(Year 1 Coursework - 2019) The Criminal Justice System, Psychology and Language Sciences

I have written this 3000-word essay as my Capstone Project at the end of my first-year of undergraduate studies (required for progression into the second year).

Insights from Social Psychology, Neuroscience, Perception and Speech Science Research

I. Social Psychology and the Criminal Justice System

Scientists formulate hypotheses and seek information and data (experimentation) to test these hypotheses. However, they can, even unintentionally, engage in forms of information processing that lead them to selectively process information confirming their hypotheses and overlook data that challenge them. This bias towards confirming one’s assumption is termed “confirmation bias”, first coined by Wason (1960) (Oswald & Grosjean, 2004). In his study (1960), Wason gave participants a sequence of 3 numbers (2-4-6) and asked them to find the rule governing this sequence, by formulating a hypothesised rule and testing it. Wason observed that most participants sought sequences that confirmed their hypothesis and that this testing technique sufficed to make them confident in their rule (Oswald & Grosjean, 2004). Most people’s testing strategy operates with the goal of confirming their assumptions or beliefs instead of trying to falsify them, although several philosophers of science and psychologists have underlined the importance of “disconfirmation” (Klayman & Ha, 1987); thus, it seems that neither scientists (extensively formulating and testing hypotheses) nor the general public is immune to the confirmation bias.

This biased search for information has important implications in the criminal justice system, and especially at the early stage of police interviewing: this stage constitutes the first point of contact of the suspect with the criminal justice system and is critical, as its outcome can determine whether the case is brought to the Court and the judge. Research into police investigations shows that, often, police officers conducting these interviews assume suspects to be guilty before the interview (i.e. they form a “guilty” hypothesis) and their goal to be the confirmation of this guilty hypothesis (i.e. obtaining a confession) (Hill, Memon & McGeorge, 2008). Given that, one may wonder whether these assumptions influence, in any way, the police interviewing process. Hill, Memon and McGeorge (2008) conducted three studies on university students to observe whether holding a presumption of guilt would impact the way they conduct interviews on mock suspects (some of which are innocents, others guilty).

In Study 1, results showed that participants in the “presumption of guilt” condition (led to expect the suspect to be guilty) generated more guilt-presumptive questions than participants in the “presumption of innocence” condition: this shows that the former group of students sought evidence confirming their expectation of guilt (confirmation bias). Also, students in the “presumption of guilt” condition rated themselves as “more confident” with their judgement (guilt or innocence) than students presuming the suspect to be innocent.

Study 3 focused on whether the style of questioning impacts the verbal behaviour of the suspect: another group of students listened to the recordings of the suspects’ responses to the questions (asked by students in Study 1), without listening to the questions themselves. The results showed that suspects responding to guilt-presumptive questions were perceived as guiltier than suspects responding to neutral questions; above all, innocent suspects were perceived as guiltier when responding to guilt-presumptive questions than guilty suspects responding to the same questions. Thus, suspects appear to have responded to guilt-presumptive questions in ways that satisfy the expectation of guilt: that is, those questions produced a self-fulfilling prophecy.

Consequently, these findings question the objectivity of police interviews: if police officers believe the suspect to be guilty beforehand (as they tend to), this could influence their questioning style (Study 1), which in turn influences the behaviour of the suspect in ways that appear to confirm their ‘guilt’; this effect on behaviour is even more pronounced if the suspect is in fact innocent (Study 3). Additionally, experienced officers seem more susceptible to confirmation bias, more often seeing the suspect as deceitful than truthful (Kassin, Meissner & Norwick, 2005). Thus, experience may not hinder the influence of the confirmation bias.

