“Moore” Than Just a Number: Why IQ Cutoffs Are an Unconstitutional Measure for Determining Intellectual Disability – Note by Courtney Johnson

From Volume 91, Number 4 (May 2018)


Moore Than Just a Number: Why IQ Cutoffs are an Unconstitutional Measure for Determining Intellectual Disability

Courtney Johnson[*]


“Yeah,” said George. “I’ll come. But listen, Curley. The poor bastard’s nuts. Don’t shoot ‘im. He di’n’t know what he was doin’.”

John Steinbeck, Of Mice and Men[1]

Bobby James Moore was twenty years old when he fatally shot a store clerk while robbing a grocery store in April 1980.[2] On paper, this is a tragic felony murder, but behind the scenes lies a different story. Bobby was not a typical twenty-yearold; he did not understand the days of the week, the months of the year, [or] the seasons.[3] Bobby could barely tell time, and he could not understand standard measurements or that subtraction is the opposite of addition.[4] Bobby suffered an “abuse-filled childhood.”[5] Bobby dropped out of high school due to his limited ability to read and write, and he lived on the streets after being kicked out of his home for being “stupid.”[6] Bobby is intellectually disabled, and despite the evidence put forth demonstrating his disability, he was sentenced to death pursuant to a set of factors used by a Texas court; these factors are largely based on stereotypes and caricatures from literature.[7] As the United States Supreme Court decided in 2017, this was a gross violation of the Eighth Amendment’s protection against cruel and unusual punishment to rely on “wholly nonclinical”[8] factors rather than the “medical community’s diagnostic framework.”[9]

Mental health and the criminal justice system consistently interact when it comes to theories of punishment and culpability. When they clash, the crime often takes center stage, while the mental health of the defendant remains ignored. Individuals suffering from intellectual disabilities, mental disorders, or both are treated unfairly as criminal defendants when their conditions, consisting largely of impairments of the ability to make rational decisions, are not taken into serious consideration at sentencing. Courts, legislatures, and the public generally have struggled to understand criminal defendants with intellectual disabilities. It was not until 2002 that the United States Supreme Court ruled it unconstitutional to execute intellectually disabled persons in the landmark decision Atkins v. Virginia.[10] The Court found there is a “national consensus” that people who suffer from “mental retardation”[11] should be exempt from the death penalty.[12] Before Atkins was decided in 2002, at least forty-four people who would have been exempt under Atkins were executed.[13]

Like many prior landmark decisions, states resisted Atkins because of its failure to define “intellectual disability” and the fact that it left to the states “the task of developing appropriate ways to enforce the constitutional restriction.”[14] Several states took this opportunity to implement harsh IQ cutoffs for determining intellectual disability in capital cases.[15] Many states passed legislation prior to Atkins defining intellectual disability as requiring an IQ score of below seventy, and following Atkins, these states began denying its exemption to any claimants with IQ scores of seventy or above. Twelve years after Atkins, the Supreme Court addressed the issue of IQ cutoffs in Hall v. Florida, in which the Court concluded that Florida’s cutoff “disregards established medical practice[s]” and that when a defendant’s IQ falls in a certain margin of error, the defendant must be able to present additional evidence of adaptive deficits.[16] Hall reinforced the need to focus on adaptive behavior in addition to IQ, but again, the states still had discretion over how to consider the behavior. Then came Moore v. Texas, the most striking example post-Hall that there remains a long-standing misperception of intellectual disabilities.

Before reaching the Supreme Court, the Texas Court of Criminal Appeals (“CCA”)[17] denied Bobby James Moore’s habeas petition claiming exemption from the death penalty under Atkins. The CCA relied on its prior decision in Ex Parte Briseno, in which it determined a defendant was essentially not disabled enough for death penalty exemption, contrasting with the classic example of a character with a severe disability, Lennie in John Steinbeck’s Of Mice and Men.[18] The CCA in Briseno held that[m]ost Texas citizens might agree that Steinbeck’s Lennie should, by virtue of his lack of reasoning ability and adaptive skills, be exempt [from the death penalty],” but because the petitioner did not fall in the category of “severely mentally retarded” like Lennie would, he was denied exemption from the death penalty.[19]

Reasoning based on a literary character is dangerous, and, as this Note will argue, intellectual disabilities cannot be boiled down to stereotypes or an isolated number from an IQ test. This Note will look to recent court decisions, state statutes, and literature from the psychological and psychiatric communities, and it will evaluate the Supreme Court’s decision in Moore v. Texas overturning the CCA. It will further consider what the Supreme Court’s decision could mean not only for the future of intellectually disabled defendants, but also for defendants with mental illnesses facing the death penalty.

Part I of this Note focuses on the legal background pre- and post-Atkins and how courts have treated individuals with intellectual disabilities. It will examine attempts to define “mental retardation” for the purpose of exemption under Atkins by looking at how various literatures, state statutes, lower court decisions, and how clinicians define it. Part II then focuses on the recent Supreme Court case on this issue, Moore v. Texas. Lastly, Part III discusses Moore’s future implications on the ability of intellectually disabled persons to argue that without Hall and Moore applying retroactively or a specific holding from the Court regarding the unconstitutionality of IQ cutoffs, such persons will continue to have their constitutional rights violated.


A.  Creating a Constitutional Exemption for the Intellectually Disabled

The Eighth Amendment prohibits the infliction of cruel and unusual punishment.[20] Initially, this prohibited archaic punishments that were considered “cruel and unusual” when the Bill of Rights was adopted, but the Supreme Court has since recognized it encompasses the “evolving standards of decency that mark the progress of a maturing society.”[21] In addressing these “evolving standards,” the Court has looked to evidence from society’s current views on punishment, most often found in state legislation and data from jury sentencing.[22] Since the adoption of the Bill of Rights, the Eighth Amendment has been expanded to prohibit the execution of juveniles,[23] individuals deemed incompetent at the time of execution,[24] and the “mentally retarded.”[25] All three protected classes represent important facets within the complexity of the death penalty issue; however, this Note focuses on the latter: how the Eighth Amendment has come to prohibit the execution of individuals with “mental retardation.”

The Supreme Court has come a long way in its treatment of the intellectually disabled in regards to the death penalty since its first decision on the matter in Penry v. Lynaugh.[26] Decided in 1989, this was the first time the Court addressed the issue of whether the Eighth Amendment’s prohibition against cruel and unusual punishment exempts intellectually disabled individuals from execution. The Court held that executing the “mentally retarded” did not violate the Eighth Amendment, but based its decision largely on the fact that only two states at the time (Georgia and Maryland) banned executions of “mentally retarded” criminals.[27] The Court found that this was not sufficient evidence of a “national consensus” that the practice violated “standards of decency.”[28]

Just as the states began to change their position on the issue, so did the Supreme Court. Following the decision in Penry, sixteen states across the country enacted statutes like those in Georgia and Maryland from the period of 1990 to 2002, totaling eighteen states with exemptions for the mentally disabled when Atkins was decided.[29] Even more significant is the fact that no states passed legislation enforcing the power to execute intellectually disabled individuals in this time period.[30] This shift in the national consensus, arising from the states, changed the way the Court viewed intellectual disabilities. The same year the eighteenth state enacted legislation exempting the “mentally retarded,” the Supreme Court handed down its landmark decision Atkins v. Virginia, in which it held it is unconstitutional and a violation of the Eighth Amendment to execute people with “mental retardation.”[31]

Atkins was largely based on this shift in national consensus (demonstrated by state’s enacting laws banning the execution of “mentally retarded” individuals), which the Court found to provide “powerful evidence that today our society views mentally retarded offenders as categorically less culpable than the average criminal.”[32] Eighteen states is still short of half the country, but the Court found it significant that in the states that still allowed the execution of intellectually disabled offenders, actually carrying out the practice is rare.[33] In fact, only five states executed individuals with an IQ of less than seventy between the time Penry and Atkins were decided.[34]

In addition to a shifting national consensus, the Court considered two goals of the criminal justice systemretribution and deterrenceand evaluated how executing the petitioner in Atkins would serve either of those interests.[35] First, with respect to retribution, the Court reasoned that a defendant deemed “mentally retarded” acts with a “lesser culpability” than the average person guilty of murder, and thus a death sentence would be disproportional and would not serve the interests of retribution.[36] Secondly, with respect to deterrence, the Court found that capital punishment would serve as a deterrent only for potential murderers with a “cold calculus that precedes the[ir] decision” to kill.[37] The theory behind deterrence rests upon the notion that the severity of the punishment, in this case death, will dissuade criminal conduct.[38] That is simply not true for individuals with “mental retardation,” as the Court found that it makes one “less likely [to] process the information of the possibility of execution as a penalty and, as a result, control their conduct based upon that information.”[39] This is also true in reverse: exempting individuals with “mental retardation” will not lessen the deterrent effect on those unprotected. Given that, in Atkins, the individual had an IQ of fifty-nine, as well as clinician testimony and school records supporting a finding of “mental retardation,” it is not likely others would attempt to claim “mental retardation” falsely to gain exemption without such evidence.

The core of the Atkins decision is its recognition that intellectually disabled individuals are less culpable because of their cognitive and behavioral impairments. This creates a “diminished ability to understand and process information, to learn from experience, to engage in logical reasoning, or to control impulses . . . .[40] Not only does this impairment affect an individual’s decisionmaking throughout the circumstances of their crime and lead to a failure to appreciate risks and consequences, but also it can inhibit one’s ability to receive proper counsel, resulting in deficient due process.[41]

The Court in Atkins felt secure in its decision given the supportive national consensus, but did note that to the extent there is any disagreement, it will be in “determining which offenders are in fact retarded.”[42] This prediction by the Court anticipated the ambiguity and difficulty states have had post-Atkins in determining which defendants qualify as “mentally retarded” to be exempt from execution. Thus, while Atkins serves as a momentous step for intellectually disabled defendants, it leaves uncertainty on how to properly protect such individuals from execution by failing to define “mental retardation.” It left “to the State[s] the task of developing appropriate ways to enforce the constitutional restriction upon . . . [its] execution of sentences,”[43] meaning that the states retained discretion in determining the criteria for who exactly is “mentally retarded” for purposes of a death penalty exemption under Atkins.

B.  Defining “Mental Retardation” after Atkins

The Supreme Court’s reasoning for leaving the definition of “mental retardation” ambiguous is that “[n]ot all people who claim to be mentally retarded will be so impaired as to fall within the range of mentally retarded offenders about whom there is a national consensus.”[44] The Court appears to refrain from defining “mental retardation” to avoid a categorical ban that would exempt some individuals who claim to be intellectually disabled, but actually are not impaired to the extent that it diminishes their culpability. The Court did however explain that the medical community defines “mental retardation” per three criteria: (1) significantly subaverage intellectual functioning,[45] (2) deficits in adaptive functioning,[46] and (3) that these deficits manifest prior to age eighteen.[47]

1.  Significantly SubAverage Intellectual Functioning

The American Association on Mental Retardation (“AAMR”) defines “significantly subaverage” intellectual functioning as having an IQ of about seventy or below.[48] An IQ can be obtained by one or more of the standardized, individually administered intelligence tests, such as the Wechsler Adult Intelligence Scales (WAIS), Wechsler Intelligence Scales for Children (WISC), the Otis-Lennon Mental Ability Test (OLMAT) and the Stanford-Binet-V (SB-V).[49] The American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders, fifth edition, (“DSM-5”) has recognized there is a measurement error of approximately five points in assessing IQ.[50] DSM-5’s guidelines ensure that no one would be diagnosed with an IQ lower than seventy if no significant defects in adaptive functioning are shown, however, that also makes it possible “to diagnose Mental Retardation in individuals with IQ’s between [seventy] and [seventy-five] who exhibit significant deficits in adaptive behavior.”[51] The DSM-5 also designates classifications of mental retardation into degrees: profound (IQ below 2025), severe (IQ 2025 to 3540), moderate (IQ 3540 to 5055), and mild (IQ 5055 to 7075).[52] Approximately 85% of all intellectually disabled persons, and the “overwhelming majority of capital defendants with mental retardation,” fall in the mild range.[53] Only 34% fall in the “severe mental retardation” category,[54] where Lennie from Of Mice and Men would likely fall. There are several moral and practical implications of setting a strict IQ cutoff at seventy for defendants bringing Atkins claims that will be discussed in Part IV.

2.  Adaptive Behavior Criteria

The American Association on Mental Deficiency (“AAMD”) defines significant impairments on adaptive functioning as “limitations in an individual’s effectiveness in meeting the standards of maturation, learning, personal independence, and/or social responsibility that are expected for his or her age level and cultural group, as determined by clinical assessment and, usually, standardized scales.”[55] In other words, this analysis focuses generally on ordinary skills the typical individual possesses to function in everyday life. The level of everyday adaptive functioning is compared to an individual’s age, gender, and socioculturally matched peers.[56]

The AAIDD and DSM-5 sum this up in three adaptive-behavior skills: conceptual, social, and practical.[57] The conceptual domain, also called academic domain, involves, among other abilities, competence in functions such as memory, language, reading, writing, math, problem solving, and ability to form judgment in novel situations.[58] The social domain focuses on personal interactions and how one reacts to them. This domain includes: awareness of others’ thoughts, feelings, and experiences; empathy; interpersonal communication skills; ability to make friends and judge social situations.[59] Lastly, the practical domain focuses on the individual’s ability to live a productive life in the world by evaluating self-management, ability to care for oneself, ability to adhere to school or job responsibilities, and ability to manage money, among other skills.[60] The DSM-5 recommends gathering evidence of skills deficits in these domains by various means, such as looking into educational, developmental, and medical history.[61]

3.  Manifesting Before Age Eighteen

 The third, and least litigated, factor of the definition of intellectual disability requires that the disability “manifest[s] before age 18.”[62] The DSM-5 requires evidence of both intellectual and adaptive deficits, the first two prongs, to be shown during this “developmental period.”[63] In most Atkins claims this is not as vigorously litigated, as courts often review evidence of childhood environment, medical histories, behavioral records, school records, and testimony of behavior from those who knew the individual as a child.[64]

 A recent Sixth Circuit decision arising out of Tennessee provides an example of how this third requirement can impact an Atkins claim. There, an Atkins claimant’s school records indicated IQ scores from eighty-three to ninety-seven from ages seven to thirteen years old, yet at age forty-five he was receiving scores of fifty-seven and sixty-nine.[65] Those scores were not evidence of lifelong mental retardation,” manifesting during childhood, and because all of his scores before he turned eighteen were higher than seventy (even considering the Flynn effect and other deviations), the defendant’s Atkins challenge was rejected.[66]

4.  State Interpretations of the Factors: Ex parte Briseno

One of the most blatant departures from the principles established in AtkinsEx parte Brisenocame out of Texas in 2004. Jose Garcia Briseno sought state habeas relief, “alleging he was mentally retarded and . . . exempt from execution” for the murder he was sentenced to death for in 1991.[67]  The Texas legislature had not adopted a statute implementing Atkins, so the Texas Criminal Court of Appeals (“CCA”) took matters into its own hands, resulting in a wholly nonclinical approach that evaluated what became known as the Briseno factors.

In the absence of any state statute, the CCA looked at the DSM-IV,[68] bills the Texas Legislature had passed, relevant case law, and finally the AAMR. First, the court examined the DSM-IV and found that mental health professionals define intellectual disability “broadly to provide an adequate safety net for those who are at the margin and might well become mentally-unimpaired citizens if given additional social services support.”[69] The broad range of intellectual disabilities is shown by [t]he DSM-IV categoriz[ing] the mentally retarded into four subcategories: mildly retarded, moderately mentally retarded, severely mentally retarded, and profoundly mentally retarded.” The court noted that “mental retardation is not necessarily a lifelong disorder” given that many individuals fall into the “mildly mentally retarded” category.[70] Further, due to the broad categorization and range of IQ numbers, the court was unsure whether the petitioner’s disability was severe or long-standing enough for exemption.

The only other source the CCA looked to was Texas House Bill 236, passed by the 77th Legislature in 2001, before the Atkins decision was announced.[71] House Bill 236 would have prohibited the execution of intellectually disabled defendants, adopting the definition of “mental retardation” found in Texas Health and Safety Code Section 591.003(13).[72] However, House Bill 236 was vetoed by the Governor, and subsequently “[t]he 78th Texas Legislature did not pass a statute implementing Atkins.[73]

The CCA thus adopted the definition set by the AAMR, the same criteria Atkins discusses, including the requirement that the defendant’s “adaptive deficits” are related to the intellectual disability.[74] The CCA settled on this definition because it closely resembles the definition under the Texas Health and Safety Code: “‘[i]ntellectual disability’ means significantly subaverage general intellectual functioning that is concurrent with deficits in adaptive behavior and originates during the developmental period.”[75] Despite adopting the same definition, state courts could vary wildly in determining what is considered “significantly subaverage” functioning.

Indeed, unlike any other jurisdiction, the CCA adopted a narrower focus than what mental health professionals recommend by looking to the “level and degree of mental retardation at which a consensus of Texas citizens would agree that a person should be exempted from the death penalty.”[76] The CCA said unless the defendant seems retarded enough to the average person in Texas, they will not be exempt. The court cited Atkins, that “[n]ot all people who claim to be mentally retarded will be so impaired as to fall within the range of mentally retarded offenders about whom there is a national consensus.”[77] The CCA took this phrase and ran with it, interpreting it to allow the execution of persons whose disability falls in the mild rangethe largest portion of intellectually disabled criminal defendants and the range in which the Atkins claimant fell. The fact that Texas did not yet have any statutory provisions applying Atkins further allowed the court to ignore the holding in Atkins and create its own “factors” test not based on any clinical determination.[78]

The Briseno court argued that adaptive behavior criteria are “exceedingly subjective” and thus set out a list of seven evidentiary factors to guide the determination of whether a defendant is considered intellectually disabled under Atkins:[79]

[1.] Did those who knew the person best during the developmental stagehis family, friends, teachers, employers, authoritiesthink he was mentally retarded at that time, and, if so, act in accordance with that determination?

