Joseph Clifton Smith v. Commissioner, Alabama Department of Corrections

67 F.4th 1335

• Court: Court of Appeals for the Eleventh Circuit
• Decided: May 19, 2023
• Docket: 21-14519
• Status: Published

Opinion

USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 1 of 38

[PUBLISH] In the United States Court of Appeals For the Eleventh Circuit

____________________

No. 21-14519 ____________________

JOSEPH CLIFTON SMITH, Petitioner-Appellee, versus COMMISSIONER, ALABAMA DEPARTMENT OF CORRECTIONS,

Respondent-Appellant.

Appeal from the United States District Court for the Southern District of Alabama D.C. Docket No. 1:05-cv-00474-CG-M ____________________ USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 2 of 38

2 Opinion of the Court 21-14519

Before WILSON, JORDAN, and ROSENBAUM, Circuit Judges. PER CURIAM: This appeal concerns whether the district court clearly erred in finding that Joseph Clifton Smith is intellectually disabled and, as a result, that his death sentence violates the Eighth Amendment. We hold that the district court did not clearly err. We therefore affirm the district court’s judgment vacating Smith’s sentence. I. A. A jury found petitioner Joseph Clifton Smith guilty of capital murder. Durk Van Dam was brutally murdered on November 23, 1997. Smith v. Campbell (“Smith III”), 620 F. App’x 734, 736 (11th Cir. 2015). Police found Van Dam’s body in an isolated area near his pick-up truck in Mobile County. Id. On the same day that police discovered Van Dam’s body, they interviewed Petitioner Joseph Clifton Smith. Id. Although Smith confessed to Van Dam’s murder, he offered two conflicting versions of the crime. Id. At first, he said that he watched Van Dam’s murder. Id. Then, he said that he participated, but that he didn’t intend to kill Van Dam. Id. A grand jury in Mobile County eventually indicted Smith for capital murder. Id. The case went to trial, and the jury found Smith guilty. Id. at 736–37. USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 3 of 38

21-14519 Opinion of the Court 3

B. During the sentencing phase of Smith’s trial, the parties pre- sented evidence of Smith’s intellectual abilities. During the sentencing phase, the parties presented evidence concerning aggravating and mitigating factors. One mitigating fac- tor was whether Smith committed the crime while he “was under the influence of extreme mental or emotional disturbance.” Ala. Code § 13A-5-51(2). Both sides presented evidence of Smith’s child- hood, family background, and intellectual abilities to contest whether that mitigating factor applied to Smith. Smith’s mother and sister testified that his father was an abu- sive alcoholic. Smith III, 620 F. App’x at 738–39. Smith’s father beat the children with belts and water hoses. Id. Smith’s mother and father divorced when Smith was nine or ten years old. Id. at 738. Soon after his parents divorced, Smith’s mother remarried to a man named Hollis Luker. Like Smith’s father, Luker beat the children and was drunk “just about every day.” Id. at 739. Smith’s neighbor testified that his mother would bring Smith and his sib- lings to the neighbor’s home to escape Luker’s beatings. Id. In the meantime, Smith struggled in school. He had been described as a “slow learner” since he was in the first grade. Smith was eight years old when he reached third grade. At that point, he still needed help to function at a first-grade level, prompting his teacher to label him an underachiever and refer him for an “intel- lectual evaluation.” During that evaluation, Smith obtained a full-scale IQ score of 75. That score meant that Smith was “functioning in the USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 4 of 38

4 Opinion of the Court 21-14519

Borderline range of measured intelligence.” Smith’s school then asked his mother for permission to do more testing. At the beginning of Smith’s fourth-grade year, which coin- cided with his parents’ divorce, his mother agreed to have the school perform additional testing. After undergoing more testing, Smith was placed in a learning-disability class. After that placement, Smith developed an unpredictable temper and often fought with classmates. His behavior became so troublesome that his school placed him in an “emotionally con- flicted classroom.” These types of classrooms hosted special-edu- cation classes for students who could not adjust to a regular class- room, according to Dr. James Chudy, a clinical psychologist. Dr. Chudy met with Smith three times after Van Dam’s murder, ad- ministered several tests, analyzed records from Smith’s past, au- thored a report about his findings, and testified during Smith’s sen- tencing phase. Id. at 738–39. Smith’s academic deficits persisted through junior high school. When he entered sixth grade, his school reevaluated his intellectual abilities. This time, he obtained a full-scale IQ score of 74, again placing him “in the Borderline range of measured intelli- gence.” By grade seven, the school determined that Smith was el- igible for the “Educable [Intellectually Disabled]” program. He went on to fail the seventh and eighth grades before dropping out of school for good. Id. at 740. Smith spent much of the next fifteen years in prison. When he was nineteen, Smith went to prison for burglary and receiving USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 5 of 38

21-14519 Opinion of the Court 5

stolen property. He was released from prison after six years. But he returned a year later when he violated the conditions of his pa- role. There he remained until his release in November 1997, just two days before Van Dam’s murder. Dr. Chudy reevaluated Smith just after Van Dam’s murder. When Dr. Chudy tested Smith’s IQ, Smith obtained a full-scale score of 72. During the sentencing phase, Dr. Chudy testified that Smith’s true IQ score could be as high as 75 or as low as 69 after accounting for the standard error of measurement inherent in IQ tests. “69 is considered clearly [intellectually disabled],” 1 he ex- plained. Either way, Smith’s raw score of 72 suggests that he func- tions at a lower level than 97% of the general population. Dr. Chudy also described Smith as “barely literate in reading.” The sentencing phase eventually came to an end, and the Al- abama trial court found that the aggravating circumstances out- weighed the mitigating ones. The court thus sentenced Smith to death. C. Smith petitioned for habeas relief and argued, among other things, that his sentence violates the Eighth and Fourteenth Amendments because he is intellectually disabled. After exhausting his direct appeals, Smith sought habeas re- lief in state court. He argued, among other things, that his

1 We alter quotations that use outdated language to describe intellectual disa- bilities. E.g., Brumfield v. Cain, 576 U.S. 305, 308 n.1 (2015); Kilgore v. Sec’y, Fla. Dep’t of Corr., 805 F.3d 1301, 1303 n.1 (11th Cir. 2015). USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 6 of 38

