A jury convicted Appellant Clayton Jerrod Ellington of murdering his wife, Berna Ellington, and their twin two-year-old sons, *110Cameron and Christian.1 The jury found two statutory aggravating circumstances related to each of the three murders and recommended three death sentences, which the trial court imposed. For the reasons set forth below, we affirm Ellington’s convictions. As to his death sentences, however, as discussed at length in Division 7 below, we hold that the trial court abused its discretion in prohibiting Ellington from asking prospective jurors in voir dire whether they would consider all three sentencing options (death, life without parole, and life with the possibility of parole) in a case involving the murder of young children, where that was clearly a critical fact in this case, as shown by, among other things, the responses of prospective jurors who knew or inferred that fact from other sources and by the way the prosecutor tried and argued the case. We cannot say that this error was harmless, so we must reverse Ellington’s death sentences and remand the case for further proceedings.
1. Viewed in the light most favorable to the jury’s verdict, the evidence presented at trial showed the following. Ellington was married to Berna Ellington, but for several months he had been having an affair with another woman. Ellington frequently spent the night with his mistress, who knew that Ellington was married but was led to believe that he had separated from his wife, was filing for divorce, and was living with a male roommate. Ellington told his mistress that he loved her and wanted to marry her, and he asked about moving in with her, but she said she was not ready for that. The weekend before the murders, Ellington and his mistress had an argument after he appeared at her house in his wife’s car. The mistress informed Ellington that she was “calling it off.” The couple remained on speaking terms, and Ellington spent some more nights at the mistress’s house, but she told him that he had to choose between her and his wife.
*111On the night of the murders, Ellington had plans to meet his mistress after she got off work. At about 7:15 p.m., he called her to see if she could get off early. At about 9:00 p.m., Ellington’s wife spoke with her sister on the phone. She did not sound like herself and said she had to hang up but would call back; she never did. Around 10:45 p.m., Ellington called his mistress again to say that he might need to travel to Augusta to help the man he claimed was his “roommate,” who was actually in Washington, D.C. at the time. About 15 minutes later, Ellington arrived at the house of a friend, Sean Fennell, as they had previously planned to watch a basketball game. While still sitting in his pickup truck, Ellington asked Fennell for a pair of shoes and a trash bag, which Fennell got for him. Ellington then drove with Fennell to the mistress’s house to watch the game. After watching the game for only a few minutes, Ellington told Fennell that he had sent a text to his wife, that she had not replied, and that he was concerned about her.
Ellington drove to his house with Fennell and told Fennell to wait in the truck while he went inside. As he entered the house through the garage door, Ellington took off the shoes he had borrowed from Fennell. After several minutes, Ellington ran out the front door, screaming for Fennell to come inside. In the foyer on the ground floor, Berna Ellington lay facedown, dead in a pool of blood. Blood spatters and smears covered the foyer and the stairway. Ellington’s twin boys, who had just turned two years old, lay dead in their blood-stained cribs upstairs. As Fennell and Ellington went up the stairs at about 11:45 p.m., Ellington called 911 and began screaming hysterically, but he refused the 911 operator’s request to put Fennell on the line, exclaiming, “You talk to me; you don’t talk to nobody else!”
At the scene, Ellington repeatedly said, “they killed them.” He agreed to have the officers drive him to the police station for an interview. At the police station and later at the jail, in a series of statements discussed in detail in Division 2 below, Ellington initially denied any involvement in the murders, but he later claimed that he walked in and found his wife beating the children with a hammer, that he took the hammer from her, and that he then beat her with the hammer to show her how it felt.
Forensic analysis of the crime scene and autopsies of the victims showed the following. Berna Ellington likely was beaten to death with both ends of a hammer, with most of her wounds inflicted with the claw-end. The attack began while she lay in bed; as she tried to flee, she was beaten as she cowered or fell near the bed, as she went down the stairs, and while upright in the foyer. She then fell to the floor, where she was beaten further while helpless; she then had a plastic bag placed over her head, perhaps to suffocate her and hasten *112her inevitable death. She had a large number of wounds spread about her head, neck, shoulders, arms, and hands, and her skull had been fractured in so many places that it was impossible to determine the number of individual wounds that she had suffered.
The twin boys had been beaten to death in their cribs, likely with the claw-end of a hammer. They were both found lying face up with their skulls fractured numerous times. There were impact blood spatter stains on the wall by the twins’ cribs, inside their cribs, and on the floor of their room, but all 41 of the blood samples taken from Berna Ellington’s camisole matched only the DNA from her own blood. The investigation revealed blood drops leading to the shower and blood residue inside the shower, indicating that the assailant had cleaned up there after the murders. Apart from the evidence of the triple murder, the house appeared undisturbed.
When viewed in the light most favorable to the verdict, the evidence presented at trial and summarized above was sufficient to authorize a rational jury to find Ellington guilty beyond a reasonable doubt of the crimes for which he was convicted. See Jackson v. Virginia, 443 U. S. 307, 319 (99 SC 2781, 61 LE2d 560) (1979). See also Vega v. State, 285 Ga. 32, 33 (673 SE2d 223) (2009) (“ ‘It was for the jury to determine the credibility of the witnesses and to resolve any conflicts or inconsistencies in the evidence.’ ” (citation omitted)); UAP IV (B) (2) (providing that, in all death penalty cases, this Court will determine whether the verdicts are supported by the evidence). We also conclude that the evidence presented at trial was sufficient to support the jury’s finding of both statutory aggravating circumstances alleged • — • (1) that each murder was committed during the commission of another capital felony (another murder) and (2) that each murder was outrageously or wantonly vile, horrible, or inhuman in that it involved torture, depravity of mind, or an aggravated battery to the victim. See OCGA § 17-10-30 (b) (2), (7); Ring v. Arizona, 536 U. S. 584, 609 (122 SC 2428, 153 LE2d 556) (2002) (holding that statutory aggravating circumstances must be proven to a jury beyond a reasonable doubt).
Evidence Suppression Issue
2. Ellington contends that statements that he made while in jail were improperly admitted at his trial. We recount Ellington’s treatment by and statements to the police after the murders to put the disputed statements in context.
After Ellington returned to his home, where his wife and twin sons lay dead, he called 911 at approximately 11:30 p.m. Upon arriving, officers observed him “screaming and crying” and saying *113“they killed them.” The admission of Ellington’s statements at the crime scene is not in dispute here. Ellington was not touched or handcuffed by the officers, and he willingly agreed to travel in a police car to a police station to be interviewed by detectives. On the way to the police station, Ellington spoke on his cell phone, apparently to relatives, saying “someone killed them.” The admission of these overheard statements also is not in dispute here.
At the station, Ellington was taken to a small interview room, where he waited until Detectives Mullner and Farmer arrived to begin the interview at about 12:30 or 1:00 a.m. According to Detective Farmer, Ellington seemed “morose” and sometimes “agitated” during the interview, but he was lucid and cooperative. He was never handcuffed, restrained, or told that he was under arrest. The trial court concluded that the statements made during the first portion of the interview by Detectives Mullner and Farmer, which were memorialized in writing by the detectives, were admissible because they were made voluntarily and before Ellington was in custody; Ellington does not challenge that ruling on appeal. However, the trial court concluded that Ellington was in custody at the time that Detectives Mullner and Farmer had him sign his memorialized statements, and, because he had not been given Miranda warnings, see Miranda v. Arizona, 384 U. S. 436 (86 SC 1602, 16 LE2d 694) (1966), the trial court suppressed the fact of his signing the document, his signature on the document, and the video recording of his interview that began as he signed the document.
At the time of a shift change at the police station at approximately 7:30 a.m., Detectives Mullner and Farmer ended their interview. At this point, Ellington had been in the presence of various officers at the crime scene and at the station for about eight hours and, based on the trial court’s ruling, he had been in custody for at least five hours. The court also found that, at this point, Ellington had likely not slept in 24 hours.
Nevertheless, two new officers, Sergeant Germano and Detective Bryant, continued Ellington’s interview. The trial court suppressed the first part of this interview because Ellington still had not been advised of his Miranda rights. Because it is relevant to our analysis of Ellington’s later statements at the jail, we note that these suppressed statements included his claim that his wife hammered the children to death and that he then hit her with the hammer to show her how it felt.
The trial court further concluded, based on the circumstances of the interview and the officers’ “confusing presentation” of the Miranda warnings, that Ellington never made a knowing and voluntary waiver of his Miranda rights, which he was not even read until about *1148:20 a.m. after he was formally arrested; the court suppressed Ellington’s subsequent statements to Sergeant Germano and Detective Bryant on that ground. Moreover, the court found that, under the totality of the circumstances, Ellington’s statements during this continued interview were involuntary and therefore were inadmissible under statutory and constitutional standards, citing OCGA § 24-3-50 and Arizona v. Fulminante, 499 U. S. 279, 285 (111 SC 1246, 113 LE2d 302) (1991).
The trial court made specific findings of fact that the police officer made promises of leniency to Ellington, that he was held incommunicado in a small room and “backed into a small corner of that room for hours,” that he had been awake for nearly 30 hours, and that officers employed “trickery” that “was intended to produce an untrue statement” from him. The court therefore suppressed all of Ellington’s statements to Sergeant Germano and Detective Bryant for all purposes. The court also suppressed the physical evidence discovered in a garbage bag to which Ellington had directed the police during this portion of the interview; according to the trial court’s order, the bag included “a bloody T-shirt, a green towel, two sets of rubber gloves, a pair of white shoes, a hammer and some hair.” Compare United States v. Patane, 542 U. S. 630, 633-634 (124 SC 2620, 159 LE2d 667) (2004) (holding that physical evidence discovered as a result of a voluntary statement obtained in violation of Miranda may be admissible); Harris v. New York, 401 U. S. 222 (91 SC 643,28 LE2d 1) (1971) (holding that a defendant’s voluntary statement obtained in violation of Miranda may be admitted solely for the purpose of impeaching the defendant’s own testimony at trial). The State did not appeal any ofthese suppression rulings before trial, see OCGA § 5-7-1 (a) (4), or seek interim review, see OCGA §§ 17-10-35.1, 17-10-35.2, nor has the State challenged them in this appeal as a basis for contesting the trial court’s other rulings. We therefore treat these rulings as sound.
After Sergeant Germano and Detective Bryant concluded their interview at about 2:12 p.m., Ellington was allowed to see his mother and was then taken to the jail. At about 8:00 p.m., nearly six hours after his interrogation ended, Ellington saw one of the officers who had first interviewed him, Detective Mullner, as she passed through the jail on her way to interview another inmate about an unrelated case. Ellington asked a jail deputy to bring Detective Mullner to his cell to speak to him. Detective Mullner found another detective to come with her as a witness and went to Ellington’s cell. There she simply told Ellington that she understood that he wanted to speak to her. Ellington said that he had previously lied to Detective Mullner and that the truth was that he had walked in on his wife as she was *115hammering their twins to death, had taken the hammer from her, and had beaten her to death with the hammer to show her how it felt. Detective Mullner then told Ellington that she was “glad that [he] got it off his chest” and that she would be willing to speak to him again in the future.