However, there might be techniques to reduce confirmation bias. For instance, Hernandez and Preston (2013) showed a reduction in this selective information processing when information is presented in a disfluent format, that is, when participants experienced some difficulty in processing the information. Thus, in the context of police interviews where officers have access to pieces of evidence prior to the interview, it is worth exploring the outcomes of changing the presentation style of evidence provided to the officers (to a disfluent format): this could promote a stronger consideration of opposing views, as suggested by Hernandez and Preston (2013), and prevent officers from holding a presumption of guilt before the interview: consequently, this could hinder the self-fulfilling prophecy, causing, in suspects, behaviours consistent with the presumption of guilt.

References: Social Psychology

  • Hernandez, I., & Preston, J. L. (2013). Disfluency disrupts the confirmation bias. Journal of Experimental Social Psychology, 49(1), 178-182. https://doi.org/10.1016/j.jesp.2012.08.010

  • Hill, C., Memon, A., & McGeorge, P. (2008). The role of confirmation bias in suspect interviews: A systematic evaluation. Legal and Criminological Psychology, 13(2), 357-371. https://doi.org/10.1348/135532507X238682

  • Kassin, S. M., Meissner, C. A., & Norwick, R. J. (2005). ‘I’d know a false confession if I saw one’: A comparative study of college students and police investigators. Law and Human Behaviour, 29, 211–227. https://doi.org/10.1007/s10979-005-2416-9

  • Klayman, J., & Ha, Y. W. (1987). Confirmation, disconfirmation, and information in hypothesis testing. Psychological Review, 94(2), 211-228. https://doi.org/10.1037/0033-295X.94.2.211

  • Oswald, M. E., & Grosjean, S. (2004). Confirmation bias. Cognitive Illusions: A Handbook on Fallacies and Biases in Thinking, Judgement and Memory, 79-96.

II. Neuroscience and the Criminal Justice System

The use of neuroscientific evidence in criminal cases has significantly increased from the beginning of the 21st century, with the introduction of such evidence to the courtroom going back to the early 20th century (Aono, Yaffe & Kober, 2019). Neuroscientists have long been informing judges and jurors in courts for crucial decisions; indeed, criminal justice history shows cases where neuroscientific evidence has played a critical role in finding the accused not guilty (Greely & Farahany, 2019). One particular criminal justice system area in which neuroscience can provide valuable insight is the cautious use of witness testimonies as evidence of guilt or innocence. Indeed, there is a common assumption in the general public, judges and jurors that memories for violent and stressful events are well-encoded and reliable ‘recordings’ of the environment; that confident witnesses’ testimonies are generally accurate reproductions of the crime scene (Lacy & Stark, 2013). However, neuroscientific research has shown that episodic memories (memories of specific events in time and space) are prone to distortions as extreme as the formation of false memories for events.

One of the earliest and most remembered neuroscientific case studies of memory, Scoville and Milner’s study on patient H.M. in 1957, showed that the bilateral removal of his medial temporal lobes (including the hippocampus) led H.M. unable to form new episodic memories (Squire, 2009). However, brain areas need not be injured to impact one’s ability to form or retrieve episodic memories accurately. The most apparent memory problems in the judicial context that one can think of are incomplete memory encoding and memory loss. However, witnesses can also be vulnerable to ‘false memory’, or the memory for an event that has actually been experienced differently or not at all (Werner, Kühnel & Markowitsch, 2013). Particularly, the early contact with additional information from the crime scene can distort witnesses’ memories. In the judicial context, memory distortions have been shown to occur due to misleading or suggestive information, one extensively studied effect being the ‘misinformation effect’. In police interviews, for example, distortion to the original memory following the exposure to misleading information (related to that memory) has been demonstrated to induce false memories in individuals (Lacy et al. 2013; Loftus, 1979). An example scenario would be an eyewitness’ memory of the face of the criminal being influenced by their exposure, prior to the testimony, to an innocent suspect’s photo.