6 Opinion of the Court 21-14519

sentence violated the Eighth and Fourteenth Amendments because he is intellectually disabled. See Atkins v. Virginia, 536 U.S. 304 (2002). Consistent with the medical community’s general consen- sus, Alabama law defines intellectual disability as including three criteria: (1) significantly subaverage intellectual functioning (i.e., an IQ of 70 or below); (2) significant or substantial deficits in adap- tive behavior; and (3) the onset of those qualities during the devel- opmental period (i.e., before the age of 18). Ex parte Perkins, 851 So. 2d 453, 456 (Ala. 2002). Applying that definition, the Alabama Court of Criminal Appeals ultimately rejected Smith’s Atkins claim, finding that he could not meet the intellectual-disability criteria based on the evi- dence adduced during his trial and sentencing phase. See Smith v. State (“Smith I”), 71 So. 3d 12, 19–21 (Ala. Crim. App. 2008) (“[T]he record in Smith’s direct appeal supports the circuit court’s conclu- sion that Smith does not meet the broadest definition of [intellec- tually disabled] adopted by the Alabama Supreme Court in Ex parte Perkins, 851 So. 2d 453 (Ala. 2002).”), cert denied, No. 1080589 (Ala. 2010) (mem.). Smith then invoked 28 U.S.C. § 2254 and pressed his Atkins claim in federal court. The district court rejected Smith’s Atkins claim without holding an evidentiary hearing, concluding that the Alabama Court of Criminal Appeals did not unreasonably apply federal law. Smith v. Thomas (“Smith II”), No. 05-0474-CG-M, 2013 WL 5446032, at *29 (S.D. Ala. 2013). In doing so, the district court USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 7 of 38

21-14519 Opinion of the Court 7

relied on Smith’s failure “to prove that his intellectual functioning was or is significantly subaverage,” id. at *29 n.1, which is the first prong for Alabama’s intellectual-disability definition and requires an IQ of 70 or below. Ex parte Perkins, 851 So. 2d at 456. The district court therefore treated “an IQ of 70 as the ceiling for significantly subaverage intellectual functioning” and held that Smith’s full-scale IQ scores of 75, 74, and 72 were “fatal to Smith’s Atkins claim.” Smith II, 2013 WL 5446032, at *28–29. Smith then appealed, and we reversed. Smith III, 620 F. App’x at 749–52. We first explained that Alabama law does not em- ploy “a strict IQ cut-off of 70” to define significantly subaverage in- tellectual functioning. Id. at 749. And that was key because Dr. Chudy’s testimony during the sentencing phase “showed that Smith’s IQ could be as low as 69 given a standard error of measure- ment of plus-or-minus three points.” Id. at 749–50 (citation omit- ted). We also noted that “other trial evidence” suggested that Smith had “deficits in intellectual functioning,” id. at 750. Based on that evidence and “the fact that Alabama does not employ a strict IQ cut-off score of 70,” we held that the Alabama Court of Crimi- nal Appeals “determination that Smith conclusively did not possess significantly subaverage intellectual functioning was an unreasona- ble determination of the facts.” Id. (citation omitted). We then turned to the Alabama Court of Criminal Appeals’s finding “that Smith did not suffer from significant or substantial deficits in adaptive behavior.” Id. This, too, was an unreasonable determination of the facts, we said, because the record contained USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 8 of 38

8 Opinion of the Court 21-14519

evidence “that would support a fact finding that Smith had signifi- cant limitations in at least” two areas of adaptive functioning: “(1) social/interpersonal skills and self-direction.” Id. We therefore said that “the record affirmatively contradicts” the Alabama Crim- inal court of Appeals’s finding that Smith did not suffer from sig- nificant defects in adaptive behavior. Id. at 750–51. For those reasons, we remanded Smith’s Atkins claim to the district court. Id. at 751. We instructed the district court “to allow Smith . . . to present an expert witness on his behalf.” Id. at 750– 51. And we also instructed the district court to consider “Smith’s requests for discovery and an evidentiary hearing.” Id. at 752. D. The district court held an evidentiary hearing to determine whether Smith is intellectually disabled. On remand, the district court held an evidentiary hearing to assess whether Smith is intellectually disabled. The district court heard lay and expert testimony and received reports from experts who evaluated Smith and analyzed his records. i. Evidence of Smith’s Intellectual Functioning Smith’s first witness was Dr. Daniel Reschly, a certified school psychologist with fifty years of experience in assessing intel- lectual disability. Since 1998, Dr. Reschly has taught at (and some- times chaired) the top-ranked Department of Special Education at Peabody College of Education and Human Development. His teaching and research focus on identifying and treating “persons with mild intellectual disability.” USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 9 of 38

21-14519 Opinion of the Court 9

Dr. Reschly testified that people with mild intellectual disa- bility exhibit “borderline and overall low intellectual performance.” “It’s important” to treat a person’s IQ score as indicating a range of scores, he said, because the medical community can only approxi- mate a person’s true IQ. This concept reflects the standard error of measurement inherent in IQ tests. So “a range of about 65 to 75” is the “level for someone’s performance on an IQ test consistent with mild intellectual disability,” Dr. Reschly explained. Smith also called Dr. John Fabian, who holds a doctorate in clinical psychology and works as a forensic psychologist. When Dr. Fabian assessed Smith’s IQ, Smith obtained a full-scale IQ score of 78. Although Dr. Fabian conceded that “a 78 is definitively above” the “70 to 75 IQ range,” he testified that Smith’s 78 does not elimi- nate the possibility that Smith is intellectually disabled. To support that answer, he cited Smith’s other IQ scores, all of which were lower than 75. Those scores, he said, “trump an overall score on one administration.” For its part, the state called Dr. Glen King, a clinical and fo- rensic psychologist who also practices law. When Dr. King assessed Smith’s IQ, Smith obtained a full-scale IQ score of 74. As a result, Smith has taken five IQ tests during his lifetime. And he has ob- tained full-scale IQ scores of 75, 74, 72, 78, and 74. Dr. King there- fore testified that Smith displayed “a very consistent pattern of in- tellectual quotient scores” on all five tests. In other words, he tes- tified that the standard error of measurement deserves less weight here because Smith’s scores all “fall in the borderline range of USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 10 of 38

10 Opinion of the Court 21-14519

intellectual functioning.” “I think that the scores speak for them- selves,” and “they are what they are,” he said. ii. Evidence of Smith’s Adaptive Behavior While intellectual functioning aims “to assess the individ- ual’s best level of functioning,” Dr. Reschly testified that adaptive behavior looks at the person’s “typical performance” and asks, “[W]hat do they do on a day-to-day basis?” A person has significant adaptive behavior limitations if he has “significant deficits in one of [three] areas: conceptual, social, and practical.” The conceptual do- main includes literacy skills, language, and financial literacy. “The social domain of adaptive behavior refers to various social compe- tencies” that a person “use[s] on an everyday basis.” “The practical domain includes a wide diverse set of behaviors that” involve “sim- ple self-care” including “eating, toileting, [and] dressing oneself.” A person who shows “significant deficits in one of those areas” meets the medical community’s standard for having significant deficits in adaptive behavior. Dr. King testified that the ABAS-3 test is the “only” test that is “appropriate” for assessing a person’s adaptive functioning. The test requires the subject to read a series of statements describing a behavior and rate, on a scale of one to three, how often they per- form that behavior without a reminder and without help. Dr. King administered the ABAS-3 test when he met with Smith before the evidentiary hearing. At the evidentiary hearing, Dr. King testified that in his “experience with capital litigation USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 11 of 38

