State v. Levin

1 U.S. Const. amend V.
2 See Miranda v. Arizona, 384 U.S. 436 (1966).
3 Rhode Island v. Innis, 446 U.S. 291, 297, 300-01 (1980);
Miranda, 384 U.S. at 444-45; Salt Lake City v. Carner, 664 P.2d
(continued...)
2006 UT 50
This opinion is subject to revision before final
publication in the Pacific Reporter.
IN THE SUPREME COURT OF THE STATE OF UTAH
----oo0oo----
State of Utah, No. 20050001
Plaintiff and Respondent,
v.
F I L E D
Ralph Levin,
Defendant and Petitioner. September 8, 2006
---
Fourth District, Provo Dept.
The Honorable Lynn W. Davis
No. 011402539
Attorneys: Mark L. Shurtleff, Att’y Gen., Erin Riley, Asst.
Att’y Gen., Salt Lake City, for plaintiff
Margart P. Lindsay, Orem, for defendant
---
On Certiorari to the Utah Court of Appeals
DURRANT, Justice:
INTRODUCTION
¶1 The Fifth Amendment to the United States Constitution
protects against self-incrimination.1 To preserve this right,
the U.S. Supreme Court has held that defendants subjected to
custodial interrogation are entitled to a Miranda warning.2
Where such a warning is not given, any incriminating statements
made by a defendant during the custodial interrogation are
excluded from evidence.3 We granted certiorari in this case to

3 (...continued)
1168, 1170 (Utah 1983).
4 State v. Levin, 2004 UT App 396, ¶¶ 11, 22-23, 101 P.3d
846.
5 Id. ¶ 7.
6 See id. ¶¶ 7, 11.
No. 20050001 2
clarify the standard of review to be applied by a Utah appellate
court in reviewing a trial court’s decision on whether a
defendant was subjected to custodial interrogation.
¶2 Following a jury trial, defendant Ralph Levin was found
guilty of possession of marijuana and possession of drug
paraphernalia. He challenged his convictions before the Utah
Court of Appeals, arguing that the trial court had erred in
failing to suppress certain incriminating statements he made to
police officers without the benefit of a Miranda warning. In
particular, he contended that the trial court had erred in
concluding that the officers had not subjected him to “custodial
interrogation.”
¶3 The court of appeals upheld Levin’s convictions. It
held that the trial court had erred when it concluded that Levin
was not interrogated but affirmed the trial court’s determination
that he was not “in custody.”4 In so doing, the court of appeals
expressly applied an “abuse of discretion” standard to the
custody determination.5 It did not specify the standard of
review that it applied to the trial court’s interrogation
determination.6 On certiorari, Levin challenges the court of
appeals custody determination, arguing that it was error for the
court of appeals to apply the deferential “abuse of discretion”
standard of review.
¶4 Because the “custody” and “interrogation” elements of a
trial court’s custodial interrogation determination overlap and
together serve to answer a single inquiry into whether a Miranda
warning was necessary in a particular case, we address the
standard of review for the trial court’s custodial interrogation
determination as a whole. A trial court’s application of the
legal concept of custodial interrogation to the facts of a
particular case presents a mixed question of fact and law.
Therefore, we select the appropriate standard of review using the
general factors and policy considerations that we have discussed

7 869 P.2d 932, 935-40 (Utah 1994).
8 See State v. Johnson, 748 P.2d 1069, 1076 (Utah 1987)
(Durham, J., concurring separately, joined by Howe & Zimmerman,
JJ.) (indicating that there will be sufficient notice of a
continuing objection if counsel renews the objection at trial
outside the presence of the jury).
3 No. 20050001
in State v. Pena7 and its progeny. However, we take this
opportunity to revise our statement of the original four factors
from Pena into a three-factor, policy-based balancing test. We
then conclude that the three factors of this revised balancing
test weigh in favor of reviewing for correctness a trial court’s
custodial interrogation determination.
¶5 Ultimately, the important policy of promoting
uniformity in police officers’ administration of Miranda warnings
and in courts’ application of Miranda to exclude a defendant’s
incriminating statements mandates greater appellate involvement
in defining the concept of custodial interrogation as it applies
to the facts of individual cases. Because the court of appeals
applied a deferential standard in reviewing the trial court’s
determination that Levin was not in custody, we remand to the
court of appeals for application of the “correctness” standard of
review.
BACKGROUND
¶6 Because our opinion is concerned only with defining the
appropriate standard of review, we limit our discussion of the
facts in this case. Although somewhat abbreviated, our factual
discussion incorporates undisputed facts established at trial as
well as those found by the trial court in connection with Levin’s
initial motion to dismiss. We consider undisputed facts from
trial because at the beginning of the trial Levin made an
appropriate continuing objection to the introduction of his
incriminating statements by renewing his earlier motion to
suppress those statements.8
¶7 Levin’s convictions for drug offenses are based on
evidence gathered during an approximately one-and-one-half hour
traffic stop on the Provo Dike Road in a rural area near Utah
Lake. Deputy Wayne Keith of the Utah County Sheriff’s Office was
on patrol when he noticed a convertible bearing expired
registration tags parked on the side of the road. Three
occupants were sitting in the convertible with the roof down.
Without activating his lights or siren, Deputy Keith parked
behind the convertible. He approached on foot and saw several

