Abstract
Tunnel vision is the tendency of actors in the criminal justice system to use short-cuts to filter evidence selectively to build a case for a suspect’s conviction. Confirmation bias tends to favor information that confirms an individual’s preconceptions independently of the information’s accuracy: the present study manipulated tunnel vision and confirmation bias. The purpose was to examine the combined effects of these biases on Israeli police investigators and laypeople in hypothetical criminal investigation situations. Results indicated that police investigators were more strongly affected than laypersons by tunnel vision. Police investigators were more confident about the suspect’s guilt than were laypeople in the presence of incriminating and exonerating information, alike. Still, investigators did not ignore exonerating information and lowered their confidence in the suspect’s guilt when exonerating information was presented. We discussed the results according to police legitimacy and self-assessed lie-detection ability, which is rated higher by police investigators than by laypeople.
Legitimacy is a core notion in police work. Legitimacy is conceptualized as the belief that a person or institution receives authority over citizens to act honestly and according to the citizen’s best interests (Tyler, 2006). Tankebe (2013) suggested four related factors that constitute police legitimacy: Distributive fairness or the perception that the authority makes fair decisions (e.g., continue questioning only those who should be questioned and avoiding disproportional questioning of minorities or poor people). Procedural fairness concerns the decision-making process, whether errors are corrected, and whether suspects are treated with respect. Effectiveness refers to obtaining accurate results about different types of crimes. Lawfulness is whether the authority exercise power according to preexisting law. Individuals comply with the law and view power as justified when they view legal authorities as legitimate. Tyler and Jackson (2014) argue that legitimacy encourages people to report a crime.
The truth bias or the tendency of people to judge messages as truthful rather than deceptive (Bond & Depaulo, 2006; Vrij, 2008) should support police legitimacy. Nevertheless, the truth bias often counters the police’s mission to keep law and order. Instead, we find a guilt-presumptive process in criminal interrogation that triggers corrupt evidence gathering and biased forensic analysis. Guilt-presumption might lead to false confessions and safeguard wrongful convictions (Scherr et al., 2020). The pseudodiagnosticity bias (Doherty et al., 1979) or the failure to identify and select diagnostically relevant information may be mentioned here. We suggest that presumption of guilt involves a tendency to select and consider data that refers only to the guilt hypothesis of the police investigator. Presumption of guilt may therefore undermine police legitimacy in the eyes of the public. To reduce the presumption of guilt attempts were made to improve police performance to detect when the suspect is telling the truth (e.g., Dando & Bull, 2011; Dando & Ormerod, 2017; Sandham et al., 2020).
We designed the present study to explore two biases related to the presumption of guilt: tunnel vision and confirmation bias. Tunnel vision in the criminal justice system has been suggested to contribute to the wrongful conviction problem. Tunnel vision (Findley & Scott, 2006; Martin, 2002) is the tendency of actors in the criminal justice system to use heuristics and short-cuts to filter evidence selectively to build a case for a suspect’s conviction while ignoring or suppressing evidence that points to the suspect’s innocence. Tunnel vision motivates investigators, prosecutors, and judges, to attribute importance and relevance to information that supports their original conclusion while information inconsistent with the presumption of guilt is easily overlooked, dismissed as irrelevant, or considered unreliable (Findley, 2012). Tunnel vision was identified in the criminal justice system as an explanation for unintentional but endured pursuit to convict innocent suspects (Dioso-Villa, 2015; Findley & Scott, 2006). Tunnel vision was clearly demonstrated in a US murder case where detectives, in the investigation stage, made a series of recorded telephone calls with a prime suspect. The impression from the conversation was that detectives had already decided on the suspect’s guilt and pursued an accusatorial agenda rather than a fact-finding one. Specifically, detectives exhibited no interest in the suspect’s denials and worked toward eliciting a confession. The suspect was later convicted and then released from prison with the help of the Wisconsin Innocence Project (Maynard & Schelly, 2017).
Tunnel vision is associated with police investigators’ tendency to presume guilt, which increases with training and experience (Meissner & Kassin, 2002). Police interrogators will often decide to pursue a suspect’s interrogation due to a preliminary presumption of guilt in which they have confidence (Masip & Herrero, 2017). Nevertheless, police presumption of guilt may vary across jurisdictions.
