Case: 17-20689     Document: 00515100039        Page: 1    Date Filed: 08/30/2019




          IN THE UNITED STATES COURT OF APPEALS
                   FOR THE FIFTH CIRCUIT
                                                                       United States Court of Appeals
                                                                                Fifth Circuit

                                                                              FILED
                                     No. 17-20689                       August 30, 2019
                                                                         Lyle W. Cayce
UNITED STATES OF AMERICA,                                                     Clerk


              Plaintiff - Appellee

v.

SONNY FLOYD PERVIS; RAYNARD GRAY,

              Defendants - Appellants




                  Appeal from the United States District Court
                       for the Southern District of Texas


Before HIGGINSON and WILLETT, Circuit Judges, and BROWN, District
Judge.*
STEPHEN A. HIGGINSON, Circuit Judge:
      This appeal arises from the attempted armed robbery and then, two days
later, the actual armed robbery of a credit union in Pasadena, Texas in July
2014. Sonny Floyd Pervis and Raynard Gray were convicted of bank robbery
under 18 U.S.C. § 2113(a) and received additional sentences under 18 U.S.C. §
924(c) for carrying firearms during the offenses. Pervis and Gray both raise
two sentencing issues: first, whether bank robbery under § 2113(a) is a “crime



      *  Debra M. Brown, United States District Judge, Northern District of Mississippi,
sitting by designation.
    Case: 17-20689    Document: 00515100039     Page: 2   Date Filed: 08/30/2019



                                 No. 17-20689
of violence” for purposes of § 924(c); and second, whether the actual robbery
was permissibly deemed “second or subsequent” in relation to the attempt two
days earlier. Binding authority controls both issues, so we affirm Pervis’s and
Gray’s sentences.
      Gray also challenges the district court’s determination that he was
competent to stand trial under 18 U.S.C. § 4241, citing evidence of intellectual
disability. The district court’s careful deliberation yielded three expert
evaluations and a record posing interpretive difficulties. Deferring to the
district court’s reasonable assessment of this complex record, we affirm.


                                        I
      On July 26, 2014, a group of armed men attempted to enter the Shared
Resources Credit Union on Highway 225 in Pasadena, Texas, but its doors
were locked. Appellants Pervis and Gray were among this group.
      Two days later, on July 28, Pervis and Gray returned to the credit union.
Pervis, who rode in one vehicle with two associates, entered and robbed the
credit union at gunpoint. Gray rode in another vehicle that provided
surveillance. A third vehicle was for those carrying out the robbery to get away.
After Pervis and his two carmates robbed the credit union, they did not get
away cleanly, however. Patrol cars caught up to them as they left the area,
spurring a high-speed chase that resulted in several arrests but also the escape
of several other persons.
      Law enforcement tracked the defendants down in the ensuing months.
Most agreed to be interviewed. In their telling, Gray organized the robbery.
Defendant Keith McGee said that Gray recruited him, supplied him with the
gun that McGee used in the robbery, and coordinated the group’s efforts on
July 28. Defendant Leroy Richardson also said that Gray recruited him and


                                       2
    Case: 17-20689    Document: 00515100039     Page: 3   Date Filed: 08/30/2019



                                 No. 17-20689
instructed the group throughout the robbery. Defendant Christopher Braziel
likewise identified Gray as the planner of the robbery.
      Pervis and Gray were two of seven co-defendants charged in a two-count
indictment in February 2015. The first count charged them with the robbery
of the Shared Resources Credit Union on July 28, 2014, a violation of 18 U.S.C.
§ 2113(a). The second count charged them with carrying and brandishing
handguns during the robbery, a violation of 18 U.S.C. § 924(c)(1)(A). A
superseding indictment in May 2016 repeated the same charges but as to only
four defendants, reflecting the fact that three had chosen to cooperate. A
second superseding indictment filed in June 2016, the operative one for our
purposes, added new counts for the attempted robbery of the credit union on
July 26 and for the carrying and brandishing of a firearm in the process.
      On June 1, 2016, Gray’s counsel moved for an evaluation and hearing as
to Gray’s competency, arguing that Gray was unable to assist in his own
defense. The motion indicated that a psychologist, Dr. Diane L. Bailey, had
evaluated Gray a week earlier. An IQ test, the Stanford-Binet Intelligence
Scale–Fifth Edition, had yielded an IQ of 61, and a test of academic ability had
indicated that Gray read at a third-grade level.
      Three evaluations followed. The first was by Dr. Tennille Warren-
Phillips, a psychologist at the Federal Detention Center (FDC) in Houston, who
evaluated Gray from mid-June to mid-July, 2016. She concluded that Gray was
not competent to stand trial, finding that Gray had an IQ score of 62 +/- 5; poor
adaptive functioning since childhood, according to an interview with Gray’s
mother; and a doubtful ability to understand legal proceedings. Dr. Warren-
Phillips also found certain results difficult to interpret. For instance, on one
test, the Validity Indicator Profile (VIP), Gray gave answers following a
pattern––11221122––that could indicate either intentionally poor effort or low
intellectual functioning.
                                       3
     Case: 17-20689       Document: 00515100039          Page: 4     Date Filed: 08/30/2019