Ellington’s statements to Detective Mullner at the jail were essentially the same as the suppressed statements that he had made earlier to Sergeant Germano and Detective Bryant at the police station, and Ellington argues that the trial court erred in declining to suppress the jailhouse statements to Detective Mullner as well. As Ellington acknowledges, however, his statements to Detective Mull-ner at the jail were admissible if there was a “sufficiently isolating break in the stream of events” between them and his earlier statements that the trial court found involuntary. Leon v. Wainwright, 734 F2d 770, 772 (11th Cir. 1984). As the United States Supreme Court has explained,
[w]hen a prior statement is actually coerced, the time that passes between confessions, the change in place of interrogations, and the change in identity of the interrogators all bear on whether that coercion has carried over to the second confession.
Oregon v. Elstad, 470 U. S. 298, 310 (105 SC 1285, 84 LE2d 222) (1985). In reviewing this issue, we defer to the trial court’s findings of fact, but we apply the law to those facts de novo. See State v. Woods, 280 Ga. 758, 758-759 (632 SE2d 654) (2006).
The trial court found that Ellington’s jailhouse statements were “[voluntary, spontaneous outbursts,” that they were made in an encounter that he initiated, that they were not made in response to any questioning, that they were made several hours after his interrogation by Sergeant Germano and Detective Bryant ended, that they were made in a different location than that interrogation, and that they were made to different officers outside the presence of Sergeant Germano and Detective Bryant. These findings of fact are fully supported by the record, and we agree with the trial court that they demonstrate a sufficient break from Ellington’s earlier, involuntary statements to render his jailhouse statements voluntary and admissible. We note in particular that Ellington himself summoned Detective Mullner and made his statements without any prompting from her.
Moreover, even if we were to assume that Ellington’s jailhouse statements were improperly admitted, the error would be harmless beyond a reasonable doubt and therefore would not require reversal. *116See Fulminante, 499 U. S. at 306-312 (holding that harmless error review applies to trial court error in admitting an involuntary statement); Chapman v. California, 386 U. S. 18, 24 (87 SC 824, 17 LE2d 705) (1967) (holding that, “before a federal constitutional error can be held harmless, the court must be able to declare a belief that it was harmless beyond a reasonable doubt”). The jailhouse statements were somewhat self-serving in that they cast Ellington as having acted impulsively in killing his wife only upon seeing her murdering the children, and a similar version of events was presented to the jury through a letter that Ellington had written from jail to his mistress about a month and a half after the killings, the admissibility of which is not questioned on appeal.
Pretrial Challenges to the Death Penalty Scheme
3. Ellington contends that Georgia’s death penalty statutes are unconstitutional for a number of reasons. Although we are reversing the death sentences the jury imposed, see Division 7 below, these challenges go to the issue of whether the State can even seek the death penalty in this case and therefore are appropriate to decide at this time. None of these arguments has merit.
(a) Ellington contends that Georgia’s statutes do not provide sufficient guidance to juries in considering possible death sentences and that this deficiency leads to arbitrary results. However, the Georgia death penalty scheme sufficiently narrows the application of the death penalty and guides the jury’s consideration of it as a possible sentence, while also affording jurors the latitude to consider all mitigating circumstances in their deliberations. See Ledford v. State, 289 Ga. 70, 75 (709 SE2d 239) (2011) (citing Gregg v. Georgia, 428 U. S. 153 (96 SC 2909, 49 LE2d 859) (1976)); Arrington v. State, 286 Ga. 335, 336-337 (687 SE2d 438) (2009). Moreover, as discussed in Division 11 below, in this case the trial court properly charged the jury on mitigating circumstances.
(b) Ellington’s equal protection claim, which alleges that Georgia’s laws lead to the discriminatory application of the death penalty, fails because he has not shown any invidious discrimination in his case. See Ledford, 289 Ga. at 75 (citing McClesky v. Kemp, 481 U. S. 279 (107 SC 1756, 95 LE2d 262) (1987)).
(c) Georgia’s death penalty statutes are not unconstitutional because they allow prosecutors discretion in selecting cases in which to seek the death penalty. See Walker v. State, 281 Ga. 157, 161 (635 SE2d 740) (2006).
(d) Georgia’s death penalty statutes are not unconstitutional because they require the jury to find the existence of at least one *117statutory aggravating circumstance unanimously and beyond a reasonable doubt before a death sentence may be considered but provide no particular standard of proof to be applied to an individual juror’s consideration of non-statutory aggravating circumstances. Unlike statutory aggravating circumstances, the finding of a non-statutory aggravating circumstance does not increase the defendant’s maximum potential punishment, and thus it is not “ ‘the functional equivalent of an element of a greater offense’ ” for constitutional purposes. Ring, 536 U. S. at 609 (quoting Apprendi v. New Jersey, 530 U. S. 466, 494, n. 19 (120 SC 2348, 147 LE2d 435) (2000)). See also Jones v. State, 282 Ga. 784, 791 (653 SE2d 456) (2007) (noting that statutory aggravating circumstances must be regarded as “elements” for purposes of the United States Constitution but holding that they remain merely sentencing factors in matters affecting only Georgia law).
(e) This Court’s proportionality review is not inadequate under statutory or constitutional standards. See, e.g., Ledford, 289 Ga. at 75; Gissendaner v. State, 272 Ga. 704, 716 (532 SE2d 677) (2000).
(f) With regard to the claims addressed in subdivisions (a), (c),
and (e) above, Ellington argues that the trial court erred by refusing to conduct an evidentiary hearing. However, he has failed to provide any record citations showing that he alleged facts in the trial court that would have supported these claims if proved. See Dick v. State, 248 Ga. 898, 899 (287 SE2d 11) (1982) (addressing when a hearing is required in the context of an extraordinary motion for a new trial). Compare Allen v. Allen, 218 Ga. 364, 366 (127 SE2d 902) (1962) (reversing in a divorce case where the appellant was denied the right to present evidence which, if proved, would have shown her motion to be meritorious). Furthermore, we have not identified any such factual allegations in our own review of the record. Indeed, the citations that Ellington does provide suggest that the only arguably relevant motions at issue were never placed in the trial record. See Kegler v. State, 267 Ga. 147, 148 (475 SE2d 593) (1996) (holding that where appellant fails to ensure that a complete record of what transpired in the trial court is before this Court, there is nothing for us to review). This enumeration is therefore unsupported.
Jury Composition Issues
4. The method that Georgia jury commissions used to form the lists from which grand and trial jurors were summoned at the time that Ellington’s grand and traverse jury pools were created involved matching the percentages of certain identifiable demographic groups to the percentages of those groups in the most recent decennial *118census. This method, which is referred to as “forced balancing,” was effectively required under the then-applicable version of the Unified Appeal Procedure, the set of rules promulgated by this Court to guide the trial courts in the administration of death penalty cases. See Williams v. State, 287 Ga. 735, 735, n. 1 (699 SE2d 25) (2010) (describing how forced balancing grew out of the Unified Appeal Procedure); OCGA § 17-10-36 (a) (directing this Court to “establish ... a new unified review procedure”).2
Ellington argues that forced balancing is an unconstitutional violation of equal protection of the law, because it employs race-based classifications. However, as we have repeatedly held, Georgia’s former practice of forced balancing is not unconstitutional, as it was adopted in an attempt to promote the fair representation in jury lists of various identifiable groups and thus served a significant state interest. See, e.g., Williams, 287 Ga. at 735.
5. Ellington argues that the pool of jurors from which his grand jury was selected under-represented various groups of persons, including women, African-American persons, and Hispanic persons.3 This Court has held that the mandate in the then-applicable version of the Unified Appeal Procedure to judge the composition of jury pools in reference to the most recent decennial census rather than more recent Census Bureau estimates is not unconstitutional. See Ramirez v. State, 276 Ga. 158, 162 (575 SE2d 462) (2003). We have also held that the most relevant comparison under the former jury selection system involves the data from the decennial census indicating the citizen population of various groups (as opposed to all residents), because only citizens are eligible to serve as jurors. See Smith v. State, 275 Ga. 715, 721-722 (571 SE2d 740) (2002).
Ellington’s expert testimony showed that the absolute disparity between the percentages of the disputed groups in the grand jury pool and the percentages of those groups in the citizen population reflected in the 2000 census was as follows: -1.84 percentage points for women; -4.28 percentage points for African-American persons; and -0.09 percentage points for Hispanic persons. These small disparities do not violate constitutional due process or fair cross-section limits or statutory requirements, nor do they even violate the prophylactic limit set forth in the Unified Appeal Procedure. See Edwards v. State, *119281 Ga. 108, 109-110 (636 SE2d 508) (2006) (noting that the five percentage point limit imposed by the UAP is a prophylactic standard and holding that this Court lacks the constitutional power to reverse a conviction where the under-representation of a cognizable group in a grand jury pool does not rise to the level of a constitutional violation); Ramirez, 276 Ga. at 161 (noting that grand jury composition challenges under the former version of OCGA § 15-12-40 are subject to “standards comparable if not identical to federal constitutional standards”); Morrow v. State, 272 Ga. 691, 692 (532 SE2d 78) (2000) (explaining that an under-representation of less than ten percentage points is “usually constitutional” and that a statutory violation exists only where there is “a wide absolute disparity”).
6. Ellington argues that the list from which his prospective traverse jurors were summoned violated Georgia statutory law, the Unified Appeal Procedure, and the United States Constitution. We see no reversible error.
(a) As explained in Division 3 above with regard to the grand jury list, the process of forced balancing, whereby the percentages of cognizable groups of persons on jury lists were made to match the percentages of those groups in the most recent decennial census, is not unconstitutional. See, e.g., Williams, 287 Ga. at 735.
(b) Ellington argues that the trial court erred by failing to order the traverse jury list modified in light of evidence that he presented showing that, when properly compared to the statistics for various groups of citizens in the most recent decennial census, women were under-represented by 5.4 percentage points and African-American persons were under-representedby 5.57 percentage points. See Ramirez, 276 Ga. at 162 (holding that the most recent decennial census is the appropriate benchmark for a jury composition claim); Smith, 275 Ga. at 721-722 (holding that statistics regarding citizens are the most relevant to a jury composition claim). As explained in Division 4 above, this degree of under-representation is not unconstitutional and does not violate Georgia statutory law. However, section II (C) (6) (b) of the version of the Unified Appeal Procedure that was in effect at the time of Ellington’s trial directed the trial court to correct any under-representation greater than five percentage points on the traverse jury list prior to trial. The trial court denied Ellington’s challenge to the composition of the traverse jury list, saying that the UAP’s five percentage point limit was merely a “prophylactic rule and [an] aspirational goal” and citing Edwards.