21-14519 Opinion of the Court 11

cases,” Smith “generated the highest scores” on the ABAS-3 that Dr. King has seen. For his part, Dr. Fabian used a different test—called the In- dependent Living Scales test—to assess Smith’s adaptive behavior. The results suggested to Dr. Fabian that Smith had “deficits in every area.” Dr. King sought to undermine those results by testifying that the Independent Living Scales test “is not a recommended device for assessing adaptive behavior.” But in other cases where he pro- vided expert testimony, Dr. King testified that the Independent Liv- ing Scales test “measures adaptive functioning in a number of dif- ferent domains,” Tarver v. State, 940 So. 2d 312, 324 (Ala. Ct. Crim. App. 2004) (Cobb, J., concurring in part and dissenting in part). Dr. Reschly discussed Smith’s “failure to acquire literacy skills at an age-appropriate level, which relates to the conceptual demand of adaptive behavior.” Dr. Fabian agreed. Smith’s school records show signs “consistent with significant limitations in at least [the] conceptual domain,” he said. Dr. Reschly and Dr. Fabian also testified that Smith exhibited deficits in the social domain of adaptive behavior. Relying on Smith’s school records, Dr. Reschly testified that Smith was poor at following rules, obeying instructions, and forming relations with his peers. Dr. Fabian agreed. Dr. Fabian assessed Smith’s communication skills using the Expressive One-Word Picture Vocabulary and the Receptive One- Word Picture Vocabulary tests. The Expressive test assessed USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 12 of 38

12 Opinion of the Court 21-14519

Smith’s ability to express through language; the Receptive test as- sessed his receptiveness to language. Both tests relate “to func- tional academics or conceptual areas of adaptive functioning and even academic achievement,” said Dr. Fabian. Smith scored in the first percentile on the expressive test and in the third percentile on the receptive test. The age equivalents for those scores are thirteen and fifteen, respectively. Those scores, according to Dr. Fabian, “are consistent with someone who is intellectually disabled.” iii. Evidence of Smith’s Developmental Period As Dr. Reschly explained, the medical community defines in- tellectual disability to include not only deficits in intellectual and adaptive functioning, but also the onset of those qualities during the developmental period. Dr. Reschly said that Smith satisfies this prong of the intellectual-disability definition because Smith was placed in an “Educable [Intellectually Disabled]” program while he was in school, the criteria for which is “largely parallel to the crite- ria used to identify mild intellectual disability today.” Dr. Reschly also testified that Smith’s school records reflect that Smith exhib- ited symptoms “consistent” with someone who has “adaptive be- havioral deficits and the intellectual functioning deficits.” Dr. Fabian also concluded that Smith exhibited behavior “consistent with mild intellectual disability” during the develop- mental period. Dr. Fabian reached that conclusion after reviewing Smith’s school records and Dr. Chudy’s report. USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 13 of 38

21-14519 Opinion of the Court 13

E. After the evidentiary hearing, the district court found that Smith is intellectually disabled and therefore granted his ha- beas petition. After the evidentiary hearing, the district court issued an or- der and found that Smith is intellectually disabled. Smith v. Dunn (“Smith IV”), No. 05-00474-CG, 2021 WL 3666808, at *1 (S.D. Ala. Aug. 17, 2021). Under Alabama law, the court explained, Smith had the burden of establishing (1) that he has significantly subaverage intellectual functioning (i.e., an IQ of 70 or below); (2) that he has significant or substantial deficits in adaptive behavior; and (3) that those qualities manifested during the developmental period (i.e., before he turned 18). Id. at *2 (citation omitted). Starting with the first prong, the district court explained that when an offender’s IQ score is close to, but higher than, 70, he “must be allowed to present additional evidence of intellectual dis- ability, including testimony of adaptive deficits.” Id. (quoting Smith v. Comm’r, Ala. Dep’t of Corr., 924 F.3d 1330, 1337 (11th Cir. 2019)). The court then noted that Smith had “scores as low as 72, which according to testimony could mean his IQ is actually as low as 69 if you take into account the standard error of measurement.” Id. At the same time, the court recognized that “all of Smith’s IQ scores” are higher than 70. Id. at *3. The court then acknowledged Dr. King’s testimony that the consistency with which Smith scored above 70 makes it more likely that his true IQ is higher than 70. Id. But the court did not find Dr. King’s testimony “strong enough” to throw out the lowest score “as an outlier” or to disregard the stand- ard error of measurement. Id. The court therefore determined USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 14 of 38

14 Opinion of the Court 21-14519

that it needed to consider additional evidence, including testimony about Smith’s adaptive deficits. Id. Then the court turned to Smith’s adaptive behavior. Id. at *4. Invoking our decision in Smith III, the court explained that evi- dence from Smith’s sentencing phase “support[ed] a fact finding that Smith had significant limitations in at least two” areas of adap- tive behavior: “(1) social/interpersonal skills and (2) self-direction.” Id. at *5 (quoting Smith III, 620 F. App’x at 750). Besides evidence, the court noted that evidence from the evidentiary hearing, like the results from Dr. Fabian’s Independent Living Scales Test, “indi- cated that Smith had deficits in most areas” of adaptive function- ing. Id. at *10. The court acknowledged Dr. King’s criticism of the Inde- pendent Living Scales test. Id. But the court “question[ed] the ve- racity of Dr. King’s criticism” because he used the Independent Liv- ing Scales test in another case and testified that the test “measures adaptive functioning in a number of different domains.” Id. (quot- ing Tarver, 940 So. 2d at 324.) In the end, the court explained that “whether Smith has sig- nificant or substantial deficits in adaptive behavior largely comes down to which expert is believed.” Id. at *11. The court then found that “Smith has significant deficits in social/interpersonal skills, self-direction, independent home living, and functional academ- ics.” Id. at *11. For that reason, the court found that “Smith has shown by a preponderance of the evidence that he has significantly USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 15 of 38

21-14519 Opinion of the Court 15

subaverage intellectual functioning and significant deficits in adap- tive behavior.” Id. The question thus became whether Smith’s deficits in intel- lectual and adaptive functioning manifested during the develop- mental period. The court noted that Smith “enrolled in [Educable Intellectually Disabled] classes in the 7th and 8th grades” and that, according to Dr. Reschly, the criteria for such classes “was largely parallel to the criteria used to identify mild intellectual disability today.” Id. at *11–12 (internal quotations omitted). The court also cited testimony from Dr. Fabian, who similarly concluded that Smith exhibited behavior “consistent with intellectual disability” during the developmental period. Id. at *12. The court therefore found “that Smith’s intellectual and adaptive functioning issues clearly arose before he was 18 years of age.” Id. For those reasons, the court granted Smith’s habeas petition and vacated his death sentence, explaining that “Smith is intellec- tually disabled and cannot constitutionally be executed.” Id. at *13. II. Whether a capital offender suffers from an intellectual disa- bility is a question of fact. Ledford v. Warden, Ga. Diagnostic & Clas- sification Prison, 818 F.3d 600, 632 (11th Cir. 2016) (quoting Fults v. GDCP Warden, 764 F.3d 1311, 1319 (11th Cir. 2014)). We thus review for clear error a district court’s finding that an individual is intellec- tually disabled. Id. (citing Conner v. GDCP Warden, 784 F.3d 752, 761 (11th Cir. 2015)). “Clear error is a highly deferential standard of review.” Holladay v. Allen, 555 F.3d 1346, 1354 (11th Cir. 2009) USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 16 of 38

