Article body


Evidence in the form of opinion, any opinion, is presumptively inadmissible. Expert opinion evidence is subject to increasingly stringent substantive and procedural requirements prior to its admission into court. This heightened gatekeeping of expert opinion has motivated parties to sidestep those requirements and admit evidence as lay opinion, which is subject to a more flexible and discretionary test. As a result, it has never been more important for courts to draw a coherent distinction between lay and expert opinion. This has not been the case with lay witnesses—often police officers—now opining on traditionally expert topics in a manner exceeding their own personal observations and qualifications. Worse still, as police officers, they are cloaked in the same raiment of authority that makes expert evidence so dangerous. We believe this is a serious problem, and one that has largely flown under the radar in Canada.[1]

Trial judges gatekeep expert evidence because of the risk that the trier of fact will uncritically accept an opinion “cloaked with an aura of expertise.”[2] This gatekeeping function has the further benefit of avoiding lengthy battles between experts when that expertise is not relevant, necessary, reliable, or qualified to begin with. This is crucially important to the criminally accused parties who are frequently underfunded in comparison to the Crown.[3]

The exception for lay opinion did not evolve to undermine these expert rules, but rather as a concession to practicality—a recognition that some experiences, like eyewitness identifications and judgments of speed, are difficult to enunciate without resorting to opinion. Evidence now admitted as lay opinion regularly goes far beyond these humble beginnings, encompassing conclusory opinions about police investigations, forensic gait analysis, and complex accounting opinions. This approach can result in manifestly unjust outcomes. For instance, under the lay rule, an investigating police officer may give lay opinion that a complex series of footsteps he viewed in the snow indicated there was a chase, and that the accused cut off the complainant.[4] The accused, wishing to rebut that evidence, cannot in turn call a lay witness because the lay opinion rule, at least notionally, requires personal observation of the facts. Thus, the accused has little choice but to call an expert, who is likely expensive and may not pass the more stringent test for expert opinion.

In what follows, we will first review the expert and lay opinion rules to establish the stakes at the center of this article: why the distinction between lay and expert opinion evidence is so important. The expert opinion rules, in response to several notorious miscarriages of justice, have become in substance and in procedure increasingly demanding over the past two decades. It is therefore important that courts correctly distinguish between lay and expert opinions, lest parties undermine expert opinion doctrine by admitting their evidence as lay opinion. We then highlight three decisions emblematic of the permissive approach to lay testimony.

We go on to detail three hazards that emerge from our review of the case law. First, lay opinion evidence is unmoored from its original conception as a succinct summary of qualia. Rather, lay witnesses now opine on areas where methodologies exist to control for unconscious biases. For example, humans are excellent pattern seekers and regularly identify illusory patterns in large sets of data. This is especially true when people have some preconceived notion or bias about what they should find, as investigating police officers often do. Yet, because the lay opinion rule is more permissive, lay witnesses need not follow methodologies recently prescribed by peak scientific bodies or generally meet the more stringent reliability requirements imposed on expert opinions.[5]

The second hazard amplifies the first: the lay witnesses providing ipse dixit are, by all appearances, not lay people. Rather, they are often police officers with some superficial investigative experience in the areas on which they are opining—for instance, an expert in police dog behaviour opining on the significance of a pattern of shoe-prints in the snow. These quasi-experts can be expected to hold sway over the trier of fact, but they are not required to prove that they are properly qualified or that their methods are reliable.

Third, courts have failed to adopt a coherent jurisprudential approach to lay opinion across Canada. The principles from the leading Supreme Court case, R. v. Graat, are inconsistently applied and sometimes ignored altogether.[6] Several decisions, even when given a charitable reading, are irreconcilable. The courts’ inconsistent approach creates confusion and injustice.

Taking these three hazards into account, we offer an improved approach to lay opinion evidence. As a guiding principle, we suggest a return to Graat, which remains the leading Supreme Court case on lay opinion which, as we demonstrate, has been widely disregarded. Our proposed approach provides a more structured analysis based on Graat’s principles, which we believe will assist trial judges in applying the proper scrutiny to lay opinion.

I. The Rising Bar for Expert Opinion

In this part, we will review the preconditions for admissibility of expert evidence: what the party tendering the expert must prove before the expert evidence is admitted. We will also discuss the current trend in expert evidence law, whereby the doctrine has grown increasingly restrictive.

Expert evidence is subject to a relatively onerous test prior to its admission. The party seeking to admit the evidence must establish its admissibility by satisfying four requirements enunciated in R. v. Mohan:

  1. Logical relevance;

  2. Necessity in assisting the trier of fact;

  3. Absence of an exclusionary rule;

  4. A properly qualified expert.[7]

Evidence meeting those requirements must then pass another more discretionary test, consisting of balancing the evidence’s benefits to the trial process against its potential risks, including consumption of time and likelihood to create confusion.[8]

The Mohan test has grown stricter over the past several years. Summarizing this body of law, the Supreme Court stated in 2015 that: “[t]he unmistakable overall trend of the jurisprudence, however, has been to tighten the admissibility requirements and to enhance the judge’s gatekeeping role.”[9] More specifically, this trend has seen heightened scrutiny of experts’ reliability, independence and impartiality, and an insistence on necessity over mere helpfulness.

Courts are increasingly attuned to the negative impact of unreliable expert opinion on the legal system. This issue was viscerally documented in the miscarriages of justices found in the Inquiry into Pediatric Forensic Pathology in Ontario and The Commission on Proceedings Involving Guy Paul Morin.[10] Both public commissions highlighted the detrimental impact of unreliable science on the criminal justice system. More generally, empirical research finds that invalid forensic science (admitted as expert evidence) is present in approximately 60% of wrongful convictions that included forensic scientific evidence.[11]

Underlying the concern over reliability is the risk that the trier of fact will uncritically adopt the expert’s ready-made inferences.[12] For example, Justice Goudge warned that experts must support their opinions with more than just their credentials: “Anecdotal evidence and authoritative claims based largely on personal experience characterized the experience-based approach, making the opinions reached largely unquantifiable and shielding them from independent verification.”[13]

In keeping with the warnings in the Kaufman and Goudge Reports, reliability is now a requirement for admissibility when the subject matter of the expert evidence is novel or contested science. In such cases, the party tendering the evidence must demonstrate its reliability according to the four factors found in the leading U.S. case, Daubert v. Merrell Dow Pharmaceuticals Inc.[14] These four factors are:

  1. Whether and how the evidence has been tested;

  2. The error rate associated with the evidence;

  3. Whether the evidence has been peer-reviewed and published;

  4. Whether the evidence is generally accepted in the field from which it comes.[15]

Although commentators have bemoaned judges’ apparent reluctance to engage in a full Mohan-Daubert analysis,[16] there are several examples of courts excluding scientific evidence for want of a reliable foundation.[17]

Even if the evidence’s reliability is beyond reproach, courts increasingly exclude it as being unnecessary.[18] For instance, Canadian courts are loathe to admit expert evidence about the limitations of eyewitness memory, excluding such evidence on the grounds that it is unnecessary common sense information.[19] This issue arises when eyewitnesses give evidence—admitted as lay opinion evidence—identifying the accused at a crime scene.[20] To discredit the identification, the defence often attempts to call an expert to explain flaws in the identification, such as use of a biased or misleading photo line-up. Courts consistently find that such expert evidence is common knowledge, and thus inadmissible.[21]

Courts are also increasingly willing—but still very reluctant—to exclude expert evidence for bias when the expert appears to be acting as an advocate, rather than someone neutrally assisting the court. This position contrasts with a more liberal standard in which questions of partiality go to weight, rather than admissibility.[22] For example, in a very thoughtful decision of the Nova Scotia Provincial Court, Justice Ross excluded a police officer for bias.[23] This officer, tendered by the Crown as an expert on the practices of drug dealers, had been so intimately involved in the investigation, that he could not provide an objective opinion.[24]