We disapprove of this reasoning, because “despite being prophylactic in nature, the UAP is clear in its mandate to the trial courts that ‘(s)ignificant under-representation of any (cognizable) group on either jury list shall be corrected prior to trial.’ ” Edwards, 281 Ga. at 110 *120(emphasis in original) (quoting former UAP II (C) (6) (b)). Our authority to remedy this issue is unsettled. We have held that this Court lacks the constitutional power to direct a trial court through the Unified Appeal Procedure to quash an indictment that was issued in compliance with Georgia statutory law, the Georgia Constitution, and the United States Constitution. See id. However, we specifically left open the question of whether requiring a trial court to prospectively correct a traverse jury list pretrial would also exceed our constitutional authority. See id. at 111. In this case, we again need not decide that constitutional question, because in any event we conclude that it is highly probable that the trial court’s minor violation of the Unified Appeal Procedure, which is Rule 34 of the Uniform Superior Court Rules, did not contribute to the jury’s guilty and sentencing verdicts. See Barker v. State, 263 Ga. 746, 747-748 (438 SE2d 625) (1994) (holding that an assumed violation of the Uniform Superior Court Rules was harmless).
(c) Ellington argues that the list from which his traverse jurors were summoned violated Georgia law and the United States Constitution because the list included no persons under the age of 20. This does appear anomalous, and the record does not explain how it came about; if it appears that this situation has continued, the DeKalb County jury commission should examine it. Nevertheless, this claim does not require reversal, because Ellington made no effort to prove that persons under 20 years old were a cognizable group in DeKalb County at the time of his trial. See Jackson v. State, 270 Ga. 494, 497 (512 SE2d 241) (1999) (“Whether an age group is a cognizable group depends on the time and location of the trial.”).
(d) Finally, Ellington argues that Hispanic persons were underrepresented on his traverse jury list. However, he neither presented evidence of the fact nor obtained the State’s stipulation to the fact that Hispanic persons were a cognizable group in DeKalb County at the time of his trial. See id. Compare Smith, 275 Ga. at 718 (“The trial court did not err by finding that Hall County Hispanics are a distinctive group.”). In any event, the degree of under-representation, only 1.10 percentage points, would not warrant relief. See Hum-phreys v. State, 287 Ga. 63, 68-69 (694 SE2d 316) (2010) (holding that an under-representation of Hispanic persons of less than five percentage points was not unconstitutional); Morrow, 272 Ga. at 695 (holding that an under-representation of Hispanic persons of 3.8 percentage points did not violate constitutional or statutory law).
*121 Jury Selection Issues
7. Ellington argues that the trial court erred in precluding voir dire questioning of prospective jurors as to whether they would automatically impose the death penalty, as opposed to fairly considering all three sentencing options, in a case involving the murder of young children. Under all the circumstances, and cautioning that our holding is limited in scope, we agree, and we therefore conclude that Ellington’s death sentences must be reversed, although his convictions can be affirmed.
(a) About four months before trial, the State filed a motion in limine, asking the trial court to broadly “prohibit the defendant from presenting or asking questions which outline real or hypothetical facts ... or require a juror to specify which sentence would be appropriate in this case.” In a hearing held on that motion, Ellington argued that he should be allowed to ask some “case-specific” questions, including the following question patterned on a decision by a federal district court: “Could you fairly consider a life sentence if the evidence showed that he killed his kids?” See United States v. Johnson, 366 FSupp.2d 822, 849 (N.D. Iowa 2005). The prosecutor strongly objected. The trial court acknowledged that some prospective jurors would find the murder of children to be “a more troubling situation” in which to consider a life sentence, but it barred Ellington from informing prospective jurors of the fact that the victims in this case included two young children. The court said that informing the jurors of that fact would invite “prejudicial prejudging” of the case but added, “If I’m wrong, obviously, you’ve got tremendous grounds for appeal if there’s a conviction.”
The court later directed the parties to submit their proposed questions for the juror questionnaire and for voir dire. Question 12 of Ellington’s proposal for a section of the questionnaire addressing the death penalty asked, “Do you believe that a life sentence can be severe enough punishment for a person convicted of the deliberate killing of a child?” The State re-filed its motion in limine.
A subsequent hearing on jury selection issues did not focus specifically on Ellington’s Question 12, instead addressing the general issue of whether the prospective jurors could be questioned regarding their impartiality with respect to a murder involving child victims. The trial court reminded Ellington of its prior ruling, saying: “I mean you are aware that you cannot bring up the fact that two of these deceased people are children in voir dire.” Ellington later submitted a brief on the issue, in which he argued extensively on both constitutional and statutory grounds that he should be allowed to ask key “case-specific” questions during voir dire, quoting the following *122pattern question approved in Johnson, 366 FSupp.2d at 849: “Could you fairly consider a death sentence if the evidence showed x.” The trial court then issued a written order forbidding voir dire questions similar to Ellington’s proposed Question 12, explaining that the court would not allow questions that would require prospective jurors “to make judgments about the purported facts of the case before the proper presentation of evidence.”
At the beginning of voir dire, the trial court noted, “[T]here’s no doubt that certain jurors may have a special sensitivity to victims of child murder which would make them unable to perform their duties as jurors regardless of the facts of the case and the instructions of the Court.” Nevertheless, instead of allowing voir dire questions that specifically addressed the jurors’ views regarding the sentencing options in a case involving young children as murder victims, the court ruled that it would ask the jurors only if they could consider all sentencing options regardless of “the status of the victim” and then allow additional voir dire if any “special sensitivities” were to “bubble up.” The court again added, “If I’m wrong and he’s convicted, you’ve got an automatic reversal.”
The voir dire process was extensive. More than 200 prospective jurors answered a lengthy questionnaire and then were questioned over the course of more than three weeks; the voir dire spans more than 3,300 pages of the trial transcript. The trial court read the indictment to the prospective jurors at the outset of voir dire, and the court asked each juror general questions about their ability to be fair and impartial and to consider all three sentencing options.
However, based on the trial court’s rulings, the prospective jurors were never informed that the case involved two child victims, nor did the court or the parties pose any general questions related to that issue. Several jurors raised the issue of child victims on their own, based on pretrial publicity about the case or deductions from the indictment, noting that all three victims had the same last name. For at least four of the jurors who were aware or surmised that the alleged murders were of children, the issue affected their professed ability to be impartial in deciding between the sentencing options, and those jurors were excused for cause.4 For example, Juror 4 indicated in his juror questionnaire that he had some religious concerns regarding the death penalty in general. However, throughout his voir dire he gave responses explaining his views on the special situation of the murder of a child, explaining, “But if a child was murdered... I would *123have no problem with the death penalty,” and adding that in such a case death would be the only sentence he would give. Another juror was in tears as she discussed the issue of child murder victims. At least one of the jurors who raised the issue, Juror 127, indicated that she nevertheless would be impartial as to the sentencing options and. was allowed to remain in the pool. See Division 8 (b) (1) below. See also Division 8 (b) (2) below (discussing Juror 175). As to the other prospective jurors, including the 12 who ultimately decided his sentence, Ellington had no information about their views regarding sentencing in a case involving young children as murder victims to cite in seeking to have them excused for cause and to consider in exercising his peremptory strikes.
Having obtained rulings barring any affirmative mention during voir dire of the fact that the case involved the murder of two young children, as soon as the jury was empaneled and the trial began, that fact became a focus of the State’s argument that Ellington should be sentenced to death. Thus, within the very first minute of his opening statement, the prosecutor informed the jury that the case involved the murder of “two young boys.” The prosecutor later emphasized the point, saying, “Two 2-year old boys a month and a week after they turned two they were found identically dressed in their identical cribs on identical mattresses with identical injuries. . . .” And throughout the State’s closing argument during the sentencing phase, the prosecutor focused on the child victims. At one point he asked the jury to look at their photographs, as he argued, “I dare say you can’t find much sweeter than those two standing there.” And at the end of the State’s presentation to the jury, just as at the beginning, the prosecutor returned to the point. He implored the jury “to just remember those babies. Remember what they will never do, remember what they will never know, remember what they will never become and why and by who that was taken from them.” The jury unanimously returned death sentences for the murder of the two children and their mother.
Ellington’s motion for a new trial argued again that it was error to prohibit voir dire that addressed the fact that two of the victims were young children because that fact “was directly relevant to juror bias and ability to serve.” The trial court denied the motion, again ruling that Ellington had sought to ask questions that required the prospective jurors to pre-judge the case. He appeals that ruling.
(b) A fair trial before an impartial factfinder is a fundamental component of due process of law. See Irvin v. Dowd, 366 U. S. 717, 721-722 (81 SC 1639, 6 LE2d 751) (1961). See also Morgan v. Illinois, 504 U. S. 719, 727 (112 SC 2222,119 LE2d 492) (1992) (“[D]ue process *124alone has long demanded that, if a jury is to be provided the defendant, regardless of whether the Sixth Amendment requires it, the jury must stand impartial and indifferent to the extent commanded by the Sixth Amendment”). Indeed, the right to criminal trial “by an impartial jury” is textually guaranteed by both the United States and Georgia Constitutions. See U. S. Const. Amend. VI; Ga. Const, of 1983, Art. I, Sec. I, Par. XI (a).
Much like cross-examination is the engine of truth in our justice system, voir dire is the engine of selecting a jury that will be fair and impartial. Thus, while recognizing that trial judges must have substantial discretion to oversee jury selection and that this subject is largely governed by state laws and practices, the Supreme Court of the United States has held that due process requires that voir dire be sufficient to allow the parties and the trial court to elicit juror bias. See Rosales-Lopez v. United States, 451 U. S. 182, 188 (101 SC 1629, 68 LE2d 22) (1981) (“Without an adequate voir dire the trial judge’s responsibility to remove prospective jurors who will not be able impartially to follow the court’s instructions and evaluate the evidence cannot be fulfilled. Similarly, lack of adequate voir dire impairs the defendant’s right to exercise peremptory challenges where provided by statute or rule, as it is in the federal courts.”) (citation omitted); Morgan, 504 U. S. at 729, 735 (holding that “part of the guarantee of a defendant’s right to an impartial jury is an adequate voir dire to identify unqualified jurors,” including jurors who would automatically impose a death sentence “regardless of the facts and circumstances” of the case); Turner v. Murray, 476 U. S. 28, 36-37 (106 SC 1683, 90 LE2d 27) (1986) (holding that “a capital defendant accused of an interracial crime is entitled to have prospective jurors informed of the race of the victim and questioned on the issue of racial bias”).
