Since it decided Warner-Lambert, the Commission has considered the imposition of corrective advertising in three adjudicated cases, all of them involving claims made for OTC analgesics. Sterling Drug, Inc., 102 F.T.C. 395 (1983), aff'd, 741 F.2d 1146 (9^th Cir. 1984); Bristol-Myers Co., 102 F.T.C. 21 (1983), aff'd, 738 F.2d 554 (2d Cir. 1984); American Home Products Corp., 98 F.T.C. 136 (1981), aff'd as modified, 695 F.2d 681 (3d Cir. 1982). In none of these cases, however, did complaint counsel prove the factual prerequisites for ordering corrective advertising - - that the deceptive advertisements substantially created or reinforced a false belief and that the belief was likely to linger - - and thus the Commission declined in each case to order corrective advertising. Because Warner-Lambert is the only adjudicated case in more than two decades in which the Commission has ordered corrective advertising, it provides the benchmark⁽⁴⁾ for determining whether the evidence proves⁽⁵⁾ the factual prerequisites for corrective advertising. I do not think that the evidence here proves these prerequisites.

B. Lingering Effect

While the 1996 NFO Study shows that 20% more Doan's users and aware non-users have the false superior efficacy belief than the users and aware non-users of other OTC analgesics, it does not prove that this level of beliefs about Doan's is the lingering effect of the deceptive advertising. Study participants were simply never asked whether they had ever seen any Doan's advertising, much less the particular deceptive advertisements at issue here. Mazis Tr. at 1642, 1644, 1786. It is not impossible that study participants saw the deceptive advertising before it was discontinued in May 1996 and formed the false superior efficacy belief as a result of exposure to this advertising, and that this belief lingered until December 1996. However, a variety of influences -- other than any particular advertising campaign -- create, reinforce, and change consumer beliefs about a product. Given that other, entirely plausible influences could well be responsible for the belief reported in the 1996 NFO Study (such as historic positioning and the introduction of new extra strength Doan's products), I am not willing to infer that the belief is the enduring effect of the discontinued deceptive advertising. Jacoby Tr. at 3005-06; Scheffman Tr. at 2618.

While inference of lingering effect may be considered in this case, the particular inferences that the majority seeks to draw are not persuasive. The majority first infers a lingering effect from the purported powerful impact of the deceptive advertising on consumers, which, in turn, is based on the majority's conclusions that the superior efficacy claim was "very salient" and was made "clearly and consistently." Consumers may have taken away the implied claim immediately after seeing the deceptive advertisements, but only a minimal proportion (between 1% and 8%) of test participants recalled the claim 24 hours or 72 hours after viewing the advertisements along with programming and other advertisements.⁽¹⁶⁾ Similarly, only a minimal proportion (0% top-of-the-mind and 2% total unaided) of consumers recalled any advertising for Doan's, including the deceptive advertisements. RX 2-O. Although consumers could conceivably form a belief about a product based on a deceptive advertisement without being able to recall the claim shortly thereafter or without being able to recall any advertising for the product, the far more plausible conclusion is that the extremely low recall of the deceptive claim and of Doan's advertising means that the deceptive advertisements had no real lasting impact because they were not memorable. Whitcup Tr. at 2123. Indeed, the conclusion that the deceptive advertisements did not have a powerful impact on consumer beliefs is corroborated by the fact that unit sales of Doan's declined during 1988 to 1993, the first five years in which the deceptive advertisements were being disseminated. RX-189-A; Scheffman Tr. at 2550-51; Stewart Tr. at 3487. I am not persuaded that an inference can be drawn that this ineffective advertising campaign caused a false belief that is likely to linger until July 2000 or beyond, more than four years after Novartis ceased disseminating the deceptive advertisements.

For commercial speech to come within [the First Amendment], it at least must concern lawful activity and not be misleading. Next, we ask whether the asserted governmental interest is substantial. If both inquiries yield positive answers, we must determine whether the regulation directly advances the governmental interest asserted, and whether it is not more extensive than is necessary to serve that interest.

447 U.S. at 566.

I agree with my colleagues that the initial portions of the Central Hudson test have been satisfied, see Warner-Lambert, 562 F. 2d at 771 (corrective advertising is intended to serve the substantial governmental interest of protecting citizens against deception), but I disagree that the corrective advertising provision here "directly advances the governmental interest asserted" and is "not more extensive than is necessary to serve that interest."

