In the
United States Court of Appeals
For the Seventh Circuit
____________________
No. 16‐3397
BRENDAN DASSEY,
Petitioner‐Appellee,
v.
MICHAEL A. DITTMANN,
Respondent‐Appellant.
____________________
Appeal from the United States District Court for the
Eastern District of Wisconsin.
No. 14‐cv‐1310 — William E. Duffin, Magistrate Judge.
____________________
ARGUED SEPTEMBER 26, 2017 — DECIDED DECEMBER 8, 2017
____________________
Before WOOD, Chief Judge, and EASTERBROOK, KANNE,
ROVNER, WILLIAMS, SYKES, and HAMILTON, Circuit Judges.*
HAMILTON, Circuit Judge. Petitioner Brendan Dassey con‐
fessed on videotape to participating in the 2005 rape and mur‐
der of Teresa Halbach and the mutilation of her corpse. The
Wisconsin state courts upheld Dassey’s convictions for these
crimes, finding that his confession was voluntary and could
* Circuit Judges Flaum and Barrett did not participate in the consider‐
ation or decision of this case.
2 No. 16‐3397
be used against him. The principal issue in this habeas corpus
appeal is whether that finding was based on an unreasonable
application of Supreme Court precedent or an unreasonable
view of the facts. See 28 U.S.C. § 2254(d).
Whether Dassey’s confession was voluntary or not is
measured against a general standard that takes into account
the totality of the circumstances. See Withrow v. Williams, 507
U.S. 680, 693–94 (1993); Gallegos v. Colorado, 370 U.S. 49, 55
(1962); see also Fare v. Michael C., 442 U.S. 707, 727 (1979) (ad‐
missibility of juvenile confession). Some factors would tend to
support a finding that Dassey’s confession was not voluntary:
his youth, his limited intellectual ability, some suggestions by
the interrogators, their broad assurances to a vulnerable sus‐
pect that honesty would produce leniency, and inconsisten‐
cies in Dassey’s confession. Many other factors, however,
point toward a finding that it was voluntary. Dassey spoke
with the interrogators freely, after receiving and understand‐
ing Miranda warnings, and with his mother’s consent. The in‐
terrogation took place in a comfortable setting, without any
physical coercion or intimidation, without even raised voices,
and over a relatively brief time. Dassey provided many of the
most damning details himself in response to open‐ended
questions. On a number of occasions he resisted the interro‐
gators’ strong suggestions on particular details. Also, the in‐
vestigators made no specific promises of leniency.
After the state courts found the confession voluntary, a
federal district court and a divided panel of this court found
that the state courts’ decision was unreasonable and that Das‐
sey was entitled to a writ of habeas corpus. We granted en banc
review to consider the application of the deferential standards
No. 16‐3397 3
of 28 U.S.C. § 2254(d) and the implications of the panel deci‐
sion for interrogations of juvenile suspects. The state courts’
finding that Dassey’s confession was voluntary was not be‐
yond fair debate, but we conclude it was reasonable. We re‐
verse the grant of Dassey’s petition for a writ of habeas cor‐
pus.
Part I provides an overview of the applicable law. Part II
sets forth the relevant facts about Teresa Halbach’s murder,
Dassey’s confession, and the court proceedings. Part III ap‐
plies the law to the relevant facts, keeping in mind the defer‐
ence we must give under § 2254(d) to state court decisions as
to which reasonable judges might differ.
I. The Applicable Law
We first discuss our standard of review under the Antiter‐
rorism and Effective Death Penalty Act of 1996 (AEDPA) and
then describe the Supreme Court’s clearly established law for
when a confession, particularly a confession by a sixteen‐
year‐old like Dassey, is deemed voluntary and admissible.
A. Deference Under AEDPA
In considering habeas corpus petitions challenging state
court convictions, “our review is governed (and greatly lim‐
ited) by” AEDPA. Hicks v. Hepp, 871 F.3d 513, 524 (7th Cir.
2017) (citation omitted). The standards in 28 U.S.C. § 2254(d)
were designed to “prevent federal habeas ‘retrials’ and to en‐
sure that state‐court convictions are given effect to the extent
possible under law.” Id., quoting Bell v. Cone, 535 U.S. 685, 693
(2002). Section 2254(d) provides that a state court conviction
cannot be overturned unless the state courts’ adjudication of
a federal claim on the merits:
4 No. 16‐3397
(1) resulted in a decision that was contrary to, or in‐
volved an unreasonable application of, clearly es‐
tablished Federal law, as determined by the Su‐
preme Court of the United States; or
(2) resulted in a decision that was based on an un‐
reasonable determination of the facts in light of the
evidence presented in the State court proceeding.
The decision federal courts look to is the “last reasoned
state‐court decision” to decide the merits of the case, even if
the state’s supreme court then denied discretionary review.
Johnson v. Williams, 133 S. Ct. 1088, 1094 n.1 (2013). In this case,
we look to the Wisconsin Court of Appeals decision that Das‐
sey’s confession was voluntary. 1
The standard for legal errors under § 2254(d)(1) was
meant to be difficult to satisfy. Harrington v. Richter, 562 U.S.
86, 102 (2011). The issue is not whether federal judges agree
with the state court decision or even whether the state court
decision was correct. The issue is whether the decision was
unreasonably wrong under an objective standard. Williams v.
Taylor, 529 U.S. 362, 410–11 (2000) (majority opinion of O’Con‐
1 On October 30, 2017, the Supreme Court heard argument in Wilson
v. Sellers, No. 16‐6855, where one question is whether federal courts in ha‐
beas cases should continue to “look through” state supreme court sum‐
mary decisions on the merits to the last state court decision that provided
an explanation. See generally Hittson v. Chatman, 135 S. Ct. 2126, 2127
(2015) (Ginsburg, J., concurring in denial of certiorari). If the Court holds
in Wilson that federal courts reviewing a state supreme court summary
denial of review should give the state courts the benefit of any merits ra‐
tionale the record could support, our review would become even more
deferential, so the outcome here would not change.
No. 16‐3397 5
nor, J.). Put another way, we ask whether the state court deci‐
sion “was so lacking in justification that there was an error
well understood and comprehended in existing law beyond
any possibility for fairminded disagreement.” Richter, 562
U.S. at 103. The existing law that applies is limited to that of
the Supreme Court of the United States, which has instructed
the lower federal courts to uphold a state court conviction un‐
less the record “cannot, under any reasonable interpretation
of the [Court’s] controlling legal standard, support a certain
ruling.” Panetti v. Quarterman, 551 U.S. 930, 953 (2007). Even if
we were to consider the approach in past Supreme Court de‐
cisions outmoded, as the dissents suggest, a state court’s deci‐
sion consistent with the Supreme Court’s approach could not
be unreasonable under AEDPA.
As a result, federal habeas relief from state convictions is
rare. It is reserved for those relatively uncommon cases in
which state courts veer well outside the channels of reasona‐
ble decision‐making about federal constitutional claims.
AEDPA deference is not conclusive, however. Where the rec‐
ord shows that state courts have strayed from clearly estab‐
lished federal law, we can and do grant relief. E.g., Richardson
v. Griffin, 866 F.3d 836 (7th Cir. 2017); Jones v. Calloway, 842 F.3d
454 (7th Cir. 2016); McManus v. Neal, 779 F.3d 634 (7th Cir.
2015); Shaw v. Wilson, 721 F.3d 908 (7th Cir. 2013); Harris v.
Thompson, 698 F.3d 609 (7th Cir. 2012); Jones v. Basinger, 635
F.3d 1030 (7th Cir. 2011).
Review of state court factual findings under AEDPA is
similarly deferential. Under § 2254(d)(2), federal courts can‐
not declare “state‐court factual determinations … unreason‐
able merely because [we] would have reached a different con‐
clusion in the first instance.” Brumfield v. Cain, 135 S. Ct. 2269,
6 No. 16‐3397
2277 (2015) (internal quotation marks and citation omitted).
AEDPA does not permit federal courts to “supersede the trial
court’s … determination” if a review of the record shows only
that “[r]easonable minds … might disagree about the finding
in question.” Id. (internal quotations and citations omitted).
But again, “deference does not imply abandonment or abdi‐
cation of judicial review, and does not by definition preclude
relief.” Id. (internal quotations and citations omitted).
B. The Law of Confessions
The Due Process Clause of the Fourteenth Amendment
forbids the admission of an involuntary confession in evi‐
dence in a criminal prosecution. Miller v. Fenton, 474 U.S. 104,
109 (1985). In deciding whether a confession was voluntary,
courts assess “the totality of all the surrounding circum‐
stances—both the characteristics of the accused and the de‐
tails of the interrogation.” Schneckloth v. Bustamonte, 412 U.S.
218, 226 (1973); see also Withrow v. Williams, 507 U.S. 680, 693–
94 (1993) (collecting relevant factors). The purpose of this test
is to determine whether “the defendant’s will was in fact over‐
borne.” Miller, 474 U.S. at 116.
The Supreme Court’s many cases applying the voluntari‐
ness test have not distilled the doctrine into a comprehensive
set of hard rules, though prohibitions on physical coercion are
absolute. See Mincey v. Arizona, 437 U.S. 385, 401 (1978) (state‐
ments resulted from “virtually continuous questioning of a
seriously and painfully wounded man on the edge of con‐
sciousness”); Brown v. Mississippi, 297 U.S. 278, 279 (1936)
(confessions extracted by “brutality and violence”). AEDPA
does not “require state and federal courts to wait for some
nearly identical factual pattern before a legal rule must be ap‐
plied” because “even a general standard may be applied in an
No. 16‐3397 7
unreasonable manner.” Panetti, 551 U.S. at 953, quoting Carey
v. Musladin, 549 U.S. 70, 81 (2006) (Kennedy, J., concurring in
the judgment); accord, Yarborough v. Alvarado, 541 U.S. 652,
663–64 (2004).
Nevertheless, applying a general standard like voluntari‐
ness “can demand a substantial element of judgment,” and
determining whether that judgment is reasonable “requires
considering the rule’s specificity.” Alvarado, 541 U.S. at 664.
“The more general the rule, the more leeway courts have in
reaching outcomes in case‐by‐case determinations.” Id. (up‐
holding state court Miranda conclusion where factors pointed
in opposite directions). The state courts had such leeway here,
and in the end, that leeway is decisive as we apply the test of
§ 2254(d)(1).
This general standard has some specific requirements to
guide courts. First, a person arguing his confession was invol‐
untary must show that the police engaged in coercive prac‐
tices. See Colorado v. Connelly, 479 U.S. 157, 164–65 (1986).
Physically abusive interrogation tactics would constitute co‐
ercion per se. Stein v. New York, 346 U.S. 156, 182 (1953) (phys‐
ical violence is per se coercion), overruled on other grounds by
Jackson v. Denno, 378 U.S. 368, 381 (1964); Brown, 297 U.S. at
286–87 (coercion and brutality); United States v. Jenkins, 938
F.2d 934, 938 (9th Cir. 1991) (physical abuse is coercion per se);
Miller v. Fenton, 796 F.2d 598, 604 (3d Cir. 1986) (same).
Interrogation tactics short of physical force can amount to
coercion. The Court has condemned tactics designed to ex‐
haust suspects physically and mentally. Such tactics include
long interrogation sessions or prolonged detention paired
with repeated but relatively short questioning. Davis v. North
Carolina, 384 U.S. 737, 752 (1966) (finding coercive the practice
8 No. 16‐3397
of repeated interrogations over sixteen days while the suspect
was being held incommunicado).
The Supreme Court has not found that police tactics not
involving physical or mental exhaustion taken alone were suf‐
ficient to show involuntariness. In several cases, the Court has
held that officers may deceive suspects through appeals to a
suspect’s conscience, by posing as a false friend, and by other
means of trickery and bluff. See, e.g., Procunier v. Atchley, 400
U.S. 446, 453–54 (1971) (suspect was deceived into confessing
to false friend to obtain insurance payout to children and step‐
children); Frazier v. Cupp, 394 U.S. 731, 739 (1969) (deceiving
suspect about another suspect’s confession). False promises to
a suspect have similarly not been seen as per se coercion, at
least if they are not quite specific. See Arizona v. Fulminante,
499 U.S. 279, 285 (1991) (rejecting language in Bram v. United
States, 168 U.S. 532 (1897), stating that a confession could not
be obtained by “any direct or implied promises,” id. at 542–
43, but finding promise to protect suspect from threatened vi‐
olence by others rendered confession involuntary); Welsh S.
White, Confessions Induced by Broken Government Promises, 43
Duke L.J. 947, 953 (1994).