                                       No. 17-20689
       Faulting Dr. Warren-Phillips’s analysis, the Government requested
another evaluation. The Government contended that she interviewed too few
people, that she failed to consider Gray’s long history of prosecutions without
a finding of incompetency, and that Gray’s adaptive functioning was not as
poor as supposed. The district court granted the Government’s motion, and
Gray was transferred to the Federal Correctional Institution (FCI) in Fort
Worth for evaluation by Dr. Lisa Bellah, a psychologist there.
       Dr. Bellah produced the report on which the district court would later
ground its competency determination. She based her report on interviews with
correctional staff, the Assistant U.S. Attorney on the case, and Dr. Warren-
Phillips; on tests specifically aimed at identifying malingering; 1 and on emails
and phone calls by Gray while at FCI–Fort Worth. Dr. Bellah picked up where
Dr. Warren-Phillips’s suspicions left off. She ran the VIP again, as well as the
Test of Memory Malingering, which confirmed Gray was malingering. A test of
legal understanding, the Inventory of Legal Knowledge, indicated the same. 2
Due to Gray’s performance in objective and structured assessments, Dr. Bellah
looked to Gray’s behavior while not under clinical observation. She spoke with
Kenneth Goldsby, a counselor in Gray’s unit at FCI–Fort Worth, 3 who told Dr.
Bellah that Gray “consistently socialized,” kept score while playing dominoes,
understood unit rules, and regularly used the phone, email, and commissary.
Gray’s calls and emails, in turn, showed he understood the competency



       1 To malinger is “[t]o pretend or exaggerate illness in order to escape duty or work; to
feign or produce physical or psychological symptoms to obtain financial compensation or
other reward.” Malinger, OXFORD ENGLISH DICTIONARY (3rd ed. 2000); see also Malingering,
OXFORD DICTIONARY OF PSYCHOLOGY (4th ed. 2015) (“Intentional feigning or exaggeration of
physical or psychological symptoms, motivated by external incentives such as . . . evasion of
criminal prosecution,” among others).
       2 Gray’s score on this test score strongly resembled “a normative group of individuals

who were instructed to purposely malinger deficits in legal knowledge.”
       3 Goldsby’s educational and professional background is not clear from the record.

                                              4
     Case: 17-20689      Document: 00515100039        Page: 5     Date Filed: 08/30/2019



                                     No. 17-20689
evaluation process, as well as basic points about criminal proceedings. In one
conversation, Gray did the arithmetic necessary to find the cost of bond set at
$40,000 or $50,000. He also seemed to understand the possible sentences he
faced, the concept of a sentencing range, the likely decisions of his co-
defendants, the charges against him, and the choice between pleading and
going to trial. Based largely on this evidence, Dr. Bellah concluded that, though
Gray exhibited “borderline” intellectual functioning, he did not exhibit a
mental defect rendering him incompetent to be tried.
      Following Dr. Bellah’s report, Gray’s counsel requested and the district
court granted another evaluation, this time by Dr. Michael Chafetz, a
neuropsychologist who had published scholarly works on malingering. 4 Like
Dr. Bellah, Dr. Chafetz detected malingering. He reached this conclusion after
conducting numerous tests, some solely for the purpose of identifying
malingering and some serving other purposes but containing “embedded
indicators” to discern intentionally-poor effort. Gray’s malingering invalidated
the IQ score of 51 yielded by Dr. Chafetz’s administration of the Wechsler Adult
Intelligence Scale. To counteract this problem, Dr. Chafetz took two steps.
First, he conducted a regression analysis of Gray’s IQ score based on the
“known relationship” between “poor effort/invalidity” and “increasingly lower
IQ scores.” This adjustment changed Gray’s IQ score from 51 to a range of 65.4
to 68.1––higher but “still in the range of Intellectual Disability.” Second, Dr.
Chafetz used the Test of Premorbid Functioning, which produces an IQ
estimate “free from any complications due to poor quality of effort” by
incorporating the test-taker’s demographic profile. This adjustment yielded an



