewer. Second, sources faced with the Scharff-technique are expected to come to believe that the interviewer holds information that is, in fact, unknown to the interviewer. If these sources, in turn, act on the “it is meaningless to withhold what the interviewer already knows” counterinterrogation strategy, it may follow that they (unintentionally) will reveal information that is new to the Participants Sixty adults (37 women and 23 males) between 19 and 66 years of age (M ⫽ 25.8, SD ⫽ 8.2) were randomly allocated to two conditions (30 participants in each condition). Participants were all students at Gothenburg University, and they were recruited by advertisements at the university’s libraries and departments. The ads used for recruitment described the study as research on “decision-making under uncertainty.” Each participant was paid an amount equivalent to €15 (approx. $20). Before the experiment, all participants read and agreed to a standard informed consent form. After the experiment, all participants were fully debriefed. Procedure Phase 1: Background and planning. First, all participants received identical instructions that they had to take on the role as a source that possessed incomplete information about an upcoming terrorist attack. They were presented with a sheet of paper containing both general and specific information about a left wing This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. ELICITING INTELLIGENCE extremist group who was planning a bomb attack against a shopping mall in the city center (for a full description of this background information, see Appendix A). Participants were asked to imagine that this was the information that they held about the attack. The information was given in the form of a coherent story that was built up by a total of 35 separate pieces of information. Of those 35 pieces, 13 pieces were already known to the interviewer, and 22 pieces were unknown to the interviewer. The participants received no information on how many or which pieces of information the interviewer already held. In addition, each participant was informed that s/he needed to consider a dilemma which involved striking a balance between (a) not providing too little (in order to assist the police and to be granted free conduct out of the country) and (b) not providing too much information (as the source had rather strong social ties to the extremist group). Translated to the experimental set-up, the participants were told that their compensation could end up somewhere between €12 and €20 depending on their ability to strike this balance. In fact, all participants were paid the exact same sum (€15). With respect to lying the participants were instructed not to fabricate information during the interview, and the rationale behind this instruction was to arrive at a cleaner comparison between the two interview techniques. However, and importantly, the participants received no instructions with respect to other forms of lying. In fact, in order to properly navigate the information management dilemma imposed the participants had to lie. Specifically, they had to withhold some of the information that they possessed, and consciously withholding information is one form of lying (Vrij, 2008). Phase 2: The Interview. Each participant was instructed to establish contact with his or her “contact” (hereafter interviewer) by making a phone call from his or her room to the interviewer (who was placed in a different room at the psychology department). The participant and the interviewer never met physically, and the participant was alone in the room during the phone conversation. The participant was allowed to have all background information present during the call (all participants used this opportunity). Hence, the participants did not have to memorize the information that they had about the attack. All phoneconversations were taped in order for us to objectively score how much and what information the participant actually revealed to the interviewer. The phone-conversation lasted between 2.47 to 11.51 minutes. An independent-sample t test showed that the Scharfftechnique interview (M ⫽ 7.37, SD ⫽ 1.25) was significantly longer than the Direct Approach interview (M ⫽ 3.85, SD ⫽ 2.33), t(58) ⫽ 5.19, p ⬍ .001, d ⫽ 1.34, 95% CI [0.78, 1.90]. The difference was mainly due to the time it took to present the already known information in the Scharff-technique. Despite the different interviewing protocols (see below), all interviews began and ended in an identical manner. The Scharff-technique. This condition started with the interviewer explaining that the phone call was highly important, and then acknowledged that he understood the problematic situation the source was in (“friendly approach”-tactic). The interviewer also explained that he intended to make the interview more effective by outlining the background information that was already known to him. Then the interviewer presented information to the source without asking any questions, in order to create an illusion 481 of already knowing most of the information about the situation (“illusion of knowing it all”-tactic). The interviewer then again acknowledged that he was well aware of the source’s situation. The interviewer also explained that he had reserved time to listen to the source and that he hoped that the source understood that he did not have any intentions to “sell him/her out” (“friendly approach”-tactic). Then, the interviewer asked an open-ended question: Okay, now I shall not talk more, because I really want to listen to what you have to say . . . so how can you help us with additional information and details? When the source had finished talking, the interviewer asked a follow-up question: Okay, do you have anything else? The interviewer never pressured the source to reveal any information (“not pressure for information”-tactic). After the source’s response to the follow-up question, the interviewer stated five claims that he sought to get confirmed— or disconfirmed— by the source (“confirmation”-tactic). These claims concerned: (i) the location of the attack, (ii) the date of the attack, (iii) if there were bomb experts in the group, (iv) if the bomb would be triggered from a distance, and (v) if the bomb would explode after the stores had closed (e.g., [i] Well, you know, when information started to come in about this, I thought it was typical that Nordstan should be subjected to bomb threats time after time . . . ?; [v] But, Nordstan doesn’t close down completely until 2am, which means that there’s still a risk for people getting killed even if the bomb explodes after the stores are closed . . . ?). All five claims regarded information that was not known to the interviewer, and the claims were made even if the source had touched upon (or fully revealed) the information as a response to the initial open-ended question. After the fifth claim, the interviewer asked a final