ROVNER, Circuit Judge.
Denise and Jaron Chlopek sued Breg, Incorporated, the manufacturer of the Polar Care 300, a device that delivers cooling therapy to postoperative patients. After nearly ten days of continuous use of the device on her right foot, Denise experienced decreased blood flow and tissue damage in her right big toe. The Chlo-peks alleged that the Polar Care 300 was defective because it contained no warning about the dangers of continuous use. However, a jury found that the device was not defective. The Chlopeks appeal, arguing that the district court erroneously denied their motion for a new trial. They assert that errors relating to evidentiary matters, the jury instructions, and the district judge’s conduct rendered their trial unfair. Having determined that none of the alleged errors resulted in an unfair trial, we affirm the judgment of the district court.
I.
After undergoing several surgeries on her right big toe in the late 1980s, Denise Chlopek functioned well for more than a decade. She was mobile, experienced minimal pain, and was able to work as a nurse. But in early 2002, she suddenly began to experience severe pain in her right foot, and she sought treatment from Dr. Andrew Pankratz, a podiatrist at the Marsh-field Clinic in Eau Claire, Wisconsin. Dr. Pankratz determined that the artificial joint that had been implanted in Chlopek’s toe in 1989 had fractured and failed. Chlo-pek’s toe was significantly deformed. After discussing two treatment options with Dr. Pankratz, Chlopek elected to have “fusion surgery,” wherein a bone fragment from her hip would be implanted in her big toe to improve stability and weight-bearing capability.
Chlopek underwent surgery at the Eau Claire Surgical Center on May 31, 2002. Dr. Pankratz, along with an orthopedic surgeon, performed the surgery, which was uneventful. After the procedure, Dr. Pankratz prescribed cold therapy for Chlo-pek to prevent swelling and treat pain. Toward that end, he prescribed a Polar Care 300, which he applied to Chlopek’s
foot himself after wrapping the foot in several layers of bandages.
The Polar Care 300, manufactured by Breg, is a device consisting of a cooler attached by a tube to a pad that is wrapped around the patient’s affected area. A pump, powered by electricity, circulates ice water from the cooler through the tube and into the pad, allowing for more cold water to cycle into the pad as heat from the body warms it. Breg markets the Polar Care as “continuous cooling therapy.” The device is a Class II medical device, which, for purposes of this case, means that the Food and Drug Administration must initially approve the device for its intended use and thereafter the manufacturer must report complaints about the device to the agency. The Polar Care 300 is not marketed to the public; rather, Breg sells the devices to medical professionals who in turn prescribe it to their patients. In 2002, the Polar Care 300 had a warning label affixed to its side with the following admonitions:
Caution: Federal law restricts this device for sale by or on the order of a licensed health care practitioner. Warning: Carefully read use instructions and warnings before operation. Warning: Always apply a dressing or other moisture barrier between the pad and the patient’s skin.
Warning: A licensed health care practitioner must consider each patient’s sensitivity to cold. A periodic inspection of the patient’s skin under the pad is recommended. If a noticeable change in skin appearance in the area of the cold application is observed such as burning, itching, blistering, discoloration, or increased swelling more than an hour after use, discontinue use of this product and consult physician immediately. Always apply a barrier or dressing between the pad and the patient’s skin. Caution should be taken during prolonged use, for children, diabetics, incapacitated patients, and those with decreased skin sensitivity or poor circulation.
Warning: Carefully read use instructions and warnings provided with the water circulating pad that is to be used with this pump.
Chlopek was discharged on May 31, the day of her surgery. Dr. Pankratz did not give her any directions with respect to the proper use of the Polar Care 300 and admittedly did not read the warnings on the unit. However, before Chlopek left the clinic, a discharge nurse gave her written instructions to keep her foot elevated, leave the bandage in place at all times, and continue using the cold therapy for two weeks.
Chlopek was due back at the clinic for a follow-up appointment on July 19.
Chlopek followed the instructions that the discharge nurse gave her, but she did not read the warning label on the Polar Care 300. Her husband did read the label, and, concluding the warnings were not inconsistent with the discharge instructions, continued to follow the nurse’s instructions. Accordingly, Chlopek used the Polar Care device continuously, removing the cooling pad only when she went to the bathroom. After nine days, she noticed that her baby toe — which, according to Chlopek, was then visible only because the dressing applied by Dr. Pankratz had shifted — was purple. (Dr. Pankratz maintains that the tips of Chlopek’s toes were
always visible.) She returned to the clinic the next day.