[2.] Has the person formulated plans and carried them through or is his conduct impulsive?

[3.] Does his conduct show leadership or does it show that he is led around by others?

[4.] Is his conduct in response to external stimuli rational and appropriate, regardless of whether it is socially acceptable?

[5.] Does he respond coherently, rational, and on point to oral or written questions or do his responses wander from subject to subject?

[6.] Can the person hide facts or lie effectively in his own or others’ interests?

[7.] Putting aside any heinousness or gruesomeness surrounding the capital offense, did the commission of that offense require forethought, planning, and complex execution of purpose?[80]

The court applied these factors to the defendant-petitioner in Briseno to determine that he did not prove, by a preponderance of the evidence, that he had significant limitations in adaptive functioning.”[81] They found that stories of him running away from home to escape the beatings from his great-grandma signified “good survival skills,” that officers testified his behavior seemed normal and appropriate in prison, and his own testimony seemed “clear, coherent and responsive.”[82] As later held in Moore v. Texas, these types of factors cannot provide an adequate basis for determining adaptive behavior because they overemphasize the strengths without considering the deficits.[83] Several courts in Texas since Briseno have dissented from the application of the factors as “decidedly non-diagnostic, giving Texas judges “amorphous latitude . . . to supply the normative judgment to say, in essence, what mental retardation means in Texas . . . for Eighth Amendment purposes.” [84]

After Briseno, it is estimated that Texas executed thirty to forty people with strong claims of intellectual disability relying on the nonclinical “we know it when see it approach” that was “as meaningless as answers given by a Magic 8 Ball.”[85] After the progressive step the Court took in Atkins, Briseno ran afoul the long-standing principle that the Eighth Amendment protects “evolving standards of decency that mark the progress of a maturing society,” making clear that Atkins was just one battle won in the fight for the constitutional rights of intellectually disabled persons.[86]

C.  Setting the Stage for Moore

In 2014, ten years after Briseno, the Supreme Court revisited the issue in Hall v. Florida, holding that a Florida capital punishment law requiring an individual claiming an intellectual disability to score seventy or below on an IQ test violates the Eighth and Fourteenth Amendments of the Constitution.[87] Freddie Lee Hall was convicted of two murders in 1978, a jury sentenced him to death, and both the Court of Appeals and Florida Supreme Court affirmed concluding that his intellectual disability could not justify or excuse his moral culpability based their interpretation of a Florida statute.[88] While the Florida statute appeared nearly identical to the three criteria in Atkins, it went further by defining “significantly subaverage” intellectual functioning as “performance that is two or more standard deviations from the mean score on a standardized intelligence test.”[89] The standard deviation is fifteen points, two deviations is then thirty points, but the Florida Supreme Court interpreted the statute as creating a strict IQ cutoff of seventy.[90]

When Hall was first sentenced, the Supreme Court had not yet decided Atkins, thus, in 2004, Hall filed a motion claiming an intellectual disability that would have exempted him from the death penalty pursuant to Atkins.[91] It took five years for Florida to hold a hearing considering his motion, and when they finally did, he presented evidence that his IQ score was seventy-one.[92] Notably, Hall actually received nine IQ tests over forty years ranging from scores of sixty to eighty, but the scores below seventy were excluded for “evidentiary reasons.”[93] Under the Florida Supreme Court’s analysis, a score of seventy-one put Hall above the mandatory cutoff, and thus other evidence could not be considered.[94]

The Supreme Court noted that Florida’s mandatory IQ cutoff disregards the medical practice because it treats IQ as “conclusive evidence” of an individual’s intellectual functioning, without considering other evidence of deficiencies in adaptive behavior.[95] In fact, the very professionals who create and run IQ tests are in consensus that they “should be read not as a single fixed number but as a range.”[96] This is because each test has a “standard error of management” (“SEM”),[97] reflecting the inconsistency and imprecision of the test. For individuals like Hall with an IQ of seventy, considering the SEM places him in a range between sixty-six and seventy-six.[98] The SEM also applies to an individual like Hallwho has taken multiple testsand must be applied to each one separately.

Turning now to the Eighth Amendment analysis of whether there is a national consensus that the practice of IQ cutoffs violates standards of decency, the Court found a “significant majority of States implement the protections of Atkins by taking the SEM into account,” reflecting the “error inherent” in using the test.[99] Additionally, only two other states had adopted a fixed score cutoff identical to Florida’s at the time of this decision.[100] There are however, nine states with statutes that could be interpreted as requiring bright-line cutoffs of seventy, but the Court found that four of them have not had courts rule on the issue.[101] In stark contrast, eighteen states have abolished the death penalty altogether and at least five states have passed legislation permitting a defendant bring a claim under Atkins, despite an IQ above seventy.[102] For all of these reasons, the Court rejected the strict cutoff.[103] Rather, the Court found significant evidence must be considered: “social and cultural environment, including medical histories, behavioral records, school test and reports, and testimony regarding past behavior and family circumstances.”[104]

After disregarding the strict cutoff under Florida’s statute, the Court’s analysis considered Hall’s school records, and testimony from his teachers, his lawyer, medical clinicians, and his family, finding them to be “substantial and unchallenged evidence of intellectual disability.”[105] The Court went beyond the IQ test’s simple number by delving into the defendant’s childhood. The opinion cites several different testimonies that demonstrated his intellectual disability. For example, his siblings testified there was “something ‘very wrong’ with him as a child,” and he was “slow with speech and . . . slow to learn.”[106] Strikingly, his mother “would strap [Hall] to his bed at night . . . [and] awaken [him] by hoisting him up and whipping him with a belt, rope, or cord” and on one occasion she “buried him in the sand up to his neck to ‘strengthen his legs.’”[107] In light of the powerful evidence presented, the Court found that “Hall’s upbringing appeared to make his deficits in adaptive functioning all the more severe.”[108]

The Court’s reasoning in Hall is significant because it acknowledges that a person is more than just a number.[109] Not only is an important constitutional protection against cruel and unusual punishment at stake, but also the rights of a group of individuals who have long suffered due to the stigma of their intellectual disabilities. What is at stake here is not automatically excusing anyone with an intellectual disability from punishment, but rather ensuring that an individual has the opportunity to present evidence of his disability, including deficits in adaptive behavior. Florida’s decision to execute Hall because his IQ score was one point above the cutoff was an extreme circumstance, and the Court properly championed the rights of intellectually disabled persons in its opinion. However, its ruling only invalidated the Florida statute under the Eighth Amendment; thus the Court has not yet categorically banned state reliance on IQ tests.


Two years after Hall v. Florida, on June 6, 2016, the Supreme Court granted certiorari in its most recent case regarding the intellectually disabled, Moore v. Texas.[110] The Supreme Court granted certiorari, taking up the specific issue of whether it violates the Eighth Amendment under Hall and Atkins to prohibit the use of current medical standards on intellectual disability and instead require the use of outdated medical standards in determining whether an individual can be exempt from the death penalty. The main issue is whether medical definitions govern how courts rule, or whether courts have discretion to develop the legal standard on their own.

In 1980, Bobby James Moore and two accomplices attempted to rob a grocery store. Moore was meant to be a look-out guard positioned at the front booth with a shotgun, but when he approached the booth, he shot and killed an employee.[111] Moore was convicted of capital murder and sentenced to death.[112] Moore brought habeas petitions in both state and federal courts, and the U.S. Court of Appeals for the Fifth Circuit found Moore received ineffective assistance of counsel during his trial and sentencing because his attorney failed to develop or present mitigating or exculpatory evidence.[113] After a new state court sentencing hearing in February 2001, Moore was sentenced to death again.

Because Moore v. Texas came out of the CCA in Texas, it evaluated Moore’s intellectual functioning pursuant to the Briseno factors. The CCA found Moore failed to meet the Briseno factors and again imposed a death sentence.[114] In doing so, the CCA reversed a lower court ruling that followed the scientific diagnostic criteria set by medical professionals, which had found that Moore did have an intellectual disability.[115] Texas is the only state that followed the Briseno factors, allowing the CCA the flexibility to interpret the three prongs in whichever way it felt the majority of Texas citizens would. Thus, the CCA side-stepped Atkins and deemed Moore not intellectually disabled enough under Briseno. This decision was made in the face of showing a history of intellectual disability that had been documented throughout Moore’s childhood. This included testimony from a clinical neuropsychologist that Moore’s “mental age at the time of the offense was no greater than fourteen years” and he had a “lack of impulse control and a diminished ability to think through the consequences of his actions.”[116] Since Moore had the same cognitive functioning of a fourteen-year-old, he had the same “diminished capacity”[117] to make decisions that makes juveniles less culpable than adults. The Supreme Court has held juveniles less culpable than adults for several reasons, including a susceptibility to peer pressures and influence by others.[118]

Demonstrating his intellectual limitations, testimony at a 2014 evidentiary hearing revealed that “when [Moore] was in second and third grade, he could not tell a $1 bill from a $5 or $10 bill.”[119] Further, Moore’s siblings testified about the “neglectful, physically and verbally abusive alcoholic” father they had.[120] According to their testimony, Moore received the harshest beatings and was thrown out of the house at age fourteen because he could not spell and his father “thought he was stupid.”[121] This testimony was corroborated by a neighbor who could attest to witnessing the beatings as well as Moore’s “haggard” and bruised appearance.[122] Experts acknowledged that emotional disturbances and environmental conditions like that of Moore’s upbringing can adversely impact an individual’s learning ability and IQ scores.[123]

Resulting from his childhood hardships at home, Moore dropped out of school around age fifteen or sixteen and started living a street life.[124] Part of this new life included smoking marijuana and taking 7 to 14 Quaalude pills per day.”[125] Moore led a troubled life as a teenager, part of which stemed from his intellectual challenges. He stated that due to his inability to read or write, he skipped school starting as early as fourth grade[126] After reviewing Moore’s entire record, Dr. Borda, the clinical neuropsychologist who originally reviewed his case, stated in a 2013 affidavit that Moore met the criteria for a[n] intellectual-disability diagnosis.[127]

The CCA in Ex parte Moore began its analysis first by evaluating Moore’s “significantly sub-average” intellectual functioning, as laid about by Atkins, but added that this is “generally shown by an IQ of 70 or less.”[128] As mentioned previously, IQ scores can vary significantly over the years, but they can also vary given the type of test administered. These cases can become particularly confusing for both judges and the public alike, given that there are so many different clinical and technical ways to determine IQ. Moore had taken several IQ tests, varying in form, and the scores before the court in its deliberation were: seventy-seven on the OLMAT at age twelve, fifty-seven on the Slosson at age thirteen, seventy-eight on the WISC at age thirteen, seventy-one on an abbreviated WAIS-R at age thirty, seventy-four on a complete WAIS-R at age thirty, eighty-five on the RCPM at age fifty-four, and fifty-nine on the WAIS-IV at age fifty-four .[129] Dr. Borda identified the score of fifty-seven as the “first and most accurate assessment” and discounted the score of seventy-eight on the WISC because it should be adjusted to 70 for the Flynn Effect.[130] The Flynn Effect is a phenomenon in which IQ of the general population is estimated to increase at a rate of three points per year, so IQ tests must be renormed.[131] Dr. Borda provided opinions as to each score and why it was not reliable and asserted that Moore was “very limited to begin with,” so being in harsh environmental conditions likely adversely affected Moore’s learning ability and IQ scores.[132]

The CCA next found that “[e]ven if [Moore] had proven that he suffers from significantly sub-average general intellectual functioning, he still could not win an Atkins claim because he did not prove the second prong, limitations on adaptive functioning by a preponderance of the evidence.[133] The state pointed to Moore’s job mowing grass and “hustling pool” as evidence of money skills, knowledge, and “self-direction” to obtain a job, however expert Greenspan did not find any of those as “adaptive” behavior.[134] Moore’s school records reflect poor grades, below-grade-level scores on academic achievement tests, and as early as kindergarten, he was considered potentially intellectually disabled.[135] The court found a “far more credible” forensic psychologist, Compton, who testified that Moore did exhibit some deficits in academic and social-interaction skills during his developmental period, but was at too high a level of adaptive functioning to support an intellectual disability diagnosis.[136] Compton further pointed to the advances Moore made while on death row as evidence that his early life problems were not caused by a disability, but derived from his difficult childhood.[137] Considering the Briseno factors, the court ultimately found that there was not enough evidence of adaptive behavior deficits due to Moore’s ability to lie, and his “forethought, planning, and moderately complex execution of purpose.”[138]

The American Civil Liberties Union (“ACLU”) argued in its amicus brief in support of the petitioner in Moore that given current clinical standards, the Briseno factors cannot stand. In its brief, the ACLU argued that the Briseno factors are based on a stereotyped view of intellectual disability derived from the character of Lennie in John Steinbeck’s Of Mice and Men, and that in practice, it subjects defendants with mild intellectual disability to the death penalty, thus violating the Constitution’s ban on cruel and unusual punishment.[139]

The Texas court’s “flawed interpretation” of Atkins allows the execution of those who fall in the mild intellectual disability range with significant deficits in adaptive behavior, but whose IQ scores are above the threshold of seventy for exemption.[140] The Briseno court mistakenly turned to the character Lennie, largely representative of a stereotype, for guidance in creating its factors and thus relied closely on a fictional character rather than clinical findings.[141] The ACLU points out how dangerous Briseno’s holding is because it allows the state to execute individuals with mild intellectual disability, even under Atkins, and the “overwhelming majority” of intellectually disabled persons fall in the mild range.[142] Even more shocking is that the individuals who fall in the more severe disability range, “rarely, if ever, have the capacity to commit capital crimes.”[143] In Atkins, the defendant-petitioner on death row was “mildly mentally retarded.”[144] For these reasons, the brief persuasively calls into question both Briseno’s reasoning and conclusion. Although Atkins left to the states how to implement the decision, that does not allow the states to make their own determination as to what the Eighth Amendment encompasses and certainly does not allow for concluding a person is not disabled enough on the basis of stereotypical ideas. Hall made clear that the inquiry should be informed by the medical community’s framework and clinical standards. The medical community and clinical authorities agree that “an individual with an IQ score above 70 may properly be diagnosed with intellectual disability if significant limitations in adaptive functioning also exist.”[145]

In oral argument on November 29, 2016, counsel for Petitioner Moore argued that Hall prohibits lower courts from ignoring current medical standards, like the court in Briseno did.[146] Texas, on the other hand, argued the Court’s long-standing view is that there is subjectivity in the medical diagnosis, and the habeas judge erred by employing the current standard rather than following the Briseno factors.[147]

Petitioner Moore’s counsel further argued that even if the Briseno factors are an acceptable framework, the CCA erroneously applied them in two ways. First, the court did not factor in the standard error of measurement in IQ tests. When it accepted Moore’s score of seventy-four as valid, the CCA treated it as a decisive number rather than applying the standard error of five points, which would bring his score down to sixty-nine and within the range for an intellectual disability.[148]

Second, Moore’s counsel argued that the CCA erroneously applied the adaptive-function prong. They pointed out that it is undisputed in the record that Moore exhibited signs of his disability that would support this prong; at age thirteen he could not understand the days of the week, months, seasons, how to tell time, and even lacked basic math skills such as subtraction, addition, and units of measurement.[149] Texas emphasized how the CCA believed Moore’s “strengths” outweighed his deficits, highlighting his ability to mow grass and play pool for money.[150] Again, Texas’ conclusion is contradicted by clinical standards which state that adaptive skill limitations often coexist with strengths.[151] This argument speaks to the misunderstanding and stereotype that individuals who are not of the “severely mentally retarded” category are not intellectually disabled. Texas grasped for evidence that Moore could function normally in everyday life, when in reality many individuals who fall in the “mildly mentally retarded” category hold jobs and appear to function normally in certain aspects of their lives.

Texas also defended the CCA’s decision by questioning how Moore became disabled. It pointed to poor nutrition, poverty, his history of poor academic performance, and depression while on death row to ultimately argue that these are not attributable to intellectual functioning, but rather are evidence of lack of a good home environment.[152] This argument is directly contradicted by current medical standards which state that intellectual disability can be derived from multiple causations. For instance, the AAMR advocates that etiology has a role in the diagnosis. The etiology approach is a multifactorial construct consisting of four categories: biomedical, social, behavioral, and educational.[153] The AAMR cites as “risk factors” for a disability, the exact reasons Texas argued are not evidence of an intellectual disability; these include malnutrition, family poverty, child abuse and neglect, and institutionalization among others.[154]

In oral argument, counsel for Texas argued that because the DSM-5 states there is an “imperfect fit” between the two concepts of subaverage intellectual functioning and adaptive behavior, states do not have to adopt the positions of current medical organizations.[155] This argument supports adhering to the Briseno factors instead to help clarify the prongs set out by the DSM-5. However, the danger of adhering to the Briseno factors is that they are based on a “consensus of Texas citizens,”[156] and thus based on the layperson’s stereotyped view of intellectual disability. Arguing that the Briseno factors trump clinical consensus violates Atkins. The entire basis for the Atkins decision was a shifting national consensus, based on clinical findings and the medical community because “[t]he [state] statutory definitions of mental retardation are not identical, but generally conform to the clinical definitions.”[157] If the three prongs are based entirely on clinical definitions, how then can a court conclude that they are free to ignore clinical consensus on how to apply those prongs? The Supreme Court’s decision in Atkins suggests an intent to follow the evolving standards of the medical community, and nowhere does the Court condone following the lay persons’ view, which has no basis in comparison to a professional clinical judgment.

