Full opinion text
ENTRY ON PLAINTIFF’S MOTION FOR PRELIMINARY INJUNCTION HAMILTON, District Judge. Introduction Defendant Abbott Laboratories advertises its Similac line of infant formulas under the banner “1st Choice of Doctors.” Abbott’s chief competitor in the infant formula market is plaintiff Mead Johnson & Company. Mead Johnson contends that the “1st Choice of Doctors” claim is false and/or misleading and thus violates Section 43(a)(1) of the Lanham Act, 15 U.S.C. § 1125(a)(1). Mead Johnson contends the claim is misleading because consumers interpret it to mean that most doctors believe Abbott’s product is medically superi- or to Mead Johnson’s product, when in fact ■most doctors do not have a preference between the two leading brands and there is no evidence that one product is medically superior to the other. Mead Johnson seeks a preliminary injunction to stop Abbott from continuing to advertise with the “1st Choice of Doctors” claim. The court held an evidentiary hearing on November 23-25, 1998. Pursuant to Rules 52 and 65 of the Federal Rules of Civil Procedure, the court now states its findings of fact and conclusions of law. As explained below, the court finds that Mead Johnson has shown a substantial likelihood of prevailing on its claim that “1st Choice of Doctors” misleads consumers with respect to Similac in two independent ways. The market research on doctors’ views of the competing brands is not consistent with consumers’ actual and reasonable interpretations of the “1st Choice of Doctors” claim in two ways. First, consumers reasonably interpret the claim to mean that a majority of doctors choose Abbott products over Mead Johnson products. A fair reading of the many doctor surveys in question shows that Abbott products consistently gain the support of only a plurality of doctors. Second, consumers reasonably interpret the “1st Choice of Doctors” claim to mean that doctors base their choices on professional judgments about the relative quality of the products. The surveys that Abbott offers in support of the claim were not designed to elicit a doctor’s exercise of professional judgment as distinct from a “top of the head” product or advertising recall. Because the other equitable factors also weigh in favor of injunctive relief, the court grants Mead Johnson’s motion for preliminary injunction and will enjoin continued misleading advertising using the “1st Choice of Doctors” claim for Abbott’s Similac brand. The relief will not extend, however, to Abbott’s soy-based brand Isomil, as to which Mead Johnson has not shown the claim is misleading. The facts in the case are voluminous, including evidence of numerous market research surveys of doctors and consumers who buy infant formulas. The evidence may be more readily understood if the court first lays out the legal framework that applies to Mead Johnson’s motion for a preliminary injunction under the Lan-ham Act claim. I. Preliminary Injunction Standard In another recent Lanham Act case involving these same parties, the Seventh Circuit set forth its definitive standard for deciding motions for preliminary injunctions: As a threshold matter, a party seeking a preliminary injunction must demonstrate (1) some likelihood of succeeding ■ on the merits, and (2) that it has “no adequate remedy at law” and will suffer “irreparable harm” if preliminary relief is denied. If the moving party cannot establish either of these prerequisites, a court’s inquiry is over and the injunction must be denied. If, however, the moving party clears both thresholds, the court must then consider: (3) the irreparable harm the non-moving party will suffer if preliminary relief is granted, balancing that harm against the irreparable harm to the moving party if relief is denied; and (4) the public interest, meaning the consequences of granting or denying the injunction to non-parties. The court, sitting as would a chancellor in equity, then “weighs” all four factors in deciding whether to grant the injunction, seeking at all times to “minimize the costs of being mistaken.” We call this process the “sliding scale” approach: the more likely it is the plaintiff will succeed on the merits, the less the balance of irreparable harms need weigh towards its side; the less likely it is the plaintiff will succeed, the more the balance need weigh towards its side. Abbott Laboratories v. Mead Johnson & Co., 971 F.2d 6, 11-12 (7th Cir.1992) (vacating denial of preliminary injunction where Abbott showed likelihood of succeeding on merits and court should have presumed irreparable harm from deceptive comparative advertising) (citations omitted). II. Likelihood of Success on the Merits Section 43(a) of the Lanham Act provides in relevant part: (1) Any person who, on or in connection with any goods or services, ... uses in commerce ... any false designation of origin, false or misleading description of fact, or false or misleading representation of fact, which— % % * # * ❖ (B) in commercial advertising or promotion, misrepresents the nature, characteristics, qualities, or geographic origin of his or her or another person’s goods, services, or commercial activities, shall be liable in a civil action by any person who believes that he or she is or is likely to be damaged by such act. 15 U.S.C. § 1125(a). This provision of the Lanham Act has become the principal legal vehicle for one competitor to challenge another’s advertising as false, misleading, and/or deceptive. The prohibitions of Section 43(a)(1)(B) apply “with equal force to (1) statements which are literally false and (2) statements which, while literally true or ambiguous, convey a false impression or are misleading in context, as demonstrated by actual consumer confusion.” Abbott Laboratories, 971 F.2d at 13 (collecting cases); accord, BASF Corp. v. Old World Trading Co., 41 F.3d 1081, 1089 (7th Cir.1994); Sandoz Pharmaceuticals Corp. v. Richardson-Vicks, Inc., 902 F.2d 222, 228-29 (3d Cir.1990); Avis Rent A Car System, Inc. v. Hertz Corp., 782 F.2d 381, 386 (2d Cir.1986); Vidal Sassoon, Inc. v. Bristol-Myers Co., 661 F.2d 272, 277 (2d Cir.1981). Mead Johnson has not tried to frame its case in terms of “literal falsity.” As explained in detail below, the claim “1st Choice of Doctors” is deeply—and almost ingeniously—ambiguous. As a result, Mead Johnson must prove first what messages consumers actually receive when they are exposed to the “1st Choice of Doctors” claim. In the advertising business, these received messages are often called the “takeaway.” Mead Johnson must prove next that the messages are false or misleading. Mead Johnson contends that Abbott’s “1st Choice of Doctors” claim is misleading because it conveys two false messages to consumers: first, that a substantial majority of doctors prefer Abbott’s products over Mead Johnson’s and other competitors’ products, and second, that doctors prefer Abbott’s products because they are medically superior for infants. Mead Johnson contends the first message is false because market surveys of physicians show consistently that fewer than half of physicians have a preference for Abbott’s products over competitors’ products, and that a significant proportion of physicians have no preference among the competing products. Mead Johnson contends the second message is false because the available medical evidence shows that both Abbott’s and Mead Johnson’s products provide excellent nutrition to infants, and neither has been shown to produce clinically better results than the other. Abbott correctly concedes that its “1st Choice of Doctors” claim implies the existence of market research to support the claim. Abbott Br. at 17; see also C.B. Fleet Co. v. SmithKline Beecham Consumer Healthcare, L.P., 131 F.3d 430, 436 (4th Cir.1997) (whether claim of superiority implies survey or tests results to support claim is a question of fact). The “1st Choice of Doctors” claim plainly implies the existence of supporting survey data. The parties have presented a large quantity of evidence, especially concerning doctors’ views on competing infant formula products. In evaluating Mead Johnson’s likelihood of success on the merits, the court addresses that evidence as follows: (1) the infant formula market and the parties’ products; and (2) marketing of infant formulas, focusing especially on the role of doctors in the marketing. The court then turns to: (3) Abbott’s use of the “1st Choice of Doctors” claim in its advertising; (4) the evidence of consumers’ perceptions of the claim; and (5) the evidence of market research among physicians relevant to the claim. 1. Infant Formulas: The Market and the Products Each year approximately 3.8 to 3.9 million babies are born in the United States. About 3.4 million of them will be fed significant amounts of infant formula in their first year of life. Ill Tr. 172. Plaintiff Mead Johnson and defendant Abbott manufacture and sell roughly 90 percent of all the infant formula sold in the United States. I Tr. 26. Total annual sales in the market are greater than $2 billion. See Ex. 506 at D04492; see also Ex. 14. Defendant Abbott Laboratories operates its Ross Division, which manufactures and sells a family of infant formula products. The flagship brand is Similac, which is sold in a version “with iron” and a version with “low iron.” Abbott also sells Isomil, a soy-based formula used for infants who do not tolerate milk products, as well as other specialty infant formulas. Plaintiff Mead Johnson manufactures a competing family of infant formulas. Its flagship brand is Enfamil, which it similarly sells in a version “with iron” and another with “low iron.” Mead Johnson's family of products also includes a soy-based formula called ProSoBee, a lactose-free formula called Lactofree, and other specialty infant formulas. All witnesses in this case agree that both parties’ products are high quality and provide excellent nutrition for infants. No one contends that the formulas are quite as good as human milk, but the manufactured formulas generally produce excellent results in terms of growth and development of infants. See also III Tr. 172 (at age of four to five months, only 10 to 15 percent of infants in the United States are exclusively breast-fed). There are some differences in the parties’ products. As between the flagship brands — Similac and Enfamil — Abbott and Mead Johnson use different mixtures of whey and casein as their principal protein sources. They also use different combinations of fats. More recently, in 1997 both companies began adding so-called free nucleotides to the formulas, and they do so in different proportions. Abbott’s chief doctor, William C. Mac-Lean, Jr., testified about some reasons he believes Abbott’s Similac is better than' Mead Johnson’s Enfamil. Two points are critical concerning that testimony. First, Abbott categorically denies making any explicit claim to medical superiority, as Dr. MacLean repeatedly acknowledged. Second, Abbott makes no claim to medical superiority because it could not support such a claim with scientific evidence. There is no scientific evidence showing a clinically significant difference between these products, whether in terms of long-term or short-term health, growth, and development of infants. On the absence of evidence showing a clinically significant difference, the testimony of Dr. Samuel Fomon was highly credible. Dr. Fomon is a physician with great expertise in infant nutrition. He wrote the authoritative book in the field used by scientists from both companies. Although Dr. Fomon testified as an expert for Mead Johnson in this case, he operates independently of both companies and both companies have turned to him for research and advice. The court credits his evaluation of the medical research. This is not to say that differences between the formulas might not eventually turn out to be medically significant. For marketers of each company, proof of such a difference would be professionally equivalent to the discovery of the Holy Grail or the coming of the Apocalypse, depending on which product turned out to be better. Well-financed research teams for both Abbott and Mead Johnson have been trying for years to prove that one product is clinically better than the other. Neither has succeeded to date. 2. The Marketing of Infant Formulas and the Role of Doctors in Marketing The United States market for infant formulas has two distinct segments, the ‘WIC” and the “non-WIC” markets. The WIC program, an acronym for Women, Infants, and Children, is a program funded by the federal government to promote nutrition among pregnant and nursing women, infants, and young children by subsidizing the purchase of nutritious foods for those who cannot afford ample nutrition. See generally 42 U.S.C. § 1786. WIC contracts in most states are handled on a sole-source basis, with the source selected on the basis of sealed bids, so that price competition is critical in choosing among otherwise acceptable products. Mothers and infants in the WIC program do not have a choice about the brand of formula that is subsidized. Mead Johnson currently has an edge in the share of the WIC market. The WIC program includes nearly half the infants born in the United States. I Tr. 26; Ex. 14. In the non-WIC market, where consumers are free to choose among the competing brands, Abbott has had the edge in recent years in terms of market share, but Mead Johnson has been gaining. The parties’ market research has