Georgia has a broadly worded statute describing the scope of voir dire in both criminal and civil cases: *125OCGA § 15-12-133 (emphasis added). This Court has held that OCGA § 15-12-133 allows voir dire questions beyond those that the Constitution would require allowing. See Legare v. State, 256 Ga. 302, 304 (348 SE2d 881) (1986) (“It has also been held that OCGA § 15-12-133 encompasses questions regarding possible racial prejudice and bias, even when such questioning would not be constitutionally required”); Tucker v. State, 249 Ga. 323, 326-328 (290 SE2d 97) (1982) (noting the same). Moreover, the language of the statute is broad enough that we would interpret it to avoid significant constitutional concerns. See Haley v. State, 289 Ga. 515, 521-522 (712 SE2d 838) (2011). Accordingly, we address the question presented under OCGA § 15-12-133, recognizing that constitutional considerations inform our view of what the statute requires. The focus of the question is the meaning of “the subject matter of the action” in OCGA § 15-12-133.5
*124In the examination, the counsel for either party shall have the right to inquire of the individual prospective jurors examined touching any matter or thing which would illustrate any interest of the prospective juror in the case, including any opinion as to which party ought to prevail, the relationship or acquaintance of the prospective juror with the parties or counsel therefor, any fact or circumstance indicating any inclination, leaning, or bias which the prospective juror might have respecting the subject matter of the action or the counsel or parties thereto, and the religious, social, and fraternal connections of the prospective juror.
*125As the parties generally agree, Georgia law is clear regarding voir dire questions at two opposing poles. On the one end, the “subject matter” of a criminal case clearly includes the fact that it is a criminal case, see Henderson v. State, 251 Ga. 398, 402 (306 SE2d 645) (1983), as well as the offenses charged in the indictment. See, e.g., id. at 401 (citing Craig v. State, 165 Ga. App. 156, 156 (299 SE2d 745) (1983) (holding in a case charging cocaine trafficking that the trial court abused its discretion in precluding defense counsel from asking prospective jurors if they or their family had ever been the victim of a drug transaction or had a problem with drugs)).
OCGA § 15-12-133 does not use the narrow term “charges” or “cause of action,” and under Morgan and Georgia cases, the “subject matter of the action” also includes the potential sentence the indictment carries, and in death penalty cases in particular, the three sentencing alternatives — the death penalty, life without parole, and life with the possibility of parole. Thus, the parties are entitled to ask a prospective juror if he or she would automatically impose a particular sentence upon conviction, regardless of the facts and circumstances of the case. See Morgan, 504 U. S. at 735 (“Any juror who would impose death regardless of the facts and circumstances of conviction cannot follow the dictates of law.”); Zellmer v. State, 272 Ga. 735, 736 (534 SE2d 802) (2000) (after the statutory amendment *126requiring the jury to determine parole eligibility in capital cases, “parole eligibility is part of the subject matter of the sentencing phase of death penalty trials, [and] we hold that criminal defendants and the State are statutorily entitled [by OCGA § 15-12-133] to examine jurors concerning their inclinations, leanings and biases regarding parole”). See siso Atlanta Joint Terminals v. Knight, 98 Ga. App. 482, 489 (106 SE2d 417) (1958) (“Under the broad sanction of [OCGA § 15-12-133] it must necessarily be held that prejudice as to the size of verdicts is as much comprehended under the subject matter of civil actions as the nature of the cause of action.”).
At the other pole, many cases hold that “pre-judgment” questions may not be asked in voir dire in an effort to have prospective jurors commit themselves to a particular outcome based on specific facts of a case. See Crowe v. State, 265 Ga. 582, 588 (458 SE2d 799) (1995) (“It is inappropriate to outline the case and then ask the juror what his or her vote would be.”). Our cases have involved situations in which the question(s) at issue involved “an outline of the case,” multiple detailed facts, wording that asked the prospective juror to commit to a particular verdict based on hypothesized facts, or requests for the jurors to opine on what facts they would consider important.6 We have explained that voir dire is not intended to pre-try the case on hypothesized facts and get jurors to commit to the outcome based on that speculative proof. See King v. State, 273 Ga. 258, 266 (539 SE2d 783) (2000) (stating that “defense counsel improperly questioned the *127juror about what sentence the juror might choose under the specific, hypothetical circumstance” laid out by counsel); Crowe, 265 Ga. at 588. The “subject matter of an action” is not easily understood to include every detail of the case, and it is entirely appropriate for jurors to shape their judgment as to the verdict only after hearing the full evidence and complete legal instructions at trial.
The Uniform Superior Court Rule on voir dire tracks this case law, providing as follows:
Hypothetical questions are discouraged, but maybe allowed in the discretion of the court. It is improper to ask how a juror would act in certain contingencies or on a certain hypothetical state of facts. No question shall be framed so as to require a response from a juror which might amount to a prejudgment of the action.
Uniform Superior Court Rule 10.1.
Under this law, many cases are easy. But many others, including this case, are harder. See Henderson, 251 Ga. at 399 (acknowledging thatOCGA § 15-12-133 “isasourceof concern toourtrial judges,both because of the method of examining prospective jurors it authorizes and because of the scope of such examination,” and noting that “the Code section is written in general terms”). The line between permissible inquiry into “prejudice” (a juror’s fixed opinion that a certain result should automatically follow from some fact, regardless of other facts or legal instructions) and impermissible questions of “prejudgment” (speculation about or commitment to the appropriate result based on hypothesized facts) can be hazy. Thus, in this area as in other areas of voir dire practice, appellate courts should give substantial deference to the decisions made by trial judges, who oversee voir dire on a regular basis, are more familiar with the details and nuances of their cases, and can observe the parties’ and the prospective jurors’ demeanor.
Since there is often a fine line between asking potential jurors how they would decide the case and questions that merely seek to expose bias or prejudice, the scope of the voir dire examination, of necessity, must be left to the sound discretion of the trial court.
Sallie, 276 Ga. at 510. Accord Curry v. State, 255 Ga. 215, 218 (336 SE2d 762) (1985).
(c) We believe that Ellington was entitled to ask whether the prospective jurors in this case would automatically vote for a death *128sentence in any case in which two murder victims were young children, regardless of any other facts or legal instructions. As to the jury’s decision on the sentences in this case, our experience in criminal justice matters and simple common sense indicate that the fact that two of the victims were young children was the critical issue. To be sure, there were other aggravating circumstances, including the brutality of the murders and the fact that Ellington committed a triple murder. But the first point the average person would note in describing this case is that Ellington murdered his two two-year-old sons. As the Ohio Supreme Court has said:
Protecting children from harm is a common human characteristic, and many people harbor strong feelings and emotions whenever a child is a victim of a violent crime. Some prospective jurors, when presented with this fact, may have been unable to remain dispassionate and impartial when deciding whether the death sentence should be imposed.
State v. Jackson, 836 NE2d 1173, 1191-1192 (Ohio 2005). Certain types of murder are generally considered so grievous that there is a label for those who commit them in the vernacular, including “baby-killer” as well as “cop-killer,” “serial-killer,” and “mass-murderer.”
The focus of Georgia’s people on protecting children is evidenced by many laws that make conduct criminal only when committed against children or enhance the punishment when children are the victims of criminal conduct. See, e.g., OCGA §§ 16-5-70 (cruelty to children), 16-6-4 (child molestation). Georgia law even seeks to protect children from having to testify about certain crimes. See OCGA § 24-3-16 (child hearsay statute); Bunn v. State, 291 Ga. 183, 189 (728 SE2d 569) (2012). Indeed, our law provides that the fact that a murder victim was a child may serve as part of the basis for finding a statutory aggravating circumstance in death penalty cases. See West v. State, 252 Ga. 156, 161-162 (313 SE2d 67) (1984) (Appendix) (supplying a pattern jury charge to inform jurors that they “may consider the age... of the victim” in determining if a murder involved “depravity of mind” under OCGA § 17-10-30 (b) (7)). Moreover, as the Amicus Curiae Georgia Association of Criminal Defense Lawyers correctly points out, federal law and statutes in at least 13 other states provide that the fact that the murder victim was a child forms all or part of a statutory aggravating circumstance.7
*129But we need not rely only on experience, common sense, and background law. We can also look to how the State actually tried this case. After strenuously objecting to any inquiry about the jurors’ views as to child victims, the prosecutor focused on that fact from opening statement in the guilt/innocence phase to closing argument in the sentencing phase as a principal reason that Ellington should receive the death sentence. The prosecutor understood the power of that fact, and indeed the State does not argue to us that it is not a fact that would cause some people (and jurors are regular people) to automatically impose a death sentence. Indeed, in the hearing held at the beginning of voir dire, the prosecutor explained that the State was trying to avoid the lengthy voir dire that had been necessary in an earlier death penalty case, in which questions about child victims had been allowed. That point actually cuts against the State’s position. If it takes extensive voir dire to ensure that prospective jurors are not biased based on a critical fact present in a case, that shows the importance of voir dire on the issue, not the need to squelch inquiry on the subject.
Morgan recognized the reality that there are some prejudices that will not be adequately exposed with the basic impartiality questions, by simply asking jurors “generally whether [they] could be fair and impartial.” Morgan, 504 U. S. at 723. To the extent the prospective jurors know what the case is about (for example, because the indictment has been read to them), they may be able to answer such a question with particular key facts in mind. But to the extent they are not focused on an issue • — • like the sentencing options in a death penalty case — they may honestly answer a general impartiality question when they would answer differently if asked a more focused question. See id. at 735 (“As to general questions of fairness and impartiality, such jurors could in all truth and candor respond affirmatively, personally confident that such dogmatic views [about the death penalty] are fair and impartial, while leaving the specific concern unprobed.”).
In this case, for example, asked after hearing the indictment read if they could be fair and impartial, prospective jurors might have honestly answered yes; and they might honestly have answered no when asked, per Morgan, if they would automatically impose the death sentence in any murder case. But if they were advised that the case involved the murder of two young children, at least some of the *130prospective jurors might have changed their answers. See Meeks v. State, 269 Ga. App. 836, 840-841 (605 SE2d 428) (2004) (holding that the trial court abused its discretion in relying on general impartiality questions in a case charging child molestation and similar offenses rather than allowing the defendant to “ask prospective jurors whether they had such strong feelings about child molestation that it would impair their judgment or make it difficult for them to judge the case,” although concluding that the error was harmless); Johnson, 366 FSupp.2d at 847-848 (discussing the district court’s experience in a companion case that “suggest [ed] the insufficiency of purely ‘abstract’ questions to determine the ability of jurors to perform their duties” in a case involving the murder of children).