      4 See, e.g., Michael D. Chafetz & Alex Biondolillo, Validity Issues in Atkins Death
Cases, 26 THE CLINICAL NEUROPSYCHOLOGIST 1358 (2012); Michael D. Chafetz et al.,
Malingering on the Social Security Disability Consultative Examination: A New Rating Scale,
22 ARCHIVES OF CLINICAL NEUROPSYCH. 1 (2007).
                                            5
    Case: 17-20689    Document: 00515100039    Page: 6   Date Filed: 08/30/2019



                                No. 17-20689
IQ score of 73 +/- 5. Dr. Chafetz’s assessment of Gray’s adaptive functioning
yielded low results; so low that Dr. Chafetz doubted their accuracy. But his
analysis of Gray’s legal understanding left him more skeptical of it than Dr.
Bellah, so he concluded that it was “unlikely” that Gray was competent to
stand trial. He recognized Gray’s malingering but reasoned that “it was not
possible to jump to the conclusion that without malingering, his scores would
be high enough to rule out a mental defect.”
      On March 1, 2017, the district court held a four-hour evidentiary
hearing, taking testimony from Dr. Bellah and Dr. Chafetz. The hearing
extensively covered the experts’ various points of convergence and divergence.
One novelty was Dr. Bellah’s conclusion that Gray was able to understand
legal proceedings and assist in his own defense. In her report, she had
concluded only that Gray did not have a mental defect. At the hearing, Dr.
Bellah drew on Gray’s recorded phone calls and other evidence to conclude that
he was able to understand the proceedings and assist his counsel.
      In a reasoned order issued March 7, 2017, the district court ruled that
Gray was competent to stand trial. The court began with the evidence of Gray’s
continual malingering and with Dr. Chafetz’s and Dr. Bellah’s “corrective
approaches.” The court favored Dr. Bellah’s “more practical than theoretical
response” to the problem of malingering––that is, her reliance on observational
evidence, rather than Dr. Chafetz’s use of adjusted IQ scores. The court
explained that Gray’s recorded phone calls demonstrated both his adaptive
functioning, in that he displayed coherent thought and could recall
information, and his legal understanding, in that he understood the
proceedings in his case.
      Two weeks later, Gray and Pervis stood trial and were convicted on all
counts. Though the jury convicted them on the gun counts, it found that they
had not “brandished” or “discharged” a firearm during the robbery on July 28.
                                      6
     Case: 17-20689      Document: 00515100039        Page: 7     Date Filed: 08/30/2019



                                     No. 17-20689
       The district court sentenced Pervis to 485 months. The court chose 125-
month concurrent sentences for both robbery offenses. The court imposed a
consecutive 60-month sentence for Pervis’s gun offense associated with the
July 26 attempt, the minimum sentence permitted by 18 U.S.C. §
924(c)(1)(A)(i). The court imposed 300 months consecutively for the gun offense
associated with the July 28 robbery, the minimum sentence for “a second or
subsequent conviction” under 18 U.S.C. § 924(c)(1)(C) (2017). 5
       The district court sentenced Gray to 510 months. The court chose 150-
month concurrent sentences for the two robbery offenses. 6 It then imposed the
same 60-month and 300-month consecutive sentences for the gun offenses that
it imposed on Pervis. The 25-month discrepancy between Gray’s and Pervis’s
robbery sentences reflected in part that Gray received a four-level
enhancement for organizing an offense involving five or more participants. 7
       Pervis and Gray, through separate counsel, timely appealed.


                                            II
                                            A
      Pervis’s and Gray’s gun convictions under 18 U.S.C. § 924(c)(1)(A)
depend on the determination that their robbery offenses constitute crimes of
violence. The statute provides that “any person who, during and in relation to
any crime of violence . . . uses or carries a firearm, or who, in furtherance of
any such crime, possesses a firearm, shall, in addition to the punishment




       5 The First Step Act of 2018, which went into effect in December 2018, changed the
language by which an offense is deemed “second or subsequent.” First Step Act of 2018, Pub.
L. No. 115-391, § 403, 132 Stat. 5194, 5221–22 (effective Dec. 21, 2018).
       6 This was well below the guideline range of 235 to 293 months and the statutory

maximum of 25 years.
       7 See U.S. SENTENCING GUIDELINES MANUAL § 3B1.1(a) (2016).