open-ended question (Is there anything else you can tell us?). After the source had finished talking, the interview was ended. The Direct Approach. This condition started with the interviewer stating in a business-like manner: Okay, shall we start talking about what we are supposed to talk about? As you surely can understand, I am very interested in what you have to tell me about this upcoming event . . . This was directly tied to the initial open-ended question: . . . so, how about you simply tell me what you know? After the source had finished talking, the interviewer asked a follow-up question: Okay, do you have anything else? After the source’s response, the interviewer went on: Okay—thank you very much—-then I have some questions I would like you to answer. These questions concerned the same information, and were asked in the same order, as the five claims in the Scharff condition (e.g., [i] Where will they attack?; [v] Will the bomb explode after the stores have closed?). All five questions regarded new information and were asked even if the source touched upon (or fully revealed) the information answering the initial openended question. After the fifth specific question, the interviewer asked a final open-ended question (Is there anything else you can tell us?). When the source finished talking, the interview was ended. Interviewers. Two interviewers were used and both were trained in conducting these particular interviews. Importantly, the interviewers were instructed to strictly follow the interview protocols and were not allowed to go beyond what was stated in the protocol (and no one did). The number of interviews conducted by each interviewer was spread evenly over the two interview conditions. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 482 OLESZKIEWICZ, GRANHAG, AND CANCINO MONTECINOS Phase 3: Post-Interview Questionnaires. After the interview the participants filled out three questionnaires. Before filling them out, it was made clear that the role-taking part of the study was now over and that the questionnaires should be answered truthfully. The first questionnaire concerned perceptions of the interview: rated on a 7-point Likert scale. The first crucial question being How easy/difficult was it for you to understand what information the interviewer was seeking to obtain?, where the scale was ranging from 1 (very difficult to understand) to 7 (very easy to understand). In addition, we asked the participants to rate how motivated they were during the interview. This was rated on a scale ranging from 1 (not at all motivated) to 7 (very motivated). The second questionnaire consisted of a checklist with all information that was available to the participant (35 units; for the checklist, see Appendix B). Participants were instructed to mark the specific information that they perceived to have revealed during the interview. The third questionnaire, administered after the second, contained the same checklist of the 35 units of information. Here, the participants were asked to mark the specific information they believed was known to the interviewer prior to the interview. Codings of the interviews. Each interview was transcribed verbatim. Each transcript was then coded in terms of how much information the source revealed (range: 0 to 35 units). Importantly, each distinct piece of “new” information revealed by the source was identified. That is, a piece of information was scored as “new” if it was not known to the interviewer prior to the interview (range: 0 to 22 units). The interview was divided into three phases. The first phase consisted of the initial open-ended question. The second phase consisted of presenting five claims (Scharff-technique) or asking five specific questions (Direct Approach). The third phase consisted of the final open-ended question. This allowed us to score the amount of new information elicited during each particular phase. Importantly, each particular piece of new information was only counted once. If the same piece of information was revealed during two different interview phases, it was counted only once (in the first phase it was mentioned). For the first phase, we scored all pieces of new information elicited as a result of the initial openended question (ranging from 0 to 22). For the second phase, we only scored the pieces of information revealed as a result of the five confirmations (Scharff) and the five specific questions (Direct Approach). Thus, the number of information units scored for this particular phase ranged from 0 to 5. Critically, the information revealed as a result of presenting claims was scored and counted only if the participant clearly affirmed a confirmation (e.g., “yes”). Ambiguous responses were not scored (e.g., “hmm”). If the participant—after listening to a claim—reacted with silence, this was not scored. Finally, for the third phase, we scored any new information revealed as a result of the final open-ended question. In addition, we scored the number of “incorrect units” told by the source (across all three phases). Furthermore, we categorized each of the 22 pieces of new information into one of eight themes. One theme could contain a maximum of three pieces of information, running from lower to higher precision (for an example, see below). For each theme, each piece of information was given a value in accordance with its degree of precision. The rule applied was the following: the higher the precision of the information, the higher the value. For example, for the theme where the bomb will be planted, the most general piece of information “in a store” was given a score of 1; the comparatively more precise piece of information “in an electronics store” was given a score of 2, and the most precise piece of information for this particular theme “in the store named Elektronik Experten” was given a score of 3. We then used this fixed table to calculate a “precision score” for each participant. For each theme, only the most precise piece of information was counted (i.e., the piece of information with the highest score). That is, the maximum precision score for an individual theme was 3. In order to arrive at a total “precision score,” we added (for each participant) the scores obtained for the eight themes. The maximum precision score that could be attained by a singular source was 23 (a max score of 3 for 7 themes, and a max score of 2 for one theme; 21 ⫹ 2 ⫽ 23). Interrater reliability. One assistant coded all the transcribed interviews (100%), while another assistant coded 30% of the transcripts (selected from both interview conditions). Based on these 30%, interrater reliability was calculated (Cohen=s ␬ ⫽ 0.93). The disagreements for the 30% of the transcripts were settled in a discussion between the coders after that kappa had been calculated. For the remaining 70% of the material, the primary coder’s scores were used. Results With respect to participants’ motivation, we