One way in which neuroscience can experimentally inform the Court is the distinction between true and false memories using neuroimaging techniques to measure brain activity during the retrieval of these memories (Schacter, Chamberlain, Gaesser & Gerlach, 2012). Most of the neuroimaging studies on this matter have relied on the ‘sensory reactivation hypothesis’ stating that true memories lead to more sensory-perceptual details than false memories at retrieval (Schacter & Loftus, 2013). Following that, an fMRI study on the neural correlates of true memories, false memories and deception revealed that brain areas involved in encoding sensory-perceptual information (left temporoparietal regions) tended to be more active during retrieval of true than false memories (Abe et al., 2008). Another study, using PET imaging, also showed significant blood flow increases in left temporoparietal regions for the veridical recognition of printed words (heard seven minutes before the test phase), compared to illusory recognition, that is, the recognition of words not heard before the test phase (Schacter et al., 1996).

Although it is tempting to apply these neuroimaging techniques and findings to the judicial setting, Schacter and Loftus (2013) have raised potential issues in this premature application: the simple materials (e.g. pictures, words) used by laboratories differ from the rich sensory information in everyday life; the participants (healthy young adults) differ from the diverse population of witnesses; relatively short periods separate the study phase (where encoding takes place) from the test phase (moment of retrieval). Further research is still needed to confirm that these neuroimaging findings can generalise to the complexity of real-life scenarios.

Given the current scope of neuroimaging studies, it is necessary for judges, jurors and police officers to be cautious about the reliability of witness testimonies. Documentation on the neural basis of memory is needed on their side: contemporary studies revealed that judges and law enforcement personnel are closer to university students than to eyewitness testimony experts, in terms of awareness of memory phenomena (Benton, Ross, Bradshaw, Thomas & Bradshaw, 2005); on the study’s 30-item questionnaire, they responded differently from the experts for 60% of items; with jurors, this percentage went up to a significantly high 87%. Thus, raising awareness on the detrimental influence of misinformation during police interviews and the possible emergence of false memories is essential in the judicial setting.

References: Neuroscience

  • Abe, N., Okuda, J., Suzuki, M., Sasaki, H., Matsuda, T., Mori, E., Tsukada, M., & Fujii, T. (2008). Neural correlates of true memory, false memory, and deception. Cerebral Cortex, 18(12), 2811-2819. https://doi-org.libproxy.ucl.ac.uk/10.1093/cercor/bhn037

  • Aono, D., Yaffe, G., & Kober, H. (2019). Neuroscientific evidence in the courtroom: A review. Cognitive Research: Principles and Implications, 4(1), 1-20. https://doi.org/10.1186/s41235-019-0179-y

  • Benton T. R., Ross D. F., Bradshaw E., Thomas W. N., Bradshaw G. S. (2005). Eyewitness memory is still not common sense: Comparing jurors, judges and law enforcement to eyewitness experts. Applied Cognitive Psychology, 20,115–129. https://doi.org/10.1002/acp.1171

  • Greely, H. T., & Farahany, N. A. (2019). Neuroscience and the criminal justice system. Annual Review of Criminology, 2, 451-471. https://doi.org/10.1146/annurev-criminol-011518-024433

  • Lacy, J. W., & Stark, C. E. (2013). The neuroscience of memory: Implications for the courtroom. Nature Reviews Neuroscience, 14(9), 649-658. https://doi.org/10.1038/nrn3563

  • Loftus, E. F. (1979). Eyewitness Testimony. Harvard; Cambridge.

  • Schacter, D. L., Chamberlain, J., Gaesser, B. & Gerlach, K. (2012). Neuroimaging of true, false, and imaginary memories. In L. Nadel & W.P. Sinnott-Armstrong (Eds.), Memory and Law (pp.233-262). New York: Oxford University Press.