A. Direct Advancement of Substantial Governmental Interest

Central Hudson requires that the restriction on commercial speech "directly advance [ ] the governmental interest asserted." 477 U.S. at 566.⁽²³⁾ This "is not satisfied by mere speculation or conjecture; rather [the government] must demonstrate that the harms it recites are real and that its restrictions will in fact alleviate them to a material degree." Edenfield, 507 U.S. at 770-71; see also 44 Liquormart, Inc. v. Rhode Island, 116 S. Ct. 1495, 1509 (1996) ("some impact" in redressing harm is not enough; ban on alcohol price advertising must "significantly reduce alcohol consumption") (emphasis in original). A restriction thus will not be sustained if "it provides only ineffective or remote support for the government's purpose." Edenfield, 507 U.S. at 770, quoting Central Hudson, 447 U.S. at 564; see also City of Cincinnati v. Discovery Network, Inc., 507 U.S. 410 (1993).

Corrective advertising is intended to prevent deception by curing the lingering false beliefs of consumers that were caused by deceptive advertising. The record before us does not demonstrate that the false superior efficacy belief here is likely to linger through the time that the corrective advertising provision will be in effect. As explained above, the only evidence that a heightened level of false superior efficacy beliefs is likely to linger until July 2000 or beyond is the inconclusive 1996 NFO Study, the unsupported and indefinite testimony of Dr. Mazis, and the unwarranted broad inferences that the majority draws. This weak evidence of lingering effect does not satisfy the Commission's burden of showing direct advancement of a substantial governmental interest, because a corrective advertising provision cannot prevent deception arising from false superior efficacy beliefs in the absence of proof that such lingering beliefs are likely to exist. See Rubin v. Coors Brewing Co., 514 U.S. 476, 490 (1995) ("anecdotal evidence" and "educated guesses" are not sufficient); Edenfield, 507 U.S. at 771 (conclusory testimony is not sufficient).⁽²⁴⁾

B. No More Extensive than Necessary

The corrective advertising requirement also violates the last prong of Central Hudson, 477 U.S. at 566, which requires that the governmental restriction be no more extensive than necessary to serve the asserted governmental interest. See also Warner-Lambert, 562 F.2d at 758 (Commission has a "special responsibility to . . . order corrective advertising only if the restriction inherent in its order is no greater than necessary to serve the interest involved"). This means that there must be a "reasonable fit" between the restriction imposed and the government interest sought to be advanced. Board of Trustees of the State Univ. of N.Y. v. Fox, 492 U.S. 469, 480 (1989). "[I]f there are numerous and obvious less-burdensome alternatives to the restriction on commercial speech, that is certainly a relevant consideration in determining whether the 'fit' between ends and means is reasonable." City of Cincinnati, 507 U.S. at 417 n.13; see also Rubin, 514 U.S. at 490-91 (no reasonable fit between restriction and governmental interest existed because less restrictive options were available). In analyzing the fit between the restriction and the governmental interest, the government must carefully calculate the costs and benefits associated with the restriction. City of Cincinnati, 507 U.S. at 417-18; Fox, 492 U.S. at 480.

The majority addresses in one short paragraph whether the corrective advertising provision here is a reasonable fit with the asserted governmental interest in preventing deception. The paragraph states that the Commission has balanced the need for correcting lingering false beliefs against Novartis's ability to broadcast effectively, the upshot of which is to exempt short television and radio advertisements from the corrective advertising requirement. Majority Op. at 37. Thus, except for not applying the corrective advertising requirement to short television and radio advertisements, the majority does not consider any less restrictive alternatives. This minimal analysis is not the careful calculation of the costs and benefits associated with alternatives that Central Hudson requires.