                                            7
     Case: 17-20689        Document: 00515100039           Page: 8     Date Filed: 08/30/2019



                                        No. 17-20689
provided for such crime,” serve a sentence of at least five years. 8 The statute
carries a hefty penalty for recidivists. At the time of Pervis’s and Gray’s
sentencing, it provided that, “[i]n the case of a second or subsequent conviction
under this subsection,” the defendant shall serve a sentence of at least twenty-
five years. 

(2017). 9
       The statutory definition of “crime of violence” employs an elements
clause as well as a residual clause. Under the former, a “crime of violence” is
“an offense that is a felony and has as an element the use, attempted use, or
threatened use of physical force against the person or property of another.” 18
U.S.C. § 924(c)(3)(A). Under the latter, a “crime of violence” is “an offense that
is a felony and that by its nature, involves a substantial risk that physical force
against the person or property of another may be used in the course of
committing the offense.” 

       In briefing and at oral argument, Pervis relied on Sessions v. Dimaya,
which declared unconstitutionally vague the residual clause of 18 U.S.C. § 16. 10

, 1213–16 (2018). The residual clauses of § 16 and of § 924(c)
contain “virtually identical” definitions. United States v. Davis, 

,
2329 (2019). In June 2019, after oral argument in this case, the Supreme Court
followed Dimaya and declared § 924(c)’s residual clause unconstitutionally
vague as well. 

       Pervis’s attention is concentrated on the residual clause, but § 924(c)’s
elements clause is also a basis for classifying robbery under § 2113(a) as a




       8   If the gun is brandished, the minimum is seven years. 18 U.S.C. § 924(c)(1)(A)(ii). If
it is discharged, the minimum is ten. 

Hence the special jury questions
on brandishing and discharging.
         9 As noted above, the First Step Act revised this language, but the change did not go

into effect until December 2018.
         10 18 U.S.C. § 16 supplies a general-purpose definition of “crime of violence” referenced

in a wide variety of federal criminal and immigration statutes.
                                                8
    Case: 17-20689         Document: 00515100039         Page: 9   Date Filed: 08/30/2019



                                       No. 17-20689
crime of violence. In United States v. Brewer, our court considered whether
robbery under § 2113(a) is a crime of violence for purposes of the Sentencing
Guidelines. 

, 713–16 (5th Cir. 2017). Section 4B1.2(a) contained
an elements clause worded identically to the clause at issue here. 

Under § 2113(a), robbery can be committed “by force and violence, or by
intimidation.” Only intimidation possibly allowed a way out of the elements
clause’s scope, but in light of our bank robbery caselaw, we concluded that
“intimidation in the bank-robbery context is inherently tied to a threatened
use of force.” 

(citing United States v. Higdon, 

, 315 (5th
Cir. 1987)). Consequently, we held that robbery under § 2113(a) constitutes a
crime of violence.
      Following Brewer, we have applied its holding to § 924(c)(3)(A) in
numerous unpublished decisions. See United States v. Cheers, 760 F. App’x
272, 274 (5th Cir. 2019); United States v. Lance, 742 F. App’x 871, 872 (5th Cir.
2018); United States v. Cadena, 728 F. App’x 381, 381–82 (5th Cir. 2018);
United States v. Ward, 698 F. App’x 178, 178 (5th Cir. 2018); United States v.
Cosner, 690 F. App’x 292, 293 (5th Cir. 2017). Pervis offers no argument that
§ 924(c)(3)(A) should be read differently than § 4B1.2(a), the guideline
construed in Brewer, and no such argument occurs to us. 11 Accordingly, we do
as we have done before and conclude that § 2113(a) is a crime of violence under
18 U.S.C. § 924(c)(3)(A).
                                              B
      Gray argues that his conviction under § 924(c) for the July 28 robbery
cannot be a “second or subsequent conviction” in relation to the July 26 attempt
because this proceeding was the first time he was ever charged with violating
§ 924(c). As Gray recognizes, his argument is foreclosed by Deal v. United


      11   Gray concedes that Brewer forecloses the issue.
                                              9
    Case: 17-20689       Document: 00515100039         Page: 10     Date Filed: 08/30/2019