No. 20050001 4
open containers of alcohol in plain view inside the convertible.
¶8 Deputy Keith asked the convertible’s occupants for
identification. Levin was in the driver’s seat, Michael Winger
was a passenger in the front seat, and Richard Johnson was
sitting in the backseat. Deputy Keith had all three men step out
of the vehicle and notified them that he was going to search for
more open containers. His search of the vehicle’s center console
uncovered an odor of marijuana and a metal “socket” tool that had
been fashioned into a pipe, which appeared to contain burnt and
unburnt marijuana. Deputy Keith also found several small bags of
marijuana in a backpack claimed by Johnson.
¶9 There is some dispute over the precise chronology of
the following events, but the record establishes that Deputy
Keith called in two deputies who were certified drug recognition
experts. Because the vehicle belonged to Levin and he had been
sitting in the driver’s seat, Deputy Keith pulled Levin aside and
personally subjected him to a sobriety test designed to identify
alcohol impairment. He passed. The drug recognition experts
then subjected Levin to additional field sobriety tests. Those
officers determined that Levin had a fast pulse rate and a lack
of convergence of the eyes. They informed Deputy Keith that they
believed Levin was under the influence of marijuana. At some
point, either before or after these tests, Deputy Keith asked
Levin at least once about the socket, and Levin asserted that he
knew nothing about it and had not smoked marijuana. Deputy Keith
also patted Levin down but found no marijuana and no scent of
marijuana on him.
¶10 However, after the drug recognition experts presented
their conclusions to Deputy Keith, Deputy Keith pulled Levin
aside and stated: “There’s no doubt in my mind that you’ve been
smoking marijuana.” Deputy Keith’s accusation was not phrased in
the form of a question, and Deputy Keith was not “in Levin’s
face.” Deputy Keith testified that he did not expect a response
because Levin had already denied using marijuana. Nevertheless,
Levin answered by saying that “he had taken a couple of hits”
with Richard Johnson but that Michael Winger had not used any
marijuana. He also added that they had smoked out of a pipe that
the officers had not located. At no time was Levin formally
arrested, handcuffed, or given a Miranda warning, although he was
issued a citation.
¶11 In addition to the investigation of Levin, the officers
questioned the two passengers. The officers briefly questioned
Winger about smoking marijuana. They read Johnson his Miranda
rights and questioned him about the marijuana found in his
backpack. Johnson admitted that he had been smoking with Levin,

9 State v. Levin, 2004 UT App 396, ¶¶ 7, 11, 22-23, 101
P.3d 846.
10 State v. Brake, 2004 UT 95, ¶ 11, 103 P.3d 699.
5 No. 20050001
but later said that he had smoked the marijuana alone. When the
officers had completed their investigation, they allowed Levin
and his passengers to leave in the convertible. As the
convertible started to drive away, one of the officers spotted a
pipe located directly under the convertible. The officers
stopped the convertible, and Deputy Keith asked if this was the
pipe they had used to smoke. Johnson stated that it was.
¶12 Levin was later charged with possession and use of
marijuana, possession of drug paraphernalia, and with having an
open container in a vehicle. Levin pleaded no contest to the
open container charge. With regard to the drug offenses, he
pleaded not guilty and then moved to suppress the incriminating
statements he had made to Deputy Keith, arguing that despite
being subjected to custodial interrogation, he had not been given
the required Miranda warning. The trial court denied the motion.
It determined that Levin had not been in custody or subject to
interrogation. At the commencement of trial, Levin renewed his
motion, which the trial court again denied. Following the trial,
a jury found Levin guilty of both possession of marijuana with a
prior conviction and possession of drug paraphernalia.
¶13 On appeal, the court of appeals concluded that Levin
had been subjected to “interrogation,” but it applied a
deferential “abuse of discretion” standard of review to the trial
court’s determination that Levin was not “in custody” and
affirmed that determination.9 It did not specify the standard of
review that it applied to the trial court’s interrogation
determination.
¶14 Levin now challenges the court of appeals’ decision,
arguing that the court of appeals erred in reviewing the trial
court’s custody determination under a deferential abuse of
discretion standard. We granted Levin’s petition for certiorari
to decide whether the court of appeals applied the correct
standard of review. We have jurisdiction pursuant to Utah Code
section 78-2-2(5).
STANDARD OF REVIEW
¶15 On certiorari, we review for correctness the decision
of the court of appeals, not the decision of the trial court.10
The correctness of the court of appeals’ decision turns, in part,