First, the majority does not analyze whether there are any narrower alternatives to imposing corrective advertising, including considering whether traditional cease-and-desist order provisions (such as those contained in Parts I and II of the Order, or triggered disclosure requirements) could be adequate to address future deception.⁽²⁵⁾ Second, assuming that some corrective advertising provision is warranted, the majority does not address in any detail whether there are narrower alternatives to this particular corrective advertising provision. The corrective advertising requirement in this case apparently is intended to closely track the requirement imposed in Warner-Lambert. The respondent in Warner-Lambert was required to make a corrective statement in all advertising until it had "expended on Listerine advertising a sum equal to the average annual Listerine advertising budget for the period of April 1962 to March 1972." 86 F.T.C. at 1515.⁽²⁶⁾ Here, Novartis is required to make a corrective statement in all of its "advertising" (except short television and radio advertisements) for "one year and until the respondents have expended on Doan's advertising a sum equal to the average amount spent annually during the eight years of the challenged campaign." The Order defines an "advertisement" broadly to include any intended inducement to sale that appears in a brochure, newspaper, magazine, free standing insert, marketing kit, leaflet, circular, mailer, book insert, letter, catalog, poster, chart, billboard, public transit card, point-of-purchase display, package insert, package label, product instructions, electronic mail, website, homepage, film, slide, radio, television, cable television, program-length commercial or infomercial, or in any other medium.

Part IV thus imposes a corrective advertising requirement that is nearly identical to the one-year, all-media requirement that the Commission imposed in Warner-Lambert.

While applying the corrective requirement to all media may have been a reasonable fit with the objective of correcting false beliefs in Warner-Lambert, it is not a reasonable fit in this case. In Warner-Lambert, the Commission was trying to correct false beliefs among the general public concerning Listerine mouthwash, and so an all-media corrective advertising provision was consistent with that objective. See Warner-Lambert, 86 F.T.C. at 1501, 1503 (false beliefs exist among "Listerine users as well as nonusers"; "long after Listerine cold efficacy advertising ceased, a substantial portion of the public would continue to believe") (emphasis added). In contrast, the Commission here is trying to correct false superior efficacy beliefs among Doan's users and aware non-users. Mazis Tr. at 1385, 1805 (back pain sufferers who are neither Doan's users nor aware non-users have no need to receive the corrective statement). Therefore, the media chosen for the dissemination of the corrective message here must be targeted to Doan's users and aware non-users if the Commission's remedy is to achieve the reasonable fit that is constitutionally required. See 44 Liquormart, Inc., 517 U.S. 484, 529 (1996) (O'Connor, J., concurring in judgment) ("The scope of the restriction on speech must be reasonably, though it need not be perfectly, targeted to address the harm intended to be regulated.") (emphasis added). Significantly, the difference between the general public as a target audience and Doan's users and aware non-users as a target audience is quite substantial, given that 31% of back pain sufferers (itself a subset of the general public) are neither Doan's users nor aware non-users. Mazis Tr. at 1793.

The corrective advertising requirement here is in no way limited to media that are likely to target Doan's users and aware non-users. One narrower alternative that would more accurately target Doan's users and aware non-users is to require the corrective statement only on product labeling and in packaging. Product labeling and packaging are sources of critical safety and efficacy information for users and potential users of Doan's, such as indications for use, directions, warnings, drug interactions, active ingredients, and inactive ingredients. See Mazis Tr. at 1607-08 (product package can affect beliefs; consumers look at the product package immediately at the point of purchase). Another narrower alternative is brochures with corrective information that would be made available to Doan's users and aware non-users through prominent displays on the drug store shelves and other locations at which Doan's and other OTC analgesics are sold. Indeed, the Commission has used similar media to target a particular group of consumers who have false beliefs to be corrected.⁽²⁷⁾ Although dissemination of a corrective statement through product packaging and point-of-sale displays, either separately or combined, is a less restrictive alternative that may well be adequate to correct the false belief among Doan's users and aware non-users, the majority does not consider the imposition of such alternatives - - much less conduct a careful calculation of their costs and benefits. Therefore, the corrective advertising requirement imposed here has not been demonstrated to be no more extensive than necessary, as Central Hudson requires.

IV. Conclusion

Because the evidence in the record does not prove that the false superior efficacy belief will linger for the requisite period of time for imposing corrective advertising under the standard set forth in Warner-Lambert, and also because the corrective advertising provision is an unconstitutional infringement on Novartis's right to engage in commercial speech under the First Amendment, I dissent from Part IV of the Order.