False promises may be evidence of involuntariness, at
least when paired with more coercive practices or especially
vulnerable defendants as part of the totality of the circum‐
stances. E.g., Lynumn v. Illinois, 372 U.S. 528, 534 (1963) (pre‐
Miranda confession found involuntary based on false promise
of leniency to indigent mother with young children, com‐
bined with threats to remove her children and to terminate
welfare benefits, along with other factors). But the Supreme
Court allows police interrogators to tell a suspect that “a co‐
operative attitude” would be to his benefit. Fare v. Michael C.,
No. 16‐3397 9
442 U.S. 707, 727 (1979) (reversing finding that confession was
involuntary). Supreme Court precedents do not draw bright
lines on this subject.
In assessing voluntariness, courts must weigh the tactics
and setting of the interrogation alongside any particular vul‐
nerabilities of the suspect. Bustamonte, 412 U.S. at 226. Rele‐
vant factors include the suspect’s age, intelligence, and educa‐
tion, as well as his familiarity with the criminal justice system.
Withrow, 507 U.S. at 693–94 (collecting factors); Michael C., 442
U.S. at 725–26 (significant criminal justice experience); Clewis
v. Texas, 386 U.S. 707, 712 (1967) (limited educational attain‐
ment); Culombe v. Connecticut, 367 U.S. 568, 620 (1961) (intel‐
lectual disability); Gallegos v. Colorado, 370 U.S. 49, 53 (1962)
(age).
The interaction between the suspect’s vulnerabilities and
the police tactics may signal coercion even in the absence of
physical coercion or threats. The Supreme Court has made it
clear that juvenile confessions call for “special care” in evalu‐
ating voluntariness. E.g., Haley v. Ohio, 332 U.S. 596, 599
(1948); see also J.D.B. v. North Carolina, 564 U.S. 261, 277 (2011);
In re Gault, 387 U.S. 1, 45 (1967); Gallegos, 370 U.S. at 54. In
juvenile cases, the law is particularly concerned with whether
a friendly adult is present for or consents to the interrogation.
In re Gault, 387 U.S. at 55–56; Gallegos, 370 U.S. at 53–54; Haley,
332 U.S. at 600. Concerns about physical exhaustion, naïveté
about friendly police in the context of an adversarial police
interview, and intellectual disability also take on heightened
10 No. 16‐3397
importance for assessing whether a juvenile’s will was over‐
borne. 2
As we detail below, Dassey’s case presents different factors
pointing in opposite directions. Those most important to our
analysis include: his age and intellectual ability; the physical
circumstances of the interrogation; the manner and actions of
the police in questioning Dassey, including bluffing about
what they knew and assuring him of the value of honesty;
Dassey’s resistance or receptiveness to suggestions by interro‐
gators; and the extent to which he provided the most incrim‐
inating information in response to open‐ended, non‐leading
questions.
2 We have reservations about the use of “suggestibility” as a factor in
this analysis, at least on these facts. Dassey relies heavily on the results of
a Gudjonsson Suggestibility Scale test measuring him as more susceptible
to fabrication than 95 people out of 100, given slight prodding by ques‐
tioners. A Gudjonsson test is administered by reading a short story aloud
to an examinee and then later asking leading questions about it. The more
answers that change in response to mild pressure, the more suggestible
the examinee is. The administration of this test for people with intellectual
disabilities has been criticized because they may have good recall of their
own lived experiences but poor recall of facts not relevant to their lives.
Paul Willner, Assessment of capacity to participate in court proceedings: a selec‐
tive critique and some recommendations, 17 Psychology, Crime & Law 117,
117 (2011). This criticism mirrors Dassey’s own testimony that his recall
was better for lived experiences. In any event, the State’s expert forcefully
contested both the administration and meaning of Dassey’s Gudjonsson
test at trial. We cannot draw conclusions from these disputed results.
No. 16‐3397 11
II. The Murder, the Interrogation, and the Convictions
A. The Murder of Teresa Halbach
With the applicable law in mind, we turn to the horrifying
murder of Teresa Halbach and then the circumstances of Das‐
sey’s confession. More detailed accounts are available in the
panel, district court, and state court opinions. See Dassey v.
Dittmann, 860 F.3d 933 (7th Cir. 2017); Dassey v. Dittmann, 201
F. Supp. 3d 963 (E.D. Wis. 2016); State v. Dassey, 346 Wis. 2d
278, 2013 WL 335923 (Wis. App. 2013) (per curiam) (un‐
published disposition); see also State v. Avery, 804 N.W.2d 216
(Wis. App. 2011) (affirming convictions of Dassey’s uncle).
In 2005, Teresa Halbach was a young photographer with
her own business based in Calumet County, Wisconsin. On
October 31, her last appointment of the day was at Avery’s
Auto Salvage to photograph a van for an advertisement. Hal‐
bach never returned from that appointment. A few days later
during a missing‐person search, her car was found at the sal‐
vage yard. Her blood stained the car’s interior. A further
search turned up Halbach’s charred remains in a burn pit on
the property, along with shell casings on the floor of Steven
Avery’s garage.
B. Dassey’s Early Police Interviews
Police investigators spoke with a number of Avery’s rela‐
tives in early November, including an hour‐long interview of
his sixteen‐year‐old nephew Brendan Dassey, who lived close
by. Dassey said he had seen Halbach taking pictures at the
salvage yard on the afternoon of October 31, but he resisted
the suggestion that she had entered Avery’s home. At that
time, he provided no other useful information.
12 No. 16‐3397
Several months later, though, investigators received word
that Dassey had been crying uncontrollably and had lost
about forty pounds of weight. They proceeded to interview
him a total of three times on February 27, 2006. In these vol‐
untary witness interviews, it became clear that Dassey knew
much more about Teresa’s murder. (Dassey was not in cus‐
tody on February 27th. He signed and initialed a Miranda
waiver, and his mother consented, though she did not sit in.)
In those interviews, Dassey admitted that on October 31st, he
had gone over to Avery’s trailer around 9:00 p.m. to help with
a bonfire. He told the police that he had seen parts of a human
body in the fire. He also said that Avery had threatened to
hurt him if he spoke to the police. When the police asked
about a pair of bleach‐stained jeans they had learned about
from another family member, Dassey admitted that he had
helped Avery clean up a spill on the garage floor late that
night. But Dassey claimed to have had nothing to do with Te‐
resa’s death.
C. The March 1st Interview and Confession
1. The Circumstances of the Interview
After those interviews, investigators thought Dassey had
been a witness to at least the aftermath of a terrible crime and
was struggling with the horror of what he had seen. On March
1st, the investigators (Mark Wiegert and Tom Fassbender) ob‐
tained his mother’s permission for another interview. They
took Dassey from his high school to a local sheriff’s depart‐
ment, where he was questioned without the presence of a
friendly adult. In the car the investigators gave Dassey stand‐
ard Miranda warnings about his right to remain silent, his
right to an attorney, and the possibility that statements he
gave could be used against him. Dassey orally acknowledged
No. 16‐3397 13
the warnings, and he initialed and signed a written Miranda
waiver form. He and the officers chatted during the ride. The
three took a short detour to Dassey’s home to retrieve his pair
of bleach‐stained jeans, which were kept as evidence. When
they arrived at the sheriff’s department, Dassey confirmed
that he understood his rights and still wanted to talk to them.
The interview took place in a so‐called “soft” interview
room equipped for videotaping. Dassey sat on a couch facing
two officers and a camera. Over the next three hours, Dassey
was repeatedly offered food, drinks, restroom breaks, and op‐
portunities to rest. At no point in the interview did the inves‐
tigators threaten Dassey or his family. Nor did they attempt
to intimidate him physically. They did not even raise their
voices. Neither investigator tried to prevent Dassey from leav‐
ing the room, nor did they use any sort of force to compel him
to answer questions. Dassey never refused to answer ques‐
tions, never asked to have counsel or his mother present, and
never tried to stop the interview.
2. The First Hour of Questioning
One officer began by telling Dassey how he could help the
investigation, since “this information and that information”
from previous accounts needed “just a little tightening up.”
Sensing that Dassey “may have held back for whatever rea‐
sons,” the officer assured Dassey “that Mark and I both are in
your corner, we’re on your side.” Acknowledging Dassey’s
potential concern that talking to the police meant he “might
get arrested and stuff like that,” the investigator urged Dassey
to “tell the whole truth, don’t leave anything out.” Talking
could be in Dassey’s best interest even though it “might make
you look a little bad or make you look like you were more
14 No. 16‐3397
involved than you wanna be,” because admitting to unfortu‐
nate facts would leave “no doubt you’re telling the truth.” The
first investigator closed by saying that “from what I’m seeing,
even if I filled” in some holes in Dassey’s story, “I’m thinkin’
you’re all right. OK, you don’t have to worry about things …
[W]e know what Steven [Avery] did … we just need to hear
the whole story from you.” The other investigator went next:
Honesty here Brendan is the thing that’s gonna
help you. OK, no matter what you did, we can
work through that. OK. We can’t make any
promises but we’ll stand behind you no matter
what you did. OK. Because you’re being the
good guy here …. And by you talking with us,
it’s, it’s helping you. OK? Because the honest
person is the one who’s gonna get a better deal
out of everything.
Supp. App. 30. After Dassey nodded in agreement, the inves‐
tigator continued:
You know. Honesty is the only thing that will set
you free. Right? And we know, like Tom said we
know, we reviewed those tapes …. We pretty
much know everything that’s why we’re ….
talking to you again today. We really need you
to be honest this time with everything, OK….
[A]s long as you be honest with us, it’s OK. If
you lie about it that’s gonna be problems. OK.
Does that sound fair?
Id. Dassey again nodded and the questioning turned to the
events of October 31st.
No. 16‐3397 15
Over the course of the next three hours, with several
breaks as the investigators conferred outside the room, Das‐
sey told an even more disturbing and incriminating story
about October 31st. In the first hour, Dassey admitted that he
received a telephone call from Avery, went over to Avery’s
garage in the six o’clock hour, and found Teresa already dead
in her car. Dassey then said he helped Avery lower Teresa’s
bound body onto a “creeper” (used to work underneath an
automobile), which he and Avery used to take her body out‐
side and throw her onto the already‐burning bonfire.
At that point, less than an hour into the interview, Dassey’s
story pivoted dramatically. Dassey revised his story to say
that he first noticed something amiss in the four o’clock hour.
Dassey volunteered that when he was out getting the mail, he
heard a woman screaming inside Avery’s trailer. Supp. App.
50. Dassey knocked on Avery’s door, ostensibly to deliver a
piece of mail, and a sweaty Avery answered the door.
Dassey said he then saw Teresa alive, naked, and hand‐
cuffed to Avery’s bed. Dassey said he went inside at Avery’s
invitation and had a soda while Avery told him that he had
raped Teresa. Dassey said that, at Avery’s urging, he then
raped Teresa, having intercourse against her will as she was
bound to the bed, and as she protested and begged him to
stop. After the rape, Dassey reported, he then watched televi‐
sion with Avery for a while. Supp. App. 55–65.
In Dassey’s telling, he next helped Avery subdue and kill
Teresa and move her to the garage. Id. at 66–76. In response to
questioning and prodding, Dassey told a confusing story
about these critical events. Dassey said that Avery stabbed Te‐
resa with a large knife, that her handcuffs were removed, and
that she was tied up with rope. He also said that Avery cut off
16 No. 16‐3397
some of her hair with that large knife, that he (Dassey) cut her
throat with the same knife, and that at some point Avery
choked or punched her. All these events reportedly happened
by 6:00 or 6:30 p.m. 3
The details and sequence of these events changed repeat‐
edly, however, as investigators pressed Dassey for more de‐
tails. This portion of the interrogation provides the most sup‐
port for Dassey’s claim that his confession was both involun‐
tary and unreliable. 4 For example, because the recovered rem‐
nants of Teresa’s skull contained trace amounts of lead, the in‐
vestigators believed that Teresa had been shot in the head.
They were eager for Dassey to describe what “else was done
to her head” besides cutting and punching. In this exchange,
Dassey did not provide the answer they were looking for. He
offered what seemed like guesses. The investigators aban‐
doned their vague admonitions to tell the truth. They lost pa‐
tience and blurted out:
Wiegert: All right, I’m just gonna come out
and ask you. Who shot her in the
head?
3 Given the damage to Teresa’s body, few of these details could have
been confirmed or contradicted by the surviving physical evidence. But
what did survive elsewhere does not necessarily vindicate Dassey. For ex‐
ample, Dassey contends that no handcuff marks were found on the head‐
board of Steven Avery’s bed, but a thin plastic film from a substance used
in rope manufacturing was found on the headboard.
4 This portion of Dassey’s confession also led to another search of Ste‐
ven Avery’s garage that uncovered perhaps the most powerful physical
evidence of the investigation: a bullet fragment with Teresa Halbach’s
DNA on it.
No. 16‐3397 17
Brendan: He did.
Fassbender: Then why didn’t you tell us that?
Brendan: Cuz I couldn’t think of it.
Fassbender: Now you remember it? (Brendan
nods “yes”) Tell us about that
then.