In addition to these changes strengthening the substantive law of expert evidence, parties seeking to admit expert evidence also face additional procedural requirements. For example, parties seeking to tender expert evidence are typically held to a strict procedural timeline for serving their expert’s report, usually far in advance of the trial.[25] Further, Ontario and British Columbia have codified the duty of experts in their respective civil procedure rules.[26] A signed acknowledgement of the duty to assist the court in a fair and non-partisan manner within the expert’s area of expertise must be submitted along with the report.[27]

II. The Lay Opinion Rule and its Lost Moorings

Given the stricter approach to expert evidence, it has never been more important that courts police the border between lay and expert opinion. As we shall demonstrate below, the opposite has occurred. Lay and expert opinion are now nearly identical in the topics they relate to, the type of reasoning they engage in, and accordingly, the prejudices they present. In reviewing this trend, we begin by presenting the historic rule. We then trace its growth through two appellate cases, and a recent trial decision building on those cases.[28]

Prior to the Supreme Court’s decision in Graat, trial courts were much more dubious of lay opinion evidence. It was only admissible if it fell into one of several categories such as the identification of handwriting known to the witness, or the emotional state of a person.[29] Over time, courts and academics noted that lay opinion in practice was often not prejudicial and could assist the trier of fact. For instance, Justice Lederman and colleagues provide the example of an eyewitness identifying a man she saw robbing a store.[30] That identification is an opinion based on a multitude of facts, many of which stem from unconscious cognitive processes beyond the witness’s reckoning:

it is apparent that the witness’ testimony is an opinion or conclusion based on physical and perceptual sensations that he or she has experienced in the past and with which the witness has compared to his or her observations of the events in issue. To rule such testimony inadmissible because it constitutes an opinion would conflict with natural mental processes which translate a person’s perception into words.[31]

This is a sensible and pragmatic reason for admitting lay opinion with the lay evidence rule that tracks the scientifically founded limit on our ability to report on our cognitive processes.[32] At the same time, the system of carve-outs, based on elements such as identity, age, sobriety, was drawing increasing criticism for being overly technical and arbitrary. Together, these pressures brought about the Supreme Court’s foundational lay opinion decision—R. v. Graat.

Graat did away with the categorical approach and adopted a more flexible analysis.[33] Specifically, it indicated that lay opinion evidence is admissible when its probative value is not overcome by the prejudice it introduces.[34] In performing this analysis, the Court considered four factors:

(1) the witness has personal knowledge of observed facts; (2) the witness is in a better position than the trier of fact to draw the inference; (3) the witness has the necessary experiential capacity to draw the inference, that is, form the opinion; and (4) the opinion is a compendious mode of speaking and the witness could not as accurately, adequately and with reasonable facility describe the facts she or he is testifying about.[35]

Graat concerned the admissibility of two police officers’ opinion that the accused was intoxicated.[36] Justice Dickson, as he then was, applied the above approach to rule the officers’ opinions admissible. Specifically, he held that the officers were (1) making personal observations of Graat’s state of inebriation, and (2) were better positioned than the judge to do so.[37] Further, the opinion (3) did not require “scientific, technical, or specialized testimony.”[38] Finally, the opinion that a person is intoxicated (4) derives from a constellation of facts difficult to enunciate separately, and thus is more usefully provided as an opinion.[39]

Although the evidence was admitted, Justice Dickson issued an important and often overlooked caution.[40] He warned that when a police officer provides lay opinion, the judge or jury may be inclined to prefer that authority figure’s opinion over the evidence of a similarly placed witness. There was a hint of this phenomenon in Graat itself, where the accused’s civilian witness was cross-examined on not having the same level of experience identifying drunkenness as do police officers. To this point, Justice Dickson noted that police testimony deserves no extra weight because it is not based on expertise.[41]

Following Graat, which remains Canada’s leading lay opinion decision, courts have widely failed to engage with the doctrine it laid down, with some disputing that Graat places any restriction on lay opinion at all.[42] This incoherence undermines the practical benefits that Graat was meant to carry. For instance, failure to distinguish between lay and expert opinion by allowing a police officer to opine on the meaning of a bite mark required Supreme Court intervention and a new trial in R. v. A.(J.).[43] In the remainder of this section, we highlight three decisions that demonstrate the rapid expansion of lay opinion. Two are appellate decisions and the third is a very recent application of Graat, which expressly builds on the doctrine flowing from the two appellate decisions.

A. A Crime Scene Cleanup in R. v. Ilina

In R. v. Ilina, the Court of Appeal for Manitoba considered the conviction of a fifty-five-year-old university researcher for the murder of her husband.[44] The basis of the conviction was circumstantial: the deceased had been bludgeoned to death in the house he shared with his wife, the accused, and then dragged out onto the carport. Bloodstained clothes were found in the couple’s washing machine. The Crown argued the accused staged the scene to look like a robbery.[45] In support of this theory, two investigating officers, Sergeant Bell and Constable Rautavuori, testified that they had observed a clear, odourless stain surrounding the bloodstains inside the house, and opined that this indicated an attempted cleanup. The Crown’s blood analysis expert, Sergeant Maclean, disagreed. Based on photographs taken by Constable Rautavuori and on his own viewing of the crime scene two days later, the expert officer concluded there was no evidence of a clean-up.[46] On cross-examination, both Sergeant Bell and Constable Rautavuori said they would defer to Sergeant Maclean’s opinion. One of the primary issues on appeal was the admissibility of the two officers’ opinion, admitted at trial as lay opinion.

The Court of Appeal upheld the trial judge’s decision to admit Sergeant Bell and Constable Rautavuori’s opinion evidence. The court minimally engaged with lay opinion law, merely noting that Graat and subsequent cases had established a more permissive approach.[47] First, the Court of Appeal simply stated it would have been “difficult” for the non-expert officers to give their testimony as factual observations rather than as opinion. The court gave no support for this conclusion, and it is unclear where the complexity lies in simply stating that a bloodstain was situated adjacent to a clear, odourless, wet spot on the carpet. The Court of Appeal then noted that the testimony in question did not call for specialized knowledge, even though Sergeant Maclean had been qualified as an expert in this very matter.[48]

Admission of the lay officer’s testimony gave rise to difficulty with respect to the judge’s instructions to the jury. Despite assuring the defence counsel he would place the officers’ testimony in the proper context, he merely instructed the jury that the expert and lay crime scene testimony would have to be weighed against one another.[49] On review, the Court of Appeal expressed disapproval, suggesting the trial judge should have reminded the jury of the lay officers’ express deferral to the expert.[50] This misstep, however, did not merit appellate intervention.[51]

Ilina raises serious questions about the trial judge’s gatekeeping function. The officers were ostensibly comparing the case-specific facts of Ilina to a generalization about their experiences with similar crimes. The latter body of knowledge, however, was not evidence in court, and thus not open to cross-examination.[52] Furthermore, the jury could be expected to have trouble discounting the officers’ opinions because there is a specious connection between their testimony and their experiences as police officers. This danger is heightened in the following case, Lee, in which the witness was an expert, but not in the area in which he gave testimony. Lee demonstrates a closely related failure of gatekeeping that occurs when the witness is qualified as an expert in one area but gives evidence beyond that topic as a lay witness.[53]

B. Footsteps in the Snow in R. v. Lee

The accused in Lee was convicted by a jury of sexual assault in relation to an act of oral sex.[54] The only issue on appeal was whether the evidence gave rise to a reasonable doubt regarding consent. As is often the case, the credibility of the accused and complainant was central to this determination.[55] Both provided accounts marked by inconsistencies.[56] The testimony of a police dog handler, Sergeant Carriere, supported the complainant’s account. Sergeant Carriere was qualified as an expert at trial to give evidence about the reactions of his dog, but also gave evidence about the pattern and size of the footsteps he saw in the snow and what they indicated about the incident. The accused contended that Sergeant Carriere’s opinion about gait and footprint identification was inadmissible as lay opinion—being the proper subject matter of expert evidence, an expertise Sergeant Carriere did not possess. The Alberta Court of Appeal split two to one, with the majority upholding the conviction and finding the shoe print evidence admissible, while Justice Berger, in a forceful dissent, would have excluded the evidence.