Bobby James Moore ultimately prevailed, with Justice Ginsburg writing the 5-3 decision vacating the CCA’s judgment on March 28, 2017.[158] The majority opinion emphasized that although the states are tasked with deciding how to enforce Atkins, their discretion is not “unfettered” and the decision must be “informed by the medical community’s diagnostic framework.”[159] First, in considering IQ score, the Court held CCA’s conclusionthat Moore’s IQ scores establish he is not intellectually disabledis “irreconcilable with Hall,” which mandates that when an IQ score is close to and above seventy, courts must account for the test’s standard error of measurement.”[160] The standard error of measurement (“SEM”) is particularly important because it “reflects the reality that an individual’s intellectual functioning cannot be reduced to a single numerical score.”[161] Accounting for the SEM, Moore’s score of seventy-four actually yields a range of sixty-nine to seventy-nine, because it can be plus or minus five points either way.[162] Justice Ginsburg’s opinion emphasizes the importance that “the Eighth Amendment [does not turn] on the slightest numerical difference in IQ score”[163] and reinforces the importance of considering adaptive behavior deficits, not strengths, when the IQ is around this range. The CCA erred by “overemphasiz[ing] Moore’s perceived adaptive strengths” such his lawn mowing and time living on the streets.[164] The medical consensus is to focus on deficits, not strengths.[165]

Chief Justice Roberts, writing for the dissent, agrees with the majority that the Briseno factors are an “unacceptable method,” but dissents because he believes “clinicians, not judges, should determine clinical standards; and judges, not clinicians, should determine the content of the Eighth Amendment.”[166] In the dissent’s view, it is the evolving standards of decency that matters for Eighth Amendment, not a medical assessment.

Despite Chief Justice Roberts best efforts, Moore held that courts must follow the medical community consensus in determining intellectual disabilitySEMs must be considered, and adaptive strengths cannot be overemphasized. However, Part III will argue that without the retroactive effect of Moore or Hall and without a ban on IQ cutoffs, this decision may not fully protect future intellectuallydisabled defendants or reach those currently sitting on death row.


Bobby James Moore received his justice in Moore v. Texas, but where does that leave the remaining intellectuallydisabled persons sitting on death row? Will future intellectuallydisabled criminal defendants be given the same treatment without a bright-line rule from the Court on IQ cutoffs? Although Moore represents a triumphant moment in the judicial system’s effort to understand mental deficiencies, these two major questions remain given the majority in Moore again left the states some discretion in determining when a defendant is intellectually disabled enough to qualify for the Atkins exemption. This Note will explore (1) whether Moore and Hall can be given retroactive effect, (2) whether a ban on harsh IQ cutoffs is possible to protect current and future intellectually disabled claimants, and (3) how evidence of mental illness must be more seriously considered along with intellectual disability for death penalty exemption.

A.  Retroactivity

States have always resisted complying with Supreme Court ruling they dislike, often finding ways to limit the decisions’ impact. One way lower courts can side-step a Supreme Court decision is to argue that the decision does not apply retroactively. Post-Moore, lower courts have begun pointing to retroactivity as a reason to continue to deny Atkins claims. For example, only a few months after Moore was decided, a federal court in Alabama held Moore is not retroactive, but rather is a mere, new application of Hall.[167] There, petitioner Smith argued that the Alabama court unreasonably applied federal law by failing to apply the SEM adjustment to his IQ score.[168] The Alabama court reasoned that because Moore had not been decided when the Alabama Court of Criminal Appeals entered its decision, there was no error in failing to consider the SEM when examining Smith’s IQ scores.[169] Without applying the SEM, if an IQ score does not go below seventy, the Alabama court can consider a claimant’s adaptive strengths more convincing than the deficits. Alabama viewed Moore as simply cautioning against over-emphasis of adaptive strengths, and with neither Moore or Hall being retroactive,[170] the court dismissed petitioner’s claim despite evidence that his IQ scores ranged from as low as sixty-four to seventy-five.[171] Petitioner Smith filed an appeal in the U.S. Court of Appeals for the Eleventh Circuit on November 9, 2017 and is currently awaiting review. The Supreme Court in Moore and its line of precedents aim to uphold and protect the Eighth Amendment right against cruel and unusual punishment. Yet despite holding in Moore that the SEM and adaptive deficits must be considered, individuals are being denied that right because they happen to appear before a court that does not agree with the Supreme Court. The Eighth Amendment applies to all, and when states to continue to uphold harsh IQ cutoffs of seventy, it goes against the premise that the Court is the “supreme law of the land.” One solution, if possible, is for the Supreme Court to give retroactive effect to Hall and Moore.

The framework for retroactivity in cases on federal collateral review was established in 1989 in Teague v. Lane.[172] As a general matter, a new constitutional rule of criminal procedure does not apply . . . to convictions that were final when the new rule was announced.[173] Teague recognizes two categories of rules that are not subject to this bar: new substantive rules of constitutional law and new “watershed rules of criminal procedure.”[174] Substantive rules include those “forbidding criminal punishment of certain primary conduct” and those “prohibiting a certain category of punishment for a class of defendants because of their status or offense.”[175] This Note seeks to determine whether Moore and Hall are new substantive rules qualifying as an exception to Teague.

The Supreme Court has done this before. In 2011, the Court held in Miller v. Alabama that mandatory life without the possibility of parole for juvenile homicide offenders violates the Eighth Amendment, and such a sentence is disproportionate for all but the “rare juvenile offender whose crime reflects irreparable corruption.”[176] In the wake of Miller, it was unclear whether its holding could be applied retroactively to juvenile offenders whose convictions were final when Miller was decided. In 2016, the Court came back to the issue in Montgomery v. Louisiana to clarify that Miller is retroactive because the Constitution requires state collateral review courts to give retroactive effect when a new substantive rule of constitutional law controls the outcome of a case.[177] Because Miller concluded a sentence of life without parole is disproportionate for the vast majority of juvenile offenders, not giving it retroactivity raised a “grave risk that many are being held in violation of the Constitution.[178] In comparison, the vast majority of intellectual disabled persons fall on the “mild” range (IQ 5055 to 7075). Applying the SEM to the 7075 range could protect individuals whose range goes as high as eighty in one direction, and as low as sixty-five in the other. Like the juveniles Montgomery sought to protect, mildly intellectually disabled persons with IQ scores of 7075 are being denied rights under Atkins, Hall, and Moore in states that cut off protection at seventy, leaving them at a “grave risk” of being detained in violation of the Constitution. Montgomery clarified that when a State enforces a penalty barred by the Constitution, the sentence is unlawful.[179] Here, if states like Alabama are continuing to enforce statutes with IQ cutoffs set at seventy, they are enforcing penalties barred by the Constitution given the Court’s holding in Hall that it violates the Eighth Amendment not to consider adaptive behavior when defendants are in the range close to, but above, 70.[180] Like how Miller’s rule controls the outcome of juvenile life without parole cases, Hall and Moore now control the outcome of Atkins claims with new guidance on how to determine intellectual disability.

The Kentucky Supreme Court agrees and in 2016 held that Hall should be applied retroactively because “[i]t is a substantive restriction on the State’s power to take the life of individuals suffering from intellectual disabilities.”[181] The Kentucky Supreme Court reasoned that Hall is a “directive that not only proscribes intellectually disabled people from being put to death, but defines the manner in which the mental deficiencies of offenders must be evaluated.”[182] Kentucky’s emphasis that following Hall, there is a new method to evaluate intellectually disabled defendants fits within Montgomery’s holding that a new substantive rule exists when it controls the outcome of a case. Hall rejected the bright-line cutoff of seventy because it “create[d] an unacceptable risk that persons with intellectual disability will be executed” and was an unconstitutional violation of the Eight Amendment.[183] Hall’s requirement to consider the SEM and its rejection of an IQ cutoff would change the outcome of many cases where a defendant with IQ scores between seventy and seventy-five brings an Atkins claim in a state like Alabama that employs a harsh cutoff of seventy.

If a petitioner with scores ranging from sixty-four to seventy-five brings a claim in Florida, under Hall courts would be required to consider adaptive deficits. If a petitioner with the same scores brings this claim in Alabama, under its current law the claim likely would be denied in part because Hall and Moore are not retroactive. Inconsistency as such among the states results in similarly situated persons receiving vastly different treatment. The decision between life and death for an individual with an intellectual disability should not hinge on whether they are in a state that follows the clinical approach of following the SEM or not.

The argument against retroactivity should not be overlooked. There is a persuasive point that Hall and Moore cannot be applied retroactively because they announce procedural rules rather than substantive rules. In Montgomery, Louisiana noted that Miller did not categorically bar a penalty for a class of offenders or type of crime, but only mandated a process to follow by creating a set of factors courts must consider when sentencing juveniles to life without parole. Although it was a losing argument for Louisiana, here lower courts might argue that Atkins barred a penalty for the class of intellectually disabled persons constituting a substantive rule, but Hall and Moore do not bar a penalty for a class, but provide guidance to courts on how to make the intellectual disability finding, and thus mandate a process as a procedural rule. It could be argued that Hall and Moore did not place any punishment beyond the State’s power to imposeand this power rests solely with the states. This argument brings the question of whether defendants who are “close to, but above 70” can be considered a “class of defendants.” Nowhere in Hall did the Court say how far above seventy this class could include. The Court applied the SEM to Bobby James Moore’s IQ of seventy-four, so is that where it stops? Applying the SEM, the majority explained it can go five points in either direction, rendering Moore with a range of 6979. If this means someone with a score of seventy-nine could still be considered intellectually disabled, perhaps the solution is to consider defendants who fall between seventy and seventy-nine a class of persons in need of protection from an unconstitutional death sentence.  

The argument against retroactivity also stresses that retroactivity may lead to an increase in frivolous intellectual disability claims. Atkins recognized that “[n]ot all people who claim to be mentally retarded will be so impaired as to fall within the range of mentally retarded offenders,so the question about which group of intellectually disabled individuals there exists a national consensus prohibiting execution is a valid concern.[184] The worry is that courts will receive an influx of frivolous or dishonest habeas petitions alleging intellectual disability. Frivolous litigation, however, is not likely a threat. A recent study found that from the time of Atkins in 2002, through the end of 2013, only 371 death row inmates or capital defendants claimed intellectual disability.[185] This study calculated the filing rate of these 371 persons to be only approximately 7.7% of persons whose lives could potentially be spared by a determination of intellectual disability”a fairly consistent number over the ten years since Atkins.[186] The empirical evidence from this study “also refutes any concern that significant numbers of frivolous claims would be filed.”[187] Given that Atkins did not generate much frivolous litigation, it follows that if retroactive, Hall and Moore would continue the trend of allowing justice for the true claims of those intellectually disabled still on death row, without generating many frivolous claims.

Moore may not need to be applied retroactively if is viewed as simply dealing with the narrow circumstance of Texas’s reliance on nonclinical factors. Yet how many cases in the above study were denied because courts relied too heavily on adaptive strengths rather than deficits like Moore now demands? There is arguably a subset of cases, which were denied pre-Moore and could now be granted post-Moore, that will not see justice because courts are not treating Moore as retroactive. Through 2013, the study found that 31% of all unsuccessful cases were denied for failure on prong one, significantly subaverage intellectual functioning, and 12% were unsuccessful for lack of adaptive deficits.[188] Of those who lost on prong one, 71% had an average IQ score over seventy-five, while the successful cases had an average score of sixty-eight.[189] However, of the successful cases finding intellectual disability, 46% of the [claimants] had at least one IQ score over seventy-five, and 20% [had] one or more IQ scores over 80.[190] It is important to note that these cases were decided before Moore held that the SEM must be applied to the overall scores. Of the 71% with an average over 75, it is possible there are cases with scores between seventy-six and seventy-nine, a range that when the SEM is applied, falls within the same range that Hall and Moore protect.

Many intellectually disabled defendants, like in Moore, have a range of IQ scores over time, but demonstrate adaptive deficits that must be considered and not ignored merely because some of the scores are above seventy. Bobby James Moore had an average IQ of seventy-four, but the Court refrained from setting a brightline rule for IQ scores. On one hand, it could be argued that courts have been generally doing it right, and there is no need for Moore to apply retroactively because the majority of denied cases were above seventy-four. On the other hand, there is still the 29% who had IQ scores below seventy-five, and it is possible that several of those cases were close calls involving evidence of adaptive deficits not properly considered. If states like Alabama continue to render decisions that fly in the face of Hall and Moore, the goals the two cases aimed to accomplish will be continuously undermined.

Finding intellectual disability is not a blackandwhite issue. With Moore requiring application of the SEM and Hall rejecting a harsh IQ cutoff of seventy, it is evident that Atkins claims require case-by-case analysis and that this is not too much to ask considering the relatively low number of claims being brought. The response to this complex analysis is to not continue allowing states to implement harsh IQ cutoffs of seventy. The variety of IQ scores, adaptive deficits, childhood trauma, and mental illnesses from which defendants suffer cannot be boxed into one number. For these reasons, IQ cutoffs, like in Hall, should be banned nationwide by giving Hall and Moore retroactive effect.

B.  Current States with IQ Cutoffs

Texas provided one example of how states made it more challenging for claimants to prevail on an Atkins claim with its nonclinical Briseno factors, but state statutes with IQ cutoffs persist post-Moore. Side-stepping Moore by treating it as not retroactive, states with IQ cutoffs are free to ignore the SEM, adaptive deficits, and claims of those who do not have overall IQ scores below seventy. Although Moore applied the SEM to Moore’s score of seventy-four, yielding a sufficient range of sixty-nine to seventy-nine,  the Court refrained from making a categorical holding as to constitutionality of IQ cutoffs. The Court left it open to the states to decide what a sufficient IQ score is. The opinion makes clear that the SEM should be considered plus and minus five points in both directions, but does not explain whether it should be applied to all IQ scores in a claimant’s life. This leaves the question on where to draw the line with IQ scores. Although Hall held that the cutoff of seventy was unacceptable in Florida, the Court did not make the decision retroactive, as discussed in Section III.A. As a result, states still enforce statutes with unconstitutional IQ cutoffseven after Moore.

Several states have already taken the general position that “[w]hile IQ tests are one of the many factors that need to be considered, they alone are not sufficient to make a final determination on this issue.[191] Currently, eleven states have statutes with IQ cutoffs in their definition of intellectual disability: Arizona, Arkansas, Kentucky, New Mexico, Nebraska, North Carolina, South Dakota, Tennessee, Washington, Idaho, and Oklahoma.[192] Of those eleven, all set the significantly subaverage functioning level at an IQ of seventy and below aside from Arkansas, which states that there is a “rebuttable presumption of mental retardation when a defendant has an intelligence quotient of sixty-five (65) or below.”[193] The Ohio Supreme Court held that there is a rebuttable presumption that a defendant is not mentally retarded if his or her IQ is above 70.”[194]

Although state statutes and procedures post-Atkins differ, if a significant number of states come to a consensus that an IQ cutoff is not sufficient by itself, it could prove to be as influential as the “national consensus” was in Atkins’s overturning Penry. Of the eleven states with IQ cutoffs, several courts have started departing from their brightline rules. Arizona’s statute was directly called into question by Hall, which explained that although it has the brightline set at seventy, another provision of the statute “instructs courts to ‘take into account the margin of error for a test administered.’”[195] Hall cited what it called the “principal Arizona case on the matter” where a defendant had an IQ score of eighty, and “all but one of the sub-parts of the IQ test were ‘above 75.’”[196]

In a case remanded from the Eighth Circuit, a United States District Court in Arkansas on March 2, 2018, performed a thorough analysis to determine the intellectual disability of a defendant claiming an Atkins exemption because a diagnosis “cannot be justified solely on the basis of a fixed score.[197] The court applied both the Flynn Effect and the SEM and considered all evidence of petitioner’s childhood, education, employment, financial abilities, and personal relationships.[198] Although the court ultimately did not find a “significant subaverage general intelligence” and upheld the 1993 Arkansas statute,[199] its analysis demonstrates a promising acceptance of Moore. Alternatively, the Supreme Court of Tennessee, before Moore was decided, held in 2011 that determining a defendant’s functional intelligence is “not limited to raw scores.”[200] The court required expert testimony to assist the determination and a “full and fair consideration” to all evidence, including the results of all IQ tests administered to a defendant.[201] To help inform its decision, the Tennessee Supreme Court reviewed all cases involving the relevant statute and found that neither litigants nor Tennessee courts in general have been limiting their consideration to raw IQ test scores, and there are even cases in which the State has argued and presented evidence challenging the accuracy of scores.[202]

An IQ score must not be the only factor considered and cannot be the only deciding factor for whether an individual qualifies for exemption under Atkins. “Because intelligence tests are indirect rather than direct measures of intelligence, experts in the field recognize that they, like other measures of human functioning, are not ‘actuarial determination[s],’ that these tests cannot measure intelligence with absolute precision and that these tests contain a potential for error.”[203] Indeed, as experts have recognized, it is dangerous to rely solely on IQ tests for proof of intellectual disability. Commonly used tests in the public domain . . . are administered in a group setting with poor or non-existent test control,” and the test itself can be “sketchy” or based on obsolete norms.”[204] The DSM-5 stresses that clinical judgment is [required] in interpreting the results.[205] Since 1959, clinicians have considered a person’s adaptive behavior in addition to IQ scores due to a decreasing confidence in the scores as the sole measure.[206] This suggests that in death penalty cases involving Atkins claims, courts should evaluate adaptive behavior on a casebycase basis, rather than implement a categorical ban like the one states are trying to impose on defendants with IQs above seventy.