11 Id.
12 869 P.2d 932, 939 (Utah 1994).
13 2004 UT 95, ¶ 14, 103 P.3d 699.
No. 20050001 6
on whether it accurately reviewed the trial court's decision
under the appropriate standard of review.11
ANALYSIS
¶16 Levin argues that the court of appeals erred in
reviewing the trial court’s determination that he had not been
“in custody” under an “abuse of discretion” standard. We agree.
The importance of uniformity in Utah courts’ application of Fifth
Amendment Miranda protections leads us to conclude that Utah
appellate courts should review for correctness trial courts’
custodial interrogation determinations.
¶17 We will begin our analysis by describing the policy-
based balancing test that guides our selection of a standard of
review for mixed questions of fact and law. We will then apply
this balancing test to the mixed question of custodial
interrogation.
I. WE DETERMINE THE STANDARD OF REVIEW FOR A MIXED QUESTION OF
FACT AND LAW BY EMPLOYING A POLICY-BASED BALANCING TEST
¶18 In selecting the deferential standard of review that it
applied to the trial court’s “custody” determination in this
case, the court of appeals engaged in an incomplete analysis of
the factors that we discussed in State v. Pena12 and did not
adequately consider the policy implications that we highlighted
in State v. Brake.13 Given the lingering difficulties in the
application of our standard of review jurisprudence, we take this
opportunity to further discuss the role of policy in our
selection of a standard of review and to refine our statement of
the balancing test that we use in selecting a standard of review
for a mixed question of fact and law.
A. Standards of Review Apportion Power Between the Trial and
Appellate Courts Based on the Courts’ Institutional Competencies
¶19 We have previously explained that “[t]he primary
function of a standard of review is to apportion power and,
consequently, responsibility between trial and appellate courts

14 State v. Thurman, 869 P.2d 1256, 1265-66 (Utah 1993).
15 Id. at 1266.
16 See Pena, 869 P.2d 932, 936 (Utah 1994) (“[O]ne can
visualize the traditional standard-of-review scheme as a
continuum of deference anchored at either end by the clearly
erroneous and correction-of-error standards, which correspond
with whether the issue is characterized as one of fact or of
law.”).
17 Thurman, 846 P.2d at 1271.
18 Id.; accord Pena, 869 P.2d at 936 (“[A]ppellate courts
have traditionally been seen as having the power and duty to say
what the law is and to ensure that it is uniform throughout the
jurisdiction.”).
19 Thurman, 846 P.2d at 1271.
7 No. 20050001
for determining an issue.”14 Standards of review should allocate
discretion between the trial and appellate courts in a way that
takes account of the “relative capabilities of each level of the
court system to take evidence and make findings of fact in the
face of conflicting evidence, on one hand, and to set binding
jurisdiction-wide policy, on the other.”15 These considerations
are critical in selecting a standard of review from along a
spectrum of deference that runs from highly deferential review
under a “clearly erroneous” standard on one end to completely
nondeferential review under a “correctness” standard on the other
end.16
¶20 Because a trial court is in a better position to
“judg[e] credibility and resolv[e] evidentiary conflicts,” an
appellate court reviews the trial court’s findings of fact for
clear error.17 Conversely, an appellate court reviews a trial
court’s conclusions of law for correctness because “a single
trial judge is in an inferior position to determine what the
legal content of [a legal concept] should be [whereas] a panel of
appellate judges, with their collective experience and their
broader perspective, is better suited to that task.”18
Additionally, the published decisions of appellate courts
“provid[e] state-wide standards that guide law enforcement and
prosecutorial officials.”19
¶21 The analytical complexity of our standard of review is
at its height when we review a trial court’s application of a
legal concept to a given set of facts. When we review so-called

20 See Pena, 869 P.2d at 938; see also Searle v. Milburn
Irrig. Co., 2006 UT 16, ¶ 16, 133 P.3d 382 (“The measure of
discretion afforded varies, however, according to the issue being
reviewed.” (internal quotation marks omitted)).
21 Drake v. Indus. Comm’n, 939 P.2d 177, 181 (Utah 1997)
(internal quotation marks omitted).
22 See Pena, 869 P.2d at 938-40.
23 Id. at 940.
24 Id. at 937-39; State v. Virgin, 2006 UT 29, ¶ 27, 137
P.3d 787.
25 See Brake, 2004 UT 95, ¶¶ 14-15; see also State v.
Hansen, 2002 UT 125, ¶ 26, 63 P.3d 650 (stating that there must
be “statewide standards that guide law enforcement and
(continued...)
No. 20050001 8
“mixed questions of fact and law,” the considerations that favor
a more-deferential standard of review and those that favor a
less-deferential standard of review compete for dominance, and
the amount of deference that results will vary according to the
nature of the legal concept at issue. Mixed questions of fact
and law involving different legal issues will often require
different standards of review.20
¶22 While we have said that, ultimately, “the legal effect
of [the] facts is the province of the appellate courts,”21 our
prior decisions recognize that, with regard to many mixed
questions of fact and law, it is either not possible or not wise
for an appellate court to define strictly how a legal concept is
to be applied to each new set of facts.22 Where the correct
application of a legal concept is difficult to explain using a
generally applicable standard, overinvolvement by an appellate
court can lead to confusing and inconsistent pronouncements of
the law.23 We have recognized that the application of such a
legal concept incorporates a de facto grant of discretion to the
trial court, and, accordingly, we review the trial court’s
decision on the mixed question of fact and law with deference
commensurate to that discretion.24
¶23 But with regard to certain mixed questions where
uniform application is of high importance, as in the context of
Fourth Amendment protections, we have held that policy
considerations dictate that the application of the legal concept
should be strictly controlled by the appellate courts.25 Thus,