Supp. App. 76. Dassey continued to do so over the whole
course of the March 1st interview, revising upwards the num‐
ber of times Teresa was shot from twice to three times, and
then up to ten times. 5 Dassey also revised the location of the
shooting, first outside the garage, then inside Teresa’s car,
then on the floor of the garage. After this shifting exchange
about the shooting, the first hour of the March 1st interview
concluded with Dassey explaining how he and Avery put Te‐
resa’s body on the fire, how they moved her car, and finally
how they cleaned up the stain in Avery’s garage before Dassey
went home.
3. The Second Hour of Questioning
The investigators then took a break to confer. During the
break, Dassey had the opportunity to rest and to use the rest‐
room. Before starting up again, Dassey and Wiegert had this
exchange, indicating that Dassey did not understand the
gravity of what he had told the investigators:
Brendan: How long is this gonna take?
5 Throughout the interview, however, Dassey resisted all suggestions
that he personally ever shot Teresa, and he described his discomfort with
guns from a young age.
18 No. 16‐3397
Wiegert: It shouldn’t take a whole lot
longer.
Brendan: Do you think I can get [back to
school] before one twenty‐nine?
Wiegert: Um, probably not.
Brendan: Oh.
Wiegert: What’s at one twenty‐nine?
Brendan: Well, I have a project due in sixth
hour.
Supp. App. 102.
In the second hour of questioning, the investigators sought
to confirm details from the first. They had only limited suc‐
cess. Dassey provided more confusing details about how Te‐
resa was killed and the status of the bonfire. But in the main,
Dassey largely confirmed his account from the first hour, es‐
pecially about the details of his sexual assault of Teresa. His
story regarding what he saw of Teresa in the fire—hands, feet,
forehead, and part of a torso—also remained mostly con‐
sistent.
Signaling that the investigators did not overwhelm his
will, Dassey resisted repeated suggestions by both investiga‐
tors that he and Avery used the wires and cables hanging in
the garage to torture Teresa. The investigators also tested Das‐
sey’s suggestibility. They told him falsely that Teresa had a tat‐
too on her stomach and asked if he had seen it. Here is the
exchange:
Fassbender: … did she have any scars, marks,
tattoos, stuff like that, that you can
remember?
No. 16‐3397 19
Brendan: I don’t remember any tattoos.
…
Fassbender: OK. (pause) We know that Teresa
had a, a tattoo on her stomach, do
you remember that?
Brendan: (shakes head “no”) uh uh
Fassbender: Do you disagree with me when I
say that?
Brendan: No but I don’t know where it was.
Fassbender: OK.
Supp. App. 150–52. In this exchange, Dassey stuck to what he
thought he knew, despite being challenged and prodded by
the investigators.
4. The Final Hour of Questioning
The investigators took another break, during which Das‐
sey ate a sandwich and briefly fell asleep. The investigators
returned to talk about the consequences Dassey was facing:
Fassbender: What do you think’s gonna hap‐
pen? What do you think should
happen right now?
Brendan: I don’t know.
Fassbender: You know obviously that we’re
police officers, OK. (Brendan nods
“yes”) And … because of what
you told us, we’re gonna have ta
arrest you. Did you kinda figure
that was coming? For … what you
did we … can’t let you go right
20 No. 16‐3397
now. The law will not let us. And
so you’re not gonna be able to go
home tonight. All right?
Brendan: Does my mom know?
Fassbender: Your mom knows.
Supp. App. 157. After briefly discussing some logistics, the
exchange continued:
Fassbender: Did you kinda … after telling us
what you told us you kinda fig‐
ured this was coming? (Brendan
nods “yes”) Yeah? (Brendan nods
“yes”)
Brendan: Is it only for one day or?
Wiegert: We don’t know that at this time,
but let me tell ya something Bren‐
dan, you did the right thing. OK.
(Brendan nods “yes”) By being
honest, you can at least sleep at
night right now … .
Fassbender: Your cooperation and help with
us is gonna work in your favor. I
can’t say what [it’s] gonna do or
where [you’re] gonna end up but
[it’s] gonna work in your favor
and we appreciate your continued
cooperation. (Brendan nods
“yes”) … .
No. 16‐3397 21
Id. 6
Dassey’s mother Barb Janda then came into the room to
speak with Brendan about his arrest and confession. Dassey,
now with his head buried in his hands, asked his mother what
would happen if Avery gave a different version of events,
such as “I never did nothin’” to Teresa Halbach “or some‐
thin’.” His mother followed up on this point, asking whether
Dassey had done anything to Teresa:
Barb Janda: Did you? Huh?
Brendan: Not really.
Barb Janda: What do you mean not really?
Brendan: They got to my head.
Barb Janda: Huh?
Brendan: … say anything.
Barb Janda: What do you mean by that?
(pause) What do you mean by that
Brendan?
Supp. App. 157. Dassey was taken into custody after this in‐
terview, which he now contends was involuntary and should
not have been used at his trial.
6 If Dassey had continued to cooperate in the case against Steven
Avery, that might well have worked in his favor. At the 2010 post‐convic‐
tion hearings, Dassey’s lawyer and the prosecutor both indicated that the
State could have advocated for more lenient punishment for Dassey if he
had testified against Steven Avery. See Dkt. 19–26 at 47–48, 99–100, 158–
61.
22 No. 16‐3397
At trial, Dassey testified and denied any knowledge of or
involvement in Teresa Halbach’s murder. He did not try to ex‐
plain what he had meant by telling his mother “not really”
and “they got to my head.” According to his lawyer’s version
of events, Brendan came home from school at 3:45 p.m. on Oc‐
tober 31st and played video games until having dinner with
his brother and mother. After the others left, Dassey claimed,
he fielded a phone call from his brother’s boss and then
shortly after that a call from Avery. At “about seven‐ish,” Das‐
sey claimed, he joined Avery for the bonfire, making four or
five trips around the salvage yard picking up discarded items
to throw on the flames. Around nine o’clock, Dassey helped
Avery clean up a spill in his garage, and after a phone call
from his mother, Dassey claimed, he returned home around
9:30 or 9:45 p.m. According to his trial testimony, none of the
incriminating events related in his March 1st confession ever
happened. 7
D. The State Courts’ Treatment of Dassey’s Confession
Before trial, Dassey moved to suppress his confession as
involuntary. After briefing and a hearing, the trial judge
stated detailed findings of fact in an oral ruling. Supp. App.
7 At trial Dassey gave no explanation for his March 1st confession be‐
yond controverted expert testimony that he was highly suggestible and a
suggestion that he had confused his own experiences on October 31st with
a book he had ostensibly read “three, four years” before called Kiss the
Girls. No scenes in either the book or the movie it inspired are remotely
similar to the crimes Dassey described on March 1st. See generally James
Patterson, Kiss the Girls (1st ed. 1995); Kiss the Girls (Paramount Pictures
1997) (fictional coast‐to‐coast hunt for serial killers) Also, in nearly six
months after March 1st, Dassey never mentioned the book or movie to his
then‐counsel.
No. 16‐3397 23
168–77. The judge noted Dassey’s age and observed that he
had “an IQ level in the low average to borderline range.” The
judge noted that school records showed that Dassey was in
regular‐track classes but had some special education help.
The judge also noted Dassey’s lack of a criminal record, the
noncustodial nature of the February 27th and March 1st inter‐
views (as the parties had stipulated), and Dassey’s Miranda
waivers from both days. The judge found that Dassey knew
he could stop answering questions and knew he could leave
the room at any time on February 27th, and that he repeatedly
indicated his continuing interest in speaking with the police
on March 1st. The judge found that both Dassey and his
mother consented to the interview on March 1st. The judge
also quoted several of the investigators’ admonitions to tell
the truth, including “honesty here is the thing that’s going to
help you,” and “honesty is the only thing that will set you
free,” upon which Dassey relies so heavily now.
Throughout the interview, the judge found, the investiga‐
tors had used “a normal speaking tone with no raised voices,
no hectoring, or threats of any kind.” “Nothing on the video‐
tape visually depicts Brendan Dassey as being agitated, upset,
frightened, or intimidated by the questions of either investi‐
gator,” and he “displayed no difficulty in understanding the
questions asked of him,” the judge found. Though at times
“prodded to be truthful,” at “no time did he ask to stop the
interview or request that his mother or a lawyer be present.”
The admonitions, the judge found, amounted to “nothing
more than a reminder to Brendan Dassey that he had a moral
duty to tell the truth.” The judge also found that Dassey was
not coerced by the “interviewers occasionally pretending to
know more than they did” because that “did not interfere
with [his] power to make rational choices.” And finally, the
24 No. 16‐3397
judge found that “[n]o frank promises of leniency were made
by the … interviewers to Brendan Dassey,” and that he was in
fact flatly told “we can’t make any promises.”
On the basis of these findings of fact, “given Brendan Das‐
sey’s relevant personal characteristics” and applying “a total‐
ity of the circumstances test, which I’m using here,” the judge
found that Dassey’s admissions in the March 1st interview
were voluntary statements and denied Dassey’s motion to
suppress. Supp. App. 177.
The March 1st confession was the most incriminating evi‐
dence at trial. The jury found Dassey guilty on all charges:
participating in rape and murder, and mutilation of a corpse.
In August 2007, Dassey was sentenced to life in prison. Das‐
sey filed detailed motions for a new trial in 2009, and the same
trial court held five days of hearings on those motions in Jan‐
uary 2010, probing Dassey’s claims that his attorneys ren‐
dered ineffective assistance.
A three‐judge panel of the Wisconsin Court of Appeals af‐
firmed Dassey’s convictions, finding that his confession was
voluntary and any ineffective assistance was not prejudicial.
State v. Dassey, 346 Wis. 2d 278, 2013 WL 335923 (Wis. App.
2013). The Court of Appeals used the trial court’s findings of
fact to summarize the circumstances of the March 1st confes‐
sion and Dassey’s claim that it was involuntary. The court then
cited the legal standard for such claims—the totality of the
circumstances—as applied by leading Wisconsin state cases.
These state cases, particularly In re Jerrell C.J., 699 N.W.2d 110
(Wis. 2005), cited and discussed several of the leading prece‐
dents on voluntariness from the United States Supreme
Court. The Court of Appeals cited Jerrell C.J. for the principle
No. 16‐3397 25
that a voluntariness “analysis involves a balancing of the de‐
fendant’s personal characteristics against the police pressures
used to induce the statements.” Wisconsin law uses a clearly
erroneous standard for appellate review of trial court findings
of voluntariness.
After summarizing the trial court’s findings, the Court of
Appeals concluded:
¶7 The court’s findings are not clearly errone‐
ous. Based on those findings, we also conclude
that Dassey has not shown coercion. As long as
investigators’ statements merely encourage
honesty and do not promise leniency, telling a
defendant that cooperating would be to his or
her benefit is not coercive conduct. State v. Berg‐
gren, 2009 WI App 82, ¶31, 320 Wis. 2d 209, 769
N.W.2d 110. Nor is professing to know facts
they actually did not have. See State v. Triggs,
2003 WI App 91, ¶¶15, 17, 264 Wis. 2d 861, 663
N.W.2d 396 (the use of deceptive tactic like ex‐
aggerating strength of evidence against suspect
does not necessarily make confession involun‐
tary but instead is factor to consider in totality
of circumstances). The truth of the confession
remained for the jury to determine.
The court went on to reject Dassey’s claims that his pre‐
trial and trial counsel provided ineffective assistance. The
Wisconsin Supreme Court denied Dassey’s petition for re‐
view. Dassey did not file a petition for certiorari in the United
States Supreme Court.
26 No. 16‐3397
E. Federal Habeas Corpus Review
Dassey filed a federal habeas corpus petition in the East‐
ern District of Wisconsin in 2014. In a detailed opinion, the
district court granted habeas relief, finding that false promises
of leniency were indeed made to Dassey and that his March
1st confession was not voluntary. Dassey, 201 F. Supp. 3d 963.
A divided panel of our court affirmed. Dassey, 860 F.3d 933.
We granted the State’s petition to rehear the case en banc and
now reverse with instructions to dismiss Dassey’s habeas pe‐
tition.
III. Applying the AEDPA Standard
A. Voluntariness Under § 2254(d)(1)
The state court decision that Dassey confessed voluntarily
was not an unreasonable application of Supreme Court prec‐
edent. The state appellate court drew on fairly detailed find‐
ings of fact, which were not clearly erroneous, and provided
a terse but sufficient explanation for why the trial court’s de‐
cision was a reasonable application of the broad totality‐of‐
the‐circumstances test.