This is not just a reasonable inference. We know that several prospective jurors who knew from pretrial publicity or deduced from the indictment that there were child victims actually expressed doubts about their impartiality as to sentencing that they had not expressed in response to the general impartiality questions. Some of those prospective jurors were excused for cause, and even for the one not so excused, Ellington could factor in that juror’s views in exercising his peremptory challenges. See Henderson, 251 Ga. at 399-400 (“It should be kept in mind that the larger purpose of [OCGA § 15-12-133] is to enable counsel to identify those prospective jurors counsel desires to remove from the panel by use of peremptory strikes as opposed to challenges for cause.”). See also Jackson, 836 NE2d at 1190 (explaining that “[e]ven without being informed that one of the victims in this case was a child, some members of the venire indicated a possible bias in favor of imposing the death penalty on murderers of children,” which put the trial court “on notice that some prospective jurors harbored a strong bias in favor of imposing death for murderers of children”). We cannot safely conclude that, when the jurors who were finally seated found out that this case involved two very young murder victims — a fact most of them learned just moments after they were sworn in and the prosecutor started his opening statement —not even one of them said internally, “Well, that’s all I need to hear.”
The trial court appears to have shared this understanding of the force of the child-victim fact. Thus, just after prohibiting any voir dire questions on the issue, the court said:
[T]here’s no doubt that certain jurors may have a special sensitivity to victims of child murder which would make them unable to perform their duties as jurors regardless of the facts of the case and the instructions of the Court.
*131The court then tried to supplement the basic sentencing option questions with a vague reference to “the status of the victim,” apparently hoping that astute jurors would announce any victim-related prejudices. Yet the trial court ruled that the issue could not be directly addressed with jurors unless it “bubble [d] up” spontaneously in a particular juror’s voir dire responses. When a few prospective jurors did bring up the child-victim issue, based on their hearing the indictment read, their exposure to pretrial publicity, or simply their own ideas about the death penalty in general, the court allowed the parties to explore it and Ellington to move to exclude jurors, sometimes successfully, based at least in part on their answers.
We therefore understand the trial court to have misperceived the scope of its discretion to allow voir dire on the child-victim issue, wrongly believing that it had none. That misperception produced an abuse of discretion.
Where a ruling of the trial court which is ordinarily one within the sound discretion of the court shows that no discretion was, in fact, exercised, and the judgment rendered is based upon an erroneous view of the law which would preclude the exercise of a discretion, a new trial results.
Watson v. Elberton-Elbert County Hosp. Auth., 229 Ga. 26, 27 (189 SE2d 66) (1972). Compare Braithwaite v. State, 275 Ga. 884, 889 (572 SE2d 612) (2002) (finding no reversible error where the trial court acted within its legal range of discretion and where “the record does not show that the court made [its] decision under a misapprehension about the scope of its discretion”).
(d) The District Attorney argues that Ellington only sought to ask one specific question — Question 12 on the proposed jury questionnaire — which was properly deemed a pre-judgment question, thus Ellington did not properly preserve any broader issue for appeal. Question 12 asked, “Do you believe that a life sentence can be severe enough punishment for a person convicted of the deliberate killing of a child?” It was arguably within the trial court’s discretion to rule that this specific question was improperly framed. It employed the word “deliberate” which might be construed to mean the disputed and complex fact of premeditation; it did not include any reminder that the jurors’ ultimate decision must be based on the evidence at trial and the court’s instruction; and it did not focus on the proper inquiry, which was whether the prospective jurors would fairly consider a life sentence for a person who murders a child. See Johnson, 366 FSupp.2d at 849. See also Gissendaner, 272 Ga. at 707-708 (holding that the *132trial court “properly admonished” defense counsel for asking a prospective juror about the detailed hypothesized circumstance of “malice murder where significant premeditation has occurred”).
But in any event, Ellington’s argument, and the trial court’s ruling, was clearly not limited to Question 12 or the arguable need to reframe its wording. Indeed, the court’s initial ruling was to grant the State’s motion in limine and preclude any inquiry into whether jurors might be biased in a case involving child victims. And the court reiterated its ruling in the hearing before voir dire, when Ellington argued, “How are we going to uncover that bias if we never mention the fact that it’s a child?” Ellington was not seeking to “outline the evidence” or commit the jurors to a specific result based on a set of hypothesized facts. Indeed, the fact at issue — that two of the murder victims were young children — was entirely undisputed. Thus, Ellington properly preserved the argument he makes to this Court.
The State contends that allowing a child-victim question is inconsistent with Georgia precedent and the majority view of courts nationwide. However, as discussed above, the Georgia cases that the State cites did not focus on a narrow question involving a particular, critical fact that could reasonably be expected to cause jurors to be biased in the case at issue. Instead, this Court and the Court of Appeals properly rejected questions that sought to outline the case. See, e.g., Blankenship, 258 Ga. at 45 (“[N] either the defendant nor the state has the right simply to outline the evidence and then ask a prospective juror his opinion of that evidence.”). By contrast, Ellington cites cases in which questions going beyond the charges in the indictment have been allowed. See Quintana v. State, 276 Ga. 731, 733, 733, n. 5 (583 SE2d 869) (2003) (holding that the trial court did not abuse its discretion in allowing the State to ask, in a case where defendant allegedly committed murder while using drugs, whether prospective jurors believed that “alcohol and/or drugs could make you impulsive” or “could make a person violent,” and allowing the defendant to ask “whether anyone held views of illicit drugs that would prevent them from being an impartial juror”); Childers v. State, 228 Ga. App. 214, 215 (491 SE2d 456) (1997) (in a case involving an aggravated battery charge arising from an incident of domestic violence, calling domestic violence the “subject matter of the action” and approving the State’s questions about prospective jurors’ “personal beliefs concerning domestic violence issues”).
We have faced a similar issue interpreting OCGA § 15-12-133 in the civil context. In a medical malpractice case involving the alleged failure to diagnose the plaintiff’s breast cancer, we did not restrict voir dire to prospective jurors who would automatically impose a plaintiff or defense verdict in any medical malpractice cause of action. *133Instead, we held that the “subject matter of the action” allowed the plaintiff to ask whether prospective jurors or their immediate families had been diagnosed with cancer, explaining that the question “may be characterized as one bearing directly on the subject matter of the litigation.” Gainesville Radiology Group v. Hummel, 263 Ga. 91, 93 (428 SE2d 786) (1993).
We also note that, despite its claim that impartiality questions relating to facts not set forth in the indictment are improper, the State sought and obtained approval to ask voir dire questions regarding the prospective jurors’ views about gruesome photographs of the murder victims — case-specific evidence that was not mentioned in the indictment and that, unlike the age of the victims, was subject to some dispute. See Division 9 below. Tellingly, the State, despite having sought approval to discuss the photographs in its voir dire questions, argued pretrial that Ellington should not be able to discuss the photographs in asking if jurors would be able to consider a sentence of less than death. Nor did the State object to the trial court reading the indictment to the jurors, including the allegation in each of the three malice murder counts that Ellington struck the victims with a hammer — a fact that was not an element of those charges.
These decisions by the prosecutor illustrate why our resolution of this issue should not be perceived as helpful in general to either party in criminal cases, including death penalty cases. While the child victims in this case may be particularly sympathetic, and voir dire into that sympathy would have assisted the defendant in removing prospective jurors who could not be impartial, the victims in other cases maybe less sympathetic, and it maybe the State that wants and needs to probe juror partiality on that point. See, e.g., United States v. Flores, 63 F3d 1342,1356 (5th Cir. 1995) (approving the trial court’s decision to excuse a juror who stated that he would never vote for death in a case involving a victim who was involved with drugs); Greg Bluestein, “Man spared death sentence in federal case,” Online-Athens, Apr. 26, 2012, http://onlineathens.com/local-news/2012-04-26/man-spared-death-sentence-federal-cáse (reporting on a death penalty case in which the jury returned a life sentence for a federal prison inmate who murdered another inmate who was a convicted child molester).
Other courts, albeit not all of them, have reached the same conclusion we reach today, particularly on the specific issue of child victims. The State cites United States v. Johnson, which noted that the “majority of courts reject [pre-judgment] questions.” 366FSupp.2d at 840. However, the cases Johnson was discussing involved the sort of more detailed questions that this Court has rejected, as shown by Johnson’s citation to our decision in Lucas. See Johnson, 366FSupp.2d *134at 841 (citing Lucas, 274 Ga. at 646). In fact, the Johnson court carefully and extensively reviewed the constitutional and decisional law on the same issue presented here and concluded that,
while Morgan does not require “case-specific” questioning of prospective jurors to satisfy constitutional requirements for life- and death-qualifying prospective jurors, “case-specific” questions are nevertheless appropriate — indeed, necessary —to empanel a fair and impartial jury in this particular case [involving child victims].
Id. at 848 (citing Morgan, 504 U. S. at 730-734). The court added that
either party is entitled to ask prospective jurors questions that are specific enough to determine if those jurors harbor bias, as to some fact or circumstance shown by the trial evidence, that would cause them not to follow an instruction directing them to determine [the] penalty after considering aggravating and mitigating evidence.
Id. at 846-847 (citation omitted).
At least two other state supreme courts also agree with our conclusion. See Jackson, 836 NE2d at 1192 (“The trial court abused its discretion by refusing defense counsel’s requests to advise prospective jurors that one of the murdered victims was a three-year-old child and by refusing to allow voir dire on that fact.”); State v. Clark, 981 SW2d 143, 147 (Mo. 1998) (reversing where the trial court refused to allow voir dire questions regarding potential bias arising from the “critical fact” that one of the victims was three years old). The Ohio Supreme Court seems to allow a broader scope of case-related voir dire questions than we have allowed in Georgia. See Jackson, 836 NE2d at 1190. However, the Missouri Supreme Court’s reasoning was similar to ours, in a case similar to ours:
Every fact need not be disclosed to prospective jurors. Only those critical facts — facts with substantial potential for disqualifying bias •— must be divulged to the venire. In the present case, the trial court completely precluded defense counsel from questioning prospective jurors on the “specifics of the case being tried,” in particular that one victim was only three years old. A case involving a child victim can implicate personal bias and disqualify prospective jurors. The trial court must strike for cause prospective jurors when they exhibit prejudicial bias because the victim is a child.
*135Due to the sweeping nature of the trial court ruling in this case, the defense could not attempt to discover that bias. Because this does not satisfy the defendant’s right to an impartial jury, the ruling constitutes an abuse of discretion.
Clark, 981 SW2d at 147 (citations omitted).