25 (...continued)
prosecutorial officials” (internal quotation marks omitted)).
26 See Brake, 2004 UT 95, ¶¶ 14-15.
27 See Pena, 869 P.2d at 937-38 (citing Maurice Rosenberg,
Judicial Discretion of the Trial Court, Viewed From Above, 22
Syracuse L. Rev. 635, 662-63 (1971)); Brake, 2004 UT 95, ¶ 14.
28 See Pena, 869 P.2d at 937-38.
29 Virgin, 2006 UT 29, ¶ 27 (quoting Pena, 869 P.2d at
937).
30 See Pena, 869 P.2d at 938-40.
31 See Brake, 2004 UT 95, ¶ 14 (“Considerations of policy
play a central part in the placement of discretionary fences.”)
9 No. 20050001
if we determine that society’s interest in establishing
consistent statewide standards outweighs other considerations, we
grant no discretion to the trial court, and we review the mixed
question for correctness.26
¶24 We have described the varying levels of discretion
afforded trial courts in Pena and Brake using Professor Maurice
Rosenberg’s pasture metaphor, which describes the discretion
given to a trial court on a particular mixed question as a
pasture bounded by fences that represent the boundaries of the
legal concept.27 Because the established boundaries of each
legal concept are unique, different mixed questions are
associated with pastures of different sizes.28 When a trial
court stays within the pasture associated with a specific legal
concept, it is free “to reach one of several possible conclusions
about the legal effect of a particular set of facts without
risking reversal.”29 Discretion is broadest--and the standard of
review is most deferential--when the application of a legal
concept is highly fact dependant and variable.30 Discretion is
most confined--and the standard of review is nondeferential--when
the legal concept is easily defined by appellate courts or when
appellate courts erect strict fences for policy reasons.31
B. The Test We Employ in Determining the Standard of Review for
Mixed Questions Balances Policy Considerations Related to Courts’
Institutional Competencies
¶25 In Pena and its progeny, we have articulated four
factors to guide Utah appellate courts in the difficult task of

32 2006 UT 29, ¶ 28, 137 P.3d 787 (“[W]e quantify [a trial
court’s] discretion by weighing the following factors: (1)
whether the facts to which the legal rule is to be applied are so
complex and varying that no rule adequately addressing the
relevance of all these facts can be spelled out; (2) whether the
situation to which the legal principle is to be applied is
sufficiently new to the courts that appellate judges are unable
to anticipate and articulate definitively what factors should be
outcome determinative; (3) whether the trial judge has observed
‘facts,’ such as a witness’s appearance and demeanor, relevant to
the application of the law that cannot be adequately reflected in
the record available to appellate courts; and (4) whether there
are policy reasons that weigh for or against granting discretion
to trial courts, such as when substantial constitutional rights
are implicated.” (citations and internal quotation marks
omitted)).
33 Id.
No. 20050001 10
selecting the appropriate standard of review for a mixed question
of fact and law from the spectrum of possible levels of deference
to a trial court. Most recently, we discussed these factors in
State v. Virgin.32 However, this four-factor test has continued
to cause some confusion. As will be explained, we therefore take
this occasion to refine the test by eliminating a factor that has
proven to be unhelpful and rephrasing the others to better
reflect the purpose of the test. Our revised test considers the
following factors: (1) the degree of variety and complexity in
the facts to which the legal rule is to be applied; (2) the
degree to which a trial court’s application of the legal rule
relies on “facts” observed by the trial judge, “such as a
witness’s appearance and demeanor, relevant to the application of
the law that cannot be adequately reflected in the record
available to appellate courts;” and (3) other “policy reasons
that weigh for or against granting discretion to trial courts.”33
¶26 As to the first factor, the greater the complexity and
variety of the facts, the stronger the case for appellate
deference. As to the second, the greater the importance of a
trial court’s credibility assessments that cannot be adequately
reflected in the record, the stronger the case for appellate
deference. The third factor requires that we take into
consideration policy factors related to the degree of deference
that should be applied. Even where a case for appellate
deference is strong under the first two factors, policy