1. Factors Pointing in Opposite Directions
A number of relevant factors, we recognize, tend to sup‐
port Dassey’s claims about the March 1st confession. He was
young. He was alone with the police. He was somewhat lim‐
ited intellectually. The officers’ questioning included general
assurances of leniency if he told the truth, and Dassey may
have believed they promised more than they did. At times it
appeared as though Dassey simply did not grasp the gravity
of his confession—after confessing to rape and murder, he
asked the officers if he would be back at school that afternoon
in time to turn in a project. Portions of the questioning also
No. 16‐3397 27
included leading and suggestive questions, and throughout
the interrogation Dassey faced follow‐up inquiries when the
investigators were not satisfied with what he had told them,
leading him at times to seem to guess. In addition, the confu‐
sion and contradictions in Dassey’s account of the crimes of
October 31st lend support to the view that his confession was
the product of suggestions and/or a desire to tell the police
what they wanted to hear.
At the same time, many other factors support the finding
that Dassey’s confession was indeed voluntary. Start with the
circumstances of the interrogation. As stipulated by both
sides, Dassey was not in custody when he admitted partici‐
pating in the crimes of October 31st. He went with the officers
voluntarily and with his mother’s knowledge and consent. He
was given Miranda warnings and understood them suffi‐
ciently. The interrogation was conducted during school hours
and in a comfortable setting. Dassey showed no signs of phys‐
ical distress. He had access to food, drinks, and restroom
breaks. The interrogation was not particularly lengthy, espe‐
cially with the breaks that were taken every hour.
Dassey was not subject to physical coercion or any sort of
threats at all. Given the history of coercive interrogation tech‐
niques from which modern constitutional standards for con‐
fessions emerged, this is important. The investigators stayed
calm and never even raised their voices. As the Wisconsin
courts found, there is no sign that Dassey was intimidated.
Turning to the techniques used in the interrogation, the in‐
vestigators told Dassey many times that they already knew
what had happened when in fact they did not. Such deception
is a common interview technique. To our knowledge, it has
not led courts (and certainly not the Supreme Court) to find
28 No. 16‐3397
that a subject’s incriminating answers were involuntary. See
Frazier v. Cupp, 394 U.S. 731, 739 (1969) (fabricating a co‐con‐
spirator’s confession is relevant, but “insufficient in our view
to make this otherwise voluntary confession inadmissible”).
Also, most of the incriminating details in Dassey’s confession
were not suggested by the questioners. He volunteered them
in response to open‐ended questions.
When Dassey’s story did not make sense, seemed incom‐
plete, or seemed to conflict with other evidence, the question‐
ers pressed Dassey with further questions. Those techniques
are not coercive. Dassey responded to such questioning by
modifying his story on some points, but he stuck to his story
on others. Those passages support the view that he was not
being pushed to provide a false story against his will. For ex‐
ample, Dassey resisted repeated suggestions that he had par‐
ticipated in shooting Teresa. He denied repeated suggestions
that he and Avery had used wires and cables in the garage to
restrain or harm her. In one telling instance, the questioners
tested Dassey by falsely telling him that Teresa had a tattoo
on her stomach and asking him if he had seen it. He told them
no. When the questioners pushed harder, he was not willing
to say he knew they were wrong, but he stuck to his recollec‐
tion that he had not seen a tattoo.
Under AEDPA, the essential point here is that the totality‐
of‐the‐circumstances test gives courts considerable room for
judgment in cases like this one, where the factors point in both
directions. Given the many relevant facts and the substantial
weight of factors supporting a finding that Dassey’s confes‐
sion was voluntary, the state court’s decision was not an un‐
reasonable application of Supreme Court precedent. This
view is similar to Yarborough v. Alvarado, 541 U.S. 652, 664–65
No. 16‐3397 29
(2004), where the Supreme Court applied AEDPA to a state
court finding that a seventeen‐year‐old suspect had not been
in custody when he confessed to murder. The custody ques‐
tion was governed by a similarly general totality‐of‐the‐cir‐
cumstances standard. The Supreme Court summarized the
array of factors pointing in opposite directions, in custody or
not in custody. Emphasizing that the more general the rule,
the more leeway courts have in reaching outcomes in case‐by‐
case determinations, the Supreme Court found that the state
court finding was not an unreasonable application of binding
precedent: “These differing indications lead us to hold that
the state court’s application of our custody standard was rea‐
sonable. The Court of Appeals was nowhere close to the mark
when it concluded otherwise.” Id. at 665.
2. The Terse State Court Opinion
Dassey criticizes the Wisconsin appellate court’s decision
for having been too terse, addressing the confession in just
two pivotal paragraphs. The relative brevity of that part of the
opinion is not a reason for granting habeas relief. Given the
volume of words that federal judges have devoted to this case,
one might assume that the totality‐of‐the‐circumstances test
requires courts to detail at length the weight they have as‐
signed to all factors and how the presence of one factor affects
the weight or relevance of other factors.
That assumption would be incorrect. The Supreme Court
itself has issued terse final determinations on voluntariness
after a recitation of relevant facts. See Greenwald v. Wisconsin,
390 U.S. 519, 519–21 (1968) (per curiam); Davis v. North Caro‐
lina, 384 U.S. 737, 752 (1966). It has ruled on voluntariness by
simply adopting the reasoning of other courts. Boulden v. Hol‐
man, 394 U.S. 478, 480–81 (1969). Section 2254(d)(1) does not
30 No. 16‐3397
authorize federal courts “to impose mandatory opinion‐writ‐
ing standards on state courts.” Johnson v. Williams, 568 U.S.
289, 300 (2013). State court decisions receive significant defer‐
ence even if they provide no reasons at all. Harrington v. Rich‐
ter, 562 U.S. 86, 98–99 (2011); Whatley v. Zatecky, 833 F.3d 762,
774 (7th Cir. 2016). In this case, the state appellate court en‐
dorsed detailed findings by the trial court that provide sub‐
stantial support for the finding that Dassey’s confession was
voluntary in the eyes of the law.
3. Juveniles and Special Care
The requirement that courts take “special care” in analyz‐
ing juvenile confessions does not call for habeas relief here.
The state appellate court met the requirements for analyzing
juvenile confessions by considering Dassey’s age, his intellec‐
tual capacity, and the voluntary absence of his mother during
the interrogation. The state court noted that the officers read
Dassey his Miranda rights and that Dassey later remembered
his rights and agreed to talk anyway. The court assessed coer‐
cion in relation to Dassey’s vulnerabilities, including his “age,
intellectual limitations and high suggestibility.” The court did
not limit its inquiry to only whether the most abusive interro‐
gation techniques were used. The court examined the tones
and volumes of the investigators’ voices, finding that the of‐
ficers “used normal speaking tones, with no hectoring, threats
or promises of leniency,” though they did prod Dassey to be
honest and sought to establish a rapport with him. The court
even considered Dassey’s physical comfort by noting he sat
on a sofa and was offered food, drink, and restroom breaks.
No. 16‐3397 31
4. Precedent
Dassey simply has not pointed to Supreme Court prece‐
dent that mandates relief under these circumstances. Even in
cases where deferential review under AEDPA does not apply,
the Supreme Court has not found a confession involuntary in
circumstances like Dassey’s March 1st confession.
Consider Boulden v. Holman, 394 U.S. 478, 480–81 (1969).
The defendant there was eighteen years old, had an I.Q. of 83,
suffered from an anxiety complex, and was “susceptible to co‐
ercion.” Boulden v. Holman, 385 F.2d 102, 104, 105 (5th Cir.
1967). He was interrogated for less than three hours after be‐
ing told he had the “right not to make a statement, and that
any statement made might be used against him.” Id. at 104.
He was “treated courteously and allowed to eat, smoke and
to use [the] toilet facilities.” Id. at 105. Though two years older
than Dassey, Boulden was apparently still dependent on his
parents. Id. Other facts of his interrogation were more troub‐
ling than those in this case. Boulden was interrogated from 10
p.m. until after midnight after several hours in custody. Id. at
104. Police had denied Boulden’s father access to him, and af‐
ter Boulden asked “whether he was supposed to have a law‐
yer,” the police said “he would not get one until he talked.”
Id. The Supreme Court “determined that although the issue is
a relatively close one, the conclusion … was justified” that
Boulden had confessed voluntarily. 394 U.S. at 480–81.
In Fare v. Michael C., 442 U.S. 707, 727 (1979), the Court
again ruled a juvenile confession was voluntary. Like Dassey,
Michael C. was sixteen years old. He claimed that the police
made promises and threats during the interrogation “in the
hope of obtaining leniency for his cooperative attitude.” Id.
Michael C. indicated that his pleas to stop the interrogation
32 No. 16‐3397
were ignored. He also claimed he feared police coercion and
pointed out that he “wept during the interrogation.” Id. De‐
spite these assertions, the Court determined that Michael C.’s
claims of coercion were “without merit.” Id.
Unlike Dassey, Michael C. apparently did not have a low
average to borderline I.Q., and Michael C. did have significant
prior experience with the criminal justice system. See id. at
726. Though the presence of those factors may have provided
room for Dassey to argue on direct appeal that Michael C.
should be distinguished, they do not show that the Wisconsin
courts’ decision here was unreasonable within the meaning of
§ 2254(d)(1). As in Michael C., the police here indicated “that a
cooperative attitude would be to [the suspect’s] benefit, but
their remarks in this regard were far from threatening or co‐
ercive.” Id. at 727.
In reviewing these cases, we remember the Supreme
Court’s admonition that determining whether a confession is
voluntary “requires more than a mere color‐matching of
cases.” Reck v. Pate, 367 U.S. 433, 442 (1961). But like the Court,
we find these comparisons helpful after “careful evaluation of
all the circumstances of the interrogation.” Mincey, 437 U.S. at
401; see Reck, 367 U.S. at 442 (finding comparison to analo‐
gous cases “not inappropriate” when determining voluntari‐
ness). AEDPA “would be undermined if habeas courts intro‐
duced rules not clearly established under the guise of exten‐
sions to existing law.” Alvarado, 541 U.S. at 666. To be sure, this
line between application and extension of existing law blurs
“when new factual permutations arise.” Id. The cases show,
however, that the Supreme Court has considered and rejected
claims similar to Dassey’s, and Supreme Court cases do not
require relief here. The Wisconsin courts did not apply the
No. 16‐3397 33
law unreasonably in finding that Dassey’s confession was vol‐
untary.
B. Factual Findings Under § 2254(d)(2)
Dassey also argues that he is entitled to relief under
§ 2254(d)(2) on the ground that the state courts made an un‐
reasonable finding of fact: that the questioners made no false
promises of leniency. Affirming the trial court, which found
“no frank promises of leniency were made,” the Wisconsin
Court of Appeals determined that the investigators’ state‐
ments “merely encourage[d] honesty and [did] not promise
leniency.” Dassey’s argument that this finding was unreason‐
able focuses on two things: his intellectual limitations and the
spots in the March 1st interrogation where he claims the in‐
vestigators implied that he would not even be arrested if he
told the truth. We reject this argument.
Because the Wisconsin appellate court accepted the trial
court’s findings of fact, we review the trial court’s factual de‐
terminations directly. See Rice v. Collins, 546 U.S. 333, 339
(2006) (indicating that AEDPA review and deference in such
a situation should extend to state trial court findings). The
trial court here highlighted the key points for both sides, in‐
cluding the warning that the questioners could not make
promises (which supports the State here) and the problematic
assurance that honesty was the only thing that would set Das‐
sey free (which helps Dassey’s claim, especially in light of his
limited intellect). Whether we treat the state court’s decision
on this point as a finding of fact or a conclusion of law, we find
nothing unreasonable about it.
As noted above, the Supreme Court has not treated gen‐
eral assurances of leniency in exchange for cooperation or
34 No. 16‐3397
confession as coercive. To the extent precedents from other
courts might be helpful in understanding a state court’s fac‐
tual findings, the cases signal that such general assurances are
not legally relevant facts for determining whether a suspect’s
will was overborne and a confession was involuntary. See,
e.g., United States v. Villalpando, 588 F.3d 1124, 1129 (7th Cir.
2009); see also United States v. Binford, 818 F.3d 261, 271–72 (6th
Cir. 2016); United States v. Corbett, 750 F.3d 245, 253 (2d Cir.
2014); United States v. Jackson, 608 F.3d 100, 103 (1st Cir. 2010);
United States v. Kontny, 238 F.3d 815, 818 (7th Cir. 2001); United
States v. Rutledge, 900 F.2d 1127, 1130 (7th Cir. 1990) (“The po‐
liceman is not a fiduciary of the suspect. The police are al‐
lowed to play on a suspect’s ignorance, his anxieties, his fears,
and his uncertainties; they just are not allowed to magnify
those fears, uncertainties, and so forth to the point where ra‐
tional decision becomes impossible.”). The state appellate
court should be understood as having said that the investiga‐
tors made no legally relevant false promises to Dassey.
The district court, the panel majority, and our dissenting
colleagues have viewed the interrogation differently, finding
psychological coercion through a form of operant condition‐
ing, where different investigative tactics combined to con‐
vince Dassey that the police had agreed to end the interroga‐
tion and to grant him leniency in exchange for confessing.
