joined by Justice WAINWRIGHT and Justice JOHNSON, dissenting.
We can all agree that (1) litigants are entitled to fair and impartial jurors, (2) voir dire should not be used as an exercise to preview the verdict, and (3) trial courts must necessarily have broad discretion when conducting voir dire. That said, I do not agree that a trial court can totally divorce the legitimate search for bias and prejudice during voir dire from all material facts in the case. I also disagree with the *764Court’s statement of the issue in this case because this case is not simply about the weight prospective jurors may attach to certain evidence but whether such jurors can follow their oath and the court’s instructions. I believe that the issue is whether the trial court abused its discretion when it cut off questioning about seat belts; specifically questions about whether members of the venire would fairly consider all the evidence in the product liability and wrongful death suit, knowing that the decedent was not wearing her seat belt at the time of the accident.
The trial court here summarily dismissed two jury panels after approximately 60% of the first panel, and 35% of the second, initially indicated that they would not listen to any other evidence in the case upon learning that Amber was not wearing a seat belt at the time of the accident. Before beginning a third voir dire, the trial court devised a new format for questioning the panel, requiring that specific questions about seat belt attitudes be asked only during individual questioning of prospective jurors, and not to the panel as a whole, as had been done previously. When the time came to ask these questions individually, however, the trial court abandoned its plan, forbidding any further questions about seat belts. The Court agrees that this was error, but concludes that the error-was not appropriately preserved for our review. I disagree and therefore respectfully dissent.
I
In order to discover disqualifying bias or prejudice, we have said that litigants must be allowed reasonable latitude in examining prospective jurors on voir dire. Babcock v. Nw. Mem’l Hosp., 767 S.W.2d 705, 709 (Tex.1989). We recently noted, however, that a litigant may not, under the guise of bias and prejudice, ask prospective jurors to pledge or speculate about how specific evidence in the case may influence their- verdict. See Cortez v. HCCI-San Antonio, Inc., 159 S.W.3d 87, 94 (Tex.2005) (attempts to preview a prospective juror’s likely vote not permitted). Such questions, commonly referred to as commitment questions, are improper because they seek to commit prospective jurors to a particular view based on selected facts or evidence disclosed by counsel. See Lassiter v. Bouche, 41 S.W.2d 88, 90 (Tex.Civ. App.-Dallas 1931, writ ref'd). Their purpose is not to uncover a specific bias or prejudice but to identify prospective jurors who are impressed by the perceived strength or weakness of a particular piece of evidence.
The parties to this appeal strongly disagree about whether the rejected question here was an improper commitment question. The Court suggests that it was, and further submits that the Court of Criminal Appeals, which has had more experience in such matters, would agree with that assessment. But the subject of commitment questions remains a fertile area for debate despite the efforts of our Court of Criminal Appeals.
It was a sharply divided Court of Criminal Appeals that recently proposed a complex analysis for determining when voir dire questions seek an improper commitment from prospective jurors. In Standefer v. State, 59 S.W.3d 177 (Tex.Crim.App.2001), the court described commitment questions as those that commit a prospective juror to resolve, or to refrain from resolving, an issue a certain way after learning a particular fact. Id. at 179. The court noted, however, that not all commitment questions are improper. Id. at 181. “When the law requires a certain type of commitment from jurors, the attorneys may ask the prospective jurors whether they can follow the law in that regard.” Id. Thus, the court noted that it might be appropriate to ask in a murder case if the *765prospective juror could consider probation because the law requires jurors in such cases to consider the full range of punishment, including probation. Id. Such a question would not be inappropriate because it merely commits the juror to follow the law. But when the law does not require the commitment, such that a juror would not be disqualified for cause if influenced by a particular fact or if possessed of a particular attitude or opinion, then the question would be improper. Id. at 181-82 (“for a commitment question to be proper, one of the possible answers to that question must give rise to a challenge for cause”). The court further cautioned that even when the question meets the challenge-for-cause requirement, the question may still be improper if it includes facts unnecessary to establish a challenge for cause. Id. at 182. Four members of the court disagreed with the analysis, complaining that “[t]he majority’s attempt to clarify what constitutes a commitment question simply muddies the issue more by attempting to create a bright-line standard.” Standefer, 59 S.W.3d at 186 (Johnson, J. dissenting).