obtained a mean score well above the midpoint of the scale (M ⫽ 5.57; SD ⫽ 1.15), and we found no differences between the conditions, t(58) ⫽ ⫺1.12, p ⫽ .267, d ⫽ ⫺0.29, 95% CI [⫺0.80, 0.22]. To detect possible interviewer effects, we conducted a two-way ANOVA that examined the effect of interviewer and interview condition on each dependent measure. There was no significant interaction between the interviewer and the interview condition on any dependent measure: the amount of new information revealed, F(1, 58) ⫽ 0.24, p ⫽ .878, d ⫽ 0.13, 95% CI [⫺0.38, 0.64]; the precision of the new information revealed, F(1, 58) ⬍ 0.01, p ⫽ 1.000; the perception of the interviewer’s knowledge, F(1, 58) ⫽ 0.24, p ⫽ .878, d ⫽ 0.13, 95% CI [⫺0.38, 0.64]; the misperceived interviewer knowledge, F(1, 58) ⫽ 0.30, p ⫽ .587, d ⫽ 0.14, 95% CI [⫺0.36, 0.65]; the perception of the interviewer’s information objectives, F(1, 58) ⫽ 0.65, p ⫽ .425, d ⫽ 0.21, 95% CI [⫺0.30, 0.72], or the perception of the amount of new information revealed, F(1, 58) ⬍ 0.01, p ⫽ .969, d ⫽ 0.01, 95% CI [⫺0.49, 0.52]. Hence, we found no interviewer effects. Information Revealed During the Interview The amount of new information. An independent-sample t test showed that the Scharff condition (M ⫽ 7.93, SD ⫽ 2.70) resulted in significantly more new information than the Direct Approach (M ⫽ 5.10, SD ⫽ 2.26) after the full interview, t(58) ⫽ 4.40, p ⬍ .001, d ⫽ 1.14, 95% CI [0.59, 1.68]. Thus, Hypothesis 1 was supported. The precision of the new information. An independentsample t test showed that the new information reported in the Scharff condition (M ⫽ 13.37, SD ⫽ 4.60) reached a significantly higher degree of precision than the new information reported in the Direct Approach (M ⫽ 8.77, SD ⫽ 4.50), t(58) ⫽ 3.91, p ⬍ .001, This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. ELICITING INTELLIGENCE d ⫽ 1.01, 95% CI [0.47, 1.55]. Thus, Hypothesis 2 was supported. As expected, we found that (a) the amount and (b) the degree of precision of the new information was highly correlated; for the Scharff condition, r(29) ⫽ .87, p ⬍ .001, and for the Direct Approach, r(29) ⫽ .89, p ⬍ .001. New information for each interview phase. In order to map how each phase of the interview contributed to the total amount of new information gathered, we conducted an independent-sample t test for each interview phase. The first test examined the outcome of the initial open-ended question and showed that the Scharff condition (M ⫽ 5.53, SD ⫽ 2.30) resulted in significantly more new information than the Direct Approach (M ⫽ 2.90, SD ⫽ 2.47), t(58) ⫽ 4.28, p ⬍ .001, d ⫽ 1.11, 95% CI [0.56, 1.65]. Thus, we found support for Hypothesis 2a. The second test examined the result of presenting confirmations (Scharff condition) or asking explicit questions (Direct Approach), and showed that the Scharff condition (M ⫽ 2.10, SD ⫽ 1.16) resulted in significantly more new information than the Direct Approach (M ⫽ 1.40, SD ⫽ 0.81), t(58) ⫽ 2.71, p ⫽ .009, d ⫽ 0.70, 95% CI [0.18, 1.22]. Thus, Hypothesis 2b was supported. The third test showed that the Scharff condition (M ⫽ 0.30, SD ⫽ 0.54) resulted in significantly less new information than the Direct Approach (M ⫽ 0.80, SD ⫽ 0.85) as a result of the final open-ended question, t(58) ⫽ ⫺2.73, p ⫽ .008, d ⫽ ⫺0.70, 95% CI [⫺1.23, ⫺0.18]. This result was not in line with our expectation. For all means and standard deviations, see Table 1. Incorrect information. In total, we found only four pieces of incorrect information reported during the interviews; all these four pieces were found in the Direct Approach. These four pieces of incorrect information constituted less than 1.7% of all information revealed in this interview condition. The Sources’ Perception of the Interview An analysis of the postinterview checklist showed that participants in the Scharff condition (M ⫽ 11.93, SD ⫽ 2.86) perceived the interviewer to have known significantly more information prior to the interview than participants in the Direct Approach (M ⫽ 5.77, SD ⫽ 3.57), t(58) ⫽ 7.38, p ⬍ .001, d ⫽ 1.91, 95% CI [1.30, 2.52]. Thus, Hypothesis 4a was supported. Relatedly, collapsing both interview conditions, we found a significant positive correlation between perceived interviewer knowledge and the amount of new information reported during the interview, r(59) ⫽ .43, p ⫽ .001. Furthermore, we found that the number of pieces of information that the sources incorrectly believed the interviewer to have held prior to the interview, did not differ comparing the Scharff condition (M ⫽ 3.57, SD ⫽ 1.63) and the Direct Approach 483 (M ⫽ 3.27, SD ⫽ 1.86), t(58) ⫽ 0.67, p ⫽ .509, d ⫽ 0.17, 95% CI [⫺0.33, 0.68]. Finally, the independent-sample t test showed no difference between the Scharff condition (M ⫽ 4.03, SD ⫽ 1.56) and the Direct Approach (M ⫽ 4.17, SD ⫽ 1.62) with respect to the sources’ understanding of what information the interviewer was seeking to obtain, t(58) ⫽ ⫺0.32, p ⫽ .747, d ⫽ ⫺0.08, 95% CI [⫺0.59, 0.42]. Thus, Hypothesis 4b found no support. Above we showed that the condition affected the amount of new information revealed. There is a possibility that the perceived knowledge of the interviewer could mediate this difference. However, a mediation analysis showed no indirect effect of interview condition on new information revealed through the sources’ perceived interviewer knowledge, b ⫽ 0.32, BCa CI [⫺0.29, 1.08], which represents a small effect, ␬2 ⫽ .09, 95% BCa CI [0.00, 0.28] (see Figure 1). Relating the Objective and Subjective Measures A mixed ANOVA with the two interview conditions as the between-subjects factor and the revealed new information scores (Objective, Subjective) as the within-subjects factor was conducted. The interview ⫻ revealed information interaction showed that the difference between the objective amount of new information revealed and the perception of the amount of new information revealed depended on the interview condition, F(1, 58) ⫽ 32.46, p ⬍ .001, d ⫽ 1.50, 95% CI [0.92, 2.07]. The interaction was analyzed further by the use of simple effects tests for each interview condition. The participants in the Scharff condition perceived to have revealed a significantly lesser amount of new information (M ⫽ 5.40; SD ⫽ 2.30) than they objectively revealed (M ⫽ 7.93; SD ⫽ 2.70), F(1, 58) ⫽ 32.15, p ⬍ .001, d ⫽ 1.49, 95% CI [0.92, 2.06]. In sharp contrast, the participants in the Direct Approach perceived to have revealed significantly more new information (M ⫽ 6.17; SD ⫽ 3.97) than they objectively revealed (M ⫽ 5.10; SD ⫽ 2.26), F(1, 58) ⫽ 5.70, p ⫽ .020, d ⫽ 0.63, 95% CI [0.11, 1.15] (see Figure 2). Thus, we found support for Hypothesis 5. For the Scharff condition a large majority of the participants (87%) underestimated the amount of new information revealed. Only a small number of participants (10%) did not over- or underestimate, and only one single participant (3%) overestimated the