16 Opinion of the Court 21-14519

(citation omitted). “Under that standard, we may not reverse just because we ‘would have decided the [matter] differently.’ A finding that is ‘plausible’ in light of the full record—even if another is equally or more so—must govern.” Cooper v. Harris, 581 U.S. 285, 293 (2017) (citations omitted). III. The question presented is whether the district court clearly erred by finding that Smith is intellectually disabled and, as a result, that his sentence violates the Eighth and Fourteenth Amendments. The Eighth and Fourteenth Amendments prohibit states from ex- ecuting intellectually disabled offenders. Atkins, 536 U.S. at 321. That prohibition stems from “a national consensus” against the practice of executing such offenders. Id. at 316. “To the extent there is serious disagreement about the execution of [intellectually disabled] offenders, it is in determining which offenders are in fact [disabled].” Id. at 317. To resolve that disagreement, the Supreme Court has granted the states some discretion to develop standards for as- sessing whether an offender is intellectually disabled. Id. (quoting Ford v. Wainwright, 477 U.S. 399, 416–17 (1986)). But states do not wield “unfettered discretion” to determine “how intellectual disa- bility should be measured and assessed.” Hall v. Florida, 572 U.S. 701, 719 (2014). Instead, a state’s assessment of whether an offender is intel- lectually disabled “must be ‘informed by the medical community’s diagnostic framework.’” Moore v. Texas, 581 U.S. 1, 13 (2017) USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 17 of 38

21-14519 Opinion of the Court 17

(quoting Hall, 581 U.S. at 721). Courts identify that framework us- ing “the most recent (and still current) versions of the leading diag- nostic manuals—the DSM-5 and the AAIDD-11.” Id. (citing Hall, 572 U.S. at 704–05, 713); see also Am. Psychiatric Ass’n, Diagnostic and Statistical Manual of Mental Disorders (5th ed. 2013) (hereinafter DSM-5); Am. Ass’n on Intell. & Dev. Disabilities, Intellectual Disabil- ity: Definition, Classification, and Systems of Support (12th ed. 2021) (hereinafter AAIDD-12). We start, then, with Alabama’s standard for determining in- tellectual disability. Under Alabama law, Smith “has the burden of proving by a preponderance of the evidence that he . . . is [intellec- tually disabled] and thus ineligible for the death penalty.” Smith v. State, 213 So. 3d 239, 252 (Ala. 2007). Carrying that burden requires Smith “to show significant subaverage intellectual functioning at the time the crime was committed, to show significant deficits in adaptive behavior at the time the crime was committed, and to show that these problems manifested themselves before the de- fendant reached the age of 18.” Id. at 249. IV. Whether Smith has significantly subaverage intellectual functioning turns on whether he has an IQ equal to or less than 70. Ex parte Perkins, 851 So. 2d at 456. But the medical community rec- ognizes “that the IQ test is imprecise.” Hall, 572 U.S. at 723. “Each IQ test score has a ‘standard error of measurement.’” Id. at 713 (citation omitted). “The standard error of measurement accounts for a margin of error both below and above the IQ test-taker’s USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 18 of 38

18 Opinion of the Court 21-14519

score.” Ledford, 818 F.3d at 640. The standard error of measure- ment thus “allows clinicians to calculate a range within which one may say an individual’s true IQ score lies.” Hall, 572 U.S. at 713. For that reason, the intellectual functioning inquiry must recognize “that an IQ test score represents a range rather than a fixed number.” Id. at 723. So when the lower end of that range is equal to or less than 70, an offender “must be able to present addi- tional evidence of intellectual disability, including testimony re- garding adaptive deficits.” Id. at 723; see also Moore, 581 U.S. at 14 (“Because the lower end of Moore’s score range falls at or below 70, the [Texas Court of Criminal Appeals] had to move on to con- sider Moore’s adaptive functioning.”). A. The district court did not err by turning to evidence of Smith’s adaptive functioning after finding that his IQ score could be as low as 69. While he was in school, Smith took two IQ tests. He ob- tained a full-scale IQ score of 75 on the first test. On the second test, he obtained a full-scale score of 74. Dr. Reschly testified that those scores are consistent with mild intellectual disability, “partic- ularly if you consider the standard error of measurement.” Dr. Chudy assessed Smith’s IQ for a third time after Van Dam’s murder. Smith obtained a full-scale score of 72 on that test. Based on that test, Dr. Chudy testified that Smith’s true IQ score could be as high as 75 or as low as 69 after accounting for the test’s standard error of measurement. He added that “69 is considered clearly [intellectually disabled].” And when he was asked whether USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 19 of 38

21-14519 Opinion of the Court 19

that finding was consistent with the results on Smith’s prior IQ tests, Dr. Chudy said, “Yes, all the scores are very much the same.” Then, before the evidentiary hearing, Smith obtained a full- scale IQ score of 74 on the test that Dr. King administered. Because that score falls within the 70 to 75 range, Dr. Fabian testified that the results of Dr. King’s IQ test are consistent with mild intellectual disability. Dr. Fabian also tested Smith’s IQ ahead of the evidentiary hearing. Smith obtained a full-scale score of 78 on that test. Alt- hough Dr. Fabian conceded that “a 78 is definitively above” the “70 to 75 IQ range,” he testified that Smith’s 78 does not eliminate the possibility that Smith is intellectually disabled. Instead, he cited Smith’s other scores, all of which were lower than 75, and said that those scores “trump an overall score on one administration.” Dr. King contradicted Dr. Fabian. According to Dr. King, Smith displayed “a very consistent pattern of intellectual quotient scores” on all five tests. Dr. King therefore testified that the stand- ard error of measurement deserves less weight because Smith’s scores all “fall in the borderline range of intellectual functioning.” “I think that the scores speak for themselves,” he said, “they are what they are.” In the end, the district court said that Dr. King’s testimony was not “strong enough” for the court to find “that the lowest score can be thrown out as an outlier or that the standard error for the tests can be disregarded.” Smith IV, 2021 WL 3666808, at *3. As the district court twice noted, Smith had an IQ score of 72, USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 20 of 38

