Defendant appeals a judgment based on a jury verdict that awarded plaintiff Richard Schwarz, the personal representative of the estate of his wife Michelle Schwarz (decedent), compensatory damages totaling $168,514.22 and punitive damages of $150 million against defendant Philip Morris Incorporated.1 Plaintiff alleged that decedent’s death resulted from metastatic lung cancer that was caused by her smoking cigarettes that defendant manufactured and promoted as “low-tar.” Defendant makes numerous assignments of error on appeal that we discuss below in detail. Plaintiff cross-appeals, challenging the trial court’s reduction of the jury’s punitive damages award from a total of $150 million to $100 million. On appeal, we affirm the portion of the judgment that finds defendant liable for compensatory damages. However, we vacate the judgment for punitive damages on all claims and remand for a new trial on the amount of those damages. We dismiss the cross-appeal as moot.

We state the facts in the light most favorable to plaintiff because of the verdict in his favor. Or Const, Art VII (Amended), § 3; Jensen v. Medley, 336 Or 222, 226, 82 P3d 149 (2003). Decedent began smoking in 1964, when she was 18 and a student in nursing school. Plaintiff and decedent were married in 1965, the year after they met. After they were married, decedent did not complete her nursing studies but instead raised the couple’s children. When decedent started smoking, she knew that there was a potential link between cigarettes and illness, including lung cancer. Her parents had discouraged her from smoking, in part because of health concerns. Plaintiff quit smoking several years after he and decedent were married and he encouraged decedent to follow his example. However, throughout their marriage, she was unable to quit despite a number of attempts to do so, and she continued to smoke about a pack of cigarettes a day.

Decedent first smoked Benson & Hedges cigarettes, a full-flavor brand manufactured by defendant.2 In 1976, *24defendant introduced Merit cigarettes to the public with an extensive advertising campaign that emphasized that Merits had less tar than full-flavor brands but, according to the advertising campaign, tasted like full-flavor brands. At about the same time, plaintiff and decedent discussed her quitting smoking. Decedent suggested that, rather than quitting, she would try a low-tar cigarette; plaintiff agreed, believing that switching brands would be a step toward weaning her off cigarettes entirely. Soon after their discussion, decedent began smoking the Merit brand instead of Benson & Hedges. According to her mother’s testimony, decedent switched to the Merit brand because she believed that “the low tar and nicotine filters are better for you,” an idea that others talked about at the time. Indeed, decedent’s stepfather had previously switched to a low-tar brand because he believed that they were safer than full-flavor cigarettes.

After switching to the Merit brand, decedent continued to smoke the same number of cigarettes as before. However, according to the testimony at trial, her method of smoking changed; she took longer puffs, inhaled the smoke more deeply, and held it longer in her lungs. She began smoking each day early in the morning and continued until late at night. She could not go without smoking a cigarette for more than an hour and a half without feeling deprived. Being without cigarettes made her act nervous, edgy, and irritable; smoking a cigarette would cause her to become calm and serene. On one long international airplane flight, she was so obviously upset from the lack of a cigarette that a flight attendant showed her how to smoke in a rest room without setting off the smoke alarm.

After switching to the Merit brand, decedent made further attempts to stop smoking entirely, including using nicotine patches, but she was always unsuccessful. She once told a physician that she had stopped smoking for six months, but plaintiff testified that she never actually quit. An expert on addictive substances testified unequivocally that decedent was addicted to nicotine. Decedent was diagnosed with a brain tumor in February 1998 that proved to be the result of metastatic lung cancer. Despite treatment and a period of temporary remission, she died on July 13, 1999, at *25the age of 53. At her son’s wedding shortly before her death, while she was in a wheelchair and on oxygen, she begged her mother for a cigarette.

Plaintiff brought this action against defendant for damages in February 2000. The case went to trial in February 2002 on plaintiffs third amended complaint, in which he alleged claims based on theories of strict products liability, negligence, and fraud.3 The gist of his allegations was that (1) the Merit brand of cigarettes was unreasonably dangerous in a manner that was not contemplated by the consumer because it was marketed as a less harmful alternative to ordinary cigarettes; (2) defendant had been negligent in the manner in which it tested, manufactured, and marketed the Merit brand; and (3) defendant had defrauded consumers by making false claims about the health effects of the Merit brand, the contents of the brand itself, and its addictive nature.

After a lengthy trial, the jury returned a verdict for plaintiff on all his claims. As to the strict products liability and negligence claims, it apportioned fault between defendant (51 percent) and decedent (49 percent). As to the fraud claim, the jury found in defendant’s favor on two specifications of fraud and in favor of plaintiff on two other specifications. It awarded compensatory damages consisting of $118,514.22 in economic damages and $50,000 in noneconomic damages. The jury also awarded punitive damages of $10 million on the strict liability claim, $25 million on the negligence claim, and $115 million on the fraud claim, for a total punitive damages award of $150 million. The trial court entered judgment for the full amount of the compensatory damages, reasoning that apportionment of fault did not apply to the fraud claim. It also reduced the total award of punitive damages to $100 million without apportioning it among the various claims. As stated above, both parties appeal.

*26On appeal, defendant makes 21 assignments of error. They include challenges to the denial of motions for directed verdicts for defendant on plaintiffs claims, challenges to the trial court’s decision to give or not to give certain jury instructions regarding those claims, a challenge to the court’s failure to grant a mistrial, and challenges to the jury instructions pertaining to, and the constitutionality of, the punitive damages awards. We reject without discussion any assignments of error and any arguments not addressed below. We begin with the assignments of error that pertain to the judgment on plaintiffs fraud claim.

I. THE FRAUD CLAIM

In his third amended complaint, plaintiff, after incorporating other allegations in the complaint, alleged:

“13.

“Defendant * * * recklessly and/or intentionally made fraudulent misrepresentations about its tobacco products, including misrepresentations about adverse health effects, the addictive nature of its tobacco products, and their contents.

“14.

“Defendant * * * engaged in an ongoing public relations effort beginning in the early 1950s, designed to manipulate public opinion by creating doubt about the adverse health effects of smoking and to provide rationalizations to help smokers keep smoking in spite of the adverse health effects. Defendant * * * made statements which were intended to and did cause cigarette smokers such as [decedent] to continue smoking cigarettes in spite of their adverse health effects. Defendant voluntarily assumed a duty to disclose all research.

“15.

“[Decedent] did not know defendant[’s] representations were false and reasonably relied on, and suffered and died as a result of defendant [’s] misrepresentations.

“16.

“Defendant [’s] misrepresentations included the following and similar misrepresentations:

*27“a. That the causal link between cigarette smoking and human disease was in doubt or ‘had not been proven’ in repeated statements during the past 50 years;

“b. That cigarettes are not addictive; and

“c. That ‘low tar’ cigarettes delivered less tar and nicotine to the smoker and were therefore safer and healthier than regular cigarettes as an alternative to quitting smoking.”

Answering specific interrogatories, the jury found in plaintiffs favor on two theories of fraud. First, it found that defendant voluntarily assumed a duty to disclose all research regarding smoking and health to consumers, that it breached that duty by concealing research from consumers, that decedent, one of defendant’s consumers, reasonably relied on defendant’s performance of its assumed duty, and that those representations and reliance were a cause of her death. Second, the jury found that defendant falsely represented that “low-tar” cigarettes delivered less tar and nicotine to the smoker and that they were therefore safer and healthier than regular cigarettes and presented an alternative to quitting smoking, that decedent relied on those representations, and that those representations and her reliance on them were a cause of her death. Specifically, the jury answered “yes” to the following questions:

“9. Did Philip Morris voluntarily assume a duty to disclose all research regarding smoking and health to [decedent]?

* * * *

“10. Did Philip Morris fail to perform its assumed duty by concealing research, did [decedent] reasonably rely upon defendant’s performance of its duty, and was such failure to perform and reliance a cause of [decedent’s] death]?]

% * * *

“13. Did Philip Morris make false representations that ‘low tar’ cigarettes delivered less tar and nicotine to the smoker and were therefore safer and healthier than regular cigarettes and an alternative to quitting smoking upon which [decedent] reasonably relied, and if so, were such *28false representations and reliance a cause of [decedent’s] death?”

Defendant first assigns error to the trial court’s denial of its motion for a directed verdict on plaintiffs fraud claim, ORCP 60, offering several arguments in support of its position. We begin by addressing the issue of preemption by federal law. On appeal, defendant first argues that the “low-tar” fraud specification is preempted. Later in its brief, in a footnote, it asserts that “[p]laintiffs fraud claim for breach of an assumed duty also is expressly preempted!.]” But defendant did not raise the issue of whether the assumed duty fraud specification is preempted in its motion for a directed verdict to the trial court. We therefore decline to address that argument as it applies to plaintiffs fraud claim for breach of an assumed duty. See ORAP 5.45(1) (embodying preservation-of-error rule); State v. Thompson, 328 Or 248, 254 n 3, 971 P2d 879, cert den, 527 US 1042 (1999) (refusing to address claim absent a “thorough and focused” analysis). We return to the issue that defendant did preserve.

As noted, the jury found that defendant fraudulently represented that low-tar cigarettes were safer than full-flavor cigarettes. As it did in its motion for a directed verdict to the trial court, defendant argues under its first assignment of error that plaintiffs “low-tar” fraud claim is preempted by federal law. It explains, “[PJlaintiff s claim necessarily boils down to this: defendant was at fault for failing to disclose information about the health risks of low-tar cigarettes.” (Emphasis in original.) The statutory provision relied on by defendant, 15 USC section 1334(b), is part of the Federal Cigarette Labeling and Advertising Act. That section provides:

“No requirement or prohibition based on smoking and health shall be imposed under state law with respect to advertising or promotion of any cigarettes the packages of which are labeled in conformity with the provisions of this chapter.”

The leading case on the subject of federal preemption of tobacco-related claims is Cipollone v. Liggett Group, Inc, 505 US 504, 112 S Ct 2608, 120 L Ed 2d 407 (1992), in which the United States Supreme Court held that the congressional preemption of state law under section 1334(b) *29“sweeps broadly and suggests no distinction between positive enactments and common law; to the contrary, those words easily encompass obligations that take the form of common-law rules.” 505 US at 521 (plurality opinion).4 The Cipollone Court considered a number of different common-law claims against cigarette manufacturers and determined whether each was preempted by section 1334(b). We consider defendant’s argument that plaintiffs low-tar specification of fraud is preempted by section 1334(b) under the guideposts set out in Cipollone, viewing the evidence relating to the fraud claim generally in the light most favorable to plaintiff.

The possibility of a connection between smoking and disease first received significant public attention in 1952 through an article in Reader’s Digest, “Cancer by the Carton,” that discussed then-recent research with mice and suggested that smoking caused lung cancer. The next year, overall cigarette sales fell significantly for the first time. As a result, the leading manufacturers in the tobacco industry, including defendant, met and agreed on a common response to the health issue. Their primary public action was to create the Tobacco Industry Research Council (TIRC). In 1954, the TIRC’s creators, including defendant, first published as an advertisement in over 400 newspapers, including newspapers in Oregon, a statement entitled “A Frank Statement to Cigarette Smokers” (the Frank Statement). In the Frank Statement, the industry stated that it recognized public concern about recent reports “that cigarette smoking is in some way linked with lung cancer in human beings.” Although the Frank Statement emphasized that the research was not conclusive, it also agreed that the research should not “be disregarded or lightly dismissed.” Instead, the industry stated that it would “meet the public’s concern aroused by the recent reports” by pledging “aid and assistance to the research effort into all phases of tobacco and health.” To accomplish that purpose, defendant and other manufacturers of tobacco products created the TIRC and placed it in charge of the *30industry’s research activities. The industry represented that “a scientist of unimpeachable integrity and national repute,” along with an advisory board of “scientists disinterested in the cigarette industry” would conduct the research effort. The signers of the Frank Statement, including defendant, explained that “we believe the people are entitled to know where we stand on this matter and what we intend to do about it.” Soon afterwards, one of defendant’s vice-presidents stated in a publicized speech that the industry would “stop business tomorrow” if it thought that its product was harming smokers.

In 1962, The Tobacco Institute, Inc., the public relations successor to the TIRC, prepared another advertisement, “Some frank words about, [sic] Smoking and Research,”5 in which it stated, in part:

“Most scientists recognized long ago that there are no simple, easy answers in cancer research. They know that the answers to fundamental questions about causation can come only through persistent scientific research.

“The tobacco industry supports and cooperates with all responsible efforts to find the facts and bring them to the public.

“In that spirit, we are cooperating with the U.S. Surgeon General and his special study group appointed to evaluate presently available research knowledge. Similar cooperation has been offered to the American Medical Association’s proposed study.

“We know we have a special responsibility to help scientists determine the facts about tobacco use and health.

“The industry accepted this responsibility in 1954 by establishing the Tobacco Industry Research Committee to provide research grants to scientists in recognized research institutions. This research program is continuing on an expanded and intensified scale.”

(Emphasis added.)

*31In 1966, the Tobacco Institute issued a press release in response to a forthcoming article on tar and nicotine in cigarettes in which it said, in part:

“Scientists throughout the world are continuing to investigate to learn the full facts about ‘tar’ and nicotine, and about questions concerning tobacco and health. The tobacco industry is supporting much of this research and will continue to do so.”

Based on the evidence before it, the jury in this case was entitled to infer that defendant and the industry as a whole did not conduct the research that they represented that they would conduct.6 Rather, the evidence produced by plaintiff shows that they focused their efforts and their health-related research on attacking research that showed the dangers of smoking in order to keep alive a public controversy over whether tobacco smoke was harmful to the human body. The jury was entitled to infer that defendant’s purpose was to give smokers what one of defendant’s executives described as a “crutch” — essentially a rationalization — that would justify their continued smoking. The jury was also entitled to find, based on the evidence produced by plaintiff at trial, that defendant knew throughout this time period that tobacco smoke was a carcinogen and that nicotine was addictive — indeed, that nicotine addiction was the primary reason why smokers continued to smoke. Rather than making its research public as it had represented that it would do, defendant publicly denied and suppressed the results of its research.