34 See Brake, 2004 UT 95, ¶ 14.
35 See Pena, 869 P.2d at 939 (listing three circumstances
in which the trial court should be given discretion, namely: “(i)
when the facts to which the legal rule is to be applied are so
complex and varying that no rule adequately addressing the
relevance of all these facts can be spelled out; (ii) when the
situation to which the legal principle is to be applied is
sufficiently new to the courts that appellate judges are unable
to anticipate and articulate definitively what factors should be
outcome determinative; and (iii) when the trial judge has
observed ‘facts,’ such as a witness’s appearance and demeanor,
relevant to the application of the law that cannot be adequately
reflected in the record available to appellate courts” (quoting
Rosenberg, supra note 27, at 662-63)).
36 See, e.g., Desert Miriah, Inc. v. B & L Auto, Inc., 2000
UT 83, ¶¶ 10, 12, 12 P.3d 580 (quoting and applying the original
three Pena factors and then mentioning in a later paragraph that
there are “no policy reasons outweighing” the first three factors
suggesting a deferential standard).
11 No. 20050001
considerations may nevertheless lead us to limit that
deference.34
¶27 While the above balancing test reflects the principles
relied upon in our opinions in Pena and its progeny and does not
significantly depart from our prior statements regarding mixed
questions of fact and law, we have rephrased the language of the
factors and have eliminated the second Pena factor--the novelty
of the situation. We have made these revisions to enhance the
analytical consistency and clarity of the balancing test to be
applied in placing different mixed questions along the spectrum
of deference and discretion.
¶28 As the first three factors for determining the standard
of review have been phrased in Pena and its progeny, their
application suggested only “yes” or “no” answers,35 making the
factors ill-suited to use in a balancing test. Further, by
quoting these first three Pena factors verbatim in the process of
transforming them into a balancing test, we have in many of our
earlier statements of the balancing test overemphasized their
importance and artificially divorced them from our central
concern with the policy implications of selecting a more-
deferential or less-deferential standard of review.36 In
contrast, we have recently reaffirmed the centrality of policy

37 2004 UT 95, ¶¶ 14-15.
38 2006 UT 29, ¶ 28.
39 869 P.2d at 939.
40 See Searle v. Milburn Irr. Co., 2006 UT 16, ¶ 17, 133
P.3d 382 (concluding that “at least some deference should be
granted to the district court’s application of the law to the
facts” where it was “exceedingly difficult to craft a uniform
rule neatly applicable in all situations”); Butler, Crockett &
Walsh Dev. Corp. v. Pinecrest Pipeline Operating, 2004 UT 67,
(continued...)
No. 20050001 12
considerations in our decision in Brake37 and treated policy
considerations as a “fourth” Pena factor in State v. Virgin.38
Thus, to clarify the appropriate test, we have rephrased the
factors here in a manner that better reflects their usefulness in
selecting a standard of review from somewhere along the spectrum
of deference.
¶29 In the process of rephrasing the test, we have dropped
the “novelty” factor because it has rarely, if ever, proven to be
helpful to our analysis. As it was phrased in Pena, the novelty
factor considered whether “the situation to which the legal
principle is to be applied is sufficiently new to the courts that
appellate judges are unable to anticipate and articulate
definitively what factors should be outcome determinative.”39
The situations in which this factor required more deferential
review were unclear. Our cases demonstrate that legal issues
involving situations that are completely new to the appellate
courts are rare. Furthermore, where a situation is novel, it is
not self-evident that the appellate courts should restrain
themselves from exercising searching review and should instead
take a wait-and-see approach to establishing a legal test. In
the language of the pasture metaphor, we are not convinced that
we should necessarily refrain from establishing fences that
restrain trial courts simply because a situation is novel and
anticipating the future development of the law may be difficult.
¶30 Furthermore, because this “novelty” factor was
prominent in the original Pena test despite its rare
applicability, it has often proven unwieldy, cluttered the
analysis, or been ignored. For example, we have sometimes stated
that the situation presented was not “new,” but then have
addressed whether we could articulate “outcome determinative
factors.”40 The later inquiry is substantially the same as the