In the year since Moore was decided, the Supreme Court has remanded several cases for additional consideration in light of Moore.[207] If this is to be a continuing trend, for pure judicial economy reasons it would make sense to have Hall and Moore apply retroactively. In October 2017, the United States Supreme Court ordered the Florida Supreme Court to reconsider, in light of Moore, a decision denying death-row inmate Tavares Wright’s intellectual disability claim.[208] All of Wright’s nine IQ tests yielded scores of seventy-five or above[209] and the Supreme Court of Florida listed several adaptive strengths such as Wright’s job as a grocery clerk, job in prison, ability to write cards, cleanliness, and understanding of social interactions[210]yet the Supreme Court found this was insufficient under Moore. This was the sixth time the Court has vacated a state or federal court’s rejection of an intellectual-disability claim and remanded for reconsideration under Moore.[211]

The Supreme Court also vacated a decision of the Alabama Court of Criminal Appeals (“CCA”) in May 2017 for Taurus Carroll because Alabama had overemphasized Carroll’s adaptive strengthsthat he had passed a GED exam . . . and . . . held . . . a job in the prison kitchen.[212] On remand, the court reconsidered evidence that Carroll was in special education as a child, twice failed first and eighth grade, had an absent father, and experienced physical and sexual abuse as young as age seven.[213] However, the Alabama CCA again denied Carroll’s claim of intellectual disability, citing evidence that he had a good memory, knew dates and times of day, and had adequate school records, despite an IQ score of seventy-one that would adjust to sixty-six to seventy-six with the SEM.[214] Despite succeeding on prong one, Carroll lost his claim on prong two’s adaptive functioning requirement. The Alabama CCA pointed to the fact that in high school, Carroll was given the Wechsler Intelligence Scale for Children twice and received scores of eighty-five and eighty-seven.[215]

A recent report from 2017 revealed that almost all twenty-six men scheduled for execution in Ohio over the next three years suffer from mental, emotional, or cognitive impairments or limitations; at least 11 have evidence of intellectual disability, borderline intellectual disability, or a cognitive impairment, including brain injury.[216] Stanley Fitzpatrick, sentenced to death for murder committed at nineteen, not only suffered from hallucinations that the devil appeared to him and he “saw demons,” but he also had a “devastatingly low [IQ of] 69.”[217] Fitzpatrick’s death sentence was affirmed by the Ohio State Supreme Court in 2004, but his defense lawyers never introduced his low IQ during the penalty phase, and neither did they pursue evaluations to argue that he qualified as intellectually disabled.[218] James Frazier failed first grade, was a “slow learner,” attended “special classes,” dropped out of high school, and has an IQ of seventy-two.[219] Frazier was brought up in a household with a weekly wage of $64, with no supervision, and he was sexually abused as a child. Frazier’s Atkins claim was denied by the Supreme Court of Ohio in 2007.[220] James Derrick O’Neal had a reported IQ score of sixty-four at age fourteen, with three other scores of sixty-three, sixty-seven, and seventy-one.[221] O’Neal’s death sentence was affirmed by the Supreme Court of Ohio in 2000, two years before Atkins was decided, and he currently sits on death row. David Sneed suffers from both mental illness and impaired intellectual functioning, about which two psychiatrists testified at the penalty phase “combined to prevent him from appreciating the criminality of his actions.”[222] Sneed’s conviction was affirmed by the Supreme Court of Ohio in 1992, and he currently sits on death row. Lastly, Angelo Fears, with an IQ of seventy-five, family history of mental illness, and traumatic childhood of beatings, sits on death row following an affirmation of his sentence by the Ohio Supreme Court in 2008.[223]

These cases represent complex issuesthe presence of mental illness, childhood trauma, and drug use, along with evidence of intellectual disabilitybut all together they show that the process of finding an Atkins exemption cannot be merely a numbers game. Based on the cases above, Ohio does not give proper consideration, as required by Moore, to other factors when IQ scores are above seventy. It is wholly inconsistent with Moore to count evidence of mental illness or trauma against a finding of intellectual disability. In Moore, the majority found that the CCA had erred in failing to appreciate his childhood trauma and requiring a showing that Moore’s adaptive deficits were not related to a “personality disorder.”[224] It is now recognized that many intellectually disabled people also have other mental or physical impairments such as depression, bipolar disorder, and autism.[225]

Neither the AAMR nor the DSM diagnostic criteria intend that a fixed IQ cutoff be used to diagnose intellectual disabilities, given the known measurement of errors. Hall and Moore make clear that when an individual is “close to, but above, 70, courts must account for the test’s standard error of measurement.[226] The difficulty lies with how far from seventy scores can deviate or how many can be over seventy when there are multiple IQ scores over the developmental period, but a bright-line rule is not the answer.

C.  A Shifting National Consensus Against the Death Penalty

The issue of how to determine intellectual disability for the purposes of Atkins rests on the fundamental idea that those individuals are less culpable due to their diminished capacity.[227] Culpability is at the center of the analysis for death penalty purposes, so it is necessary to take a step back and look at the arguments in favor and against the death penalty in general, without narrowing it to the intellectually disabled. The Supreme Court has expressed a trend away from imposing the death penalty, and today “academic defenders of the death penalty are few and far between.[228] Simply put, the death penalty is “almost universally agreed . . . at worst barbaric and at best a waste of money.”[229]

The few academic defenders of the death penalty take the position that the death penalty, in certain cases, can be “morally required . . . to prevent the taking of innocent lives.”[230] This argument assumes the death penalty has a deterrent effect. A study in 2003 found that “each execution prevents some eighteen murders, on average, which supports defenders of the death penalty and the theory of deterrence.[231] As discussed in Part II, deterrence is not necessarily at play when criminal defendants are intellectually disabled, and in the last decade, opinions have changed and new studies have come out. In fact, a 2012 report by the National Research Council found that studies claiming the death penalty has a deterrent effect on murder rates have “fundamental flaws” and should not be relied on for policy decisions.[232]

Further, simply because the death penalty fits a theory of punishment does not mean it can be justified all things considered. For instance, torture could potentially fulfill the purposes of punishment, but as many Americans agree, it might not be permissible on other moral [or] legal grounds.[233] While the death penalty may fit a theory of punishment in general terms, here, in the case of intellectually disabled persons, it does not because of their inability to consider the risks and consequences of their actions, resulting in their overall diminished culpability. To otherwise justify the death penalty for intellectually disabled defendants is to move outside the theories of punishment, against Atkins’ position that “[u]nless the imposition of the death penalty on a mentally retarded person ‘measurably contributes to one or both of these goals, it is nothing more than the purposeless and needless imposition of pain and suffering, and hence an unconstitutional punishment.[234] By failing to consider adaptive behavior in determining exemption under Atkins, states are failing to uphold Atkins by allowing individuals like Stanley Fitzpatrick, James Frazier, and the others on Ohio’s death row to be subjected to needless pain and suffering.

Because the death penalty does not contribute to the goals of deterrence or retribution as applied to intellectually disabled, it must go beyond these theories as the Court did in Atkins by considering the national consensus. States are trending away from executing criminal defendants and are abolishing capital punishment altogether. The numbers are striking. In 2012, forty-three executions took place, thirty-nine in 2013, thirty-five in 2014, and twenty-eight in 2015.[235] Only twenty individuals were executed in 2016.[236] Further, the number of death sentences per year has also dropped dramatically from 279 in 1999, to only thirty-one in 2016.[237] These statistics encompass all criminal defendants and could include murderers and rapists with IQs well over seventy and the ability to fully understand the consequences of their actions, yet states are trending against sentencing those individuals to death. If states are shifting that way, they certainly should agree that defendants with even lesser culpability should not be executed either.

D.  Expanding the Exemption

With a constitutional ban on executing intellectually disabled persons, it follows that there should be a serious conversation regarding the execution of mentally ill defendants to find an avenue for exemption other than the rarely used insanity defense.

Following the Court’s decision in Atkins, the American Bar Association (“ABA”) has taken several steps towards enforcing and expanding its precedential value. The ABA established a Task Force on Mental Disability and the Death Penalty, which deliberated from 2003 to 2005.[238] The Task Force was comprised of roughly twenty-four lawyers and mental health practitioners, as well as members of the American Psychiatric Association and American Psychological Association. The Task Force successfully put together a proposal which became the ABA’s official recommendation on the death penalty exemption postAtkins.[239] The ABA formally takes a position that goes beyond the scope of Atkins, calling for exemption from the death penalty of not only individuals with an intellectual disability, but also those with serious mental illnesses.[240] The ABA put forth this recommendation in its 2006 122A Recommendation, outlining two scenarios in which defendants should not be executed: those with significant limitations in both intellectual functioning and adaptive skills, and those with severe mental disabilities.[241] The language from the Recommendation is as follows:

1. Defendants should not be executed or sentenced to death if, at the time of the offense, they had significant limitations in both their intellectual functioning and adaptive behavior, as expressed in conceptual, social, and practical adaptive skills, resulting from mental retardation, dementia, or a traumatic brain injury;

2. Defendants should not be executed or sentenced to death if, at the time of the offense, they had a severe mental disorder or disability that significantly impaired their capacity (a) to appreciate the nature, consequences or wrongfulness of their conduct, (b) to exercise rational judgment in relation to conduct, or (c) to conform their conduct to the requirements of the law.[242]

The first paragraph is essentially the ABA’s definition of “mental retardation” as taken from the American Association of Mental Retardation, and is thus in line with Atkins except that it also encompasses dementia and traumatic brain injury. This highlights an important and often overlooked problem in the three-prong criteria set out for defining an intellectual disability: the requirement that the deficiencies manifest prior to eighteen does not consider later dementia or brain injury. For instance, what happens to a defendant who grew up at a normal level of intellectual functioning, but was in a car accident after turning eighteen in which they suffered traumatic brain injury rendering them unable to function at the same level? This recommendation would serve as an important step towards expanding protection to those rare individuals who suffer from late on-set deficiencies.

The second paragraph, strikingly, is an unprecedented call for exemption of defendants with severe mental disorders or disabilities. The ABA elaborates that this narrowly refers to only those with “severe” disorders, meaning those disorders that mental health professionals would consider to be on “Axis I diagnoses.”[243] Among these include schizophrenia and other psychotic disorders, mania, major depressive disorder, and dissociative disorders, all of which are associated with delusions, hallucinations, disorganized thinking, and disruption of consciousness, memory, and perception of the environment.[244] Although this is seemingly a shift towards a more inclusive mental health law system, the ABA was sure to impose limitations in its recommendation by requiring a “significant impairment” requirement for individuals with severe mental disorders or disabilities. This requires that the disorder significantly impair cognitive or volitional functioning at the time of the offense.”[245]

The recommendation is in line with the goals of the criminal justice system. Defendants with established disorders that are considered “severe” enough to fall on the “Axis I diagnoses” certainly lack the ability to make rational decisions in comparison to the average offender without a mental disorder. In fact, the effect that a serious mental disorder has on a person’s culpability is essentially the same as the effect that being a juvenile or having an intellectual disability has on a person’s culpability. If the individual meets the “significant impairment” requirement discussed above, then the individual certainly lacked proper decision-making abilities at the time of their offense. Thus, like the Court reasoned in Roper and Atkins, punishing a person with lesser or low culpability due to a mental disorder does not properly serve the goals of deterrence since they lack the capacity to understand what they did or lacked the capacity to fully understand the consequences of their actions at the time of their offense.

Expanding the exemption to serious mental disorders is also important because the “scientific and clinical definitions emphasize that individuals with mental retardation often have mental disorders as well.”[246] This was an issue in Williams v. Quarterman, where an individual meeting the IQ cutoff of seventy faced a different problem: the Fifth Circuit interpreted evidence of his social and practical skill deficits as “bizarre and antisocial conduct,” demonstrating characteristics that are “just as easily seen as attention-getting behaviors as they are evidence of mental retardation.”[247] The Fifth Circuit found that these characteristics “could be explained by anti-social personality rather than mental retardation.”[248] This is clearly not in line with the rationale the Supreme Court has been implementing in its series of cases, from Roper to Moore, because someone exhibiting effects of both an intellectual disability and mental disorder absolutely lacks the mental culpability necessary to impose the death penalty. To discount an intellectual disability due to evidence of mental illness would be a grave practice threatening the constitutional rights and liberty of a class which has long suffered under perpetuated stereotypes of mental disability and illness.


The Supreme Court in Moore reinforced the long-standing theme that “[t]o enforce the Constitution’s protection of human dignity, we look to the evolving standards of decency that mark the progress of maturing society.”[249] Intellectual disability involves complex factors that cannot be reduced to a single IQ score. An evaluation of state statutes, current litigation, and the Supreme Court’s stance since Atkins reveals a trend away from bright-line IQ rules of seventy, and towards taking a holistic, case-by-case approach to Atkins’ claims. The individuals who need protectionwhom Atkins seeks to protectare the “mildly retarded” individuals who live successfully in the community, either independently or in supervised settings, and who have jobs, families, maintain a home, and even raise children. A mere number cannot define whether someone is intellectually disabled;[250][t]he term ‘intellectual disability’ does not refer to a single disorder or disease, but rather to a heterogeneous set of disabilities that affect the level of a person’s functioning in defined domains.”[251] These are people who engage in actions with lesser culpabilities than normallevelfunctioning people, but who are not recognized as needing protection because of lasting stereotypes that intellectually disabled persons are only the severe, Lennie character types.

[*] *. Managing Editor, Southern California Law Review, Volume 91; J.D. 2018, University of Southern California Gould School of Law; B.A. Political Science 2015, University of San Diego. I would like to thank Professor Saks for her invaluable guidance and feedback on earlier drafts of this note. In addition, I would like to thank the staff and editors of the Southern California Law Review for their excellent work.


 [1]. John Steinbeck, Of Mice and Men 263 (Penguin Books 1993) (1937).

 [2]. Moore v. Texas, 137 S. Ct. 1039, 1044 (2017).

 [3]. Id. at 1045.

 [4]. Id.

 [5]. Id. at 1047.

 [6]. Id. at 1045.

 [7]. See id. at 1044. See also Ex parte Briseno, 135 S.W.3d 1, 6 (Tex. Crim. App. 2004).

 [8]. Moore, 137 S. Ct. at 1053.

 [9]. Id. (citation omitted).

 [10]. Atkins v. Virginia, 536 U.S. 304, 321 (2002).

 [11]. Id. at 316-17 At the time the Supreme Court first heard cases on this issue, the terminology used was “mental retardation,” however today the DSM-5 has changed the term to “intellectual disability.” See Am. Psychiatric Ass’n, Diagnostic and Statistical Manual of Mental Disorders 33 (5th ed. 2013) [hereinafter DSM-5].

 [12]. Atkins, 536 U.S. at 316–17.

 [13]. Denis Keyes, William Edwards & Robert Perske, People with Mental Retardation Are Dying, Legally: At Least 44 Have Been Executed, 40 Mental Retardation 243 (2002).

 [14]. Atkins, 536 U.S. at 316–17 (citation omitted).

 [15].                             States that Have Changed Their Statutes to Comply with the Supreme Court’s Decision in Atkins v. Virginia, Death Penalty Info. Ctr., (Mar. 2, 2018), https://deathpenaltyinfo.org/states-have-changed-their-statutes-comply-supreme-courts-decision-atkins-v-virginia.

 [16]. Hall v. Florida, 134 S. Ct. 1986, 1995, 2001 (2014).

 [17]. The CCA is Texas’s court of last resort in criminal cases. See Tex. Const. art. V, § 5.

 [18]. See Ex parte Briseno, 135 S.W.3d 1, 5–6 (Tex. Crim. App. 2004).

 [19]. See id.

 [20]. U.S. Const. amend. VIII.

 [21]. Penry v. Lynaugh, 492 U.S. 302, 330–31 (1989).

 [22]. Id. at 331.

 [23]. See Roper v. Simmons, 543 U.S. 551, 578 (2005).

 [24]. See Ford v. Wainwright, 477 U.S. 399, 401 (1986).

 [25]. See Atkins v. Virginia, 536 U.S. 304, 316–17 (2002).

 [26]. Penry, 492 U.S. at 333–34.

 [27]. Id.

 [28]. Id.

 [29]. See Atkins, 536 U.S. at 314­–15. The seventeen states listed include: Kentucky and Tennessee in 1990; New Mexico in 1991; Arkansas, Colorado, Washington, Indiana, and Kansas in 1993 and 1994; New York in 1995, Nebraska in 1998; South Dakota, Arizona, Connecticut, Florida, Missouri, North Carolina, and Texas. When Atkins was decided, Texas had just passed a similar bill, and both Virginia and Nevada had similar bills passed in at least one house, but it was not yet law in these states. Id.

 [30].               See id. at 315–16.

 [31].               See id.

 [32]. Id.

 [33]. See id. (noting that New Hampshire and New Jersey were two states that continued to allow execution sentences, but had not actually carried one out in decades).

 [34]. Id.

 [35]. Id. at 318–20.

 [36]. Id. at 319.

 [37]. Id. (citation omitted).

 [38]. Id. at 320.

 [39]. Id.

 [40]. Id.

 [41]. See id. at 320–21.

 [42]. Id. at 317.

 [43]. Id. (citation omitted).

 [44]. Id.

 [45]. Examples of “deficits in intellectual functions” can be demonstrated by the level of an individual’s “reasoning, problem solving, planning, abstract thinking, judgment, academic learning, and learning from experience.” DSM-5, supra note 11, at 33.

 [46]. Limitations in adaptive skills refer to “the inability to learn basic skills and adjust behavior to changing circumstances.” Hall v. Florida, 134 S. Ct. 1986, 1994 (2014).

 [47]. Atkins, 536 U.S. at 318.

 [48]. Am. Ass’n on Intellectual and Developmental Disabilities, Intellectual Disability: Definition, Classification, and Systems of Supports 41 (11th ed. 2010) [hereinafter AAIDD].

 [49]. See Bryan Lester Dupler, Capital Cases Involving Mental Retardation, in 93 American Jurisprudence Trials 1, §§ 17–19 (Westlaw 2018) (2004).

 [50]. DSM-5, supra note 11, at 37.

 [51]. Dupler, supra note 49, § 11.

 [52]. Id. § 12 (citation omitted) (Persons with mild retardation “‘typically develop social and communication skills during the preschool years . . . have minimal impairment in sensorimotor areas, and often are not distinguishable from children without mental retardation until a later age.’ They can acquire basic academic skills up to about the sixth grade level.”).

 [53]. Id.

 [54]. Id.

 [55]. AAIDD, supra note 48, at 3, 11.

 [56]. DSM-5, supra note 11, at 37.

 [57]. AAIDD, supra note 48, at 44; DSM-5, supra note 11, at 37.

 [58]. DSM-5, supra note 11, at 37.

 [59]. Id.

 [60]. Id.