  • Schacter, D. L., & Loftus E. F. (2013). Memory and law: What can cognitive neuroscience contribute? Nature Neuroscience, 16(2), 119–123. http://dx.doi.org.libproxy.ucl.ac.uk/10.1038/nn.3294

  • Schacter, D., Reiman, E., Curran, T., Yun, L. S., Bandy, D., McDermott, K. B., & Roediger, H. L. (1996). Neuroanatomical correlates of veridical and illusory recognition memory: Evidence from positron emission tomography. Neuron, 17(2), 267-274. https://doi-org.libproxy.ucl.ac.uk/10.1016/S0896-6273(00)80158-0

  • Squire, L. R. (2009). The legacy of patient HM for neuroscience. Neuron, 61(1), 6-9. https://doi.org/10.1016/j.neuron.2008.12.023

  • Werner, N. S., Kühnel, S., & Markowitsch, H. J. (2013). The neuroscience of face processing and identification in eyewitnesses and offenders. Frontiers in Behavioral Neuroscience, 7, 1-12.

III. Perception and the Criminal Justice System

The visual world around us is extremely rich in details; however, our stable representations across views do not rely on such fine details (Simons & Levin, 1997). Further, research has shown that we commonly experience a perceptual phenomenon called ‘change blindness’, i.e. the inability to detect changes to an object or a scene, as defined by Levin and Simons (1997) in their review. In one example research study, Simons and Levin (1997) observed the extent of change blindness in ten undergraduate students viewing a short video depicting a conversation between two actors: across the cuts of the video, at least one continuity error was made (e.g. the disappearance of a colourful scarf). After a first viewing, participants were asked whether they noticed any changes between cuts, and then proceeded to watch the video again. Before the second viewing, they were told that there was a total of nine changes in the scene and instructed to catch them. After the first viewing, only one participant noticed a change (and only one change out of nine); after the instructed second viewing, on average, participants only noticed two changes. Therefore, Simons and Levin (1997) concluded that sensory information across views are generally not integrated, although ‘effortful encoding’ (which occurred during the second viewing that was preceded by instructions) seems to produce longer-lasting visual representations.

These findings, replicated by many further studies in realistic interaction settings, both by Simons and Levin and other researchers (Simons & Rensink, 2005), are of significant importance in the judicial context of eyewitness testimonies. Indeed, eyewitnesses to criminal or violent scenes process rich visual information, often in relatively short time periods; following that, they are requested to recall the event with as many details as possible and to identify the criminal from line-ups including suspects. However, unlike the participants in the study described previously, as the crime proceeds, eyewitnesses are typically neither ‘told’ (i.e. not ‘instructed’ beforehand) nor particularly trying to intentionally catch changes to the crime scene.

In line with this context, Fitzgerald et al. (2016) investigated change blindness in eyewitnesses for a multiple-actors criminal event. In their study, 180 undergraduate students viewed a video beginning with an innocent person walking through a building and ending with another person committing a theft. Fitzgerald et al. (2016) were interested in the differences in accuracy and confidence ratings between participants who detected the change (‘change detection’ group) and those who did not (‘change blindness’ group). The two video actors were relatively dissimilar in appearance (the innocent weighted 50 lbs heavier than the culprit), which is often the case in real-word crime scenes. During the line-up identification task in the questionnaire (containing black-and-white photographs, either culprit-present or culprit-absent and with five additional filler innocent suspects), participants had several choices: identifying either the video innocent (in the culprit-absent line-up), the culprit (in the culprit-present line-up), or a filler person as the thief, or writing that the thief was not present (rejection).

Fitzgerald et al. (2016) reported that only 36.1% of all participants detected the change between the video innocent and the video culprit, demonstrating the prevalence of change blindness in participants’ visual experience. Line-up accuracy (correct identification or rejection) was significantly lower in the change blindness group (28.7%) than in the change detection group (53.1%), showing that awareness of the change has a direct impact on subsequent line-up choices. Further, in the culprit-absent line-up condition, only 31% of participants in the change blindness group correctly rejected the line-up, compared to a significantly higher percentage (69%) in the change detection group. Misidentification rates (innocent or fillers as the culprit) were also higher in the change blindness group.