                                      No. 17-20689
States, 

(1993), in which the Supreme Court construed the statute
to permit a “second or subsequent conviction” to be charged in the same
indictment and adjudicated in the same proceeding as the first instance of the
offense. The Court interpreted “conviction” to mean the finding of guilt, rather
than the entry of a final judgment of conviction, including determination of the
sentence. Compare 

with 

(Stevens, J., dissenting). 12
Consequently, Deal defeats Gray’s argument.
                                             C
       Pervis challenges the 25-year sentence for his second gun conviction on
a different ground, arguing that the district court impermissibly made the
“second or subsequent” determination itself, rather than putting it to the jury
as ostensibly required by Alleyne v. United States, 

(2013). Pervis
acknowledges that plain-error review would apply here because he did not
object contemporaneously. The nuances of plain-error review do not matter,
however, because no error occurred. Pervis overlooks Almendarez-Torres v.
United States, 

(1998), which controls this issue.
       Almendarez-Torres concerned the illegal reentry statute, 8 U.S.C. § 1326,
which generally caps sentences at two years but raises the maximum to twenty
years for defendants previously deported “subsequent to a conviction for
commission of an aggravated felony.” 

(quoting 8 U.S.C. §
1326(b)(2) (1996)). The Supreme Court ruled that the fact of a prior
aggravated-felony conviction is a “sentencing factor” that a court, rather than
a jury, can determine. 

239–46. Two years later, the Court preserved
Almendarez-Torres when it decided Apprendi v. New Jersey, 

       12 The First Step Act’s revised language codifies the dissenting view in Deal:
permitting imposition of the enhanced sentence only in a subsequent proceeding. See 18
U.S.C. § 924(c)(1)(C) (imposing the 25-year minimum sentence “[i]n the case of a violation of
this subsection that occurs after a prior conviction under this subsection has become final”)
(emphasis added); see 

n.1 (construing the revised language).
                                             10
    Case: 17-20689     Document: 00515100039      Page: 11    Date Filed: 08/30/2019



                                   No. 17-20689
(2000). The Court ruled that “any fact that increases the penalty for a crime
beyond the prescribed statutory maximum must be submitted to a jury, and
proved beyond a reasonable doubt,” with the exception of “the fact of a prior
conviction.” 

The Court preserved Almendarez-Torres again when it
decided Alleyne. 

n.1 (“Because the parties do not contest
that decision’s vitality, we do not revisit it for purposes of our decision today.”).
It remains the case that the district court can resolve the question of a prior
conviction, and so we reject Pervis’s argument.


                                         III
      Competency to stand trial in federal court is a matter of both statutory
application and constitutional significance. In federal court, the defendant and
the Government have the statutory right to move for a hearing to determine a
defendant’s mental competency. 18 U.S.C. § 4241(a). “If, after the hearing, the
court finds by a preponderance of the evidence that the defendant is presently
suffering from a mental disease or defect rendering him mentally incompetent
to the extent that he is unable to understand the nature and consequences of
the proceedings against him or to assist properly in his defense,” the court shall
commit the defendant to federal custody for further examination, treatment,
and, if possible, restoration of his competency. 

The Government
bears the burden of proving that the defendant is competent. United States v.
Hutson, 

, 1018 (5th Cir. 1987). “[T]he criminal trial of an
incompetent defendant violates due process.” Cooper v. Okla., 

,
354 (1996) (quotation omitted).
       In evaluating competency, the district court may consider various
sources of evidence, “including, but not limited to, its own observations of the
defendant’s demeanor and behavior; medical testimony; and the observations
of other individuals that have interacted with the defendant.” United States v.
                                         11
    Case: 17-20689       Document: 00515100039           Page: 12    Date Filed: 08/30/2019



                                       No. 17-20689
Porter, 

, 380 (5th Cir. 2018) (quoting United States v. Simpson,

, 306 (5th Cir. 2011)). Our court reviews a district court’s
competency determination “using a ‘species of clear error’ review.” 

(quoting 

). Our “task is . . . to take a ‘hard
look’ at the facts to determine whether the district court’s competency finding
was ‘clearly arbitrary or unwarranted.’” 