amount of new information revealed. Whereas for the Direct Approach, more than half of the participants (57%) overestimated the amount of new information revealed. Only a small number of participants (13%) did not over- or underestimate the amount of new information revealed, and the remaining participants (30%) underestimated the amount of new information revealed. Table 1 Means (and SD) for New Information Revealed for the Two Interview Conditions When Asking (A) the Initial Open-Ended Question, (B) the Confirmations/Specific Questions, (C) the Final Open-Ended Question, and, (D) for the Complete Interview Initial open-ended question M Direct approach Scharff-technique Note. SD a 2.90 5.53b 2.47 2.30 Confirmations/specific questions M SD a 1.40 2.10b 0.81 1.16 Different superscripts indicate that means are significantly different (p ⬍ .05). The final open-ended question M SD a 0.80 0.30b 0.85 0.53 Total new Information M SD a 5.10 7.93b 2.26 2.70 OLESZKIEWICZ, GRANHAG, AND CANCINO MONTECINOS b = 3.08, p < .001 b = 0.10, p = .313 Perceived Interviewer Knowledge Interview Condion New Informaon Revealed Direct effect, b = 1.10, p = .017 Indirect effect, b = 0.32, CI [-0.29, 1.08]. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Figure 1. Model testing if perceived interviewer knowledge mediates the relationship between interview condition and the new information revealed. Discussion The current study is one of the very first examining the comparative efficacy of human intelligence gathering techniques. In essence, we found that the Scharff-technique outperformed the Direct Approach on most of the measures used for tapping efficacy. New Information Revealed During the Interview The Scharff-technique resulted in significantly more new information compared to the Direct Approach. This finding was expected for two reasons that were assumed to work in tandem. The first reason pertained to the information management dilemma that the sources had to navigate. In brief, a source faced with the Scharff-technique—who strived to be perceived as sufficiently willing—needed to provide information beyond that which was disclosed by the interviewer who attempted to create the “illusion of knowing it all.” Information given by the sources in the Scharfftechnique is therefore likely to be new to the interviewer. In contrast, for the Direct Approach the information provided was assumed to be a mix of new and already known information. The second reason drew on observations from real life interrogations that show that sources seem to be particularly keen to disclose information that they believe the interviewer already has (Soufan, 2011; Toliver, 1997). By doing so, they can pretend to be willing sources, without revealing any new information. This counterinterrogation strategy will, however, only work if the source correctly predicts the information that the interviewer already holds. In the study sources in both interview conditions ascribed the interviewer to hold information that, in fact, the interviewer did not hold. However, since the amount of knowledge incorrectly ascribed to the interviewer did not differ across conditions, this cannot explain the superiority of the Scharff-technique. Hence, in the current study, it is likely that only our first reason accounts for the difference in the amount of new information elicited (i.e., that sources faced with the Scharff-technique had to provide information beyond the interviewer’s “illusion of knowing it all”). These results may point toward a possible unused potential of the Scharff-technique. Should future attempts prove successful in having sources incorrectly ascribe comparatively more knowledge to the interviewer by, for example, establishing a more powerful “illusion of knowing it all,” this might further add to the observed successes of the Scharff-technique. As we have data for each phase of the interview, we can offer some further information on how we arrived at the impressive difference between the two techniques. The gap between the techniques was established already after the first phase of the interview; after the initial open-ended question. Having played the three Scharff-tactics “friendly approach,” “not pressure for information,” and the “illusion of knowing it all” before leaving the word to the source, resulted in almost double the amount of new information compared to the Direct Approach. We find it reasonable to speculate that the “illusion of knowing it all” tactic was central for this outcome. For the second phase of the interview the Scharff-technique utilized claims in the form of “confirmations,” whereas the sources in the Direct Approach were faced with explicit questions. Importantly, the claims and the direct questions “asked” for the same five pieces of information. We predicted and found that the Scharfftechnique outperformed the Direct Approach also for this interview phase. Interestingly, this finding shows that a less (vs. a more) obtrusive way of interviewing can be more effective in terms of eliciting information. It should also be acknowledged that this finding emerged despite the fact that the Scharff-technique resulted in almost double the amount of new information already after the first interview phase. For the final interview phase, the Direct Approach resulted in significantly more information than the Scharff-technique. This finding was not expected but is rather easy to explain in hindsight. The outcome might simply be a reflection of the fact that the Scharff-technique had resulted in 75% more new information compared to the Direct Approach after the first two phases. Differently put, there was much more room for the sources in the Direct Approach to add new information in response to the final open question. Importantly, the amount of new information revealed during the third and final phase was comparatively small, and did not cloud the overall pattern. Furthermore, in line with our hypothesis, we found that for the Scharff-technique the level of precision reached for the new information elicited was higher than for the Direct Approach. We consider this an important finding as more precise information 9 Amount of New Information 484 8 Subj 7 Obj 6 5 4 3 2 Direct Approach Scharff-technique Interview Condition Figure 2. Illustrating the interaction effect for the subjective and objective scores of new information revealed for two interview conditions. This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. ELICITING INTELLIGENCE often equals more actionable information. Our analysis showed that the amount of new information and the degree of precision were highly correlated. This does not come as a surprise as the more information a source reveals, the higher the likelihood that the source will mention some very specific pieces of information. However, the two measures are not necessarily related as a source may reveal (a) a small, but highly precise, portion