Moreover, plaintiff offered evidence that defendant avoided studying the health effects of smoking at the very time that it was insisting on the need for additional research on that subject. For example, defendant’s scientists were not allowed to conduct studies on actual production cigarettes or to conduct tests of the effects of those products on animals. Any research conducted by defendant that could involve the “biological activity” of tobacco smoke (a euphemism for its carcinogenic properties) was conducted at a laboratory in *32Europe. The results of that research were forwarded to only one supervising scientist in the United States at his home address; thus, the research never became part of defendant’s official records, nor was it publicized as defendant had represented would occur. In fact, that particular scientist told a subordinate that his role was to keep the controversy over the health effects of smoking alive.

Rather than focusing its research on issues of health, as it had represented that it would do, defendant focused on research that led to the modification of the tobacco in the cigarettes it produced. One of those research efforts included adding substances that would increase the impact of the nicotine in the cigarettes on the smoker, thereby increasing their addictive effect. Although defendant manipulated the tobacco itself and added a number of flavorants, the most significant additives were ammonia and urea. Those substances, according to plaintiffs evidence, have the effect of increasing the proportion of “free base” nicotine in the smoke from the cigarette; that reduces the amount of nicotine that is bound to particulates and is more difficult for the body to absorb. The result is an increase in the effect of the nicotine on the human body, leading to greater addiction.

According to plaintiffs experts, smokers continue smoking because of nicotine’s addictive qualities. The jury could find based on the evidence from defendant’s internal files that defendant was well aware of that fact. Defendant’s research indicated that smokers develop a certain “comfort level” of nicotine and will smoke until that level is reached. Nicotine, according to the evidence, does not itself cause cancer; rather, it is the “tar” — the solid matter in the cigarette smoke — that produces that result. The amount of tar that a cigarette delivers, however, tends to be proportionate to the amount of nicotine in the smoke. During the 1940s and 1950s, the Federal Trade Commission (FTC), working with the tobacco industry, developed what remains the standard method for measuring the amount of tar and nicotine in a cigarette. The FTC created a machine that smokes a cigarette in a standard fashion to a standard length. It is possible through the use of the machine to measure the tar and nicotine that a particular cigarette produces and to use those *33measurements to develop a relative ranking of the amounts of tar and nicotine in each brand of cigarettes. Although the FTC does not classify specific brands as high or low in tar content, the FTC measurements that the machine produces have appeared on cigarette packages for decades, including the packages for the Merit brand.

One inherent problem with the FTC measurements is that they do not account for the manner in which smokers actually ingest the smoke from a cigarette. For example, a person who takes longer or more frequent puffs than the machine takes will receive a higher dose from each cigarette than the machine indicates. According to plaintiffs evidence, that problem in the FTC measurement process is significant with regard to low-tar cigarettes such as Merit because smokers compensate for the lower nicotine level that such cigarettes contain. The jury could have found that defendant’s understanding that smokers tend to compensate for lower nicotine levels played a major role in the development of its Merit brand.

Based on its research, defendant made a number of modifications to its low-tar cigarettes, including using expanded tobacco, reconstituted tobacco, specially treated cigarette paper, and certain additives, and placing microscopic ventilation holes in the paper.7 The cumulative effect of those modifications was to increase the amount of air in each standard puff that the FTC machine measured and to reduce the amount of tar and nicotine that it measured. Those modifications resulted in more favorable comparisons with other brands regarding tar content. However, the jury could also have found from the evidence that defendant knew that the FTC measurements would mislead consumers because of the ways in which smokers compensate for the reduction in tar and nicotine in brands like Merit. We therefore discuss in more detail what the evidence showed about *34defendant’s research regarding how smokers compensate for reductions in nicotine levels.

The jury could have found from the evidence that defendant knew that smokers who smoke cigarettes that are low in tar and nicotine compensate for those reductions by smoking more cigarettes, by taking longer, deeper, and more frequent inhalations of smoke, by holding the smoke in their lungs longer, and by subconsciously holding the cigarettes in a way that blocks microscopic ventilation holes. Those behaviors are prompted by the need of smokers to reach the level of nicotine in their bodies where they feel physically comfortable. Until that level is reached, smokers will continue to smoke in order to receive the amount of nicotine that their bodies require. Thus, smokers who switch to low-nicotine and low-tar cigarettes such as Merit may compensate for the low content, often unconsciously. Because the ratio of nicotine to tar in a cigarette remains relatively constant, a person who smokes to obtain the effect of increased nicotine — the addictive portion of the smoke — will ingest more tar. Thus, the effect of compensation is that a person who smokes a low-tar cigarette may not actually reduce his or her exposure to the harmful substances in tobacco smoke.

Defendant was aware of the concept of compensation by 1961, at least in broad outline. Its subsequent research confirmed the significance of compensation as it continued to create strategies to meet the onslaught of other research that suggested that cigarette smoking was hazardous to human health. In 1971, defendant conducted a study of smokers in which it asked participants to smoke cigarettes that delivered varying amounts of tar and nicotine. That study showed defendant that, as tar delivery decreased, cigarette consumption increased, with the result that the smoker’s daily intake of tar remained constant. A similar effect occurred with nicotine. The study concluded that its “findings support the hypothesis that the smoker does have daily intake quotas for tar and/or nicotine; and that he titrates his smoke intake to meet his quotas.” In addition, a 1975 memorandum from one of defendant’s leading researchers described research that showed that smokers smoked Marlboro Lights — a Philip Morris low-tar brand — differently from how they smoked the full-flavor Marlboros. Among other things, the study showed *35that each inhalation was greater in volume, with the result that Marlboro Light smokers received higher proportions of the available tar. Although many of defendant’s senior scientists knew about the results of research regarding compensation, defendant neither published that information nor shared it with regulatory agencies or the public health community.

Defendant specifically developed the Merit brand of cigarettes to respond to the health concerns that became especially prominent after the 1964 Surgeon General’s report on smoking and health. Defendant’s goal in developing Merits was to produce a cigarette that consumers would perceive as healthy. The intended marketing target was smokers who believed that smoking was not healthy, but who sought some reason to continue smoking. Advertisements for the Merit brand described it as a “light” or “low-tar” cigarette. Although defendant did not expressly assert in its advertising campaigns that Merit cigarettes were healthy or safe for consumption, the jury could find from the evidence presented by plaintiff that it otherwise promoted that message.

At the time that defendant began promoting the Merit brand, the public health community generally believed that low-tar cigarettes were safer than full-flavor brands. Indeed, the 1981 Surgeon General’s report suggested that smokers who could not quit smoking could diminish the risk to their health by switching to a low-tar brand. At the time of that report, the extent and importance of the concept of compensation was not generally known to the relevant scientific community, in part because defendant kept its research confidential and unpublished. For instance, the person who drafted the 1981 Surgeon General’s report testified at trial that, if he had known at the time what defendant knew, he would not have recommended in that report that smokers switch to a low-tar cigarette as an alternative to quitting.

As stated above, defendant argues that plaintiffs low-tar fraud claim is preempted by the Federal Cigarette Labeling and Advertising Act. It argues that plaintiffs claim is, at bottom, “a claim that defendant should have said more about low-tar cigarettes.” (Emphasis in original.) According *36to defendant, “the Labeling Act bars any claim that defendant [ ] should have provided additional information about the health risks of low-tar cigarettes.” Plaintiff responds that “[federal preemption does not relieve defendant of its duty not to deceive.” For the reasons discussed below, we agree with plaintiff.

Cipollone makes it clear that section 1334(b) does not preempt all common-law claims. Rather, the Court explained, it is necessary to “look to each of [the] common-law claims to determine whether it is in fact pre-empted.” 505 US at 523 (footnote omitted). The plaintiff in Cipollone alleged two theories of fraudulent misrepresentation. Id. at 527. His first theory of fraud was that the defendant used its advertising to neutralize the federally mandated warnings under the statute. That claim was predicated on a state-law prohibition against advertising statements and promotional information that minimized the health hazards associated with smoking. The Court held that that claim was preempted because the statute preempts both warning requirements and prohibitions. Id.

In contrast to the above theory alleged by the plaintiff in his first theory in Cipollone, his second theory was that the defendant had committed fraud by “false misrepresentations of a material fact [and by] conceal [ment] of a material fact.” The Court reasoned that that theory was not preempted by the statute because it was predicated not on a duty based on smoking and health, “but rather on a more general obligation — the duty not to deceive.” 505 US at 528-29. The Court relied on the fact that the legislation enacted by Congress expressly reserved the FTC’s authority to identify and punish deceptive advertising practices. According to the Court, “[t]his indicates that Congress intended the phrase ‘relating to smoking and health’ [in the statute] * * * to be construed narrowly so as not to proscribe the regulation of deceptive advertising.” Id. at 529 (footnote omitted). The Cipollone Court “conclude [d] that the phrase ‘based on smoking and health’ fairly but narrowly construed does not encompass the more general duty not to make fraudulent statements.” Id.

*37We conclude that plaintiffs “low-tar” fraud specification does not constitute an allegation that is preempted by section 1334(b). Plaintiff alleged that defendant deceived the smoking public by falsely representing that low-tar cigarettes are safer than regular cigarettes and thus constitute a reasonable alternative to quitting smoking altogether. Based on the evidence offered by plaintiff under its fraud specification, the jury could properly have found that, apart from any labeling of its product, defendant affirmatively represented to the smoking public and the public health community that it would reveal the results of its research as part of a common quest to determine the health hazards of smoking. Instead, it hid the results of its research and offered the Merit brand .to the public, knowing that the public health community generally believed that low-tar cigarettes were safer than full-flavor brands. In other words, defendant permitted the public health community to promote brands like Merit on its behalf, knowing that, in fact, they were not safer than full-flavor brands. Plaintiffs fraud specification, when understood in this light, does not allege that defendant attempted to neutralize the federally required warnings or that it failed to give additional warnings concerning particular kinds of cigarettes, claims that would have been preempted under the federal act. Rather, it alleges the violation of the general duty not to deceive and, as the Court in Cipollone held, such a claim is not preempted.

We turn to defendant’s other arguments relating to plaintiffs fraud claim. In its second assignment of error, defendant challenges the court’s jury instruction that, if the jury found that defendant voluntarily assumed a duty to decedent as a member of the consuming public, it must disclose all material matters of which it had knowledge. In its third assignment of error, defendant claims that the court erred in failing to instruct the jury that the Frank Statement was not a representation on which decedent relied and that, therefore, the statement could not be the basis for plaintiffs fraud claim. Defendant supports those assignments in part with arguments that go beyond what it raised to the trial court; we therefore do not consider those portions of its arguments because the trial court did not have the opportunity to consider them. See ORAP 5.45. Nonetheless, some of the *38unpreserved arguments provide background to our discussion of the arguments that defendant did preserve and that it now raises on appeal. Accordingly, we refer to them below only to provide context.

Defendant first argues — with regard to plaintiffs claim that it failed to disclose its research to the public — that there is no legally cognizable claim in Oregon for a fraud perpetrated on the public based on an assumed duty. It points out that all the Oregon cases that discuss liability for an assumed duty involve claims in which the plaintiff attempted to hold the defendant liable in negligence for responsibilities that the defendant assumed on behalf of a particular person or class of persons. See, e.g., Quackenbush v. PGE, 134 Or App 111, 118, 894 P2d 535, rev den, 322 Or 193 (1995) (assumed duty to make tree safe for decedent to prune it); Peterson v. Mult. Co. Sch. Dist. No. 1, 64 Or App 81, 93-94, 668 P2d 385, rev den, 295 Or 773 (1983) (assumed duty to make safety recommendations to high school football coaches). Defendant’s assertion is correct as a matter of a survey of existing case law, but its assertion does not resolve the issue whether liability arises from a fraudulent representation by a defendant that has assumed a duty to the consumer public.

In Williams v. Philip Morris Inc., 182 Or App 44, 48 P3d 824 (Williams 7), adh’d to on recons, 183 Or App 192, 51 P3d 670 (2002) (Williams IP, rev den, 335 Or 142, vac’d and rem’d, 540 US 801, 124 S Ct 56, 157 L Ed 2d 12 (2003), on remand, 193 Or App 527, 92 P3d 126 (2004) (Williams III), aff'd, 340 Or 35, 127 P3d 1165 (2006), we considered a similar issue.8 In that case, the defendant argued that Oregon did not recognize a common-law claim of fraud arising out of representations made to the consumer public rather than to a specific individual. We disagreed, observing that the “plaintiffs theory fits comfortably within traditional common-law principles.” Williams I, 182 Or App at 53. We apply a similar analysis here.

The elements of common-law fraud are:

*39“(1) a representation; (2) its falsity; (3) its materiality; (4) the speaker’s knowledge of its falsity or ignorance of its truth; (5) his intent that it should be acted on by the person and in the manner reasonably contemplated; (6) the hearer’s ignorance of its falsity; (7) his reliance on its truth; (8) his right to rely thereon; (9) and his consequent and proximate injury.”

Conzelmann v. N. W. P. & D. Prod. Co., 190 Or 332, 350, 225 P2d 757 (1950); see also Meader v. Francis Ford, Inc., 286 Or 451, 456, 595 P2d 480 (1979) (holding that a claim in fraud need not be founded on a contractual obligation but will lie based on the knowledge of a falsity of a representation and the intention to mislead). As we said in Williams I, “the fundamental character of fraud is the communication of a misimpression to induce another to rely on it.” 182 Or App at 54. One species of a misimpression that will give rise to an actionable claim in fraud is a promise made with the knowledge that it will not be performed or with reckless disregard about whether it will be performed. Elizaga v. Kaiser Found. Hospitals, 259 Or 542, 548, 487 P2d 870 (1971). It follows that, if defendant made a promise to the consumer public without intending to perform it, a traditional element of common-law promissory fraud would be satisfied.