produced conflicting results as to which company currently has the lead in that market. The court need not and does not make a finding on that point. In the non-WIC market, both Abbott and Mead Johnson use highly sophisticated marketing and sales techniques, and appear to compete vigorously. Both have medical research programs aimed at developing improved products and gaining a competitive edge. Both Abbott and Mead Johnson focus their marketing efforts in the non-WIC market on doctors and other health care professionals. Abbott and Mead Johnson also use direct advertising to consumers (expectant mothers and mothers of newborn infants), plus advertising and marketing efforts directed at major retailers to secure channels for retail distribution with prominent product placement. Abbott and Mead Johnson agree that doctors have significant influence over their patients’ brand selections of infant formula. See, e.g., II Tr. 278-79 (Abbott’s Gottenborg: more than half of mothers are influenced by doctor recommendations); Ex. 17 at A047702 (doctor recommendations have a powerful influence on brand choice); Ex. 43 (Abbott promotion to retailers stating that doctor’s recommendation is strongest influence on brand choice); I Tr. 50-51 (Mead. Johnson’s McCabe: doctor recommendation is the strongest advertising message in the market and sells more product than any other message); III Tr. 85 (Abbott’s market research expert Dr. McDonald: “no doubt that’s one of the most powerful claims that can be made *** it’s a very powerful claim. And obviously it’s used as often as manufacturers can legitimately substantiate it.”). The infant formulas are treated like medical products, and especially first-time mothers may not be confident that they know enough to make a good decision for their babies. Both Mead Johnson and Abbott view new mothers as emotionally vulnerable, especially with the birth of their first children. See I Tr. 33 (Mead Johnson’s McCabe: first time mothers “seek out someone with authority to help them make those choices”); Ex. 43 (Abbott promotion to retailers: “high emotional significance of purchasing infant formula products”); III Tr. 85 (McDonald: doctor’s ratification of brand choice relieves consumer of worrying about choice). Both competitors direct major market research efforts to understand the interactions between doctor and new mother. Both competitors pursue major marketing and advertising programs to influence those intéractions. The doctor’s influence on brand selection is important both when the doctor makes an express recommendation (of one brand or more than one brand) and when the doctor gives her approval to the patient’s selection, as when a patient asks her doctor: “Is it all right if I use [Brand A]?” and the doctor says yes. Also, Abbott and Mead Johnson strive mightily to have hospitals and doctors distribute free samples of their products to new mothers. The doctor or hospital staff may not perceive distribution of free sam-pies as a recommendation, but market research shows that mothers perceive it as such. See, e.g., I Tr. 29-30 (McCabe); see also Ex. 469 at B2328 (recommendations and samples from doctors and hospitals have significant influence on brand selection). In the past, doctors often made specific brand recommendations to their patients. In recent years, however, fewer and fewer doctors have been recommending one particular brand with a high degree of frequency. Most doctors perceive both Abbott and Mead Johnson products as high quality and suitable for infants, which is consistent with the available medical evidence summarized above. Those doctors generally either make no brand recommendation or recommend two or more brands. Mead Johnson and Abbott agree that doctors have become less willing to direct their patients to one specific brand of infant formula. See, e.g., I Tr. 28 (Mead Johnson’s McCabe); Ex. 73 at A016286 and A016290 (Abbott market research); Ex. 74 at A050734 (same). As doctors have become more passive in brand selection, mothers have become more active in the product choice, so both companies have increased their advertising aimed directly at new mothers. I Tr. 28-29 (McCabe). 3. Abbott’s Use of the “1st Choice of Doctors” Claim Abbott features the “1st Choice of Doctors” claim prominently in its advertising, marketing, and packaging. Abbott places a blue flag with the words “1st Choice of Doctors” on all its infant formula product labels. That specific phrasing first appeared late in 1995. Abbott had included similar claims, such as “First choice of more physicians,” on earlier labels beginning in 1989. See Exs. 405 & 406. The uncontradicted testimony of Mead Johnson’s McCabe shows that Abbott recently has been increasing the amount of money it spends to promote the “1st Choice of Doctors” claim. See I Tr. 55-56. Mary Gottenborg is the group product director for Abbott’s line of infant formulas and is responsible for the overall marketing of the products. She described Abbott’s marketing program for infant formulas. She described the first step as developing a “brand vision,” a long-term goal for the product brand. The next step is to develop a “brand positioning statement,” which is a summary of what the marketer wants to communicate to consumers about the product. The brand positioning statement then helps the marketer identify potential 'advertising claims for the product that might be included in advertising for the product. II Tr. 256-57. Gottenborg testified that the brand positioning statement for Similac has three components. The first is that “we are the first choice formula of doctors.” The second is that “we help babies grow and develop.” The third is that “by feeding Similac, we want mothers to have trust and confidence in our product.” II Tr. 260. All advertisements should be consistent with the brand positioning statement. Id. Because the “1st Choice of Doctors” claim reflects all three components of the brand positioning statement, Abbott has featured it prominently in advertising for Similac and the other Abbott infant formulas. Exhibit 22 is a March 1998 report to Abbott on a test of consumer reactions to several informational advertising pamphlets on Similac. Among six pamphlets, the one stressing the “1st Choice of Doctors” claim scored highest in terms of consumers’ likelihood of purchase. II Tr. 273; Ex. 22 at A022986. The report concluded: “Doctor recommendations and the ‘science’ behind the formula appeared to drive purchase interest for this concept, as well as the other concepts tested,” and use of similar pieces emphasizing the claim was “highly recommended.” Ex. 22 at A022983. The advertisement scored very high in terms of convincing consumers that Similac “is the formula doctors prefer.” Ex. 22 at A022990. That market research led to the development and use of Exhibit 39, Abbott’s direct consumer advertising piece that leads with the “1st Choice of Doctors” claim. Ill Tr. 13. The piece then links .the “1st Choice of Doctors” claim directly to a mother’s desire for “only the best for your baby,” and then emphasizes that “doctors know the science of good nutrition.” The message and linkage between doctor preference and superior quality are plain. Abbott also uses the “1st Choice of Doctors” claim in advertising directed at doctors, hospitals, and retailers. See Exs. 38, 43; II Tr. 261, 269 (Gottenborg); I Tr. 52 (McCabe). 