 [61]. Dupler, supra note 49, § 11.

 [62]. Atkins v. Virginia, 536 U.S. 304, 318 (2002).

 [63]. DSM-5, supra note 11, at 33.

 [64]. See Hall v. Florida, 134 S. Ct. 1986, 1994 (2014); Moore v. Texas, 137 S. Ct. 1039, 1055–56 (2017).

 [65]. Black v. Carpenter, 866 F.3d 734, 738 (6th Cir. 2017).

 [66]. Id. at 748­–49.

 [67]. Ex parte Briseno, 135 S.W.3d 1, 1 (Tex. Crim. App. 2004).

 [68]. At the time Briseno was decided in 2004, DSM-5 was not yet published and DSM-IV was the current edition.

 [69]. Briseno, 135 S.W.3d at 6.

 [70]. Id. at 5–6 (citation omitted).

 [71]. Id. (citation omitted).

 [72]. Id. (“‘Mental Retardation’ means significant subaverage general intellectual functioning that is concurrent with deficits in adaptive behavior and originates during the developmental period.”).

 [73]. Id. at 6–7. Texas did however make another attempt with Tex. H.B. 614, 78th Leg., R.S. (2003), but its definition of “mental retardation” did not significantly differ from Tex. H.B. 236, 77th Leg., R.S. (2001), and “[n]either of [the] bills addressed the issue of determining mental retardation claims on a post-conviction habeas corpus writ brought by inmates sentenced to death before the Supreme Court decision in Atkins.” See id. at 7 n.22.

 [74]. Id. at 7–8, 13.

 [75]. Tex. Health & Safety Code Ann. § 591.003(7-a) (West 2015).

 [76]. Briseno, 135 S.W.3d at 6.

 [77]. Id. at 5 (quoting Atkins v. Virginia, 536 U.S. 304, 317 (2002)).

 [78]. See id. at 6.

 [79]. Id. at 8.

 [80]. Id. at 8–9.

 [81]. Id. at 14.

 [82]. Id. at 15, 18.

 [83]. Moore v. Texas, 137 S. Ct. 1039, 1059–60 (2017).

 [84]. Brief for the American Civil Liberties Union & the ACLU of Texas as Amici Curiae Supporting Petitioner at 28, Moore v. Texas, 137 S. Ct. 1039 (2017) (No. 15-797) [hereinafter ACLU] (citation omitted).

 [85].  Robin M. Maher, Moore v. Texas: The Supreme Court Limits State Discretion to Make the ‘Protection of Human Dignity’ a Reality for the Intellectually Disabled, Geo. Wash. L. Rev. On the Docket (Apr. 9, 2017), http://www.gwlr.org/moorevtexas.

 [86].  Atkins v. Virginia, 536 U.S. 304, 311–12 (2002) (citation omitted).

 [87]. Hall v. Florida, 134 S. Ct. 1986, 1990 (2014). The Court uses “intellectual disability” to mean the same as “mental retardation,” and noted that the change in terminology is used by professionals and approved by DSM-5. Id.

 [88]. Id. at 1991.

 [89]. Id. at 1994 (citation omitted).

 [90]. See id. (“[The Florida Supreme Court] has held that a person whose test score is above 70, including a score within the margin for measurement error, does not have an intellectual disability and is barred from presenting other evidence that would show his faculties are limited.”).

 [91]. Id. at 1990–92.

 [92]. Id. at 1992.

 [93]. Id.

 [94]. Id. at 1994.

 [95]. Id. at 1995.

 [96]. Id.

 [97]. Id. The SEM is considered “[o]ne of the most important concepts in measurement theory,” because “[a]n individual’s IQ test score on any given exam may fluctuate” for several reasons including: the person’s health, how many tests they have taken in the past and thus can remember how to do well on them, the environment they take the test in, the behavior of the examiner administering it, the “subjective judgment involved in scoring certain questions,” and even simple luck. Id.

 [98]. Id.

 [99]. Id. at 1996.

 [100]. Id. (citations omitted). However, the Court notes that “Arizona, Delaware, Kansas, North Carolina, and Washington have statutes which could be interpreted to provide a bright-line cutoff leading to the same result that Florida mandates in its cases.” Id.

 [101]. Id. at 1997.

 [102]. Id.

 [103]. See id. at 1998 (citation omitted) (“The rejection of the strict 70 cutoff in the vast majority of States and the ‘consistency in the trend,’ toward recognizing the SEM provide strong evidence of consensus that our society does not regard this strict cutoff as proper or humane.”).

 [104]. Id. at 1994.

 [105]. See id. at 1990–91 (discussing how Hall’s teachers described him as “mentally retarded,” his lawyer testified that he “[c]ouldn’t really understand anything [Hall] said” and compared him to his four-year-old daughter, and medical clinicians testified that he was “significantly retarded.”).

 [106]. Id. at 1991.

 [107]. Id.

 [108]. Id.

 [109]. Id. at 1990–91 (“This rigid rule, the Court now holds, creates an unacceptable risk that persons with intellectual disability will be executed, and thus is unconstitutional.”).

 [110]. Case Files:               Moore v. Texas, SCOTUSblog, http://www.scotusblog.com/case-files/cases
/moore-v-texas (last visited May 24, 2018).

 [111]. Ex parte Moore, 470 S.W.3d 481, 490 (Tex. Crim. App. 2015), cert. granted in part sub nom Moore v. Texas, 136 S. Ct. 2407 (2016), vacated and remanded by 137 S. Ct. 1039 (2017).

 [112]. Id. at 484.

 [113]. Id.

 [114]. See id. at 489.

 [115]. Id. at 528.

 [116]. Id. at 495, 506.

 [117]. Id. at 542.

 [118]. Miller v. Alabama, 567 U.S. 460, 476–78 (2012) (discussing the hallmark features of youth which include immaturity, impetuosity, failure to appreciate risks and consequences, and what the child’s family and home environment is like from which they cannot usually extricate themselves).

 [119]. Moore, 470 S.W.3d at 509–10.

 [120]. Id. at 495.

 [121]. Id. at 496.

 [122]. Id.

 [123]. Id. at 515.

 [124]. Id. at 506.

 [125]. Id.

 [126]. Id. at 507.

 [127]. Id. at 511.

 [128]. Id. at 486, 513.

 [129]. Id. at 514.

 [130]. Id.

 [131]. See Dupler, supra note 49, § 18.

 [132]. Moore, 470 S.W.3d at 515.

 [133]. Id. at 520.

 [134]. Id. at 520–21 (Greenspan also denied the following as evidence of adaptive skills: “(1) in preparation for his new punishment trial, consulting with counsel about whether to inform the jury that he had been on death row; (2) concealing a shotgun in a shopping bag when entering a store to rob it; (3) attempting to conceal his appearance during the offense by wearing a wig and sunglasses, and after the offense, changing his appearance by shaving his head; (4) arguing with accomplices over how to divide the proceeds of the crime; (5) deciding to stipulate that he had prior criminal convictions … (6) writing four letters to his appellate lawyer …[(7)] hustling pool; and [(8)] working as a barber and a porter in prison.”).

 [135]. See id. at 526.

 [136]. Id. at 524, 526.

 [137]. Id. at 526.

 [138]. Id. at 527.

 [139]. See ACLU, supra note 84, at 3–6.

 [140]. See id. at 2.

 [141]. See id. at 19–24 (discussing each factor and how they were modeled after the character Lennie to exemplify a stereotype based off someone with severe disability).

 [142]. Id. at 9.

 [143]. Id.

 [144]. Atkins v. Virginia, 536 U.S. 304, 308 (2002).

 [145]. Hall v. Florida, 134 S. Ct. 1986, 1994–95 (2014).

 [146]. Transcript of Oral Argument at 9–11, Moore v. Texas, 137 S. Ct. 1039 (2017) (No. 15–797) [hereinafter Transcript] (explaining that the Briseno court viewed the medical standards that were current at the time of its decision as “exceedingly subjective,” and instead came up with nonclinical factors based on “lay stereotypes”).

 [147]. Id. at 11–12.

 [148]. Id. at 16–17.

 [149]. Id. at 20.

 [150]. Ex parte Moore, 470 S.W. 3d 481, 520 (Tex. Crim. App. 2015).

 [151]. AAIDD, supra note 48, at 45.

 [152]. Transcript, supra note 146, at 17.

 [153]. AAIDD, supra note 48, at 61.

 [154]. Id. at 60.

 [155]. Transcript, supra note 146, at 30–31.

 [156]. Ex parte Briseno, 135 S.W.3d 1, 6 (Tex. Crim. App. 2004).

 [157]. Hall v. Florida, 134 S. Ct. 1986,1999 (2014) (citation omitted).

 [158]. See generally Moore v. Texas, 137 S. Ct. 1039 (2017).

 [159]. Id. at 1047 (citation omitted).

 [160]. Id. at 1049.

 [161]. Id. (citation omitted). See also AAIDD, supra note 48, at 2223; DSM-5, supra note 11, at 37.

 [162]. Moore, 137 S. Ct. at 1060.

 [163]. Id. at 1061.

 [164]. Id. at 1050.

 [165]. AAIDD, supra note 48, at 47 (“[S]ignificant limitations in conceptual, social, or practical adaptive skills [are] not outweighed by the potential strengths in some adaptive skills.”).

 [166]. Moore, 137 S. Ct. at 1054 (Roberts, J., dissenting).

 [167]. See Smith v. Dunn, No. 2:13-CV-00557-RDP, 2017 U.S. Dist. LEXIS 113862, at *10–13 (N.D. Ala. July 21, 2017). An appeal has been filed to the Eleventh Circuit.  

 [168]. Id. at *3.

 [169]. Id. at *10–13.

 [170]. Id. at *13–16.

 [171]. Smith v. State, 112 So. 3d 1108, 1128–29 (Ala. Crim. App. 2012).

 [172]. See               Teague v. Lane, 489 U.S. 288, 300 (1989).

 [173]. Montgomery v. Louisiana, 136 S. Ct. 718, 728 (2016) (discussing Teague, 489 U.S. at 307).

 [174]. Id. See also Teague, 489 U.S. at 311.

 [175]. Montgomery, 136 S. Ct. at 728 (citation omitted).

 [176]. Miller v. Alabama, 567 U.S. 460, 479–80 (2012).

 [177]. Montgomery, 136 S. Ct. at 729.

 [178]. Id. at 736.

 [179]. See infra Section III.B for a full discussion on state statutes similar to Alabama’s.

 [180]. Moore v. Texas, 137 S. Ct. 1039, 1049 (2017).

 [181]. White v. Commonwealth, 500 S.W.3d. 208, 215 (Ky. 2016).

 [182]. Id.

 [183]. Hall v. Florida, 134 S. Ct. 1986, 1990 (2014).

 [184]. Atkins v. Virginia, 536 U.S. 304, 317 (2002).

 [185]. John H. Blume, et al., A Tale of Two (And Possibly Three) Atkins: Intellectual Disability and Capital Punishment Twelve Years After the Supreme Court’s Creation of a Categorical Bar, 23 Wm. & Mary Bill of Rts. J. 393, 396 (2014).

 [186]. Id. at 396–97.

 [187]. Id. at 397.

 [188]. Id. at 400–01.

 [189]. Id. at 402.

 [190]. Id. at 404.

 [191]. State v. Lott, 779 N.E.2d 1011, 1014 (2002) (citing Murphy v. State, 54 P.3d 556, 568, 573–74 (Okla. Crim. App. 2002)).

 [192]. See Ariz. Rev. Stat. Ann. § 13-735(F) (2017) (validity called into question by Hall v. Florida, 134 S. Ct. 1986, 1996–97 (2014)); Ark. Code Ann. § 5-4-618(a)(2) (2017); Idaho Code § 19-2515A(1)(b) (2017); Ky. Rev. Stat. Ann. § 532.130(2) (West 2017) (limited by White v. Commonwealth, 500 S.W.3d 208, 214 (2016) (holding in light of Hall, “trial courts in Kentucky must consider an IQ test’s margin of error” as well as “additional evidence of intellectual disability”)); Neb. Rev. Stat. § 28-105.01(3) (2017); N.M. Stat. Ann. § 31-9-1.6(E) (2018); N.C. Gen. Stat. § 15A-2005(a)(1)(c) (2017); Okla. Stat. tit. 21, § 701.10b(A)(3) (2017); S.D. Codified Laws § 23A-27A-26.2 (2017); Tenn. Code Ann. § 39-13-203(a)(1) (2017); Wash. Rev. Code § 10.95.030 (2017) (held unconstitutional on other grounds by State v. Bassett, 198 Wash. App. 714 (2017)).

 [193]. Ark. Code Ann. § 5-4-618 (2017).

 [194]. Lott, 779 N.E.2d at 1014.

 [195]. Hall v. Florida, 134 S. Ct. 1986, 1996 (2014) (citation omitted).

 [196]. Id. at 1996–97 (citation omitted).

 [197]. Sasser v. Kelley, No. 4:00-CV-04036, 2018 WL 1147102, *10–12 (W.D. Ark. Mar. 2, 2018).

 [198]. Id.

 [199]. Id. at *12.

 [200]. Coleman v. State, 341 S.W.3d 221, 221, 224 (Tenn. 2011).

 [201]. Id. at 241–42 (citation omitted).

 [202]. Id. at 247–48.

 [203]. Id. at 245 (citation omitted).

 [204]. Dupler, supra note 49, § 15.

 [205]. DSM-5, supra note 11, at 37.­­

 [206]. AAIDD, supra note 48, at 43–44.

 [207]. See generally Long v. Davis, 663 Fed. Appx. 361 (5th Cir. 2016), judgement vacated, 138 S. Ct. 72 (2017); Weathers v. Davis, 659 Fed. Appx. 778 (5th Cir. 2016), judgement vacated, 138 S. Ct. 315 (2017); Martinez v. Davis, 653 Fed. Appx. 308 (5th Cir. 2016), judgement vacated, 137 S. Ct. 1432 (2017); Henderson v. Stephens, 791 F.3d 567 (5th Cir. 2015), judgement vacated, 137 S. Ct. 1450 (2017).

 [208].               Supreme Court Directs Florida to Reconsider Intellectual Disability Decision in Death Penalty Case, Death Penalty Info. Ctr., https://deathpenaltyinfo.org/node/6902 (last visited May 14, 2018) [hereinafter Florida, Death Penalty Info.].  See also Wright v. State, 213 So. 3d 881 (Fla. 2017), cert. granted, vacated, 138 S. Ct. 360 (2017).

 [209]. Wright, 213 So. 3d at 897.

 [210]. Id. at 899–901.

 [211]. Florida, Death Penalty Info. Ctr., supra note 208.

 [212]. Supreme Court Tells Alabama to Reconsider the Factors It Has Used to Determine Intellectual Disability, Death Penalty Info. Ctr., https://deathpenaltyinfo.org/node/6756 (last visited May 24, 2018).

 [213]. Caroll v. State, No. CR-12-0599, 2017 WL 6398236, at *4–5 (Ala. Crim. App. Dec. 15, 2017).

 [214]. Id. at *5–6.

 [215]. Id. at *6.

 [216]. New Report: Prisoners on Ohio’s Execution List Defined by Intellectual Impairment, Mental Illness, Trauma, and Young Age, Fair Punishment Project (Aug 29, 2017) http://fairpunishment.org

 [217]. Id. Pre-Atkins, Fitzpatrick’s lawyers did not assert the defense of intellectual disability in the last review of his case. See also State v. Fitzpatrick, 810 N.E.2d 927, 932 (2004).

 [218]. Fair Punishment Project, supra note 216.

 [219]. Id.

 [220]. State v. Frazier, 873 N.E.2d 1263, 1263 (2007).

 [221]. Fair Punishment Project, supra note 216. See also State v. O’Neal, 721 N.E.2d 73 (Ohio 2000).

 [222]. Fair Punishment Project, supra note 216. See also State v. Sneed, 584 N.E.2d 1160 (Ohio 1992).

 [223]. Fair Punishment Project, supra note 216. See also               Fears v. Bagley, No. 1:01-cv-183, 2008 WL 2782888 (S.D. Ohio July 15, 2008), aff’d, 462 F. App’x 565 (6th Cir. 2012).

 [224]. Moore v. Texas, 137 S. Ct. 1039, 1051 (2017).

 [225]. Id. (citation omitted).

 [226]. Id. at 1049.

 [227]. Atkins v. Virginia, 536 U.S. 304, 349–50 (2002) (Scalia, J., dissenting) (citation omitted) (“[T]he ‘diminished capacities’ of the mentally retarded raise a ‘serious question’ whether their execution contributes to the ‘social purposes’ of the death penalty, viz., retribution and deterrence.”).

 [228]. Chad Flanders, The Case Against the Case Against the Death Penalty, 16 New Crim. L. Rev. 595, 596 (2013).

 [229]. Id.

 [230]. Cass R. Sunstein & Adrian Vermeule, Is Capital Punishment Morally Required? Acts, Omissions, and Life-Life Tradeoffs, 58 Stan. L. Rev. 703, 705, 711 (2005) (citing a study used from 3,054 U.S. counties between 1977 and 1996, in which it was “found that the murder rate is significantly reduced by both death sentences and executions,” and that “each execution results in eighteen fewer murders”). See also Hashem Dezhbakhsh et al., Does Capital Punishment Have a Deterrent Effect? New Evidence from Postmoratorium Panel Data, 5 Am. L. & Econ. Rev. 344, 344 (2003) (arguing “that capital punishment has a strong deterrent effect”).

 [231]. See Sunstein & Vermeule, supra note 230, at 706.

 [232]. See Comm. on Deterence and the Death Penalty, Nat’l Research Council, Deterrence and the Death Penalty 4 (Daniel S. Nagin & John V. Pepper eds., 2012). The study provides three reasons for why these studies are fundamentally flawed: (1) they ignore impact of noncapital punishment; (2) the studies use unrealistic assumptions to model potential murderers’ responses to the possibility of the death penalty; and (3) the statistical models assume (without reason) that the effect will be the same across states and years. See generally id.