(quoting
United States v. Joseph, 

, 589 (5th Cir. 2003)). Though we are to
take a hard look at the record, “[i]t is not our task, as an appellate court, to
relitigate the battle of the experts.” 

defects or diseases may render a defendant
incompetent to stand trial. E.g., United States v. McKown, --- F.3d ----, 

, at *1 (5th Cir. July 22, 2019) (“grandiose and persecutory
delusional disorder”); United States v. Brennan, 

, 212 (2nd Cir.
2019) (severe alcoholism); United States v. Arenburg, 

, 166 (2nd
Cir. 2010) (“paranoid schizophrenia”); United States v. Filippi, 

,
650 (1st Cir. 2000) (“vascular dementia”); United States v. Shawar, 

, 858 (7th Cir. 1989) (“mental retardation”).
       Of intellectual disability, the Supreme Court has said––in a related but
distinct     context––that     it   includes      “not   only   subaverage       intellectual
functioning, but also significant limitations in adaptive skills such as
communication, self-care, and self-direction that became manifest before age
18.” Atkins v. Virginia, 

, 318 (2002). 13 The analysis of each expert




       13 The Supreme Court’s Eighth Amendment caselaw on competency to be executed
includes general discussion of the nature and assessment of intellectual disability. See also
Moore v. Texas, 

(2017); Brumfield v. Cain, 

(2015); Hall v.
Florida, 

(2014). But it is important not to confuse or conflate this jurisprudence
with caselaw on the competency to stand trial. The latter is structured by a statutory
standard, the former by constitutional doctrine. The latter is tied to the defendant’s ability
to participate in legal proceedings, while the former reflects multifarious concerns regarding
                                             12
    Case: 17-20689        Document: 00515100039          Page: 13     Date Filed: 08/30/2019



                                       No. 17-20689
who evaluated Gray followed this distinction between intellectual and adaptive
functioning. Each also separately examined Gray’s legal knowledge and
understanding, the touchstone of the statutory standard.
       The experts’ objective assessments of Gray’s intellectual functioning
would be substantial and compelling if the test results were trustworthy. Dr.
Warren-Phillips found that Gray’s “overall intellectual ability” was “extremely
low,” based principally on the Wechsler Adult Intelligence Scale–Fourth
Edition, which yielded a full scale IQ of 62 +/- 5. Subtests indicated “extremely
low” verbal abilities, “low average” nonverbal abilities, low ability to sustain
attention, and borderline mental “processing speed.” 14 Dr. Chafetz found a
score of 51, which he did not consider accurate, and then produced adjusted
scores that placed Gray in the range of borderline to low intellectual
functioning: 65.4 to 68.1, using a “symptom validity scale” to account for Gray’s
effort to lower his scores; and 73 +/- 5, using the Test of Premorbid Functioning
to account for his demographic profile.
       This spread of scores illustrates that IQ scores must be “read not as a
fixed number but as a range,” owing to “the inherent imprecision” of IQ testing.
See Hall v. Florida, 

, 712 (2014) (describing the consensus of
“professionals who design, administer, and interpret IQ tests” on this point).
As the Supreme Court explained in Hall, “an individual’s IQ test score on any
given exam may fluctuate for a variety of reasons,” including “the test-taker’s
health; practice from earlier tests; the environment or location of the test; the
examiner’s demeanor; the subjective judgment involved in scoring certain
questions on the exam; and simple lucky guessing.” 

Accordingly, we look only in a limited and careful manner to the
Court’s general discussion of intellectual disability in these cases.
        14 The IQ score that Dr. Warren-Phillips found comports with the IQ score found by

Dr. Bailey, 61, which triggered the district court’s competency review. But we have no expert
report from Dr. Bailey in the record, and Gray does not urge the score she found as significant.
                                              13
    Case: 17-20689        Document: 00515100039          Page: 14      Date Filed: 08/30/2019



                                        No. 17-20689
       Salient here is another factor: Gray’s intentionally poor effort. Gray’s
first evaluator suspected malingering, and the second and third evaluators
confirmed it. The results of the VIP administered by Dr. Warren-Phillips
suggested that Gray might have been intentionally performing poorly. To
corroborate this suspicion, Dr. Bellah used the VIP as well, instructing Gray
on the need for good effort. Rather than oblige, Gray completed the nonverbal
portion, which normally takes roughly forty-five minutes, in fewer than five.
In the process, he missed some of the “easiest . . . items on the test,” an unlikely
result for anyone making a serious effort. Dr. Bellah then used the Test of
Memory Malingering, a memory test that shows the test-taker fifty simple
pictures of everyday objects. The test-taker is then shown the same pictures
one-by-one alongside “distractor” pictures and has to choose which of the two
he has seen before. This is done three times. The simplicity and repetition of
the images are such that virtually anyone, even those with severe mental
impairment, will do reasonably well and will do better in successive trials.
Malingering can therefore be assessed with reference to a cutoff score––below
which no earnest test-taker would fall––and to the trend of the test-taker’s
scores. Dr. Bellah found that Gray’s results on the test fell below the cutoff
score and trended downward over the three trials––strong indications of
malingering. 15
       During Dr. Chafetz’s evaluation, Gray repeatedly engaged in the same
behavior, identified by indicators embedded in the tests that Dr. Chafetz used.
This left Dr. Chafetz reliant on necessarily more speculative adjustments to