of information, or (b) a rather large amount of very general (nonspecific) information. The number of incorrect pieces of information reported during the interactions was very low (4 pieces), and all these pieces were found in the Direct Approach condition. We were not surprised by this very low number as (a) the sources had full access to the background information during the conversation, and (b) the sources were discouraged to fabricate information during the interview. The Sources’ Subjective Perceptions of the Interview In line with our prediction, the sources in the Scharff condition perceived the interviewer to have held significantly more information prior to the interview compared to the sources in the Direct Approach. It is noteworthy that about 30% of all the information that the sources in the Scharff condition believed that the interviewer knew prior to the interaction, was in fact information that was not known to the interviewer. That is, this information went beyond what the interviewer stated when trying to establish the “I already know it all” illusion. That the corresponding proportion was even larger for the sources in the Direct Approach is considered much less noteworthy; the reason for this is that these sources received no information on the interviewer’s prior knowledge and therefore had a very difficult task assessing what information the interviewer had at the outset of the interview. As 60% of the information was new to the interviewer (22 of the 35 pieces of information), it is not surprising that their assessment resulted in a rather large proportion of incorrectly categorized information (i.e., information units incorrectly scored as already known to the interviewer). The sources in the Scharff- and the Direct Approach condition did not differ with respect to their understanding of what information the interviewer was seeking to obtain. In fact, the sources in both conditions reported to have a rather good understanding of what information the interviewer was seeking to obtain. For the Direct Approach, this may not come as a surprise as the sources were asked direct questions. However, that the sources in the Scharff condition also perceived to have a good understanding of what the interviewer was after was not in line with our expectation and is a more difficult to explain. However, it should be acknowledged that a source’s perception of the interviewer’s information objectives can be more or less correct. For our study, we did not probe the sources for information on what specific information they thought the interviewer was after. Hence, we do not know whether the sources had a correct understanding of the interviewer’s information objectives. Future studies might profit from matching (a) the sources’ perception of the interviewers’ information objectives and (b) the interviewers’ actual objectives. 485 Relating Objective and Subjective Measures An effective information elicitation technique should result in new information, while leaving the source with the impression that s/he revealed less new information than s/he actually did. For the present study, the Scharff-technique fulfilled this critical criterion. Noteworthy, almost 90% of the sources in the Scharff condition underestimated the amount of new information revealed. Furthermore, as much as 30% of the new information revealed in the Scharff condition was information that was provided unknowingly by the sources. These findings are at the heart of the Scharff-technique; to elicit critical information without the source being aware that he or she is contributing with any new information. We attribute this finding to the tactics used for the Scharfftechnique, in particular to the “confirmation” tactic. In brief, confirming claims might result in sources underestimating their contribution of new information. This is further supported by the second phase of the Scharff-interview (posing claims), where a substantial amount of new information was gathered. The finding that the sources faced with the Direct Approach significantly overestimated the amount of new information that they actually revealed was expected. In brief, answering direct and explicit questions might lead to the perception that all (or almost all) information revealed is information that is new to the interviewer. The Scharff-Technique: The Basic Principles At the heart of the Scharff-technique is the counterinterrogation strategies used by the sources and the interviewer’s awareness of these strategies (Granhag, 2010; Scharff, 1950; Toliver, 1997). Importantly, the sources’ counterinterrogation strategies can be linked to the basic psychology of self-regulation theory; a social– cognitive framework for understanding how people control their behavior to steer toward desired goals and steer away from undesired outcomes (for a recent review, see Carver & Scheier, 2012). For the current context, this is mirrored in the information management dilemma the sources had to navigate; to reveal enough information to receive the help needed (the desired outcome), but to avoid revealing too much or too little information about the attack (the undesired outcome). The Scharff-technique is anchored in the following basic principles: (a) a source typically forms a hypothesis on how much and what information the interviewer already holds, (b) the source’s perception will affect his or her counterinterrogation strategies, and (c) the counterinterrogation strategies employed will affect how much and what information the source reveals. An interviewer who fully grasps the relation between the basic principles can utilize perspective-taking to anticipate the specific reactions of an individual source and better predict the case-specific outcome. The act of perspective-taking is to be placed at the metalevel, whereas the basic principles are to be found at the object-level (Nelson, 1992). Limitations Both the current study and the Scharff-technique as such come with a number of limitations. Needless to say, some aspects of a typical human intelligence interaction are very difficult to mirror This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 486 OLESZKIEWICZ, GRANHAG, AND CANCINO MONTECINOS in a laboratory setting. A first limitation is that our study is based on a student sample; a sample very different from a typical source. However, we believe it is reasonable to argue that the Scharfftactics might be more (not less) effective in real-life settings compared to lab-settings. Our argument is based on the assumption that many real-life sources will be very motivated to evoke counterinterrogation strategies (Soufan, 2011; Toliver, 1997), and much more motivated to do so than students participating in an experiment. Hence, as the Scharff-technique is tailored to counteract these counterinterrogation strategies, the technique might, in fact, be even more effective in real-life settings. Furthermore, although