Defendant also argues that there is no evidence that it made a promise to disclose research that it did not intend to keep at the time it was made. However, plaintiff points to the Frank Statement as evidence that defendant made promises that it did not intend to keep. In the 1954 statement, defendant promised to give “aid and assistance to the research effort into all phases of tobacco and health.” In the 1962 statement, the Tobacco Institute, on behalf of defendant, stated that it “supports and cooperates with all responsible efforts to find the facts and bring them to the public” and recognized that the industry has “a special responsibility to help scientists determine the facts about tobacco use and health,” a responsibility that it accepted in 1954 by creating the TIRC to fund scientific research. In 1966, defendant, again through the Tobacco Institute, emphasized its support for research to discover the full facts about tobacco and health. Defendant and the industry made similar statements in later years. We hold, based on all of the above evidence, that the jury could *40reasonably infer that defendant engaged in a continuing course of conduct in which it represented to the public that it would conduct research and disclose its results to the public over a period of time.9 The jury could also infer from the above evidence that, at the time that defendant made those representations, it did not intend to perform them.

Defendant argues, however, that there is no evidence that decedent relied on any promise made by defendant to reveal the results of its research. It points out that there is no evidence that decedent ever saw either the Frank Statement or the 1962 statement. Indeed, decedent was eight years old when defendant published the Frank Statement, and plaintiff testified that he had no reason to believe that she had heard, seen, or cared about industry advertising in 1954. Plaintiff agrees with defendant’s assertion about the state of the evidence, but responds that it is not necessary to show that decedent was aware of defendant’s assumed duty; he points out that the plaintiff in Peterson did not know that the defendant had undertaken the duty that it negligently failed to perform in that case.

In Peterson, the plaintiff became a quadriplegic after suffering a neck injury at a high school football practice. 64 Or App at 83. He asserted that, because the Oregon School Activities Association (OSAA) had voluntarily undertaken to make and disseminate safety recommendations to the high school that he attended and had failed to perform that duty, it was liable for negligence. Id. Plaintiffs argument in this case, however, fails to distinguish between the different requirements for proving a claim of negligence and for proving a claim of fraud. The fraud theories are intended to protect different interests from the negligence theories. The plaintiffs lack of knowledge about an assumption of a duty by the OSAA in Peterson did not matter to his claim for negligence because the assumed duty of care by the OSAA extended to the plaintiffs interest in participating safely in a particular activity.

We conclude that Peterson is inapposite. Unlike in a negligence claim where a defendant may assume a duty to *41more than one person or to a class of persons, the intrinsic nature of a claim of promissory fraud is such that it protects the interests only of those to whom a promise is made specifically and directly. That is because the requisite intent to mislead in a fraud claim consists of a defendant misrepresenting a material fact for the purpose of misleading the other party. U.S. National Bank v. Fought, 291 Or 201, 225, 630 P2d 337 (1981). That kind of requirement does not exist in a negligence claim where a duty of reasonable care is assumed generally on behalf of a class of persons participating in a particular activity. We conclude therefore that, in the absence of evidence that decedent relied on defendant’s representations that it would reveal its research to the public, the trial court erred when it submitted that theory to the jury. Yet, as explained below, because we find no error regarding the “low-tar” specification of fraud, that error does not require reversal. See Moe v. Eugene Zurbrugg Construction Co., 202 Or App 577, 588, 123 P3d 338 (2005) (if jury returns special verdict and court upholds at least one specification, any error regarding other specifications does not require reversal).

We return to the issue whether the trial court properly submitted the “low-tar” specification of fraud to the jury, having already determined that it was not preempted by federal law. Defendant argues that it never represented publicly that low-tar cigarettes were safer; rather, throughout the time that decedent smoked Merits, it was the public health community, not defendant, that gave smokers the information that low-tar cigarettes were safer and that suggested that they switch from full-flavor brands. It follows, according to defendant, that it cannot be held liable for fraud when smokers relied on recommendations from the public health community rather than on what it represented in its promotion of low-tar cigarettes. Plaintiff counters that there is evidence that defendant’s development and promotion of low-tar cigarettes, in particular, the Merit brand, was part of its continuing fraudulent effort to dissuade the public health community about the hazards of smoking, an effort that began in 1954 and continued thereafter.

The flaw in defendant’s argument is its failure to recognize the extent to which the public health community relied on what the jury could have found to be defendant’s *42deceitful failure to fulfill the promises that it made to the public at large, including the public health community, in the Frank Statements and elsewhere. The jury could have inferred from the evidence that defendant knew from its undisclosed research that the FTC machine results on which the public health community relied were misleading but that defendant and other members of the industry worked to retain that standard of testing because it “gave low numbers.” Defendant was also aware of the significance of the concept of compensation and that people who smoke low-tar cigarettes typically do not reduce their intake of tar and nicotine. Although defendant publicly represented that it was working with scientists to discover the truth about tobacco addiction and health, it did not share the essential information that it had learned from its research, either with the public or with non-Philip Morris scientists, as it had promised. In particular, the evidence shows that decedent switched to the Merit brand of cigarettes in 1976 because she believed that “the low tar and nicotine filters are better for you.” The jury could infer from the evidence that decedent’s belief arose because of the mistaken belief of the public health community that low-tar cigarettes were safer than full-flavor brands. The jury could also find that smokers like decedent were ultimately the intended recipients of defendant’s deceit perpetrated on the public health community.

Defendant cannot shield itself from liability merely because it used a third party to deliver its misleading message. In American National Bank of Denver v. Tonkin, 286 Or 73, 77, 592 P2d 1008 (1979), the plaintiff agreed to lend $300,000 to Baron if the defendant would guarantee the loan. The plaintiff told Baron that the loan would be for two and one-half years at nine percent interest. Baron communicated that statement to the defendant, who agreed to guarantee the loan on those terms. In fact, without telling the defendant, the plaintiff made the loan for 365 days with renewal up to two and one-half years and at interest rates higher than nine percent. When the plaintiff sued to collect on the guaranty, the defendant raised the affirmative defense of fraud. The Supreme Court held that there was sufficient evidence that the bank made the statements to Baron with the intent that he would convey them to the defendant to support the *43affirmative defense. Relying on earlier cases, it held that one who makes representations to another, intending for the recipient to communicate them to a third person so that the third person may act on them, may be liable to the third person for fraud. Id. at 81-83, 85-86.

In Williams I, we held, relying on Oregon Supreme Court decisions, that “a person who makes a misrepresentation may be liable to the intended recipients of a misrepresentation without regard to whether the person making the representation intends to defraud a particular person.” 182 Or App at 53. What matters is whether the representation was intended ultimately for the public or for a particular class of persons to which the plaintiff belongs. Id. at 53-55. We adhere to those principles in this case. Here, there is evidence from which the jury could find that defendant concealed the results of its research from the public health community with the intent that low-tar cigarettes would receive a favorable recommendation from that community and that members of the consumer public would rely on those recommendations. That evidence is sufficient to support plaintiffs claim that defendant misrepresented to smokers like decedent that low-tar cigarettes were safer.

Our conclusion that the trial court did not err in denying defendant’s motion for a directed verdict on plaintifPs claim that defendant committed fraud by misrepresenting to the consumer public that low-tar cigarettes were safer requires us to consider defendant’s thirteenth assignment of error, which affects all of plaintiffs claims — the trial court’s failure to grant a mistrial during the jury’s deliberations. During the trial, the court admitted hundreds of documentary exhibits offered by both parties. Each party occasionally highlighted portions of the exhibits to show the jury during a witness’s testimony or as part of opening statements or closing arguments. Before the jury began deliberating, defendant’s counsel gave the clerk copies of defendant’s exhibits that did not include any highlighting so that the clerk could give them to the jury; they believed that plaintiffs counsel had done the same with plaintiffs exhibits. On the second day of deliberations, however, defendant discovered that approximately 100 of the exhibits that plaintiffs counsel gave the clerk for delivery to the jury contained highlights *44that plaintiffs attorneys had made on them. One exhibit also contained an attached note in which plaintiff’s attorney included selective quotations from a different exhibit. When defendant objected to the actions of plaintiffs counsel, they explained that they were “just trying to direct [the jury’s] attention to the appropriate portions of the documents.” Defendant moved for a mistrial, arguing that plaintiffs counsel’s actions were misconduct that prejudiced its ability to receive a fair trial.

Rather than granting defendant’s motion for a mistrial, the trial court ultimately decided to give a curative instruction that addressed both the highlighting and the fact that the jury had accidentally received several exhibits that had not been admitted into evidence. After telling the jury to disregard those exhibits, the trial court told the jury:

“And also the Court — it has been brought to the Court’s attention that in some of the exhibits from the plaintiff they’re highlighted with yellow markings. You are not to give any special weight to any highlighting in a document.

‘You have the right to review that document and it’s your responsibility to determine what weight you will give to various portions of that document. And I will give you an example. I could read a novel, and I could go through that novel and I could highlight portions that I thought were very important to me.

“I could give that same novel to you to read, and you read it, ‘Why was this guy highlighting this right here? It doesn’t mean anything to me.’ You might highlight other areas of the novel. You give it to a third person, a third person might highlight other areas.

“We may have to go through a large number of people before we start finding people that agree with what we, as individuals, highlighted. Just because counsel for the plaintiff highlighted an area, that [does not] mean that you are bound to accept that as being an important area for this trial.

‘You might read another paragraph or another page, and that might be more important to you. As a juror that’s your responsibility to do that.”

*45Defendant argues that the court’s instruction was insufficient to cure the effect of the highlighting, arguing that there is a presumption of prejudice when a jury receives unauthorized communications. The cases on which defendant relies all involve the jury receiving information that was not admitted in evidence; none involves emphasizing certain aspects of the admitted evidence. In addition, they all involve review of the jury’s verdict, not actions that the trial court took before the verdict to cure any prejudice. See, e.g., State v. Mapel, 54 Or App 795, 797, 636 P2d 445 (1981) (jurors in drug case attended seminar during trial at which prosecutor and state witness spoke); State v. Holmes, 17 Or App 464, 476, 522 P2d 900 (1974) (court provided dictionary for jury to use during deliberations); Stephens v. South Atlantic Canners, Inc., 848 F2d 484, 486-89 (4th Cir), cert den, 488 US 996 (1988) (prejudicial comments written on exhibits). Assuming, however, that a presumption of prejudice arose under the circumstances of this case, it does not necessarily follow that the trial court’s only choice was to grant a mistrial. The question is whether the instruction that the court gave was sufficient to overcome any prejudice that plaintiffs actions may have created.

We review a trial court’s denial of a motion for mistrial for an abuse of discretion because the trial judge is in the best position to evaluate the prejudicial effect, if any, of the challenged actions and whether there are alternatives to a mistrial. State v. Lotches, 331 Or 455, 496-97, 17 P3d 1045 (2000), cert den, 534 US 833 (2001). In this case, plaintiff’s counsel’s action in highlighting exhibits that counsel intended the clerk to submit to the jury may have been improper. The effect of that action was to direct the jury’s attention to the portions of the documents that plaintiff thought persuasively supported his contentions. Advocacy, however, ends when the jury retires to consider its verdict. At that point it is the jury’s role, not that of the parties, to determine which portions of the evidence are important and which are not. The jury should have been given unaltered exhibits.

The question, however, is whether it was within the trial court’s discretion to respond to those circumstances in some way other than ordering a mistrial. In exercising its discretion, the court could have considered that the highlighting *46was a small aspect of a long and hard-fought trial during which both parties had emphasized to the jury the portions of the evidence that they considered important. The trial court also stated more than once during the course of the trial that the jury was particularly attentive and involved in the case. The curative instruction that the court gave was straightforward, and it illustrated the court’s point with an appropriate example. We perceive no prejudice from the highlighting that could not be cured by the instruction. See State v. Terry, 333 Or 163, 177, 37 P3d 157 (2001) (“ ‘Jurors are assumed to have followed their instructions, absent an overwhelming probability that they would be unable to do so.’ ” (Quoting State v. Smith, 310 Or 1, 26, 791 P2d 836 (1990).)). Under the circumstances, it was within the court’s discretion to conclude that the jury would follow the direction of the curative instruction to disregard the highlighting. We therefore conclude that the trial court did not act outside of its range of discretion in giving a curative instruction, rather than granting defendant’s motion for a mistrial.

We turn now to the award of punitive damages on the fraud claim. Defendant assigns error to the trial court’s failure to give its requested jury instruction that “[y]ou are not to punish a defendant for the impact of its conduct on individuals in other states.” The issue framed by defendant’s assignment implicates constitutional principles under federal due process law and trial practices governed by the law of the State of Oregon. As background, we begin our discussion with two United States Supreme Court cases that established the applicable constitutional principle: State Farm Mut. Automobile Ins. Co. v. Campbell, 538 US 408, 123 S Ct 1513, 155 L Ed 2d 585 (2003), and BMW of North America, Inc. v. Gore, 517 US 559, 116 S Ct 1589, 134 L Ed 2d 809 (1996).

In Gore, an automobile purchaser brought an action against an automobile manufacturer for the failure to disclose that a purportedly new automobile had been repainted after being damaged before delivery occurred. 517 US at 563. The manufacturer acknowledged that it had a nationwide policy of not advising customers of predelivery damage to new cars when the cost of repair did not exceed three percent of the car’s suggested retail price. Id. The jury returned a *47$4 million punitive damage award against the manufacturer, which was reduced to $2 million by the Alabama Supreme Court. The United States Supreme Court held, however, that even the reduced award violated the Due Process Clause because it was grossly excessive in relation to the state’s legitimate interests in punishing and deterring unlawful wrongful conduct. Id. at 573-74. The Court observed that a state may protect its citizens by prohibiting deceptive trade practices, but that neither the jury nor the trial court was presented with any evidence that the manufacturer’s out-of-state conduct was unlawful. Indeed, there was testimony that approximately 60 percent of the vehicles that were repainted were sold in states where such repairs did not constitute unfair trade practices. Id. at 573. Although the plaintiff argued that the manufacturer’s conduct was particularly reprehensible because the nondisclosure of repairs regarding his car was part of a nationwide pattern of tortious conduct, the Court held that the Alabama Supreme Court “properly eschewed reliance on BMW’s out-of-state conduct,” id., and that “a State may not impose economic sanctions on violators of its laws with the intent of changing the tortfeasors’ lawful conduct in other States.” Id. at 572.