40 (...continued)
¶ 47, 98 P.3d 1 (concluding that the second factor supported
additional deference to a trial court’s beneficial use
determination, even though the beneficial use doctrine “has roots
dating back to the turn of the last century”); Jeffs v. Stubbs,
970 P.2d 1234, 1245 (Utah 1998) (concluding that although “the
unjust enrichment doctrine has ancient roots,” the second factor
supported a “broad degree of discretion in applying the law”
where “the court’s ability to state clearly the outcome-
determinative factors remains elusive”).
13 No. 20050001
inquiry that we make under the first factor, namely: the degree
to which the variety and complexity of the facts make it
difficult to articulate a legal test or factors that are outcome
determinative. Such mixing of the separate analytical inquiries
from the first and second original Pena factors has only served
to confuse the analysis and may have led appellate courts to
place too much weight on the difficulty of articulating a rule.
Ultimately, in the rare instances where this “novelty of the
situation” factor may be important, it could fall under the
umbrella of other policy considerations.
¶31 In making these changes to the way that we articulate
the established standard, our intent is to improve upon our
statement of the test that we apply to mixed questions of fact
and law without changing its core substance. As before, our goal
in applying the above balancing test is to allocate tasks between
the trial and appellate courts based on their institutional roles
and competencies.
II. WHETHER A DEFENDANT WAS SUBJECTED TO CUSTODIAL
INTERROGATION IS A MIXED QUESTION OF FACT AND LAW THAT
WE REVIEW FOR CORRECTNESS.
¶32 Having set forth the balancing test to be used in
selecting an appropriate standard of review for a mixed question
of fact and law, we now apply the test to answer the question at
hand; namely, what is the standard appellate courts apply in
reviewing a trial court’s determination that a person was or was
not subjected to custodial interrogation for the purpose of Fifth
Amendment Miranda protections? To do so, we will first outline
the legal concept of custodial interrogation in the context of
our Fifth Amendment jurisprudence. We will then apply the three-
factor balancing test to the mixed question of custodial
interrogation, discussing each of the three factors in turn.

41 Rhode Island v. Innis, 446 U.S. 291, 297, 300-01 (1980);
Miranda v. Arizona, 384 U.S. 436, 444-45 (1966).
42 Innis, 446 U.S. at 298-99 (quoting Miranda v. Arizona,
384 U.S. 436, 444 (1966); accord Stansbury v. California, 511
U.S. 318, 323 (1994).
43 Berkemer v. McCarty, 468 U.S. 420, 440 (1984) (internal
quotation marks omitted); see also State v. Mirquet, 914 P.2d
1144, 1147 (Utah 1996).
44 Stansbury, 511 U.S. at 324 (internal quotation marks
omitted); accord Mirquet, 914 P.2d at 1147.
No. 20050001 14
A. The Legal Concept of Custodial Interrogation
¶33 To apply the three-factor mixed question test set forth
above, we must first understand the legal concept of custodial
interrogation, which trial courts apply to the facts of each
case. The Fifth Amendment to the United States Constitution
guarantees that a person shall not be “compelled in any criminal
case to be a witness against himself.” We protect this right by
excluding from a defendant’s criminal trial any incriminating
statement that the defendant made to police officers while under
custodial interrogation if the officers did not give a Miranda
warning.41
¶34 Generally, custodial interrogation consists of
questioning or use of other techniques of persuasion “‘initiated
by law enforcement officers after a person has been taken into
custody or otherwise deprived of his freedom of action in any
significant way.’”42 Thus, custodial interrogation occurs where
there is both (1) custody or other significant deprivation of a
suspect’s freedom and (2) interrogation. These two elements are
interrelated.
¶35 We often describe the first element as an inquiry into
whether a suspect was “in custody.” A person is in custody when
“[the person’s] freedom of action is curtailed to a degree
associated with formal arrest.”43 The inquiry is objective and
considers “how a reasonable man in the suspect’s position would
have understood his situation.”44 A suspect may understand
himself or herself to be in custody based either on physical
evidence or on the nature of the officer’s instructions and
questions. Therefore, we focus on both the evidence of restraint

45 See Berkemer, 468 U.S. at 442; Salt Lake City v. Carner,
664 P.2d 1168, 1170 (Utah 1983).
46 Stansbury, 511 U.S. at 325.
47 See Mirquet, 914 P.2d at 1148 (indicating that
accusatory questioning is relevant, but does not necessarily
establish a coercive environment); Carner, 664 P.2d at 1170
(recognizing import of accusatory statements); State v. Snyder,
860 P.2d 351, 357 (Utah Ct. App. 1993)(same).
48 664 P.2d 1168 (Utah 1983).
49 See Mirquet, 914 P.2d at 1147 n.2 (explaining that
although Carner was decided under Article I, section 12 of the
Utah Constitution, the same test applies under the Fifth
Amendment of the United States Constitution).
50 Carner, 664 P.2d at 1171.
15 No. 20050001
and on objective evidence of the officers’ intentions.45 As
stated by the U.S. Supreme Court,
[A]n officer’s views concerning the nature of
an interrogation, or beliefs concerning the
potential culpability of the individual being
questioned, may be one among many factors
that bear upon the assessment whether that
individual was in custody, but only if the
officer’s views or beliefs were somehow
manifested to the individual under
interrogation and would have affected how a
reasonable person in that position would
perceive his or her freedom to leave.46
¶36 For instance, when investigatory questioning shifts to
accusatory questioning, the existence of custody is likely
because this often indicates to the defendant that he or she is
not free to leave. By making accusations, the police officer
indicates that there are reasonable grounds to believe that a
crime has been committed and that the defendant committed it.47
In Salt Lake City v. Carner,48 we set forth four factors that aid
in determining whether a defendant is “in custody” for purposes
of the Miranda protections:49 “(1) the site of interrogation; (2)
whether the investigation focused on the accused; (3) whether the
objective indicia of arrest were present; and (4) the length and
form of interrogation.”50