The best case for the State’s position appears to be United States v. McVeigh, 153 F3d 1166 (10th Cir. 1998), disapproved on unrelated grounds by Hooks v. Ward, 184 F3d 1206, 1227 (10th Cir. 1999). There the Tenth Circuit upheld the district judge’s refusal to allow the defense in Timothy McVeigh’s trial for bombing the federal building in Oklahoma City — which murdered 168 people, including children (and law enforcement officers) — to ask prospective jurors if they had formed an opinion regarding the appropriate punishment in light of the specific facts of the case, which many prospective jurors knew about from pretrial publicity. See McVeigh, 153 F3d at 1208 (‘Morgan ... is designed to illuminate a juror’s basic beliefs ‘regardless of the facts and circumstances of conviction,’ not to allow defendants to pre-determine jurors’ views of the appropriate punishment for the particular crime charged. Morgan does not require that the questions at issue be asked.” (citation omitted)). We might also conclude that the actual questions at issue in McVeigh were inappropriate because they were too detailed, failed to remind jurors that ultimately the case should be decided on the evidence presented at trial, and sought a final prejudgment of the appropriate punishment rather than the jurors’ views as to whether they could consider all the sentencing options. More generally, however, we believe that the Tenth Circuit misread Morgan to prohibit all case-specific questions, when Morgan holds only that due process requires the trial court to allow questions probing a prospective juror’s willingness to consider a sentence of less than death upon the conviction of a defendant for an abstract murder. See Johnson, 366 FSupp.2d at 834-835. We also note that McVeigh involved a trial in federal court, where voir dire is traditionally more restrictive than it is in Georgia under OCGA § 15-12-133. We therefore believe that our holding is not out of line with other courts on the precise question at issue.
(e) In sum, we hold that the “subject matter of the action” as to which voir dire is permitted under OCGA § 15-12-133 extends beyond the crimes charged in the indictment and the sentences the charges carry to other “critical facts” of the case that experience, reason, and common sense indicate will be so influential for at least some prospective jurors that they will be unable to consider all of the evidence in the case in light of the court’s instructions on the law and render *136a fair and impartial verdict. It is important to recognize three limitations of this holding.
First, the issue is not whether the prospective juror will consider a critical fact to be very important or worthy of great weight. Indeed, it is often entirely appropriate for jurors to consider such critical facts as carrying great weight in making their judgments about the appropriate verdict in a case. There is nothing wrong, for example, with jurors who consider carefully, and even assign dispositive weight to, the fact that the victim of a murder was a young child in deciding to return a verdict of death — so long as that judgment is rendered fairly, at the end of the case, after the juror has heard not only that fact (assuming it is proved) but all the other evidence and after the juror has considered that evidence in light of the court’s legal instructions. The problem is with prospective jurors who, if asked about a critical fact involved in a case, admit that they would automatically return a certain verdict, regardless of other facts and regardless of the law. Such prospective jurors are not impartial, and the parties are entitled to ask appropriate questions in voir dire to identify their bias.
We note in this regard that, as with other issues about a case, prospective jurors who, when advised of a critical fact, initially indicate some leaning based on that fact may be rehabilitated by additional questions that reveal that they will not make up their minds until they hear all the evidence and the law. See Pace v. State, 271 Ga. 829, 834 (524 SE2d 490) (1999) (“A prospective juror is not subject to excusal for cause for merely leaning for or against a death sentence.”); Head v. State, 276 Ga. 131, 133 (575 SE2d 883) (2003) (holding that a juror with a leaning regarding the defendant’s guilt was nevertheless qualified because his opinion was not “fixed and definite”). The problem is not the prospective juror who says, “Of course I would consider the fact that a murder victim was a baby very heavily in deciding on the sentence.” It is the juror whose answers reveal, “and nothing else would matter to me,” rather than showing, “but I will reach my verdict in the case only after considering all of the evidence and the law.”
Second, and following from the first point, the voir dire questions must be framed properly to reveal the prospective juror’s general view on the critical fact and whether that view is so strong that it would substantially impair the juror in considering all three sentencing options; the questions must not seek to commit the juror to voting a certain way based on that fact. Thus,
questions must be in the form of whether or not the prospective juror “could fairly consider” [each option of life, life *137without parole, and death], not whether the prospective juror would vote for life or death in light of particular facts.
Johnson, 366 FSupp.2d at 849. Ideally, such questions will include a reminder that the ultimate decision must be based on the evidence at trial and the court’s instructions, such as “if the evidence shows.” See id. The question must also be based only on critical facts that are likely to be proved at trial or will genuinely be in dispute. See id. Examples of proper questions about critical facts are: “Would you automatically reject a life sentence if the evidence showed x?” and “Could you fairly consider a death sentence if the evidence showed x?” Id. The trial court retains its broad discretion to control the form and number of voir dire questions. See Uniform Superior Court Rule 10.1 (providing that “the form, time required and number of such questions is within the discretion of the court”); King, 273 Ga. at 271 (holding that the trial court has discretion in controlling the form and number of questions regarding possible racial bias); Jackson, 836 NE2d at 1192 (holding that the trial courts in Ohio have discretion regarding “the form and number of questions” like those at issue here).
Finally, as emphasized above, decisions as to what, if any, facts of a particular criminal case beyond the charges and sentencing options qualify as “critical” in terms of risking juror partiality can be difficult and context-specific. The trial court’s decision on this point should be given significant deference in appellate review, and it should be reversed only for a manifest abuse of discretion. In holding that trial courts have discretion in allowing such voir dire questions, we do not suggest in any way that such questions should be routine, even in death penalty cases. Questions are appropriate only when not asking them runs a real risk that juror partiality driven by the fact at issue will not otherwise be identified in voir dire. As an example, a trial court considering a case in which the victim was a much older child than the twin toddlers in this case would need to exercise its discretion to decide the potential impact of that fact among the particular circumstances of that case, and we would accord considerable discretion to that decision. But trial courts must understand that they have such discretion; and denying important voir dire without such an understanding is error, such as occurred in this case.8
*138(f) We must next consider the effect on the trial court’s judgment of our holding that the court erred in not allowing Ellington to ask voir dire questions to elicit potential juror partiality based on the fact that two of the murder victims were children. To the extent that the court’s ruling violated Ellington’s statutory rights under OCGA § 15-12-133, the error may be deemed harmless if the State can show that it is highly probable that the error did not affect the result of the trial. See Legare, 256 Ga. at 303-304 (addressing a violation of the statutory right to adequate voir dire and reversing because it was not “highly probable” that the error was harmless); Henderson, 251 Ga. at 403 (same). To the extent that the ruling impaired Ellington’s constitutional right to due process, we could affirm only if the error was harmless beyond a reasonable doubt. See Chapman, 386 U. S. at 24 (holding that, “before a federal constitutional error can be held harmless, the court must be able to declare a belief that it was harmless beyond a reasonable doubt”). Our conclusions below are the same under either standard.
We see no reason to believe that any juror partiality toward the death penalty for defendants who murder young children would have undermined the jury’s guilty verdict on the charges in this case. See Nance v. State, 272 Ga. 217, 224 (526 SE2d 560) (2000) (affirming convictions where an error in qualifying a juror did not affect the guilt/innocence determination). Ellington did not argue at trial that he was entitled to ask the jurors about child victims with regard to their guilt-innocence determination, and he does not explain on appeal how the jury’s views about child murder victims would make them more likely to find him guilty of such murders. To the contrary, Ellington’s principal defense at trial was that he killed his wife in a fit of rage because she killed their children. To the extent that any juror believed that story, and was unalterably of the view that anyone who kills a child deserves to be executed, the juror would have been more inclined to accept Ellington’s defense. See Nance, 272 Ga. at 224 (affirming convictions where an error in juror qualifications did not affect the guilt/innocence determination).
Ellington’s death sentences are another matter. His death sentences for the murders of his two children were directly affected by the trial court’s error and must therefore be reversed. We reach the same conclusion as to Ellington’s death sentence for the murder of his *139wife. If this case did not involve the murder of children, a reasonable jury might still conclude that a death sentence was appropriate for Berna Ellington’s murder, which was brutal and was committed in the course of two other murders. But this case does involve two child murders, and as discussed above, that fact was the critical non-statutory aggravating circumstance emphasized by the State at trial. We cannot say with the necessary confidence that not a single juror on Ellington’s trial jury would have automatically imposed the death sentence for any child murder or that the jury would have unanimously imposed a death sentence for Berna Ellington’s murder if this case had not also involved the murder of two children — or at least that one of the jurors selected to serve would not have expressed a view on the child victim issue that would have led Ellington to use a peremptory strike to remove him or her.
We therefore reverse all three of Ellington’s death sentences and remand the case to the trial court for resentencing. Because we hold that there was sufficient evidence to support the jury’s findings that two statutory aggravating circumstances existed as to each murder, see Division 1 above, the State is not barred from seeking the death penalty again in a new sentencing trial before a properly qualified jury. See Bryant v. State, 288 Ga. 876, 901.(708 SE2d 362) (2011).
8. Ellington raises other jury selection issues, many of which are unnecessary to resolve in light of our decision to reverse his death sentences. However, we will address Ellington’s arguments to the extent they concern matters potentially affecting his convictions or matters likely to recur if he is retried as to sentencing.
(a) Ellington contends that the trial court erred by prohibiting voir dire questions seeking to learn whether the jurors would be swayed by the concept of mercy in the event that they found no other mitigating evidence sufficient to justify a sentence of less than death. Both of the jurors discussed in Ellington’s brief made perfectly clear that they would consider mitigating evidence and would be willing to impose any of the three possible sentences.9 We reject Ellington’s argument that the Georgia Code divorces the consideration of mercy from the consideration of mitigating circumstances. Instead, a jury’s consideration of whether to grant mercy is meant to be conducted in light of the mitigating and aggravating circumstances. See OCGA § 17-10-2 (c) (providing that “the jury shall retire to determine whether any mitigating or aggravating circumstances, as defined in Code *140Section 17-10-30, exist and whether to recommend mercy”). Thus, the trial court properly sustained the State’s objection to Ellington’s attempt to obtain a prejudgment on the ultimate question of mercy under hypothesized circumstances of what the jurors, if selected to serve, would ultimately consider to be mitigating factors. See Sallie, 276 Ga. at 509.
(b) Ellington argues that the trial court erred by declining to excuse two prospective jurors based on their responses to questions on various points of potential bias not related exclusively to the sentencing options.10 The trial court’s findings regarding the potential bias of a juror are afforded deference on appeal. See Ledford, 289 Ga. at 80-81 (addressing the trial court’s findings regarding various possible sources of bias under an abuse of discretion standard); Hyde v. State, 275 Ga. 693, 696 (572 SE2d 562) (2002) (“Aconclusion on an issue of juror bias is based on findings of demeanor and credibility which are peculiarly in the trial court’s province, and those findings are to be given deference.”). The question is whether the juror has a fixed and definite opinion, not whether he or she leans one way or another but nevertheless has an open mind. See Pace, 271 Ga. at 834; Head, 276 Ga. at 133. The trial court must exercise its discretion carefully in this area, however, because we have held that a defendant is entitled to a full panel of qualified jurors at the beginning of peremptory strikes and that “the erroneous qualifying of a single juror for the panel from which the jury was struck” requires reversal. Lance v. State, 275 Ga. 11, 15 (560 SE2d 663) (2002).