According to Chlopek, when Dr. Pan-kratz removed the bandage on her foot, he exclaimed, “This looks like frostbite!”. Dr. Pankratz instructed Chlopek to discontinue use of the Polar Care 300, but he advised her to apply ice behind her knee three times per day. His treatment notes for the day of his visit do not say anything about frostbite or thermal injury, but he noted blistering and “grayish discoloration” characteristic of ischemia (decreased blood flow) on Chlopek’s toes. Over the next month, Chlopek returned to the clinic for a series of follow-up appointments. The damaged tissue in Chlopek’s toe was beyond repair, and so Dr. Pankratz waited until a clear demarcation became evident between the damaged and healthy tissue. At that point, on July 18, he amputated the damaged tissue, which amounted to Chlo-pek’s entire right big toe.
After the amputation, the Chlopeks sued Breg and its insurer in the Circuit Court of Eau Claire County, Wisconsin. They alleged that the Polar Care 300 was defective because its warning label did not caution consumers that continuous use could lead to thermal injuries. After service the defendants removed the case to federal district court. The Chlopeks survived Breg’s motion for summary judgment, and a jury trial was scheduled. Over the plaintiffs’ objection, the district court decided to bifurcate the trial; only if the jury found during the first phase that Breg’s product was defective would a trial on the Chlo-peks’ damages be held.
Before trial, the district court resolved a number of evidentiary motions. As relevant here, the court granted in large part the defendants’ motion to exclude evidence — in the form of Breg’s records of customer complaints — of other thermal injuries potentially caused by the Polar Care 300 and its sister unit, the Polar Care 500.
The court admitted one of ten incident reports related to the Polar Care 300 upon concluding that the incident occurred under “substantially similar” circumstances to those leading to Chlopek’s injury; the court further decided that admitting evidence involving the Polar Care 500 would waste time and confuse the issues. The district court also excluded evidence that Breg changed the warnings on the Polar Care 300 sometime after Chlopek’s injury; it determined that the evidence was barred under Federal Rule of Evidence 407 as a later remedial measure.
Jury selection began on April 24, 2006. During voir dire, the court asked the potential jurors a number of standard questions relating to their knowledge of the parties, the attorneys, and the facts of the case, and their participation in lawsuits in the past as parties or witnesses. The court inquired generally about the potential jurors’ attitudes toward “the commencement of lawsuits, the administration of justice generally, or jury awards,” but the court did not did not ask whether the potential jurors had strong opinions about tort reform or civil damages. After completing its voir dire, the district court asked the parties if they had additional questions for the potential jurors. The court made a few additional inquiries at the defendants’ behest, and then the parties exercised their challenges. A seven-member jury was empaneled.
The two-day trial began with the testimony of Denise and Jaron Chlopek. The plaintiffs also called Lee Sapetti, a mechanical engineer who testified about the design of the Polar Care 300 and opined
that the product lacked an adequate warning about potential hazards from continuous use. Finally, Dr. Pankratz testified for the plaintiffs, opining that the continuous use of the Polar Care 300 was one of three possible causes of Chlopek’s injury (the others being existing vascular damage from prior surgeries and decreased blood flow caused by cigarette smoking). Among Breg’s witnesses were two experts. The first, an engineer with expertise in thermal science, testified that the Polar Care 300 cannot create skin temperatures cold enough to cause frostbite. The second was Dr. Lance Silverman, an orthopedic surgeon, who opined that the use of the Polar Care 300 was not a cause of the ischemia that led to the amputation of Chlopek’s toe. At one point during the cross-examination of Dr. Silverman, the plaintiffs attorney, Drew Ryberg, grew frustrated with the witness for giving answers when no question had been asked, or for giving unresponsive answers to counsel’s questions. As Dr. Silverman responded to a yes-or-no question with a narrative, Ryberg interjected: “Excuse me, Doctor. This is done in a question- and-answer format and please deal with the questions I ask as we proceed.” In response, the district judge chastised Ry-berg and instructed the jury to disregard his comments:
Please do not argue to the jury. I wish you would disregard counsel’s continual testifying and arguing to the jury. There is a simple procedure which you haven’t learned yet. When an answer comes in that’s nonresponsive, you object thereafter. That’s the gentlemanly way to do it, that’s the ethical way to do it, and that’s the way to do it before this jury.