       15 Random selection would yield a score around 25: half of the 50 pictures correctly
identified. Gray’s scores were 21 in the first trial, 26 in the second, and only 18 in the third.
Dr. Bellah supplied three comparison scores: an individual with “severe learning disability”
and other impairments who scored 48-50-50; an individual with a “closed head injury from
motor vehicle accident” who scored 38-48-50; and an individual with Huntington’s disease
and borderline IQ, who scored 42-50-50.
                                              14
    Case: 17-20689        Document: 00515100039           Page: 15      Date Filed: 08/30/2019



                                        No. 17-20689
Gray’s IQ score, exacerbating the imprecision inherent in IQ testing. One of
these adjustments overlapped only partially with the confidence interval of Dr.
Warren-Phillips’s score; the other did not overlap at all. “An IQ score is an
approximation, not a final and infallible assessment of intellectual
functioning.” 

. The district court’s “studied skepticism” of
Gray’s scores, 

was therefore both understandable and appropriate. 16
       To gauge Gray’s adaptive functioning, Dr. Warren-Phillips conducted a
structured interview of Gray’s mother, who reported that Gray had a difficult
childhood: a five-month coma at age one due to an adverse reaction to medicine;
developmental delays; academic difficulty; a “significant head injury” at some
point; school dropout in his mid-teens; and limited ability to care for himself.
As with assessments of Gray’s intellectual functioning, however, Dr. Warren-
Phillips found some of these claims “questionable,” because Gray presented
with fewer limitations than his mother asserted. Dr. Warren-Phillips also used
the Minnesota Multiphasic Personality Inventory, which includes validity
scales to suss out inaccurate or inconsistent responses. Though the test showed
Gray had limitations, its validity scales also indicated that Gray’s responses
exhibited “over-reporting of a broad range of psychological, cognitive, and
somatic symptoms.” Dr. Warren-Phillips thus regarded the results with
“caution.”
       Given Gray’s behavior in objective assessments, Dr. Bellah looked to
evidence of Gray’s behavior while not under observation. As noted, this
included the input of the counselor on Gray’s unit, Kenneth Goldsby, who saw
Gray socializing, playing games, following unit rules, and making use of
resources in the unit. Dr. Bellah also inferred from Gray’s behavior, phone



       16 We certainly do not rule out IQ testing as significant. Indeed, “[i]t is of considerable
significance.” 

. We simply acknowledge its limitations.
                                               15
    Case: 17-20689       Document: 00515100039        Page: 16     Date Filed: 08/30/2019



                                      No. 17-20689
calls, and email that Gray and his mother had exaggerated his adaptive
deficits to Dr. Warren-Phillips. 17
       Dr. Chafetz used the Vineland II–Adaptive Behavior Scales to assess
Gray’s adaptive functioning. The Vineland scales depended on Gray’s answers
to interview questions, and Dr. Chafetz obtained an improbably low score for
Gray. This led him to suspect “negative bias or disengagement” in Gray’s
answers and to put more stock in prior assessments of Gray’s adaptive
functioning than in his own.
       The district court found the observations and recordings of Gray’s
behavior from FCI–Fort Worth to be most reliable. We note that, in a decision
issued after the district court’s competency determination, the Supreme Court
cast doubt on the informational value of behavior observed in correctional
settings and on analysis emphasizing adaptive strengths to the exclusion of
adaptive deficits. See Moore v. Texas, 

, 1050 (2017). We also
acknowledge that reasonable people may judge the significance of observed
behavior––like playing dominoes––differently. But in this instance, Gray’s
intentionally poor effort and the dubiousness of other available information
left the district court little better material with which to evaluate Gray’s
adaptive functioning.
       To assess Gray’s ability to understand and assist in legal proceedings,
Dr. Warren-Phillips used the Competence Assessment for Standing Trial for
Defendants with Mental Retardation (CAST-MR). It is a three-part test: “basic
legal concepts”; “skills to assist in defense”; and “understanding of case events.”
On the first, Gray’s score was average for those deemed incompetent to stand