The above rule was reiterated in Campbell, a case in which the insureds brought an action in Utah against an automobile liability insurer to recover for a bad-faith failure to settle within policy limits. Evidence of the insurer’s practices both in and out of state was offered to show that the insurer’s actions were tortious — that is, to show that its practice was deliberate and to rebut the insurer’s assertion that its practices were inadvertent or mistakes in judgment — and to support an award of damages to punish the insurer for its nationwide practices. The jury awarded the plaintiffs $2.6 million in compensatory damages and $145 million in punitive damages, which the trial court reduced to $1 million and $25 million, respectively. 538 US at 415. The Supreme Court, in reversing the award of punitive damages, stated that, as a general rule, a state does not

“have a legitimate concern in imposing punitive damages to punish a defendant for unlawful acts committed outside of the State’s jurisdiction. Any proper adjudication of conduct that occurred outside Utah to other persons would require *48their inclusion, and, to those parties, the Utah courts, in the usual case, would need to apply the laws of their relevant jurisdiction.”

538 US at 421-22. The Court also admonished:

“Lawful out-of-state conduct may be probative when it demonstrates the deliberateness and culpability of the defendant’s action in the State where it is tortious, but the conduct must have a nexus to the specific harm suffered by the plaintiff. A jury must be instructed, furthermore, that it may not use evidence of out-of-state conduct to punish a defendant for action that was lawful in the jurisdiction where it occurred. Gore, 517 U.S., at 572-573 (noting that a State ‘does not have the power ... to punish [a defendant] for conduct that was lawful where it occurred and that had no impact on [the State] or its residents’). A basic principle of federalism is that each State may make its own reasoned judgment about what conduct is permitted or proscribed within its borders, and each State alone can determine what measure of punishment, if any, to impose on a defendant who acts within it jurisdiction.”

538 US at 422 (emphasis added; bracketed material by Campbell Court).

The issue underlying defendant’s assignment of error is whether the trial court erred — in light of the principle announced by the above cases that a state cannot impose punitive damages for harm caused to nonresidents by out-of-state conduct — when it refused to give defendant’s requested instruction that “it was not to punish defendant for the impact of its conduct on individuals in other states.” That issue also implicates a common-law rule regarding Oregon trial practice. In general, a party in a civil action is entitled to jury instructions on its theory of the case only if its requested instruction accurately states the law that governs the trial of the case. Hernandez v. Barbo Machinery Co., 327 Or 99, 106, 957 P2d 147 (1998). At least three corollary principles necessarily flow from that rule. First, the rule contemplates that instructions be based on the evidence adduced at trial and not on abstract principles that do not engage with the evidence.10 Second, evidence admitted at trial may be probative *49on more than one issue, and the jury’s use of it, once admitted and absent some limitation by the court, is unrestricted.11 For instance, in this case, plaintiff offered evidence of defendant’s out-of-state conduct to demonstrate the ongoing pattern of deceit over many years regarding the harmful effects of its tobacco products. That evidence was probative to demonstrate the reprehensibility and the deliberativeness of defendant’s conduct. On the other hand, the jury was not entitled to use that evidence to punish defendant for the impact of that out-of-state conduct on non-Oregon residents under the Due Process Clause. Third, no party under Oregon law is required to request a jury instruction that advances the use of evidence in a way that benefits the party’s adversary. Rather, Oregon law allocates the responsibility of each party to request a jury instruction on its theory of the case, not on the other party’s theory of the case.12 Each of those principles is embodied in the general rule, and each is implicated here, as more fully explained below.

To recapitulate: Under the Due Process Clause, a state can punish a defendant for in-state conduct that causes in-state harm. A state also has an interest in punishing a defendant for its out-of-state conduct that causes in-state harm. Both of those interests were at issue in this case because of the evidence that plaintiff adduced. As the Court held in Gore and Campbell, however, a state may not punish a defendant for out-of-state conduct that causes out-of-state harm. The final construct, punishing a defendant for in-state conduct that causes out-of-state harm, also necessarily depends on the state of the evidentiary record. For instance, Oregon could make a legitimate policy choice — that is, a choice that comports with the Due Process Clause — to punish *50an actor who is conducting a fraudulent advertising campaign from Oregon that causes harm in a neighboring state. But the applicability of that construct, like the other constructs, depends on the evidence adduced at trial and the legal theories of the parties regarding that evidence. We turn, then, to the evidence in this case.

In this case, plaintiff offered evidence of harm that was caused to nonresidents by the consumption of defendant’s product. Indeed, one witness testified that there are approximately 500,000 deaths nationwide each year that could be attributed to cigarette smoking. Plaintiffs theory regarding the use of that kind of evidence was that defendant’s “unlawful misconduct in 50 states is subject to punitive damages in this courtroom.” In turn, defendant sought to limit the jury’s use of the evidence of defendant’s out-of-state conduct. It told the trial court, “Philip Morris cannot be punished in this state for conduct which occurred out of state. [The requested instruction] informs the jury we cannot be punished for out of state harms.” Plaintiff countered that, unlike in Gore, “defendant’s conduct in this case before us * * * would be tortious in any state.” The court ruled that the requested instruction “is out.”

Defendant correctly understood the import of the Court’s holding in Gore, and its requested instruction accurately addressed the issue that was before the trial court. Plaintiffs counter argument was legally incorrect.13 Significantly, the language used in defendant’s requested instruction follows almost word-for-word the relevant language in Campbell. In effect, the parties and the court understood that the requested instruction would have placed a limitation on the jury’s use of the evidence of defendant’s nationwide policies and conduct to punish defendant for purposes of punitive damages. As plaintiff concedes in his brief on appeal, “[t]he constitution prohibits plaintiff from urging the jury to multiply punitive damages by the number of others similarly harmed.” Yet, that is just what plaintiff did in his arguments to the jury. Had the court given defendant’s requested instruction, it would have been given in the context of other instructions regarding punitive damages, and it would have *51prevented the jury from considering the evidence of out-of-state harm for purposes of punishing defendant.

The consequence of the trial court’s failure to give defendant’s requested instruction in this case is a violation of defendant’s due process rights. As the Court aptly explained in Campbell:

“For a more fundamental reason, however, the Utah courts erred in relying upon this and other evidence [of out-of-state harm]: The courts awarded punitive damages to punish and deter conduct that bore no relation to Campbell’s harm. A defendant’s dissimilar acts, independent from the acts upon which liability was premised, may not serve as the basis for punitive damages. A defendant should be punished for conduct that harmed the plaintiff, not for being an unsavory individual or business. Due process does not permit courts, in the calculation of punitive damages, to adjudicate the merits of other parties’ hypothetical claims against a defendant under the guise of the reprehensibility analysis, but we have no doubt the Utah Supreme Court did that here.”

538 US at 422-23. Consequently, the trial court erred in refusing to give defendant’s requested instruction.

The dissenting members of this court disagree. In their view, the requested instruction did not accurately state the principle of constitutional law that plaintiff acknowledges in his brief. However, the arguments made by them in support of their dissents are not arguments that were made by plaintiff at trial or on appeal. We have previously summarized plaintiffs position at trial and explained why it was incorrect. On appeal, plaintiff first argues that “|h]arm to others in or out of state is certainly 'relevant to the determination of reprehensibility of the defendant’s conduct.’ ” (Quoting Gore, 517 US at 574 n 21.) As explained above, we agree. Plaintiffs next assertion is that it did not argue to the jury that defendant ought to be punished for harm to nonresidents caused by its out-of-state conduct. As we indicated above, the transcript of the closing arguments to the jury undercuts that assertion. Finally, plaintiff summarily argues, “Defendant’s requested instructions were improper and correctly rejected.” In light of plaintiffs arguments to the *52trial court and on appeal, we question whether it is appropriate for this court to consider in this case the abstract notions of law advanced by the dissents.

Nevertheless, we respond to the dissents’ arguments as follows. Judge Armstrong appears to make three arguments, two of which are interrelated. He argues initially that, because of the mobility of the nation’s population, principles of due process do not prevent a state from punishing a defendant for harm caused to people in other states. Judge Armstrong also argues that “defendant’s proposed instruction was flawed because it did not distinguish between defendant’s conduct in Oregon and its conduct in other states.” 206 Or App at 74-75 (Armstrong, J., concurring in part, dissenting in part). Finally, he argues that defendant’s proposed instruction cut too broadly, because a jury can in fact award punitive damages to “reflect the fact that the defendant’s unlawful conduct in other states had harmful effects in those states.” Id. at 73 (Armstrong, J., concurring in part, dissenting in part).

We begin with Judge Armstrong’s two interrelated arguments concerning the mobility of the nation’s population and his assertion that the Due Process Clause does not prevent a state from punishing a defendant for harm caused to people in other states. In general, the concern in Campbell and Gore is that when a state punishes a defendant for harm caused in another state — in contrast to using evidence of actions in other states to show the reprehensibility of the defendant’s actions that affected people within the state — it acts outside its territorial jurisdiction. A state has an interest in protecting its own consumers and its own economy; however, a state does not have a valid interest under the Due Process Clause in punishing a defendant for harm that is caused and occurs outside its jurisdiction. That is a limitation under Campbell and Gore that a jury must be informed about when it determines the amount of an award of punitive damages. Simply put, Judge Armstrong’s construct that the mobility of the nation’s population permits a state to punish a defendant for harm caused to people in other states is not supported by the reasoning in Campbell and Gore because it would offend the principles of federalism on which the Due Process Clause is based.

*53There is, however, a more fundamental problem with Judge Armstrong’s analytical construct: it is completely divorced from the evidence actually introduced in this case. At trial, plaintiff offered considerable evidence of defendant’s unlawful conduct, but he offered no evidence of the number of Oregon residents who had switched to Merits in another state and then moved to Oregon. Moreover, no party at trial made a distinction between defendant’s in-state conduct and its nationwide conduct generally. Rather, after the trial court refused to give defendant’s requested instruction, plaintiff argued extensively to the jury about the import of defendant’s nationwide conduct and why defendant should be punished for that conduct.

The parties’ discussion of defendant’s proffered instruction at trial also makes clear that no one at the trial level contemplated the distinctions that Judge Armstrong makes under the Due Process Clause, which, given the state of the evidence, is not surprising. In discussing defendant’s proffered instruction, neither counsel suggested that mobility of the population was relevant or that the instruction should distinguish — as Judge Armstrong asserts — between in-state and out-of-state conduct. Apparently, in Judge Armstrong’s view, to constitute a legally correct instruction, defendant’s requested jury instruction should have said, “You may not punish defendant for the impact of its out-of-state conduct on individuals in other states unless you find that those individuals have moved to the state of Oregon.” But defendant cannot be held responsible for the failure to make the distinction that Judge Armstrong asserts should have been made unless there is evidence to support such a distinction; here, there was no evidence of defendant’s instate conduct that caused harm to nonresidents, including nonresidents who later moved to Oregon.14

Nonetheless, Judge Armstrong asserts that decedent’s “experience is not unique, and the jury reasonably could infer that many people in Oregon have contracted *54tobacco-related diseases and incurred health care costs as a result of defendant’s conduct toward those people when they lived in other states.” 206 Or App at 73 (Armstrong, J., concurring in part, dissenting in part). We think that Judge Armstrong’s example makes our point for us. Decedent began smoking full-flavor cigarettes when she was a nursing student in Missouri. She opted, rather than quitting smoking, to switch to defendant’s Merit brand after she moved to Oregon. The gravamen of plaintiffs punitive damages claim against defendant focused on defendant’s nationwide conduct while she was a resident of Oregon and while she was considering quitting smoking. But plaintiff sought in the trial of this case to punish defendant for all of the harm suffered out-of-state by nonresidents as a result of the tobacco industry’s promotion of its products. That position runs afoul of the due process concept that a defendant can be punished in a particular case only for the kind of harm suffered by the plaintiff in that case. There is simply no evidence in the record that would permit a reasonable factfinder to draw the inference that Judge Armstrong says would have been available to the jury.

We conclude our response to Judge Armstrong’s dissent with several observations. Under Judge Armstrong’s view, Oregon could punish a defendant for harm caused by out-of-state conduct merely because the person harmed became a resident of Oregon after the out-of-state harm occurred. That construct punishes a defendant for harm that has already occurred; it does not in any way further the purpose of an award of punitive damages to deter a defendant’s future conduct. Moreover, if Judge Armstrong’s view were correct, neither Gore nor Campbell, based on their facts, would have come out the way they did. Essentially, the ramifications of Judge Armstrong’s proposed construct would eviscerate the holdings in Gore and Campbell; there would be no limitation on punishing a defendant for out-of-state conduct.

We turn now to Judge Rosenblum’s dissent. She believes that defendant’s instruction “would have misled the jury into thinking that it could not consider defendant’s out-of-state conduct for any purpose.” 206 Or App at 86 (Rosenblum, J., concurring in part, dissenting in part) (emphasis in original). In particular, she faults the requested *55instruction for not telling the jury that they could consider evidence of out-of-state conduct for purposes of assessing the reprehensibility of defendant’s conduct and for using the word “impact” instead of the word “harms” to describe the effect of defendant’s out-of-state conduct on “people in other states.” Id. at 87 (Rosenblum, J., concurring in part, dissenting in part).

Judge Rosenblum’s concerns about the jury being misled by the language of defendant’s requested instruction do not engage with the record of this case. Her concerns focus on the word “punish” and on the phrase “the impact of its conduct on individuals in other states.” The evidentiary record is unequivocal as to how the jury would have understood the import of those words. As stated above, under plaintiffs theory, the specific harm suffered by decedent was that she was deceived by defendant into believing that, by switching to the Merit brand of cigarettes, she would be reducing the harmful effects of consuming tobacco products. Although plaintiff offered considerable evidence of defendant’s deceitful practices, it offered no specific evidence of how many smokers were deceived in the way that plaintiffs decedent was deceived or how many smokers were harmed in the way that decedent was harmed (by not quitting smoking, but rather switching from full-favor brands). Campbell’s admonition is clear: a defendant should be punished by an award of punitive damages for the conduct that harmed the plaintiff, and not for other tortious conduct that bore no relation to the harm caused to the plaintiff. It is in that context that we inquire whether the jurors would have been misled by defendant’s requested instruction.

