delivered the Opinion of the Court.
T1 In this case and two companion cases decided today, Gibbons v. People, 2014 CO 67, 328 P.3d 95, and Fain v. People, 2014 CO 69, 329 P.3d 270, we consider whether a trial court must inform a jury that a mistrial will be declared if it cannot reach a unanimous verdict when giving a modified-Allen instruction.
12 A modified-Allen instruction is a supplemental jury instruction designed to encourage, but not coerce, a deadlocked jury into reaching a unanimous verdict. To accomplish this, the instruction informs the jury that it should attempt to reach a unanimous verdict; that each juror should decide the case for himself or herself; that the jurors should not hesitate to reconsider their views; and that they should not surrender their honest convictions solely because of others' opinions or to return a verdict. See CJI-Crim. 38:14 (1983 & Supp. 1993); see also Chief Justice Directive No. 14 (1971). We approved this instruction as non-coercive in our own "Allen" case, Allen v. People, 660 P.2d 896 (Colo.1983).
13 Since approving this instruction, a line of authority has developed in the court of appeals adding another component: "In addition, the [modified-Allenr] instruction must inform the jurors that if it appears to the trial court that a unanimous decision cannot be reached, they will be excused and a mistrial will be declared." People v. Raglin, 21 P.3d 419, 423 (Colo.App.2000) (emphasis added). The court of appeals here rejected Martin's argument that Raglin required the trial court to instruct the jury about the possibility of a mistrial: "The absence of language telling jurors that a mistrial would be declared if they could not reach a verdict did not render the court's instruction-which tracked language repeatedly approved by the Colorado Supreme Court-coercive." People v. Martin, No. 06CA1509, slip op. at 8, 2011 WL 1909107 (Colo.App. May 19, 2011) (not selected for official publication).
14 In another case announced today, we rejected Raglin's mistrial advisement requirement and another division's per se rule prohibiting such advisements. Gibbons, ¶¶ 3, 26. We held that a trial court is not required to provide a mistrial advisement when giving a modified-Allen instruction. The trial court has discretion to instruct a deadlocked jury about the possibility of a mistrial when, considering the content of the instruction and the context in which it is given, the instruction will not have a coercive effect on the jury. The court should consider exercising its discretion in rare circumstances, for example when a jury has actually indicated a mistaken belief in indefinite deliberations. Id. at (1 4, 88.
T5 To apply that holding to this case, we recount the relevant facts and procedural history.
I. Facts and Procedural History
T6 Lori Martin shot her husband in the back of the head, killing him, during a dispute about her moving to another state with their daughter. The prosecution charged her with first degree murder and two crimes of violence.
T7 Martin argued that she killed her husband in self-defense. At trial, she testified that she suffered from battered women's syndrome and post-traumatic stress disorder after years of psychological abuse. Martin also claimed that her husband had admitted to killing four people and that he told her to "give him a head start" if she ever decided to reveal that information. (Police investigated the alleged murders but found no evidence of them.)
18 On the night of the shooting, Martin testified that she had told her husband "consider this your head start," insinuating that she would tell the authorities about the murders. When he approached Martin with a threatening expression, she shot and killed him.
T 9 The trial began in earnest on March 7, and lasted nearly three weeks. After about a day of deliberations, the jury told the clerk that it was "having difficulty" reaching a verdict. The court did not respond. Later that day, the jury again told the clerk that it was having trouble reaching a verdict. De*250fense counsel preferred "that the Court not get involved at this point," and the prosecution agreed. But the court expressed concern that it should respond to the jury, given that "we've been told twice now that they are deadlocked." The court then instructed the jury to "continue to deliberate, reviewing all of the evidence, and applying the instructions of law to the facts as you determine them to be." The court told the parties that, if the deadlock persisted, it would "consider the modified Allen instruction."
10 The next morning, Friday, March 24, the jury informed the court that it was in a "hopeless deadlock":
We ... have found ourselves at a hopeless deadlock and do not believe we will ever reach a unanimous decision. We believe any further deliberation would prove fruitless and counter-productive. ...
111 The prosecution argued in favor of giving a modified-Allen instruction. Defense counsel agreed that a supplemental instruetion was appropriate, given that the jury's "fruitless and counter-productive" description had "answered what [would have been] the next stage for the Court"-a reference to the practice of inquiring about the "likelihood of progress towards a unanimous verdict" before giving a modified-Allen instruction. See People v. Schwartz, 678 P.2d 1000, 1012 (Colo.1984) (stating that the trial court should "determine whether there is a likelihood of progress towards a unanimous verdict" before giving a modified-Allen instruction). Without objection, the court then read a modified-Allen instruction that substantially tracked the pattern instruction.1 See CJI-Crim. 38:14.