 [233]. Flanders, supra note 228, at 598.

 [234]. Atkins v. Virginia, 536 U.S. 304, 319 (2002) (citation omitted).

 [235]. Facts About the Death Penalty, Death Penalty Info. Ctr., (Mar. 17, 2018), http://www.deathpenaltyinfo.org/documents/FactSheet.pdf.

 [236]. Id.

 [237]. See id.

 [238]. Section of Individual Rights and Responsibilities, et al., Am. Bar Ass’n, Recommendation 3 (Am. Bar Ass’n 2006) http://www.deathpenaltyinfo.org/documents

 [239]. See id.

 [240]. Id. at 3–5.

 [241]. Id.

 [242]. Id. at 1, 20.

 [243]. Id. at 5, 7.

 [244]. Id. at 7.

 [245]. Id. (emphasis added).

 [246]. John H. Blume, Sheri Lynn Johnson & Christopher Seeds, Of Atkins and Men: Deviations from Clinical Definitions of Mental Retardation in Death Penalty Cases, 18 Cornell J.L. & Pub. Pol’y 689, 692 (2009).

 [247]. Id. (citation omitted).

 [248]. Id.

 [249]. Moore v. Texas, 137 S. Ct. 1039, 1048 (2017).

 [250]. DSM-5, supra note 11, at 37. (“IQ test scores are approximations of conceptual functioning but may be insufficient to assess reasoning in real-life situations and mastery of practical tasks. For example, a person with an IQ score above 70 may have such severe adaptive behavior problems in social judgment, social understanding, and other areas of adaptive functioning that the person’s actual functioning is comparable to that of individuals with a lower IQ score.”).

 [251]. Coleman v. State, 341 S.W.3d 221, 230 (Tenn. 2011) (citation omitted).

Eyewitness Identifications: Recommendations to the Third Circuit – Note by Brady Witbeck

From Volume 91, Number 3 (March 2018)


Eyewitness Identifications: RecoMmendations to the Third Circuit

Brady Witbeck[*]


Just before two o’clock in the afternoon on October 22, 1991, two high school students, Chedell Williams and Zahra Howard, ascended the steps of the Fern Rock train station in North Philadelphia, planning to take a train back to their homes.[1] Seemingly out of nowhere, two men appeared, blocked the girls way up to the station, and demanded Chedells earrings. Terrified, the girls bolted in opposite directions. The two men followed Chedell. They soon caught her and tore out her earrings. Then “[o]ne of the men grabbed her, held a silver handgun to her neck, and shot her.”[2] The perpetrators fled. Chedell was pronounced dead within the hour.[3]

Police soon focused their investigation on James Dennis, who lived relatively close to the train station in the Abbotsford Homes projects. Detectives would later explain that they heard rumors that Dennis was involved in the shooting, though they were at that time “unable to identify the source of the rumors.”[4] The detectives obtained preliminary descriptions of the perpetrators from three eyewitnesses.[5] These initial descriptions did not align well with Dennis’s actual appearance. Nonetheless, a few eyewitnesses identified Dennis during subsequent photo lineups, live lineups, and the trial.[6] In presenting the government’s case, the prosecution relied heavily on these eyewitness identifications.[7] Dennis was found guilty of “first-degree murder, robbery, carrying a firearm without a license, criminal conspiracy, and possession of an instrument of a crime.”[8] He was sentenced to death.

Then, after spending twentyfour years challenging his conviction, Dennis was granted a conditional writ of habeas corpus.[9] In Dennis v. Secretary, Pennsylvania Department of Corrections, the Third Circuit Court of Appeals found that prosecutors had improperly withheld evidence that bolstered Dennis’s alibi and implicated another man in Chedell’s death.[10]

Dennis is most notable not for unearthing aberrant prosecutorial misconduct, but for Chief Judge Theodore McKee’s lengthy concurrence, which illuminated endemic failures by courts and police departments to understand and mitigate the unreliability of eyewitness identification evidence.[11] Shortly after issuing its decision in Dennis, the Third Circuit formed a task force instructed to “make recommendations regarding jury instructions, use of expert testimony, and other procedures and policies intended to promote reliable practices for eyewitness identification and to effectively deter unnecessarily suggestive identification procedures, which raise the risk of a wrongful conviction.”[12] The Task Force will rely on scientific research and is co-chaired by Chief Judge McKee.[13] By establishing this Task Force, the Third Circuit recognized that not only is there a problem with the way the criminal justice system deals with eyewitness identification evidence, but also that unreliable identifications correspond to false convictions. Chief Judge McKee’s concurrence in Dennis and the commissioning of the Task Force demonstrate that the legal system is opening up to implementing scientifically proven methods to lessen the problem of false identifications and convictions.[14]

This Note will concentrate on how system variables impact the reliability of eyewitness identifications.[15] System variables are the procedures and practices law enforcement use to elicit eyewitness identifications.”[16] Because system variables are generally within the exclusive control of law enforcement, they present the most straightforward method through which the criminal justice system can make eyewitness identifications reliable, thus decreasing the risk of false convictions. This Note will focus on how the criminal justice system can improve the eyewitness identification process. In particular, this Note evaluates suggested reforms for photo arrays, live lineups, and jury instructions.

This Note will present simple, scientifically proven approaches to reform that will lead to a more just system and more accurate identifications and convictions. The Third Circuit Task Force should adopt recommended methods found in the volumes of psychological research written on eyewitness identification and analyzed in detail in this Note. Through a combination of legislative and judicial action, the system can be dramatically improved with minimal cost and inconvenience. Part I of this Note will examine Dennis in-depth and demonstrate how failures on the part of the criminal justice system led to false identifications and Dennis’s conviction. Part II will analyze the scientific research concerning system variables as well as the intersection of science and the criminal justice system. Part III will discuss current procedures for photo arrays, live lineups, and jury instructions, and their deficiencies. Part IV will discuss how different states have tried to solve these problems. Part V will make recommendations to the Third Circuit Task Force.

I.  DENNIS V. SECRETARY, Pennsylvania Department of Corrections

As Chief Judge McKee examined the data and scientific research on eyewitness identifications, he came to the conclusion that cases like Dennis are not mere anomalies; instead, they are serious miscarriages of justice that occur too frequently and should be rectified by the judiciary.[17] Even when multiple eyewitnesses identify a person, those identifications can be unreliable and “[a]lmost without exception, eyewitnesses who identify the wrong person express complete confidence that they chose the real perpetrators.[18] Even though three people identified the defendant as the perpetrator in Dennis, the way those identifications were obtained raised “serious questions about the accuracy of those identifications.”[19] Perhaps most troubling, the jury had no way of knowing the unreliable nature of the identifications, and as a result, an innocent man spent more than twenty years on death row.[20]

On the day Chedell Williams was murdered, the police obtained initial reports from eyewitnesses to the crime.[21] Five eyewitnesses claimed they could identify the shooter.[22] These five eyewitnesses were at varying distances from the shooter when the crime took place. The eyewitnesses said the shooter wore a red sweat suit and wielded either a dull silver gun or a shiny, chrome-plated gun. One of the key eyewitnesses told police that he “would be able to identify the shooter if he saw him again,” as he was only “about six feet from the perpetrators” and looked directly at the shooter as the shooter ran away.[23]

After the police heard rumors that the shooter was Dennis, they arranged for several eyewitnesses to see if they could identify Dennis as the shooter by placing his picture in a photo array.[24] The police “compiled three arrays of eight photographs each.”[25] The first array was used to identify the shooter, the second to identify the accomplice, and the third to give the eyewitnesses an opportunity to identify a suspect. Police composed the photo arrays with pictures of seven innocent fillers and a recently taken photo of Dennis. They then individually showed the photo arrays to each eyewitness and instructed each witness to “[s]ee if you recognize anyone.”[26] Four of the nine eyewitnesses stated that Dennis looked familiar, but no eyewitness expressed a high degree of confidence in their identification at the time of the photo array.[27] Following at least two of these uncertain identifications, the photo array administrator asked the eyewitnesses if they were confident in their identifications; when responding to this question, two eyewitnesses reported greater confidence in their identifications.[28] The remaining five eyewitnesses were not able to identify the shooter with any degree of certainty.[29]

Around a month and a half later, police conducted a live lineup, which included six persons: Dennis and five fillers.[30] Only the four eyewitnesses that identified Dennis in the prior photo arrays were present at this live lineup,  and [t]he police had those four witnesses view the lineup at the same time, in the same room.[31] The police gave instructions to each eyewitness to carefully look at all of the lineup participants to see if they recognized any one of them as the suspect, and they also instructed that none of the eyewitnesses had to make an identification if they could not recognize the suspect in the lineup.[32] Two of the eyewitnesses somewhat confidently pointed out Dennis, one eyewitness was less sure, and one—the eyewitness who initially claimed that he was so close to the perpetrator that he could easily make an identification—identified a filler.[33] Later at the trial, the prosecution put three eyewitnesses on the stand, all of whom confidently pointed at Dennis, “even though all three had expressed doubt in their earlier identifications.”[34]


The debate about what role science should play in eyewitness identification evidence is not new. In 1908, Hugo Münsterberg, a pioneering psychologist, published On the Witness Stand.[35] In it, Münsterberg profiles different judicial and police practices and analyzes them to see how the judicial system could improve with respect to eyewitness testimony.[36] To illustrate the need to incorporate science into the criminal justice system, he detailed an experience that occurred after his family home was burglarized.[37] As an eyewitness at the burglary trial, he recounted various details about the robbery.[38] But after comparing his testimony to the crime-scene evidence, Münsterberg realized there were significant errors in his testimony and that, despite his best intentions, some of his memories were distorted.[39] He emphasized that human memory is inherently faulty and that perhaps the greatest impediment to justice is not intentional lies on the part of the eyewitness but the unintentional failings of memory.[40] Finally, he spoke to the issue of how the judicial system has failed to put into practice the scientific research of the time.[41]

Münsterberg was repudiated by John Wigmore for what Wigmore viewed as an uncouth attack on the legal profession, an attack that was not justified by the scientific research Münsterberg touted.[42] Wigmore viewed Münsterberg as a popular scientist—someone more interested in fame than properly integrating science and the law.[43] While Wigmore criticized Münsterberg, he himself was a strong proponent of the use of psychology in the legal profession.[44] This debate, which took place over a century ago, demonstrates that even among those who believe science should play a greater role in evidence, it is difficult to achieve a consensus on the specifics.

Today, there are some in the legal profession who believe few or no reforms are necessary to bring science and evidence together. For example, Justice Antonin Scalia, in his concurrence in Kansas v. Marsh, rejected the idea that the way the justice system handles eyewitness identifications is deeply flawed.[45] He wrote his Marsh concurrence primarily as a response to Justice Souter’s dissent, in which Souter acknowledged the primary risk of capital punishment: that the defendant is innocent.[46] Scalia claimed that because Souter did not list an instance when an innocent person was put to death, this risk is overstated.[47] Scalia went further, claiming that DNA evidence has confirmed guilt more often than it has proved the innocence of convicted persons.[48] He also claimed that the recent reversals of false convictions are not the result of “the operation of some outside force to correct the mistakes of our legal system, rather than as a consequence of the functioning of our legal system.”[49] Scalia stated that capital cases are actually given heightened judicial scrutiny, which leads to better and more accurate results, as the appeals process can be very lengthy in these cases.[50] Scalia echoed the claims of many who say that while the system may be imperfect, it still functions at a high rate of accuracy and needs minimal reform, if any.[51]

Scalia’s assessment contrasted with that of Chief Judge McKee in Dennis, in which McKee tied together the best psychological research on eyewitness identifications and recognized that the criminal justice system must improve in order to be more accurate.[52] McKee began by quoting Justice Brennan, who had stated over three decades prior that juries are likely to believe eyewitness testimony over other types of evidence, especially when the eyewitness is confident.[53] “James Dennis was sentenced to death because three eyewitnesses appeared at trial and confidently pointed their fingers at him when asked if they saw Chedell Williams’ killer in the courtroom.”[54] Because the jury was not properly instructed as to how to handle eyewitness identifications by the court and the police department was not properly trained, an innocent man was sentenced to death.[55] His purpose in writing his lengthy and thoughtful concurrence was to push the law to catch up with the science and persuade both police departments and juries to reform.[56] Reform is critical as mistaken identifications “‘erode public confidence in the criminal justice system as a whole.’”[57]

The Supreme Court itself recognized the problems inherent in eyewitness identifications in United States v. Wade.[58] In Wade, the Court declared that “the vagaries of eyewitness identification are well-known; the annals of criminal law are rife with instances of mistaken identification.”[59] The Court cited “the degree of suggestion inherent in the manner in which the prosecution presents the suspect to witnesses for pretrial identification” as a major factor contributing to misidentification.[60] The Court’s opinion recognized the danger that once an eyewitness has identified someone during a lineup, that eyewitness’s confidence in that identification can be artificially inflated. Improper suggestions and poor lineup construction during the lineup process can taint an entire trial.[61]


In the United States, eyewitnesses identify some 77,000 suspects annually.[62] Often, juries place great weight on eyewitness identifications and, accordingly, the identifications provide powerful evidence against a defendant.[63] Despite the importance of eyewitnesses, their accounts are generally less accurate than most people—including judges, jurors, and attorneys—would assume.[64]

In one study, 590 participants were tested to determine if, after having spoken to a woman for fifteen seconds, they could later identify that same woman.[65] During a live lineup where the woman was present, only forty-nine percent of the participants were able to correctly identify the woman.[66] While 62% of participants correctly refrained from making an identification when the target was absent from the lineup, the remaining thirty-eight percent of participants made an identification.[67]

Police departments generally use three types of methods to obtain identifications from eyewitnesses: showups, photo arrays, and live lineups.[68] But how police departments administer these three methods varies greatly and lacks uniformity across jurisdictions.[69] With thousands of police departments and courts, it is difficult to obtain a clear picture of how different jurisdictions obtain eyewitness identifications.[70] Many police departments have no standing procedures or policies, and many police officers are not aware of how system variables, which police control, can influence the reliability of this type of evidence.[71]

This Section describes how photo arrays, live lineups, and jury instructions function and how these processes often fall short of their objective to obtain reliable identifications. Showups, ad hoc procedures where law enforcement officers bring eyewitnesses to a location to show them a suspect, will not be discussed at length.[72]

A.  Photo Arrays and Live Lineups Defined

Photo arrays and live lineups constitute important ways in which police can obtain eyewitness identifications.[73] Police regularly use both photo arrays and live lineups in their investigative efforts.[74] Though live lineups are generally considered more accurate than photo arrays, they are conducted less frequently.[75] In most photo arrays, the eyewitness is presented with a number of photographs and instructed to identify the photo of the person who the eyewitness believes committed the crime.[76] A defendant does not have the right to have an attorney present during a photo array.

Like photo arrays, live lineups are used by police either to determine or confirm the identity of a suspect.[77] In a live lineup, an eyewitness is presented with a number of people and asked to identify the person the witness believes to be the suspect.[78] Live lineups can occur either before or after an indictment.[79] Most live lineups in the United States contain around five participants.[80] Eyewitnesses either view the lineup participants sequentially or simultaneously.[81] Sequential lineups compel the eyewitness to make an absolute judgment of identity, while simultaneous lineups allow the eyewitness to make a relative judgment of identity.[82]

In a sequential lineup the eyewitness views the suspect and fillers one at a time. . . . In the original sequential lineup for each person (i.e., the suspect and fillers) the eyewitness either identifies the person as the culprit or not. If the eyewitness makes an identification the procedure ends. If no identification is made then the next person is shown to the eyewitness.[83]

In a simultaneous lineup, the eyewitness is presented with all the lineup participants at one time.[84]

B.  System Variables and Accuracy

As Chief Justice McKee stated in Dennis, system variables are within the control of law enforcement.[85] Because police departments control the practices and procedures used to acquire eyewitness identifications, the Third Circuit Task Force (“Task Force”) should examine the scientific research concerning the accuracy of those procedures. Studies have identified simple, cost-effective ways to adjust system variables to improve the reliability of eyewitness identification evidence.