Although I empathize with the Court of Criminal Appeals’ efforts to define standards for this area of the law, I ultimately must agree with the dissent that Standefer provides no bright-line test for distinguishing an improper commitment question from a proper bias question. Nor can I propose an alternative test because I agree with the Court that the process “does not lend itself to formulaic management.” 189 S.W.3d at 750 (quoting 4 ROY W. MCDONALD & ELAINE A. GRAFTON CARLSON, TEXAS CIVIL PRACTICE § 21:19, at 116 (2d ed.2001)).
Clearly, a prospective juror who expresses disqualifying bias for or against a party or the litigation’s subject matter cannot serve, but no single rule can explain the extent to which specific evidence in the particular case may be used to test for such bias. In one sense, the facts of the case are inconsequential because disqualifying bias or prejudice will exist irrespective of the strength or weakness of one’s case. But bias and prejudice cannot be probed in a vacuum, and therefore some discussion of the evidence is inevitable. And it may occasionally happen that a material piece of evidence which strongly favors one party coincides with a bias or prejudice of a particular prospective juror. Because no one rule fits all circumstances, the burden must inevitably fall on the trial court to reasonably manage the voir dire, while guaranteeing the constitutional imperative of a fair trial. Thus, the overriding principle here is that the trial court is vested with broad discretion to control the voir dire, and its rulings should not be disturbed absent a clear abuse of that discretion. Babcock, 767 S.W.2d at 708 (citing Texas Employers Ins. Ass’n v. Loesch, 538 S.W.2d 435, 440 (Tex.Civ.App.-Waco 1976, writ ref'd n.r.e.)).
The question rejected by the trial court here inquired of the venire if they could be fair and impartial to the claim of the un-belted decedent. Admittedly, the question is vague although the Court finds it clear enough that the question’s purpose was to preview the verdict rather than to probe for bias or prejudice. But the trial judge must have understood the question to touch upon disqualifying bias or prejudice at one time or she would not have struck the first two panels. .
Before the third voir dire began, the trial judge indicated that she would again permit the Vasquezes to ask about seat belt bias during individual questioning of the venire. She later changed her mind, however, sustaining Hyundai’s objection to the aforementioned question and stating: “We are not going to go any further into seat belts.” The Vasquezes’ attorney ob*766jected that the court’s ruling prevented him from exploring a relevant bias. The trial judge acknowledged that she understood counsel’s objection, but did not modify her ruling.
A disqualifying bias or prejudice exists whenever a prospective juror’s opinions for or against a party or the subject matter of the litigation are so strong that the juror’s decision will be based on those opinions rather than on the evidence presented at trial. See Cortez v. HCCI-San Antonio, 159 S.W.3d 87, 94 (Tex.2005); see also 1 WILLIAM V. DORSANEO III & EARL JOHNSON, JR., TEXAS CIVIL TRIAL GUIDE § 10.03[2], at 10-15 (2005) (citing Shepherd v. Ledford, 962 S.W.2d 28, 34 (Tex.1998)). The purpose of voir dire is to identify this type of bias or other disqualifying cause and, failing that, to make intelligent use of peremptory challenges. Babcock, 767 S.W.2d at 709. And although a trial court has discretion to impose reasonable limitations on voir dire and may control the form of any question seeking legitimate information, it does not have discretion to simply foreclose a proper line of questioning. See Barajas v. State, 93 S.W.3d 36, 38 (Tex.Crim.App.2002) (abuse of discretion to prevent “a proper question about a proper area of inquiry”). As the Court observes, the trial court could not have denied the Vasquezes the right to ask the venire if anyone had a firm bias against product liability claims that would prevent them from considering the plaintiffs’ claims. 189 S.W.3d at 761. Similarly, any prospective juror in an automobile product liability action who had such strong feelings about unbelted occupants as would preclude that juror from listening to all the evidence and following the court’s instructions would be subject to challenge.
I do not disagree that the question rejected by the trial court, as it was phrased, was an impermissible attempt to pre-test the weight jurors would attach to Amber not being belted. However, this improper question did not authorize the trial court to foreclose the entire area of properly-phrased seat belt usage questions. As the record reflects, there were (and are) a number of ways to pose the subject to the jury. See 119 S.W.3d at 851-54. Not all of them are improper commitment or pretesting questions. But, when a trial court tells a party’s attorney that there will be no further inquiry into a particular subject matter, as was done in this case, that party should not be prejudiced by taking the court’s ruling at face value.
