Some notes on Constructing Evidence and Educating Juries: The Case for Modular, Made-In-Advance Expert Evidence about Eyewitness Identifications and False Confessions (Jennifer L Mnookin)

At the recommendation of one of the editors of the Osgoode Hall Law Journal, I recently read Jennifer Mnookin’s excellent article on “modular” expert evidence. In our forthcoming paper, Will Crozier and I suggested that expert evidence about false confessions and the fallibility of eyewitness memory is excluded on the basis of a misunderstanding of human psychology. In short, Canadian courts deem expertise unnecessary because it simply duplicates the knowledge and experience and the factfinder. We disagree: people are often not aware of how their memory works and how strong the impact of the situation is on their behaviour (including those forces that produce false confessions).

We provided some ideas for how to more appropriately deal with eyewitness identifications and confessions, but it is a difficult question. For example, while we contend that the stated legal reason for excluding experts in these cases is mistaken, we do not think that experts are always going to be the appropriate corrective. This is because they are expensive and time consuming. Moreover, there simply may not be a qualified expert available.

Mnookin approached the same problem and made the (strong, in my view) case for pre-packaged modular evidence on eyewitness memory and confessions (and she did so in her characteristically lucid manner). The idea is that – because expert evidence in these cases is rarely fact-specific – expert opinion about, say, the effect of weapon-presence on memory can be pre-written and -recorded by a neutral and sophisticated body. The same can be done for the cross-examination of this opinion.

In making a case for modularized evidence, it’s important to understand the strengths and weaknesses of the current ways in which we regulate eyewitness identifications and confessions. In other words, is there really a need for a new corrective? I’ll run through those now.

Exclude unreliable identifications & confessions, the pros and cons

  • The pros

    • Excluding an unreliable eyewitness identification or confession ensures they cannot go on to be used by the factfinder and contribute to a miscarriage of justice

    • This nudges police towards using the most reliable procedures because they do not want to lose their evidence.

    • This is an already established role of the trial judge: to exclude unreliable evidence.

  • The cons

    • (in response to the arguments that exclusions nudge police). With eyewitness identifications, many of the factors that produce mistaken identifications are out of the control of the police (i.e., estimator variables like whether the remembered person was of the same or different race).

    • There is still a need to educate factfinders about identifications that are reliable enough to admit but may still be tenuous.

    • Can judges decide, with any accuracy, which identifications are unreliable and which ones are not? On this point, Mnookin notes that even psychological experts don’t have the tools to opine on the reliability of any particular eyewitness identification or confession. They can only discuss the factors that impact them. Similarly, even a judge that is very familiar with the social scientific literature will not be able to precisely say which identifications and confessions are unreliable.

    • Mnookin provides two related other “cons”. First, it is typically the jury’s job to judge credibility and assign weight. We generally ought not encroach on that without good reason. And second, the juries will look at the evidence holistically rather than in a vacuum. For instance, if an identification is corroborated by other evidence, then they may assign more weight to it. Judicial gatekeeping may not be able to perform this more sensitive task.

Here, I might push back on Mnookin’s thesis a bit. First, just because judges do not have a finely-tuned way to determine which identifications and confessions are unreliable, that does not mean that there are not clear-cut cases that should be excluded. For instance, we certainly cannot say that a long delay, plus a cross-racial identification, plus a police officer who knew the suspect equals a 95% chance of a mistaken identification. But just as we exclude, on a case-by-case basis, hearsay and (notionally) expertise for indicia of unreliability, why not apply a similar (albeit somewhat noisy) standard to confessions and eyewitness identifications (and note that Canadian courts at least do already apply reliability scrutiny to Mr. Big confessions and, to some degree, traditional confessions)? Further, the fact that jurors weigh evidence in light of other evidence might actually be a reason for stronger gatekeeping. Research shows that unreliable (but admissible) evidence can bolster a juror’s view of independent evidence (see Hasel & Kassin, 2009, “On the Presumption of Evidentiary Independence: Can Confessions Corrupt Eyewitness Identifications?)”.

Judicial instructions, the pros and cons

  • The pros

    • Judges may be more credible to a jury than an adversarial witness.

    • They are relatively quick and economical.

    • They can be based on pre-prepared instructions developed by non-partisan scientific bodies (just like modules).

  • The cons

    • They can be misleading (this is Mnookin’s primary concern with them). In particular, they present social scientific findings as common sense. In this way, they may say it’s normal human experience to have poorer recognition of other-race faces. But, as Mnookin notes, this may not have been the person’s (conscious and remembered) experience and thus they may not understand the instruction or discount it. In other words, they might not have noticed that their memory for other-races is poorer than that for same-races. Instead, these should be couched as what they are – social scientific findings.

    • And, following from the last point, the appropriate source of social scientific findings is a social scientist (i.e., an expert witness). These findings are not the appropriate subject matter of judicial notice. Here, I’ll quote Mnookin (at 1841-1842): “The fundamental problem is that these matters are, in fact, evidence. They are not incontestable. They are not certain. They are also not findings that are ideally reduced to a single sentence or two.” The result is what Mnookin describes as a “category error” – courts are trying to transform evidence into instructions. Therefore, there is no chance for parties (in in direct or cross-examination) to get into the subtleties, weaknesses and controversies of the science.