Though photo arrays and live lineups are most accurate when administered blindly—that is, when the person administering the lineup does not know the identity of the suspect—very few police departments conduct blind lineups and arrays.[86] In an experiment, students were randomly assigned to play the role of either a lineup administrator or a mock eyewitness.[87] The mock eyewitnesses were shown a video of a theft in which they were exposed to the perpetrator’s face for twenty-five seconds.[88] One group, who was assigned the role of lineup administrator, was told the identity of the suspect, while the other group was not.[89] The researchers found that the non-blind administrators often smiled when the mock eyewitness viewed the suspect in the photo array and smiled after the eyewitness identified the suspect.[90] The non-blind photo arrays resulted in significantly more false identifications than the photo arrays that were administered blind.[91] The researchers also found that the non-blind administrators affected eyewitnesses’ confidence in their selections.[92]

Feedback from a non-blind administrator can manipulate eyewitnesses’ confidence in their identifications.[93] This is even true when the eyewitness mistakenly identifies the wrong person; feedback confirming an eyewitness’s mistaken identification impairs the eyewitness’s memory of the original perpetrator.[94] When an administrator makes statements like “[w]e thought this might be the one,” [t]hat’s the one you picked out in the photo,” or even more subtle, non-verbal communications, eyewitnesses’ confidence can increase and their ability to recognize the actual suspect can decrease.[95] “Relative to a no feedback condition, witnesses who received good-memory feedback expressed higher post-identification confidence in a subsequent lineup identification, whereas those who received poor-memory feedback evinced lower confidence.”[96]

Jurisdictions, as well as experts, disagree as to the advantages of using sequential lineups instead of simultaneous lineups. Some jurisdictions have reformed their procedures in order to have eyewitnesses make absolute judgments of identity, while others have cited evidence that claims relative judgments of identity are more reliable.[97] A study that purports to go against the grain of recent evidence—in that it supports simultaneous lineups—found that eyewitness identifications based on relative judgments are less reliable than those based on absolute judgments.[98][A] witness using an absolute judgment makes an identification of a lineup member if the match between that lineup member and the witness’s memory of the perpetrator is sufficiently high,” while a relative judgment can be made when the match is relatively better than any other member of the lineup.[99] The study also found that “witnesses’ reliance on relative judgments undermines the reliability of the identification evidence, and increases the relative risk of a false identification that can ultimately lead to a wrongful conviction.”[100] Despite this recent study, the scientific community is still somewhat divided on this issue, with some studies claiming that there is little difference in reliability between the two approaches.[101]

Lineup instructions given to eyewitnesses before they make identifications impact the reliability of any identification that follows.[102] Biased instructions occur when the lineup administrator fails to explicitly instruct the eyewitness that the perpetrator may not be present in the lineup and that it is permissible to identify no one.”[103] In some instances, biased instructions “compel[] witnesses to adopt a lower criterion for accepting their sense of recognition of the most familiar-looking lineup member as correct . . . and thereby enhances their confidence in making a positive identification of that lineup member,” and can also artificially increase eyewitnesses confidence in their identifications because they may assume the suspect is in the lineup.[104] In one study, participants viewed a video of a mock theft and were instructed to identify a suspect from both a thief-present and thief-absent live lineup.[105] One group was given biased instructions before they attempted to make an identification, while the other group was not. The mock eyewitnesses’ confidence was then measured. The results found that “[b]iased instructions and positive feedback increased confidence and ratings of eyewitnessing conditions.”[106] The study also found that eyewitnesses’ confidence in their identifications only modestly relates to the accuracy of those identifications.[107]

C.  Jury Instructions

Jury instructions that provide the jury with information on how to use eyewitness identifications could improve a jury’s evaluation of eyewitness evidence, thus improving the deliberation process. Instructions can inform the jury how memory works, how an identification was obtained, and dismiss the myth of the infallibility of the identification process.[108] Jury instructions regarding eyewitness identifications and their use in trials typically contain some qualifications about their accuracy, but these instructions are often generic and do not properly convey scientific realities.[109] Some experts claim that most current jury instructions do not increase a jury’s sensitivity to possible errors in eyewitness testimony.[110] This is because jurors weigh eyewitness evidence too heavily and because they are “often uncritical of the reliability of the testimony.”[111]

For example, the instructions received by the jury in Dennis were “plain vanilla” and unhelpful.[112] The instructions were long, confusing, and did not include any “explanation of the relevant system or estimator variables that so crucially impact the reliability of witness identifications.”[113] Jurors are often not aware, or at least do not receive instructions from the court, of possible inaccuracies of eyewitness testimony generally and of eyewitness identifications specifically.[114] Studies have shown that jurors do not understand how memory functions or how memory can be influenced and manipulated.[115] Juries have limited knowledge about memory and rely on eyewitness confidence, an eyewitness’s memory for minor details, and the consistency of an eyewitness’s testimony, while ignoring the impact system variables have on the reliability of eyewitness identifications.[116] The myth that people can never forget a face or that an encounter with an armed suspect increases or enhances one’s ability to identify a suspect can lead to a jury overvaluing an identification during its deliberations.[117]


In the past decade, a few jurisdictions have reformed procedures with the goal of improving the reliability of eyewitness identification evidence. This Section will discuss three states in particular: New Jersey, Oregon, and North Carolina. These jurisdictions used scientific research to improve how police departments obtain identifications and to ensure that courts only admit into evidence eyewitness identifications that have indicia of truth and reliability. New Jersey and Oregon addressed eyewitness identification procedures through their respective Supreme Courts.[118] North Carolina’s legislature instituted reforms statutorily.[119] In addition to analyzing the reforms adopted by these states, this Section will evaluate proposals from experts in the fields of law and psychology who have proposed procedures and practices to increase the reliability of eyewitness identification evidence.

A.  New Jersey

Recently, the Supreme Court of New Jersey attempted to improve the reliability of eyewitness identification evidence with its decision in State v. Henderson.[120] The court overhauled its test for the admission of eyewitness identification evidence.[121] The decision called for blind administration of photo arrays and live lineups, new pre-lineup instructions, the creation of rules for lineup construction, and new record keeping procedures.[122] The court also determined that jury instructions needed to improve in order to better equip juries in their process of deliberation.[123]

In Henderson, an eyewitness to a crime was shown a photo array that included eight photographs—one of the suspect and seven of innocent fillers.[124] Before the photo array was administered, the eyewitness was given instructions that were standard in New Jersey police departments.[125] He was informed that an administrator would show him photos sequentially and the perpetrator’s photo was not necessarily included in the array.[126] The eyewitness also was instructed that the suspect could have either gained or lost weight since the incident and that facial hair could easily be altered.[127] The photos were shown to the eyewitness in an order that was random to the administrator.[128] During the photo array, the eyewitness narrowed the photos down to two, but he could not make a clear identification.[129] Police later testified that during this point in the photo array, the eyewitness was excited, so police removed him from the room, calmed him down for one to five minutes, and then showed him the eight photos again.[130] Police claimed that the eyewitness was then quickly and confidently able to identify the police suspect. The eyewitness later testified that he felt pressured to make an identification and that police pushed him to identify the suspect.[131]

The jury instructions provided at trial did not inform the jury about the influence suggestive police behavior can have on the reliability of identifications.[132] The instructions were long, confusing, and included scientific language most likely unfamiliar to jurors, and to determine whether the identification was reliable, the instructions asked jurors to consider a number of competing and seemingly contradictory factors.[133]

To improve the reliability of eyewitness identifications, the New Jersey Supreme Court addressed system variables within the control of the criminal justice system that it believed would best improve the reliability of identification evidence.[134] The court determined that because even subtle, non-intentional suggestions by police during the identification process can influence memory, photo arrays and live lineups should be administered blindly.[135] Because police departments have limited resources, the court suggested that departments could use the “envelope method” for the administration of photo arrays.[136] With the envelope method, “an officer who knows the suspect’s identity places single lineup photographs into different envelopes, shuffles them, and presents them to the witness. The officer/administrator then refrains from looking at the envelopes or pictures while the witness makes an identification.”[137] This method would decrease the likelihood of improper suggestion by the police.[138]

The court ordered that before administering a photo array or a live lineup, New Jersey police must always instruct the eyewitness that the person who committed the crime may or may not be present and that the eyewitness should not feel pressure to make an identification.[139] In order to decrease the possibility of an eyewitness simply guessing the identity of the suspect, every lineup should be composed of fillers who look similar to the suspect, so the suspect does not stand out.[140] This is so an eyewitness’s confidence is not artificially inflated by a perception that the identification process was “easy.”[141] There should be at least five fillers in a live lineup, and lineups should not feature more than one suspect.[142] The court also reminded police departments that all lineups should be recorded and preserved so that courts can later determine if the lineup was properly constructed.[143]

In order to avoid improper feedback from police that could inflate eyewitnesses’ confidence in their identification, the court held that “law enforcement officers should make a full record—written or otherwise—of the witness’ statement of confidence once an identification is made.”[144] Officers should not allow eyewitnesses to view the suspect multiple times, as this can artificially increase confidence in their identification.[145] The New Jersey Supreme Court took no position on whether police departments should favor sequential or simultaneous lineups.[146] The court believed that there was insufficient scientific evidence to show a preference for either and that more studies needed to be conducted before the court could state a preference.[147]

To better help jurors understand the eyewitness identification process, the court reformed jury instructions.[148] Lay people, on the whole, do not understand how memory works.[149] The court identified the common misconceptions that memory is similar to a video recording and that memory cannot be contaminated or distorted by outside influence.[150] Juries also tend to give disproportionate weight to the confidence of the eyewitness.[151] In order to better equip the jury to evaluate eyewitness identifications, jury instructions need to clearly and comprehensively inform the jury about the science of eyewitness identification and the nature of memory.[152] However, jury instructions should not overwhelm the jury and must be helpful to jurors.

B.  Oregon

In State v. Lawson, the Supreme Court of Oregon overhauled its test for determining the admissibility of eyewitness identifications.[153] In Lawson, the court consolidated two cases, in which the admissibility of eyewitness identification evidence was at issue.[154] Two defendants were separately tried and convicted, at least in part because of eyewitness identifications that “had been subject to an unduly suggestive police procedure in the course of identifying” the defendants.[155]

The test used by Oregon courts during the defendants’ trials to evaluate the admissibility of eyewitness identification evidence was fairly permissive, and it failed in its purpose of preventing suggestive and inaccurate identifications from being admitted into evidence.[156] The test was comprised of generic, unhelpful factors that attempted to make sure the time between the event and the identification was minimized, the certainty of the eyewitness was high, and the eyewitness had a chance to clearly see the suspect before the identification was admitted.[157]

In one of the cases consolidated in Lawson, a victim was shot in the chest and admitted to the hospital, where she was questioned by police as to the identity of her attacker.[158] The victim was shown a black-and-white photo array while heavily medicated, sedated, and restrained in her hospital bed.[159] Moreover, because her injuries necessitated a breathing tube, the victim could only respond to police questioning by nodding or shaking her head. At first, the victim did not identify anyone from the photo array; however, she eventually nodded “yes” to leading questions regarding the suspect’s identity.[160] The victim later had no recollection of this interview.[161]

Approximately two weeks later, when the victim could speak, she said that she was not able to identify the person who shot her; the following month, she was not able to pick the defendant out of another photo array, but shortly thereafter the police informed her that she had identified someone during her stay in the hospital.[162] After hearing this, the victim said she recognized the man police had identified as a suspect; however, she stated that she was not certain he was the perpetrator.[163] At a much later date, and after police repeatedly exposed the victim to the suspect’s photo, the victim identified the suspect at a live lineup and even testified at trial that she “always knew it was him.”[164] Based in part on this evidence, the defendant was convicted.[165] On appeal, the Oregon Supreme Court held that the identification should not have been admitted into evidence as it was subject to suggestive police procedure.[166]

In the case, the Oregon Supreme Court examined scientific research about system variables that the court believed could prevent false or unreliable identifications from being admitted into evidence.[167] Based on their examination of the science, the court mandated judicial and police department reforms.[168] Additionally, the Lawson court shifted the defendant’s burden to prove suggestibility onto the prosecution.[169]

In order to improve the reliability of eyewitness identification evidence, the Oregon Supreme Court found that the criminal justice system needed to improve several system variables, which are in the exclusive control of the justice system.[170] The court called for the blind administration of photo arrays to prevent an administrator from improperly influencing an identification.[171] When police administer photo arrays or live lineups, the administrator should inform eyewitnesses that they do not have to make an identification, as the perpetrator may not be in the lineup or array.[172] The court called for live lineups to be constructed using fillers that look physically similar to the suspect so the suspect does not stand out.[173] Furthermore, live lineups and photo arrays in Oregon must now be conducted sequentially so that the eyewitness makes an absolute judgment of identity instead of a relative judgment.[174]

The fact that the victim in Lawson viewed the suspect multiple times was a major factor in determining the identification was unreliable.[175] When police continually expose a victim to images of one suspect, the victim tends to become more familiar with the suspect’s face; this can result in the victim eventually identifying that suspect with confidence, even if initially the victim was unsure of the perpetrator’s identity. Because police continually exposed the victim to images of the suspect, the victim became more familiar with his face, so much so that the victim could eventually identify him with confidence, even though initially the victim was unsure of the perpetrator’s identity.[176] For this reason, after Lawson, Oregon police are required to avoid multiple viewings when conducting photo arrays and live lineups.[177]

The opinion did not elaborate in-depth about how Oregon courts should craft jury instructions on how to evaluate eyewitness identifications, but the court suggested that future jury charges should include reference to system variables that influence reliability.[178] The court cited an Oregon evidence rule that stated identifications must be helpful to the trier of fact.[179] Therefore, identifications, when admitted, should not serve to confuse the jury but should help the jurors with their fact-finding, thus providing another reason to improve the reliability of eyewitness identifications.[180]

C.  North Carolina

While legislatures lack some of the sophisticated legal experience of the courts, passing laws to regulate police conduct can be an effective way to quickly and authoritatively adjust system variables. North Carolina took this approach with the North Carolina Eyewitness Identification Reform Act (“the Act”).[181] The Act, passed in 2007, attempts to incorporate scientific advances in the field of eyewitness identifications to better assure reliability and bolster the truth finding function of the criminal justice system in North Carolina.[182] To further this goal, it provides instructions for police departments on how to administer identifications according to the best available practices.[183]

The Act calls for independent administrators, who are not aware of the suspect’s identity, to carry out both photo arrays and live lineups.[184] The independent administrator will give instructions that inform the eyewitness that the perpetrator may or may not be in the lineup or photo array, and will also state that the investigation does not hinge on the eyewitness making an identification, so the eyewitness should not feel undue pressure to make one.[185]

Under the Act, both photo arrays and live lineups should contain at least five innocent fillers who resemble the suspect.[186] Lineups and photo arrays with more than one suspect are prohibited, and eyewitnesses are separated from others who are making an identification to prevent them from conferring with one another before or during the live lineup or photo array.[187] Eyewitnesses are not be provided any information about the suspect, and police make a video recording of the process or an audio recording if a video recording is not feasible.[188] The Act also proposes that lineups could be administered by a computer program as an alternative method to keep the administrator from seeing the photo in front of the witness.[189]

In order to facilitate these reforms, law enforcement officers are required to go through training programs so that they know how to conduct lineups and photo arrays in compliance with this statute.[190] The Act calls for the creation of materials and classes to facilitate the training of law enforcement officers.[191] Two preexisting North Carolina police-training agencies, the North Carolina Criminal Justice Education and Training Standards Commission and the North Carolina Sheriffs’ Education and Training Standards Commission, were made responsible for creating these programs and materials.[192]

D.  Scholarly Proposals

Legal scholars have proposed reforms that often go further than the changes made in states like New Jersey, Oregon, and North Carolina. For example, the National Academy of Sciences issued a report addressing the reliability of eyewitness identification evidence.[193] The academy’s goal was to digest the current scientific research on the subject and present it to law enforcement and the legal community, and the academy called for greater cooperation among the law enforcement and scientific communities so that identification procedures can improve across the country.[194] Scholars hope that the training of law enforcement as to how memory works and how law enforcement can unintentionally influence identifications will allow police to see why reform is necessary.[195]

The report made recommendations as to how jurisdictions can improve the reliability of eyewitness testimony.[196] It called for blind administration of lineups, uniform and “easily understood instructions” to be provided to the eyewitness prior to an identification, and careful documentation of eyewitnesses confidence in their identifications.[197] These instructions should inform the eyewitness that “the perpetrator may or may not be in the photo array or lineup and that the criminal investigation will continue regardless of whether the witness selects a suspect.”[198] The report suggested that due to a lack of consensus as to the merits of sequential versus simultaneous lineups, neither method should be preferred.[199]

The academy acknowledged that some police departments are hesitant to make changes that would require them to stretch their limited resources.[200] In response, the committee suggested that “departments consider procedures and new technologies” that would alleviate this concern.[201] For example, if a non-blind administrator is not available, a department could use either a “computer-automated presentation of lineup photos”[202] or the envelope method that is employed in New Jersey.[203] The eyewitness identification process should also be videotaped, even though doing so could increase costs and burden eyewitnesses’ privacy interests.[204] However, when these concerns arise, departments can videotape the process non-intrusively, and in fact, many departments already have the technology that would allow them to document these procedures.[205]

The report called for the “use of clear and concise jury instructions” to assist jurors in their fact-finding mission.[206] Jury instructions can convey the most important underlying aspects of the identification process in clear language.[207] This would allow the jury to properly give weight to eyewitness identification evidence in its deliberations.[208] “Appropriate legal organizations, together with law enforcement, prosecutors, defense counsel, and judges, should convene a body to establish model jury instructions regarding eyewitness identifications.”[209]

Going forward, the academy recommended that a national research initiative be established to increase our understanding of the science of eyewitness identifications.[210] The research initiative would allocate future funds for research, formulate new policy positions, review research, advocate future policy changes, and provide formal assessments of reforms across the country.[211]

Separately, Dan Simon, in his book In Doubt, proposed a series of reforms that could fix the systematic errors inherent in the identification process.[212] He proposed a series of reforms that would “provide best-practice protocols andare directed at the twofold goal of maximizing the accuracy of identifications and the transparency of the procedures used to elicit them.[213] Simon’s reforms include:

2. . . . [L]ive and video lineups should be preferred over photographic arrays.

3. Suspects should not be placed in identification procedures absent an appreciable threshold of guilt.

4. Prior to the lineup, witnesses should not be exposed to any identifying information about the suspect from any source.

5. Lineups should be conducted as soon as possible after the witnessed event.

6. Lineups should include only one suspect and five or more fillers whose innocence is beyond doubt.

7. Fillers should match the witness’s description of the perpetrator and not be noticeably dissimilar from the suspect.

8. The suspect should be allowed to determine his place in the lineup and to change places between lineups.

9. The witness should be instructed that the perpetrator “may or may not be” in the lineup, and that it is appropriate to respond “perpetrator is not present,” and “don’t know.”

10. Targets should be presented sequentially (rather than simultaneously).

11. All identification procedures should be “double blind”: the administrator must be kept unaware of the identity of the suspect; the witness should be informed that the administrator does not know the suspect’s identity.

12. The administrator should refrain from any communication or behavior that could be interpreted as suggestive or revealing of the identity of the suspect.

13. The witness should announce his recognition or nonrecognition, followed immediately by a confidence statement. The witness should not be given any feedback before completing the statement.

14. The time it took the witness to announce recognition should be measured and recorded. . . .

16. Witnesses who at any time pick someone other than the suspect should not be allowed to provide any identification testimony about the suspect.

17. Witness [sic] who fail to identify the suspect, make a hesitant decision, or express low confidence at the initial identification should be deemed to have a weak memory of the suspect.