Dassey, 860 F.3d at 963, 974. As the panel explained, in its view
of the interrogation, the investigators offered Dassey multiple
assurances and “sounded the theme of ‘truth leads to free‐
dom’” culminating in the “direct promise, ‘honesty is the only
thing that will set you free.’” Id.
The state courts did not view these tactics the same way.
Their view was not unreasonable. The state courts saw and
No. 16‐3397 35
read, as we have, exactly what the questioners told and asked
Dassey in the interview and how he responded. AEDPA
leaves room for reasonable disagreement between state and
federal courts. Disagreement on a particular judgment call
does not show that the state court found the facts unreasona‐
bly. Collins, 546 U.S. at 341–42.
In denying Dassey’s suppression motion, the state trial
court weighed the same statements that concerned the district
court and the panel. The judge quoted four separate instances
where investigators prodded Dassey by stating that honesty
would help him, and the judge noted that these were “but a
few example[s] of admonitions to be honest.” The state court
also recounted four quotations and other “similar statements”
where investigators assured Dassey that they were behind
him and in his corner. It viewed these statements as an “at‐
tempt to achieve a rapport” rather than “frank promises of le‐
niency.” These findings are reasonable and consistent with the
evidence and the relevant law. Habeas review does not permit
us to “use a set of debatable inferences to set aside the conclu‐
sion reached by the state court.” Collins, 546 U.S. at 342.
C. Police Best Practices and the Law
The concerns expressed by our dissenting colleagues and
the district court about the potential coercive effects of the po‐
lice tactics here are understandable. Critics of Dassey’s inter‐
rogation see evidence of fabrication through the confession’s
inconsistencies and lack of solid corroborating physical evi‐
dence. Some of the confession’s inconsistencies are startling,
particularly Dassey’s shifting answers on the location of the
shooting (outside the garage, on the garage floor, and in the
car inside the garage), and his failure to recall consistently the
order of attacks in the bedroom (stabbing, hair‐cutting, and
36 No. 16‐3397
throat‐slicing). Also, during the dialogue about Teresa’s
shooting, the investigators prodded Dassey and injected some
critical facts into the discussion that corroborated evidence
they already knew.
The state courts did not address these factual inconsisten‐
cies or the alleged lack of corroborating evidence, though it is
not clear how they should have approached the question, if at
all. United States Supreme Court precedent on this point is
not unequivocal. In Blackburn v. Alabama, 361 U.S. 199 (1960),
the Court considered the “unreliability of the confession” in
determining that a mentally ill defendant’s confession was not
voluntary. Id. at 207. The very next year the Court indicated
that “the reliability of a confession has nothing to do with its
voluntariness” because extrinsic evidence that a confession is
true can confound the inquiry into “whether a defendant’s
will has been overborne.” Jackson v. Denno, 378 U.S. 368, 384–
85 (1964), citing Rogers v. Richmond, 365 U.S. 534, 545 (1961).
The Court later seemed to signal another direction, writing in
Colorado v. Connelly that whether a confession is reliable, as
distinct from voluntary, “is a matter to be governed by the ev‐
identiary laws of the forum … and not by the Due Process
Clause of the Fourteenth Amendment.” 479 U.S. 157, 167
(1986).
Analysis of a confession’s reliability as part of the totality
of the circumstances may survive the instruction in Connelly,
but it is not unreasonable to interpret Connelly as foreclos‐
ing—or at least not requiring—this line of inquiry before trial.
We cannot fault the Wisconsin courts for failing to measure
the inconsistency of Dassey’s confession in this context. In ad‐
dition, the contradictions as to some details do not necessarily
No. 16‐3397 37
undermine the reliability of the core incriminating admis‐
sions. See Dassey, 860 F.3d at 993–94 (Hamilton, J., dissenting).
The concerns about reliability echo the opinions of schol‐
ars who believe that certain interrogation tactics tend to pro‐
duce false confessions. Some police departments and experts
have acknowledged this criticism and have changed their in‐
terrogation practices in response. We must note, though, that
some of the interrogation tactics used in this case—like the re‐
peated challenges to explain details that seem implausible—
reflect practices advocated by such reformers. See, e.g., Saul
Kassin et al., Interviewing Suspects: Practice, Science, and Future
Directions, 15 Legal & Criminological Psychology 39, 47 (2010)
(describing as “non‐coercive” the practice of investigators
“challeng[ing] suspects’ accounts, often by pointing out con‐
tradictions and inconsistencies”); Kassin, The Psychology of
Confessions, 2008 Annual Rev. of Law & Soc. Sciences 193, 208
(favoring interrogation technique where investigators “ad‐
dress discrepancies that may appear in the suspect’s narrative
account” to determine if the suspect is fabricating).
These debates over interrogation techniques have not re‐
sulted in controlling Supreme Court precedent condemning
the techniques used with Dassey. Absent a clear declaration
from the Court, we may not create new constitutional re‐
straints on habeas review. See Kernan v. Cuero, 138 S. Ct. —, —
(2017) (circuit precedent does not satisfy § 2254(d)(1), “[n]or,
of course, do state‐court decisions, treatises, or law review ar‐
ticles”).8
8 Judge Rovner’s dissent cites studies of exonerated defendants show‐
ing that false confessions are more common among juveniles and mentally
ill or intellectually deficient suspects. See post at 60–65; Dassey, 860 F.3d at
38 No. 16‐3397
D. Ineffective Assistance of Counsel
Finally, Dassey has also pursued his separate claim that his
original lawyer provided ineffective assistance of counsel on
the theory that the lawyer was operating under an actual con‐
flict of interest prohibited by Cuyler v. Sullivan, 446 U.S. 335
(1980). On this point the state and federal courts have agreed.
The Wisconsin appellate court rejected this claim. The district
court also considered this claim carefully and rejected it, cit‐
ing the limits placed on Sullivan claims by Mickens v. Taylor,
535 U.S. 162, 175, 176 (2002). Dassey, 201 F. Supp. 3d at 989.9
We agree for substantially the reasons set forth by the district
court. Id. at 987–93. In this case there was no actual conflict of
952–53 (panel majority). False confessions are a real phenomenon, and
even one is very troubling. Yet we should not conclude from these studies
of exonerated defendants that there is an epidemic of false confessions, as
might be inferred by looking at studies of only demonstrably wrong con‐
victions. The more relevant fraction uses as the denominator the number
of all confessions. That number is not easy to estimate, but we can estimate
a conservative lower boundary for the number of confessions to violent
felonies. Bureau of Justice Statistics reports on Felony Defendants in Large
Urban Counties tally violent felony convictions by guilty plea (i.e., by con‐
fessions of guilt) in just the nation’s 75 largest counties. (The most recent
report is Brian A. Reaves, U.S. Depʹt of Justice, Bureau of Justice Statistics,
Felony Defendants in Large Urban Counties, 2009—Statistical Tables
(2013), https://www.bjs.gov/content/pub/pdf/fdluc09.pdf.) The dissent’s
statistics report 227 demonstrably false confessions from 1989 to 2016. Post
at 60. From the BJS reports, we can estimate that over that period, in just
those 75 largest counties, there were more than 1.5 million guilty pleas to
violent felonies. The relevant fraction may thus be estimated conserva‐
tively as 227/1,500,000. For every one demonstrably false confession over
those years, there were more than 6,500 guilty pleas to violent felonies in
just those counties.
9 The panel majority did not reach the issue. 860 F.3d at 983.
No. 16‐3397 39
interest and no multiple or concurrent representations that
could have resulted in an actual conflict of interest.
Conclusion
Given the state courts’ reasonable findings of fact and the
absence of clearly established Supreme Court precedent that
compels relief for Dassey, the district court’s grant of habeas
relief is REVERSED. The case is REMANDED to the district
court with instructions to dismiss the petition.
40 No. 16‐3397
WOOD, Chief Judge, and ROVNER and WILLIAMS, Circuit
Judges, dissenting. Psychological coercion, questions to which
the police furnished the answers, and ghoulish games of
”20 Questions,” in which Brendan Dassey guessed over and
over again before he landed on the “correct” story (i.e., the one
the police wanted), led to the “confession” that furnished the
only serious evidence supporting his murder conviction in
the Wisconsin courts. Turning a blind eye to these glaring
faults, the en banc majority has decided to deny Dassey’s
petition for a writ of habeas corpus. They justify this travesty
of justice as something compelled by the Antiterrorism and
Effective Death Penalty Act (AEDPA). If the writ, as limited
by AEDPA, were nothing more than a dead letter, perhaps
they would be correct. But it is not. Instead, as the Supreme
Court wrote in Harrington v. Richter, 562 U.S. 86 (2011), “[t]he
writ of habeas corpus stands as a safeguard against
imprisonment of those held in violation of the law.” Id. at 91.
It is, the Court went on to say, “a guard against extreme
malfunctions in the state criminal justice systems.” Id. at 102
(citation and internal quotation marks omitted).
As the district court and the panel majority recognized, we
have before us just such an extreme malfunction. Dassey at
the relevant time was 16 years old and had an IQ in the low
80s. His confession was coerced, and thus it should not have
been admitted into evidence. And even if we were to overlook
the coercion, the confession is so riddled with input from the
police that its use violates due process. Dassey will spend the
rest of his life in prison because of the injustice this court has
decided to leave unredressed. I respectfully dissent.
No. 16‐3397 41
I
As the Wisconsin Court of Appeals correctly noted, the
question whether a confession is voluntary (i.e., not coerced)
is assessed in light of the totality of the circumstances. The age
and sophistication of the person being questioned are critical
factors. When the suspect is a minor, courts must review the
confession and record with “special care.” J.D.B. v. North
Carolina, 564 U.S. 261, 280–81 (2011); In re Gault, 387 U.S. 1, 45
(1967); Gallegos v. Colorado, 370 U.S. 49, 53–55 (1962); Haley v.
Ohio, 332 U.S. 596, 599 (1948). Courts also must take the
suspect’s intellectual capacity into account. Culombe v.
Connecticut, 367 U.S. 568, 620, 625 (1961) (opinion of
Frankfurter, J., joined by Stewart, J.); 639 (Douglas, J., joined
by Black, J., concurring); 641–42 (Brennan, J., joined by
Warren, C.J., and Black, J., concurring). Dassey, as the majority
concedes, was a mentally limited 16‐year‐old. It was thus
incumbent on the state courts to evaluate his “confession” in
light of those traits.
The Wisconsin courts failed to take this essential step.
When asked at oral argument where one might find evidence
that the state appellate court took the required special care,
counsel for the state came up dry. All counsel could do was to
point out a brief mention in the state court’s opinion of
Dassey’s age and mental capabilities. But so what? The
Supreme Court has never said or implied that the totality of
the circumstances are beside the point as long as the state
court simply jots down a fact without a hint about if or how
that fact influenced the outcome. There is nothing “special”
(or even meaningful) about a naked word on a page. The
reader has no idea whether the state court mentioned the
word meaning to indicate that it found the factor irrelevant
42 No. 16‐3397
(which would have been inconsistent with the clear Supreme
Court precedent listed above), or exculpatory, or damning.
Notably, even though the Wisconsin Court of Appeals gave a
nod to the totality test, it made no mention of the special‐care
standard for juvenile confessions.
To be sure, Harrington v. Richter, 562 U.S. 86 (2011), holds
generally that federal courts may not draw any dispositive
conclusions from a state court’s silence. But by the same token,
the state court’s silence cannot be leveraged into any assur‐
ance that the court went the extra mile required by the
U.S. Supreme Court and gave Dassey’s age and limited men‐
tal ability particularized care. The majority’s finding to the con‐
trary has no support in the record. Worse, the majority writes
off in a footnote Dassey’s extreme suggestibility by casting
doubt on the applicability of a formal test (Gudjonsson).
Ante at 10 n.2. As the painstaking review of the record re‐
flected in Judge Rovner’s panel opinion reveals, even a lay‐
person could see readily that Dassey yielded to any sugges‐
tion the person in authority made. 860 F.3d 933 (7th Cir. 2017)
(Dassey I). More generally, no court is entitled to pick and
choose which evidence to consider when evaluating the total‐
ity of the circumstances. Clearly established U.S. Supreme
Court decisions compelled the Wisconsin court to pay special
attention to Dassey’s age and intellectual abilities, including
his high level of suggestibility. Its failure to do so is one reason
why it erroneously concluded that Dassey’s “confession” was
not coerced.
If the Wisconsin Court of Appeals had done what it should
have, it could not reasonably have concluded that Dassey’s
confession was either voluntary or reliable (both of which are
required for the use of a confession to be consistent with due
No. 16‐3397 43
process). Nevertheless, first the state and now the en banc ma‐
jority have culled a sentence here and there and have at‐
tempted to craft a coherent confession from them. The video
recording of the police interrogation of Dassey, however, tells
another story—one that is diametrically opposed to the state’s
tidy and selective summary. Among the many red flags are
the following:
Dassey’s answers to questions frequently changed at
the detectives’ prodding.