The decision to pursue interrogation may rely on several aspects of deception judgments, such as the plausibility of the testimony. Hartwig and Bond (2011) have noted that “it seems that people are likely to judge communications as deceptive if they provide implausible, illogical accounts with few details” (p. 654). Aamodt and Custer (2006) similarly noted that in real-world situations, judgments about deception are often made on such factors as the story not making logical sense. Tunnel vision increases reliance on factors such as illogical statements.
A recent study (Elaad, 2019) examined differences between the preferences of police investigators and laypeople for lies over implausible truths when assigned to the role of innocent suspects in simulated police investigation scenarios. Specifically, police investigators and laypeople were asked to report how they would behave in four imaginary implausible crime scenarios, given their role as innocent suspects. Results showed that police investigators tended to lie more and present more plausible stories when lying compared to laypeople. Police investigators’ tendency to tell lies more than laypeople even when they were asked to simulate the role of an innocent suspect may reflect internal knowledge of how to appear innocent.
Similarly, police investigators may be committed to the Crime Control Model (Packer, 1964), which requires them to secure appropriate sentences of persons guilty of crimes. The crime control model is an affirmative model that emphasizes the existence and exercise of official power. In this context, attributes of the tunnel vision serve the police well when the exercise of power fit in the police culture, low trust, and zero-sum game thinking. Accordingly, police investigators fear false-negative errors (releasing guilty suspects) more than false-positive faults (apprehending innocent persons) and use accusatorial methods in interrogations to force confessions from suspects (Kassin, 2005).
Tunnel vision is associated with various common cognitive biases such as confirmation bias, hindsight bias, and outcome bias which are frequently observed among actors in the criminal justice system (Findley, 2012). Confirmation bias is an effect of favoring information that confirms an individual’s preconceptions or hypotheses independently of the information’s truthfulness or falsity. For example, forensic confirmation bias may corrupt experts’ judgments in various forensic domains and professional forensic sciences. For example, the National Academy of Sciences (NAS, 2009) reported concerns regarding standardization, reliability, accuracy, errors, and the potential for related biases in multiple forensic disciplines. In addition, forensic experts who know the nature and details of the crime then may be biased by the pressure of interrogators, by working for the prosecution, and by using computer lists that feature some suspects ahead of others (Kassin et al., 2013). These influences may increase their vulnerability to biases, including confirmation biases.
Tunnel vision is over-reliance on internal sources while ignoring accurate external information. Confirmation bias is over-reliance on external sources that are not necessarily accurate. People tend to trust their sources in clear-cut cases, while people typically trust external sources in ambiguous cases. Both operate simultaneously and might affect each other. We designed the present study to examine further the combined effects of tunnel vision and confirmation bias on police investigators and laypeople in hypothetical criminal investigation situations.
The present study adapted the two most implausible scenarios used by Elaad (2019). The scenarios were modified so that investigators were asked to act as investigators in familiar situations. One scenario is coupled with intelligence information indicating the suspect’s innocence, while the other is accompanied by intelligence information indicating the suspect’s guilt. The implausible scenarios facilitate presumptions of guilt, and the added information confirms or contradicts those beliefs. The tunnel vision effect predicts that investigators will respond to the incriminating information and regard the exonerating information as irrelevant. Laypeople with no investigation experience may be willing to treat exonerating information more seriously than do investigators. Specifically, for police investigators, tunnel vision and confirmation bias operated in the same direction and strengthened the assumption of guilt. For laypeople, this is not necessarily the case.
Based on these accounts, we hypothesized that police investigators would rely on their internal knowledge (tunnel vision) of how an investigator should act in an investigation and rely on the story’s plausibility. On the other hand, laypeople will rely more on external information.
The presumption of guilt relates to investigators’ perceived ability to detect lies accurately (Elaad, 2009). In practice, people differ little from one another in their ability to detect lies (Bond & DePaulo, 2008), and therefore most people can be regarded as good as others concerning their lie-detection abilities. However, self-assessments of lie-detecting abilities rely on intuitive judgments that are often biased (e.g., Elaad, 2009, 2019). Previous accounts have shown that people tend to rate their ability to detect lies higher than the abilities of other individuals (Ekman & O’Sullivan, 1991; Elaad, 2003). The bias fits general human assumptions that most communications are truthful, and if not, they can easily detect their lack of veracity. Elaad (2019) found that police investigators and laypeople differ in their perceived ability to tell and detect lies, tell the truth convincingly, and detect when other people are truthful. Note that lie-detection and truth-detection are completely distinct abilities because lie-detection requires more significant mental effort than truth detection (Levine et al., 1999). Competence in lie detection engages different cognitive and emotional resources than truth detection (Stewart et al., 2019).