T12 The jury continued deliberating and then told the court it had reached a verdict. A few minutes past 5:00 p.m., the jury found Martin guilty of second degree murder-heat of passion. Although the foreperson had signed the verdict form, when polled, she said that the verdiet was not hers.
1 13 The court then provided another supplemental instruction that tracked CJI-Crim. 88:15, Colorado's pattern instruction directing the jury to continue deliberations after polling indicates that the jury did not reach a unanimous verdict:
In the polling of one of the jury members, one of your members made an answer which indicates that you may not have reached a unanimous verdict.
For this reason, the Court will be asking you to return to the jury room for further reconsideration of your verdict.
Whenever you have reached a unanimous verdict, you may return it to the Court. If you are not unanimous, then, you should continue your deliberations.
After you return to the jury room, any member is free to change his or her vote on any issue submitted.
1 14 The jury continued deliberations, and defense counsel expressed concern that the foreperson's resolve would be "beaten down by the rest of the jurors." She suggested "further instruction" but did not elaborate or provide specific language. Around 5:45 p.m., or about a half hour after the jury's first verdict, the jury returned the same verdict it had returned earlier, finding Martin guilty of second degree murder-heat of passion.
€15 On appeal, Martin argued that the trial court erred by giving the deadlocked jury the two instructions to continue deliberations. As pertinent here, Martin contended that the trial court erred by failing to give a mistrial advisement with the modified-Allen *251instruction, per Raglin. The court of appeals recognized that Colorado's "Allen" case "requires the district court to discharge the jury if it appears there is no reasonable probability of agreement," but "it does not require the court to expressly inform the jury of this possibility." Martin, slip op. at 7. It then concluded that the trial court did not coerce the jury into rendering a compromise verdict by failing to give a mistrial advisement. Id. at 8.
{16 Martin sought certiorari, and we agreed to review whether Raglin's mistrial advisement requirement is consistent with our precedent.2 For the reasons stated in Gibbons, ¶¶ 23-27, we conclude that it is not.
117 We hold that a trial court is not required to provide a mistrial advisement when giving a modified-Allen instruction. The trial court has discretion to instruct a deadlocked jury about the possibility of a mistrial when, considering the content of the instruction and the context in which it is given, the instruction will not have a coercive effect on the jury. The court should consider exercising its discretion in rare circumstances, for example when a jury has actually indicated a mistaken belief in indefinite deliberations.
II. Application
T18 Martin argues that the trial court erred by failing to provide a mistrial advisement and that its failure to do so coerced the jury into rendering a compromise verdict3
119 This case differs from Gibbons and Fain. In those cases, the juries returned verdicts soon after the trial court gave modified-Allen instructions. Here, by contrast, the jury returned a verdict, but, when polled, the foreperson said she disagreed with it. The trial court then instructed the jury to deliberate for "further reconsideration of your verdict," in accordance with Colorado's pattern instructions. In addition, the trial court told the jury that, "Iilf you are not unanimous, then, you should continue your deliberations."
{20 Martin contends that the court's see-ond supplemental instruction was "particularly coercive" because it gave the jury only two options: return a unanimous verdict or continue deliberating. Martin points out that the second instruction did not reiterate or reference the modified-Allen instruction, essentially supplanting it. And, because the jury appeared to contain a lone holdout after polling, Martin argues that the instruction effectively singled out that juror for admonition.
121 The second instruction did not supplant the court's modified-Allen instruction; it supplemented it. "We consider all of the instructions given by the trial court together to determine whether they properly informed the jury." Riley v. People, 266 P.3d 1089, 1093 (Colo.2011). We thus read the court's second supplemental instruction in conjunction with its first to mean that the jury should continue to deliberate "for further reconsideration of your verdict," but that the jurors should not "surrender your honest conviction ... solely because of the opinion of your fellow jurors, or for the mere purpose of returning a verdict."
122 For the same reason, we reject Martin's argument that the second supplemental instruction singled out the lone holdout for admonition. For one, the instruction stated that "any member is free to change his or her vote on any issue submitted." (Emphasis added.) The modified-Allen instruction was likewise directed to all of the jurors, and it made clear that the jurors should "reexamine [their] own views" but not "surrender [their] honest conviction[s]." We *252presume that the jurors understood and followed the trial court's instructions. See Domingo-Gomez v. People, 125 P.3d 1043, 1053 (Colo.2005).
123 We recognize that the foreperson's disagreement during polling distinguishes this case from Gibbons and Fain, but we disagree with Martin that that distinction requires a different result. Rather than show a propensity to be "beaten down" by her fellow jurors, the foreperson's resolve during polling demonstrated a willingness to stand by her then-held convictions. And, when polled a second time, the verdict was unanimous, suggesting that the foreperson reexamined her views, as the modified-Allen charge instructs.