Ryberg returned to his cross-examination without objecting or asking for a sidebar to protest the judge’s remarks.
At the close of evidence, Breg moved for judgment as a matter of law, and the district court denied the motion. The judge instructed the jury and sent them to deliberate with a special verdict form. The form, which was used over the plaintiffs’ objection, first asked the jury to decide whether the Polar Care 300 was defective. If the jury answered the question in the negative, it was to proceed no further. If it concluded that the Polar Care 300 was defective, it was to go on to determine whether the defect was a cause of Chlopek’s injury. Again, if the jurors answered yes, they were to proceed to a series of questions related to the issue of comparative fault on the part of Chlopek, her husband, Dr. Pankratz, and the Marshfield Clinic.
The jury found at the first step that the Polar Care 300 was not defective and did not answer the remaining questions. On the basis of the special verdict, the district court entered judgment for the defendants. Shortly thereafter the Chlopeks filed a motion for a new trial.
See
Fed. R.Civ.P. 59(e). They argued that the trial was unfair for several reasons. First, they asserted that they had been prejudiced by the bifurcation of the trial and the district court’s failure to explain the procedure to the jury. Second, they challenged the court’s exclusion of evidence of other accidents and of the changes to the Polar Care 300’s warning label. Third, they argued that they had been prejudiced by the judge’s rebuke of attorney Ryberg during the cross-examination of Dr. Silverman. Finally, they maintained that the special verdict form confused the issues in the case and unfairly invited the jury to answer the first question in the negative and “go home early.” The district court, unpersuaded by any of the arguments, denied the motion.
II.
On appeal the Chlopeks renew most of the arguments they made in their
motion for a new trial, although they address them in a rather cursory manner. We review the denial of the motion for a new trial for abuse of discretion.
See Naeem v. McKesson Drug Co.,
444 F.3d 593, 605 (7th Cir.2006). We note that, although the substance of this case was governed by Wisconsin law, the issues the Chlopeks have raised on appeal are all evidentiary or procedural, and we therefore apply federal law.
See Schindler v. Seiler,
474 F.3d 1008, 1010 (7th Cir.2007);
Bevolo v. Carter,
447 F.3d 979, 982 (7th Cir.2006).
The Chlopeks first argue that the district court committed reversible error when it excluded most of their proffered evidence of other injuries involving the Polar Care 300 and all the evidence they sought to introduce of injuries involving the Polar Care 500. In a products-liability case, evidence of other accidents is relevant to show the existence of a danger, the defendant’s notice of the danger, and the cause of the accident.
Weir v. Crown Equip. Corp.,
217 F.3d 453, 457 (7th Cir.2000) (citing
Nachtsheim v. Beech Aircraft Corp.,
847 F.2d 1261, 1268 (7th Cir.1988)). However, before the evidence will be admitted, the proponent must show that the other accidents occurred under substantially similar circumstances.
Id.
As we explained in
Nachtsheim,
this showing is necessary because the probative force of evidence of other accidents decreases “[a]s the circumstances and conditions of the other accidents become less similar to the accident under consideration.”
Nachtsheim,
847 F.2d at 1268. At the same time, the risk of unfair prejudice remains, along with potential costs in terms of time and distraction.
Id.
The district court did not abuse its discretion by excluding the evidence. First, we note that the court held a hearing at which it individually considered each accident report relating to the Polar Care 300 that the plaintiffs sought to admit. These ten reports, comprising Exhibit 10, depicted complaints of redness, blistering, “burns,” “nerve damage,” “frostbite,” and other symptoms, on the patients’ knees, feet, or legs. The district court admitted Exhibit 10-C, the report of an injury under circumstances the court deemed most similar to Chlopek’s: a complaint of frostbite on the foot and ankle area after the patient used the Polar Care 300 continuously after surgery. The remaining incidents, the court determined, were not similar enough, either because the injury was to another body part, the type of injury was unclear or not of the same nature as Chlopek’s, or the nature of the complaint was different (for example, some consumers simply complained of the absence of temperature control). The Chlopeks do not address the reports individually; they simply state that the district court “incorrectly determined that [the other accidents] were not a of a similar kind” and that Exhibit 10 “should have been admitted in its entirety.” Such a cursory argument cannot support a reversal of the district court’s decision, particularly where the court discussed each report separately and gave sound reasons for distinguishing the other incidents from the one at issue. Moreover, we note that the plaintiffs’ theory of the case was failure-to-warn, but no information was supplied about the warnings given in the other cases; a crucial point of comparison was lacking.