(1) Juror 127 said that her daughter had been the victim of spousal abuse and had two children. The juror indicated that she had learned from pretrial publicity that the victims in this case were a mother and two children, that the victims had been killed with a hammer, and that it was not clear whether the mother had killed the children and was then killed by someone else or whether someone else had killed all three victims. She also had seen a photograph of Ellington that she had assumed was a “mug shot” taken as a result of his arrest. However, the juror said that she did not always believe what she saw in the media, that she could set aside what she had been exposed to pretrial, and that she would base her decision solely on the evidence presented at trial. The juror at first said that she was “not 100 percent sure” that she could be impartial in a case involving the murder of two children, but she also indicated that she would decide *141the case based upon the evidence at trial and that she thought that she could be impartial, and she later indicated that she was “100 percent sure” that she could be fair. The juror also noted that she had attended some rallies in support of children who had been murdered, that she had “put a couple bucks in” the collections that were taken at the rallies, and that the issue of the murder of children was dear to her heart. Considering Juror 127’s voir dire as a whole, we conclude that the trial court did not abuse its discretion by finding her qualified to serve.
(2) Juror 175 said that she had learned from a friend that there was a pending case in DeKalb County involving a man charged with killing his family and she had been relieved when she wrongly concluded that she had not been summoned as a potential juror for that case. She indicated consistently that she would consider all three sentencing options. She explained that, particularly given the fact that she had children, she would have “mixed [e] motions” about serving as a juror in a case involving children as murder victims, and she did not know whether she would be able to endure seeing gruesome photographs of the victims. Ellington argues that this juror should have been disqualified for cause because she thought that she might become “emotional” during the trial and because she showed signs of such emotions during her voir dire. However, she also said, “And I feel that I’m an intelligent person and I have the ability to listen and pay attention and understand evidence and be fair.” Although a juror’s emotional issues may rise to such a level that they would interfere with his or her sworn duties, the trial court did not abuse its discretion in finding Juror 175 qualified to serve.
Guilt/innocence Phase Issues
9. The trial court did not abuse its discretion in the type and number of crime scene photographs it admitted at trial. See Riley v. State, 278 Ga. 677, 686 (604 SE2d 488) (2004) (“The photographs admitted were relevant and admissible to show the nature and extent of the injuries to the victims and the locations and positions of the bodies____”). See also Stinski v. State, 281 Ga. 783, 785-786 (642 SE2d 1) (2007) (noting a trial court’s discretion in weighing the probative value of photographs against any “undue prejudice”).
10. Ellington contends that the prosecutor made several improper arguments to the jury at the conclusion of the guilt/innocence phase. Most of Ellington’s claims were waived because he failed to object at trial. But in any event, even if some of the State’s arguments were improper, there was no reversible error.
*142(a) Ellington contends that the following argument by the prosecutor amounted to an improper comment on Ellington’s exercising his right not to testify and urged the jury to shift the burden of proof to him rather than placing it on the State:
You know, [defense counsel] said [the State] didn’t do this or [the State] didn’t do that.
You know, a criminal defendant never and I mean never has a burden to produce any evidence, to testify, witnesses, nothing. Nothing. The burden is always on the State 100 percent. We have to prove guilt beyond a reasonable doubt. But that doesn’t mean they can’t. And to say we didn’t do something or the evidence wasn’t shown you — because if it exists — ask why they didn’t.
An argument “that the defendant has not rebutted or explained the State’s evidence” is not an improper burden-shifting argument. Arrington, 286 Ga. at 346. Aprosecutor also may comment on the lack of evidence offered by the defendant to rebut the evidence the State presented without thereby commenting on the defendant’s right not to testify. See Smith v. State, 245 Ga. 205, 207 (264 SE2d 15) (1980); Jordan v. State, 239 Ga. 526, 527-528 (238 SE2d 69) (1977). See also LeMay v. State, 265 Ga. 73, 75 (453 SE2d 737) (1995) (“Reversal for improper comment by the prosecutor requires a finding either that 1) the prosecutor’s manifest intention was to comment on the accused’s failure to testify, or 2) the remark was of such a character that a jury would naturally and necessarily take it to be a comment on the accused’s failure to testify.”). However, the prosecutor’s argument at issue was an impermissible reference to the fact that Ellington “didn’t” offer evidence on certain points about which the State had failed to present evidence, see Al-Amin v. State, 278 Ga. 74, 84-85 (597 SE2d 332) (2004), although any reference to his failure to testify was more tenuous. Even assuming this matter was fully preserved for appeal based on the burden-shifting objection Ellington made at trial, the error was harmless beyond a reasonable doubt in light of the totality of the evidence presented and the nature of the prosecutor’s remark, which was brief and prefaced with an emphasis on the defendant’s right not to testify or offer evidence and on the State’s burden of proof. See id. at 85-86.
(b) Ellington claims that the prosecutor made an impermissible “golden rule” argument, urging the jurors to imagine themselves in the place of the victims during the murders and thus improperly evoking the jury’s passion, sympathy, and personal identification *143with the victims. See Braley v. State, 276 Ga. 47, 55 (572 SE2d 583) (2002). The prosecutor argued as follows:
You know the last thing that Berna felt, the last thing that Cameron felt, and the last thing that Christian felt was hard cold steel crashing through flesh and bone from someone who vowed to love and cherish and from someone they called daddy.
He also said, “On behalf of Berna and in memory of Berna, of Cameron, of Christian and on behalf of their family I want to thank you . . . for your attention in this case.”
Ellington did not object to these arguments at trial, and he therefore waived the issue for purposes of appellate review of his convictions. See Gissendaner, 272 Ga. at 713-714. In any event, although the first argument treads close to being improper by referring to what the victims “felt,” we conclude that the argument focused the jury’s attention on the evidence showing the nature of the acts Ellington committed against the victims rather than directly asking jurors to place themselves in the position of the victims, and therefore it was not impermissible. The second quoted argument was unobjectionable. Compare Braley, 276 Ga. at 55 (holding that it was improper for the prosecutor to ask the jurors to “imagine what it was feeling like for [the victim]”).
(c) Ellington contends that the prosecutor made an argument that was contrary to the State’s own expert testimony by suggesting that the fact that Christian Ellington’s eyes were open after his death was an indication that he was awake during his murder. Ellington also failed to object to this argument at trial and therefore waived the issue for purposes of appellate review of his convictions. See Gissendaner, 272 Ga. at 713-714. In any event, we conclude that the argument was not improper, even assuming it was not correct, because it was based upon evidence showing that Christian Ellington’s eyes were open after his death, including crime scene photographs and testimony by the medical examiner. As to what inferences may be drawn from the evidence admitted at trial, we have said that a closing argument may even be “absurd” as long as it does not introduce facts not otherwise in evidence, and the trial court has broad discretion in controlling the scope of closing arguments. See Morgan v. State, 267 Ga. 203, 203-204 (476 SE2d 747) (1996). It was for the jury to decide what inferences from the evidence were reasonable.
*144(d) Ellington argues that the prosecutor made an impermissible, inflammatory argument by characterizing him as “the monster everyone fears” and as “a monster with a hammer,” and by displaying a store-bought hammer to the jury. Because Ellington did not object to these matters at trial, they are waived for purposes of appellate review of his convictions. See Gissendaner, 272 Ga. at 713-714.
Furthermore, although we have characterized arguments using metaphors for a defendant such as “animal” and “snake” as “unnecessary and undesirable,” we have held that allowing them is not reversible error. Simmons v. State, 266 Ga. 223, 228-229 (466 SE2d 205) (1996). See also Brannan v. State, 275 Ga. 70, 85 (561 SE2d 414) (2002) (finding no reversible error in the prosecutor’s characterizations of the defendant as “a coward, a beast, and an animal, and as wicked and evil,” even if they had been objected to at trial). We also see no merit in Ellington’s contention that the brief characterization of him as a “monster everyone fears” improperly injected the issue of his possible future dangerousness, which was irrelevant in the guilt/innocence phase. See McClain v. State, 267 Ga. 378, 383 (477 SE2d 814) (1996) (“The issue of a defendant’s future dangerousness, although relevant to the jury’s sentencing decision, is irrelevant to the question of his guilt.”). Nor was it improper for the prosecutor to display a hammer during his closing arguments, while informing the jury that the hammer was not the actual murder weapon, because the evidence strongly suggested that a hammer had been used as the murder weapon. See Laney v. State, 271 Ga. 194, 198 (515 SE2d 610) (1999) (holding that the use of a prop during closing arguments was not improper where the prop was related to the evidence presented at trial).
(e) In rebuttal to Ellington’s suggestion that Sean Fennell had testified falsely in order to gain the dismissal of a charge against him for making a false statement to investigators in connection with Ellington’s case, the prosecutor argued as follows:
What was done? Sean Fennell’s case was dismissed. I did it. Looked at the facts of the case and it was [eminently clear, just like the facts that [Ellington] killed his family, that Sean Fennell was no more than a patsy.
Why should that man face a felony charge because he’s a patsy? That’s not fair. It’s not right. So, yeah, we dismissed Sean Fennell’s and he’s been in Morehouse and he’s still going and he should graduate soon. And that’s right.
*145Because Ellington did not object to this argument at trial, it is waived for purposes of appellate review of his convictions. See Gissendaner, 272 Ga. at 713-714. Furthermore, although this argument was an improper statement of the prosecutor’s personal opinion about a witness, see Powell v. State, 291 Ga. 743, 746 (733 SE2d 294) (2012), it would not warrant reversal even if Ellington had obj ected to it at trial. It is highly probable, in light of the strong evidence of Ellington’s guilt and of Fennell’s lack of involvement in or prior knowledge of the murders, that this argument did not contribute to the verdict. See Alexander v. State, 263 Ga. 474, 477-478 (435 SE2d 187) (1993) (concluding that a prosecutor had improperly stated her opinion regarding the credibility of a witness but further concluding that it was highly probable that the trial court’s error regarding the argument did not affect the verdict). Compare Bolden v. State, 272 Ga. 1, 2 (525 SE2d 690) (2000) (reversing where an objection was overruled regarding the solicitor’s improper comment on the credibility of the State’s only witness).
Sentencing Phase Issues That are Likely to Recur on Retrial
11. Ellington argues that the pattern jury charges given by the trial court on mitigating evidence were constitutionally inadequate. The jury instructions challenged on appeal are essentially identical to the pattern charges currently in effect, which will likely apply if Ellington faces a new sentencing trial. See Suggested Pattern Jury Instructions, Vol. II: Criminal Cases, §§ 2.15.30, 2.15.50 (2012). Ellington’s contentions are meritless.