These findings demonstrate the robustness of ‘change blindness’, experienced by the majority of participants, despite the controlled dissimilarity between the video culprit and the video innocent. They also give a valuable insight into the effect of change blindness on line-up choices, with eyewitnesses that experienced change blindness being more susceptible to misidentify the culprit.
Given that change blindness in eyewitnesses is beyond the control of law enforcement and courts, judges and jurors need to be aware of the existence and impacts of this common perceptual phenomenon (Gibbs, Davies & Chou, 2016). They should be cautious about the outcomes of line-ups and eyewitness testimonies. Gibbs et al. (2016) also underline the importance of reviewing and standardising the interview process, and the need for further research on the relationships between personal characteristics, perceptual disabilities, mental disorders and the vulnerability to change blindness.

References: Perception

  • Fitzgerald, R. J., Oriet, C., & Price, H. L. (2016). Change blindness and eyewitness identification: Effects on accuracy and confidence. Legal and Criminological Psychology, 21(1), 189-201. https://doi.org/10.1111/lcrp.12044

  • Gibbs, R., Davies, G., & Chou, S. (2016). A systematic review on factors affecting the likelihood of change blindness. Crime Psychology Review, 2(1), 1-21. https://doi.org/10.1080/23744006.2016.1228799

  • Levin, D. T., & Simons, D. J. (1997). Failure to detect changes to attended objects in motion pictures. Psychonomic Bulletin & Review, 4(4), 501-506. https://doi.org/10.3758/BF03214339

  • Simons, D. J., & Levin, D. T. (1997). Change blindness. Trends in Cognitive Sciences, 1(7), 261-267. https://doi.org/10.1016/S1364-6613(97)01080-2

  • Simons, D. J., & Rensink, R. A. (2005). Change blindness: Past, present, and future. Trends in Cognitive Sciences, 9(1), 16-20. https://doi.org/10.1016/j.tics.2004.11.006

IV. Speech Science and the Criminal Justice System

The criminal justice system gives considerable weight to witness testimonies and line-up identifications, especially in Court where these can be used as robust pieces of evidence; among witnesses, earwitnesses are those who have not seen but claim to have heard the voice of the criminal during the crime. Earwitnesses participate in earwitness line-ups, in which they try to identify the voice of the criminal among a range of voices. Voice identification evidence, from non-expert earwitness line-ups, can potentially constitute the only means of identifying the criminal, especially when no speech recordings of the criminal are available to the expert investigation. In this case, they can play a critical role in courts. Consequently, awareness of the factors that might bias the identification of criminals’ voices during earwitness line-ups is crucial, both for the police officers conducting these line-ups and the decision-makers in the courtroom (judges, jurors) that interpret their outcomes. Experts can inform on such factors; a particular area of high relevance is forensic phonetics, i.e. the application of the knowledge and methods of phonetics (e.g. the phonetic study of speaker differences) to the legal and judicial settings (Jessen, 2008).

One obvious factor that might lead to inaccurate speaker identifications in earwitness line-ups is the intentional attempt of the criminal to modify their voice during the crime to hide their original voice: common strategies include pitch variation (heightening or lowering) and denasalisation (Künzel, 2000). Künzel’s study (2000) showed a correlation between the speaker’s initial pitch and their personal pitch-variation strategy to disguise their voice: speakers with higher-than-average fundamental frequencies tended to disguise their voice by increasing it even more, and conversely for speakers with lower-than-average fundamental frequencies. This strategy might have important judicial implications for the identification of disguised voices: earwitnesses’ descriptions of pitch might inform forensic phoneticians on the original pitch of the criminal. However, factors beyond the intentional effort of the criminal to change their voice can bias the identification in earwitness line-ups. For instance, earwitnesses seem to be biased towards voices with specific pitches. Indeed, studies have demonstrated significant distortion effects for pitch (termed ‘pitch exaggeration’): when the original voice has a low pitch, earwitnesses are likely to select a voice (from the line-up) with an even lower pitch, and conversely for high-pitched voices (Mullennix et al., 2010). Thus, earwitnesses cannot reliably inform phoneticians on criminals’original pitches.