4. Consumer- Perception of the “1st Choice of Doctors” Claim Lanham Act' cases charging false or misleading advertising turn on the messages conveyed by the advertising claims in question. Where the plaintiff contends a claim is misleading but not literally false, the plaintiff must prove what messages consumers actually receive from the advertising. (a) Abbott’s Evidence on the Message Conveyed Abbott’s group product director Mary Gottenborg testified that Abbott intends the advertising claim “1st Choice of Doctors” claim to communicate “trust and confidence in the brand.” II Tr. 261. She explained that Abbott wants to convince mothers that they can have trust and confidence in the brand because its products are the “first choice of doctors.” II Tr. 266. “Our intent on communication in this piece was the same as it is on our label; that is, it’s the first choice of doctors and we want to convince mothers that they can have trust and confidence in using our brand.” Id. Gottenborg testified that Abbott viewed the claim as “straightforward,” and that being the “first choice of doctors” implies “trust and confidence” — that is “the take-away that one should perceive from it.” II Tr. 281. She later added that the “1st Choice of Doctors” claim “conveys a special sense of confidence.” II Tr. 298. The court does not agree that the meaning of the “1st Choice of Doctors” claim is straightforward or self-evident. For example, does the claim refer to all types of doctors in all types of practices, or to some particular types of practices? More to the point here, does “1st Choice” refer to nearly all doctors, a substantial majority of doctors, a bare majority of doctors, or perhaps only a bare plurality of doctors? How, if at all, does the claim take into account the possibility that many doctors have no “choice” or preference among the top brands? At a more fundamental level, what does the claim mean by “choice?” Does it refer to a doctor’s exercise of her professional judgment on behalf of patients, or does it refer to a more-superficial brand recognition or top-of-the-head impression about brands? Abbott’s assertion that it intends to communicate “trust and confidence” raises as many questions as it answers. Why should consumers perceive “trust and confidence” from the “1st Choice of Doctors” claim? Abbott agreeg that the claim is one of the most powerful it could use in marketing infant formulas. Ill Tr. 85 (McDonald). Why should that be so? One possible answer is that consumers expect doctors to have used their professional judgment in making any “choices” among competing brands. As Dr. McDonald testified, the claim “says to consumers that a professional who is knowledgeable has ratified it [the choice of brand] and, therefore, allows a consumer not to think anymore about it.” Id. Also, unlike a blander, non-comparative claim such as “trusted by doctors” (see Ex. 417 — Enfamil advertisement for “the formula mothers and doctors have trusted for nearly 100 years”), Abbott’s claim to be “1st Choice of Doctors” is an inherently comparative (actually, a superlative) claim. Abbott’s assertion that it seeks to communicate trust and confidence simply overlooks the comparative or superlative nature of its claim. See III Tr. 85 (McDonald acknowledging that some consumers may interpret the claim as a claim to be “better”). Although Abbott acknowledges that its ubiquitous “1st Choice of Doctors” advertising claim is a powerful claim, Abbott itself has undertaken no consumer surveys to test the actual “takeaway” of the claim. That is, Abbott has not presented any direct evidence that the actual takeaway is the “trust and confidence” message Abbott says it intends to communicate. (b) Dr. Ross Survey of Consumers At the preliminary injunction stage, Mead Johnson was not required to come forward with “full-blown consumer surveys” to prove actual consumer confusion. See Abbott Laboratories v. Mead Johnson, 971 F.2d at 15 (upholding district court finding of likely actual consumer confusion at preliminary injunction stage even though Abbott had not conducted such surveys); accord, United Industries Corp. v. Clorox Co., 140 F.3d 1175, 1183 (8th Cir.1998). In this case, however, Mead Johnson has conducted a systematic survey of consumers’ actual “takeaway” from the “1st Choice of Doctors” claim. That survey was conducted by Dr. Ivan Ross, who was employed by Mead Johnson’s attorneys to conduct such a study. It is the only systematic study of consumers’ actual perceptions of Abbott’s claim. The results, set forth in Exhibit 1, are central to this case. Dr. Ross supervised a survey of consumers through face-to-face interviews at shopping malls in 16 cities across the United States. Participants were screened so as to be women between the ages of 18 and 44 who either had purchased infant formula in the last three months or intended to purchase infant formula within the next 12 months for one or more of their own children. Women in the WIC program were excluded from the survey because they ordinarily are permitted no meaningful brand choice. Dr. Ross used a face-to-face survey because the participants were asked to respond to a “visual stimulus,” a sample package of Abbott’s Similac with the blue flag stating “1st Choice of Doctors.” Neither the interviewers nor the participants were told the purpose or the sponsor of the survey. For the first question to each participant, the interviewer pointed to the blue flag on the label claiming “1st Choice of Doctors” and then asked: “Please tell me what you understand this part of the label to communicate to you.” Ex. 1, Tab A at A-4. The interviewer recorded the response verbatim, then continued to probe: “Anything else that communicates to you?” The interviewer was instructed to probe, that is, to repeat the “anything else” question, until the response was “unproductive,” meaning the respondent said “no,” or “nothing else.” Some respondents were asked that question several times. If the respondent answered by essentially repeating or “playing back” the words “first choice,” the interviewer was then instructed to ask “what do you mean by that?” or words to that effect. See II Tr. 204. The second question was a further follow-up: “Although you may already have mentioned this, what do you understand the wording ‘first choice’ to communicate to you?” The interviewer was again instructed to probe until the response was “unproductive.” The third question asked: “And although you may already have mentioned this, if you have an opinion, what reason or reasons do you think this product was ‘first choice of doctors’?” The interviewer was again instructed to probe for additional reasons until the response was “unproductive.” The interviewer then asked “why do you say that?” and probed the response once with: “Any other reasons you say that?” Ex. 1, Tab A at A-5. The fourth question asked the participants about the types of doctors referred to in the “1st Choice of Doctors” claim. The survey then probed a central problem posed by doctors’ approaches to these competing brands — the fact that a substantial proportion of doctors have no significant preference between them. The fifth question asked: “Now, suppose that a survey of doctors had been conducted by the company which makes this product and that they were relying on the results of that survey to make the statement you see here [“1st Choice of Doctors”]. If you have an opinion, in order for this statement [“1st Choice of Doctors”] to be true, what percent of those doctors would have had to say this product was their first choice?” Interviewers were instructed to probe once: “And why do you say that.” Responses were recorded verbatim. See Ex. 1, Tab A at A-7. The sixth question went a step farther. The interviewer handed the participant a card showing: DISPLAY CARD FOR POSSIBLE SURVEY OF DOCTORS QUESTION: What brand of infant formula, if any, is your first choice? RESULTS: 50% — Said they had no first choice 30% — Said Similac was their first choice 20% — Said another brand was their first choice 100% — Total doctors surveyed Ex. 1, Tab A at A-20. The interviewer then asked: “Suppose that in that possible survey, the doctors were asked, ‘what brand of infant formula, if any, was your first choice?’ And suppose that the results were that 50 percent of the doctors said they had no first choice, 30 percent said Similac was their first choice, and that 20 percent said their first choice was another brand. If those were the results of the survey, would you say that this statement [“1st Choice of Doctors”] is accurate or would you say that this statement is not accurate, or don’t you have an opinion about that one way or the other?” Ex. 1, Tab A at A-7. Interviewers probed with “why do you say that?” Id. Because several questions were open-ended and responses were recorded verbatim, Dr. Ross had to “code” the responses. Dr. Ross and an associate separately coded all responses and then compared their results as a means of quality control. The final results reflect the reconciled coding after Dr. Ross and his associate compared their results. In response to the first question about what the claim “1st Choice of Doctors” communicated, 52.6 percent of respondents gave answers that Dr. Ross and his associate coded as a reflection of qualitative superiority: superior, better, best, etc. Ex. 1, Tables at T-3. Abbott’s market research expert criticized this calculation because Dr. Ross included answers referring only to doctors’ “preferences.” The criticism was valid because a consumer perception of doctor “preference” does not necessarily show that the consumer perceived the preference in terms of a view of product quality. In response to the criticism, Dr. Ross did a new calculation showing that if the answers referring only to doctors’ “preferences” are subtracted, as they properly should have been, the relevant figure is that 41.3 percent understood the claim as making some claim of superiority. See II Tr. at 200-02; Ex. 114, Table 9. In response to questions about doctors’ reasons for choosing Similac, Dr. Ross found that 54.9 percent of participants responded in terms of the quality of the product and/or superiority to other products. Ex. 1, Tables at T-7. The Ross survey did not differentiate between claims of long-term medical superiority defined by clinically proven results, which is the standard Mead Johnson contends is relevant, and less dramatic claims of short-term convenience and tolerance effects. See II Tr. 198-99. In response to the fifth question about how many doctors would need to have identified Similac as their “first choice” for the statement to be true, 84 percent of the respondents said it would have to be more than 50 percent of doctors. Ex. 1, Tables at T-l. Only 1.4 percent of respondents gave answers of less than 50 percent as supporting the claim. Id. The sixth question introduced the hypothetical survey in which 50 percent of doctors had no preference, 30 percent preferred Similac, and 20 percent preferred another brand. That question therefore introduced to participants the problem posed by those doctors having no preference and the possibility of a plurality of doctors preferring Similac. In response to the sixth question, 64.8 percent of respondents said those results would make the “1st Choice of Doctors” statement not accurate, while 19.8 percent said the statement would be accurate, and 15.4 percent had no opinion. Ex. 1, Tables at T-9. Dr. Ross concluded from the results of his study that more than 80 percent of the consumers understood the “1st Choice of Doctors” claim to convey that a majority of doctors choose or prefer Similac. Ex. 1 at 23. He also concluded that more than 40 percent understood the claim as making some claim of product superiority. With respect to the issue of doctors’ reasons for choosing or preferring Similac, more than half attributed doctors’ reasons to the doctors’ views of the quality or superiority of the product. More than 90 percent of those respondents who expressed any opinion about doctors’ likely reasons for their choices thought that the doctors’ opinions would be based on product quality or performance. Abbott criticizes the Ross consumer survey on several major grounds. Abbott contends that the Ross survey used excessive probing, which can introduce biases into a study by forcing responses from a person who has no meaningful response to give. Second, Abbott criticizes the attempt to ask consumers to speculate about why doctors might choose Similac over other brands. Third, Abbott criticizes the attempt to have consumers quantify the proportion of doctors who would have to choose Similac for the “1st Choice of Doctors” claim to be true. The court is not persuaded by these criticisms. First, probing