Some context about the allegations is useful to understand the importance of Sergeant Carriere’s testimony. On the night of the incident, the accused, the complainant, and her friend got into an argument at a nightclub. Eventually, the friend left, and the complainant performed oral sex on the accused behind a nearby school. Shortly after, the complainant appeared at a nearby home of a stranger and reported a sexual assault. [57] Sergeant Carriere’s analysis of the footprints in the snow entering and leaving the area supported the complainant’s account.

Several aspects of Sergeant Carriere’s testimony are worth highlighting. First, the Crown attempted to buttress the sergeant’s opinion by emphasizing expertise that he had not been qualified for: “Is that part of your training, to notice when you have a chance to see footprints visibly and you notice something about them, are you trained to come to conclusions about that?”[58] Sergeant Carriere responded, detailing his experience beyond that of a dog expert to someone with an apparent expertise in interpreting the patterns left by shoe prints: “For sure. And over six and a half years it’s been my experience that . . . using that expertise is following footprints at times.”[59] Sergeant Carriere went on to rely on this expertise to provide a narrative of what, to a naïve reader, resembles a Rorschach-like pattern of shoe prints and other markings in the snow: “There was no set pattern in them, but you could see how the outer layer would be the small foot impressions and then the larger ones were, of course, trying to cut that person off.”[60] This account threatened the accused’s credibility and was relied on by the trial judge for that purpose.[61]

On appeal, the majority held that Sergeant Carriere’s opinion was admissible as lay opinion evidence. Notably, they did not engage in a substantive discussion of the requirements for lay opinion evidence, noting rather that there was no rule: “Whatever rule there may have been against a lay witness giving opinion evidence, it has not survived the decision in Graat.[62] Although they did not acknowledge the existence of a rule, the majority did state that Sergeant Carriere’s interpretation represented a common sense inference and a compendious statement of fact. To these two points, the majority relied on, without comment, just one modern Canadian case.[63]

Justice Berger, dissenting, concluded that the footprint analysis was not lay opinion, but rather expert opinion requiring scrutiny under the Mohan-Daubert test.[64] Interpreting the shoe prints required scientific or specialized expertise that Sergeant Carriere lacked.[65] Furthermore, the subject matter of Sergeant Carriere’s opinion did meet the lay opinion exception. His inferences were “not of the same character”[66] as the more typical lay opinions, such as those concerning a person’s age or mental state.

In the final case in this section, R. v. Colpitts, the trial judge relied heavily on Ilina and Lee to allow the evidence of a police officer who opined on various forensic accounting analyses. Colpitts puts on display the same dangers present in Ilina and Lee, while further expanding lay evidence rule by allowing the lay witness to testify beyond his own personal observations.

C. The Lay Forensic Accountant in R. v. Colpitts

In Colpitts, a lay opinion issue arose on the twenty-fourth day of what would prove a lengthy criminal fraud trial. The accused parties were “charged with unlawfully affecting the public market price of Knowledge House Incorporated (KHI)”.[67] On that twenty-fourth day, the defence applied to limit the evidence of one of the Crown’s key witnesses, a veteran police officer named Ian Black.[68] Officer Black intended to provide evidence from his investigation, which forensically linked the accused parties to the alleged crime.[69] The Crown had initially tendered Officer Black as an expert witness,[70] but resiled from that position, advising that “none of Mr. Black’s testimony will constitute expert opinion.”[71] This tactic proved successful: the evidence was admitted.

Officer Black’s opinion resulted from a series of analyses and inferences, ultimately connecting the accused parties to the illegal stock manipulation. First, Officer Black’s evidence affirmed the accuracy of an expert report prepared by a securities investigator.[72] To this end, Officer Black had gathered data on the trading of KHI stocks from “account statements and the Toronto Stock Exchange.”[73] After confirming the report’s accuracy, Officer Black compared its findings with the timing of emails sent by various suspects, including the accused parties. Based on the “strength of that comparison,”[74] he compiled a narrower group of suspects, the connected group, which included the accused. Finally, Officer Black performed analyses on that connected group. He compared their KHI trading activity with the rest of the market, and he analyzed the amount of late-day trading they engaged in using “established criteria”.[75]

The defence argued that Officer Black’s opinion was too complex to properly be the subject of lay opinion.[76] The Crown disagreed: he was merely synthesizing thousands of pages of seized documents and his testimony would therefore be of immense help to the court.[77] Justice Coady of the Nova Scotia Supreme Court ultimately agreed with the Crown. The evidence was admissible as a “compendious expression of technical data.”[78] In coming to his decision, Justice Coady relied primarily on Ilina and Lee.

Justice Coady found Ilina analogous because the officers in that case, as in Colpitts, had privileged access to the evidence in question: they saw the fluids before they dried.[79] Justice Coady also relied heavily on Lee. He did not draw any express analogy, but after excerpting Lee, he noted that “Justice Berger concluded the disputed evidence was neither expert evidence nor novel science,” but rather a common sense inference.[80] Though mistakenly referring to Justice Berger, who dissented in Lee, Justice Coady’s assessment of the majority’s decision is accurate.

Colpitts is emblematic of the hazards that flow from the rapid expansion of lay opinion doctrine since Graat. For one, unlike in Graat, Officer Black’s opinion went well beyond a judgment of intoxication to a complicated series of inferences requiring reference to accounting standards and statistical methodologies that would safeguard against the Officer’s subjective biases. Moreover, the Nova Scotia court did not recognize the danger posed by Officer Black’s status as an authority figure with no proven qualifications in accounting. Finally, Ilina and Lee are dubious authorities for the admission of Black’s evidence. While Officer Black indeed had access to data from the alleged offences, the nature of that access was vastly different from that in Ilina, wherein evidence became inaccessible as fluids dried over time. In contrast, Officer Black was analyzing permanently stored data, much of it gathered by other investigators and thus not gleaned from his own personal observations. We will now delve more deeply into these hazards.

III. The Hazards of Lay Experts

In this section, we will explore three hazards that flow from the rapid expansion of lay evidence documented above. The first hazard is unconscious bias and the fact that lay witnesses need not demonstrate they relied on a methodology to control that bias. A lack of methodology would not be so problematic if not for the second hazard, a specious nexus between the witness’s authority and his or her evidence. A trier of fact might be expected to discount a lay person’s evidence if it were brought by a true lay person. As seen in the last section, however, lay opinion evidence is often brought by police officers with experience in fields only speciously related to the evidence. Thirdly, many of the issues we found are attributable to a jurisprudence that has developed incoherently, with little heed paid to cases with similar facts and the general principles of evidence law. We will address these hazards in order, beginning with unconscious bias and the importance of methodology in controlling for that bias.

A. Methodology versus Qualia

The first principle is that you must not fool yourself—and you are the easiest person to fool.[81]

As we described in Part III, the justification for admitting lay opinion is one of necessity: the facts underlying these opinions are unconscious mental processes that cannot be put into words.[82] Justice Dickson confirmed this interpretation in Graat, adopting a passage from Cross on Evidence stating that lay opinion should be admissible when that opinion was made “without conscious ratiocination”.[83] In this subsection, we will first demonstrate that this justification accords with the current psychological scientific understanding of unconscious cognitive and affective processes, which finds they are indeed unverbalizable. However, these unconscious processes are often biased in ways we cannot know or report on. We will document several forms of unconscious bias that may have infected the opinions in Ilina, Lee, and Colpitts. We then suggest that in cases like these, where methodologies exist to control for unconscious bias, there is danger in exempting witnesses from a legal standard that would demand they follow such procedures.