See Weir,
217 F.3d at 458.
As for Exhibit 11, the plaintiffs do not seriously challenge the exclusion of the accident reports related to the Polar Care 500; they do not explain why any of the reported incidents was substantially similar nor do they respond to the district court’s reasons for exclusion. They state only that the Polar Care 500 was equally capable of producing frostbite and was
comparable for purposes of the trial. In fact, the district court agreed that the Polar Care 500 was in many respects similar to the Polar Care 300 but determined nevertheless that it was inadvisable to “try the 500” when the issue of the case was the adequacy of the warnings on the Polar Care 300. We see no abuse of discretion in the district court’s determination. As the court suggested, exploring injuries potentially caused by the Polar Care 500 could have confused the jury and unnecessarily prolonged the trial. See Fed. R.Evid. 403;
Guy v. Crown Equip. Corp.,
394 F.3d 320, 329 (5th Cir.2004).
The final evidentiary decision the plaintiffs challenge is the exclusion of evidence that Breg changed the warning label on the Polar Care 300 some time after Chlopek’s injury; the new label cautions against continuous use for more than 12 hours. They seek to sidestep Federal Rule of Evidence 407 by insisting that the change was not a subsequent “remedial” measure because, according to the affidavit of a Breg executive, the change was not prompted by safety concerns. But Breg’s
motive
for making the change is irrelevant. All the rule requires is that the measure “would have made the injury or harm less likely to occur.” Fed.R.Evid. 407. Regardless of Breg’s stated reason for the change, the plaintiffs undoubtedly wanted the
jury
to conclude that Breg added the warning because the product was unsafe without it. That is precisely the type of inference that Rule 407 forecloses, in order to avoid discouraging the taking of remedial measures.
See Pastor v. State Farm Mut. Auto. Ins. Co.,
487 F.3d 1042, 1045 (7th Cir.2007);
Lust v. Sealy, Inc.,
383 F.3d 580, 585 (7th Cir.2004).
The plaintiffs get no further by arguing that the evidence of the changed warning nevertheless was admissible because it was relevant to causation. The fact of a new warning does not tend to prove that the absence of an adequate warning caused Denise Chlopek’s injury; rather, it is relevant to whether continuous use of the product can cause injury. However, Breg never argued that its product could not cause an injury like Chlopek’s; its position was that with proper use (including heeding the existing warnings) on appropriate candidates the product was safe. In any event, the district court found that evidence of a changed warning label was excludable for the additional reason that it would be unfairly prejudicial,
see
Fed.R.Evid. 403, and this determination was not an abuse of discretion.
Next the plaintiffs challenge the bifurcation of the trial and, relatedly, the omission of an instruction to help the jury “delineate the role of the two trials.” Under Federal Rule of Civil Procedure 42(b), a district judge may separate claims or issues for trial if the separation would prevent prejudice to a party or promote judicial economy.
See Houseman v. U.S. Aviation Underwriters,
171 F.3d 1117, 1121 (7th Cir.1999). If one of these criteria is met, the district court may order bifurcation as long as doing so will not prejudice the non-moving party or violate the Seventh Amendment.
Krocka v. City of Chi.,
203 F.3d 507, 516 (7th Cir.2000);
Houseman,
171 F.3d at 1121. The Chlo-peks do not dispute that bifurcation served the interest of judicial economy and did not result in a violation of the Seventh Amendment, and so the only issue is prejudice.
The Chlopeks contend that the bifurcation prevented them from presenting evidence that was “mixed in character,” namely, evidence “that other aspects of tissue below the knee were affected.” But, contrary to the Chlopeks’ assertion, evidence about the full extent of Denise’s injuries was not relevant to the only fact at issue in the first phase of trial: whether
the Polar Care 300 was defective because of inadequate warnings.