20 Opinion of the Court 21-14519

meaning that his IQ could be “as low as 69 if you take into account the standard error of measurement.” Id. at *2; id. at *3. The court therefore “conclude[d] that additional evidence must be consid- ered, including testimony on [Smith’s] adaptive deficits.” Id. at *3. In reaching that conclusion, the district court merely applied the Supreme Court’s decisions in Hall and Moore, which hold that a district court must move on to consider an offender’s adaptive func- tioning when the lower end of his lowest IQ score is equal to or less than 70. We start with Hall, which arose after the Florida Supreme Court denied Freddie Lee Hall’s Atkins claim that he could not be put to death because he was intellectually disabled. Hall “had re- ceived nine IQ evaluations in 40 years, with scores ranging from 60 to 80,” Hall, 572 U.S. at 707. Because “the sentencing court ex- cluded the two scores below 70 for evidentiary reasons,” that left only seven “scores between 71 and 80.” Id. And because none of those scores were equal to or lower than 70, the Florida Supreme Court rejected Hall’s Atkins claim and affirmed his death sentence. Id. (citation omitted). The Supreme Court reversed. It said that when an offender’s “IQ test score falls within the test’s acknowledged and inherent margin of error, the [offender] must be able to present additional evidence of intellectual disability, including testimony regarding adaptive deficits.” Id. at 723. Because Hall had obtained an IQ score as low as 71, the Court held that “the law require[d] that he have an USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 21 of 38

21-14519 Opinion of the Court 21

opportunity to present evidence of his intellectual disability, includ- ing deficits in adaptive functioning over his lifetime.” Id. at 724 Now for Moore, which arose after the Texas Court of Crimi- nal Appeals denied Bobby Moore’s Atkins claim. Moore, 581 U.S. at 5. Although Moore had obtained IQ scores of 74 and 78, 2 the Texas Court of Criminal Appeals “discounted the lower end of the stand- ard-error range associated with those scores” and concluded that Moore functioned above the intellectually disabled range. Id. at 10 (citation omitted). Again, the Supreme Court reversed, this time explaining that “Moore’s score of 74, adjusted for the standard error of measure- ment, yields a range of 69 to 79,” id. at 14. “Because the lower end of Moore’s score range falls at or below 70,” the Supreme Court said that the Texas Court of Criminal Appeals “had to move on to consider Moore’s adaptive functioning.” Id. In sum, then, both Hall and Moore hold that when an of- fender’s lowest IQ score, adjusted for the test’s standard error of measurement, is equal to or less than 70, a court must move on and consider evidence of the offender’s adaptive deficits. See Hall, 572 U.S. at 707, 724 (holding that “the law require[d]” that Hall have an “opportunity to present evidence” concerning his “adaptive func- tioning” when his lowest score was a 71, even though he also

2 Although the habeas court credited seven of Moore’s IQ scores, the Texas Court of Criminal Appeals rejected five of those scores as unreliable and “lim- ited its appraisal to Moore’s scores” of 78 and 74. Id. at 8, 10. USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 22 of 38

22 Opinion of the Court 21-14519

obtained six other IQ scores, including an 80); Moore, 581 U.S. at 14 (holding that the Texas courts “had to move on to consider Moore’s adaptive functioning” when his lowest score, “adjusted for the standard error of measurement, yield[ed] a range of 69 to 79”); see also Jackson v. Payne, 9 F.4th 646, 654 (8th Cir. 2021) (disregarding a habeas petitioner’s IQ score of 81 and holding that “the district court ‘had to move on to consider [the petitioner’s] adaptive func- tioning’” when his lowest score’s score range was less than 70 (quoting Moore, 581 U.S. at 14)). And that is exactly what the district court did here. It first noted that Smith “had IQ test scores as low as 72,” suggesting that “his IQ is actually as low as 69 if you take into account the standard error of measurement.” Smith IV, 2021 WL 3666808, at *2. The court then court declined to treat that score as an outlier. Id. at *3. And as a result, the court “conclude[d] that additional evidence must be considered, including testimony” concerning Smith’s “adaptive deficits.” Id. B. Alabama’s arguments to the contrary are unpersuasive. Alabama argues that the district court erred in three ways. We’ll start with Alabama’s argument that the district court clearly erred when it found that Smith suffers from significantly subaver- age intellectual functioning. That finding was clear error, Alabama says, because all Smith’s IQ scores “place him in the borderline range of intelligence.” Given that consistency, Alabama contends that the standard error of measurement warrants less weight. USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 23 of 38

21-14519 Opinion of the Court 23

This argument ignores Hall and Moore. Just as Smith scored between 72 and 78 on five IQ tests, Freddie Lee Hall scored be- tween 71 and 80 on seven IQ tests. Hall, 572 U.S. at 707. 3 Relying on the lowest of those scores, the Supreme Court mandated that Hall “have the opportunity to present evidence of his intellectual disability, including deficits in adaptive functioning over his life- time.” Id. at 724. The Supreme Court reached this conclusion, even though Hall’s highest score was an 80—two points more than Smith’s highest score here. Heeding Hall’s command, the district court relied on Smith’s lowest score and turned to “additional evi- dence” including testimony concerning Smith’s adaptive deficits. Smith IV, 2021 WL 3666808, at *3. Alabama contends that we have read Hall in a way that per- mits the district court to ignore the lower end of an offender’s standard-error range. Alabama is not wrong. In Ledford, 4 we sug- gested that Hall’s “consideration of the standard error of measure- ment ‘is not a one-way ratchet.’” Ledford, 818 F.3d at 641 (quoting

3 In fact, Hall had nine IQ scores between 60 and 80, “but the sentencing court excluded the two scores below 70 for evidentiary reasons,” id. 4 Although Alabama also relies on our decision in Jenkins v. Commissioner, Ala- bama Department of Corrections, 963 F.3d 1248 (11th Cir. 2020), we declined to apply Hall retroactively in that case. See id. at 1275 (declining to apply Hall because “our Circuit has specifically held that Hall is not retroactive to cases on collateral review”). We need not address Hall’s (or Moore’s) non-retroac- tivity here (1) because we already set aside the Alabama court’s denial of Smith’s Atkins claim, see Smith III, 620 F. App’x at 746–52; and (2) because this is Smith’s first § 2254 petition. USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 24 of 38

24 Opinion of the Court 21-14519

Mays v. Stephens, 757 F.3d 211, 218 n.17 (5th Cir. 2014)). Instead, we said that “the standard error of measurement is merely a factor to consider when assessing an individual’s intellectual functioning— one that may benefit or hurt that individual’s Atkins claim, depend- ing on the content and quality of expert testimony presented.” Id. at 640–41; but see United States v. Wilson, 170 F. Supp. 3d 347, 366 (E.D.N.Y. 2016) (“[T]he facts in Hall require lower courts to con- sider evidence of adaptive functioning if even one valid IQ test score generates a range that falls to 70 or below.”). Our decision in Ledford predates Moore, though. And Moore rejects Ledford’s assertion that a district court can consider anything other than the lower end of an offender’s standard-error range. See Moore, 581 U.S. at 10, 14; see also Jackson, 9 F.4th at 655 n.8. 5 Indeed, Moore requires courts to move on and consider adaptive deficits when the lower end of an offender’s standard-error range is equal to or less than 70. And to the extent that Ledford holds otherwise,