we assessed both the amount and the precision of the information revealed by the sources, we did not assess the relative importance of each individual theme. This might be viewed as a limitation since in real-life information on one particular theme is often more critical than information on other themes. Hence, for future studies it might be worthwhile to introduce a coding system that not only accounts for the amount and precision of the new information revealed across all themes, but also accounts for the relative importance of different themes. Third, although the Scharff-technique outperformed the Direct Approach both in terms of the amount and the precision of the new information, our findings also show that there is room for further improvement. The Scharff-technique— on average— elicited (a) about one third of all new information that was possible to obtain and (b) information that accounted for about 60% of the maximum precision score. On a different note, these results could be interpreted as indicating that the sources took on the information management dilemma in a serious manner (i.e., to navigate between not revealing too little or too much information). A fourth possible limitation is that the sources were discouraged to fabricate information (but were allowed to lie in other ways). The rationale for this was that our focal interest was to examine to what extent the techniques differed with respect to eliciting new information. As the techniques differed with respect to the number of explicit questions asked, this might have resulted in a relatively higher frequency of fabrications for the Direct Approach (relatively more explicit questions). In turn, such a result might have clouded the overall pattern. Here it should be underscored that all sources lied in the sense that they were withholding (more or less) information from the interviewer, and it is reasonable to argue that this form of lying is very common in intelligence contexts (e.g., Soufan, 2011). We believe that future studies might profit from allowing the sources to lie also in the sense of fabricating information. On a related note, although the current article provides a template for future experimental work in this emerging domain, such research would also profit from having access to a palette of different scenarios (management dilemmas) from which proper samples could be drawn. Fifth, we used rather short interviews over the phone, during which the sources had access to a piece of paper with all relevant background information. Future studies should examine the efficacy of intelligence gathering techniques where the source and the interviewer meet face-to-face, and where the source needs to memorize his or her background information. Finally, we would like to acknowledge a few limitations that pertain to the Scharff-technique as such. First, there are situations where the Scharff-technique might be difficult to use. For example, to properly paint the “I-already-know-it-all” illusion demands that the interviewer is in possession of a certain amount of accurate information. On the other hand, to build the reference system needed for this tactic is a much easier task today than it was during WWII. Here it should also be noted that there are examples where Scharff (Toliver, 1997)—and more modern interrogators like Ali Soufan (2011)— have managed to build this illusion by using only a single piece of critical information. A further limitation is that for some situations it would be a clear tactical mistake to reveal to the source how much and what intelligence the interviewer holds on a certain topic. For example, if the source is not in custody, he or she may inform the individual (or cells of individuals) to which the intelligence pertains, and/or networks planning future illegal actions could deploy false sources in order to tap how much intelligence is held on them. Third, there are many different forms of human intelligence interactions, and the Scharff-technique is primarily aimed for settings where the source is expected to be questioned. Such interactions might take place in a voluntary context (as mirrored in the current study), or in a nonvoluntary, custodial setting (similar to the context in which Hanns Scharff developed his technique). Other human intelligence interactions are characterized by a clandestine component and may occur as a seemingly causal conversation (e.g., at a bar or on a bus). We do not claim that the results of the current study can be generalized to such covert intelligence gathering. A Note on Policy Names like Guantanamo and Abu Ghraib lead to the epicenter of one of the most debated and politically charged issues within American policymaking after 9/11; the use of “advanced interrogation methods” and extraordinary renditions (e.g., The Constitution Project, 2013; Skerker, 2010; McKelvey, 2007). In an attempt to end the use of coercive interrogation methods— of which some amounted to torture—President Obama made it clear that American interrogators can only use the methods described in the Army Field Manual (Executive Order 13491). Leaving the issue of the still unbridged gap between official interrogation policies and actual interrogation practices (Brandon, 2011), we believe that the current article is policy relevant on two different levels. On the more general level, very few interrogation methods listed in the Army Field Manual have been subjected to scientific evaluation, and the current study contributes to remedy this shortcoming. Furthermore, and still on the more general level, the current article offers several measures that we argue are highly relevant when assessing the efficacy of a human intelligence gathering technique. On the more specific level, the current article offers a much fuller description of the Scharff-technique than the one found in the current version of the Army Field Manual. Finally, and most importantly, the current study shows that the Scharff-technique can be more effective for eliciting intelligence than the more commonly used Direct Approach. We believe that the Scharfftechnique has wide applications; it can be used for a number of different types of sources (e.g., prisoners of war, detainees, and informants). In brief, the current article can be used to inform policymakers on both general and specific issues with respect to interrogation practices. Conclusions In the wake of the terrorist attacks in New York, Madrid, London, and Boston the increased threat of terror worldwide, the This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. ELICITING INTELLIGENCE need for effective human intelligence techniques is more critical than ever (Brandon, 2011; Loftus, 2011). In light of this, it is nothing short of remarkable that there is little to no research that can inform on the efficacy of different human intelligence gathering techniques. The present study is one of the first attempts to remedy this