See Krocka,
203 F.3d at 516;
Hydrite Chem. Co. v. Calumet Lubricants Co.,
47 F.3d 887, 890-91 (7th Cir.1995) (explaining that in bifurcated proceedings, “the fact of injury belongs in the first trial and the quantification of the injury by means of an assessment of damages in the second”). As the district court noted when denying the Chlopeks’ motion for a new trial, the jury was well aware that Denise suffered injuries to her toe resulting in its amputation. An in-depth presentation about the full extent of her injuries was best left for the damages phase so as not to cloud the issues or prejudice the defendant. And the plaintiffs’ contention that bifurcation results in a proceeding that is “too expensive” is a non-starter; they were not financially prejudiced by a second phase of trial that never occurred.
As for the contention that the district court erroneously failed to instruct the jury on the issue of bifurcation, we note that the plaintiffs argue from the faulty premise that Rule 42(b) requires the district court to instruct the jury
about
bifurcation. The jury instructions in this case appropriately focused on the issue of liability and omitted any mention of damages; the instructions were therefore sufficient to ensure that the jury during the liability phase knew its role. In any case, the plaintiffs did not propose an instruction on bifurcation, and so they are not entitled to any relief.
See Penn v. Harris,
296 F.3d 573, 577 (7th Cir.2002).
The plaintiffs present as a related error the district court’s use of a special verdict form that, they say, distorted the issues and contained “random inquiries” that prejudiced their case. According to the Chlopeks there was “no evidence” presented at trial regarding the negligence of Jaron Chlopek, Dr. Pankratz, or the Marshfield Clinic, and so the special ver-diet form should not have mentioned them. The Chlopeks also take issue with the format of the document, which, they assert, invited the jury “to end work early.” The plaintiffs refer to the instruction to proceed no further if the jury answered “no” to the question of whether the Polar Care 300 was defective.
The formulation of special verdict questions is a matter of the trial court’s discretion. Fed.R.Civ.P. 49(a);
Mattson v. Schultz,
145 F.3d 937, 939 (7th Cir.1998). However, the verdict form must “accurately, adequately, and clearly state the relevant issues,” and “[a]mbiguous, biased, misleading, or confusing questions may warrant reversal.”
Mattson,
145 F.3d at 939. Here, the special verdict form first asked the jury to decide if the Polar Care 300 was “in a defective condition so as to be unreasonably dangerous to a prospective user.” An answer of “no” would end deliberations, while an affirmative response would lead the jury to decide whether the defect was a cause of Denise Chlopek’s injury. Only after answering “yes” to the first two questions would the jury then confront issues of comparative fault. The district court’s formulation of the questions did not exceed the bounds of its discretion. Although the first question could have presented the plaintiffs’ theory of the case more specifically by linking the defective condition to the absence of adequate warnings, the jury instructions covered that ground. By inquiring first into the condition of the Polar Care 300 and second into whether the product caused Denise Chlopek’s injury, the special verdict form focused the jury on the plaintiffs theory, and so the questions “accurately, adequately, and clearly” stated the main issue of the case.
The plaintiffs’ argument that the verdict form improperly invited the jury to answer the first question in the negative
and “go home early” is confusing. Surely they do not mean to contend that the jury should have gone on to answer questions related to causation and comparative fault after deciding that the Polar Care 300 was not defective. And, contrary to the plaintiffs assertion that “no evidence” was presented regarding the negligence of any parties with the possible exception of Denise Chlopek, the jury reasonably could have inferred from the testimony at trial that the negligence of third parties caused or contributed to Chlopek’s injuries. The inquiries, therefore, were not confusing and “random” as the plaintiffs contend, and the mere presence of these questions did not prejudice them. The plaintiffs point to no evidence to overcome the strong presumption that juries follow instructions,
see 3M v. Pribyl,
259 F.3d 587, 600 (7th Cir.2001), and we will not conclude that the jury approached the verdict form out of order, or with an eye towards an early dismissal, based solely on the plaintiffs’ speculation.