In particular, Judge Rosenblum takes issue with the word “punish” as used in defendant’s requested instruction.15 She speculates that the instruction, if given, would have precluded the jury from considering defendant’s out-of-state conduct for the purpose of demonstrating its reprehensibility. However, that argument takes the requested instruction out of the context in which defendant proposed that it be given. The instruction would have been given in the context of all *56the instructions on punitive damages, which, in language and in effect, told the jury that their consideration of an award of punitive damages was for the purpose of punishing defendant. For example, one of the instructions that the court did give told the jury that the factors it could consider in determining whether to punish defendant included “the total deterrent effect of other punishment imposed on the defendant as a result of the misconduct.” Moreover, nothing prevented plaintiff from requesting an instruction based on his theory of the probative value of the evidence. Indeed, plaintiff was not concerned about the meaning of the word “punish” in defendant’s requested instruction or that the jury could have been misled by the word if the trial court had given the instruction: Plaintiff told the trial court that the law permitted the jury to punish defendant for its conduct in all 50 states.

Judge Rosenblum’s concern about the use of the word “impact” instead of the word “harm” in the requested instruction is similarly divorced from the record made in the trial court. The only evidence of out-of-state impact that bore on the issue of punitive damages in this case was evidence of harm to nonresidents. The parties could have had no doubt in their minds that the word “impact” referred to the out-of-state harm caused by defendant’s nationwide conduct, as defendant’s counsel explained to the trial court in the presence of plaintiff’s counsel. And that understanding affected the presentation of the case to the jury. With due respect, one can always resort on appeal to a dictionary to find alternative meanings for words used in a jury instruction or disagree about whether one word should be used in place of another, but that practice does little to inform the issue of whether the instruction, as worded, would have misled the jury unless those definitions engage with what actually happened at trial. Here, the parties and the trial court knew precisely what defendant’s requested instruction referred to as revealed by the record in this case, and there is no legally cognizable basis for asserting that the jury would have understood otherwise.

In summary, plaintiff offered evidence at trial of the impact of defendant’s nationwide conduct. He effectively took the position that the jury could consider all evidence of the harm caused by defendant to nonresidents for the purpose of *57punishing defendant, regardless of where that harm originated, telling the court that “defendant’s conduct * * * would be tortious in any state.” Plaintiff either disagreed with or missed the point of defendant’s requested instruction that a state has no interest under the Due Process Clause in punishing a defendant for conduct that occurs and causes harm outside its territorial jurisdiction. Throughout the colloquy between the parties and the trial court with regard to defendant’s requested instruction, only “out-of-state” conduct was referenced, and significantly, plaintiff never objected to the requested instruction on any of the grounds advanced by the dissents. It follows that the dissents’ arguments that defendant’s proposed limiting instruction was too broad in scope are unsupported by the record made in the trial court.

Finally, it is important to keep in mind that trials in Oregon are adversarial proceedings. Defendant’s requested instruction was a correct statement of law as to the subject that it actually addressed, but the dissents’ position would effectively require defendant to include in its requested jury instruction provisions that would have benefitted only plaintiff. No party is required under Oregon law to make an argument for another party or to propose a use of evidence that benefits an adversary. Rather, the responsibility for requesting an instruction that is based on a party’s theory of the case is on that party. Hernandez, 327 Or at 106. That general principle applies here. Plaintiff was the only party who could benefit from the jury’s consideration of the evidence of defendant’s out-of-state conduct, and as previously noted, he had the burden of proof on the issues of liability and punitive damages. The above principle properly allocates the burden of requesting appropriate jury instructions to the party who advances a particular legal theory or use of the evidence. In sum, had plaintiff wanted to address the concerns of the dissents, he could have proposed his own instructions. The dissents’ assertions to the contrary in this case would constitute a radical shift in the practice of law in Oregon’s trial courts if adopted by this court.

We conclude for all of the above reasons that the trial court erred in failing to give defendant’s requested instruction, an instruction that due process of law required. Consequently, we must vacate the award of punitive damages on the fraud claim as well on the other claims. Our disposition— *58remand for a new trial on the amount of punitive damages— renders moot plaintiffs cross-appeal challenging the reduction in the amount of punitive damages.

II. THE NEGLIGENCE CLAIM

In his third amended complaint, plaintiff alleged, under the heading “Second Claim for Relief — Negligence,” seven specifications of negligence. In its sixth assignment of error, defendant claims that the trial court erred in denying its motion for a directed verdict on what it describes as plaintiffs “assumed duty5’ negligence specification on the ground that

“there is no evidence from which a reasonable juror could conclude that [decedent] ever saw, heard, or relied on any representations made by Philip Morris regarding smoking and health, much less the 1954 Frank Statement, upon which plaintiff relies to suggest that Philip Morris assumed a duty to [decedent].”

We are uncertain to which specification of negligence defendant’s motion was directed. After plaintiff alleged six specifications of negligence in his complaint, he alleged in paragraph 11 (still under the negligence heading) that defendant “voluntarily assumed a duty to disclose research, cooperate closely with public health authorities and hold people’s health paramount to all other business considerations.” He did not explain how that allegation related to any of the specifications of negligence. To the extent that defendant is arguing that plaintiff was required to demonstrate that decedent had to have knowingly relied on defendant’s failure to disclose its research, its argument is answered by our earlier distinction between negligence and promissory fraud claims; that is, plaintiff is a member of a group of people for whom defendant assumed a duty of reasonable care. In other words, for the purpose of plaintiffs negligence claim, he was not required to demonstrate additionally that decedent expressly relied on defendant’s promise to disclose its research in order to hold defendant liable in negligence, as defendant apparently contends.

In its seventh assignment of error, defendant posits a separate ground for its contention that the trial court erred in denying its motion for a directed verdict on plaintiffs negligence claim. Defendant argued to the trial court that it was *59entitled to a directed verdict because “there is no evidence from which a reasonable juror could conclude that Philip Morris was negligent or that such negligence caused plaintiffs harm and because plaintiffs negligence claims are preempted by federal law.” Defendant thus sought a directed verdict on all of plaintiffs specifications of negligence; it did not move for directed verdicts on each specification individually or otherwise separately seek to remove each specification from the jury’s consideration. As a result, as plaintiff points out, the trial court did not err if there is sufficient evidence to support any one of plaintiffs specifications, Mauri v. Smith, 324 Or 476, 479, 929 P2d 307 (1996); NW Pac. Indem. v. Junction City Water Dist., 296 Or 365, 372 n 1, 677 P2d 671 (1984), and that specification is not preempted by federal law.

Defendant responds by relying on the rule in Whinston v. Kaiser Foundation Hospital, 309 Or 350, 788 P2d 428 (1990). In that case, the Supreme Court adopted the “we can’t tell” rule for reviewing jury verdicts. 309 Or at 358-59. Under that rule, if a court submitted more than one specification of negligence to the jury and there was sufficient evidence to support at least one specification but insufficient evidence to support another specification, and if it was impossible to tell which specification was the basis for the jury’s verdict, the appellate court could not affirm the judgment based on the verdict. Rather, the appellate court was required to order a new trial. According to defendant, we can affirm the negligence judgment in this case under the Whinston rule only if there is evidence to support all of plaintiffs specifications and none of them is preempted.

The problem with defendant’s argument is, as defendant recognizes, that, after the trial in this case, the Oregon Supreme Court overruled Whinston in Shoup v. Wal-Mart Stores, Inc., 335 Or 164, 166, 174, 61 P3d 928 (2003). Defendant argues, however, that the parties relied on Whinston in submitting this case to the jury and that applying Shoup to this appeal in a retroactive fashion would be unfair. However, even if we were to apply the Whinston rule, defendant could not prevail. The court in Whinston emphasized that, in order for a party to avail itself of the “we can’t tell” rule, the “party must have taken some action at trial to remove the unsupported allegations from the jury’s purview.” 309 Or at *60359. The court suggested that the preferable method of doing so would be to seek a suitable peremptory instruction on each element or allegation on which the party believed that it was entitled to prevail as a matter of law. Id. at 359-60. It specifically stated that, “[b]ecause a motion for directed verdict against a party’s entire case is properly denied if any allegation is supported by the evidence,” a blanket motion for a directed verdict is an insufficient basis for invoking the “we can’t tell” rule on appeal. Id. at 360 (emphasis in original).

In this case, defendant made a blanket motion for a directed verdict on plaintiffs negligence claims.16 It therefore did not provide a basis for invoking the Whinston rule even if we were to hold that the rule applies to this case.17 It follows that the issue before us is whether any one specification of negligence is supported by the evidence. Plaintiff alleged generally that defendant was negligent in “designing, testing, controlling, processing, manufacturing, assembling, distributing and supplying the Merit brand of cigarettes * * * in one or more of the following ways.” In his third specification of negligence, plaintiff alleged that defendant was negligent in

“selling and distributing cigarettes which it knew or should have known contained poisonous and carcinogenic substances capable of causing and likely to cause numerous serious and fatal injuries and diseases, including but not limited to cancer of the lungs [.]”

As we have described in the section addressing the fraud claim, there is evidence from which the jury could infer that defendant supplied its Merit brand to the public, knowing that its cigarettes contained carcinogenic substances and that smokers would compensate for the brand’s lower tar and nicotine content by consuming more cigarettes to satisfy their level of addiction. From that evidence the jury could *61find that defendant negligently supplied a hazardous product to the public.

Nonetheless, defendant argues that permitting the above specification of negligence will, in effect, hold it liable for the very act of selling cigarettes, a legal product. It quotes the United States Supreme Court as having held that “Congress * * * has foreclosed the removal of tobacco products from the market.” FDA v. Brown & Williamson Tobacco Corp., 529 US 120, 137, 120 S Ct 1291, 146 L Ed 2d 121 (2000). In Brown & Williamson, the Court considered whether the Food and Drug Administration (FDA) had the authority to regulate nicotine as a “drug” and cigarettes and smokeless tobacco as “devices.” 529 US at 131. In deciding that the FDA did not have that authority, the Court noted that, if the FDA were to regulate those products, it would have no alternative except to ban them because it would be impossible to prove that they were safe for their intended use. Id. at 137. The Court then went on to analyze other federal legislation that, it concluded, foreclosed the FDA from doing so. That legislation included 7 USC section 1311(1)— which states that the marketing of tobacco is one of the great basic industries of the United States — and various other statutes in which Congress has addressed issues of tobacco and health. The Court concluded from the multiplicity of legislation it reviewed that Congress has made a conscious decision not to ban the sale of tobacco but instead to regulate it. Id. at 137-39. Thus, when read in context, the statement on which defendant relies merely summarizes the Court’s conclusion that Congress has decided that it will not, as a matter of federal law, forbid the sale of tobacco.18 That conclusion says little or nothing about a congressional intent to foreclose state-imposed liability for the sale of tobacco products.

Moreover, defendant does not cite any federal statute that expressly preempts state authority over tobacco *62sales.19 In the absence of such a statute, we do not believe that either Congress’s recognition of the economic importance of tobacco or its decision to prohibit the FDA from regulating nicotine or smokeless tobacco implicitly preempts state authority. Neither action indicates that state regulation of tobacco sales conflicts with federal law or that federal law so thoroughly occupies the field that we can infer that Congress has left no room for the states to act. See Cipollone, 505 US at 516 (describing criteria for determining congressional preemption of state action). Even if permitting awards of damages based on plaintiffs specifications of negligence would have the effect of banning a legal product in a particular state, we find nothing in federal law that prevents state action in that regard.20

With respect to defendant’s argument that holding it liable for negligence will effectively hold it liable for selling a legal product, there is a difference in concept between the outright banning of a product for sale and permitting the sale of a product but creating liability for the harm that it causes. Plaintiff alleged that defendant sold and distributed cigarettes when it knew or should have known that its products contained harmful substances capable of causing injury and disease. When properly viewed, holding defendant liable for that conduct falls short of the complete banning of the sale of a product; rather, such an adjudication constitutes the mere exercise of the state’s regulatory power over a product in the *63event that it does cause harm. We reject defendant’s arguments and conclude that, because there is sufficient evidentiary support for at least one specification of negligence and because defendant did not ask the court to rule on each specification separately, the trial court did not err in submitting plaintiffs negligence claim to the jury.

Defendant’s eighth through eleventh assignments of error involve the trial court’s refusal to give certain instructions on negligence and its giving other instructions over defendant’s objections. The trial court gave instructions that are usually approved for negligence cases, but defendant asserts that the special circumstances of this case required special instructions. The legal principles embodied in the instructions that the trial court gave have broad applicability, and we perceive nothing in them that caused prejudice to defendant. A further discussion of defendant’s assignments in that regard would not benefit the bench or the bar. In sum, we hold that the trial court did not err in submitting plaintiffs specifications of negligence to the jury.

The jury also awarded punitive damages on plaintiffs negligence claim. That award must be vacated for the reason expressed earlier regarding defendant’s fraud claim.

III. THE STRICT PRODUCTS LIABILITY CLAIM

In its twelfth assignment of error, defendant challenges the trial court’s denial of its motion for a directed verdict on plaintiffs strict products liability claim. In his complaint, plaintiff alleged, in part:

“8.

“At all material times, the cigarettes sold by defendant were defective and unreasonably dangerous in one or more of the following respects:

“a. The cigarettes contained added ammonia to increase the effects of nicotine;

“b. The cigarettes or their smoke contained altered pH so as to increase the effects of nicotine;

“c. At the time defendant’s light cigarettes were sold, the product was dangerous and in a condition not contemplated by the ultimate consumer in that it *64was manufactured, marketed, and sold as a less harmful alternative to ordinary cigarettes.”

The jury answered “yes” to the following question: “Was defendant Philip Morris’ Merit brand cigarette defective and unreasonably dangerous to [decedent] when sold and was the defect a cause of [decedent’s] death?”