       17For instance, Dr. Warren-Phillips had learned from the defense’s investigator that
Gray did not have a driver’s license, which in isolation might imply he does not know how to
operate a car. Gray’s statements in recorded phone calls, however, indicated he does know
how. Gray also told Dr. Chafetz he knows how to drive.
                                            16
   Case: 17-20689    Document: 00515100039      Page: 17   Date Filed: 08/30/2019



                                 No. 17-20689
trial; on the second, “markedly improved,” but still lower than average for
those deemed competent. On the third, Gray turned in “his best performance,”
achieving a result stronger than the average for all groups. Overall, the CAST-
MR put him below average for those deemed competent and above average for
those deemed incompetent. Dr. Warren-Phillips also explored Gray’s legal
understanding in interviews, finding that he displayed “limited” or “very
limited” understanding of the legal process.
      Dr. Bellah employed the Inventory of Legal Knowledge, and Gray’s
malingering again complicated the inquiry. Gray’s score differed markedly
from scores of both those found competent and those found incompetent to
stand trial. Instead, his score strongly resembled “a normative group of
individuals who were instructed to purposely malinger deficits in legal
knowledge.” Dr. Bellah also used the Evaluation of Competence to Stand Trial,
a “semi-structured clinical interview” examining legal understanding and
rational ability. Dr. Bellah found that Gray reported not understanding the
charges against him or the legal process generally. But she also observed that
he “often simply refused to answer” certain questions and that he “appeared to
disengage completely from the interview.” Gray’s recorded phone calls
provided cause to doubt the lack of understanding he manifested in the
interview. One conversation, for instance, showed he understood that a
superseding indictment had added to the charges against him. He also
understood the strategic choices made by his co-defendants to cooperate with
the Government.
      Dr. Chafetz used the same test as Dr. Warren-Phillips, the CAST-MR.
He found that Gray made a genuine effort on at least the first two parts of this
test. As when Dr. Warren-Phillips administered the CAST-MR, Gray’s score
on the “basic legal concepts” portion was akin to incompetent test-takers, while
his “skills to assist defense” score was above the incompetent average but
                                      17
    Case: 17-20689          Document: 00515100039         Page: 18     Date Filed: 08/30/2019



                                         No. 17-20689
below the competent average. Notably, Dr. Chafetz did not score the third part,
“understanding of case events,” on which Gray had previously gotten his best
result. When Gray reached this portion of the test, he began giving “shutdown
answers,” a behavior attributable to fatigue but also to malingering.
       The Government played excerpts of Gray’s recorded calls at the
competency hearing, 18 and the district court relied on them in finding Gray
competent. Gray puts forward different interpretations of the calls’ contents.
For instance, Gray makes the point––not without force––that the recordings
showed Gray did not fully understand the proceedings against him. 19 But, as
with Gray’s intellectual and adaptive functioning, his malingering during
objective assessments left this observational evidence as the best available,
and it is susceptible to different interpretations. The district court’s
interpretation was reasonable, and we decline to set it aside.
                                               ***
       The assessment of intellectual disability is a complicated task––one for
which the typical judge, no more than an educated layperson in these matters,
is perhaps not best equipped. But it is one that the law assigns us. The district
court addressed its assigned task with care and commitment. Its inquiry
revealed that even experts faced challenges in assessing Gray, making findings
and reaching conclusions not wholly in harmony with one another. Their
compiled findings comprised a record riven with complication and open to
inference. Gray’s continual malingering, though not itself dispositive evidence



       18   The district court therefore had the benefit of listening to these calls, whereas we
have not.
       19  Gray knew that his co-defendants were getting a deal in exchange for their
cooperation, and his counsel had talked to him about taking a plea as well, with 30 years as
the likely sentence. The recording captured Gray arguing to an unidentified interlocutor that
he had to go to trial to avoid that result. But Gray ultimately received a sentence of forty-two
and a half years and was exposed to even more. In Gray’s view, articulated through counsel,
this showed he did not understand the risks of trial.
                                               18
   Case: 17-20689    Document: 00515100039      Page: 19   Date Filed: 08/30/2019



                                 No. 17-20689
of competency, made it all the more difficult for those experts and the court to
develop such evidence.
      Consequently, after the hard look at this record that our precedent
requires, we cannot say the district court’s conclusion was clearly arbitrary or
unwarranted. We thus affirm the district court’s determination that Gray was
competent to stand trial.


                                      IV
      For the foregoing reasons, the judgment of the district court is
AFFIRMED.




                                      19