Following Elaad (2019), we suggest that police investigators feel more able than laypeople as successful lie-detectors. Such confidence may be relevant to tunnel vision in the criminal context, and therefore measures of self-assessed abilities are included in the current study.
It is hypothesized that police investigators will show enhanced self-assessed lie-detection abilities than laypeople.
Design
Two equally sized groups of participants, Israeli police investigators and laypeople from the community were tested in a within-subjects design. Four experimental conditions were created by combining scenario (“robbery scenario” and “drug deal scenario”) and intelligence information (“incriminating” and “exonerating”). Scenario/information combinations were counterbalanced across participants in each group. Participants were asked to indicate their confidence in the imaginary suspect’s guilt, the likelihood that they may believe the suspect, and the frustration they experienced on the task. Confidence in the suspect’s guilt, likelihood not to believe, and lack of frustration were used to signify tunnel vision.
Methods
Statistical Power and Participants
The ethical committee of Ariel University approved the study, and the Israeli police authorities granted permission to police investigators to participate in the study.
We based the present effect size on a previous study that used similar hypothetical scenarios and showed a strong effect when police investigators and laypeople were compared about considering these scenarios (Elaad, 2019). A GPower analysis indicated that a sample of 79 participants is adequate to detect an anticipated medium effect size (0.35) with a power of 0.95 and α = .05.
Forty Israeli police investigators (nine females) with a mean age of 38.1 years (SD = 9.0) volunteered to participate in the study. Investigators were asked to state their interrogation experience in years, 39 complied. Experience ranged from 1 to 35 years (M = 11.0, SD = 8.6). Forty laypeople (11 females) with no past interrogation experience were recruited to the study from the local community. Their mean age was 37.6 years (SD = 10.3). The two participant groups’ roughly equivalent ages and gender proportions were planned. All participants signed a consent form and were assured of confidentiality and anonymity.
Hypothetical Crime Scenarios: Plausibility Assessments
The present study used the two most implausible hypothetical crime scenarios developed by Elaad (2019), a robbery and a drug-deal scenario, described below. Elaad (2019) provided plausibility scores for four criminal scenarios, which had been rated by 40 community members from the local community (20 females) on an 11-point scale ranging from 0 (not at all plausible) to 10 (very plausible). Of the four scenarios, the robbery and the drug deal scenarios were awarded the lowest plausibility scores (M = 2.99, SD = 2.63 and M = 2.82, SD = 2.63, respectively).
Hypothetical Crime Scenarios Description
Robbery scenario
You are an investigator in charge of investigating a robbery. The suspect tells you the following: Two days prior to the robbery, he was walking on the street when two masked men suddenly jumped him and forced him into a car. They covered his eyes and after half an hour’s drive, he found himself led into an isolated house. Two days later, with his eyes blindfolded, they forced him into the car again. He felt one of his kidnappers put something in his pants pocket during the ride. The kidnappers then threw him out of the car and drove off. He removed his blindfold and found himself in the middle of an unfamiliar town. After walking for a few minutes, he was approached by two police officers who asked him to identify himself. They then checked the police database and discovered his criminal record. They instantly notified him that he was under arrest on suspicion of committing a robbery a short time ago in the vicinity. They searched him and found a gun in his pocket. He claims that the gun is not his and asks you to believe him.
Drug deal scenario
You are an investigator in charge of investigating a drug deal. The suspect has previous drug convictions and was arrested when he was caught with a commercial quantity of cocaine. The suspect tells you the following: He is a drug addict. Two days earlier, he went to buy cocaine for his use. His regular dealer made a mistake and handed him a larger package of cocaine than usual. He went home, not noticing the mistake. Police detectives arrested him near his home. They searched him and found the commercial quantity of the cocaine on him. He was detained, and now he is being interrogated.