Much of the current research on unconscious thought was inspired by Richard Nisbett and Timothy Wilson, who found that humans cannot report on these processes.[84] Nisbett and Wilson made their landmark finding in the context of judgment and decision-making.[85] They asked consumers to choose between several sets of stockings and justify their choice. The most popular choice was the set on the far right, with most consumers giving persuasive reasons, such as its softness and superior knit. The rub was that all of the stockings were the same: previous studies had shown an unconscious bias toward items on the consumer’s right (likely driven by the custom of shopping from left to right).[86] The participants’ reasons could not have been accurate and were rather post hoc justifications for their unconscious bias.

If so much of judgment and decision-making is girded by unverbalizable (i.e. unconscious) cognitive processes,[87] what types of opinions deserve the lay opinion rule’s concession to unconscious ratiocination? We suggest that the unconscious ratiocination justification holds true for many of the forms of lay opinion identified in Graat, for example eyewitness identifications, judgments of speed, and distance.[88] But the opinions in Lee, Ilina, and Colpitts are distinguishable because they sit in fields, in which methodologies exist to control for unconscious biases. We do not mean to imply that lay witnesses giving identifications or judgments of speed are not biased. In fact, they very likely are biased in ways they cannot know.[89] In fact, despite most eyewitnesses’ best intentions and their own confidence, eyewitness identifications are demonstrably unreliable and account for 79 % of wrongful convictions.[90] Similarly, we fully believe the police officers in Ilina, Lee, and Colpitts had every intention to honestly and accurately fulfill their legal duty. The distinction is that those police officers were giving evidence on matters that can be subjected to scientific methodologies.

Methodologies, unlike qualia, are transparent, verbalizable, and thus open to scrutiny. In this spirit, Richard Feynman’s quote, which began this subsection, is often related to scientists to encourage them to continuously question their own practices and adhere to rigorous methodology.[91] In the remainder of this subsection, we discuss some biases that may have affected the judgment of the witnesses in Ilina, Lee, and Colpitts. In all these cases, the witnesses’ opinions concerned fields governed by varying degrees of methodological principle and thus, methodology should have been demanded. Or, at the very least, they should have been subjected to a rule of evidence that would require such questions be asked.

Two unconscious biases can deeply influence opinion evidence in ways the witness is unaware of: confirmation bias and contextual bias. Confirmation bias describes the process by which we force incoming information to fit a pre-existing theory we have about the world, for example the Earth is the centre of the galaxy, or my in-laws are a pain-in-the-neck. More specifically, it is “the seeking or interpreting of evidence in ways that are partial to existing beliefs.”[92] Contextual bias, on the other hand, has less of a motivational flavour, and is simply “where individuals are influenced by irrelevant background information.”[93] For instance, with respect to contextual bias, Itiel Dror and colleagues have found that exposure to emotional case details can bias fingerprint matching judgments.[94]

Both effects are pervasive, influencing the questions we ask, how we ask them, and the weight we attribute to the data we collect to answer those questions. Very importantly, we are typically unaware that our judgment is being influenced.[95] A recent leading report of a committee of the National Academies of Sciences found that several forensic sciences do not properly control for confirmation and contextual bias.[96]

Various methodological safeguards can help counter confirmation and contextual bias.[97] One of the most fundamental methodological safeguards is “blinding” the technician to the source of the sample. For instance, a 2016 report from the U.S. President’s Council of Advisors on Science and Technology recommended that fingerprint analysts avoid knowledge of other facts of the case and disclose when they have such knowledge.[98] Inadequate blinding was implicated as a source of a high-profile misidentification of the perpetrator of a 2004 bombing in Madrid. In that case, an FBI examiner had reported a 100 % certainty of a fingerprint match.[99]

More recently, other safeguards have gained traction, such as making procedures open to scrutiny and pre-defining a methodology so that it cannot be altered during the analysis to favour one conclusion over another.[100] In this vein, and also in the context of fingerprints, the PCAST Report suggested analysts should pre-commit—in a manner that cannot be changed after the fact—to the features of the found fingerprint, the “latent” fingerprint, that will be matched to the known fingerprint.[101] This method helps safeguard against “fishing expeditions” driven by a preconceived notion that there should be a match to a certain known fingerprint.

Sergeant Carriere’s testimony in Lee admitted of many opportunities for confirmation and contextual bias. As lay testimony, it was not held to a reliability standard that would have demanded the Crown to address such issues. Sergeant Carriere interpreted a complex set of footprints and other markings “with no set pattern” across a wide area. He somehow extracted a very specific narrative of a pursuer cutting off another, a fall, and then an apparent running getaway.[102] Moreover, as part of the investigating team, he was not blind to Lee’s identity, the Crown’s theory of the case, or other emotionally evocative facts about the accusation.[103]

Pattern-matching engages one of the key unconscious processes underlying confirmation bias, what psychologists call meaning maintenance.[104] Meaning maintenance researchers find that humans readily seek and find patterns in noise, and then ascribe a narrative to these patterns. The term for this phenomenon is apophenia, or the “tendency to see patterns in random data.”[105] In one study, Daniel Randles and colleagues found that participants experiencing a feeling of meaninglessness were more apt to find patterns in random strings of letters.[106] This process is fast and unconscious, with a lack of meaning causing discomfort and resolution proving satisfying.[107]

Even experts in the field on which Sergeant Carriere was opining have been criticized for lack of methodological rigour. Indeed, the NAS Report warned against conclusions from shoeprint analyses because it is impossible to know, over time, how often the expert is right or wrong: “[I]t is difficult to avoid biases in experience-based judgments, especially in the absence of a feedback mechanism to correct an erroneous judgment.”[108]

For these types of subjective judgments like the above, the PCAST Report suggests black box studies to determine if the methodology is itself valid.[109] The term “black box” refers to the black box of the examiner’s mind where, without objective standards, it is impossible to know exactly how the examiner is coming to his or her conclusion. Black box studies would then expose a sample of examiners to patterns in which the ground truth is known, for example if, in fact, a struggle between two people produced the pattern of footprints, and determine if examiners can discover this truth and with what error rate. If the methodology is found to be valid, individual examiners should be “proficiency tested” periodically to determine if they can reliably apply this method.[110] Characterizing Sergeant Carriere as a lay witness sidesteps any questions regarding the validity of his method.

Consider also a related manifestation of unconscious bias—the illusory correlation: a “tendency of individuals to perceive statistical associations that are objectively absent, or at least to perceive statistical associations that are more pronounced than objectively exist.”[111] As the NAS Report explained, it “can lead one to formulate overly simple models of reality and thus to read too much significance into coincidences and surprises. More generally, human intuition is not a good substitute for careful reasoning when probabilities are concerned.” [112]

For example, there is a widely held illusory correlation between arthritic pain and the weather.[113] Illusory correlations are perpetuated by various mechanisms, such as generalizing from small sample sizes, spurious coexistence of distinctive events, and over-attendance to confirmatory data, that is confirmation bias. In the pain-weather example, a person has a lifetime of data from which to pick out distinctive instances of the correlation (e.g., “that time it rained all week on vacation and my knees ached.”) Widespread belief in the myth makes the pain sufferer especially attuned to confirmatory instances, whereas disconfirmatory instances, such as the absence of pain, are less salient and more easily forgotten. In the casino and the stock market domain, illusory correlations have been termed the “gambler’s fallacy”,[114] to reflect the manner in which individuals read patterns into random winning and losing streaks.[115]

Scott Lilienfeld and colleagues have noted that forensic situations are particularly fertile grounds for illusory correlations because of the lack of objective benchmarks and salience of confirmatory data. In other words, the instances in which the technique uncovers the criminal, and everyone celebrates are more salient than inconclusive cases.[116] This is especially true in equities trading, where there is a great deal of data from which to find spurious patterns and produce the gambler’s fallacy. Paul Slovic, whose work has informed much of the modern approach to behavioural economics, made precisely this observation in explaining why myths about being able to beat the market persist.[117] Subsequent empirical evidence supported his hypothesis.[118]

Similarly, Officer Black’s opinion in Colpitts could be entirely based on a cocktail of the gambler’s fallacy, illusory correlation, and confirmation bias. Recall that he compared the timing of the accused parties’ trades and emails with broader market activity and concluded they were involved in manipulating the stock price. However, sifting through the trade data, there may be many instances when the timing of their trades deviated from the market average purely by chance. As with the example of arthritic pain, the trades were particularly salient to the witness. Further, Officer Black entered into a long and difficult investigation focused on the accused parties, who were notorious Halifax businessmen and legal professionals. It would have been difficult for him to shield himself from contextual and confirmation bias.