5 Moore arose after the Texas Court of Criminal Appeals “discounted the lower end of the standard-error range associated” with Moore’s lowest admissible score (a 74). 581 U.S. at 10 (citation omitted). Instead of focusing on the stand- ard-error range associated with Moore’s 74, the Texas court cited Moore’s ac- ademic history and his depression and suggested that those factors “might have hindered his performance” on the IQ test that generated the 74. Id. (ci- tation omitted). But the Supreme Court reversed, explaining that “the pres- ence of other sources of imprecision in administering the test to a particular individual cannot narrow the test-specific standard-error range.” Id. at 14 (cleaned up). Because the lower end of Moore’s score range fell at or below 70, the Texas court “had to move on to consider Moore’s adaptive function- ing.” Id. USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 25 of 38

21-14519 Opinion of the Court 25

see Ledford, 818 F.3d at 641 (suggesting that “the standard error of measurement is a bi-directional concept that does not carry with it a presumption that an individual’s IQ falls at the bottom of his IQ range”), Ledford is no longer good law. In sum, the district court did not clearly err by considering Smith’s adaptive deficits. To the contrary, Hall and Moore required the district court to turn to evidence of Smith’s adaptive deficits because the lower end of his standard-error range was 69. See Smith IV, 2021 WL 3666808, at *3. Alabama also argues that the district court erred by failing “to require Smith to prove by a preponderance of the evidence that he has significantly subaverage intellectual functioning.” On this view, the district court’s order “focused only on the testimony of ” Dr. King and Dr. Chudy, 6 “both of whom found that Smith func- tions in the borderline range of intelligence.” We disagree, though, because Smith carried his burden un- der the intellectual prong through Dr. Chudy’s testimony. To

6 The district court’s order never says that Dr. King’s and Dr. Chudy’s testi- mony was the only evidence it considered when assessing Smith’s intellectual functioning. So Alabama’s argument builds from an incorrect premise, for “we assume all courts base rulings upon a review of the entire record.” Haynes v. McCalla Raymer, LLC, 793 F.3d 1246, 1249 (11th Cir. 2015) (quoting Funchess v. Wainwright, 722 F.2d 683, 694 (11th Cir. 1985)). So regardless of what evi- dence the district court’s order did or did not cite, we will not find clear error when “the district court’s account of the evidence is plausible in light of the record viewed in its entirety,” Anderson v. Bessemer City, 470 U.S. 564, 674 (1985) (emphasis added). USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 26 of 38

26 Opinion of the Court 21-14519

satisfy the intellectual-functioning prong, as we have observed, Smith needed to prove only that the lower end of his standard-error range is equal to or less than 70. And while Dr. Chudy found that Smith functions in the borderline range of intelligence, Dr. Chudy explained that functioning in the borderline range “means that [Smith] operates between the Low Average and [intellectually dis- abled] range.” Smith III, 620 F. App’x at 740. In other words, Dr. Chudy treated Smith’s IQ score “not as a single fixed number but as a range.” Hall, 572 U.S. at 712. And Dr. Chudy found that the lower end of that range was 69. Smith III, 620 F. App’x at 738. “69 is considered clearly [intellectually dis- abled].” Id. at 738. Alabama’s final argument is that the district court commit- ted legal error by failing to make a finding concerning Smith’s in- tellectual functioning. But of course, the district court did make a finding concerning Smith’s intellectual functioning—it found that Smith “had IQ test scores as low as 72” and that a score of 72 “is actually as low as 69 if you take into account the standard error of measurement.” Smith IV, 2021 WL 3666808, at *2. As a result, the district court had to move on to assess Smith’s adaptive deficits. See Moore, 581 U.S. at 14 (requiring the Texas courts “to move on” and “consider Moore’s adaptive functioning” when his lowest score, “adjusted for the standard error of measurement, yield[ed] a range of 69 to 79”); Hall, 572 U.S. at 724 (requiring that Hall have an “op- portunity to present evidence” concerning his “adaptive function- ing” when his lowest score was a 71). USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 27 of 38

21-14519 Opinion of the Court 27

V. We turn now to the adaptive-functioning prong. To satisfy this prong, Smith needed to demonstrate “significant or substantial deficits in adaptive behavior.” Ex parte Perkins, 851 So. 2d at 456; see also Carroll v. State, 300 So. 3d 59, 65 (Ala. 2019) (noting that “assess- ments of adaptive functioning must adhere to the ‘medical com- munity’s current standards’” (quoting Moore, 581 U.S. at 20)). This criterion refers “to how well a person meets community standards of personal independence and social responsibility, in comparison to others of similar age and social background.” DSM-5, at 37; AAIDD-12, at 29 (“Adaptive behavior is the collection of concep- tual, social, and practical skills that have been learned and are per- formed by people in their everyday lives.”). “Adaptive functioning involves adaptive reasoning in three domains: conceptual, social, and practical.” DSM-5, at 37. Deficits in any one of those domains satisfies the adaptive-functioning prong. See Moore, 581 U.S. at 15–16 (citation omitted); DSM-5 at 38 (explaining that the adaptive-functioning criterion “is met when at least one domain of adaptive functioning—conceptual, social, or practical—is sufficiently impaired” such that “ongoing support” is necessary “for the person to perform adequately in one or more life settings at school at work, at home, or in the community”); AAIDD-12, at 31 (explaining that “the ‘significant limitations in adaptive behavior’ criterion” requires “an adaptive behavior score that is approximately 2 standard deviations or more below the mean in at least one of the three adaptive behavior domains, con- ceptual, social, or practical”). USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 28 of 38

28 Opinion of the Court 21-14519

After the evidentiary hearing, the district court found that “Smith has significant deficits in social/interpersonal skills, self-di- rection, independent home living, and functional academics.” Smith IV, 2021 WL 3666808, at *11. That conclusion aligns with the one we reached before the evidentiary hearing, when we said that the record contained evidence “that would support a finding of fact that Smith had significant limitations in at least two” areas: “(1) so- cial/interpersonal skills and self-direction.” Smith III, 620 F. App’x at 750. 7 And the evidentiary hearing only reinforced that conclu- sion. Dr. Fabian used the Independent Living Scales test to assess Smith’s adaptive behavior. “The ILS is probably the most readily used adaptive functioning one-on-one test used nationally in foren- sic psychology,” said Dr. Fabian. The test required Smith to answer questions like “what the purpose of a will is, what would he do if he had a pain in his chest,” how would he fix things in his home, and how would he use a map “to drive from point A to point B.” Based on that assessment, Dr. Fabian concluded that Smith had “deficits in every area” of adaptive functioning. To be sure, Dr. King testified that the ILS test “is not a rec- ommended device for assessing adaptive behavior.” But Dr. King uses the ILS test to evaluate whether someone “can manage them- selves personally.” “That really is what the device was designed to