was appropriate here because of the ambiguity of the “1st Choice of Doctors” claim. In market surveys probing can be a powerful tool, but it can also introduce biases and produce artificial responses from a participant who has run out of meaningful ideas on a subject. Compare L & F Products v. Procter & Gamble Co., 845 F.Supp. 984, 996 (S.D.N.Y.1994) (repeating questions “also serves the purpose of clarifying otherwise ambiguous first responses by probing the implications of previous answers”), aff'd, 45 F.3d 709 (2d Cir.1995), with American Home Products Corp. v. Procter & Gamble Co., 871 F.Supp. 739, 748 (D.N.J.1994) (excessive probing undermined credibility of survey evaluating arguably implicit messages); see also Johnson & Johnson-Merck Consumer Pharmaceuticals, 19 F.3d at 133-34 & n. 13 (Dr. Susan McDonald defended survey using repeated probes against attack based on “excessive probing”). Mead Johnson has the burden of showing what consumers understand that claim to mean. A more directive approach would no doubt have been criticized as introducing other biases into the study. Repeated probing was a fair means of determining the “takeaway” from the claim. Also, Abbott’s criticism of the probing rings a little hollow. An advertiser is responsible for both explicit messages in advertising and implicit messages communicated to a reasonable proportion of the audience. Abbott itself has not undertaken any study of the message or “takeaway” that consumers draw from the “1st Choice of Doctors” claim. The National Advertising Division of the Better Business Bureau (the “NAD”) provides an informal and voluntary forum for resolving disputes over advertising, and the parties here have cited its decisions as persuasive authority. The NAD has taken the approach that an advertiser is responsible for supporting “any reasonable interpretation of its claims in an advertisement.” Church & Dwight Co., NAD Case Reports ■% 213, 215. The Advertising Research Foundation similarly emphasizes the need for determining consumers’ perceptions (or “takeaway” in the jargon) from advertising claims. See II Tr. 55-56; Ex. 25. The same responsibility is the foundation of the court decisions recognizing that an advertiser may be found liable for the implied messages or claims in advertising if consumers understand those claims in ways that make them misleading or deceptive. E.g., Abbott Laboratories, 971 F.2d at 13; Sandoz Pharmaceuticals v. Richardson-Vicks, 902 F.2d at 228-29; Avis Rent A Car System, Inc. v. Hertz Corp., 782 F.2d at 386; Vidal Sassoon v. Bristol-Myers Co., 661 F.2d at 277. Abbott’s criticism on this score might have more force if Abbott had offered some conflicting evidence tending to show consumers do not infer something about product quality from the “1st Choice of Doctors” claim. It has offered no such evidence. Abbott’s second criticism — that Dr. Ross invited consumers to speculate about why doctors might prefer Similac — is also unpersuasive. See III Tr. 86-87 (McDonald). Some doctors may have brand preferences for reasons that have nothing to do with product quality. Perhaps they prefer the service from one brand’s sales representative, perhaps they like one company’s support for medical research, or perhaps they have any of a variety of other reasons for their preferences. See, e.g., Ex. 2 at Tab C-29 to C-34 (some verbatim responses from doctors about reasons for preferences); Ex. 456 at PAR00216 (good service and rapport with representative cited by doctors more often than other reasons for preferring Similac). In the court’s view, however, it is perfectly obvious to all the marketing and market research witnesses who testified that, the reason the “1st Choice of Doctors” claim is so powerful is that many consumers assume that such preferences are based on doctors’ professional judgments about product quality. The NAD similarly recognizes that professional recommendations of health-care products are powerful because consumers interpret them as based on professional assessments of prod-uet quality. See Denmat Corp., NAD Case Reports % 156, 161 (dental hygienists’ recommendation of toothpase: where hygienists are presented as experts, their “evaluation and selection of a certain product is expected to be an expression and exercise of their expertise”); Braun, Inc., NAD Case Reports % 18, 20 (“When dental professionals are given competing products to use under the same circumstances and asked a limited question, i.e. which, if either, they prefer, it can be assumed that professional expertise has been exercised in the comparison and expression of preference (if any).”); see also Church & Dwight Co., NAD Case Reports 1j/jb 213, 215 (advertisement linking product feature to dentists’ recommendation did not support claim where survey of recommendations did not show that the feature was the reason for the recommendations). Dr. Ross’s questions on the perceived reasons for doctors’ choices were less an invitation to speculate than a means of confirming what is common wisdom among those in the business. The strength of that wisdom is confirmed by the fact that, among consumers who offered an opinion on the subject, more than 90 percent answered in terms of product quality or superiority. See Ex. 1 at 23 and T-8. Abbott’s third principal criticism of Dr. Ross’s consumer survey is aimed at the questions about the proportion of doctors who would have to have preferred Similac to support the claim. Abbott argues the questions were “grossly unfair” because they ask lay people difficult problems about statistics and patterns of preferences. Ill Tr. 88 (McDonald). As Dr. McDonald observed, the consumers who were surveyed probably do not spend a lot of time thinking about “the base” and the difference between a majority and a plurality. Id. The court would not hold Abbott to a standard of 90 percent preference simply because one-third or so of the respondents thought that is what “first choice” meant, at least initially, before they were asked about the hypothetical survey of doctors. See Ex. 1 at F-13 (total of 32.8% answered 90% or more). The critical difference between a majority and a plurality, however, is well-recognized in the advertising business. See, e.g., Ex. 25. When that difference was introduced expressly in the survey, an overwhelming proportion of respondents believed the “1st Choice of Doctors” claim would not be accurate if it were supported by only a plurality. Ex. 1 at F-15 (64.8 % — not accurate; 19.8 %— accurate; 15.4 % — no opinion). The court agrees with that overwhelming majority, (c) Perception of Medical Superiority ? The Ross survey of consumers shows persuasively that the actual takeaway, the actual message