Procedure
Participants were first asked to provide personal information (gender, age, religiosity, and investigators were asked to provide their interrogation experience in years). They then completed a short 4-item questionnaire individually, about their lie/truth-related abilities. Specifically, participants were asked to self-assess their abilities to tell lies, tell the truth, detect lies, and detect truths relative to other people. Responses were given on a scale ranging from 0 (much worse than other people) to 100 (much better than other people) with a middle point of 50 (as good as others).
Participants were also presented with two hypothetical criminal scenarios (the robbery and the drug deal scenarios), each related to an implausible fictional offense and an imaginary suspect. Both investigators and lay participants were asked to imagine that they were interrogating the suspect in each scenario. At the bottom of each scenario, a note indicated that intelligence information implies that the suspect might be guilty or innocent.
Half of each group (investigators and laypeople) received information implying guilt for the robbery scenario and information implying innocence for the drug deal scenario. The second half of each group received the reverse information.
Like in Elaad (2019), participants were asked to indicate on an 11-point scale: (a) their confidence in the suspect’s guilt or innocence. Responses ranged from 0 (absolutely innocent) to 10 (utterly guilty); (b) the likelihood that they may believe the suspect’s story. Responses ranged from 0% (not at all) to 100% (very much); (c) Frustration level in the interrogation situation. Responses ranged from 0 (no frustration at all) to 10 (very frustrated). Frustration is an emotion that occurs in situations where the individual is hindered from reaching a preferred outcome. In the present study, investigators acted in a familiar domain and were not blocked from making good decisions. Laypeople found themselves in an unfamiliar situation where making decisions may be frustrating. For example, an implausible scenario followed by intelligence information indicating innocence should frustrate laypeople more than investigators who are used to such inconsistencies. Group differences between investigators and laypeople in frustration are, therefore, expected.
Results
Effects of Scenario and Added Information on Guilt Assessments
Participants were asked to express their confidence in the suspect’s guilt or innocence (Table 1). A 2 × 2 × 2 ANOVA, scenario (robbery and drug deal), information (incriminating and exonerating), and group (investigators and laypersons), was conducted on participant’s confidence ratings. Scenario and information factors served as within-subject factors and the group factor as a between-subject factor. Results revealed a significant scenario effect, F (1, 38) = 15.2, p < .001, η2 p = .29, indicating that participants were more confident about the guilt of the drug deal suspect than about the guilt of the robbery suspect. A significant information effect, F (1, 38) = 21.0, p < .001, η2 p = .36, was also obtained, showing that participants were more confident to ascribe guilt to a suspect when incriminating information was added and was less confident about the suspect’s guilt when exonerating information was included. Finally, a significant main effect for group was obtained, F(1,38) = 11.2, p = .002, η2 p = .23, implying that police investigators assigned higher confidence scores about the suspect’s guilt than did laypeople. A closer look at this difference reveals that investigators assigned higher confidence scores to their presumption of guilt than laypeople, when incriminating, t(78) = 2.72, p = .008, d = 0.61, and exonerating information, t(78) = 2.06, p = .036, d = 0.48, were presented (Table 1). No significant interaction effects were observed.
Means (and SDs) of Confidence Ratings in the Suspect’s Guilt by Investigators and Laypersons.
Note. Higher scores indicate guilt ratings. Lower scores indicate innocence.
Likelihood to Believe the Imaginary Suspect
Participants were asked to indicate the likelihood that they may believe the suspect in each scenario. The means and standard deviations of the likelihood scores are displayed in Table 2. A 2 × 2 × 2 ANOVA, with one between-subject factor (group) and two within-subject factors (scenario and information), was conducted on the likelihood scores. The analysis revealed a significant main effect for scenario F(1, 38) = 12.5, p = .001, η2 p = .25, indicating that the robbery scenario received higher likelihood scores than the drug deal scenario. A significant information effect, F(1, 38) = 13.2, p = .001, η2 p = .26, shows that participants tended to assign a higher likelihood score when exonerating information was added, and lower likelihood to believe when incriminating information was included. Finally, a significant main effect for group, F(1, 38) = 37.5, p < .001, η2 p = .50, implies that police investigators assigned lower likelihood scores than laypeople. To further examine the last difference, we used t-tests for independent groups. It appeared that laypeople were more willing to believe the suspects than investigators irrespective of the added information – incriminating information, t(78) = 5.2, p < .001, d = 1.16, and exonerating information, t(78) = 5.7, p < .001, d = 1.28. No significant interaction effects were found.