2. "Advertising" is defined in the Order to include claims made in a brochure, newspaper, magazine, free standing insert, marketing kit, leaflet, circular, mailer, book insert, letter, catalog, poster, chart, billboard, public transit card, point-of-purchase display, package insert, package label, product instructions, electronic mail, website, homepage, film, slide, radio, television, cable television, program-length commercial or infomercial, or in any other medium."

3. See California SunCare, Inc., 123 F.T.C. 332, 391 (1997) (Statement of Commissioner Roscoe B. Starek, III, concurring in part and dissenting in part) (Warner-Lambert imposes a "more demanding standard for corrective advertising" than traditional fencing-in relief, such as affirmative disclosure requirements.).

4. The majority states that the Commission "has frequently noted that the amount of evidence in Warner-Lambert was unusually strong and far exceeded the threshold needed to impose corrective advertising." Majority Op. at 30. As discussed below in the text, the Commission has simply recognized that inference, not direct evidence, may be used in appropriate cases. The availability of inference does not relieve complaint counsel of the burden of proving lingering effect by a preponderance of the evidence. Moreover, Warner-Lambert did set the standard for corrective advertising, and the evidence in that case is the only benchmark that we have for assessing the sufficiency of evidence supporting corrective advertising. See E. Levi, An Introduction to Legal Reasoning 2 (1949) (the extension of a rule of law to new facts "depends upon a determination of what facts will be considered similar to those present when the rule was first announced").

5. Complaint counsel has the burden of proving facts in Commission adjudications by a preponderance of the evidence. Carter Products, Inc. v. FTC, 268 F.2d 461, 487 (9^th Cir. 1959); ABA Antitrust Section, Antitrust Law Developments 617 (4^th ed. 1997) ("The burden of proof in a Commission proceeding is on complaint counsel to establish its case by a preponderance of the evidence.") (footnotes omitted); see 5 U.S.C. § 556(d) ("[e]xcept as otherwise provided by statute, the proponent of a[n] * * * order has the burden of proof.").

6. I am assuming for the sake of argument that the majority is correct that the false superior efficacy belief was caused substantially by the deceptive advertising at issue, rather than by some other entirely plausible factor such as the introduction of new, extra strength Doan's products or the nine decades of positioning Doan's product as an effective remedy for back pain. Compare Sterling Drug Co., 102 F.T.C. at 798-99 (concluding that it was not clear that deceptive advertising campaign was a substantial cause of false efficacy belief because "the longer a brand has been in existence, the less its image stems from one particular advertising campaign," since "[f]or a brand like Bayer, which has been on the market for years, familiarity is the primary influence on brand image").

7. See R. Pitofsky, Beyond Nader: Consumer Protection and the Regulation of Advertising, 90 Harv. L. Rev. 661, 697 (1977) (hereinafter "Pitofsky, Beyond Nader") (false belief must continue to "influence purchasing decisions up to the date of the entry of a final Commission order, and [be] likely to continue to be influential for a substantial segment of potential purchasers even if the false claims [are] no longer disseminated by the seller").

8. Commission cease and desist orders, including their corrective advertising provisions, become final 60 days after service unless the Commission or a court has granted a stay. Section 5(g) of the FTC Act, 15 U.S.C. § 45(g).

9. The corrective advertising provision could last substantially longer than one year because it is required to continue for "one year and at least until the respondent has expended on Doan's advertising a sum equal to the average amount spent annually during the eight years of the challenged campaign" (emphasis added). For instance, although the corrective advertising provision in Warner-Lambert was similarly prescribed to last until the respondent had spent the same amount on advertising as its average recent annual advertising expenditure, the provision was in effect for at least 18 months. Mazis Tr. at 1798.

10. The majority takes the ALJ to task for purportedly requiring that the lingering effect must be proven with certainty. Majority Op. at 21. The ALJ stated that "there is no certainty that the belief at issue requires corrective advertising." ID at 64. While the ALJ's language could have been more precise, the more reasonable understanding of his statement is that the evidence presented as to lingering effect was too uncertain, not that complaint counsel have not accomplished the obviously impossible task of proving lingering effect with certainty.

11. Among users, 62.3% of Advil users, 51.4% of Aleve users, 41.3% of Bayer users, 78.9% of Doan's users, 61.4% of Motrin users, and 43.8% of Tylenol users stated that their own brand was superior for back pain relief. CX-421-V. Among aware non-users, 31.2% of Advil aware non-users, 19.9% of Aleve aware non-users, 27.1% of Bayer aware non-users, 44.6% of Doan's aware non-users, 35% of Motrin aware non-users, and 22.4% of Tylenol aware non-users stated that the brand that they were aware of (but did not use) was superior for back pain relief. Id.