(a) First, we reject the claim that the pattern charges are improper because they conflate the terms “mitigating” and “extenuating,” since these terms are essentially synonymous. See, e.g., Black’s Law Dictionary (9th ed. 2009) (defining “extenuate” as “to mitigate”). In addition, the pattern charges direct jurors to consider any “mitigating or extenuating” circumstances that may exist, so any .distinction between the two concepts would be irrelevant. Suggested Pattern Jury Instructions, Vol. II: Criminal Cases, § 2.15.30 (2012) (emphasis added).
(b) Second, the pattern charges do not improperly require the jury to find that any mitigating evidence must be connected directly to the crime itself to be considered. See Tennard v. Dretke, 542 U. S. 274, 287 (124 SC 2562,159 LE2d 384) (2004) (holding that mitigating evidence lacking a “nexus” with the charged offense nevertheless must be admitted). Instead, the pattern charges emphasize that mitigating and extenuating circumstances are factors that do not “constitute a justification or excuse for the offense” but are simply *146factors related to the appropriate “punishment” for the offense and to the general question of the defendant’s “degree of moral culpability or blame.” Suggested Pattern Jury Instructions, Vol. II: Criminal Cases, § 2.15.30 (2012).
(c) Finally, the pattern jury charges are not inadequate in failing to inform the jury that it need not agree unanimously on the existence of particular mitigating circumstances. The charges properly indicate that no mitigating circumstances whatsoever are required to authorize a sentence of less than death; the jury may impose such a sentence for “any reason” or “without any reason.” Suggested Pattern Jury Instructions, Vol. II: Criminal Cases, § 2.15.50 (2012). See Rhode v. State, 274 Ga. 377, 384 (552 SE2d 855) (2001).
12. The trial court did not err by declining to give a jury charge specifically identifying residual doubt as a possible mitigating circumstance. See id.
13. Although the evidence supported the jury’s verdict on the (b) (7) statutory aggravating circumstance, there was a legal deficiency in that verdict and in the trial court’s jury charge, and those deficiencies should be corrected if Ellington is retried as to sentencing. See OCGA § 17-10-30 (b) (7) (“The offense of murder . . . was outrageously or wantonly vile, horrible, or inhuman in that it involved torture, depravity of mind, or an aggravated battery to the victim.”). The (b) (7) aggravating factor has been interpreted as follows:
This statutory aggravating circumstance consists of two major components, the second of which has three subparts, as follows: (I) The offense of murder was outrageously or wantonly vile, horrible or inhuman (II) in that it involved (A) aggravated batter[y] to the victim, (B) torture to the victim, or (C) depravity of mind of the defendant.
Hance v. State, 245 Ga. 856, 860 (268 SE2d 339) (1980). As in several prior cases, the jury’s sentencing verdict here referred to “torture, depravity of mind, or an aggravated battery to the victim.” (Emphasis added.) Because these three subparts were listed in the disjunctive rather than the conjunctive and were not separated on the verdict form to require a finding by the jury as to the individual subparts, it cannot be determined whether the jury agreed unanimously on any one of the three subparts. See, e.g., Rivera v. State, 282 Ga. 355, 366 (647 SE2d 70) (2007).
This defect in the verdict form provided to and used by the jury *147was compounded by the following erroneous charge by the trial court:
Decided November 19, 2012. Jenner & Block, David W. DeBruin, CarlP. Greenberg, Gladys H. Pollard, for appellant. Robert D. James, Jr., District Attorney, Donald P. Geary, Daniel J. Quinn, Assistant District Attorneys, Samuel S. Olens, Attorney General, Patricia B. Attaway Burton, Deputy Attorney General, Theresa M. Schiefer, Dana E. Weinberger, Assistant Attorneys General, for appellee. Hogue & Hogue, Franklin J. Hogue, James C. Bonner, Jr., J. Scott Key, amici curiae.Your verdict should also reflect your finding, if you so find, that the murder involved at least one of the following: torture, depravity of mind[,] or aggravated battery of the victim.
Absent from this charge was the further charge, contained both in the pattern charges that applied at the time of Ellington’s trial and in the pattern charges that will likely apply if he is retried as to sentencing, which directs the jury to “specify which of these was involved in the murder.” Suggested Pattern Jury Instructions, Vol. II: Criminal Cases, § 2.15.30 (2012). This omission should be rectified if this case is retried as to sentencing.
14. We do not address Ellington’s remaining enumerations of error addressing sentencing phase issues, because they relate to issues that are not likely to recur in the event of a retrial.
Judgment affirmed in part and reversed in part, and case remanded.
All the Justices concur, except Melton, J., who concurs in judgment only as to Division 11 (c).The crimes were committed on May 17, 2006. A DeKalb County grand jury indicted Ellington on June 12,2006, charging him with three counts of malice murder. On June 21,2006, the State filed notice of its intent to seek the death penalty. The trial court ruled that interim review was inappropriate in this case, see OCGA § 17-10-35.2 (granting trial courts the discretion to determine that interim review is not appropriate), although several issues discussed in this opinion appear appropriate for such review. Jury selection began on September 8, 2008. The jury convicted Ellington on all three counts on October 18, 2008, and it recommended a death sentence for each of the murders on October 21,2008. The following day, the trial court imposed three death sentences in accordance with the jury’s sentencing verdict. See OCGA § 17-10-31 (a) (“Where a statutory aggravating circumstance is found and a recommendation of death is made, the court shall sentence the accused to death.”). Ellington filed a motion for a new trial on November 13, 2008, which he amended on February 11, 2011, and again on July 11, 2011. After a hearing, the trial court denied Ellington’s amended motion for a new trial on August 19, 2011. After obtaining a 30-day extension of time, Ellington filed a timely notice of appeal on October 18, 2011. The case was docketed for the April 2012 term of this Court, and the case was orally argued on July 9, 2012.
The Jury Composition Act of 2011, Ga. L. 2011, pp. 69-70, § 1-16, replaced the forced balancing approach with a new system for creating grand and traverse jury pools, which took effect on July 1,2012. See also OCGA § 15-12-40.1 (as amended in 2011); UAPII (C) (6), (E) (as amended in 2012).
We assume for purposes of this division that Hispanic persons were a cognizable group. But see Division 6 (d) below.
Their responses on this point may not have been the sole reason for excusing three of these jurors, but it was the only reason given for excusing Juror 32.
Because it is not at issue in this case, we will not discuss the scope of appropriate voir dire questions in a civil case or regarding “the counsel or parties” and “the religious, social, and fraternal connections of the prospective juror,” except to the extent that our decisions on these issues shed light on the question presented here. See, e.g., Napier v. State, 276 Ga. 769, 773-774 (583 SE2d 825) (2003) (holding that the trial court abused its discretion under OCGA § 15-12-133 by prohibiting questions about whether prospective jurors would harbor bias toward the defendants because of their status as prison escapees, although the error was harmless).
See, e.g., Nance v. State, 280 Ga. 125, 127-128 (623 SE2d 470) (2005) (holding that a question “that listed the specific circumstances of [the defendant’s] case and then inquired of the prospective juror whether she could vote for a life sentence under those circumstances” improperly sought a “prejudgment”); Sallie v. State, 276 Ga. 506, 509-510 (578 SE2d 444) (2003) (giving the following as an example of a pre-judgment question that was properly denied in a death penalty case, “In reaching the ultimate decision are you going to be able to factor in matters like... (the defendant’s) age; what the guy’s done in the last 10 years since the offense; what the guy’s military record was like; has he shown indication that the conduct at issue that led to the criminal charges was out of context.. . Would any of those matters, could you take them into account in the sentencing decision(?)”); Lucas v. State, 274 Ga. 640, 646 (555 SE2d 440) (2001) (holding that the trial court did not err in “limiting [the defendant’s] voir dire of prospective jurors on the subject of their willingness to consider specific items of allegedly-mitigating evidence” and preventing the defendant from asking what type of mitigating evidence the jurors would find compelling); Zellmer, 272 Ga. at 736 (holding that proper questions about prospective jurors’ views on parole eligibility as part of a potential sentence in capital cases do not include “[e]xposure to the complexities of the future role of the Board of Pardons and Paroles should the sentence allow for the possibility of parole”); Blankenship v. State, 258 Ga. 43, 45 (365 SE2d265) (1988) (“[N] either the defendant nor the state has the right simply to outline the evidence and then ask a prospective juror his opinion of that evidence. Nor is it permissible to ask a juror to describe the kind of case that, in the juror’s opinion, would warrant a death sentence.”); Atlanta Joint Terminals, 98 Ga. App. at 488 (disapproving of “hypothetical questions embodying a substantial outline of the case proposed” but allowing “questions which only seek to ascertain that the juror’s mind is free of preconvictions”).
See 18 USC § 3592 (c) (11); Ariz. Rev. Stat. Ann. § 13-751 (F) (9); Ark. Code Ann. § 5-10-101 (a) (9) (A); Del. Code Ann. tit. 11, § 4209 (e) (1) (s); Fla. Stat. § 921.141 (5) (1); Idaho *129Code Ann. § 19-2515 (9) (h); Ind. Code § 35-50-2-9 (b) (12); La. Code Crim. Proc.Ann. Art. 905.4 (A) (10); Mont. Code Ann. § 46-18-303 (1) (a) (vi); 42 Pa. Cons. Stat. § 9711 (d) (16); S.C. Code Ann. § 16-3-20 (C) (a) (10); S.D. Codified Laws § 23A-27A-1 (6); Tenn. CodeAnn. § 39-13-204 (i) (1); Wyo. Stat. Ann. § 6-2-102 (h) (ix).
In oases where the State is seeking the death penalty, rather than risking reversal of the case if the trial court is deemed on direct appeal to have abused its discretion, it would be appropriate for the trial court to elicit proposed questions in this area well before voir dire commences and to rule on the issue pretrial so that it could be considered by this Court on interim review if the trial court has any doubt about its rulings. See Wagner v. State, 282 Ga. *138149, 151 (646 SE2d 676) (2007) (addressing a voir dire issue on interim review); Zellmer, 272 Ga. 735 (addressing the proper scope of voir dire on interim review); OCGA§§ 17-10-35.1, 17-10-35.2. In cases not involving the death penalty that present particularly difficult issues in this area, trial court rulings may be appropriate for interlocutory appeal under OCGA § 5-6-34 (b).
Our review of the record reveals that one of these jurors was excused for hardship at the conclusion of his voir dire, rendering any error regarding the scope of his voir dire harmless. See Butts v. State, 273 Ga. 760, 763 (546 SE2d 472) (2001) (“Because it appears that Butts’s suggestion that the juror was ultimately found qualified to serve is false, we find no error.”).
Because we are reversing Ellington’s death sentences, we need not address his contentions about two other prospective jurors, which relate only to their answers regarding the sentencing options.