Means (and SDs) of Likelihood to Believe the Suspect by Investigators and Laypersons.
Note. Higher scores indicate more believing.
Frustration Ratings
Participants were also asked to rate the degree of frustration they experienced when reading each scenario and making decisions of guilt or innocence. Mean frustration ratings are presented in Table 3. Frustration scores were analyzed using a similar 2 × 2 × 2 ANOVA (scenario, information, and group). A significant group effect, F(1,38) = 30.5, p < .001, η2 p = .45, shows that laypersons were much more frustrated than investigators. The laypeople’s frustration can be explained by the fact that they acted in an unfamiliar domain. A significant group difference appeared when both incriminating, t(38) = 4.8, p < .001, d = 1.53, and exonerating, t(38) = 4.1, p < .001, d = 1.29, information was added. A significant scenario effect, F(1, 38) = 9.8, p = .003, η2 p = .21, indicated that the robbery scenario caused greater frustration than did the drug deal scenario. No other significant effects were obtained.
Means (and SDs) of Frustration Ratings by Investigators and Laypersons.
Note. Higher scores indicate higher levels of frustration.
A closer look at the laypeople’s frustration reveals a significant positive correlation between the likelihood to believe and frustration scores, r (40) = .32, p = .047. Results may suggest that the more laypeople believed the suspect, the frustration grew larger. No significant link between the likelihood to believe and frustration ratings was found for investigators.
Self-Assessed Abilities to Tell and Detect Lies and Truths
The means and standard deviations computed for the self-assessed abilities are presented in Table 4. To examine group differences in the self-assessments of these abilities, a 2 × 4 MANOVA was used with group (investigators and laypeople) as the between-subject variable, and abilities (the four self-assessed abilities) as the dependent variable. The MANOVA showed a significant overall group effect (λ = .84, F(4,75) = 3.6, p = .01, η2 p = .16). Results imply that investigators exhibited significantly higher assessments of their lie-detecting ability than did laypeople (F(1,78) = 10.5, p = .002, η2 p = .12), and significantly higher assessments of their truth-telling ability (F(1,78) = 3.9, p = .05, η2 p = .05). However, they showed significantly lower truth-detecting ability assessments than did laypeople, (F(1,78) = −6.5, p = .013, η2 p = .08). Finally, the lie-telling ability assessments showed no group difference (F(1,78) = 0.3, p = ns).
Means (and SDs) of Self-Assessed Abilities to Tell and Detect Lies and Truths.
Note. Higher scores indicate higher self-assessed abilities.
Discussion
The present study was designed to highlight factors that undermine police legitimacy in the eyes of the public. First, we argued that police legitimacy is challenged by dominant biases in the police interrogation process. Two of these biases, tunnel vision, and confirmation bias, were manipulated to explore how these biases operate together and affect professional police investigators and laypeople’s attitudes toward imaginary suspects who presented implausible stories in hypothetical criminal interrogation situations.
Results showed that investigators were more strongly affected than laypersons by tunnel vision. Specifically, implausible scenarios triggered three tunnel vision facets: (a) Confidence in the suspects’ guilt. It was shown that police investigators were more confident about the suspect’s guilt than laypeople. (b) Likelihood to disbelieve the suspect’s story. Investigators were less likely to believe the suspect’s story than laypeople. (c) Lack of frustration on the task. Laypeople showed a much higher frustration than investigators, when they decided to believe the suspect’s account. The lower frustration level of investigators may be attributed to the familiar criminal context in which they found themselves.
Confidence in the suspect’s guilt, rejecting any option to believe the suspect, and low frustration on the task, indicate that the tunnel vision that was created was strong. We suggest that investigators and laypeople alike presumed guilt when reading and judging the implausible scenarios, but investigators created more robust tunnel vision than laypeople. Investigator’s strong presumption of guilt may be associated with their elevated self-assessed lie-detection ability, significantly higher than laypeople’s self-assessment.
Presumption of guilt is an apparent bias in criminal interrogations. Kassin (2005) noted that police investigators believe that the suspects they interrogate in custody are not innocent. In a study manipulating involvement in the commission of a mock crime and interrogator’s presumption of guilt or innocence, Kassin et al. (2003) reported that investigators who presumed guilt asked incriminating questions, were aggressive, and tried hard to extract a confession from the suspect.