The Chlopeks next challenge the district court’s failure to specifically inquire during voir dire about the potential jurors’ “perceptions ... regarding the propriety” of civil damages cases, “particularly in light of the assault on the civil jury system conducted by many politicians throughout America.” The plaintiffs do not cite any authority for the proposition that a judge must investigate potential jurors’ views on tort reform. Indeed we have rejected this very argument, as have other courts.
See Alcala v. Emhart Induss., Inc.,
495 F.3d 360, 363(7th Cir.2007);
Smith v. Tenet Healthsystem SL, Inc.,
436 F.3d 879, 884 (8th Cir.2006);
Smith v. Vicorp, Inc.,
107 F.3d 816, 818 (10th Cir.1997). The trial court has broad discretion over the selection of questions to potential jury members, and the parties have no right to have a particular question posed.
See Gardner v. Barnett,
199 F.3d 915, 920-21 (7th Cir.1999). The parties may request that the court explore certain biases, but the court is under no obligation to root them out
sua sponte
— with certain exceptions not relevant here.
See United States v. Montenegro,
231 F.3d 389, 394 (7th Cir.2000). In this case, the district court asked generally if any potential juror had an opinion about “the commencement of lawsuits, the administration of justice generally, or jury awards which would in any way affect your respective abilities to serve as a fair an impartial juror in this case.” Receiving no affirmative responses, the judge did not inquire further. The questions that the plaintiffs proposed before trial were only slightly more specific than the judge’s on the subject of tort reform, and the court’s failure to formulate more detailed questions on its own was hardly an abuse of discretion.
Next the plaintiffs argue that they were prejudiced by the district judge’s criticism — in the presence of the jury — of attorney Ryberg’s cross examination of Breg’s expert, Dr. Silverman. They contend that by suggesting that Ryberg was not handling the witness’s unresponsiveness in an “ethical” or “gentlemanly” way, the judge irreparably damaged the plaintiffs’ case. We rarely will reverse a judgment because of the district court’s conduct at trial.
See Cooper v. Casey,
97 F.3d 914, 918 (7th Cir.1996). The rules of evidence give a district judge wide latitude to control proceedings so as to “(1) make the interrogation and presentation effective for the ascertainment of the truth, (2) avoid needless consumption of time, and (3) protect witnesses from harassment or undue embarrassment.” Fed.R.Evid. 611(a);
see United States v. Reynolds,
189 F.3d 521, 529 (7th Cir.1999). Generally speaking, a judge should be impartial and should reserve criticism for times when the jury is not present,
United
States v. Mohammad,
53 F.3d 1426, 1434 (7th Cir.1995), but reversible error occurs “only when the judge so impairs the lawyer’s credibility in the eyes of the jury as to deprive the client of a fair trial,”
Cooper,
97 F.3d at 919.
We are not persuaded that the one instance of judicial intemperance the plaintiffs cite evinced bias or resulted in prejudice. Although the judge’s comment perhaps was itself unfortunate, it was hardly the “lacerating critique” or the “attack” the plaintiffs label it. The court’s underlying point — that attorney Ryberg should have moved to strike Dr. Silverman’s unresponsive answer rather than chastise the witness in front of the jury — is well-taken, although the suggestion that Ryberg’s conduct was not “ethical” was regrettable. More importantly, the primary setback that the plaintiffs identify is not borne out by the record. The Chlopeks assert that counsel was prevented from pursuing questions designed to bring to light Dr. Silverman’s “bias.” Having reviewed the transcript, however, we are convinced that counsel adduced that Dr. Silverman charged exorbitant fees to render his opinion that the Polar Care 300 did not cause Chlopek’s injury. The judge’s scolding did not obscure this point — indeed, counsel had already moved on. We also note that counsel quite eloquently asked the jury during his closing argument to focus on the substance and not the form of his questioning of Dr. Silverman. This statement, combined with a jury instruction admonishing the jurors not to take any of the judge’s statements as a signal of his opinion on the merits of either party’s case, reduced the possibility of prejudice.
See Reynolds,
189 F.3d at 529. Although the judge’s response to counsel’s misstep could have been more measured, it did not result in prejudice.
See Mohammad,
53 F.3d at 1434.
III.
The Chlopeks have not identified any errors that necessitate a new trial. Accordingly, we affirm the judgment of the district court.