The officers laid a trail of crumbs (indeed, large sign‐
posts) to the confession they sought.
Whenever Dassey went off‐course, the investigators
would shepherd him back in the desired direction—at
times with the use of fatherly assurances and gestures,
and frequently by questioning his honesty.
On both February 27 and March 1 the detectives mis‐
leadingly conveyed to Dassey, whose ability to think
abstractly was minimal, that his “honesty” was the
“only thing that will set [him] free.”
Through subsequent questioning it became clear that
“honesty” meant “what the investigators wanted to
hear.”
Dassey’s age and mental limitations made him
particularly susceptible to this psychologically manipulative
interrogation. Many of the officers’ tactics appear to be drawn
from the “Reid Technique,” which was for some time the most
widely used interrogation protocol in the country.
Miriam S. Gohara, A Lie for a Lie: False Confessions and the Case
for Reconsidering the Legality of Deceptive Interrogation
Techniques, 33 FORDHAM URB. L.J. 791, 808 (2006). The
44 No. 16‐3397
technique heavily relies on false evidence ploys and other
forms of deceit. Id. at 809. It follows a nine‐step approach:
[A]n interrogator confronts the suspect with as‐
sertions of guilt (Step 1), then develops
“themes” that psychologically justify or excuse
the crime (Step 2), interrupts all efforts at denial
(Step 3), overcomes the suspect’s factual, moral,
and emotional objections (Step 4), ensures that
the passive suspect does not withdraw (Step 5),
shows sympathy and understanding and urges
the suspect to cooperate (Step 6), offers a face‐
saving alternative construal of the alleged guilty
act (Step 7), gets the suspect to recount the de‐
tails of his or her crime (Step 8), and converts the
latter statement into a full written confession
(Step 9).
Saul M. Kassin, On the Psychology of Confessions: Does
Innocence Put Innocents at Risk?, 60 AM. PSYCHOLOGIST 215,
220 (2005); see Edwin D. Driver, Confessions and the Social
Psychology of Coercion, 82 HARV. L. REV. 42, 51–55 (1968)
(explaining the social psychological impact of the Reid
tactics). Investigators are encouraged to start by accusing the
suspect while emphasizing the importance of telling the truth.
FRED E. INBAU ET AL., CRIMINAL INTERROGATION AND
CONFESSIONS 213 (4th ed. 2001). They learn ways to build false
empathy with suspects, such as shifting the moral blame for
the offense to another person or expressing understanding for
the suspect’s actions. Id. at 213, 241–42. Investigators are
encouraged to sit physically near the suspect, maintain
“soft and warm” eye contact, and speak sincerely. Id. at 214,
349. When a suspect makes an admission implying guilt,
No. 16‐3397 45
investigators are directed to make statements of
reinforcement. Id. at 366. The technique builds in
confirmation bias; the instructions assure investigators that
while an innocent suspect will stay resolute in her denials, a
guilty person will submit to the “theme” the investigator
presents. Id. at 213; see Christian A. Meissner &
Melissa B. Russano, The Psychology of Interrogations and False
Confessions: Research and Recommendations, 1 CANADIAN J.
POLICE & SECURITY SERVS. 53, 56–57 (2003).
Courts have long expressed concern about approaches
such as the Reid Technique that rely on psychological
coercion. Just four years after the first edition of the manual
was published, INBAU ET AL., supra, at ix, the Supreme Court
in Miranda v. Arizona, 384 U.S. 436 (1966), “repeatedly cited
and implicitly criticized” the Reid approach. Gohara, supra, at
808 n.93; Miranda, 384 U.S. at 457 (“To be sure, this is not
physical intimidation, but it is equally destructive of human
dignity.”). Miranda commented that the Court for decades had
“recognized that coercion can be mental as well as physical,
and that the blood of the accused is not the only hallmark of
an unconstitutional inquisition.” Id. at 448 (quoting Blackburn
v. Alabama, 361 U.S. 199, 206 (1960)). Nothing in that respect
has changed: the Court continues regularly to hold that
psychological coercion can render a confession involuntary.
Arizona v. Fulminante, 499 U.S. 279, 287–88 (1991); Miller v.
Fenton, 474 U.S. 104, 109 (1985); Schneckloth v. Bustamonte, 412
U.S. 218, 226 (1973).
Following the Supreme Court’s guidance, we too have
repeatedly recognized that “psychological coercion alone can
result in an involuntary confession … .” United States v.
Lehman, 468 F.2d 93, 100 (7th Cir. 1972) (conceding that “subtle
46 No. 16‐3397
psychological ploys” can render a confession involuntary);
Etherly v. Davis, 619 F.3d 654, 663 (7th Cir. 2010) (considering
possible psychological coercion as part of the totality test,
while noting the need to distinguish between coercion, on the
one hand, and encouragement to tell the truth, on the other);
United States v. Villalpando, 588 F.3d 1124, 1128 (7th Cir. 2009)
(“[A] false promise of leniency may render a statement
involuntary … .”); United States v. Dillon, 150 F.3d 754, 757
(7th Cir. 1998) (“A confession is voluntary if, in light of the
totality of the circumstances, the confession is the product of
a rational intellect and free will and not the result of physical
abuse, psychological intimidation, or deceptive interrogation
tactics that have overcome the defendant’s free will.”); Burns
v. Reed, 44 F.3d 524, 527 (7th Cir. 1995) (describing the “body
of due process case law, which generally proscribes the
physical or psychological coercion of confessions” as “well‐
established, albeit heavily fact‐dependent”). Outside the
courtroom, our nation has long acknowledged through its
international commitments that mental mistreatment can be
just as bad as its physical counterpart. Convention against
Torture and Other Cruel, Inhuman or Degrading Treatment
or Punishment art. 1, Dec. 10, 1984, 1465 U.N.T.S. 85 (defining
torture to encompass physical and mental pain and
suffering).
The majority opinion downplays this reality by refusing to
acknowledge anything more than mental exhaustion and
false promises. But far worse than that was going on. Dassey’s
investigators refused to leave him alone until he gave them an
“honest” answer—where “honest” meant the answer that the
officers wanted to hear. One aspect, though by no means the
only one, of the coercion was the false promise that “honesty”
would “set him free.” But there was so much more. A brief
No. 16‐3397 47
review of what went on shows that these tactics fell decisively
on the “coercion” side of the line.
The majority finds some significance in the notion that the
detectives’ tactics were not per se coercive, but that is a red
herring. These cases cannot be assessed based on one
sentence, or one restroom break, or the comfort (or lack
thereof) of one room. The Supreme Court has instructed that
the voluntariness inquiry requires a full consideration of the
compounding influence of the police techniques “as applied
to this suspect.” Miller, 474 U.S. at 116. Many of the factors the
majority cites as evidence leaning in favor of a finding of
voluntariness—the soft interview room, offers of food and
drink, normal speaking tones—viewed in the context of the
types of questions and answers the investigators were
demanding and Dassey’s conceded intellectual disabilities,
were coercive. Psychological literature makes this clear. See
Saul M. Kassin, The Psychology of Confession Evidence, 52 AM.
PSYCHOLOGIST 221, 223–24 (1997) (criticizing the Reid
Technique’s maximization methods, or scare tactics, such as
the false evidence ploy, in addition to its minimization
methods, which “impl[y] an offer of leniency,” where police
lull a suspect into a “false sense of security” by expressing
sympathy, blaming an accomplice, and underplaying the
gravity of the situation); see also Meissner & Russano, supra,
at 57–60 (discussing the “coercive” nature of the Reid
interrogation techniques and particular concerns for minors
and suspects with low intelligence).
The state and majority brush aside even the possibility of
psychological coercion as applied to Dassey. They claim that
Dassey’s March 1 confession revealed certain “critical” details
that were corroborated by independent evidence, some of
48 No. 16‐3397
which law enforcement never publicly disclosed. I have sev‐
eral responses to that argument. First, it rests on the false idea
that if a confession is “accurate,” that indicates that it was not
coerced. See Conner v. McBride, 375 F.3d 643, 652–53 (7th Cir.
2004) (considering, under the totality test, the reliability of a
confession to support a conclusion that the confession was
voluntary). But coercion and reliability are two different
things. A confession can be coerced yet reliable, or it can be
voluntary but unreliable. Yet even if it were true that Dassey’s
confession revealed “critical” details, the confession would
not be admissible in evidence if the totality of the circum‐
stances demonstrated that it was not voluntary.
Just as importantly, a closer examination of the
supposedly reliable facts on which the majority relies shows
that they are no such thing. Without reliable facts, there is no
way to draw the Conner inference (i.e., to base a finding of
voluntariness on the reliability of the facts), questionable
though that link might be. This justifies a look at the reliability
of Dassey’s confession, even if for present purposes lack of
reliability is not a stand‐alone theory. A look at how some of
these “key” facts emerged instills no faith in either their
reliability or their knowing and voluntary quality. For ease of
reference, I have summarized in the following chart how the
investigators extracted the “critical” details they were looking
for from Dassey. It shows that there was nothing to ensure
that Dassey was offering his own independent recollection.
Instead, the officers used a combination of leading questions,
coaching, and refusal to accept one of Dassey’s guesses as the
“final” answer until it matched what they wanted to hear.
No. 16‐3397 49
“Critical” Why It Is Not How It Was Coerced
Fact Critical
Halbach Dassey was fed “Tell us, and what else
was shot in this fact through a did you do? Come on.
the head. leading question Something with the head.
after unsuccessful Brendan?”
guessing. SA 73– After Dassey guesses
76. cutting her hair,
punching her, and
cutting her throat, “All
right, I’m just gonna
come out and ask you.
Who shot her in the
head?”
Dassey’s This evidence Dassey testified that his
jeans were corroborates jeans became stained
stained Dassey’s trial with bleach while he
with bleach. testimony. helped his uncle clean up
R. 19‐21: 32–37. what looked like an
automotive fluid spill.
The RAV4’s Dassey was fed “With, how’s, the license
license this fact through a plates were taken off the
plates were leading question. car, who did that?”
removed. SA 90; R. 19‐24: On February 27, the
23. investigator also asked,
“Did he tell you if he did
anything with the license
plates?”
50 No. 16‐3397
“Critical” Why It Is Not How It Was Coerced
Fact Critical
Dassey The physical There was no evidence of
sexually evidence does not handcuffs chafing against
assaulted corroborate this the headboard. The
Halbach fact. R. 19‐17: 96– handcuffs and leg irons
while she 97; R. 19‐15: 214– found in Avery’s room
was 17. contained no fingerprints
handcuffed or DNA from Dassey or
to the bed. Halbach.
The physical The plastic film found on
evidence found the bed’s spindle was
on Avery’s bed is polypropylene, which,
not probative. according to the state’s
R. 19‐16: 246. forensic scientist, is found
in garments, in addition
to plastic containers and
rope manufacturing.
This detail was Dassey testified that he
drawn from concocted this detail from
popular media. Kiss the Girls (1995), a
R. 19‐21: 65–67. book he read, where a
woman is restrained
during a sexual assault.
Halbach The media widely This fact appeared in
was in the publicized that news stories.
back of the Halbach’s blood
RAV4. was found in the
back of the car.
RSA 70.
No. 16‐3397 51
“Critical” Why It Is Not How It Was Coerced
Fact Critical
The RAV4’s Dassey was fed “OK, what else did he do,
battery the fact that he did somethin’ else,
cables were Avery went under you need to tell us what
discon‐ the RAV4 hood he did, after that car is
nected. through a leading parked there. It’s
question after he extremely important.
unsuccessfully (pause) Before you guys
guessed. SA 92. leave that car.”
After Dassey responded
that Avery left the gun in
the car, “That’s not what
I’m thinkin’ about. He did
something to that car. He
took the plates and he, I
believe he did something
else in that car.”
“I don’t know.”
“OK. Did he, did he, did
he go and look at the
engine, did he raise the
hood at all or anything
like that?”
52 No. 16‐3397
“Critical” Why It Is Not How It Was Coerced
Fact Critical
Halbach In addition to “Was she ever in the
was shot in being fed that she garage?”
the garage. was shot, Dassey “No.”
was fed that she
was shot in the Investigators lead him,
garage, after saying “Again, we have,
initially denying w‐we know that some
she was ever in things happened in that
there. SA 81–86. garage, and in that car,
we know that. You need
to tell us about this so we
know you’re tellin’ us the
truth.”
Shortly after, they ask,
“Tell us where she was
shot?”
“In the head.”
“No, I mean where in the
garage.”
After Dassey answered
that she was shot in the
truck and not on the
garage floor, “[C]ome on,
now where was she shot?
Be honest here.”
“The truth.”
“In the garage.”
