Jean J. Cabell, University of Nevada, Reno
Sarah A. Moody, University of Nevada, Reno
Yueran Yang, University of Nevada, Reno

This article proposes an effect taxonomy to organize and categorize the interrogation techniques that police often use to elicit confessions from suspects during custodial interrogations. The effect taxonomy categorizes techniques into one of four categories according to how interrogation techniques differently influence guilty and innocent suspects: confession-prone, differentiating, assimilating, and nonconfession-prone. To apply the effect taxonomy, we employ the interrogation decision-making model (Yang, Guyll, & Madon, 2017) to analyze the 16 interrogation techniques used in Kassin and colleagues’ (2007) national survey on police interrogation practices. Our analysis revealed that the majority of the interrogation techniques are confession-prone, thus increasing the likelihood of confessing for both guilty and innocent suspects. Our analysis also revealed several differentiating techniques, which increase the likelihood of confessing only for guilty suspects but not for innocent suspects. In addition, we also identified the interrogation duration as a potentially differentiating factor for guilty and innocent suspects, such that limiting the length of an interrogation will protect innocent suspects from falsely confessing. The findings from the effect taxonomy have important implications for the legal system: Police should be aware of the effects of the interrogation techniques to avoid using confession-prone interrogation techniques and employ differentiating techniques that increase the likelihood of confessing for guilty suspects but not innocent suspects.

police interrogation, interrogation techniques, confessions, decision-making, taxonomy

“Confessions are powerful sources of incriminating evidence, so powerful that “the introduction of a confession makes the other aspects of a trial in court superfluous”. To elicit confessions, police investigators typically bring suspects into an isolated interrogation room and use multiple interrogation techniques. These techniques, however, might not only lead guilty suspects to confess but might also persuade innocent suspects to falsely confess. […] The frequency of false confessions emphasizes a need for the legal system to evaluate the effects of interrogation techniques, particularly the effects of techniques on both guilty and innocent suspects.” (p. 154)

“Based on the idea of evaluating an interrogation technique’s effects on both guilty and innocent suspects’ decisions to confess, we propose a new taxonomy of interrogation techniques: an effect taxonomy. First, we review previous research findings on interrogation techniques and confessions. Second, we briefly review previous taxonomies of interrogation techniques. Third, we propose a new taxonomy of interrogation techniques—the effect taxonomy—and provide examples of how to employ the interrogation decision-making model to categorize interrogation techniques into the effect taxonomy. Depending on whether an interrogation technique increases the likelihood of confession for guilty and innocent suspects, we categorize it into one of four categories: confession-prone, differentiating, assimilating, and nonconfession-prone. Lastly, we discuss the strengths, implications, limitations, and future directions of the effect taxonomy.” (p. 154)

“The Reid technique, which is used in a majority of large American police departments (i.e., serving cities with populations over 150,000), has been criticized for its association with false confessions. […] Although the authors of the Reid technique claim these techniques only elicit confessions from guilty suspects and not from innocent suspects, researchers have expressed concern that this set of techniques increases the likelihood of false confessions by implying leniency or exaggerating consequence.” (p. 155)

“Taxonomies have not only facilitated theory development across scientific disciplines (e.g., classifications in biology), but have also created organized frameworks for measuring human behavior, facilitating comparisons, and communicating results through a common infrastructure. Taxonomies of interrogation techniques, similarly, can help researchers and policymakers achieve a systematic understanding of interrogation processes and improve decision-making.” (p. 155)

“So far, three major taxonomies have been proposed to categorize interrogation techniques. Minimization versus maximization taxonomy. One of the first attempts to categorize police interrogation techniques used the dichotomy of minimization and maximization. […] Information-gathering versus accusatorial taxonomy. Another dichotomous taxonomy categorizes interrogation techniques as information-gathering or accusatorial techniques. […] Six-domain taxonomy. […] To address the lack of midlevel categorizations, Kelly and colleagues (2013) identified and categorized 71 interrogation techniques into six domains according to the conceptual goals of the techniques: (a) rapport and relationship building […] (b) context manipulation […] (c) emotion provocation […] (d) confrontation and competition […] (e) collaboration, […], and (f) presentation of evidence.” (p. 155)

“The taxonomies described above are useful tools for the legal system and researchers to categorize and understand existing techniques. One shortcoming of these taxonomies, however, is that they do not capture the effects of interrogation techniques on suspects’ confession decisions. […] In addition, the previously described taxonomies are generally descriptive rather than normative.” (pp. 155–156)