received by consumers, includes a perception that doctors prefer Abbott’s products because doctors believe they are superior to Mead Johnson’s products. That is a reasonable interpretation of the claim, and that is why the claim is so powerful. But Mead Johnson has tried to take the results of the Ross survey a step farther, beyond the consumer perception that doctors prefer Abbott products because they are superior or of higher quality. Mead Johnson argues that this perception by consumers means that “1st Choice of Doctors” amounts to a claim by Abbott that the Abbott products have been proven through scientific research to produce clinically better outcomes for the infants fed with those products. Mead Johnson has devoted a great deal of energy to proving that no such proof of clinical superiority exists. The court is not persuaded by Mead Johnson’s effort to transform the “1st Choice of Doctors” claim into an establishment claim for clinically proven superiority. First, Mead Johnson’s argument on this point would effectively erase all distinctions between different kinds of advertising claims. Under Mead Johnson’s approach, any advertising that effectively persuades consumers to buy a health care product (perhaps so long as it did not emphasize price) would be treated as implicitly communicating a claim of scientific proof of superiority. Second, and more important here, Mead Johnson has not supported this argument with the evidence of consumers’ actual perceptions of the Abbott claim. The Ross survey shows that consumers “takeaway” general perceptions of product superiority from the “1st Choice of Doctors” claim. It did not go so far as to show that consumers perceived the “1st Choice of Doctors” claim in terms of superiority as specifically as Mead Johnson has argued — long-term clinically proven medical superiority. A doctor may have a significant professional preference between competing products based on perceptions of quality that are not supported by sophisticated clinical trials testing long-term results. Thus, the weight of the evidence presented thus far shows that the “1st Choice of Doctors” claim communicates two key messages to consumers: (1) a majority of doctors choose Similac or other Abbott brands over Enfamil and other competitors; and (2) the doctors’ choices are the result of the exercise of the doctors’ professional judgment about the quality of the products. The court next takes up the question whether Mead Johnson has shown that those messages are not fairly supported by the evidence about doctors’ choices or preferences among the competing brands. 5. The Evidence on the Market Research Among Doctors The central facts affecting all the market research and advertising in dispute in this case are these: surveys of doctors (especially pediatricians) show consistently (a) that a substantial proportion of doctors do not have a professional preference between Mead Johnson’s Enfamil brand and Abbott’s Similac brand, but (b) that among doctors who do have a brand preference, more favor Abbott’s Similac over Mead Johnson’s Enfamil brand. The precise numbers vary from study to study. The different results arose from differences in survey questions and methodology, such as the strength of the “filter” question to screen out doctors who have no preference, how hard participants were pressed to express a preference for only one brand, and the extent to which the survey sought to evoke a thoughtful exercise of professional medical judgment as opposed to a doctor’s quick, top-of-the-head recall of brands. Nevertheless, despite significant differences in the exact numbers from survey to survey, the overall pattern is consistent. In challenging the validity of Abbott’s claim to be “1st Choice of Doctors,” Mead Johnson focuses on the lack of preference to argue that the claim is misleading. Abbott counters by focusing on the fact that, among doctors with a brand preference, Abbott products score higher than competing products. Based on the evidence of consumers’ perceptions of the claim, the court reasons that the “1st Choice of Doctors” claim could legitimately be supported only by evidence that a majority, not a plurality, of doctors express a professional “choice” or “preference” for Abbott’s products over competitors’ products. . To be fair and not deceptive or misleading, the base against which the majority is measured must include doctors who do not have a choice or preference for a single brand. See Gillette Co. v. Norelco Consumer Products Co., 946 F.Supp. 115, 124-125 & n. 7 (D.Mass.1996) (in evaluating claim of superiority based on consumer test, responses favorable to competitor and “no preference” responses must be combined because all are responses contrary to claim of advertiser’s superiority). The parties have presented evidence on numerous surveys of doctors, and the court must examine those surveys in detail. (a) Dr. Ross Survey of Physicians After conducting his survey of consumer perceptions of the “1st Choice of Doctors” claim, Dr. Ivan Ross also conducted a telephone survey of physicians concerning their preferences for infant formula brands. The Ross survey used a “strong” filter question: “First thinking only about two brands of nonspecialized milk-base infant formula with iron, Similac With Iron and Enfamil With Iron, do you or don’t you have a preference between them?” Ex. 2, Tab A at 1. In this survey, 64 percent of the physicians said they had no preference between the two brands. Ex. 2, Tab E at Table 1. Those who said they did have a preference were asked: ‘Which brand do you prefer?” If the answer was inconclusive (“it depends”), the interviewer asked: “If you usually have a preference, which brand do you usually prefer?” Using the total number of participants as the base, including the 64 percent with no brand preference, 25 percent preferred Si-milac, and 10 percent preferred Enfamil. Ex. 2 at 2; Ex. 2, Tab E at Tables 1 & 2. The doctors were then asked: “Now, thinking only about two brands of soy-based infant formula, Isomil and ProSo-Bee — do you or don’t you have a preference between them?” Ex. 2, Tab A at 3. In response to that question, 57 percent said they had no preference, 25 percent preferred Abbott’s Isomil, and 15 percent preferred Mead Johnson’s ProSoBee. Ex. 2, Tab E at Tables 11 & 12. Dr. Ross concluded from his survey that Abbott’s claim to be “1st Choice of Doctors,” as it is understood by consumers, is not supported by the views of doctors. Abbott criticizes Dr. Ross’s physician survey on several grounds. First, Abbott argues that Dr. Ross used too strong a “filter” question by asking first whether the doctor had a preference between the two leading brands, and only then asking which