7 According to Dr. Reschly and Dr. Fabian, self-direction is a subcategory that falls within the conceptual domain. USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 29 of 38

21-14519 Opinion of the Court 29

do.” Of course, whether a person “can manage themselves” is at the very core of adaptive functioning. See DSM-5, at 37; AAIDD- 12, at 29. So Dr. King’s own testimony contradicts his criticism of the ILS test. In fact, the district court “question[ed] the veracity of Dr. King’s criticism” of the ILS test—not because his testimony in this case contradicted his criticism of the ILS test, but because his testimony in another case also contradicted his criticism of the ILS test. See Smith IV, 2021 WL 3666808, at *10 (observing that Dr. King has previously testified that ILS test “measures adaptive function- ing in a number of different domains” (quoting Tarver, 940 So. 2d at 324 (Cobb, J., concurring in part and dissenting in part))). Because we cannot disturb the district court’s finding that Dr. King’s criticism of the ILS test lacked credibility, see, e.g., Beren- guela-Alvarado v. Castanos, 950 F.3d 1352, 1357 (11th Cir. 2020), it fol- lows that the conclusion that Dr. Fabian drew from the ILS test— that Smith had “deficits in every area” of adaptive functioning— supports the district court’s conclusion about Smith’s adaptive def- icits. 8 The record also reveals that Smith struggled to communi- cate effectively, which supports the district court’s finding that

8 Alabama also criticizes the district court for failing to make “findings con- cerning Dr. Fabian’s reliance” on the ILS. But as we’ve explained, see supra n.6, our task is to determine whether the district court’s conclusion—that “Smith has significant deficits in social/interpersonal skills, self-direction, in- dependent home living, and functional academics,” Smith IV, 2021 WL 3666808, at *11—“is plausible in light of the record viewed in its entirety,” An- derson, 470 U.S. at 573–74 (emphasis added). USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 30 of 38

30 Opinion of the Court 21-14519

Smith has deficits in the “functional academics” realm. Smith IV, 2021 WL 3666808, at *11. Functional academics is a subcategory within the conceptual domain, which also includes communication skills. See DSM-5, at 37 (explaining that the conceptual domain in- volves “language, reading, writing, math reasoning,” and other ac- ademic skills); AAIDD-12, at 30 (listing difficulty communicating effectively as an example of significant deficits in the conceptual domain). Dr. Reschly, Dr. Chudy, and Dr. Fabian all testified that Smith’s illiteracy suggests that he suffers significant deficits in the conceptual domain. For his part, Dr. Reschly discussed Smith’s “failure to acquire literacy skills at an age-appropriate level, which relates to the conceptual demand of adaptive behavior.” Indeed, Dr. Chudy’s administered a WRAT-3, an achievement test used to gauge scholastic abilities, which revealed that “Smith is barely lit- erate in reading.” That test is “consistent with significant limita- tions in at least [the] conceptual domain,” according to Dr. Fabian. Dr. Fabian also evaluated Smith’s communication skills us- ing the Expressive One-Word Picture Vocabulary and the Receptive One-Word Picture Vocabulary tests. These tests relate “to func- tional academics or conceptual areas of adaptive functioning and even academic achievement,” said Dr. Fabian. Smith scored in the first percentile on the expressive test and in the third percentile on the receptive test. The age equivalents for those scores are thirteen and fifteen, respectively. Those scores, according to Dr. Fabian, “are consistent with someone who is intellectually disabled.” USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 31 of 38

21-14519 Opinion of the Court 31

Contending that Smith does not struggle with communica- tion skills, Alabama repeatedly describes Smith as “savvy” and says that he “had no problem understanding or appropriately respond- ing to questions” during the evidentiary hearing. But the record contradicts that description of Smith’s testimony. Take, for in- stance, an exchange between Smith and his attorney. During this exchange, Smith read a prompt that described a behavior. Smith was then asked to rate, on a scale from zero to three, whether he was able to perform that behavior and, if so, how often he per- formed that behavior without reminders and without help. A zero would convey that he was unable to perform that behavior while a three would convey that he always or almost always performed that behavior without reminders and without help: A: “Name 20 or more familiar objects.” Q: Would you give yourself a rating of zero, one, two or three? A: Yeah, I would. Q: Would you? A: Yeah, yeah, I would. Q: What would that rating be? A: Huh? Q: What rating would you give yourself for that? A: I don’t—I don’t know. I don’t understand the question. Why would I name 20 or more—oh, USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 32 of 38

32 Opinion of the Court 21-14519

it says familiar. I thought it said—“name 20 or more familiar objects.” One. Q: But you can name familiar objects to yourself; correct? A: Huh? Q: You can name familiar objects to yourself; cor- rect? A: I can. Q: Okay. Do you think you could name 20 things? A: Yeah. Q: So would the more correct response to that be a three? A: Yeah, if you ask—if I can, yeah. As that excerpt demonstrates, the record refutes Alabama’s claim that Smith “had no problem understanding or appropriately re- sponding to questions” during the evidentiary hearing. Indeed, that example adds to the mountain of evidence that suggests Smith struggles to communicate effectively and therefore suffers deficits in the conceptual domain of adaptive functioning. And because deficits in any one domain satisfy the adaptive-func- tioning criteria, see Moore, 581 U.S. at 15–16 (citation omitted); DSM-5 at 38; AAIDD-12, at 31, we cannot say that the district court did clearly erred by finding that Smith satisfied the adaptive-func- tioning prong. USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 33 of 38

21-14519 Opinion of the Court 33

Resisting that conclusion, Alabama advances three addi- tional arguments as to why the district court clearly erred by find- ing that Smith satisfied the adaptive-functioning prong. First, Ala- bama argues that the district court clearly erred by failing to make any findings concerning the ABAS-3, 9 a test that Dr. King adminis- tered to assess Smith’s adaptive functioning. Based on the results from that test and his interview with Smith, Dr. King concluded that Smith lacked “any serious problems with adaptive function- ing.” But contrary to Alabama’s claim, the district court addressed and discredited Dr. King’s adaptive-functioning findings because they relied “solely” on “Smith’s self-reports.” See Smith IV, 2021 WL 3666808, at *7–8. Unlike the other tests we’ve described, 10 the ABAS-3 relies on “an individual giving a report on himself.” And as the district court explained, Dr. King’s reliance on Smith’s self- reports made his findings unreliable for two reasons.