As with the negligence claim, defendant did not move separately against each of plaintiffs specifications. Thus, we must affirm the trial court’s ruling if any single specification is legally sustainable and if there is evidence to support it. Defendant moved for a directed verdict on the ground that

“there is no evidence from which a reasonable juror could conclude that Merit cigarettes after the repose date, February 1990, were dangerously defective or that such a defect caused plaintiffs harm, and because plaintiffs strict liability claims are preempted by federal law.”

The legal basis for strict liability for defective products in Oregon is ORS 30.920, which the legislature adopted in 1979. In that statute, the legislature adopted the substantive terms of Restatement (Second) of Torts § 402A (1965), and stated its intent that courts interpret the statute consistently with comments a through m of the Restatement. ORS 30.920(3). Under the statute, the seller of a product that is in a condition that is unreasonably dangerous to the user or consumer is liable for physical harm to the user if the seller is engaged in the business of selling that product and if the product reaches the consumer without substantial change in the condition in which the seller sold it. ORS 30.920(1).

Defendant emphasizes that it is liable under the statute only if Merit cigarettes were unreasonably dangerous when it sold them; it points out that comment i to section 402A states:

“The rule stated in this Section applies only where the defective condition of the product makes it unreasonably dangerous to the user or consumer. Many products cannot possibly be made entirely safe for all consumption, and any food or drug necessarily involves some risk of harm[.] * * * The article sold must be dangerous to an extent beyond that which would be contemplated by the ordinary consumer *65who purchases it, with the ordinary knowledge common to the community as to its characteristics. Good whiskey is not unreasonably dangerous merely because it will make some people drunk, and is especially dangerous to alcoholics; but bad whiskey, containing a dangerous amount of fuel oil, is unreasonably dangerous. Good tobacco is not unreasonably dangerous merely because the effects of smoking may be harmful; but tobacco containing something like marijuana may be unreasonably dangerous.”

(Emphasis added.) According to defendant, it follows that it cannot be held liable to plaintiff because all that plaintiff demonstrated was the harmful effects of smoking. We disagree for the reasons that follow.

In McCathern v. Toyota Motor Corp., 332 Or 59, 75-76, 23 P3d 320 (2001), the Supreme Court emphasized, consistently with comment i, that ORS 30.920 creates a “consumer expectation” test for determining when a product is defective. “[T]he plaintiff must prove that, when the product left the defendant’s hands, the product was defective and dangerous to an extent beyond that which the ordinary consumer would have expected.” Id. at 79 (emphasis added); see also Benjamin v. Wal-Mart Stores, Inc., 185 Or App 444, 460-61, 61 P3d 257 (2002), rev den, 335 Or 479 (2003) (applying McCathern analysis). The fact that cigarettes are dangerous does not, as comment i indicates, necessarily make them defective for purposes of the statute. In other words, a reasonable consumer would expect cigarette products to be dangerous, in part because of the federally required warning and in part because of other readily available information. However, that is not the end of the question. If a product is dangerous to an extent beyond that which consumers would have reasonably expected, then liability arises under the statute. We turn then to the record to determine if there is evidence from which a reasonable juror could infer that the Merit brand was more dangerous than a reasonable consumer would have expected.

Plaintiff alleged, and offered evidence, that the Merit brand is defective and unreasonably dangerous for three reasons. Two of those reasons are that the “cigarettes contained added ammonia to increase the effects of nicotine” and that the “cigarettes or their smoke contained altered pH *66so as to increase the effects of nicotine [.]” There is evidence both that defendant added ammonia and urea (which produces ammonia on combustion) and that the result of doing so was to alter the normal pH of the smoke from the burning tobacco in a way that freed up additional “free base” nicotine.21 There is also evidence that defendant added other substances to the tobacco for a variety of reasons, such as adding flavor or binding the tobacco together, and that it manipulated much of the tobacco it used to turn it into expanded tobacco or reconstituted tobacco.

Defendant argues, nevertheless, that none of those changes made the Merit brand more dangerous to decedent. It points out that plaintiffs own expert testified that decedent was addicted to tobacco within two or three years after she began smoking, which was long before she switched to Merit. Thus, according to defendant, whatever changes it made could not have harmed decedent. That argument fails to consider that the jury could have found from the evidence that decedent switched to the Merit brand as “a halfway step to get off the medicine — off the cigarettes.” Moreover, if, as the jury could have found, using the Merit brand caused decedent to smoke more than if she had continued smoking full-flavor brands, then it could have also found that defendant’s modifications made it more difficult for consumers such as decedent to quit smoking entirely.

Finally, defendant argues that plaintiffs claim that the Merit brand is dangerously defective because of the added ammonia and urea is preempted because such a claim is tantamount to imposing a duty to give warnings in addition to those that the federal law requires. We disagree. As the Cipollone Court held, not all state law claims are preempted by 15 USC section 1334(b). Apart from the issue of the wording of label and advertisement warnings being preempted by federal law, there is also the issue under ORS 30.920(1) of the effect on reasonable consumers of defendant’s nondisclosure of its research results to the public health community and the effect of that nondisclosure on the *67recommendations made by that community to consumers. The jury could have viewed defendant’s conduct in that regard as distinct from its conduct relating to compliance with federal warning mandates. Defendant could have avoided liability under the statute in this case, not by making additional warnings, but by disclosing what it knew about its product to the public health community and, ultimately, the consumer public. We conclude therefore that the trial court did not err in denying defendant’s motion for a directed verdict on plaintiffs strict products liability claim for all the reasons expressed above.

For the reasons expressed with regard to plaintiffs fraud claim, the award of punitive damages on the strict liability claim must also be vacated.

On appeal, award of punitive damages vacated and remanded for new trial limited to determination of amount of punitive damages; otherwise affirmed. Cross-appeal dismissed as moot.

concurring.

I agree with the majority that the trial court erred in failing to give the jury an instruction that would properly channel its consideration of the evidence regarding the effect of defendant’s conduct on people in other states. I also agree that the error requires a reversal and a remand. In my view, that remedy follows regardless of the correctness of defendant’s proffered instruction.

The situation here is analogous to State v. George, 337 Or 329, 97 P3d 656 (2004). There, in his trial to a jury for aggravated murder, the defendant raised an insanity defense. As a result, ORS 161.313 required the trial court to “instruct the jury in accordance with ORS 161.327,” which sets out the consequences of a verdict of guilty except for insanity. The defendant objected to the uniform instruction designed to comply with ORS 161.313 and proffered an alternative instruction that the defendant believed was more accurate. The trial court, however, concluded that the jury should not be advised at all of the consequences of a guilty except for insanity verdict. The trial court therefore refused to give any instruction on the point.

*68On appeal, the defendant challenged the trial court’s refusal to give his requested instruction. Our court took a narrow view of the defendant’s argument. We concluded that he had raised only the issue of whether his proffered instruction should have been given, not the issue of whether the trial court erred in failing to give any instruction on the point. Because the defendant’s proffered instruction was not an accurate statement of the law, we rejected his challenge to the trial court’s failure to give it. State v. George, 183 Or App 583, 589-90, 54 P3d 619 (2002).

On review, the Supreme Court agreed that the defendant’s requested instruction was not accurate. That court, however, believed that the issue was broader and that the defendant had adequately raised the related question of whether the trial court failed in its obligation to give some instruction that comported with ORS 161.313. George, 337 Or at 336-37. Addressing the merits, the court concluded:

“The answer, on the merits, is inescapable. ORS 161.313 provides that, when a defendant’s sanity is an issue, the trial court ‘shall instruct the jury in accordance with ORS 161.327.’ As we already have noted, that statute unequivocally requires the trial court to give an instruction in accordance with ORS 161.327. The trial court failed to so instruct the jury and, in failing to do so, erred.”

Id. at 340 (emphasis in George).

In reaching that conclusion, the court rejected the state’s argument that the defendant’s failure to proffer an accurate instruction precluded the defendant’s challenge. The court did so for two independent reasons:

“Although the state’s position accurately reflects the approach that we ordinarily take regarding preservation of jury instruction issues, we reject that argument here for two reasons. First, it cannot be squared with the fact that ORS 161.313 unequivocally places the responsibility for giving the required instruction on the trial court, without regard to whether the defendant wants or requests such an instruction, much less offers one that is a correct statement of the law. Second, the state’s argument ignores the fact that, at the relevant time, the trial court already had announced that * * * it would not give the uniform instruction to the jury, ‘nor would [it] give a variation of it.’ In view *69of that announcement, defendant reasonably could assume that attempting to formulate a revised instruction that comported with the requirement of ORS 161.313 would have been an exercise in futility. Our requirements respecting preservation do not demand that parties make what the record demonstrates would be futile gestures.”

Id. at 339.

The same result should follow here. Defendant preserved an objection to the trial court’s failure to give such an instruction and did so in two ways. First, defendant objected to the punitive damages instruction requested by plaintiff,1 urging, among other objections, that “it allows punishment for out-of-state conduct.” In other words, defendant objected to the correctness of plaintiffs instruction on the theory that it contained no limitation of the kind required by federal due process principles.2 In addition, defendant proffered a requested instruction that would have limited the jury’s consideration of defendant’s out-of-state conduct.

Under the court’s analysis in George, the correctness of defendant’s proffered instruction is of no moment, because the trial court was required by the federal constitution to give a correct instruction regardless of defendant’s request or the correctness of its proffered instruction. As the majority in this case explains, in State Farm Mut. Automobile Ins. Co. v. Campbell, 538 US 408, 422, 123 S Ct 1513, 155 L Ed 2d 585 (2003), the United States Supreme Court held that, as a requirement of due process, a jury “must be instructed” on the limitations of its consideration of evidence of a defendant’s out-of-state conduct. The Court’s declaration in that regard was unqualified. That is, the Court did not suggest that such an instruction is necessary only if requested. Rather, the Court held that such an instruction is necessary so that the jury’s deliberations are constrained consistently with the federal due process limitations on a state’s power to punish a defendant’s conduct. Id. Just as the trial court in *70George was required to instruct the jury on the consequences of a guilty except for insanity verdict, so too was the trial court in this case required to guide the jury’s consideration of the evidence of out-of-state conduct and out-of-state harm.

The correctness of defendant’s proffered instruction also is of no moment for the second reason identified in George — that is, the trial court was unwilling to give any instruction limiting the jury’s consideration of the out-of-state evidence. The issue, as framed by the parties, posed an all or nothing proposition. Plaintiffs position was that no limitation was required, because defendant’s conduct was unlawful in all jurisdictions. Defendant’s position was that the jury was not entitled to award punitive damages based on its out-of-state conduct and harm to individuals in other states. The trial court agreed with plaintiff. As in George, given the trial court’s reasons for rejecting defendant’s proffered instruction, any effort by defendant to reformulate its requested instruction would have been an exercise in futility.

In sum, in failing to give a limiting instruction required by the Due Process Clause, the trial court erred. Under George, regardless of the correctness of the instruction that defendant requested, the case must be remanded for a new determination of the amount of punitive damages.

concurring in part, dissenting in part.

I agree with the majority in its disposition of all of defendant’s assignments of error except for one. The majority erroneously concludes that one of defendant’s proposed instructions on punitive damages correctly stated one of the limits imposed by the Due Process Clause on the award of punitive damages in this case. The proposed instruction did not correctly state the applicable limit. Hence, the trial court did not err in refusing to give it.

Defendant asked the court to give the following instruction:

“You are not to punish a defendant for the impact of its conduct on individuals in other states.”

*71The majority concludes that the instruction stated the legal principle, established in BMW of North America, Inc. v. Gore, 517 US 559, 116 S Ct 1589, 134 L Ed 2d 809 (1996), and later confirmed in State Farm Mut. Automobile Ins. Co. v. Campbell, 538 US 408, 123 S Ct 1513, 155 L Ed 585 (2003), that the Supreme Court requires to be conveyed to a jury at a defendant’s request. See 206 Or App at 46, 50-51. Contrary to the majority’s view, the instruction did not state the principle that the Court established in Gore.

Gore established that a state “does not have the power * * * to punish [a defendant] for conduct that was lawful where it occurred and that had no impact [on the state] or its residents.” 517 US at 572-73 (footnote omitted). The Court confirmed in Campbell that a jury must be told of that principle in appropriate cases:

“A jury must be instructed, furthermore, that it may not use evidence of out-of-state conduct to punish a defendant for action that was lawful in the jurisdiction where it occurred. Gore, 517 U.S. at 572-573 (noting that a State ‘does not have the power ... to punish [a defendant] for conduct that was lawful where it occurred and that had no impact on [the State] or its residents’).”

538 US at 422.

Defendant did not request an instruction that stated that principle. In fact, plaintiff argued to the trial court that Gore stood for the principle that evidence of defendant’s lawful conduct in other states could not be used to punish defendant in Oregon, but that the conduct at issue in this case was unlawful in every state, so an instruction stating the principle that Gore established would not be appropriate. It was that argument that appears to have led the trial court to reject defendant’s proposed instruction. Gore was solely concerned with one state’s ability to punish a defendant for the defendant’s lawful conduct in other states. In that light, it should be evident that the majority is simply wrong to say that defendant’s proposed instruction is an instruction required by Gore.

The Court in Gore specifically did not decide “whether one State may properly attempt to change a tortfeasor’s unlawful conduct in another State.” 517 US at 573 n 20 *72(emphasis in original). The Court took a step toward answering that question in Campbell, which was decided after the trial in this case. The Court explained in Campbell that “each State alone can determine what measure of punishment, if any, to impose on a defendant who acts within its jurisdiction.” 538 US at 422 (citation omitted).

It is not entirely clear how that principle applies in a case such as this. Campbell involved claims by Utah plaintiffs against their insurance company over the way in which the insurance company had defended them against tort claims arising from an automobile accident. The plaintiffs sought and recovered punitive damages against the insurance company based, in part, on its conduct in handling insurance claims involving its insureds throughout the country. The nature of that conduct was such that its effect on insureds in one state could never extend to other states or to people in them. For example, the harm that the Campbell plaintiffs suffered in Utah as a result of the insurance company’s tortious conduct was harm that would not have an effect outside of Utah.1 It was in that context that the Court said in Campbell that it is for “each State alone [to] determine what measure of punishment, if any, to impose on a defendant who acts within its jurisdiction.” Id.