18. The procedure should be recorded in its entirety, preferably on videotape. Recording should include the images used and the instructions given. The witness should be videotaped throughout the procedure.[214]

Simon also suggests that the composition of lineups be computerized to altogether remove the human error element from the equation.[215]

Simon does recognize that the implementation of most of these ideas is uncontroversial, but also that there is an inherent trade off “between the intended objective of reducing false identifications and the unintended effect of losing correct identifications.”[216] Despite this, Simon argues that these proposed reforms would provide a net gain for the judicial system.[217] Providing a complete record of identification procedures is critical for minimizing “the effects of memory decay, contamination, and any other biases induced by the investigation and pretrial procedures” and providing fact finders and other decision makers with the best possible information for assessing the reliability of the identifications.”[218]


The Third Circuit should borrow the best and most practicable reforms undertaken by North Carolina, New Jersey, and Oregon. These states have taken steps toward integrating scientific research into the judicial system, thus making eyewitness identification evidence more reliable. The Task Force has the opportunity to combine the best ideas of these states to lower the risk of wrongful convictions in the Third Circuit. These reforms can further serve as a model for other jurisdictions to reform their policies and procedures. Though live lineups generally produce more reliable evidence than photo arrays, the Task Force should recommend reforms for both photo arrays and live lineups given the impracticality of having a live lineup for every identification. The Task Force should also address jury instructions.

Because the composition of a lineup can greatly influence the reliability of the resulting identification, the Task Force should provide clear guidelines on how and when lineups should be conducted. Lineups and photo arrays should be conducted close in time to when the crime took place, so the eyewitness is more likely to remember the suspect. In many of the cases discussed above, police conducted lineups months or even a year after an event occurred, which led to decays in memory and ultimately false identifications.

The Task Force should adopt a policy similar to that of North Carolina, which requires that live lineups include at least five fillers.[219] These fillers should be similar in race, height, age, and facial structure to the suspect. Ideally, lineups will be composed by a computer program to ensure similarity among the lineup participants. If the suspect has a unique feature, such as a mole or a tattoo, lineup administrators should select photos of other suspects with the same features or alter the filler photos so that the unique feature is present in all or most of the photographs.[220] Photos of the suspect should not be more than a year old[221] and, whenever possible, should not be photos where the suspect has different facial hair than during the time the incident took place.[222] Lineups should never include more than one suspect.[223] As in New Jersey, all live lineups and photo arrays should be recorded and preserved, so that if the reliability of the identification is brought into question, a court can use the recording to help determine if the identification was reliable.

Whenever possible, photo arrays and live lineups should be administered in isolation, away from third parties who could influence the evidence. Lineup administrators should select quiet, separate areas of police precincts and ensure that the eyewitness is separated from other police officers and other eyewitnesses. The lineup administrator should ensure that the eyewitness does not have any access to case materials, including “information about the case, [and] the progress of the investigation.”[224] Eyewitnesses should not be allowed to see images of the suspect outside of the lineup administration, including wanted posters of the suspect that may be hanging in the police department where the photo array or live lineup is being administered.[225]

The Third Circuit should mandate that police departments administer both photo arrays and live lineups blindly. Blind administration increases the accuracy of eyewitness identifications and lowers the risk of feedback from the lineup administrator.[226] Because police resources are limited, the task force should recommend that even where the photo array administrator is not blind to the suspect’s identity, the police department should follow the envelope method employed in New Jersey. This method will prevent the administrator from seeing the photographs before eyewitnesses make an identification, removing the risk that the administrator could influence the eyewitnesses beforehand.[227] However, because the administrator could provide feedback to the eyewitness post-identification, the envelope method should only be used when a fully blind test is impractical.

The Task Force should recommend that police departments change the way they instruct eyewitnesses prior to administering either a photo array or a live lineup. Because biased instructions lead to false identifications and artificially increased confidence in those identifications, it is critical that police departments give uniform, unbiased instructions to eyewitnesses.[228] Lineup administrators should explicitly state that the suspect may or may not be in the photo array or live lineup and that the entire case does not rely on the eyewitness making an identification. Police should try to ensure that eyewitnesses do not feel pressure to make an identification and that they are aware they can say that they do not know if the suspect is in the lineup. These eyewitness instructions are important because an eyewitness should not assume that the suspect is in the lineup. Furthermore, multiple viewings of the suspect by eyewitnesses should not be allowed, so as not to inflate their confidence in the identification.

Once an eyewitness makes an identification, police should immediately record the level of confidence the eyewitness has in that identification. Although juries often overvalue eyewitness confidence, it can serve a role at trial, especially if the confidence is measured immediately after an identification.[229] Eyewitnesses’ confidence in their identifications can be used as a factor to determine if the evidence is admissible. When an eyewitness identifies a suspect without hesitation and without prompting by the lineup administrator, that identification is more likely to be reliable. The eyewitness should also confirm in writing the identification. This provides an additional failsafe to ensure that the eyewitness was not coerced into making an identification and allows for a statement of confidence to be in writing.

Because the scientific community is split on whether sequential or simultaneous viewing of a lineup results in the most reliable identifications, the Task Force should not state a preference for either.

The Task Force should improve existing jury instructions. If a jury were equipped to properly weigh eyewitness evidence and were aware of how and why some identifications are unreliable, police departments could internally strive to improve system variables knowing that a jury may discard improperly obtained identification evidence. Some jurisdictions use expert testimony to inform the jury about eyewitness identifications; however, this method generally appears unsuccessful.[230] Because current instructions do not assist the jury in properly evaluating eyewitness identifications, new, standard instructions should be implemented.

As in Dennis, the jury instructions in New Jersey prior to judicial reform were confusing and muddled.[231] This led the New Jersey Supreme Court to implement new jury instructions. Because the Third Circuit’s jury instructions are similar to those previously used in New Jersey—in that they are too long and do not explain simply how eyewitness identifications can be inaccurate and unreliablethe Task Force should also implement better jury instructions.[232] Proper instructions give juries a tool to compensate for their limited knowledge of how memory functions. Instructions should encourage a jury to examine various factors to determine not only if police procedure leading up to the identification was proper, but also if the eyewitness’s memory shows indicia of reliability. As in New Jersey, the Task Force should inform juries that they should refrain from assigning undue weight to eyewitness confidence; however, they should also be wary of overwhelming the jury with scientific information.

Finally, the Third Circuit Task Force should recommend a training program for police departments that will help implement these reforms. When implementing its legislative reforms, North Carolina recognized that training was essential to increase the reliability of eyewitness identification evidence.[233] Police officers should be instructed that following these procedures will not necessarily result in fewer convictions, but will help ensure that investigations are conducted in a manner most conducive to truthfinding. The Task Force could appoint a team of experts to travel to conferences and individual police departments to train police on how best to implement the proposed reforms. In order to ensure compliance, the Task Force should require periodic reports from both trial courts and police departments as to how the proposals are being implemented and if any modifications to the reforms are necessary in the future. The Task Force should reconvene in five years to reexamine scientific evidence and suggest any further changes.


The investigative procedures used in Dennis that caused such an unjust outcome are employed in many jurisdictions across the country. The Third Circuit Task Force on Eyewitness Identifications has been presented with the unique opportunity to examine every facet of the eyewitness identification process and recommend changes that will serve to decrease the risk of false convictions. New Jersey, Oregon, and North Carolina, among others, provide a path forward that the Task Force should follow. Through blindly administrated lineups, correct pre-lineup instructions, proper construction of lineups, helpful jury instructions, and other reforms analyzed above, the Third Circuit can serve as an example of how scientific research can be implemented into the justice system to produce both fair and just results.


[*] *. Executive Articles Editor, Southern California Law Review, Volume 91; J.D. Candidate 2018, University of Southern California Gould School of Law; B.A. History 2015, Brigham Young University. I would like to thank Professor Dan Simon and Professor Sam Erman for valuable guidance and feedback on earlier drafts of this note. In addition, I would like to thank the staff and editors of the Southern California Law Review for their excellent work.

 [1]. Dennis v. Sec’y, Pa. Dep’t of Corr., 834 F.3d 263, 269 (3d Cir. 2016).

 [2]. Id.

 [3]. Id.

 [4]. Id.

 [5]. Id. at 270.

 [6]. Id. at 270–71.

 [7]. Id.

 [8]. Id. at 275.

 [9]. Id. at 269.

 [10]. Id. at 275, 287.

 [11]. See generally id. at 313–44 (McKee, C.J., concurring).

 [12]. Order Establishing Third Circuit Task Force on Eyewitness Identifications (Sept. 9, 2016), http://www.ca3.uscourts.gov/sites/ca3/files/TFEyewitnessIdOrder_11042016.pdf.

 [13]. Id.

 [14]. See               Dennis, 834 F.3d at 313 (McKee, C.J., concurring); Order, supra note 12.

 [15]. While there are other factors that contribute to unreliable eyewitness identification evidence, this Note will only focus on system variables.

 [16]. Dennis, 834 F.3d at 321 (McKee, C.J., concurring).

 [17]. Id. at 313–16.

 [18]. Id. at 315.

 [19]. Id.

 [20]. Id. at 316.

 [21]. Id. at 317.

 [22]. Id.

 [23]. Id.

 [24]. Id. at 318. Police never clarified where the rumors originated or why the detectives decided to further investigate the rumors. Id.

 [25]. Id.

 [26]. Id.

 [27]. Id.

 [28]. Id.

 [29]. Id. at 319.

 [30]. Id.

 [31]. Id.

 [32]. Id. at 320.

 [33]. Id.

 [34]. Id.

 [35]. Hugo Münsterberg, On the Witness Stand: Essays on Psychology and Crime (1908).

 [36]. See generally id.

 [37]. Id. at 39–44.

 [38]. Id. at 39.

 [39]. Id. at 39–40.

 [40]. See id. at 40, 67–68.

 [41]. See generally id.

 [42]. See James M. Doyle, Ready for the Psychologists: Learning from Eyewitness Errors, 48 Ct. Rev. 4, 4–5 (2012).

 [43]. See id. at 4.

 [44]. Id.

 [45]. See Kansas v. Marsh, 548 U.S. 163, 182–99 (2006) (Scalia, J., concurring).

 [46]. Id. at 185–86.

 [47]. Id. at 188.

 [48]. Id.

 [49]. Id. at 193.

 [50]. Id. at 198.

 [51]. Id.

 [52]. See generally Dennis v. Sec’y, Pa. Dep’t of Corr., 834 F.3d 263, 313–44 (3d Cir. 2016) (McKee, C.J., concurring).

 [53]. Id. at 313.

 [54]. Id.

 [55]. Id.

 [56]. Id.

 [57]. Id. at 316 (quoting Comm. on Sci. Approaches to Understanding and Maximizing the Validity and Reliability of Eyewitness Identificationin Law Enforcement and the Courts et al., Identifying the Culprit: Assessing Eyewitness Identification 22 (2014)).

 [58]. See generally United States v. Wade, 388 U.S. 218 (1967).

 [59]. Id. at 228.

 [60]. Id.

 [61]. Id. at 231–33.

 [62]. Dan Simon, In Doubt: The Psychology of the Criminal Justice Process 50–51 (2012).

 [63]. Id.

 [64]. Id. at 51, 53.

 [65]. A. Daniel Yarmey, Eyewitness Recall and Photo Identification: A Field Experiment, 10 Psychol. Crime & L. 53, 53 (2004).

 [66]. Id.

 [67]. Id.

 [68]. Simon, supra note 62, at 51.

 [69]. Id.

 [70]. Id.

 [71]. Id. at 76.

 [72]. For an in-depth discussion of show-up procedures, see id. at 70–71, 77–78.

 [73]. Id. at 51–52.

 [74]. Id. at 52.

 [75]. Id. at 69.

 [76]. Id. at 51–52.

 [77]. Id. at 70.

 [78]. Id.

 [79]. Id. at 81.

 [80]. Id. at 72.

 [81]. Id. at 71.

 [82]. Id.

 [83]. Daniel B. Wright, The Impact of Eyewitness Identifications from Simultaneous & Sequential Lineups, 15 Memory 746, 748 (2007).

 [84]. Simon, supra note 62, at 71.

 [85]. Dennis v. Sec’y, Pa. Dep’t of Corr., 834 F.3d 263, 321 (3d Cir. 2016) (McKee, C.J., concurring).

 [86]. Steve D. Charman & Vanessa Quiroz, Blind Sequential Lineup Administration Reduces Both False Identifications and Confidence in Those False Identifications, 40 Law & Hum. Behav. 477, 477, 483–84 (2016).

 [87]. Id. at 477.

 [88]. Id. at 480.

 [89]. Id. at 477.

 [90]. Id.

 [91]. Id.

 [92]. Id. at 484.

 [93]. Laura Smalarz & Gary L. Wells, Confirming Feedback Following a Mistaken Identification Impairs Memory for the Culprit, 38 Law & Hum. Behav. 283, 283 (2014).

 [94]. Id.

 [95]. Id.

 [96]. Michael R. Leippe, Donna Eisenstadt & Shannon M. Rauch, Cueing Confidence in Eyewitness Identifications: Influence of Biased Lineup Instructions and Pre-Identification Memory Feedback Under Varying Lineup Conditions, 33 Law & Hum. Behav. 194, 197 (2009).

 [97]. See infra Part IV (discussing how various jurisdictions have either adopted or rejected this reform).

 [98]. Steven E. Clark, Michael A. Erickson & Jesse Breneman, Probative Value of Absolute and Relative Judgments in Eyewitness Identification, 35 Law & Hum. Behav. 364, 364 (2011).

 [99]. Id.

 [100]. Id. at 377.

 [101]. See, e.g., Comm. on Sci. Approaches to Understanding & Maximizing the Validity & Reliability of Eyewitness Identificationin Law Enf’t & the Courts et al., Identifying the Culprit: Assessing Eyewitness Identification 104 (2015) [hereinafter Comm. on Sci.].

 [102]. Leippe, supra note 96, at 196, 204.

 [103]. Id. at 196.

 [104]. Id.

 [105]. Id. at 194.

 [106]. Id.

 [107]. Id.

 [108]. See Dennis v. Sec’y, Pa. Dep’t of Corr., 834 F.3d 263, 341–44 (3d Cir. 2016) (McKee, C.J., concurring).

 [109]. Id. at 342.

 [110]. See, e.g., Richard A. Wise et al., An Examination of the Causes and Solutions to Eyewitness Error, Frontiers Psychiatry, Aug. 14, 2014, at 1, 4.

 [111]. Wright, supra note 83, at 747.

 [112]. Dennis, 834 F.3d at 342 (McKee, C.J., concurring).

 [113]. Id.

 [114]. Id.

 [115]. See generally Wise, supra note 110.

 [116]. Id. at 1–2.

 [117]. Dennis, 834 F.3d at 342–43 (McKee, C.J., concurring).

 [118]. See generally State v. Lawson, 291 P.3d 673 (Or. 2012); State v. Henderson, 27 A.3d 872 (N.J. 2011).

 [119]. N.C. Gen. Stat. § 15A-284.50–.53 (2015).

 [120]. See Henderson, 27 A.3d at 896–903.

 [121]. See id.

 [122]. Id. at 896–900.

 [123]. Id. at 878.

 [124]. Id. at 880–81.

 [125]. Id.

 [126]. Id.

 [127]. Id. at 881.

 [128]. Id.

 [129]. Id.

 [130]. Id.

 [131]. Id.

 [132]. See id. at 882–83.

 [133]. Id.

 [134]. See id. at 892, 896.

 [135]. Id. at 896–97.

 [136]. Id. at 897.

 [137]. Id.

 [138]. Id.

 [139]. Id.

 [140]. Id. at 897–98.

 [141]. Id. at 898.

 [142]. Id.

 [143]. Id.

 [144]. Id. at 900.

 [145]. Id. at 900–01.

 [146]. Id. at 901–02.

 [147]. Id.

 [148]. Id. at 910–11.

 [149]. Id. at 910.

 [150]. Id. at 894–95.

 [151]. Id. at 910–11.

 [152]. See id. at 910–11, 924–25.

 [153]. See State v. Lawson, 291 P.3d 673, 688 (Or. 2012).

 [154]. Id. at 678.

 [155]. Id.

 [156]. Id. at 683–84, 688–89.

 [157]. Id. at 683–84.

 [158]. Id. at 678–79.

 [159]. Id. at 679.

 [160]. Id.

 [161]. Id.

 [162]. Id.

 [163]. Id. at 679–80.

 [164]. Id. at 680.

 [165]. Id.

 [166]. Id. at 698.

 [167]. Id. at 685–88.

 [168]. Id. at 698.

 [169]. Id. at 693–94.

 [170]. Id. at 685.

 [171]. Id. at 686.

 [172]. Id.

 [173]. Id.

 [174]. Id.

 [175]. See id. at 698.

 [176]. Id. at 686–87.

 [177]. Id.

 [178]. Id. at 688.

 [179]. Id. at 693–94.

 [180]. Id.

 [181]. N.C. Gen. Stat. § 15A-284.50–.53 (2015).

 [182]. See id. § 15A-284.51–.52.

 [183]. Id. § 15A-284.52.

 [184]. Id. § 15A-284.52(b).

 [185]. Id.

 [186]. Id.

 [187]. Id.

 [188]. Id.

 [189]. Id. § 15A-284.52(c).

 [190]. Id. § 15A-284.53.

 [191]. Id.

 [192]. Id.

 [193]. See generally Comm. on Sci., supra note 101.

 [194]. See id. at xiii–xiv.

 [195]. Id. at 106.

 [196]. Id. at 105–12.

 [197]. Id. at 104.

 [198]. Id. at 107.

 [199]. Id. at 104.

 [200]. Id. at 106–07.

 [201]. Id. at 106.

 [202]. Id. at 107.

 [203]. Id. See also State v. Henderson, 27 A.3d 872, 897–99 (N.J. 2011).

 [204]. Comm. on Sci., supra note 101, at 109.

 [205]. See id.

 [206]. Id. at 112.

 [207]. See id.

 [208]. Id.

 [209]. Id.