9 We just described the ABAS-3 test; it requires the subject to read a descrip- tion of a behavior and rate, on a scale of zero to three, whether the subject can perform that behavior and, if so, how often the subject performs that behavior without reminders and without help. Dr. King administered the ABAS-3 to Smith before the evidentiary hearing. 10 The ILS test, for example, requires Smith to show (rather than tell) his adap- tive abilities by requiring him to answer questions like what is the purpose of a will, what would you do if you had chest pains, how do you fix things in your home, and how do you use a map to get from point A to point B. The administering professional then assesses the test taker’s answers to evaluate his adaptive abilities. USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 34 of 38

34 Opinion of the Court 21-14519

First, the district court explained that the AAIDD “cautions against reliance on self-reporting.” Id. at *7. The AAIDD warns against “using self-report[ing] for the assessment of adaptive behav- ior” because self-reporting “may be susceptible to biased respond- ing.” AAIDD-12, at 40–41. To that end, Dr. Fabian testified that Smith “has not wanted to be found intellectually disabled.” In Dr. Fabian’s opinion, Smith is “embarrassed/offended by this.” Second, and relatedly, the district court explained that much of the information Smith reported to Dr. King was demonstrably untrue: For instance, Smith’s mother was 63 (not 69) when she died, and Smith’s father was 64 (not 70) when he died. Dr. King also acknowledged that Smith told him that he had not attended school beyond the sixth grade, but records show he did not leave school until he was in the eighth grade. Smith also reported to Dr. King that he was drinking on a daily basis from the age of 20 until age 27 when he was arrested. But Smith was actually incarcerated from age 19 to 26 and then again at 27. Smith IV, 2021 WL 3666808, at *7. We also note a third reason to doubt Dr. King’s reliance on the ABAS-3 test: Smith took that test twice and reported different answers each time, and as we’ve mentioned (see supra at 32–33), a review of Smith’s responses the second time he took the test USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 35 of 38

21-14519 Opinion of the Court 35

(during the evidentiary hearing) reveal that it’s not clear he under- stood what was being asked. As we’ve explained, the ABAS-3 test required Smith to rate, on a scale from zero to three, whether he was able to perform a particular behavior and, if so, how often he performed that behav- ior without reminders and without help. Smith first took the ABAS-3 test when he met with Dr. King before the evidentiary hearing. Then, Smith’s counsel administered the ABAS-3 test dur- ing the evidentiary hearing. During the second administration of the test, Smith reported different ratings than the ones he reported when Dr. King administered the test before the evidentiary hearing. When Dr. King administered the ABAS-3, for example, Smith gave himself a three for the following prompt: “Answers the telephone by saying ‘Hello.’” In other words, Smith reported that he always performs that behavior. But when he read that same prompt dur- ing the evidentiary hearing, Smith said, “I don’t answer no tele- phone.” Similarly, Smith gave himself a one at the evidentiary hear- ing in response to the prompt that reads: “Nods or smiles to en- courage others when they are talking.” But Smith gave himself a three in response to the same prompt when Dr. King administered the test before the evidentiary hearing. The court ultimately discredited Dr. King’s testimony con- cerning Smith’s adaptive deficits. See id. at *11 (explaining that “whether Smith has significant or substantial deficits in adaptive be- havior largely comes down to which expert is believed”). We USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 36 of 38

36 Opinion of the Court 21-14519

cannot say that the district court clearly erred in doing so given the problems with Dr. King’s testimony. For the same reason, we must reject Alabama’s second argu- ment as to why the district court clearly erred when finding that Smith satisfied the adaptive deficits prong. To support this argu- ment, Alabama contends that Dr. King “found that Smith had strengths in his home living and functional academics.” This argu- ment fails because, as we have observed, the district court discred- ited Dr. King’s testimony concerning Smith’s adaptive deficits. But even if the district court had credited Dr. King’s testimony, this piece of testimony does not help Alabama to show clear error, for “‘the medical community focuses the adaptive-functioning inquiry on adaptive deficits,’ not strengths.” Carroll, 300 So. 3d at 63 (quot- ing Moore, 581 U.S. at 16). Finally, Alabama claims that the district court improperly “discounted” Dr. King’s reliance on records from the Alabama De- partment of Corrections about Smith’s behavior in prison. Those records were “significant,” Alabama claims, “because there was no indication that Smith has a mental disability or psychiatric prob- lems, and because the records indicated that he functioned nor- mally.” But the Supreme Court has explained that “[c]linicians . . . caution against reliance on adaptive strengths ‘in a controlled set- ting,’ as a prison surely is.” Moore, 581 U.S. at 16; see also DSM-5, at 38 (“Adaptive functioning may be difficult to assess in a controlled USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 37 of 38

21-14519 Opinion of the Court 37

setting (e.g., prisons, detention centers)[.]”). So the prison records do not allow Alabama to show clear error. In sum, we cannot say that the district court clearly erred by finding that Smith satisfied the adaptive-functioning prong. We have already explained that the record contains evidence “that would support a finding of fact that Smith had significant limita- tions in at least two” domains. Smith III, 620 F. App’x at 750. Dr. King’s testimony is the only new evidence that has undermined that conclusion. But the district court discredited Dr. King’s testi- mony. As a result, the district court did not clearly err. VI. Finally, we turn to the district court’s finding that “Smith’s intellectual and adaptive functioning issues clearly arose before he was 18 years of age.” Smith IV, 2021 WL 3666808, at *12. While in school, Smith took two IQ tests and obtained scores of 74 and 75. As a result, the school recommended placing Smith in the “EMR program.” EMR at that time referred to “educable [intellectually disabled],” according to Dr. Reschly,11 who added that “the criteria for identifying someone with educable [intellectual disability] at

11 Alabama asks us to hold that the district court clearly erred by refusing to discredit Dr. Reschly’s testimony. On this view, “Dr. Reschly made his diag- nosis that Smith was intellectually disabled as a child . . . without personally evaluating him.” But because we cannot go back in time, it was impossible for Dr. Reschly (or anyone else, for that matter) to “personally evaluat[e]” whether Smith exhibited deficits in intellectual and adaptive functioning be- fore turning 18. USCA11 Case: 21-14519 Document: 46-1 Date Filed: 05/19/2023 Page: 38 of 38

38 Opinion of the Court 21-14519

that time was largely parallel to the criteria used to identify mild intellectual disability today.” Those criteria were an IQ score “be- low 75” and “documented deficits in adaptive behavior.” Dr. Fabian shares Dr. Reschly’s “understanding” that EMR is “pretty con- sistent with modern day intellectual disability mild.” In sum, then, the record supports the district court’s conclu- sion that Smith’s deficits in intellectual and adaptive functioning “were present at an early age.” Id. As a result, we cannot say that the district court clearly erred by finding that Smith satisfied the final prong of his Atkins claim. VII. We hold that the district court did not clearly err in finding that Smith is intellectually disabled and, as a result, that his sen- tence violates the Eighth Amendment. Accordingly, we affirm the district court’s judgment vacating Smith’s death sentence. AFFIRMED.