Careful observational coding of LE interrogations with 75 convicted terrorism suspects (the largest international corpus of field data of its kind) revealed that detainee engagement (and disengagement) was impacted by interviewer behavior. In legal terms, this broadly relates to the concept of detainees exercising their right to silence, while also respecting, where they wish to, their right to talk. Although results reveal that MI skills encourage engagement, even more pronounced was the finding that any approach antithetical to MI (accusation, assumption and confrontation) had a profoundly negative impact on detainee engagement. By accounting for individual differences between detainees, the results reveal that these effects are found regardless of how cooperative (or not) a detainee may be generally. This suggests that detainees previously willing to talk are likely to be discouraged from doing so by an interviewer working too hard to convince them to talk (e.g., through accusatory statements, prejudging their answers, confronting them too forcefully with evidence, or rational cornering). These approaches appear to cause reactance and thus encourage detainees to avail themselves of the right to silence, whereas previously, they may have been considering their right to speak.
In previous studies of MI, clients’ contemplation of behavioral change (indicated by articulating their reason for change) has been the main predictor of a successful outcome (i.e., reducing substance misuse; Apodaca & Longabaugh, 2009). In this sample of HVDs, contemplating engagement (‘Should I talk or not talk to this interviewer?’) determined whether they would reveal information. The results suggest that many of the HVDs do contemplate engagement rather than the often-held legal assumption that it is always in their best interest to say nothing. Additionally, the results showed the use of MI skills and commitment to creating an accepting, empathic atmosphere was associated with increased engagement. Interviewing officers and attorneys or solicitors need to consistently bear in mind that it is not their job to convince individuals either way what is in their best interest. As enshrined in U.K. and U.S. law, it is detainees’ choice whether to avail themselves of their right to silence or their right to speak. Even gentle persuasion on the part of the interviewing officer to speak could push individuals into a decision to not speak. Equally, expert legal representatives recognize that their advice is just that—advice—and not an instruction to not speak.
Assumptive questioning, judgmental summaries, and accusatory challenges (behaviors inconsistent with MI) caused detainees to disengage and stop talking. This supports MI research that shows how therapist MIIC behaviors are associated with higher levels of resistance, lower client engagement, and worse outcomes (Apodaca & Longabaugh, 2009). It also supports investigative interviewing research that has found accusatory, confrontational approaches (e.g., disallowing denials and asserting authority) are associated with increased resistance and decreased cooperation (Kelly, Miller, & Redlich, 2016). The current study did not examine whether particular interrogator and/or detainee characteristics predicted the use of MIIC behaviors, but this may be an interesting area to explore.
In therapeutic settings, the use of MIIC behaviors can lead to particularly negative outcomes (i.e., increased substance misuse) with angry and/or highly reactive patients (Karno & Longabaugh, 2004, 2005). Reactance theory posits that a person’s reaction to a loss of freedom will be greater the more important it is perceived to be and when several freedoms are threatened. Because all detainees in the sample had been arrested, were being held in police custody at the point of interview, and were having to be interviewed, it is likely that many of them were experiencing reactance in response to their loss of freedoms. Consequently, MIIC approaches directed at pressurizing or persuading detainees to talk may have increased detainee reactance further and thus reinforced their motivation to resist (their way of regaining some freedom of choice). For ambivalent detainees, such approaches may have removed any doubt they had about whether to cooperate, strengthening their resolve not to speak. Because a controlling interviewer behavior can arouse reactance, officers must work hard to avoid such techniques and abstain from language that builds the illusion of limiting choice (Place & Meloy, 2018). This may prove a difficult task for interrogators if their well-established interviewing style incorporates MIIC behaviors such as controlling or accusatory language. Research in the therapeutic literature has shown that although therapists from other counseling backgrounds can successfully learn MI, it is considerably harder for them to stop using MIIC behaviors (e.g., directing, persuading, confronting; W. R. Miller & Mount, 2001). As such, it is suggested that interrogator training should first and foremost focus on identifying and removing MIIC behaviors from interrogators’ repertoires.