Although there are several obvious ways for the criminal to intentionally change their speech voice at the moment of the crime or during a line-up (voice disguise, e.g. pitch variation, imitating another speaker’s voice), the speech of a single speaker is by definition extremely variable. Uncontrollable factors such as linguistic content, speaking style, emotional state, health condition and acoustic environment affect the speaking voice of the criminal at the moment of the crime, and might differ in another context (e.g. during the line-up). It is thus essential to consider these factors carefully when weighting an earwitness’ claim of identification. For instance, emotional states have been shown to impact speakers’ voices in various studies: one such study investigated the acoustic attributes of several emotional states (fear, anger, sadness and joy), showing, for instance, ‘significant differences’ in volume variance (intensity of voice) and average pitch between ‘anger and ‘fear’ states (Sobin & Alpert, 1999). Thus, pitch heightening or lowering in the voice of the criminal might also emerge unintentionally; one may think of the likely scenario in which the voice of a violent and angry criminal during the crime would sound substantially different to the earwitness during the line-up in which ‘anger’ might not affect the criminal’s voice as much as ‘fear’ (e.g. of being identified) or another emotional state.

Finally, it is worth noting that there is no consensus among expert phoneticians and speech scientists on whether earwitness line-ups should be conducted in the first place, with some experts advocating for its absolute ban. Reasons for this viewpoint include the problematic way of constructing and conducting these line-ups that lacks clear guidelines (McGorrery & McMahon, 2017). Moreover, possessing speech recordings of the criminal does not guarantee subsequent accurate voice identification, even if their examination is left to phoneticians. Often, legal authorities claim that spectrographic evidence can accurately and reliably identify speakers, considering spectrograms as unique to an individual as their fingerprint (‘voiceprints’); however, most experts consider spectrograms to be extremely variable, thus unreliable (Rafferty, 1999). Fundamental factors such as sample duration and acoustic quality of speech recordings may greatly influence spectrographic evidence (Eriksson, 2005).

References: Speech Science

  • Eriksson, A. (2005). Tutorial on forensic speech science. Part I: Forensic phonetics. In INTERSPEECH-2005. Proceedings of the 9th European Conference on Speech Communication and Technology (pp. 4-8). Lisboa, Portugal.

  • Jessen, M. (2008). Forensic phonetics. Language and Linguistics Compass, 2(4), 671-711. https://doi.org/10.1111/j.1749-818X.2008.00066.x

  • Künzel, H. J. (2000). Effects of voice disguise on speaking fundamental frequency. Speech, Language and the Law, 7(2), 149-179.

  • McGorrery, P. G., & McMahon, M. (2017). A fair ‘hearing’: Earwitness identifications and voice identification parades. The International Journal of Evidence & Proof, 21(3), 262-286. https://doi.org/10.1177%2F1365712717690753

  • Mullennix, J. W., Stern, S. E., Grounds, B., Kalas, R., Flaherty, M., Kowalok, S., May, E., & Tessmer, B. (2010). Earwitness memory: Distortions for voice pitch and speaking rate. Applied Cognitive Psychology: The Official Journal of the Society for Applied Research in Memory and Cognition, 24(4), 513-526. https://doi.org/10.1002/acp.1566

  • Rafferty, L. (1999). Anything you say can and will be used against you: Spectrographic evidence in criminal cases. The American Criminal Law Review, 36(2), 291-311.

  • Sobin, C., & Alpert, M. (1999). Emotion in Speech: The Acoustic Attributes of Fear, Anger, Sadness, and Joy. Journal of Psycholinguistic Research, 28(4), 347-365. https://doi.org/10.1023/A:1023237014909

Acknowledgments

This piece of work has been written by me (Dilay F Ercelik), as part of my undergraduate studies at University College London (UCL).

Written on December 2, 2020