serious shortcoming. At the heart of this study is the Scharff-technique; a noncoercive yet psychologically sophisticated technique, revolving around the concept of perspective-taking. The results show that the Scharff-technique outperformed the Direct Approach on a number of measures at the very core of an effective information elicitation technique. Specifically, we found that the Scharff-technique was superior both in terms of the breadth (amount) and depth (precision) of the new information elicited, and that the absolute majority of sources faced with the Scharfftechnique underestimated how much new information they actually revealed during the interview. We argue that the combined evidence speak to the Scharff-technique as a promising tool for eliciting intelligence. References Brandon, S. E. (2011). Impacts of psychological science on national security agencies post-9/11. American Psychologist, 66, 495–506. doi: 10.1037/a0024818 Carver, C. S., & Scheier, M. F. (2012). A model of behavioral selfregulation. In P. A. M. Van Lange, A. W. Kruglanski, & E. T. Higgins (Eds.), Handbook of theories of social psychology (Vol. 1, pp. 505–525). Thousand Oaks, CA: Sage. Clemens, F., Granhag, P. A., & Strömwall, L. A. (2011). Eliciting cues to false intent: A new application of strategic interviewing. Law and Human Behavior, 35, 512–522. doi:10.1007/s10979-010-9258-9 Dujmovic, N. (2005). Fifty years of studies in Intelligence: Building an ‘Intelligence literature’. Studies in Intelligence, 49, 6 –13. Evans, J. R., Meissner, C. A., Brandon, S. E., Russano, M. B., & Kleinman, S. M. (2010). Criminal versus HUMINT interrogations: The importance of psychological science to improving interrogative practice. Journal of Psychiatry & Law, 38, 215–249. Exec. Order No. 13491, Ensuring Lawful Interrogations, 3 C.F.R. page 4894 (2009). Galinsky, A. D., Maddux, W. W., Gilin, D., & White, J. B. (2008). Why it pays to get inside the head of your opponent: The differential effects of perspective taking and empathy in negotiations. Psychological Science, 19, 378 –384. doi:10.1111/j.1467-9280.2008.02096.x Goldsmith, M., Koriat, A., & Weinberg-Eliezer, A. (2002). Strategic regulation of grain size in memory reporting. Journal of Experimental Psychology: General, 131, 73–95. doi:10.1037/0096-3445.131.1.73 Granhag, P. A. (2010). The Scharff-technique: Background and first scientific testing. Professional Development Seminar, High-value Detainee Interrogation Group (HIG, FBI). Washington, DC. Granhag, P. A., Cancino Montecinos, S., & Oleszkiewicz, S. (2013). Eliciting intelligence from sources: The first scientific test of the Scharff-technique. Legal and Criminological Psychology. Advance online publication. doi:10.1111/lcrp.12015 487 Granhag, P. A., & Hartwig, M. (2008). A new theoretical perspective on deception detection: On the psychology of instrumental mind-reading. Psychology, Crime & Law, 14, 189 –200. doi:10.1080/ 10683160701645181 Granhag, P. A., & Hartwig, M. (in press). The Strategic Use of Evidence (SUE) technique: A conceptual overview. In P. A. Granhag, A. Vrij, & B. Verschuere (Eds.), Deception detection: current challenges and cognitive approaches. Chichester, England: Wiley. Granhag, P. A., Mac Giolla, E., Strömwall, L. A., & Rangmar, J. (2012). Counter-interrogation strategies among small cells of suspects. Psychiatry, Psychology and Law, 20, 1– 8. doi:10.1080/13218719.2012.729021 Granhag, P. A., Strömwall, L. A., Willén, R. M., & Hartwig, M. (2013). Eliciting cues to deception by tactical disclosure of evidence: The first test of the Evidence Framing Matrix. Legal and Criminological Psychology, 18, 341–355. doi:10.1111/j.2044-8333.2012.02047.x Justice, B. P., Bhatt, S., Brandon, S. E., & Kleinman, S. M. (2010). Army Field manual 2–22.3 interrogation methods: a science-based review. Unpublished manuscript. Loftus, E. F. (2011). Intelligence Gathering Post-9/11. American Psychologist, 66, 532–541. doi:10.1037/a0024614 McKelvey, T. (2007). Monstering. Inside America’s policy of secret interrogations and torture in the terror war. New York, NY: Carrol & Graf Publishers. Nelson, T. O. (1992). Metacognition: Core readings. Boston, MA: Allyn & Bacon. Redlich, A. D. (2007). Military versus police interrogations: Similarities and differences. Peace and Conflict: Journal of Peace Psychology, 13, 423– 428. doi:10.1080/10781910701665741 Redlich, A. D., Kelly, C., & Miller, J. (2011). Systematic survey of the interview and intelligence community. Final report submitted to the Federal Bureau of Investigation/High-Value Detainee Interrogation Group, Washington, DC. Scharff, H. J. (1950). Without torture. Argosy, 39, 87–91. Shoemaker, D. P. (2008). Unveiling Charlie: U.S. Interrogator’s creative success against insurgents. In National Defense Intelligence College. (Ed.), Interrogation: World War II, Vietnam and Iraq (pp. 77–146). Washington, DC: NDIC Press. Skerker, M. (2010). An ethics of interrogation. Chicago, IL: University of Chicago Press. doi:10.7208/chicago/9780226761633.001.0001 Soufan, A. H. (2011). The black banners: The inside story of 9/11 and the war against Al-Qaeda. New York, NY: Norton. The Constitution Project, Task Force on Detainee Treatment. (2013). Report of The Constitution Project’s Task Force on Detainee Treatment. Retrieved from http://detaineetaskforce.org/report/download/ Toliver, R. F. (1997). The interrogator; the story of Hanns Joachim Scharff master interrogator of the Luftwaffe. Atglen, PA: Schiffer Publishing Ltd. United States Army. (2006). Human intelligence collector operations, Field manual 2–22.3 (FM 34 –52). Washington, DC: Headquarters, Department of the Army. Varouhakis, M. (2013). What is being published in intelligence? A study of two scholarly journals. International Journal of Intelligence and Counterintelligence, 26, 176 –189. doi:10.1080/08850607.2012.705564 Vrij, A. (2008). Detecting lies and deceit: Pitfalls and opportunities (2nd ed.). Chichester, England: Wiley. (Appendices follow) 488 OLESZKIEWICZ, GRANHAG, AND CANCINO MONTECINOS Appendix A Instructions to the Source This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. Background Imagine that economic problems, not caused by yourself, made you participate in the robbery of a cash transport van in the fall of 2007. The actual robbery went fine, but three months ago, the other three involved in the act got arrested. The only one who is still free is you, but you feel that this is only a matter of time. You know where most of the stash (approximately 45 million SEK) is kept. You understand that your time is scarce, and you immediately need to get the stash and move yourself and your money out of the country. Some time ago you got an idea of how it could all be solved, and briefly your plan is as follows: Through a close friend you have come by information that a radical political group in Sweden has future plans to perform a bomb attack in Gothenburg, around