12. Another possible method of proving lingering effect would be through a series of comparable consumer surveys conducted over the course of years demonstrating that the belief is durable. In Warner-Lambert, for example, the Commission concluded that a false cold and sore throat efficacy belief concerning Listerine would persist based on numerous, identical quarterly market research reports over an eight-year period demonstrating that consumers had consistent levels of the belief and that the belief did not diminish substantially during periodic cessations of the advertising during the summer months. 86 F.T.C. at 1472-76, 1503-04. Other than the 1996 NFO Study, the only other extrinsic evidence that purports to show the false superior efficacy belief is the 1993 Brand Equity Study. Like the ALJ, I do not believe that the 1993 Brand Equity Study is probative because the question posed was unclear as to whether participants were being asked if Doan's was very effective in an absolute sense or if Doan's was more effective than other OTC analgesics. FF 246. Consequently, unlike Warner-Lambert, there is no series of comparable tests over the course of years in this case that proves the existence of a stable and enduring false superior efficacy belief.

13. Dr. Mazis also relied on consumer research studies purportedly showing lingering false beliefs about Listerine mouthwash and Hawaiian Punch fruit drink in the 1970s. He provided no analysis of the reasons why the results of these studies are applicable to the specific facts of this case -- false superior efficacy beliefs about an OTC analgesic in the 1990s. Mazis Tr. at 1256-63. Consumers of OTC analgesics may well be subject to significantly different influences than consumers of mouthwash or fruit punch; for example, advertising for OTC analgesics is much more competitive than advertising for mouthwash or fruit punch. Scheffman Tr. at 2603-04, 2626, 2647. Consumers of products in the 1990s also may well be subject to significantly different influences than in the 1970s because of new media, such as cable television, electronic mail, and websites. Without a cogent analysis of why the results of these consumer research studies are applicable to current consumer beliefs about Doan's, I am not persuaded by Dr. Mazis's testimony that these studies prove lingering effect.

14. As an example of how indefinite are Dr. Mazis's testimony and the other evidence on the issue of the duration of the false superior efficacy belief, one need look no further than the disagreement between the majority and complaint counsel over the suitable length of the corrective advertising remedy: the majority has concluded that the evidence warrants a one-year period for corrective advertising, while complaint counsel have argued that (if a fixed period is imposed) the evidence warrants an eight-year period for corrective advertising. CCRB at 40 n. 55.

15. It is extremely difficult to infer any particular duration of a lingering effect from other facts. For example, in this case, what are the differences in length of lingering effect among a material claim, a salient claim, and a very salient claim? What are the differences in length of lingering effect for an implied claim, a nearly express claim, a clear and consistent claim, and an express claim? What are the differences in length of lingering effect among a ten-year, $45 million advertising campaign; an eight-year, $65 million advertising campaign; and a five-year, $75 million advertising campaign? The indeterminate duration of any inferred lingering effect indicates that the case in which inference will support corrective advertising is likely to be the exception, not the rule.

16. FF 141, 148, 153, 157, 164. While these studies may understate the level of advertising claim communication because they are designed primarily to test the memorability of advertisements, not claims in advertisements, see Kraft, Inc., 114 F.T.C. at 126 n.13, they nevertheless raise serious doubt as to whether the deceptive advertisements had the claimed powerful impact on consumer beliefs.

17. In determining whether the deceptive advertisements were so extensive that an inference of lingering false belief can be drawn, the majority rejects any consideration of the extent of advertising by other competitors in the marketplace. Majority Op. at 31. However, in assessing the effects of a deceptive advertising campaign, the Commission should not treat deceptive advertising, especially comparative deceptive advertising,

as if it takes place in a vacuum. For instance, assume that Company A spent $20 million over five years on advertisements making the deceptive claim that Product A is better than Product B, while Company B spent $500 million over the same five years on advertisements making the claim that Product B is better than Product A.

In determining if it can be inferred that Company A's campaign is likely to create the lingering false belief that Product A is superior, the Commission should consider the nature and extent of the advertising campaigns of both Company A and Company B.