51 Rhode Island v. Innis, 446 U.S. 291, 298-301 (1980).
52 Id. at 300-02.
53 Pena, 869 P.2d at 939.
54 Compare Mirquet, 914 P.2d at 1148 (noting that one
factor indicating custody was location of questioning inside the
confines of the front seat of a police car) with Berkemer v.
McCarty, 468 U.S. 420, 438 (1984) (suggesting that because the
typical traffic stop is conducted where passersby may witness the
interaction, a motorist does not feel completely at the mercy of
the police).
No. 20050001 16
¶37 Once a trial court determines that the defendant was
“in custody,” it must then decide whether the incriminating
statement was the product of interrogation.51 Interrogation is
“either express questioning or its functional equivalent” and it
incorporates any “words or actions on the part of police officers
that they should have known were reasonably likely to elicit an
incriminating response.”52
B. Application of the Balancing Test to the Mixed
Question of Custodial Interrogation
¶38 Applying our three-factor balancing test to the legal
concept of custodial interrogation, we first consider the degree
of complexity and variety in the facts that are involved in
custodial interrogation determinations. We have said that
additional deference to the trial court is warranted where the
facts “are so complex and varying that no rule adequately
addressing the relevance of all these facts can be spelled
out.”53
¶39 We initially note that the rules defining both the
custody and interrogation prongs of custodial interrogation are
well defined and adequate. We set forth the four Carner factors
for determining custody more than twenty years ago and they
continue to guide us today. Furthermore, in light of the
objective nature of both prongs of the custodial interrogation
test, the relevant facts are typically not particularly complex
and can usually be identified with specificity. The location of
the interrogation usually can be found in the record, and the
significance of the location is often intuitive. Places that are
confined or isolated are more likely to indicate custody than
those that are public and open.54 The length of the
interrogation can usually be closely approximated and compared
with the length of ordinary investigative detentions. As for

55 See State v. Wood, 868 P.2d 70, 83 (Utah 1993); Carner,
664 P.2d at 1171.
56 See Pena, 869 P.2d at 939.
57 State v. Virgin, 2006 UT 29, ¶ 28 (citing Pena, 869 P.2d
at 938-39).
17 No. 20050001
indicia of arrest, we generally look to whether handcuffs, drawn
guns, locked doors, threats, or coercion are present.55 The
question of whether the defendant was a “focus” of the
investigation depends on whether the investigators’ actions
indicated that they had identified the defendant as a likely
criminal culprit. Finally, the facts that are relevant to the
objective legal question of whether the police officers should
have known that their words or actions were likely to elicit an
incriminating response consist of the words or actions
themselves, their meaning, and their likely impact. In sum,
although the facts relating to custodial interrogation will vary
from case to case, this first factor of the balancing test weighs
against appellate deference because such facts typically are not
“so complex and varying that no rule adequately addressing the
relevance of all these facts can be spelled out.”56
¶40 Second, we consider the degree to which the application
of the legal rule relies on “facts” observed by the trial judge--
such as a witness’s appearance and demeanor--that cannot be
adequately reflected in the record available to appellate courts.
The greater the significance of these types of facts, the greater
the case for appellate deference. Although the trial court’s
superior position to make credibility determinations can be
important with respect to custodial interrogations, given the
objective nature of the test to be applied, it is typically less
important than in other contexts. The necessary facts regarding
the site, length, and focus of the interrogation and the indicia
of arrest are facts that generally can be adequately reflected in
a cold record. Similarly, facts that show that a police
officer’s statement or act is one that the officer should have
known would likely elicit an incriminating response generally can
be identified with specificity and written into the record.
Thus, this second factor of the balancing test does not weigh in
favor of granting greater discretion to the trial court.
¶41 Third, and lastly, we consider whether policy
considerations favor more or less appellate control over the
factual application of the concept of custodial interrogation.57
Because the custodial interrogation inquiry is the crux of the
test that determines when a suspect’s Fifth Amendment right must