A court abuses its discretion when it renders an arbitrary decision, lacking support in the facts and circumstances of the case. Goode v. Shoukfeh, 943 S.W.2d 441, 446 (Tex.1997); Johnson v. Fourth Court of Appeals, 700 S.W.2d 916, 917 (Tex.1985). The trial court controls the process, Cortez, 159 S.W.3d at 92, and may in its discretion reject questions which pose unnecessary factual detail or otherwise attempt to fish for a prospective juror’s reaction to evidence, but it abuses its discretion when its rulings are arbitrary and go beyond that necessary to protect the integrity of the process. The trial court could have required the Vasquezes’ question to be more precisely drawn, but the court did not have discretion to forbid the area of inquiry altogether. Cf. Howard v. State, 941 S.W.2d 102, 108 (Tex.Crim.App.1996) (litigant must rephrase or waive an improperly phrased question where trial court places no absolute limitation on the underlying substance of the voir dire). The trial court accordingly abused its discretion when it cut off any further questions about seat belts.
The Court, however, concludes that the Vasquezes’ objection to this ruling was not specific enough to preserve error. But rules are not traps. See McConnell v. *767Southside Indep. Sch. Dist., 858 S.W.2d 337, 346 (Tex.1993) (Hecht, J. dissenting) (“When rules are divorced from the basic principles they effectuate, the resulting structure is deformed and arbitrary, and its purpose — achieving justice — is thwarted.”). Preservation rules exist to inform courts of error in a plain and timely manner so that they can make informed rulings. See, e.g., State Dept. of Highways & Pub. Transp. v. Payne, 838 S.W.2d 235, 241 (Tex.1992). That happened in this case when the court acknowledged that it understood the basis for the Vasquezes’ objection and ruled that it would permit no further questions on seat belts. Moreover, Hyundai’s counsel acknowledged during argument to this Court that the trial court was fully apprised of the types of questions the Vasquezes wanted to pursue from the preceding voir dires.
II
Having concluded that the trial court erred in restricting voir dire in this case, I must next consider whether such error was harmful. Hyundai argues that any error was harmless because the Vas-quezes had a fair opportunity to discover bias or prejudice by questioning the panel about their personal seat belt habits. Moreover, Hyundai submits that the court of appeals erred in presuming harm rather than considering whether the error either “probably caused rendition of an improper judgment” or “probably prevented the appellant from properly presenting the case to the court of appeals.” TEX. R. APP. P. 44.1(a); see also TEX. R. APP. P. 61.1(a).
The Vasquezes respond that general questions regarding personal seat belt habits were too broad to be useful in uncovering an existing bias against an injured party who did not use her seat belt. They further respond that presumed harm is the appropriate analysis when a trial court errs in seating a jury during voir dire.
First, I agree that questions about personal seat belt habits were not a reasonable substitute for a direct question about seat belt non-use bias. Just as our Court of Criminal Appeals has cautioned against injecting unnecessary specifics into voir dire questioning, so too has the Supreme Court condemned another extreme — the use of such general inquiries as to deny a party’s right to an adequate voir dire. Morgan v. Illinois, 504 U.S. 719, 734-35, 112 S.Ct. 2222, 119 L.Ed.2d 492 (1992). Too little information in the question can be as damaging to the process as too much. See id. at 729, 112 S.Ct. 2222 (“part of the guarantee of a defendant’s right to an impartial jury is an adequate voir dire to identify unqualified jurors”). Without some connection to the case to “lay bare the foundation of [a] challenge for cause,” the right not to be tried by jurors who harbor a disqualifying bias, or inability to follow the law, is “nugatory and meaningless.” Id. at 733-34, 112 S.Ct. 2222.