The tunnel vision hypothesis was examined against intelligence information about the suspect’s guilt or innocence. Exonerating information was coupled with one scenario, and incriminating information was added to the other scenario. According to tunnel vision predictions, investigators who developed a firm belief in the suspect’s guilt were expected to render the exonerating information irrelevant, while laypeople were expected to consider the exonerating information more seriously than investigators. However, the results did not support this suggestion. While all participants were confident to ascribe guilt to the imaginary suspects, irrespective of the added information (incriminating or exonerating), this was particularly true for police investigators. Specifically, police investigators were more confident about the suspect’s guilt than laypeople in the presence of incriminating and exonerating information. Still, investigators did not ignore exonerated information and lowered their confidence in the suspect’s guilt when exonerating information was presented.
Similar effects were found concerning the likelihood of believing the imaginary suspects. The general tendency was not to believe. Still, police investigators were more unwilling to believe the suspect’s story than laypeople, irrespective of the added information. However, police investigators did not ignore the exonerating information and responded by attenuating their reluctance to believe the suspects.
In summary, investigators and laypeople accepted the exonerating information as a reliable external source of information, moderated their confidence in guilt judgments, and enhanced their likelihood of believing the suspect.
The creation of tunnel vision and confirmation bias is based on global impressions rather than on single, well-defined behaviors (Hartwig & Bond, 2011). People tend to instantly make intuitive lie/truth judgments based on very little information rather than invest effort in careful consideration of the available information. This implies that people generally rely on simple heuristics when deciding whether to trust a message. Police investigators are not exceptional and can be susceptible to simple observer biases when in doubt whether to pursue the interrogation of a suspect.
Beliefs are generally formed in two ways: by inner sources such as experiences, inferences, and deductions or by accepting what an external source tells us to be true. The belief that the suspect is guilty because the scenario does not make sense is an example of a belief influenced by an inner source. Investigators have much more experience handling similar cases than laypeople and therefore find it easier to generate a firm belief about the suspect’s guilt. This notion is supported by Hartwig and Bond (2011), who noted that implausible, illogical accounts with few details trigger distrust. Similarly, Aamodt and Custer (2006) suggested that a story that does not make logical sense is often judged in real-world situations as being deceptive.
It has already been demonstrated that police investigators are sensitive to implausible scenarios (Elaad, 2019). Such sensitivity was replicated in the present study. Elaad (2019) suggested that innocent suspects should try to prepare a plausible version of their statement. A plausible statement is especially important because previous results show that while guilty suspects prepare themselves for an interrogation, innocent suspects feel no need to invest efforts in such preparations because they believe their innocence must be apparent (Granhag & Hartwig, 2008). The present results support this recommendation.
On the other hand, people at large are anchored to the truth-bias, or the tendency of receivers to judge messages as being truthful rather than deceptive (e.g., Kohnken, 1989; Vrij, 2008). In a meta-analysis of lie detection studies, Bond and DePaulo (2006) found that 56% of the messages were judged as truthful and only 44% as deceptive, although participants were presented with equally frequent truths and lies. The truth bias may explain the laypeople’s tendency to be less committed to the presumption of guilt than were investigators. Still, believing an implausible scenario frustrated laypeople.
It is well known that preexisting beliefs influenced decisions of law enforcement personnel such as investigators (Kassin et al., 2003), jurors (Lange et al., 2011), and judges (Halverson et al., 1997). However, ambiguity has a moderating effect on this relationship. Preexisting beliefs affect forensic expert judgments in the criminal justice system when the judged case is ambiguous and not when it is conclusive. Elaad et al. (1994) demonstrated the effect among police polygraph examiners who received inconclusive charts and were asked to make a lie/truth decision. Biasing information noted that the examinee had later confessed or someone else had later confessed. The added information influenced decisions. However, when charts indicating deception were presented, judgments were not influenced by biased information that denoted innocence. It appears that tunnel vision was set on when conclusive charts were presented.