No. 16‐3397 53
“Critical” Why It Is Not How It Was Coerced
Fact Critical
Halbach’s Dassey was fed On February 27, “Did he
camera and this fact through a tell ya anything about …
phone were leading question her other possessions …
burned in a on February 27. she probably had her cell
barrel. Then on March 1, phone, a camera to take
he guessed that pictures.”
these items were
After Dassey denied
burned. R. 19‐24:
putting anything in the
36; SA 109–11.
burn barrel or knowing
whether she had a purse,
cell phone, or camera, he
was pressed about what
happened to these items
and guessed, “[Avery]
burnt ‘em.” The only
possessions he said he
saw in the burn barrel
were those fed to him
(“Like a cell phone,
camera, purse.”).
54 No. 16‐3397
“Critical” Why It Is Not How It Was Coerced
Fact Critical
Halbach’s Fed fact and On February 27,
remains media reports. investigators said, “I find
were R. 19‐24: 5–6, 9; it quite difficult to believe
burned in RSA 69. that if there was a body
the bonfire in that [fire] Brendan that
pit. you wouldn’t have seen
something like a hand, or
a foot, a head, hair,
something.” Media had
reported her remains
were found there.
Dassey Dassey’s response “We know that Teresa
resisted the seems to accept had a, a tattoo on her
suggestion the suggestion stomach, do you
that that she had a remember that?”
Halbach tattoo. SA 151–52. (shakes head “no”) “uh
had a uh.”
tattoo.
“Do you disagree with
me when I say that?”
“No but I don’t know
where it was.”
The majority concedes that AEDPA does not require a
“nearly identical factual pattern” to find that a decision
involved an unreasonable application of law. Panetti v.
Quarterman, 551 U.S. 930, 953 (2007) (citation omitted). But
that is in essence what the majority has demanded. In arguing
No. 16‐3397 55
that even non‐AEDPA cases have found confessions
voluntary under similar circumstances, the majority cites two
decisions. But as it concedes, Fare v. Michael C., 442 U.S. 707
(1979), is critically different: Michael C. was of average
intelligence and had many prior interactions with the
criminal justice system. Id. at 726. While Boulden v. Holman,
394 U.S. 478 (1969), may superficially appear to be more
similar to Dassey’s case, it is of dubious relevance given the
fact that it was decided (along with Michael C.) decades before
the Supreme Court instructed lower courts to recognize the
unique psychological vulnerabilities of youth stemming from
their incomplete neurological development. See, e.g., Graham
v. Florida, 560 U.S. 48, 68 (2010); Roper v. Simmons, 543 U.S. 551,
569–70 (2005).
The Wisconsin Court of Appeals failed reasonably to
apply in any meaningful way at least three principles that the
Supreme Court has clearly established: (1) special care for
juvenile confessions, (2) consideration of the totality of the
circumstances, and, most importantly, (3) prohibition of
psychologically coercive tactics. This led to the kind of
extreme malfunction in the adjudication of Dassey’s case for
which section 2254(d)(1) provides a remedy. By turning a
blind eye to these problems, the majority has essentially read
habeas corpus relief out of the books.
II
There is a second, independent, reason why the district
court correctly granted Dassey’s habeas corpus petition and
our original panel was correct to uphold that ruling: the
Wisconsin Court of Appeals made unreasonable factual
determinations. See Brumfield v. Cain, 135 S. Ct. 2269, 2276
(2015) (granting habeas corpus relief under section 2254(d)(2),
56 No. 16‐3397
without needing to reach petitioner’s section 2254(d)(1)
argument). The district court, whose factual assessments
deserve some deference from us, found that the Wisconsin
Court of Appeals erroneously concluded that investigators
made no promises of “leniency.” According to the district
court, though no statement in particular rendered the
confession involuntary, the cumulative effect of investigators’
tactics overbore Dassey’s free will.
The majority dismisses this concern because there was no
“specific” promise of lenience. But as the district court
concluded, when examining the totality of the circumstances,
it is clear that Dassey was guessing at what he thought the
investigators wanted to hear so that he could leave. Dassey
was reassured across two days of interviews that being
“honest” would allow him to go “free.” Although an adult of
average intelligence might recognize the Biblical allusion, see
John 8:32 (“You will know the truth, and the truth will set you
free.”), Dassey was not an adult and not of average
intelligence. Instead, he was a mentally limited teenager who
did not understand abstractions. Playing their “20 Questions”
game, the officers forced Dassey to try out different answers
until he stumbled upon the answer they wanted—defined by
them as the answer that was sufficiently truthful. And what
was Dassey’s response after all this? He asked if he was free
to go back to school to turn in a project that was due, and
when told that he could not, he indicated that he thought he
would be in jail for just one day. No more conclusive evidence
of his literalism and his lack of understanding is needed.
By finding no promises of lenience were made and that the
confession was voluntary, the Wisconsin Court of Appeals
No. 16‐3397 57
made an unreasonable determination of fact in light of the
clear and convincing weight of the evidence.
III
Under AEDPA, the role of the federal courts in reviewing
Dassey’s petition for habeas relief is quite limited. But AEDPA
does not paralyze us in the face of a clear constitutional
violation. The Due Process Clause and the right against self‐
incrimination demand that, in order to be admissible in
evidence, a suspect’s confession must be voluntary. Dassey’s
was not. Because the detectives used coercive interrogation
tactics on an intellectually disabled juvenile, Dassey’s will was
overborne during his March 1 interrogation. Without this
involuntary and highly unreliable confession, the case against
Dassey was almost nonexistent. This court should be granting
his petition for a writ of habeas corpus and giving the state an
opportunity to retry him, if it so desires. I respectfully dissent.
58 No. 16‐3397
ROVNER, Circuit Judge, and WOOD, Chief Judge, and
WILLIAMS, Circuit Judge, dissenting. I continue to believe, as I
explained in the panel opinion, and as Chief Judge Wood’s
dissent so persuasively argues, that the state court failed to
fulfill the Supreme Court’s mandate to review juvenile con‐
fessions with special care, and unreasonably held that Das‐
sey’s confession was voluntary. And for all of the reasons
upon which Chief Judge Wood has expounded and those set
forth in the original panel opinion in Dassey v. Dittmann, 860
F.3d 933 (7th Cir. 2017), rehʹg en banc granted, opinion vacated
(Aug. 4, 2017), I too respectfully dissent. I write separately
simply to point out the chasm between how courts have his‐
torically understood the nature of coercion and confessions
and what we now know about coercion with the advent of
DNA profiling and current social science research.
Although I write in the hope of encouraging courts to
update their understandings of the factual nature of coer‐
cion, my conclusion about the proper outcome of Dassey’s
habeas petition does not depend on any change in law. Cur‐
rent Supreme Court precedent requires that a court view the
totality of the circumstances of any interrogation, and to take
special care when evaluating the confessions of juveniles. To
comply with the command of the Supreme Court, therefore,
a court must include within its evaluation of the totality of
the circumstances the impact of coercive interrogation tech‐
niques upon the particular vulnerabilities of the individual
subject to those techniques. The state court did not do so in
considering Dassey’s appeal. For this reason, Dassey’s con‐
viction cannot stand. Unfortunately, four members of the
seven‐member en banc panel of this court do not agree—a
decision that I believe has worked a profound injustice.
Nevertheless, I hope to convince my colleagues throughout
No. 16‐3397 59
the courts that reform of our understanding of coercion is
long overdue. When conducting a totality of the circum‐
stances review, most courts’ evaluations of coercion still are
based largely on outdated ideas about human psychology
and rational decision‐making. It is time to bring our under‐
standing of coercion into the twenty‐first century.
Half a century ago the Supreme Court held that police
misrepresentations during interrogations, although relevant
to a totality of the circumstances inquiry, were not in and of
themselves sufficient to render an otherwise voluntary con‐
fession inadmissible. Frazier v. Cupp, 394 U.S. 731, 739 (1969).
In other words, police may deceive, trick, conceal, imply,
and mislead in any number of ways, provided that, under a
totality of the circumstances evaluation, they do not destroy
a suspect’s ability to make a rational choice. See id. (finding
an interrogator’s lie that a fellow suspect had confessed in‐
sufficient to make an otherwise voluntary confession inad‐
missible); Procunier v. Atchley, 400 U.S. 446, 454 (1971) (de‐
termining that it was not per se coercive for police to send in
a cooperating insurance agent to deceive the defendant into
confessing to obtain insurance payments for his children);
see also United States v. Villalpando, 588 F.3d 1124, 1128 (7th
Cir. 2009) (“Trickery, deceit, even impersonation do not ren‐
der a confession inadmissible”); United States v. Rutledge, 900
F.2d 1127, 1131 (7th Cir. 1990) (noting that “the law permits
the police to pressure and cajole, conceal material facts, and
actively mislead—all up to limits”).
These cases, however, were born in an era when the hu‐
man intuition that told us that “innocent people do not con‐
fess to crimes” was still largely unchecked. This belief is
rooted in the mind’s tendency to assume that statements
60 No. 16‐3397
made to a police officer that are against one’s self interest
can be trusted or, to put it simply, the thought that most of
us have that “I would never confess to a crime I did not
commit.”1 Peer‐reviewed studies confirm that jurors tend to
have hard‐to‐dislodge beliefs that a suspect who is innocent
could not be manipulated into confessing.2 And, in fact, this
false notion is precisely what the state implored the jurors in
Dassey’s trial to believe, arguing in closing that “[p]eople
who are innocent don’t confess.” R. 19‐23 at 144. We know,
however, that this statement is unequivocally incorrect. In‐
nocent people do in fact confess, and they do so with shock‐
ing regularity. As of June 7, 2016, The National Registry of
Exonerations had collected data on 1,810 exonerations in the
United States since 1989 (that number as of December 4, 2017
is 2,132), and that data includes 227 cases of innocent people
who falsely confessed.3 This research indicates that false con‐
fessions (defined as cases in which indisputably innocent in‐
dividuals confessed to crimes they did not commit) occur in
approximately 25% of homicide cases.4
1 Saul M. Kassin et al., Police‐Induced Confessions: Risk Factors and Rec‐
ommendations, 34 L. & Hum. Behav. 49, 51 (2010).
2 Iris Blandón‐Gitlin et al., Jurors Believe Interrogation Tactics Are Not
Likely to Elicit False Confessions: Will Expert Witness Testimony Inform Them
Otherwise?, 17 Psychol., Crime & L. 239, 256 (2011).
3 Samuel Gross et al., For 50 Years, You’ve Had “The Right to Remain Si‐
lent,” The National Registry of Exonerations, False Confessions (June 12,
2016), http://www.law.umich.edu/special/exoneration/Pages/false‐
confessions.aspx.
4 Samuel Gross et al., Exoneration in the United States, 1989‐2012: Re‐
port by the National Registry of Exonerations, 58, 60,
No. 16‐3397 61
In a world where we believed that “innocent people do
not confess to crimes they did not commit,” we were willing
to tolerate a significant amount of deception by the police.
Under this rubric, the thinking went, the innocent person (or
at least the vast majority of healthy, sane, innocent adults of
average intelligence) would not confess even in response to
deception and cajoling. And so our case law developed in a
factual framework in which we presumed that the trickery
and deceit used by police officers would have little effect on
the innocent.
If it is true that, except in extreme cases, innocent people
do not confess, what difference does it make if detectives
Fassbender and Wiegert made false assurances and used de‐
ception in interrogating Dassey? So what if they gave gen‐
eral assurances of leniency, used leading questions, fed Das‐
sey information, lied about how much information they had,
told Dassey that they were on his side, implored him that
“honesty is the only thing that will set you free,” suggested
answers, and even went so far as to tell a confused and
floundering Dassey that Teresa had been shot in the head?
“Dassey was not subject to physical coercion or any sort of
threats at all,” the majority tells us, and “[g]iven the history
of coercive interrogation techniques from which modern
constitutional standards for confessions emerged, this is im‐
portant.” Ante at 27.
But what do we do when the facts that supported our
“modern constitutional standards” come from a fifty‐year‐
old understanding of human behavior, and when what we
https://www.law.umich.edu/special/exoneration/Documents/exoneration
s_us_1989_2012_full_report.pdf.
62 No. 16‐3397
once thought we knew about the psychology of confessions
we now know not to be true? Our long‐held idea that inno‐
cent people do not confess to crimes has been upended by
advances in DNA profiling. We know now that in approxi‐
mately 25% of homicide cases in which convicted persons
have later been unequivocally exonerated by DNA evidence,
the suspect falsely confessed to committing the crime.5 The
majority points out that the number of known false confes‐
sions is low compared to the total number of guilty pleas to
violent felonies. Ante at 37–38 n.8. This comparison is inap‐
propriate for two reasons. First, the number of guilty pleas is
the wrong denominator. Defendants plead guilty in all man‐
ner of situations, not only after interrogations by the police,
as was the case with Dassey. Many defendants, for example,
accept a plea after carefully weighing their options with a
lawyer without ever having been subject to a coercive inter‐
rogation—the only type of confessions with which we are
concerned in this case. Moreover, and more importantly, in
the numerator, the statistics for false confessions include on‐
ly those who have been exonerated based on some form of
objective evidence (DNA, impossibility, the confession of
another, etc.). The universe of people who falsely confess is
undoubtedly larger than the subset of people who have con‐
fessed and then been fortunate enough to have been exoner‐
ated by objective, irrefutable evidence. But most important‐
ly, as the majority concedes, even one coerced false confes‐
sion is “very troubling.” Ante at 37–38 n.8. Indeed any co‐
erced false confession is an affront to due process and cannot
stand.