That principle is consistent with a principle that was critical to the decision in Gore. The Court was careful in Gore to make clear that the limitation that it imposed was a limitation on the power of a state “to punish [a defendant] for conduct that was lawful where it occurred and that had no impact on [the State] or its residents.” 517 US at 573 (emphasis added). Campbell extended that principle to apply to the imposition of punitive damages by a state for unlawful conduct in other states that had no impact on the state or its residents.

In a tobacco case such as this, however, defendant’s conduct in other states that causes harm to people in those states can have an effect on Oregon and its residents. The conduct at issue in this case was conduct that was nationwide *73in scope, it extended over decades, and it endangered the health of people throughout the country. Because of the mobility of people in our country, each state can be adversely affected by the health-care costs incurred in the state as a result of tobacco-related diseases contracted by people who move to the state after having been induced to purchase and smoke cigarettes in other states by defendant’s unlawful conduct in those states.

For example, Michelle Schwarz began smoking full-flavor cigarettes when she was a nursing student in Missouri. She later moved to Oregon, where she eventually switched to Merit cigarettes, contracted lung cancer, and died. Her experience is not unique, and the jury reasonably could infer that many people in Oregon have contracted tobacco-related diseases and incurred health-care costs as a result of defendant’s conduct toward those people when they lived in other states. Consequently, Oregon and its residents can be affected by defendant’s unlawful conduct in other states that affected people in other states. Because Oregon and its residents can be affected in that way, it may be permissible under the Due Process Clause for Oregon to punish defendant for the effect of its unlawful conduct on people in other states. In other words, in a case such as this, it may be permissible for Oregon to punish defendant to deter its unlawful conduct in other states toward people in those states. If so, then defendant’s proposed instruction was unquestionably an erroneous instruction, and the trial court did not err in refusing to give it.

The majority understands Campbell to sweep more broadly and to stand for the principle that an Oregon jury cannot, under any circumstances, punish a defendant for the effect in other states of the defendant’s unlawful conduct in those states. Even if that understanding of Campbell is correct, that does not mean that a jury cannot award punitive damages that reflect the fact that the defendant’s unlawful conduct in other states had harmful effects in those states. Gore recognized that distinction in explaining the relationship between the imposition of punishment for lawful conduct in other states and the manner in which a state can impose punishment that reflects the fact that the defendant’s unlawful conduct in other states caused harm in those states:

*74“Our cases concerning recidivist statutes are not to the contrary. Habitual offender statutes permit the sentencing court to enhance a defendant’s punishment for a crime in light of prior convictions, including convictions in foreign jurisdictions. A sentencing judge may even consider past criminal behavior which did not result in a conviction and lawful conduct that bears on the defendant’s character and prospects for rehabilitation. But we have never held that a sentencing court could properly punish lawful conduct. This distinction is precisely the one that we draw here.”

517 US at 573 n 19 (emphasis in original; citations omitted).

For example, a court imposing punishment on a defendant who has committed manslaughter in two states cannot sentence the defendant for the manslaughter in the other state, but the punishment that the court imposes can be based on the fact that the defendant committed manslaughter in two states. In this case, that means that the punishment imposed by an Oregon jury for the effect on plaintiff of defendant’s unlawful conduct can be based on the fact that that conduct affected people in other states in the same way that it affected plaintiff, subject to the limit imposed by the Due Process Clause on the size of the award. In other words, the jury could punish defendant more severely in Oregon because of the fact that the unlawful conduct that harmed plaintiff also harmed people in other states. What the jury could not do is to punish defendant specifically and independently for the effect of its unlawful conduct on people in other states. Defendant’s proposed instruction would have misled the jury by telling it that it could not punish defendant at all for the effect of its conduct on people in other states. Because the instruction would have misled the jury in that way, the trial court did not err in refusing to give it. See, e.g., Beglau v. Albertus, 272 Or 170, 179, 536 P2d 1251 (1975) (“It is fundamental that a request for an instruction may properly be denied, without error, unless the requested instruction is clear and correct in all respects, both in form and in substance, and unless it is altogether free from error.”); see also Bennett v. Farmers Ins. Co., 332 Or 138, 153, 26 P3d 785 (2001).

Finally, defendant’s proposed instruction was flawed because it did not distinguish between defendant’s *75conduct in Oregon and its conduct in other states. The Due Process Clause restricts a state’s authority to punish a defendant for conduct occurring outside the state that does not affect the state or those in it. As noted above, the conduct at issue in this case was nationwide in scope and extended over decades. Because of the mobility of people in this country and the period over which defendant’s unlawful conduct occurred, many states could punish defendant for the harm suffered by a particular individual. For example, a Merit smoker who lived in Idaho but regularly vacationed and purchased Merit cigarettes in both Oregon and Washington and who contracted lung cancer and died in Idaho would be a person for whose injuries any of the three states could impose punishment, because defendant’s conduct in each state could have contributed to the person’s injuries.2 Consequently, the people affected by defendant’s conduct whose injuries could provide a basis for Oregon to impose punishment would not be limited to Oregon residents who purchased and smoked Merits, nor to Merit smokers who are diagnosed with or die of a tobacco-related disease in Oregon. There is evidence in this record that over 500,000 people die in the United States each year from tobacco-related causes, and that 6,522 of those deaths occurred in Oregon in 1999, the year that Michelle Schwarz died. In light of that evidence, defendant’s instruction could have misled the jury to believe that it could punish defendant only for the deaths caused by Merit cigarettes in Oregon. Because the instruction could have misled the jury in that way, that is a further reason to conclude that defendant’s proposed instruction was flawed and, hence, that trial court did not err in refusing to give it.

For all of the foregoing reasons, the majority is wrong to conclude that the trial court erred in failing to give defendant’s proposed instruction. It follows that it is wrong to remand the punitive damage award for a new trial.3

*76My conclusion that the proper disposition of defendant’s appeal is to affirm on defendant’s assignments of error *77necessarily means that I disagree with the majority that plaintiff’s cross-appeal is moot. Plaintiff assigns error to the trial court’s decision to reduce plaintiffs award of punitive damages from the $150 million that the jury awarded to $100 million. For the following reasons, I conclude that the trial court erred in reducing the award.

After the jury returned its verdict, defendant moved to eliminate or reduce the punitive damage award. Defendant’s central argument was that the $150 million punitive damage award violated the federal Due Process Clause because it was grossly excessive.4 The trial court agreed with defendant and reduced the award to $100 million.

Punitive damages in Oregon “have a salutary effect in two respects [: they] visit the wrong-doer with wholesome punishment, and afford an example calculated to deter others from the commission of malevolent acts[.]” Sullivan v. Oreg. Ry. & N. Co., 12 Or 392, 404, 7 P 508 (1885); see also Martin v. Cambas, 134 Or 257, 261, 293 P 601 (1930) (“The generally accepted doctrine is that [punitive] damages are awarded by way of punishment to the offender and as a warning to others, or, according to some authorities, by way of example.”). However, the United States Supreme Court has explained that “[e]lementary notions of fairness enshrined in our constitutional jurisprudence dictate that a person receive fair notice not only of the conduct that will subject him to punishment, but also of the severity of the penalty that a State may impose” for that conduct. Gore, 517 US at 574. Thus, the Due Process Clause “imposes substantive limits ‘beyond which penalties may not go.’ ” TXO Production Corp. v. Alliance Resources Corp., 509 US 443, 453-54, 113 S Ct *782711, 125 L Ed 2d 366 (1993) (quoting Seaboard Air Line R. Co. v. Seegers, 207 US 73, 78, 28 S Ct 28, 52 L Ed 108 (1907)). Where a state imposes a “grossly excessive” punitive damage award against a tortfeasor, it transgresses those constitutional limits. Gore, 517 US at 562. A court’s task on review of a punitive damage award is to subject it to “[e]xacting appellate review” to determine whether it is grossly excessive, Campbell, 538 US at 418, which presents a legal rather than a factual issue. Parrott v. Carr Chevrolet, Inc., 331 Or 537, 555, 17 P3d 473 (2001).

The United States Supreme Court has identified three guideposts to assist courts in assessing whether a punitive damage award is grossly excessive: “(1) the degree of reprehensibility of the defendant’s misconduct; (2) the disparity between the actual or potential harm suffered by the plaintiff and the punitive damage award; and (3) the difference between the punitive damages awarded by the jury and the civil penalties authorized or imposed in comparable cases.” Campbell, 538 US at 418 (citing Gore, 517 US at 575). Those guideposts are designed to serve the constitutional concerns about notice and fairness. I examine the punitive damage award in this case in relation to each guidepost in turn.

The “most important indicium of the reasonableness of a punitive damages award is the degree of reprehensibility of the defendant’s conduct.” Gore, 517 US at 575. The United States Supreme Court has explained that a court should consider the following in analyzing the reprehensibility of a defendant’s conduct:

“[(1) whether] the harm caused was physical as opposed to economic; [(2) whether] the tortious conduct evinced an indifference to or a reckless disregard of the health or safety of others; [(3) whether] the target of the conduct had financial vulnerability; [(4) whether] the conduct involved repeated actions or was an isolated incident; and [(5) whether] the harm was the result of intentional malice, trickery, or deceit, or mere accident.”

Campbell, 538 US at 419.

Defendant’s conduct in this case is remarkably similar to the conduct of the defendant in Williams v. Philip Morris Inc., 340 Or 35, 63, 127 P3d 1165 (2006) (WilliamsIV), *79that the Oregon Supreme Court concluded was “extraordinarily reprehensible.” In that case, the court held that a punitive damages award of $79.5 million against a tobacco company was not grossly excessive. The evidence supporting the underlying fraud and negligence claims in Williams IV is largely similar to the evidence in this case. Williams IV, however, did not involve allegations that the marketing of low-tar cigarettes constituted fraud. Rather, the claims in that case were based on the defendant’s

“40-year publicity campaign * * * to undercut published concerns about the dangers of smoking. [The defendant] and the tobacco industry had known for most of those 40 years, if not all of them, that smoking was dangerous. Nevertheless, they tried to create in the public mind the impression that there were legitimate reasons to doubt the danger of smoking. [The defendant] and the tobacco industry did so to give smokers a reason to keep smoking (or, perhaps more accurately, to undermine one of the main incentives for smokers to stop smoking).”

Id. at 39 (citations omitted). On the facts in that case, the Oregon Supreme Court concluded that the defendant’s conduct met four of the five reprehensibility factors that had been identified by the United States Supreme Court. Id. at 56.

Defendant’s conduct in this case satisfies those same four reprehensibility factors. The harm was plainly physical rather than economic — as a result of defendant’s conduct, plaintiff developed lung cancer, which metastasized to her brain and ultimately caused her death. As did the defendant in Williams IV, this defendant “showed indifference to and reckless disregard for the safety not just of [decedent], but of countless other Oregonians, when it knowingly spread false or misleading information to keep smokers smoking.” Id. Furthermore, defendant’s conduct was repeated and, because it was fraudulent, was the result of deceit. There can be no serious debate on the question of the reprehensibility of defendant’s conduct.

The second guidepost is the “the disparity between the actual or potential harm suffered by the plaintiff and the punitive damages award.” Campbell, 538 US at 418. At trial, *80as on appeal, defendant emphasized the disparity between the $150 million punitive damage award and the $168,514.22 compensatory damage award. From the trial court’s oral ruling on defendant’s motion to reduce or eliminate the punitive damage award, it appears that the court’s motivation for reducing the punitive damage award was to address that disparity.

“The second * * * guidepost examines the ratio between the punitive damage award and the actual or potential harm to the plaintiff.” Williams IV, 340 Or at 48 (citing Campbell, 538 US at 424) (emphasis added). However, the United States Supreme Court has “consistently rejected the notion that the constitutional line is marked by a simple mathematical formula, even one that compares actual and potential damages to the punitive award.” Gore, 517 US at 582 (emphasis in original). In Campbell, the Court “decline [d] again to impose a bright-line ratio which a punitive damages award cannot exceed.” 538 US at 425.

The numerator in the ratio in this case is $150 million dollars. That much is simple. Determining the denominator presents more of a challenge.

In Williams IV, the Oregon Supreme Court made clear that, in calculating the denominator, a court should look only at the harm to this plaintiff and not at harm to others. 340 Or at 61. However, a court may consider the actual as well as the potential harm to this plaintiff. Id. at 60.

The jury valued the actual harm to plaintiff at $168,514.22. Michelle Schwarz’s medical bills totalled $118,514.22, and the jury awarded her $50,000 in general damages for pain and suffering. Michelle Schwarz was diagnosed with lung cancer in February 1998 and died in July 1999. Hence, she battled cancer for a period of approximately 17 months. Had she lived longer, her medical bills could have accumulated to at least $250,000. See Williams TV, 340 Or at 60 (noting that the plaintiffs medical bills for lung cancer treatment “could easily have been 10 or more times the [$25,000] awarded here”). Similarly, “[o]nly chance saved [defendant] from a much higher [general] damage award.” Id. We know from the Williams IV decision that death by *81lung cancer has the potential to cause pain and suffering that a jury could value at at least $800,000. Id. at 44.

Thus, at a minimum, the denominator in the ratio in this case is $1,050,000 ($250,000 in potential medical expenses and $800,000 in potential general damages). Of course, the denominator could have been an even larger number. Causing a human being to develop lung cancer and die is a risky business that is not for the faint of heart; it is impossible to predict exactly how much pain and how many medical bills that person will be forced to endure. We could very easily say that the potential harm to Michelle Schwarz from defendant’s fraud and negligence was $37.5 million, based on a compensatory damage award that a jury returned in a Florida case. See Harold C. Reeder, The ‘Law of Tobacco’ Is a Major Contributing Factor that Hampers Effective Resolution to the Country’s Tobacco Problem, 6 Fla Coastal L Rev 17, 48 (2004) (describing the $37.5 million in compensatory damages awarded to John Lukács).