“A 2 X 2 contingency table can display the possible outcomes when considering whether an interrogation technique can increase the likelihood of confessing for guilty or innocent suspects […] The four proposed categories [confession-prone, differentiating, assimilating, nonconfession-prone] are solely focused on suspects’ decisions to confess, not the qualities of a confession.” (p. 156)

“Confession-prone techniques. If an interrogation technique increases the likelihood of confessing for both guilty and innocent suspects, we define such a technique as confession-prone. Using such techniques can increase the likelihood that that police will accomplish their goal of incriminating the guilty suspect. However, using confession-prone techniques can also increase the likelihood that police will incriminate an innocent suspect.” (p. 156)

“Differentiating techniques. If an interrogation technique increases the likelihood of confessing for guilty suspects but not for innocent suspects, we consider the technique to create a differentiating effect. In other words, these techniques differentiate between the decisions of guilty and innocent suspects. Differentiating techniques would be the most ideal type for police to use when trying to elicit confessions from suspects because these techniques’ effects are consistent with the police goal of incriminating guilty suspects but not innocent suspects.” (p. 156)

“Assimilating techniques. If an interrogation technique increases the likelihood of confessing for innocent suspects but not for guilty suspects, we consider the technique to create an assimilating effect. This category of techniques would be the least ideal type for police to use because these techniques not only increase the likelihood of incriminating innocent suspects, but also decrease the likelihood of incriminating guilty suspects.” (p. 156)

“Nonconfession-prone techniques. It is possible that an interrogation technique does not increase the likelihood of confessing for either guilty or innocent suspects. […] These techniques typically serve to initiate an interrogation, facilitate interactions and communications, and elicit information from suspects. Although not directly increasing the likelihood of confession, these techniques could possibly moderate the effects of other types of interrogation techniques on suspects’ confession decisions.” (p. 156)

“By separately examining the effects of an interrogation technique on guilty and innocent suspects, the effect taxonomy provides a useful tool for researchers and the legal system to evaluate and understand the effects of interrogation techniques. To illustrate the use of the effect taxonomy, we analyze and classify 16 interrogation techniques (Kassin et al., 2007) into the four categories identified in the effect taxonomy in the following section. Specifically, we analyze how the techniques influence suspects’ decisions to confess using the interrogation decision-making model.” (p. 156)

“The interrogation decision-making model provides a normative analytic framework to understand and examine suspects’ decision-making processes during custodial interrogations. Drawing upon expected utility theory, this model proposes that suspects decide to confess or deny guilt by evaluating and comparing the expected utilities of their choices. To evaluate the expected utilities, suspects weigh the immediate and future outcomes of their choices by the utilities and probabilities of the outcomes, with a tendency to discount future outcomes. […] Researchers and the legal system can use this theoretical framework to analyze how specific interrogation techniques influence guilty and innocent suspects’ decision-making, thereby categorizing them into the effect taxonomy.” (pp. 156–157)

Confession-prone techniques from Kassin and colleagues (2007): isolating suspect, identifying contradictions, appealing to self-interest, offering excuses, interrupting denials, bluff, minimizing seriousness, expressing negative emotions, consequence threatening, false-evidence ploy, physical intimidation (p. 158)

Differentiating techniques from Kassin and colleagues (2007): building rapport, presenting guilt evidence, conscience appeals, showing photos (p. 158)

Assimilating techniques from Kassin and colleagues (2007): none (p. 158)

Nonconfession-prone techniques from Kassin and colleagues (2007): manipulating room size (p. 158)

“Police interrogators seldom use only one type of interrogation technique over the course of an interrogation. […] Correspondingly, the decision-making process of interrogated suspects is dynamic rather than static. Suspects do not make one single decision; they make a series of decisions, each of which reflects their continuous reevaluation of the situation. […] As interrogators use more interrogation techniques, the effects of these techniques are expected to accumulate, progressively reducing suspects’ evaluation of the utility of a denial and increasing their evaluation of the utility of a confession. Flipping from a denial to a confession, therefore, becomes more likely as the duration of the interrogation increases.” (p. 161)

“Based on the cumulative effects of interrogation techniques, we want to point out one more differentiating factor beyond the previously described interrogation techniques: the length of an interrogation. Guilty and innocent suspects initiate their decision-making at very different points during an interrogation.” (p. 162)