Next, regarding harm analysis in this type of case, we said in Babcock that the trial court’s erroneous refusal to permit a party to ask an appropriate question about subject matter bias was harmful because it denied that party’s “constitutional right to trial by a fair and impartial jury.” Babcock, 767 S.W.2d at 709. For this proposition, we cited Texas & Pacific Ry. v. Van Zandt, 159 Tex. 178, 317 S.W.2d 528, 531 (1958), in which we said that the “mere denial [of the constitutional right of trial by jury] raisefd] an inference of probable harm.” Similarly, in another case in which the jury was seated in an improper manner, we adopted the following harmless error analysis as our own:
It would be impossible for appellant to demonstrate with any degree of certainty that he in fact suffered injury as a *768result of the manner in which the jury panel was selected, and he has made no effort to show specific harm or injury, but we think that harm within contemplation of the so-called harmless-error rule ..., must be presumed in the circumstances, if such rule or rules can be said to be applicable to the situation at all. Approval of the judgment would be tantamount to denying appellant his constitutional right of a trial by jury, because trial by a jury that has at least been selected in substantial compliance with law is what is guaranteed him by both the federal and our state constitutions.
Heflin v. Wilson, 297 S.W.2d 864, 866 (Tex.Civ.App.-Beaumont 1956, writ ref'd); cf. Rivas v. Liberty Mut. Ins. Co., 480 S.W.2d 610, 612 (Tex.1972) (harm not presumed when error is procedural and does not violate fundamental right of trial by jury)-
An improperly impaneled jury is also akin to the denial of one. In this latter situation, we have held that the failure to grant a jury trial, when properly requested, is harmless only when “the record shows that no material issues of fact exist and an instructed verdict would have been justified.” Halsell v. Dehoyos, 810 S.W.2d 371, 372 (Tex.1991); see also Mercedes-Benz Credit Corp. v. Rhyne, 925 S.W.2d 664, 667 (Tex.1996). More recently, we noted that an appellate court must presume harm when an objectionable person is erroneously included on a jury because the court “cannot know for certain that his inclusion did not affect the verdict.” Cortez, 159 S.W.3d at 91.
Just as in Cortez, it is impossible to know here whether the verdict would have been different had the jury been seated properly. But as in other cases involving the right to trial by jury, our harm analysis must reflect the importance our justice system accords this precious right. See Gen. Motors Corp. v. Gayle, 951 S.W.2d 469, 476 (Tex.1997) (right to jury trial is “one of our most precious rights”); see also McDaniel v. Yarbrough, 898 S.W.2d 251, 253 (Tex.1995) (“Denial of the constitutional right to trial by jury constitutes reversible error.”). I conclude then that the court of appeals correctly relaxed its harmless error analysis under the circumstances of this case.
Ill
Finally, I find it highly improbable that all of the prospective jurors who responded affirmatively to the seat belt bias questions during the failed voir dires actually harbored disqualifying bias or prejudice against unbelted claimants. It is more likely that their responses were the result of confusion or misunderstanding as the trial court itself acknowledged after the second voir dire. A prospective juror does not demonstrate disqualification merely by raising his or her hand in response to a question asking about bias or prejudice. The court or parties should ordinarily follow up to confirm or dispel any initial notion of disqualification. See Cortez, 159 S.W.3d at 93 (“If the initial apparent bias is genuine, further questioning should only reinforce that perception; if it is not, further questioning may prevent an impartial veniremember from being disqualified by mistake.”).
The voir dire in this case occurred before our decision in Cortez, which may explain the court’s management of the process. No attempt was made to discover what the prospective jurors actually intended when 29 of them, about 60% of the first panel, raised their hands in response to the seat belt question and when 18 of them, about 35% of the second panel, responded similarly. 189 S.W.3d at 748 n. 10. After the second voir dire, the court recognized the problem: “The problem ... *769is I was automatically excluding some people who may have [misjunderstood [the seatbelt] questions ...” The court’s solution to that problem was to permit individual questioning of the prospective jurors, but it abruptly abandoned that plan before it was ever implemented. Instead, the court solved its problem of “automatic exclusion” by simply forbidding any further inquiry into seat belts. In Cortez, however, we admonished that “trial judges must not be too hasty in cutting off examination that may yet prove fruitful.” Cortez, 159 S.W.3d at 92.
Because I believe the trial court’s solution of forbidding further inquiry into the issue of seat belt bias was arbitrary and made without reference to the principles which should have guided the court’s discretion, I would affirm the judgment of the court of appeals and remand this case for trial. Because the Court does not, I dissent.