4 24 Martin also calls attention to the fact that the jury rendered its verdict after 5:00 p.m. on Friday, March 24-the date on which the trial court supposedly informed the jury the trial should end.4
A court may not impose a deadline on deliberations that prevents the jury from reaching a well-considered verdict. Allen, 660 P.2d at 898. And we recognize, as did the court of appeals, that discussing scheduling pressures with the jury may be coercive if those discussions effectively impose a deadline for the jury to end its deliberations with a verdict or have a mistrial declared. See Key v. People, 865 P.2d 822, 825 (Colo.1994) (citing Allen, 660 P.2d at 898-99); see also People v. Urrutia, 893 P.2d 1338, 1343 (Colo.App.1994) ("Discussing scheduling problems with the jury may ... be coercive if those scheduling problems ere-ate an impression that the jury is under a short time limit to reach a verdict.").
126 There is no record support for Martin's contention that the trial court ever "discussed" scheduling problems with the jury. Nor do we interpret informing the jury about the anticipated length of trial as imposing a deadline or as coercive. The court merely informed the venire that the trial could last three weeks, presumably so venire members could raise scheduling conflicts or other issues to the court. Even defense counsel during jury selection recognized that, "[slince it's such a lengthy trial, there might be an exorbitant amount of hardships." And we note that defense counsel requested that the trial begin on March 7 to accommodate her (and the trial judge's) scheduled vacations.
T27 But in Martin's view, this confluence of cireumstances-the jury's difficulty reaching a unanimous verdict, the foreperson's disagreement during polling, and the fact that the jury returned a verdict after 5:00 p.m. on a Friday-rendered her case "one in which the court was required to tell the jury it would be discharged at some point if it could not reach a unanimous verdiet."
1 28 But as our holding makes clear, a trial court is not required to give a mistrial advisement. Instead, it should consider exercising its discretion to do so in rare cireum-stances, for example when a jury has actually indicated a mistaken belief in indefinite deliberations. The trial court here could not have erred, let alone plainly erred. So for that reason alone, Martin's argument fails.
129 And, in any event, we disagree with Martin that the jury was "laboring under the basic misapprehension that the trial cannot end without a verdict" The trial lasted nearly three weeks, involved voluminous evidence, and raised difficult issues about battered women's syndrome, post-traumatic stress disorder, and self-defense. That the jury appeared to struggle with these issues is not surprising.
30 We conclude that the trial court did not err by failing to instruct the jury about the possibility of a mistrial.
III. Conclusion
31 For the reasons stated, we affirm the judgment of the court of appeals.
*253JUSTICE COATS concurs in part and concurs in the judgment, and CHIEF JUSTICE RICE and JUSTICE EID join in the concurrence in part and concurrence in the judgment.. In its oral instruction, the court erroneously told the jurors "to consult with one another and to deliberate with a view to reaching a verdict, if you can do so without violence to each other." The written instruction, however, properly told the jurors "to consult with one another and to deliberate with a view to reaching a verdict, if you can do so without violence to individual judgment." (Emphasis added.)
Martin argues that the trial court's oral instruction suggested to the jurors that "they had no choice but to continue deliberations until they reached a verdict, so long as the deliberations did not result in a physical confrontation."
We question that premise. No juror would think that a trial judge would keep the jury captive in a deliberation room unless a danger of physical violence arose. And Martin did not object to the oral instruction at trial, suggesting that it was not as noticeable or confusing as she makes it out to be on appeal. The jury had available the correct written instruction during deliberations. Under these circumstances, the trial court's oral mistake did not amount to reversible error.
. We granted certiorari in this case, as well as in Gibbons v. People, 2014 CO 67, 328 P.3d 95, and Fain v. People, 2014 CO 69, 329 P.3d 270, to consider the following issue:
Whether Colorado's modified-Affen instruction requires a trial court to inform the jury that if it cannot reach a unanimous verdict then the jury will be dismissed and a mistrial will be declared.
. The parties disagree about whether Martin preserved her contentions by objecting in the trial court. Martin argues that she did, so that we should review any errors under the constitutional harmless error standard. The People argue that plain error applies. We need not resolve this dispute because we conclude that the trial court did not err, let alone plainly err.
. At a pretrial conference to discuss jury selection, the court agreed that "three weeks should be more than enough" time to try the case, and it noted that was "not trying to impose limits." The court suggested telling the venire that the "case is set to commence on March the 7th, to conclude on March the 24th." But the parties do not refer us to any point in the record during which the trial court actually told the venire about the March 24 date or discussed scheduling concerns in the jury's presence, nor have we found any.