18. The majority states that I am emphasizing "the duration of the advertising campaign and the dollars spent in these cases." Majority Op. at 32 n.44. I have addressed the length of deceptive advertising campaigns and the amounts spent during these campaigns simply because they are some of the facts from which the majority is drawing an inference of lingering effect.

19. In Sterling Drug, the Commission did not order corrective advertising because "it ha[d] not been shown that [the deceptive] advertising created or reinforced the public's image of Bayer," 102 F.T.C. at 799, and, therefore, the Commission did not reach the issue of lingering effect.

20. 1997 Annual Report: R.J. Reynolds Tobacco Co. (1997) ("Winston's comprehensive marketing program includes eye-catching billboards and print ads that speak straight to adults with a twist of humor. Point-of-sale displays cut through the marketplace clutter, and new packaging - with distinctive wraparound graphics - reflects the "No Bull" attitude."); American Lung Association, American Lung Association News, "Winston Campaign Attacked by Health Groups" (Aug. 25, 1997) (R.J. Reynolds launched a "massive national advertising campaign to reposition Winston. Ads * * * appeared in such widely circulated publications as People, Glamour, and Inside Sports magazines. Billboards, bus shelters, and other outdoor advertising proclaim Winston as the new cigarette with nothing but tobacco.").

21. Resort to inference is more likely in the context of consent agreements than in adjudicated cases. Extrinsic evidence and expert testimony often are not available to the Commission when it considers a consent agreement, which makes the use of inference more probable. See Eggland's Best, 118 F.T.C. 340, 365 n.3 (1994) (Statement of Commissioner Roscoe B. Starek, III, concurring) ("It is certainly unrealistic to think that we will have [extrinsic evidence of lingering effect] when the respondents enter into a consent agreement before a complaint is filed."). Moreover, because the Commission applies a "reason to believe" standard to consent agreements and a "preponderance of the evidence" standard to adjudicated cases, inference is more likely to suffice in connection with consent agreements than adjudicated cases.

22. The corrective advertising remedy mandates that Novartis make a statement that it finds objectionable in part because its competitors in the highly competitive OTC analgesic market do not have to make such statements. Therefore, the corrective advertising remedy here is a form of compelled speech that is to be analyzed under the Central Hudson test. See Glickman v. Wileman Bros. & Elliott, Inc., 117 S. Ct. 2130, 2139 (1997) (Central Hudson test applies to compelled commercial speech that requires advertisers to "repeat an objectional [sic] message out of their own mouths").

23. The government has the burden of proving that a corrective advertising requirement meets the Central Hudson standard because "[i]t is well-established that '[t]he party seeking to uphold a restriction on commercial speech carries the burden of justifying it.' " Edenfield v. Fane, 507 U.S. 761, 770 (1993), quoting Bolger v. Youngs Drug Products Corp., 463 U.S. 60, 71 n. 20 (1983); see also Ibanez v. Fla. Dept. of Bus. & Pro. Regulation, 512 U.S. 136, 142 n.7 (1994).

24. Similarly, it is unclear that the corrective advertising provision will in fact correct any remaining false superior efficacy beliefs (and thereby prevent deception) to any material degree in the approximately one year that it will be in effect. While testifying that the remedy will correct beliefs much more quickly than if it were not imposed, Dr. Mazis also acknowledged that "[w]e don't know how much faster" and no one "can measure with any precision how long a corrective notice for this particular case should be run." Mazis Tr. at 1975, 1382.

25. In other cases, the Commission analyzed whether other cease-and-desist provisions would substantially prevent deception before concluding that corrective advertising was the "least restrictive means of achieving a substantial and important governmental objective." Warner-Lambert, 562 F. 2d at 770-71; see also American Home Products Corp., 98 F.T.C. at 411 (corrective advertising was not needed in part because a triggered efficacy disclosure would be sufficient to prevent deception).

26. When it issued its decision in 1975, the Commission concluded that the false belief about Listerine would linger "well into the 1980's," 86 F.T.C. at 1504, that is, at least five years after the Commission's order became final. The Commission imposed an approximately one-year corrective advertising requirement to address this lingering effect. This demonstrates an effort to carefully craft a remedy that was not overbroad.