Scientists attempt to combat illusory correlations with inferential statistical methodologies. In fact, this was a key component of the arthritic pain and weather study referenced above. The authors, Donald Redelmeier and Amos Tversky, first established that there was no statistically significant correlation between the weather and joint pain.[119] Similarly, inferential statistics could have helped determine the unlikeliness of the timing of the impugned trades and emails in Colpitts, or if they were consistent with chance. And even these statistical tests would have had to be performed with a high degree of rigour. For instance, statistician Adriaan De Groot describes the mistake even seasoned statisticians can make when they sift through a large dataset:

he will be on his guard for spurious correlations; but nevertheless he still attempts, by means of a procedure that consists of searching, trying, and selecting, to “extract from the material what is in it”. Of course, this means that he will also extract that which is in there accidentally.[120]

For example, imagine flipping a coin hundred times without predefining what you will accept as evidence the coin is not fair. Even if, in total, the experiment returned fifty heads and fifty tails, the experimenter might say: “well, for flips 10–19, there were 8 heads and 2 tails; this is evidence it’s not a fair coin.” In other words, if the researcher does not predefine what he or she is looking for—as fingerprint analysts must do in light of the PCAST recommendations reviewed above––the results lose much of their probative value.

It is unclear whether Officer Black performed these tests and with what degree of rigour. Perhaps more importantly, the lay opinion rule did and does not require he be qualified to do any of this. And, at the risk of sounding like a broken record, the lay test does not provide an opportunity to exclude Officer Black on this basis. By contrast, forensic accountants, qualified as experts, must perform what is known as an econometric analysis, which determines if the event at issue in the case affected a market price, statistically controlling for other possible effects.[121]

Similar challenges are present in Ilina. The officers opining that there was evidence of a cover-up were basing their conclusions on their experience with examples that readily came to mind.[122] Limited in their qualifications, they had no way of knowing how often a bloodstain appears next to a different stain and how often there is an innocent explanation. Their conclusion was particularly troubling when compared to the contrary opinion of the crime scene investigator, who would be expected to see more investigations through to the end, and thus know more about the accuracy of the correlation. In fact, researchers have found that feedback about the accuracy of one’s judgments is a key component in building expertise.[123]

Finally, and to bring the discussion full circle, Kathleen Kennedy and Emily Pronin describe the “bias blind spot”, or in other words, the tendency for all the unconscious biases discussed herein to go unnoticed.[124] Indeed, as we began this subsection, it is easy to fool ourselves.[125] Again, we do not think any of the police witnesses in the cases we have highlighted intentionally misled their respective judges and juries. Rather, as Nisbett and Wilson demonstrated, despite our best efforts, we cannot report on how we are being unconsciously influenced. The bias blind spot renders the police experts difficult to cross-examine and unduly persuasive because they can confidently say they believe their account is accurate.[126] Indeed, empirical research finds that jurors and judges are swayed by confidence. Witness confidence accounts for 50 % of the variance in jury decisions about the believability of eyewitnesses,[127] and judges tend to prefer confident experts.[128] In the cases we described, confidence may be diagnostic of nothing but blithe inadvertence to unconscious bias.

It should be no surprise at this point that unconscious biases have been at the heart of many prominent miscarriages of justice.[129] The commissions and researchers who have studied these miscarriages of justice unanimously recommend stricter adherence to methodologies designed to combat unconscious biases. Methodologies transform qualia into testable assumptions.[130] It is possible to evaluate a methodology’s reliability by observing its error rate: how many false positives and negatives does it produce?[131] It is also possible to cross-examine a methodology for bias. For instance, does the methodology contain safeguards for unconscious bias, such as keeping the experimenter blind to whether the sample came from a control or experimental group?[132] The National Academy of Sciences recognized the value of methodology in minimizing unconscious bias:

All of these sources of bias are well known in science, and a large amount of effort has been devoted to understanding and mitigating them. The goal is to make scientific investigations as objective as possible so the results do not depend on the investigator. [133]

Without methodology, the putative lay testimony in Ilina, Lee, and Colpitts was ipse dixit and, in fact, the worst kind of ipse dixit: the witnesses were not qualified to begin with.

B. Specious Nexus between Authority and Evidence

Experience in Canada and elsewhere teaches that wrongful convictions are often traceable to evidence that is either unreliable or prejudicial. When the two combine, they make for a potent mix — and the risk of a wrongful conviction increases accordingly. Wrongful convictions are a blight on our justice system and we must take reasonable steps to prevent them before they occur.[134]

If a true lay person had brought the evidence in the cases described above, it would have been easier for the judge or jury to discount it. For instance, and at the risk of alienating a substantial portion of our audience, what if a local law professor had passed by the putative crime scene in Lee and observed the impressions in the snow? This professor might have come to a similar conclusion with the same lack of training and methodology. But with the professor, it would have been clearer to the trier of fact that this was a mere compendious statement of fact rather than an expert opinion.

The potential impact of Sergeant Carriere’s opinion on the jury in Lee was wildly different. He was an authority figure in a closely related field, that is the behaviour of tracking dogs. The jury might have easily jumped to the conclusion that he was an expert in the field at issue, despite the fact he was expressly not qualified to make such conclusions.[135] Cognitive scientists have long warned about generalizing expertise from one domain to a closely related one: “Critically, superior performance in a particular domain does not guarantee superior performance in another, even when the domains seem similar.”[136] Moreover, police officers receive training in courtroom presentation and are often persuasive and confident witnesses.[137] The result, as the Supreme Court warned in the quote from Hart above, is a volatile mix of unreliability and prejudice.

Safeguards surrounding expert evidence emerged to manage precisely the risk of having the trier of fact unduly swayed by the mere authority of the witness.[138] This authority has been described in various ways, from “impressive credentials”,[139] to the “mystique of science”[140] when the witness is a scientist, and the “superficial attractiveness”[141] of police witnesses. And over the past several years, both in Canada and in the United States, the rules have tightened in response to such arguments from authority causing demonstrable miscarriages of justice.[142]

There is scientific support for this caution: research demonstrates that people lacking knowledge in a given field attorn to the opinions of experts. In other words, people focus on the extrinsic qualities of the message, such as the qualifications of the speaker and how confident he or she seems, as opposed to the content of the message.[143] Such shortcuts may work in some cases, but they are especially problematic when there is a specious nexus between the qualifications of the speaker and the topic he or she is opining on. Further, unconscious and unverbalizable biases give rise to overconfident witnesses, who in turn may “usurp the functions of the trier of fact.”[144]

Ilina, Lee, and Colpitts all present striking examples of the specious nexus between authority and evidence. As noted, Lee is especially problematic because the police officer wore two hats in closely related fields: dog handling, and shoe print and gait analysis. A juror would have understandable difficulty in distinguishing between the police officer’s expert evidence in one field and lay evidence in the other.