Based on a numerator of $150 million and a denominator of $1,050,000, the ratio between the punitive damages award and the actual and potential harm to this plaintiff in this case is 143 to l.5 As the Oregon Supreme Court observed in Williams IV, the second guidepost is not met where the ratio “substantially exceed[s] the single digit ratio (9:1) that the [United States Supreme] Court has said ordinarily will apply in the usual case.” 340 Or at 62 (citing Campbell, 538 US at 425). Thus, the question will ultimately turn on whether, in light of where this award stands in relation to all three guideposts, this is a “usual case.” However, before answering that question, it is necessary to address the third guidepost — “the difference between the punitive damages awarded by the jury and the civil penalties authorized or imposed in comparable cases.” Campbell, 538 US at 418.

As the Oregon Supreme Court has explained, analyzing the third guidepost

*82“requires three steps. First, courts must identify comparable civil or criminal sanctions. Second, courts must consider how serious the comparable sanctions are, relative to the universe of sanctions that the legislature authorizes to punish inappropriate conduct. Third, courts must then evaluate the punitive damage award in light of the relative severity of the comparable sanctions. The guidepost may militate against a significant punitive damage award if the state’s comparable sanctions are mild, trivial, or nonexistent. However, the guidepost will support a more significant punitive damages award when the state’s comparable sanctions are severe.”

Williams IV, 340 Or at 58. For the reasons that follow, I conclude that the state’s comparable sanctions are severe and, thus, justify a more significant punitive damage award.

The parties have not identified comparable civil sanctions under state law. However, defendant’s conduct in this case is not unlike the conduct prohibited by the Oregon Unlawful Trade Practices Act. That act provides that a person engages in an unlawful trade practice when it, among other things, “[Represents that * * * goods * * * have sponsorship, approval, characteristics, ingredients, uses, benefits, quantities or qualities that they do not have * * *.” ORS 646.608(l)(e) (emphasis added). Here, the jury could have concluded from the evidence that defendant represented Merit brand cigarettes to be a less dangerous alternative to full-flavor cigarettes — a characteristic or quality that Merit cigarettes do not, in fact, have. Under ORS 646.632, a prosecuting attorney can bring an action against a person engaging in unlawful trade practices. In such an action, the prosecuting attorney can recover, on behalf of the state, a civil penalty of $25,000 for each violation, “if the court finds that a person is willfully using or has willfully used [an unfair trade practice].” ORS 646.642(3). The facts in this case could easily support a conclusion that defendant willfully misrepresented that Merit brand cigarettes were a less risky alternative to full-flavor cigarettes when it knew that they were not. The record suggests that Michelle Schwarz smoked a pack a day of Merit brand cigarettes from 1976 until at least 1998— when she was diagnosed with cancer. Thus, defendant sold 365 packs of Merit brand cigarettes to decedent per year for 22 years, for a total of 8,030 packs of Merit brand cigarettes. *83Each sale of a pack of cigarettes constituted a violation of the Unlawful Trade Practices Act. Thus, a prosecuting attorney could have sought a civil penalty of $200,750,000 against defendant for its sales of Merit brand cigarettes to Michelle Schwarz. Thus, the comparable civil sanctions are severe indeed.

The comparable criminal sanctions are as well. Based on conduct similar to defendant’s conduct in this case, the Oregon Supreme Court concluded in Williams IVthat the conduct of the defendant in that case “would constitute at least second-degree manslaughter, a Class B felony.” 340 Or at 59 (citing ORS 163.125(l)(a)). As the Oregon Supreme Court noted, “[i]ndividuals who commit Class B felonies may face up to 10 years in prison and a fine of up to $250,000.” Id. at 59-60 (citing ORS 161.605(2) and ORS 161.625(c)). Thus, the legislature has seen fit to prescribe incarceration for individuals who behave as defendant has in this case, and the deprivation of one’s liberty is a severe penalty indeed. Of course, one cannot incarcerate a corporation. Instead, “[corporations that commit a felony of any class may be fined up to $50,000, or required to pay up to twice the amount that the corporation gained by committing the offense.” Id. (citing ORS 161.655(l)(a), (3)). Hence, the criminal penalty that could be imposed against defendant for its conduct in causing Michelle’s death is at least $50,000.

In short, comparable civil and criminal sanctions for defendant’s conduct could exceed $150 million. As a result, the third guidepost supports a similar punitive damage award.

Having reviewed the punitive damage award in relation to the three guideposts, we find ourselves in the same situation in which the Oregon Supreme Court found itself in Williams IV. That is, of the three guideposts, “two support a very significant punitive damage award. One guidepost — the ratio — cuts the other way.” Williams IV, 340 Or at 62-63. Because of the similarity between the two cases, it is useful to set forth the Oregon Supreme Court’s analysis:

“The * * * guideposts are not bright-line tests. See, e.g., Campbell, 538 US at 425 (‘there are no rigid benchmarks that a punitive damages award may not surpass’); see also *84Gore, 517 US at 582 (‘we have consistently rejected the notion that the constitutional line is marked by a simple mathematical formula’). In other words, the guideposts are only that — guideposts. Gore also referred to them as indicia. 517 US at 575 (reprehensibility is ‘most important indicium’); id. at 580 (ratio is ‘second and perhaps most commonly cited indicium’); id. at 583 (comparable sanctions ‘provides a third indicium for excessiveness’). Campbell specifically contemplated that some awards exceeding single-digit ratios would satisfy due process. See [538 US] at 425 (‘in practice, few awards exceeding a single-digit ratio between punitive and compensatory damages, to a significant degree, will satisfy due process’). Single-digit ratios may mark the boundary in ordinary cases, but the absence of bright-line rules necessarily suggests that the other two guideposts — reprehensibility and comparable sanctions— can provide a basis for overriding the concern that may arise from a [higher] ratio.

“And this is by no means an ordinary case. [The defendant’s] conduct here was extraordinarily reprehensible, by any measure of which we are aware. It put a significant number of victims at profound risk for an extended period of time. The State of Oregon treats such conduct as grounds for a severe criminal sanction, but even that did not dissuade [the defendant] from pursuing its scheme.

“In summary, [the defendant], with others, engaged in a massive, continuous, near-half-century scheme to defraud the plaintiff and many others, even when [the defendant] had reasons to suspect — and for two or more decades absolutely knew — that the scheme was damaging the health of a very large group of Oregonians — the smoking public — and was killing a number of that group. Under such extreme and outrageous circumstances, we conclude that the jury’s $79.5 million punitive damage award against [the defendant] comported with due process, as we understand that standard to relate to punitive damage awards.”

340 Or at 63-64 (emphasis in original).

Although defendant’s low-tar fraud may be of more recent vintage than the fraud at issue in Williams IV, defendant’s conduct with regard to low-tar cigarettes is no less “extreme and outrageous” than the conduct at issue in Williams IV. In fact, defendant’s low-tar fraud was even more “extraordinarily reprehensible” in that it demonstrated *85another level of sophistication in defendant’s campaign of deception. That is, not only did defendant try to create doubt in the mind of the consumer about the dangers of smoking, but it also offered those smokers whose concerns lingered a purportedly less hazardous alternative — an alternative that defendant knew was no less hazardous. Furthermore, defendant’s conduct exposed it to civil and criminal liability in the hundreds of millions of dollars. For those reasons, I would conclude that the reprehensibility of defendant’s conduct and the comparable civil and criminal sanctions override any concern that arises from the ratio between compensatory and punitive damages in this case. The jury’s $150 million punitive damage award was not grossly excessive and comported with the requirements of due process. Thus, the trial court erred in reducing the punitive damage award in this case from $150 million to $100 million.

In sum, the jury’s original $150 million punitive damages award complied with the federal Due Process Clause. The trial court erred in reducing the award. Therefore, the proper disposition of plaintiffs cross-appeal is to reverse the trial court’s judgment and remand with instructions to enter a judgment in accordance with the jury’s verdict. The majority errs in failing to do that.

concurring in part, dissenting in part.

I join in Judge Armstrong’s opinion, but write separately to emphasize one of his points and to clarify my understanding of it — namely, that the proposed instruction was properly rejected because it would very likely have produced confusion in the minds of the jurors and misled them about the use they could make of evidence of out-of-state harms and of defendant’s out-of-state conduct. Under State Farm Mut. Automobile Ins. Co. v. Campbell, 538 US 408, 427, 123 S Ct 1513, 155 L Ed 2d 585 (2003), and BMW of North America, Inc. v. Gore, 517 US 559, 116 S Ct 1589, 134 L Ed 2d 809 (1996), in assessing punitive damages, a jury is permitted to consider out-of-state conduct in determining the reprehensibility of the defendant’s conduct. See Gore, 517 US at 574 n *8621; see also Campbell, 538 US at 419 (“ ‘[T]he most important indicium of the reasonableness of a punitive damages award is the degree of reprehensibility of the defendant’s conduct.’ ” (Quoting Gore, 517 US at 575; brackets in Campbell.)). The proposed instruction stated, “You are not to punish a defendant for the impact of its conduct on individuals in other states.” That instruction would have misled the jury into thinking that it could not consider defendant’s out-of-state conduct for any purpose.

The majority questions whether that issue is properly before us. 206 Or App at 52. It notes that the trial court did not exclude the proposed instruction on the ground that it was misleading and suggests that plaintiff has not adequately advanced the argument on appeal as an alternative basis for affirmance. I disagree. Plaintiff argued to this court that the instruction was properly excluded because it did not reflect the principle that “out of state conduct and harms caused thereby are * * * relevant to reprehensibility.” Thus, the argument is properly before us. I turn to it now.

I find two serious interrelated flaws in the proposed instruction. The first lies in the word “punish.” I believe that, in the minds of the jurors in a case of this type, “punitive damages” and “punishment” are synonymous. Although, in Campbell, the Supreme Court distinguished between “punishing” a defendant directly for harms occurring out-of-state and using evidence of out-of-state conduct in assessing reprehensibility, the proposed instruction in this case made no such distinction. Had the jury been given that instruction without further guidance, it would not have been likely to draw that distinction for itself. Thus, had the jury been instructed that it could not “punish a defendant for the impact of its conduct on individuals in other states,” the jurors would likely have inferred that they could not consider such conduct at all in determining the appropriate amount of punitive damages. See White v. Ford Motor Co., 312 F3d 998, 1016 n 69 (9th Cir 2002) (“In some cases the distinction between using the evidence as it bears on reprehensibility but not as a measure of damages might be so gossamer as to be difficult for a jury to apply * * *.”); see also Steven R. *87Hamlin, Punitive Damages after Campbell: The Role of Out-of-State Conduct, 28 Campbell L Rev 63, 75 (2005) (questioning how holding that out-of-state conduct may be used to determine reprehensibility can be squared with admonition that such conduct may not be used to punish the defendant).

The second flaw in defendant’s proposed instruction lies in the word “impact.” It appears from the instruction conference with the trial court that defendant intended to convey to the jury that it could not punish defendant for the harms that its out-of-state conduct wrought on people in other states. But that is not what the proposed instruction said. The proposed instruction said that the jury could not punish defendant for the “impact of its conduct on individuals in other states.” (Emphasis added.) “Impact” is not an especially precise word. It could mean, for example, “the act of impinging or striking (as of one body against another or of a stream squarely against a fixed or moving surface).” Webster’s Third New Int’l Dictionary 1131 (unabridged ed 2002). Defendant might have intended its proposed instruction to convey that direct and immediate sort of impact — in other words, to inform the jury that it could not impose punitive damages as retribution for particular harms suffered by people in other states.

However, regardless of what defendant intended, the jury could have inferred that the instruction had a broader meaning. “Impact” can also mean “the effect or influence of one person, thing, or action on another.” The New Oxford American Dictionary 851 (1st ed 2001). Under that definition, the proposed instruction could have been understood to be even more restrictive. The jury may have understood the proposed instruction to prevent them from considering anything occurring out-of-state in assessing reprehensibility.

For example, the jury would have understood the proposed instruction to apply to the factors that were laid out in the punitive damages instruction,1 which provided, in part:

*88“Punitive damages, if any, shall be determined and awarded based on the following:

“No. 1, the likelihood at the time that serious harm would arise from defendant’s misconduct;

“No. 2, the degree of the defendant’s awareness of that likelihood;

“No. 3, the profitability of the defendant’s misconduct;

“No. 4, the duration of the misconduct and any concealment of it;

“No. 5, the attitude and conduct of the defendant upon discovery of the misconduct;

“No. 6, the financial condition of the defendant;

“And, No. 7, the total deterrent effect of other punishment imposed on the defendant as a result of the misconduct, including, but not limited to, punitive damages awards to persons in situations similar to the claimant’s and the severity of criminal penalties of the defendant [that have] been or may be subjected.”2

The impact of defendant’s out-of-state conduct is certainly relevant, for example, to defendant’s awareness of the likelihood of serious harm. The jury was permitted to conclude that defendant’s conduct was reprehensible because defendant was aware that it was likely to have an impact on people — that is, to cause serious harm — nationwide. Likewise, the jury could conclude that the conduct was reprehensible because defendant’s concealment activities were likely *89to have a nationwide impact. The impact of defendant’s conduct on all of those people — not just those who resided in Oregon — is relevant to the reprehensibility of defendant’s conduct. It follows that using the word “impact,” like using the word “punish,” in the proposed instruction would have misled the jury into thinking that it could not impose punitive damages based on the reprehensibility of defendant’s conduct, as determined, in part, by its out-of-state conduct.3

The majority insists that it was not defendant’s burden to propose an instruction that would clarify for the jury how evidence of defendant’s out-of-state conduct should play into its assessment of punitive damages, because defendant was under no obligation to propose an instruction that would benefit plaintiff. 206 Or App at 57. In my view, however, it was defendant’s burden, in crafting a limiting instruction, to propose one that was not overly broad or misleading. Because defendant’s proposed instruction would have conveyed to the jury that it could not consider evidence that it was entitled to consider in assessing punitive damages, the trial court did not err in refusing to give the instruction.

I respectfully dissent.