Christmas, 2011. Your plan is to reveal information about this bomb attack to the special police force (SÄPO), and in favor of the information receive free conduct out of Sweden. Ten days ago you contacted the special police (anonymously of course) and carefully asked if there was any interest in talking further about this matter. SÄPO said that they were very interested in talking more thoroughly with you, and it is this call you are now about to make. The group that is planning this bomb attack is called MDA16, and consists of a loosely assembled network of approximately 10 members. You are a close friend to one of those members, and you feel some sympathy for the group’s opinions. After a lot of consideration, you have decided to reveal some pieces of information about the planned bomb attack to the police. You do understand that it is possible that the police already have some information about the planning—partly because SÄPO have conducted their own investigations, and partly because you have heard, from your friend, that a few of the members in MDA16 suspects that their phones have been tapped (but this is nothing they know for certain). In brief, you don’t know what the police actually know about the planned attack (or if they know anything at all). But before the phone call, you have a very important additional dilemma to reflect upon: When speaking to the police you should absolutely not tell them everything you know. First of all you have, to say the least, a negative attitude toward the police. Also, if you would reveal everything you know about the planning, it could jeopardize the entire existence of MDA16, including your close friend, and might get them convicted for planning a very serious crime. If you tell too much, there is also an obvious risk that they will find out that it was you who “sold them out,” which means that you will be hunted by the entire group (and you are not prepared to go that far). On the other hand, you cannot reveal too little, because if you do so, there is a risk that the police won’t find your contribution to be significant enough to grant you free conduct out of Sweden. You also realize that you must avoid lying to the police because if they find out that you are lying, they might come to believe that you are trying to trick them (that is, to receive free conduct out of Sweden by revealing information that will deceive the police). In order to be taken seriously, and appear trustworthy, you have to show some degree of good will and cooperation. In sum, you need to find a good balance—neither revealing too much or too little information. In spite of all the effort you have put into thinking this through, you still feel very hesitant about talking to the police at all, but nonetheless you have decided to give it a try. However, you have not fully decided what specific information (and how much) you will reveal to the police. This decision is partly held open, and you will in some degree allow the development of the upcoming conversation to direct this matter. What you know about the planning of the upcoming attack is as follows: General You know that the group planning the attack is called MDA16, it consists of approximately 10 members and is located in Gothenburg. You also know that the group has been around since 2002 and came to existence as a result of the EU riots in Gothenburg 2001. You know that the group, in cooperation with two Danes, had plans to execute a bomb attack during 2006 against a conference center in Malmö where a political top meeting was held at the time. But that operation was cancelled due to internal conflicts. This conflict resulted in one of the leading figures of the group, Jari Tapio, leaving MDA16. Your Relationship to MDA16 Petter Jönsson, who is your close friend, and Jari Tapio founded MDA16. You know the names of most of the members of MDA16: Martin, Johannes, Erik, Sara, Pär, Sigge, Lisa, but have no further personal information about them. You know the background of the internal conflict that occurred in Malmö. In brief, Jari Tapio wanted to increase the effect of the attack with human casualties, something the Danes refused to go along with. Since the other members sided with the Danes, this dispute led to Jari leaving MDA16. Jari and Petter are currently bitter enemies, as it was Petter who introduced the Danes to MDA16. (Appendices continue) ELICITING INTELLIGENCE Specific Details About the Upcoming Attack This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. You know that five persons are working more specifically with the planning of the attack. Among these five there are two Danes (a male and a female) who are both experts on explosives. You also know that these two Danish bomb experts participated in the planning of the bomb attack that would have been performed in Malmö (2006), which was cancelled. You know that the shopping mall subjected for the planned attack is Femmanhuset in Nordstan, and you know that the attack will take place during the Christmas holiday sales, namely the 27th of December (2011). 489 You also know that the plan is to plant the bomb during daytime, and that the bomb will be detonated at 11PM via an advanced remote detonator. The bomb will be placed in an old-fashioned TV, which will be brought for repairs at 5.55PM the 27/12. That is, five minutes before closing time. The store, Elektronik Experten, where the TV will be repaired is centrally located in the mall’s basement. You do not know what kind of bomb it is. You do not know where the bomb is located at the moment (or if it is manufactured yet). Appendix B Checklist for the 35 Units of Information Which specific pieces of information did you reveal to your police contact? Mark the alternatives that describe the information you told the police during the conversation The group Previous planning Current planning Location Date Where the bomb is planted When the bomb is planted How the bomb is planted When the bomb is detonated How the bomb is detonated ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ ▫ Called MDA16 10 members People from Gothenburg Founded in 2002 Founded after the EU-riots 2001 Have planned a previous bomb attack Planned a bomb attack in Malmö The previous bomb attack was cancelled Was cancelled due to an internal conflict Some people left the group after the conflict Jari Tapio left the group after the conflict 5 people are planning the current attack 2 persons are Danish There are bomb experts The Danes are the bomb experts Centrum Nordstan Femmanhuset Around Christmas After Christmas 27th of December In the basement In a store In an electronics store The store Elektronik Experten During daytime Around closing time 5:55PM Placed in some kind of apparatus Placed in a TV Apparatus/TV brought for repairs During the evening After closing time Around 11PM Advanced remote detonator Received March 27, 2013 Revision received January 16, 2014 Accepted January 18, 2014 䡲