In contrast, the use of MIC skills were associated with increased detainee engagement. Specifically, reflective listening, balanced summaries, and rolling with resistance contributed to creating a nonjudgmental, supportive atmosphere in which interviewers genuinely sought to understand detainees’ perspectives and respected their right to choose to talk or not. It was in the presence of this atmosphere of communication that detainees chose to engage with the interviewers and provide information. This supports research that suggests that MI’s success can be attributed to its macrolevel approach—known as MI spirit—based on collaboration between therapists and clients, honoring clients’ autonomy, and evoking clients’ own motivations for change (Copeland, McNamara, Kelson, & Simpson, 2015; W. R. Miller & Rollnick, 2013). It also supports the view that evoking MI spirit is aided by employing the use of specific skills (Resnicow & McMaster, 2012). In mastering these, interviewers remained neutral and open to hearing detainees’ versions of events, rather than prejudging them and assuming a preformed version of events. This created an atmosphere conducive to communication and encouraged engagement without external pressure. In doing so, interviewers adhered to their central goal of collecting information neutrally and as a search for truth.
It should be noted that the IRR scores for two of the MIC scales (summaries and rolling with resistance) achieved lower scores at the interval level, which could have influenced the results. However, categorical coding (i.e., presence or absence) of these variables achieved fair agreement and the rater percentage agreements were high (both above 70%). Future research should be mindful that it may be more difficult to apply subtler scaling to these variables.
It is interesting that interviewer use of developing discrepancies appeared to operate differently from the use of other MIC skills. In MI, developing discrepancies between clients’ values and current behavior is vital in encouraging behavioral change (Westra & Aviram, 2013). However, within a LE context and the current sample of individuals convicted for terrorism offenses (i.e., eventually deemed guilty in a court of law), discrepancies existed mostly between the detainees’ accounts and evidence. Hence, even where interviewers developed discrepancies in a nonjudgmental, neutral manner, these individuals were more likely to choose to disengage. The same neutral challenges may have had a very different effect on innocent detainees, because they may have perceived the exact same option to tell the truth and explain the inconsistency as an opportunity to engage and clarify rather than, as here, shut down. In subsequent research we intend to test this hypothesis.
Multilevel analysis allowed us to understand how much variance in each variable was accounted for by individual differences between detainees (Level 2) and differences between interviews with the same detainee (Level 1). The results revealed that, for detainee engagement, yield, Global MI, and MIC scores, variance was fairly equally spread between both levels. This suggests that although there are individual differences in detainees that contribute substantial variance to these measures, what happens across interviews is equally important. It is interesting that variance in interviewer MIIC behavior was attributed predominantly to differences between interviews rather than to differences between detainees. This suggests that interviewers exhibit these behaviors across interviews with a range of detainees. Future research should explore what factors within an interview are associated with interviewers displaying MIIC behaviors, in order to stop them doing so. Additionally, future research could explore other Level 2 predictors that may explain some of the variance between detainees (e.g., the terrorist organization or group that detainees are affiliated with and thus, how much [if any] counterinterrogation training they have received; Alison et al., 2014).
This study supports the use of neutral, objective, and compassionate approaches to interviewing detainees. However, there are many contexts around the world in which detainees (especially HVDs) are still held and treated inhumanely. By providing empirical support for an ethical, objective, and compassionate approach, we hope to encourage interviewers around the world to move away from coercive practices. Regarding the use of MI in this context, we must reemphasize that the essence of MI lies in its macrolevel approach (i.e., spirit) based on an egalitarian relationship between interviewer and client or detainee and honoring client or detainee autonomy (W. R. Miller & Rollnick, 2013). It is not just a set of tactics or tricks to be used on someone (Arkowitz & Miller, 2008). MI strategies require a context in which detainees’ rights and autonomy are respected. Thus, we suggest that using MI techniques alongside coercion, persuasion, or manipulation (i.e., inconsistent with the ethos of MI) is ethically dubious and is not in the spirit of MI (i.e., as soon as such influence tricks are used alongside it, it ceases to be MI).
Although this study was based on a sample of HVDs, we predict that similar results would be found with other suspect populations. Crucially, our results reveal the detrimental use of behaviors counter to the ethos of MI (e.g., pressurizing, confronting, and judging) and that these increase resistance and reduce engagement. Hence, we echo Alison et al.’s (2013) assertion that, although it may not always be possible to engage a highly resistant detainee, using accusatory, pressurizing techniques always makes things worse. In highlighting this, we hope to encourage LE interviewers to first and foremost eliminate techniques that disengage detainees (i.e., remove behaviors antithetical to MI) and thereafter seek to adopt a set of behaviors that are more positively inclined to generate and display objectivity, compassion, and empathy.