The lay officers in Ilina pose a similar problem. Jurors were likely influenced by the fact that Sergeant Bell and Constable Rautavuori were police officers who had surely observed several crime scenes, presenting what Justice Goudge called “authoritative claims based largely on personal experience.”[145] But lay opinion law does not demand an inquiry into witnesses’ qualifications and reliability. The trial judge struggled with this very issue in his jury instructions, attempting to explain how the jury should balance the evidence of the three police officers, two in their capacity as mere observers and one, as an expert. And recall, the Court of Appeal disapproved of the trial judge’s approach to instructing the jury but ultimately found, it was not a serious enough error to order a new trial.[146]

Finally, Colpitts is also problematic. In Colpitts, Officer Black, was a senior police officer with over twenty years of experience investigating commercial crimes.[147] The Crown, in fact, originally sought to qualify him as an expert. Having been deeply involved in the investigation, Officer Black could speak knowledgably and confidently about it. In short, these factors likely made him a persuasive witness. They did not, however, demonstrate his ability to perform inferential statistics to determine the likelihood that the accused’s trading activity meaningfully differed from the market as a whole and in relation to forensic accounting standards.

Courts have largely failed to recognize the specious nexus when admitting lay evidence, but there is a basis in the jurisprudence for caution in such cases. Justice Dickson recognized it in Graat, warning that an officer’s opinion of a person being intoxicated could easily be overweighed by a jury.[148] Similarly, at the admission stage, the British Columbia Court of Appeal in R. v. Ratté[149] used a contextual approach to exclude the lay opinion of a police officer investigating the murder of the accused’s wife.[150] The officer’s testimony summarized her lengthy investigation but also included a conclusory opinion regarding the whereabouts of the wife. In holding that this lay opinion should have been excluded, the Court of Appeal factored in the prejudicial impact of the witness’s status as police officer and her experience in the field.[151] Ratté also provides an example of how the jurisprudential approach to lay opinion has failed (i.e., carefully developing the boundaries of admissible lay opinion through the case law). It analogizes closely to Colpitts but was not referenced in that case. We will return to this issue in the following section.

Overall, when lay opinion strays from a summary of unverbalizable qualia to authoritative claims about the conclusion of a police investigation, it is nothing more than the ipse dixit warned about in countless expert evidence cases. Furthermore, it is an especially pernicious variety of ipse dixit because the witness’s experience may have nothing to do with the area on which he or she is opining. This phenomenon has been made possible by a fractured and incoherent lay opinion jurisprudence.

C. Failures in the Lay Opinion Jurisprudence

Graat adopted a more flexible and principled approach to lay opinion, which has, over time, resulted in the very uncertainty the Court sought to avoid. In many ways, the decision was a logical correction to the previous rules—a system of carve outs—which negatively impacted judicial economy,[152] widely diverged across provinces, and created a great deal of uncertainty.[153] Principled approaches, however, can devolve into unstructured exercises in judicial discretion.[154] This appears to be the case with the post-Graat lay opinion rule, in reference to which the majority of judges of the Alberta Court of Appeal, in Lee, said: “Whatever rule there may have been against a lay witness giving opinion evidence, it has not survived the decision in Graat.”[155]

Indeed, Lee may be the most egregious of the cases we reviewed in its failure to advance the law incrementally. Recall that in Lee, the police officer developed a narrative to explain a complex pattern in the snow. The majority in Lee only analogized to one modern Canadian case, Powell, in which the police officer giving lay opinion testified that a footprint and the shoe appeared to be of the same size and tread pattern.[156] To make matters worse, Powell only relied on one authority, a trial decision admitting the lay opinion of two civilians that a truck driver was intoxicated, that is a Graat-type fact pattern.[157] These leaps are vast: in just two cases, permissible lay opinion in Alberta expanded from the assessment of drunkenness by a civilian to a series of inferences about a complicated pattern by an authority figure. Neither court provided an explanation or principled justification for the expansion.

Colpitts then relied on Lee and Ilina for the general proposition that, “[g]iven the developments in the law”,[158] “police officers may give opinion evidence on observations made during their investigations.”[159] However, Officer Black in Colpitts was not recounting a firsthand experience, as was the case in Graat, Ilina, and Lee. Rather, he was analyzing data gathered by others and, in fact, affirming the accuracy of an expert report.[160] His judgment called for accounting procedures and standards that were not in evidence. These facts not only diverge from the precedents in Lee and Ilina, but also seem to run afoul of Graat’s personal observation factor, which states that the witness should have personally observed the case-specific facts.[161] As we will detail below, several other decisions—none relied on in Ilina, Lee ,or Colpitts—have excluded lay opinion for this failure of personal observation.[162] This inadvertence represents a second failure of the lay opinion jurisprudence: disregarding of relevant authority.

The parties in Colpitts may not have drawn Justice Coady’s attention to Toronto Dominion Bank v. Cambridge Leasing Ltd.,[163] which likely represents its closest factual analogy. In TD, the New Brunswick Court of Queen’s Bench considered whether a financial projection could be the appropriate subject matter of lay opinion. The court ultimately excluded the evidence on the basis of Graat’s principle that the evidence be a matter of the witness’s personal experience.[164] A projection, by its prospective nature, cannot be a matter of personal experience.[165]

TD is one of several cases that have drawn a harder line at Graat’s personal knowledge principle. For instance, the Court of Appeal for Ontario in Marchand excluded a nurse’s lay evidence because it deviated from her own experience and instead delved into the actions of another nurse and the standard of care for nurses.[166] The court in American Creek relied on Marchand for precisely this point.[167] American Creek concerned a breach of contract claim in the mining industry. The parties disputed whether certain expenses were properly characterized as exploration costs. The impugned lay witness would have opined on whether drilling patterns indicated exploration or another purpose. The court characterized the testimony as specialized knowledge rather than lay opinion because it was not a summary of the witness’s own observations:

The evidence his counsel seeks to adduce does not consist of everyday inferences from observed facts which he, as the observer, was in a better position to make than I am because he was there. Rather, his evidence will concern matters of specialized, technical expertise upon which he proposes to comment on the basis of his review of documentation and reports, taking into account his own experience.[168]

TD, Marchand, and American Creek all bear a strong similarity to the facts in Colpit, with witnesses analyzing data relating to occurrences they did not personally observe. Furthermore, American Creek being an appellate decision, it should carry as much weight in Nova Scotia as the extra-provincial appellate decisions relied on in Colpitts.[169] In any event, we are surprised the court in Colpitts did not recognize an earlier—high profile—decision of the very same court. In that case, Elwin v. Nova Scotia Home for Coloured Children, the Nova Scotia Supreme Court excluded evidence from an orphanage’s director because it pertained to financial information outside of her personal experience.[170]

Stricter adherence to Graat’s “personal observation” factor, besides finding purchase in the cases above, correlates with recent American academic work. There, Edward Imwinkelried has lucidly detailed a key distinction between lay and expert opinion.[171] First, he notes that both varieties of opinion concern a comparison between a generalization and case-specific facts. For instance, a lay handwriting identification witness has viewed the defendant’s handwriting many times before (i.e., the generalization) and then compares that to a contract in evidence at trial (i.e., the case-specific fact). Similarly, an expert psychologist compares a body of knowledge about a disease (i.e., the generalization) to the symptoms of the plaintiff (i.e., the case specific fact).

The distinction, according to Imwinkelried, is how the generalization and case-specific facts are acquired.[172] The expert may, and is often expected to, go beyond his or her personal observations in coming to a generalization. Rather than remaining confined to personal experience, the expert draws from a body of knowledge, most or all of it discovered by other experts. The lay witness, on the other hand, is expected to stay within his or her personal observations. The same distinction holds for case-specific facts: the expert may also go beyond personal observations, to considering hypotheticals and second-hand reports, whereas the lay witness is restricted to personal observations.[173] We will revisit this analysis in Part V, in which we propose a revised framework for admitting lay opinion evidence into court.