Significant scenario effects were found in the present study, indicating that the suspect in the drug deal scenario was judged more strictly than the robbery scenario suspect. Specifically, the drug deal suspect was deemed guilty more confident than the robbery suspect. Furthermore, the drug deal suspect was believed less, with lower frustration involved, than was the robbery suspect. Such differences in interrogations are common. There are straightforward and complex cases, and investigators are expected to solve both successfully. There are so-called jailhouse informers or prisoners who claim to have heard a confession from a (usually high profile) suspect and testify against the suspect in exchange for some benefit. These informers are notoriously unreliable. Because of tunnel vision, their testimony continues to be implicated in wrongful convictions. Investigators are expected to unravel such cases.
A meta-analytical review (Elaad, 2018) showed that people generally overrate their lie-detection ability. This overestimation is especially true for law enforcement officers (Elaad, 2003) and police investigators (Elaad, 2009). We explained the bias by the desire of some investigators to avoid an appearance of being easily deceived and therefore unprofessional. Nevertheless, we may suggest that many police investigators believe that if a suspect is lying, they can easily detect these lies (Williams & Gilovich, 2008). To this end, interrogators use learned deception cues, including the significant cue of message plausibility.
Tunnel Vision Distortions in the Criminal Investigation Process
The number of identified wrongful convictions and false confessions is mounting, and in almost every wrongful conviction, the problem of tunnel vision is present (Findley, 2012; Findley & Scott, 2006; Martin, 2002). According to Martin (2002), tunnel vision distortion is particularly damaging in the investigative process. Investigator’s misconduct becomes prevalent in note and record-keeping, witness interviews, the interrogation of suspects, and the conduct of searches. The consequences are often the presence of incomplete or even entirely inaccurate evidence in the trial, which increases the risk of miscarriages of justice. Martin (2002) further emphasized that tunnel vision amplifies the risk of using the informer’s evidence. Police work relies on informers, either people who report a crime or provide information to the police for some profit. According to Martin (2002), informers misleading information are often present in wrongful convictions. The informer’s problem is particularly severe in the case of prisoners who claim to have heard a confession from a suspect and profit from testifying against that suspect. Such testimony is notoriously unreliable. Tunnel vision enhances the likelihood that the testimony is heard in court and sometimes leads to a wrongful conviction.
Future Research
Future research is advised to pay attention to other tunnel vision features such as the belief perseverance phenomenon (e.g., Ross et al., 1975). According to the belief perseverance phenomenon, investigators tend to preserve their guilt judgments and refuse to change their views. For example, despite the decisive DNA evidence that the semen taken from a rape-murder scene could not have come from the suspect, police investigators might continue to look for evidence explaining how the suspect is guilty.
Limitations
The simulated scenarios may have appeared somewhat artificial to police investigators. Specifically, the impact of the present scenarios is less significant than actual performance would suggest, and police investigators may behave differently in real-life situations. Therefore, caution is recommended when generalizing the present results to actual police performance. An experimental design that includes rewards for successfully identifying of suspects by police investigators and laypeople should be considered in future studies.
It may be argued that the presented intelligence information came across as contrived, and therefore results may say little about real-world cases in which investigators rely on specific hard evidence when developing their beliefs. Indeed, when hard evidence exists, the decision to pursue a suspect’s interrogation is based on the available hard evidence. However, even when hard evidence is scarce, interrogators are not always inclined to release the suspect. In such cases, police investigators may rely on less decisive evidence that may prompt judgmental biases. For example, police investigators may rely on the plausibility of the suspect’s account when deciding whether to pursue or terminate the interrogation.
Finally, we did not control for aphantasia, or the inability to form mental images of people that are not present, which may be relevant to the present scenarios. However, no participant complained about such difficulties. Therefore, it seems that aphantasia was not a problem here.
Conclusions
Two opposing biases potentially affect police investigator’s decisions in criminal interrogations. One is over-reliance on external sources that are not necessarily accurate. The other is tunnel vision, which is over-reliance on own sources while ignoring external information. Such over-reliance might be similarly misleading. Extensively studied, external sources are typically trusted in ambiguous cases, while tunnel vision is shared in clear-cut cases. However, most cases are neither clear nor unclear, which leaves room for both tendencies to operate simultaneously. The current study demonstrated more robust tunnel vision and enhanced self-assessed lie-detection abilities among police investigators than laypeople. However, reliance on external information was firm among investigators and laypeople alike. Future research aims to find ways to moderate the documented biases and be more cautious when presuming guilt. Specifically, to enhance awareness about the investigator’s over-reliance on internal and external sources that may be misleading.
Footnotes
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