5 Id. at 58.
No. 16‐3397 63
Certainly human intuition makes it almost inconceivable
to imagine that someone might falsely confess to the murder
of one’s own child. Yet in October 2004, Kevin Fox of Wil‐
mington, Illinois did just that. He confessed to sexually as‐
saulting his daughter, placing duct tape over her mouth,
drowning her in the river, and then going home to sleep.6,7
His confession was detailed and included accounts of her
moving and kicking in the water and struggling to remove
the duct tape as she drowned. He quickly rescinded his con‐
fession, but spent eight months in prison until DNA testing
ruled him out as a suspect and the State of Illinois dropped
the charges. See generally Fox v. Hayes, 600 F.3d 819 (7th Cir.
2010). Not only did the DNA alone exclude him as a suspect,
but for any who had remaining doubts, the conviction of an‐
other man six years later made it unequivocally certain that
his confession had been false. In 2010, Scott Eby, who was in
prison for raping a relative, confessed to the murder.8 At the
time of the murder he had been living not far from the Fox
home. While drunk and high on cocaine Eby decided to rob
some houses, and when he happened upon a sleeping three‐
year‐old Riley Fox, he abducted her, sexually assaulted her,
and then drowned her to cover his crime. His DNA matched
that found on the duct tape used to bind Riley. A pair of
boots, which had been found at the scene, photographed,
6 Bryan Smith, Kevin Fox, in TRUE STORIES OF FALSE CONFESSIONS 107
(Rob Warden et al. eds., 2009).
7 Bryan Smith, The Nightmare: A Look at the Riley Fox Case, Chi. Mag.,
July 3, 2006.
8 Steve Schmadeke, I’m the ‘Lowest Kind of Slime,’ Killer of 3‐Year‐Old
Confessed. Court Records Outline Investigators’ Path to Scott Wayne Eby, Chi.
Trib., Feb. 26, 2011.
64 No. 16‐3397
and then ignored for years, had the name “Eby” written on
the tongue.
Five decades ago, when the Supreme Court issued its
opinions allowing interrogator deception, there was no DNA
evidence that could demonstrate with such clarity that inno‐
cent people were confessing to crimes they had not commit‐
ted at a surprising rate, and therefore, only a limited body of
psychological science explaining why this happens.
Even now, despite the overwhelming evidence regarding
the coercive nature of constitutionally permissible interroga‐
tion techniques, we have not changed our understanding of
how to view the facts surrounding coercion when evaluating
the totality of the circumstances. Yet we now have a robust
and growing body of rigorous, peer‐reviewed, legal and
psychological research demonstrating how current interro‐
gation tactics influence people, and particularly juveniles
and intellectually impaired people, to act against their own
self‐interest in such a seemingly irrational manner.9
Some of the factors that induce false confessions are in‐
ternal. Studies have demonstrated that personal characteris‐
tics such as youth, mental illness, cognitive disability, sug‐
gestibility, and a desire to please others may induce false
confessions.10 A survey of false confession cases from 1989–
2012 found that although only 8% of adult exonerees with no
known mental disabilities falsely confessed to crimes, in the
population of exonerees who were younger than 18 at the
time of the crime, 42% of exonerated defendants confessed to
9
See Saul M. Kassin, False Confessions, 8 WIREs Cogn Sci. e1439
(2017).
10 Blandón‐Gitlin et al., supra note 2, at 240.
No. 16‐3397 65
crimes they had not committed, as did 75% of exonerees
who were mentally ill or mentally disabled.11 Overall, one
sixth of the exonerees were juveniles, mentally disabled, or
both, but they accounted for 59% of false confessions.12 In‐
deed, youth and intellectual disability are the two most
commonly cited characteristics of suspects who confess
falsely.13 Dassey suffered under the weight of both charac‐
teristics.
In addition to the factors specific to the suspect, some of
the factors that induce false confessions are externally im‐
posed. These include “isolation, long interrogation periods,
repeated accusations, deception, presenting fabricated evi‐
dence, implicit/explicit threats of punishment or promises of
leniency, and minimization or maximization of the moral
seriousness or legal consequences of the offence.”14 “Maxi‐
mization” describes the technique whereby the interrogator
exaggerates the strength of the evidence and the magnitude
of the charges.15 Dassey’s interrogators employed maximiza‐
tion by constantly reminding Dassey, “We already know
everything.” See, e.g., R. 19‐25 at 17, 19, 23, 24, 26, 28, 30, 31,
36, 37, 41, 44, 47, 48, 50, 54, 55, 60, 63, 69, 71. “Minimization”
describes tactics that are designed to lull a suspect into be‐
11 Gross, Exonerations 1989–2012, supra note 4, at 60.
12 Id.
13 Samuel R. Gross et al., Exonerations in the United States 1989 through
2003, 95 J. Crim. L. & Criminology 523, 545 (2005).
14 Blandón‐Gitlin et al., supra note 2, at 240.
15 Saul M. Kassin et al., Police Interrogations and Confessions: Communi‐
cating Promises and Threats by Pragmatic Implication, 15 L. & Hum. Behav.
233, 234–35 (1991).
66 No. 16‐3397
lieving that the magnitude of the charges and the serious‐
ness of the offense will be downplayed or lessened if he con‐
fesses.16 Studies demonstrate that minimization causes sus‐
pects to infer leniency to the same extent as if an explicit
promise had been made, increasing not only the rates of true
confessions (from 46% to 81% in one experiment) but also
the rate of false confessions (from 6% to 18%).17,18 Although a
court must exclude a confession obtained by direct promise
of leniency (see, e.g., United States v. Villalpando, 588 F.3d
1124, 1128 (7th Cir. 2009)), the research demonstrates that
minimization techniques are the functional equivalent in
their impact on suspects.19 The investigators in this case em‐
ployed classic minimization techniques by repeatedly telling
Dassey that it was not his fault that he committed the crime
because his uncle, Steven Avery, had made him do it. See,
e.g., R. 19‐25 at 28, 47, 50, 60, 62. As Chief Judge Wood
points out in her dissent, interrogators in this case, as in
most police forces in the United States, used the Reid Tech‐
nique to obtain Dassey’s confession. This technique involves
isolation, confrontation, maximization and minimization—
the psychological strong‐arm tactics that are known to pro‐
duce coerced confessions even in adults of average intelli‐
gence.
16 Id. at 235.
17 Id. at 241, 248.
18
Melissa B. Russano et al., Investigating True and False Confessions
Within a Novel Experimental Paradigm, 16 Psychol. Sci. 481, 484 (2005).
19 Kassin, Police Interrogations and Confessions, supra note 15, at 241,
248.
No. 16‐3397 67
Dassey’s interrogation thus combined a perfect storm of
these internal and external elements. He was young, of low
intellect, manipulable, without a friendly adult, and faced
repeated accusations, deception, fabricated evidence, implic‐
it and explicit promises of leniency, police officers disingen‐
uously assuming the role of father figure, and assurances
that it was not his fault.20
For many years, the Reid technique has been criticized by
scholars and experts for increasing the rate of false confes‐
sions. 21 As far back as Miranda, the Supreme Court warned
20 The majority has reservations about the use of the Gudjonsson
Suggestibility Scale and thus states that it can make no conclusions from
the disputed expert testimony about the results. Ante at 10 n.2. Whatever
one might make of the Gudjonsson Suggestibility Scale, the interrogation
speaks for itself. Dassey is almost frantic in his desire to find the story
the investigators seek. For example, in response to the question about
what happened to Teresa’s head, Dassey guessed at every possible injury
or injustice to a head (hitting, punching, throat cutting, hair cutting) hop‐
ing to please the officers until, in frustration, they finally informed him
that Teresa had been shot in the head. R. 19‐25 at 60–63. In response to
pressure from the investigators, he changes the locale of the crime from
the house to the garage (Id. at 72–73), the color of Teresa’s clothes (Id. at
20, 31–32), the location of the knife (Id. at 80–81, 121; R. 19‐34 at 23–24,
27), whether Teresa was standing on the porch after school (R. 19‐25 at
19–20, 27–28, 90–91), whether Avery went under the hood of Halbach’s
car (Id. at 77–80), when the fire occurred (Id. at 23, 32–33; R. 19‐34 at 55),
and whether he cut her hair (R. 19‐35 at 60–61; R. 19‐34 at 36–37, 65–66,
98). Even under the state’s theory of the case, the naïve Dassey, who had
never been in trouble with the law and had never had a sexual experi‐
ence with a woman, was readily manipulated by his uncle into partici‐
pating in a repulsive and heinous crime. One does not need the Gudjons‐
son Suggestbility Scale to conclude, under either party’s theory of the
case, that Dassey was highly suggestible and manipulable.
21 Kassin, False Confessions, supra note 9, at 8.
68 No. 16‐3397
that “[e]ven without employing brutality, the ‘third degree’”
used in the Reid technique “exacts a heavy toll on individual
liberty and trades on the weakness of individuals,” and
“may even give rise to a false confession.” Miranda v. Arizo‐
na, 384 U.S. 436, 455 & n.24 (1966). Recently, Wicklander‐
Zulawski & Associates, one of the nation’s largest police
consulting firms, said it will stop training detectives in the
method it has taught since 1984, stating that it “is not an ef‐
fective way of getting truthful information.”22 After a spate
of high‐profile false confession cases in the 1980’s, Great
Britain transitioned from an accusatorial and coercive Reid‐
like approach to an investigative model of interviewing
which prohibits deception, coercion, and minimization.23
Meta‐analyses of twelve different laboratory experiments
indicate that the accusatorial approach increased both true
and false rates of confessions, while the information‐
gathering approach increased the rate of true confessions
without also increasing false confessions.24
No reasonable state court, knowing what we now know
about coercive interrogation techniques and viewing Das‐
sey’s interrogation in light of his age, intellectual deficits,
and manipulability, could possibly have concluded that
22 Eli Hager, The Seismic Change In Police Interrogations: A Major Player
In Law Enforcement Says It Will No Longer Use A Method Linked To False
Confessions, The Marshall Project (March 7, 2017, 10:00 p.m.),
https://www.themarshallproject.org/2017/03/07/the‐seismic‐change‐in‐
police‐interrogations.
23 Kassin, False Confessions, supra note 9, at 8.
24 Christian A. Meissner et al., Accusatorial and Information Gathering
Interrogation Methods and Their Effects on True and False Confessions, A Me‐
ta‐Analytic Review, 10 J. Exp. Criminology 459, 481–82 (2014).
No. 16‐3397 69
Dassey’s confession was voluntarily given. Although it is my
hope that our courts will, when evaluating the totality of the
circumstances, engage with the more current understanding
of coercion, as I noted at the start, Dassey does not need a
change in our existing Supreme Court precedent or any ex‐
isting law to prevail on his habeas petition. What has
changed is not the law, but our understanding of the facts
that illuminate what constitutes coercion under the law.
Moreover, even under our current, anachronistic under‐
standing of coercion, Dassey’s confession was so obviously
and transparently coercively obtained that it is unreasonable
to have found otherwise. Dassey, however, need not rely on
this finding either. Existing Supreme Court precedent allows
for significantly deceptive and manipulative interrogation
techniques, but those very techniques must then be evaluat‐
ed, in a totality of the circumstances analysis, for what they
are.
The requirement that confessions must be voluntary is a
principle at the heart of our legal system. Although psycho‐
logical and physical torture and coercion are commonplace
in some countries as a means of obtaining “confessions,” our
system of justice rejects the notion that convictions can be
obtained through such abuse. We refuse to accept such con‐
duct as a means of obtaining information, not only because it
impacts the veracity of the confession, but because it is con‐
duct that we as human beings cannot tolerate from our gov‐
ernment. In a case such as this one, where investigators are
faced with a crime of horrific brutality and the loss of a
treasured life, the impulse to coerce a confession from a sus‐
pect may be particularly strong. As judges, we are entrusted
with the responsibility to protect against such abusive ac‐
70 No. 16‐3397
tions, and uphold those principles that our Constitution pro‐
tects even in the darkest of times.
What occurred here was the interrogation of an intellec‐
tually impaired juvenile. Dassey was subjected to myriad
psychologically coercive techniques but the state court did
not review his interrogation with the special care required
by Supreme Court precedent. His confession was not volun‐
tary and his conviction should not stand, and yet an im‐
paired teenager has been sentenced to life in prison. I view
this as a profound miscarriage of justice. I respectfully dis‐
sent.