Finally, a stronger jurisprudential approach should attempt to align expert and lay opinion case law. In other words, if the impugned evidence has typically been admitted as expert evidence, then according to Mohan, this indicates courts have found that expertise is necessary to assist the trier of fact on those topics.[174] How then could the same evidence be brought by laypeople in the cases we surveyed? Moreover, recall that Graat distinguished between lay opinion and cases where “scientific, technical, or specialized testimony” was necessary.[175] Surprisingly, this type of analysis is uncommon in the lay opinion jurisprudence.[176]

Had the majority in Lee considered the case law involving gait[177] and shoeprint[178] analysis, they would have found Justice Berger was correct. These situations are regularly the subject matter of expert opinion. Similarly, forensic accountants and securities investigators are frequently qualified as experts in prosecutions like that in Colpitts.[179] And crime scene investigators are called as experts to provide the forensic evidence of the sort given by the two lay police officers in Ilina.[180]

This confusion between lay and expert topics was particularly stark in Ilina and Colpitts, wherein lay and expert witnesses provided substantively the same evidence within the same trial. In Ilina, the Crown’s own crime scene expert gave evidence that contradicted that of the lay officers, a phenomenon that proved problematic when the judge instructed the jurors about how to weigh the three opinions.[181] This was especially problematic for the accused because the lone expert opinion, which was favourable to her case, faced two speciously lay opinions. And towards the end of the Colpitts trial, a battle was fought over qualifying an expert securities investigator to opine on the same subject matter as Officer Black,[182] with the defence arguing that the expert’s opinion was unreliable because he uncritically accepted Officer Black’s lay opinion.[183] The experience in Colpitts therefore demonstrates both the inconsistency in defining lay opinion and the unfairness this inconsistency can cause. The evidence was admitted over the defence’s protestations, allowing the possibility that unreliable lay opinion could provide the foundation for expert evidence.

Failure to enforce the limits of lay opinion evidence also undermines expert evidence law. As the courts in Abbey, Sekhon, and Bingley held, it is incumbent on the trial judge to define the nature and scope of expert opinion before assessing admissibility.[184] Allowing expert opinion to be admitted as lay opinion evidence renders this rule ineffective because evidence beyond the scope could simply be admitted as lay opinion. Indeed, the evidence in the cases we discussed would be vulnerable to exclusion under the expert rules.

If applied, Mohan’s properly qualified expert requirement would likely exclude some of the cases we have discussed. The lay witnesses in Ilina and Lee were law enforcement officers with some on-the-job experience in the areas they were opining on. This, however, may not be enough expertise under Mohan, which requires not just experience but specialized knowledge. For instance, general medical practitioners are excluded from opining on specialized disciplines.[185] The same holds for police officers. In the retrial of R. v. Abbey, for instance, a police officer with extensive experience investigating gangs was not permitted to testify about the meaning of a particular tattoo in gang culture because his knowledge in that specialized area was merely anecdotal.[186] Similarly, the police witnesses in Ilina and Lee had only limited experience in the areas they opined on. Indeed, this may be why the Crown elected to call them as lay rather than expert witnesses—they would have failed the Mohan test.

The police witnesses in Ilina, Lee, and Colpitts would also have faced potential exclusion for bias under the expert rules.[187] Sergeant Carriere in Lee and Officer Black in Colpitts would have been especially at risk because of their extensive investigatory involvement. For instance, in Tremblett and Kovats, deep involvement in an investigation was found to raise an apprehension of bias such that the officers were excluded.[188] Still, exclusion for bias remains a high bar when it comes to police witnesses.[189] The mere risk, however, that the witness would be excluded for bias might motivate investigating teams to control for unconscious bias by keeping the potential witness blind to contextual features that might impact his or her analysis. Therefore, there is still deterrence value in subjecting the witness to a rule of evidence that is sensitive to bias.

With regard to reliability, none of the lay witnesses in Ilina, Lee, or Colpitts seems to have taken any methodological measures to establish their reliability and thus remove or measure the unconscious biases we documented above. As a result, they may have been excluded if they were called as experts because expert witnesses must provide a reliable basis for their opinions.[190] With regard to forensic scientific evidence, the party proffering it must demonstrate that the methodology has been tested and possesses a low error rate.[191] Specialized knowledge must also be reliable but is subject to a more flexible inquiry into whether the expert followed the appropriate standards in the field.[192] For example, in Myles, a police officer’s expert evidence about gangs was excluded because he could not explain his methodology.[193]

Finally, expert evidence is subject to a discretionary gatekeeping analysis in which its costs and benefits are evaluated.[194] The evidence at issue in the cases we have reviewed would also likely run into trouble at this stage. In particular, the consumption of court time and prejudice involved in the witnesses providing duplicative evidence in Ilina and Colpitts would militate toward exclusion.[195]

IV. Lay Opinion Evidence, The Next Generation

Courts and academics have devoted a great deal of thought on how expert evidence should be harnessed to assist the trier of fact, while simultaneously avoiding undue prejudice. This work has already demonstrated its value,[196] with several instances of unreliable evidence excluded from reaching the ultimate decision.[197] It is time to devote some attention to lay opinion, which poses its own practical and epistemic challenges, thus requiring deeper consideration than it has been given in the years since Graat. In this section, we will map out a path forward for lay opinion evidence law and suggest where that line may be drawn. We will begin with principles and then lay out a framework for the admission of lay opinion evidence.

We are not suggesting a drastic change in Supreme Court jurisprudence or the introduction of new legislation. Rather, the root of the problem, as we noted in our discussion of the failed jurisprudential approach, is disagreement between courts about what exactly Graat stands for. For instance, consider the contrast between the majority in Lee’s statement that there is no rule,[198] with a 2016 statement from an Ontario trial court: “However, although the court in Graat eased considerably the rule against lay opinions, it did not discard it entirely.”[199] These approaches not only contradict each other, but are also difficult to reconcile with decisions that fully engage with all four of the factors[200] that Justice Dickson considered in Graat.[201]

In fact, many of the hazards we identified flow from a simple failure to fully consider the existing authority. For instance, Justice Dickson distinguished between the lay opinion found in the facts of Graat and opinions that derive from specialized and scientific knowledge.[202] Further, his caveat about police officers serving as lay witnesses finds a home in our discussion of the specious nexus. In light of the Supreme Court’s recent warning about the injustices that flow from the “potent mix”[203] of unreliability and prejudice, it may be time to elevate Justice Dickson’s caveat to a more prominent place in lay opinion doctrine.

Finally, the “personal knowledge of observed facts” component of Graat accords with Imwinkelried’s suggestion that lay and expert witnesses primarily differ in whether witnesses are permitted to base their reasoning on information beyond their firsthand observations. There is an expectation that experts are standing on the shoulders of giants in developing their opinions; indeed, that is a large part of their value. The same cannot be said for lay witnesses whose function is to provide an opinion about what they personally observed.[204]

This failure to adhere to Graat in distinguishing between lay and expert opinion disproportionately affects the criminally accused. In the domain of expert evidence law, researchers have noted that failures to gatekeep forensic science unduly burden the accused, who are ill-equipped to rebut unreliable forensics.[205] And, ironically, trial judges reserve their strictest reliability scrutiny not for forensic evidence led by the prosecution, but for evidence led by plaintiffs in civil trials.[206] This pattern seems to hold for lay evidence as well. Of the cases we found, in which lay opinion was excluded, the majority were from the civil arena.[207] And as we noted in our introduction, an extreme imbalance can occur when the Crown’s police witnesses are admitted under the lay rules. In such cases, there is no one the accused can call on to rebut that evidence other than, potentially, other police officers present at the scene of the crime. There is little recourse other than hiring an expensive expert and attempting to qualify him or her under the relatively onerous Daubert-Mohan test.[208]

To counter this imbalance in the expert sphere, some researchers have prescribed a contextual approach informed by criminal justice principles such as the presumption of innocence.[209] In our view, courts should also take a contextual approach to lay opinion, especially when it so closely resembles expert opinion. For instance, in Lee, Sergeant Carriere’s lay opinion about the meaning of the footsteps in the snow would be very difficult for the accused to rebut without an expert. Graat itself did not seem to foresee situations like Lee because in Graat, the accused had two lay witnesses of his own who also made personal observations. In cases like Lee, we suggest courts should be contextually sensitive and thus loath to admit the Crown’s witness.