58 2004 UT 95, ¶¶ 14-15, 103 P.3d 699.
59 Id.
60 Id.
61 Id.; see also State v. Hansen, 2002 UT 125, ¶ 25, 63
P.3d 650; State v. Warren, 2003 UT 36, ¶ 12, 78 P.3d 590.
No. 20050001 18
be protected through a Miranda warning, there is a strong
interest in promoting clarity and consistency in our state’s
jurisprudence. Clarity and consistency in our courts’
application of the Miranda protections will benefit the accused
by offering predictable constitutional protections, and it will
benefit the State by providing better guidance to the police
officers in their administration of Miranda warnings. Thus, the
third factor in our balancing test weighs strongly in favor of
nondeferential review.
¶42 Our application of the balancing test leads us to
conclude that nondeferential appellate review of custodial
interrogation determinations is mandated. Specifically, we hold
that the first two factors of the balancing test do not favor
deferential review because the facts involved in a custodial
interrogation determination are usually relatively simple, and
the custodial interrogation determination does not typically rely
heavily on credibility determinations or other subtle factual
determinations that are the prerogative of the trial court.
Moreover, even if the application of these factors made a
stronger case for deferential appellate review, they would be
outweighed by the need for uniformity in the custodial
interrogation standard.
¶43 Our reasoning on these points is consistent with that
from our recent decision in State v. Brake.58 In Brake, we
decided that we would review for correctness mixed questions of
fact and law in the context of Fourth Amendment search and
seizure cases.59 We grounded this decision in the substantial
constitutional issues at stake, determining that the variety of
fact patterns in those search and seizure cases was not
unmanageable and did not outweigh the need for uniform legal
rules.60 We concluded that, in the context of Fourth Amendment
search and seizure cases, the need for a consistent body of case
law that would set statewide standards demanded that we review
each of these determinations for correctness.61
¶44 Like the law governing Fourth Amendment protections,
custodial interrogation determinations define the boundaries of a

62 868 P.2d 70 (Utah 1993).
63 Id. at 83.
64 See State v. Teuscher, 883 P.2d 922, 929 (Utah Ct. App.
1994) (determining that a trial court’s custody determination
should be granted some deference because the inquiry is fact-
sensitive); see also State v. Riggs, 1999 UT App 271, ¶ 7, 987
P.2d 1281 (granting a measure of discretion to trial court’s
custodial interrogation determination); State v. Strausberg, 895
P.2d 831, 834 n.5 (Utah Ct. App. 1995) (granting a measure of
discretion to trial court’s custody determination “unless such
determination exceeds established legal boundaries”). But see
State v. Brandley, 972 P.2d 78, 81 (Utah Ct. App. 1998)
(reviewing for correctness whether, given the underlying facts,
defendant “was in custody for Miranda purposes”).
19 No. 20050001
substantial constitutional right--the Fifth Amendment right to
avoid self incrimination--and should be defined and applied
uniformly for the benefit of the State as well as for the benefit
of the criminal suspect. As in Brake, these concerns outweigh
countervailing factors and require nondeferential review of this
mixed question of fact and law. Indeed, the facts that we
consider in the context of a Fifth Amendment custodial
interrogation determination are generally simpler and more
manageable than the facts that go into determining search and
seizure issues under the Fourth Amendment.
¶45 Finally, we note that although we have not previously
applied an express balancing analysis to the mixed question of
custodial interrogation, our announcement of a correctness
standard of review for custodial interrogation determinations is
consistent with our prior precedent. Since we directly addressed
standards of review for mixed questions in Pena, this court has
discussed the standard of review for the mixed question of
custodial interrogation in only one case, State v. Wood.62 In
Wood, we stated that when facts are undisputed we review
custodial determinations for correctness.63 While our
correctness standard does conflict with the deferential standard
applied in some of our court of appeals’ cases,64 that court has
never expressly considered the policy implications discussed in
Brake or conducted a full balancing analysis using the factors
that we discussed in Pena and its progeny. In sum, our
application of the balancing test to the mixed question of
custodial interrogation, particularly in light of the policy
favoring uniformity in the application of Fifth Amendment Miranda
protections, leads us to conclude that we should apply a
correctness standard for such questions.

No. 20050001 20
CONCLUSION
¶46 We hold that custodial interrogation determinations
should be reviewed for correctness. We arrive at this conclusion
by applying a revised three-factor balancing test that is based
on the factors and policy considerations discussed in State v.
Pena and its progeny. First, the facts attendant to custodial
interrogation determinations are generally not so complex and
varied as to preclude the articulation of a rule. Second,
credibility determinations generally do not weigh heavily in such
determinations. And third, there is a strong policy interest in
establishing predictable standards to guide both the courts and
police officers in their administration of Fifth Amendment
Miranda protections. Having answered the question that is before
us on certiorari, we remand to the court of appeals with
directions to apply a correctness standard to the trial court’s
determination that Levin was not subjected to custodial
interrogation and for further proceedings consistent with this
opinion.
---
¶47 Chief Justice Durham, Associate Chief Justice Wilkins,
Justice Parrish, and Justice Nehring concur in Justice Durrant’s
opinion.