Admit expert evidence, the pros and cons

  • The pros

    • Experts can explain the science and any weaknesses can be explored in cross-examination.

    • Experts will almost always know more about the subject matter of their expertise than judges.

    • There is no category error – experts witnesses are presenting expert evidence.

    • This is a bit removed from the question, but I think it’s also probably helpful for a lawyer to consult with an expert. The expert may notice social scientific issues that the lawyer didn’t know were there.

  • The cons

    • It’s repetitive. Many cases will include exactly the same evidence. This is because it’s framework evidence, rather than individuating. Experts can only speak to the factors that influence identifications and confessions.

    • It’s expensive and time consuming. As a matter of justice, many parties will not be able to find an expert or afford one.

Modular expert evidence

In response to the above pros and cons, Mnookin proposes pre-prepared, modular expert evidence. The idea is that this social scientific framework evidence can still be presented as expert evidence (no category error), but in a more efficient way. A sophisticated neutral party would produce direct and cross-examination expert modules to be presented to the factfinder. For instance, if there’s an issue of cross-racial identification, the court can play the cross-racial identification module (and cross) for the jury. This would take less court time and have almost no financial cost to the parties.

There are some drawbacks to modular testimony. For instance, it requires someone to go out and make it. And it will have to eventually be accepted by courts. As Mnookin notes, however, prosecutors may not actually be very resistant to modularization. This is because the current trend towards science-based instructions (e.g., Henderson) results in information that is immune to cross-examination. Modules that include cross-examination may therefore be preferable to the prosecution. Further, while experts generally should not give individuating opinions (e.g., the reliability of a particular identification), the modules may be too general. For instance, an expert giving more bespoke testimony may be able to expressly link a case factor to research findings (e.g., police officer X saying “are you sure it’s not this person?” during the identification and the research that has shown the effect of that). That level of specificity will not be possible with premade modules.

There may also be cases in which specialized testimony could be useful in providing a more nuanced understanding of the psychological processes at play. For example, I can imagine cases in which the prosecution overreaches, saying that the witness is an intelligent adult with no history of memory problems (the implicit message being that they would be immune from many of the memory issues being discussed). Here, an expert may be useful in clarifying that the situational forces that impact memory have been demonstrated in healthy and intelligent adults. Moreover, there may be more complicated cases in which studies have found interaction effects (e.g., one variable only impacts memory if another variable is present). These interactions between factors would be hard to anticipate in modules.

Some stray thoughts

As I’ve indicated, I think Mnookin’s proposal shows a lot of promise. In fact, some colleagues and I have discussed something similar with forensic scientific framework evidence (i.e., meta-expertise). It can be difficult to admit “exogenous” knowledge about the uncertainties and limitations (see Edmond et al, below) of forensic science. My colleagues and I have thought about writing non-partisan primers to assist courts (as the Royal Society sometimes does). These might be introduced as Mnookn’s modules.

I would, however, take some issue with Mnookin’s portrayal of judicial instructions about eyewitness memory and false confessions. She’s right that it’s not what we typically think of as within the realm of judges (i.e., the category error). But, I don’t think that’s necessarily so problematic. As the Court in Henderson said, these findings are about as reliable as they come: “Experimental methods and findings have been tested and retested, subjected to scientific scrutiny through peer-reviewed journals, evaluated through the lens of meta-analyses, and replicated at times in real-world settings.”

In other words, if we keep judicial instructions to the very non-controversial aspects of psychology, then I don’t see a big problem with them resembling evidence more than instruction. The drafters of such instructions should, however, be very careful in the findings they rely on (note this is an issue for modules as well).

Finally, I was hoping Mnookin would deal more with the behavioural research comparing the effectiveness of judicial instructions and expert evidence on the juror’s assessment of eyewitness testimony (see footnote 116 of her piece). This is an unsettled and very interesting area of research. Based on my reading of the studies, neither instructions nor expert evidence has been shown to be effective in sensitizing jurors to the factors that produce mistaken identifications. Rather, they seem to just make eyewitnesses more generally skeptical of identifications. Will and I ran into the same problem in our paper – there simply wasn’t space to delve into this area of research. Still, the studies showing no benefit for expert witnesses challenge both the view that expertise – modular or otherwise – should see widespread use. This is topic that could really be the subject of an entire article.


Gary Edmond, David Hamer & Emma Cunliffe, “A little ignorance is a dangerous thing: engaging with exogenous knowledge not adduced by the parties” (2016) 25:3 Griffith Law Review 383-413.

Lisa Dufraimont, “Regulating Unreliable Evidence: Can Evidence Rules Guide Juries and Prevent Wrongful Convictions?” (2007-2008) 33 Queen’s Law Journal 261.

Jason M Chin & William Crozier, “Rethinking the Ken Through the Lens of Psychological Science” (forthcoming) Osgoode Hall Law Journal.

Jennifer L Mnookin, “Constructing Evidence and Educating Juries: The Case for Modular, Made-In-Advance Expert Evidence About Eyewitness Identifications and False Confessions” (2015) Texas Law Review 1811.

State v Henderson, 208 NJ 208 (2011), 27 A3d 872.