Courts should be similarly sensitive to the role the evidence plays in the totality of the case.[210] In fact, there is precedent for this approach when police witnesses are being admitted as experts. In Van Bree, for example, Justice Annis excluded portions of a police officer’s opinion for bias because the officer’s evidence was so central to the case.[211] This represents a thoughtful approach that is live to the prejudices flowing from police testimony. There is no reason the lay opinion test should not be as contextually sensitive.

A. Lay Opinion 2.0

Drawing on the above discussion, we will now offer our vision of an improved framework for trial judges to follow in deciding to admit or exclude lay opinion. Our goal is to retain Graat’s principles and flexibility by overlaying a structured approach on top of this leading case. We drew inspiration from Justice Doherty’s two-step reconceptualization of the Mohan test in Abbey, which gained acceptance amongst trial judges across Canada and was eventually adopted by the Supreme Court.[212]

Lay Opinion 2.0 structures the admissibility inquiry by asking key questions to ground the reasoning and identify the issues, and then builds on those insights in the final discretionary balancing stage. It follows four steps:

  1. Identification;

  2. Characterization of the reasoning process and witness;

  3. Categorical exclusion of opinion that is not the result of the witness’s own observations;

  4. Discretionary gatekeeping of unduly prejudicial evidence.

Evidence failing our test should either be excluded or resubmitted as expert evidence.

1. Step 1: Identification

First, the trial judge should identify “the generalization the ... witness is relying on ... the case-specific fact or facts” and the comparison between those elements.[213] For example, in Ilina, the generalization was the officers’ previous viewings of crime scenes with clear fluids adjacent to blood. The case-specific fact was the observation of the crime scene. The comparison was the judgment of similarity between the previous crime scenes and the instant one. This identification step is an opportunity for precision, something that is notoriously absent from the current approach. It is also integral to the next stage, characterization.[214]

2. Step 2: Characterization

Second, the trial judge should characterize both the elements from step 1 and the witness providing the opinion. This characterization exercise will typically map onto the following questions.

  1. Personal or Second-hand Observation: Are the elements identified at step 1 based on personal observations? In Ilina, if the officers had indeed been making their own observations instead of relying on secondhand anecdotes, then their generalizations were personal observations. However, it may be that they had never seen other similar crime scenes, but that they were relying on anecdotes relayed by other police officers. In such cases, their opinions were no more than guesses lacking any foundation and would be excluded in Step 3. Similarly, in Colpitts, one of the case-specific facts was a report that Officer Black did not prepare, and the comparative exercise was expressly guided by professional standards, rather than Officer Black’s own judgment.[215]

  2. Methodology or Qualia: Are the elements of the opinion subject to professional or scientific methodology? Drawing from the case law on expert evidence and numerous syntheses of it, the trial judge can determine if the opinion at issue is one typically given by experts and guided by methodology.[216] Further, what are the unconscious biases that these methodologies seek to control? Alternatively, is the judgment predominantly based on unconscious processes? Here, eyewitness identifications founded on qualia clearly contrast with the forensic accounting in Colpitts and the footprint analysis in Lee.

  3. Source of the Evidence: Who is providing the opinion? Is it a person with authority in the eyes of the trier of fact? How closely related is that authority to the evidence (i.e., is there a specious nexus)? Trial judges should be aware of the prejudices that flow from admitting a witness with an air of authority, but minimal verifiable ability to provide a reliable opinion, a factor for consideration at step 4.

3. Step 3: Exclusion of Secondhand Evidence

At this third step, evidence that is not based on personal observation should be categorically excluded and the witness’s evidence properly scoped to exclude such evidence. This scoping is a direct extension of the Supreme Court’s directive in Sekhon and Bingley that the permissible scope of an expert’s testimony should be established at as early a stage as possible.[217]

Given that modern lay opinion often strays into prejudicial areas, it should be scoped as well. This direction may save a great deal of time in cases like Colpitts, where the lay evidence could also have been scoped to include only the evidence that Officer Black personally observed. For instance, he would have been permitted to give a description of the content and timing of the accused’s emails. Such evidence could then have been relied on by the Crown’s expert, who was admitted later in the course of the trial.

We prefer to consider the presence of methodology in a more discretionary manner at step 4 because lay and expert opinion frequently overlap in their subject matter. In other words, lay and expert witnesses both opine on topics like voice and handwriting identification. Therefore, the failure to use methodology should be sensitive to context; it is much more problematic when an apparent expert fails to use methodology to avoid bias than when a true layperson does the same.

4. Step 4: The Lay Opinion Gatekeeper

In the tradition of the rule in Mohan, evidence that survives steps 1 through 3 should be subjected to a discretionary weighing of the benefits and costs to the trial.[218] In terms of costs, prejudice from the specious nexus found in step 2 should weigh heavily in the trial judge’s calculus. Put slightly differently, the trial judge should take into account the degree to which the witness’s authority gives a false sense of security about the evidence’s reliability. Further, failure to follow methodology to counter unconscious bias is a significant impediment to a fair trial and militates against admitting the evidence.

Benefits, in keeping with original purpose of the lay opinion rule, derive from utility to the trier of fact in making an informed decision. The trial judge should therefore consider admitting the opinion when the facts underlying it are unverbalizable qualia that cannot be admitted any other way. Similarly, and in keeping with Graat, a witness, better placed than the judge to draw the inference and one having the necessary experiential capacity, should be favoured over those who cannot establish those advantages.[219]

Identifying benefits also demands consideration of whether the witness seeks to go beyond giving a compendious statement of facts. For example, the police officers in Ilina did not provide their opinion that there was a cleanup out of any form of necessity. Rather, the exact same information—everything except the opinion—could have been conveyed by simply describing what they saw: a pool of blood adjacent to a clear liquid. In this respect, Ilina is distinguishable from cases where the witness cannot provide as accurate a comparison, as identified in step 1, without resorting to opinion—for example, an eyewitness identification.

Elements of a contextual approach should also be considered at step 4. For instance, the fact that Sergeant Carriere’s footprint analysis was effectively impossible to rebut by the accused should militate toward its exclusion or application of the more searching expert rules. Likewise, in Colpitts, the fact that Officer Black’s forensic evidence was duplicative and cut to the heart of the Crown’s case should also have provided a strong reason to exclude it.


Surveying the boundaries of the lay opinion exception poses a distinct challenge for courts and legal scholars. It is a task that draws on the purpose of the exception, the limits of human psychology, and the trial judge’s duty to gatekeep another type of evidence—that brought by experts. This challenge has, by and large, not been met. As we demonstrated in our survey of the case law, courts across Canada widely disagree over the content of the lay opinion rule, and sometimes over whether a rule exists at all. Moreover, this is an issue with significant stakes, especially to the criminally accused who regularly face adversarial lay witnesses bearing all the hallmarks of expert witnesses. We hope that we have shed an analytic light on this often overlooked problem. But this is just the first step. Canadian courts should adopt a more structured framework to the admission of lay opinion, one that will both prevent unfairness and accurately determine truth.