“Compared with innocent suspects, for example, guilty suspects might anticipate experiencing cognitive dissonance for the act of denying guilt, leading to negative emotions such as stress, anxiety, guilt, or remorse. Guilty suspects might also perceive future punishment as highly likely even if they deny guilt, presumably because future evidence will be collected to confirm their guilt and thus ensure a conviction. Moreover, guilty suspects might anticipate experiencing catharsis for the act of admitting guilt, whereas innocent suspects might anticipate experiencing cognitive dissonance for the same act because of the damage to their innocence and integrity.” (p. 162)

“Because of these preexisting differences, the initial discrepancy in suspects’ evaluations between a denial and a confession could be much larger for innocent suspects than for guilty suspects. […] The following two inferences can be derived from this initial discrepancy in suspects’ decision-making processes. First, guilty suspects would be more likely than innocent suspects to confess during an interrogation. […] Second, it would take innocent suspects longer than guilty suspects to reach their “breaking point”—the point at which suspects decide that the expected utility of a confession is large enough to flip from a denial to a confession. […] If an interrogation continues long enough, both guilty and innocent suspects might confess. However, by controlling the length of the interrogation, police may still be able to elicit confessions mostly from guilty but not innocent suspects” (p. 162)

“Our findings suggest that an interrogation is typically filled with confession-prone techniques that urge both guilty and innocent suspects to confess, which is contrary to the claim that modern interrogation techniques work only against guilty suspects but not the innocent. Interrogations are confession-prone in nature, for both guilty and innocent suspects. In addition to these interrogation techniques, our analysis also revealed the interrogation length as a differentiating factor, even if police have to rely mostly on confession-prone techniques during an interrogation.” (p. 162)

“Although most interrogation techniques are confession-prone, it is noteworthy that police do tend to use differentiating techniques relatively frequently. Police should be aware of the effects of the interrogation techniques they use, avoid using interrogation techniques that are confession-prone, and employ interrogation techniques that differentiate the confession decisions between guilty and innocent suspects.” (p. 162)

“Because of its normative function, the effect taxonomy has several important policy implications for the legal system. First, the effect taxonomy highlights the direction for better police interrogation practices, which is to use differentiating techniques but not confession-prone techniques. Indeed, our analysis reveal a gap between the current police practices and the ideal practice because the majority of current interrogation techniques are confession-prone, which encourage both guilty and innocent suspects to confess. To protect innocent suspects from falsely confessing, therefore, the legal system should identify and develop more differentiating techniques rather than relying on confession-prone techniques.” (p. 162)

“Second, the effect taxonomy illustrates the importance of imposing time limits for interrogations. Even if police interrogators have to rely mostly on confession-prone techniques, our analysis suggests that interrogations with time limits might still be able to differentiate the responses between guilty and innocent suspects. Alternatively, if an interrogation lasts too long, it might lose its differentiating power. Eventually all suspects might confess, no matter if they are guilty or innocent. Therefore, police interrogators should be aware of the increased risk of false confessions over time.” (p. 163)

“Third, the effect taxonomy is useful to analyze the techniques used in actual interrogations, particularly when video records are available. For example, if police watch an extraordinarily long video-recorded interrogation and identify the use of primarily confession-prone techniques, they can conclude the confession was likely coercive. Such scientific-based analyses resonate with the goal of the video-recording interrogation reform; it not only helps judges and juries identify potentially coercive confessions but also could help the legal system create training protocols for novice police officers and provide evaluation criteria for experienced police interrogators.” (p. 163)

“Several limitations of the effect taxonomy and our analysis warrant discussion. First, some of the interrogation techniques might be complex and ambiguous to categorize. […] Alternatively, the effects of some interrogation techniques could also depend on the personality of suspects.” (p. 163)

“A second limitation is that the effect taxonomy focuses on the directions of the effects but not so much on the magnitudes of the effects. […] Third, some of our analysis relied solely on theoretical speculation due to a lack of empirical data. […] Finally, our analysis assumes a level of rationality for decision-making, which might not hold true for certain populations.” (p. 163)

“As revealed in our analysis, interrogation research is lacking empirical data examining the effects of many interrogation techniques. […] Empirical research with multiple interrogation techniques will permit an understanding of potential interaction effects when police use multiple interrogation techniques in conjunction.” (p. 163)

“Our analysis, which serves primarily as an example of applying the effect taxonomy, only examined a small set of techniques. Future research could apply the effect taxonomy to a broader set of interrogation techniques. […] The effect taxonomy could also be applied to evaluate the effects of the techniques in the PEACE model.” (pp. 163–164)

As always, please join the discussion below if you have thoughts or comments to add!