brand the doctor preferred. Abbott contends that this strong filter was intended to increase the “no preference” category to support Mead Johnson’s position. Abbott further contends that the strong filter has the undesirable effect, for market research purposes, of suppressing subtle or slight brand preferences that are nevertheless important. The strong filter clearly had the effect of increasing the “no preference” category. In fact, the “no preference” group was larger in Dr. Ross’s survey than in any of the other surveys introduced into evidence. Both sides agree that some filter must be used so that the respondent is told explicitly that a response of “no preference” or “no opinion” is perfectly acceptable. If Dr. Ross’s survey were the only evidence tending to show that Abbott can win only a plurality among doctors, and not a majority, Abbott’s criticism on this point would be more powerful. The court does not base its decision today on the doubtful suggestion that only a “strong filter” question would be sufficient in a survey used to support the claim. As shown below, however, the great weight of doctor surveys that test doctors’ professional views — not merely their casual brand familiarity— show that Abbott’s Similac cannot win a majority among doctors. In light of all this evidence, the court finds that Dr. Ross’s use of the strong filter did not undermine the validity of his survey here. Abbott’s point about the strong filter suppressing subtle but important brand preferences goes to the heart of a critical issue here. Dr. McDonald testified for Abbott that it is often important for market researchers to probe deeply, to tease out even slight and perhaps irrational and/or unconscious brand preferences. For many purposes, that view is no doubt correct. The issue here is different. The issue here concerns the measurement of doctors’ brand preferences for the purpose of persuading consumers to follow the views of the authoritative experts. Consumers value a doctor’s recommendation so highly because they believe it reflects an expert’s professional evaluation of the product’s quality. Consumers do not value information about doctors’ preferences based on the doctors’ relationships with one brand’s sales representative or based on other factors not related to product quality. The possibility that the strong filter in Dr. Ross’s survey suppressed some slight preferences not related to perceptions of product quality also does not undermine the value of the survey for purposes of this case. Abbott sharply criticizes Dr. Ross for using the word “preference” in his questions instead of the phrase “first choice.” Abbott contends that any test of an advertising claim should use language as close as possible to the language in the claim itself. Dr. Ross explained that he used the word “preference” because of the possibility that, as a result of Abbott’s extensive advertising with the phrase, the phrase “first choice” would trigger some responses that merely “played back” the advertising because of the association of the phrase with Abbott’s product. II Tr. 135-37. Dr. McDonald disagreedj arguing that doctors are not so easily swayed by advertising slogans. Ill Tr. 68. There is some validity to both Dr. Ross’s and Dr. McDonald’s views on this problem, and there is no perfect solution to it. Even if most doctors are not swayed by a mere advertising slogan, Abbott does use the “1st Choice of Doctors” slogan in advertising aimed at doctors, see Ex. 38, II Tr. 261, 269; I Tr. 52, presumably because it •works. Also, in terms of consumer “takeaway,” the differences between “first choice” and “preference” and “recommend” appear to be very slight. For example, in testing consumers’ “takeaway” from a proposed advertisement emphasizing “1st Choice of Doctors,” the results were reported to Abbott in terms that Similac is “the formula doctors prefer” Ex. 22 at A022990 (emphasis added), and that “Doctor recommendations” drove interest in the concept, id. at A022983. The court does not believe the use of the term “preference” in Dr. Ross’s survey undermines its validity as a test, even if it is not the one definitive test, of Abbott’s claim that Simila is “1st Choice of Doctors.” Thus, the court is not persuaded by Abbott’s criticisms of Dr. Ross’s survey of doctors. The survey provides substantial support for Mead Johnson’s contention that Abbott’s ambiguous “1st Choice of Doctors” claim is misleading. (b) The Glickman Surveys To support its claim to be “1st Choice of Doctors,” Abbott relies primarily on surveys conducted by Glickman Research Associates and Burke Marketing Research. Those surveys require close scrutiny. Glickman carried out so-called “check studies” in 1995 and 1996. Interestingly, for each of these studies, Glickman reported: “The objective of this study is to provide justification for [ie., not to “test”] the brand claim with regard to physician first choice recommendation of both Similac ... and Isomil.... ” Ex. 54 at A044972 (1995 study); Ex. 61 at A045016 (1996 study). In these studies, Glickman mailed checks in the amount of $1.00 to thousands of pediatricians. On the back of each check was printed a three-question questionnaire. The pediatricians were instructed to complete the questionnaire when they endorsed the check's. By cashing the checks, the pediatricians returned the questionnaires to Glickman through banking channels. With respect to milk-based formulas, the questionnaires on the back of the checks asked: “Which of the following routine infant formulas, if any, is your ‘first choice ’ for a MILK-based feeding? (Please check only one.)” The available answers in order were: Bonamil Carnation Good Start Enfamil Gerber Baby Formula Similac SMA Other (Please Specify) None, do not have a “first choice” See Ex. 54 at A044978; Ex. 61 at A045025. Glickman reported the following results to Abbott for the milk-based products: 1995 1996 Similac 51% 52% Enfamil 29% 34% SMA 7% 3% Carnation Good Start 1% 1% Bonamil , — — Gerber Baby Formula — — Other ■— 1% None, do not have a “1st Choice” 20% 17% See Ex. 54 at A044974; Ex. 433 at A045019. For soy-based formulas, the questionnaires asked: ‘Which of the following specialty infant formulas, if any, is your first choice ’ for a SOY-based feeding? (Please check only one).” In the 1995 survey, the available choices were Abbott’s Isomil, Nursoy, Gerber Soy, Soyalac, I-Soyalae, Other, and “None, do not have a. ‘1st choice.’ ” The 1995 survey did not include Mead Johnson’s ProSoBee brand, which is the principal competitor for Abbott’s Iso-mil. See Ex. 54 at A044978. The 1996 survey of soy-based formulas included as available choices: Abbott’s Isomil, Mead Johnson’s ProSobee, Nursoy, Carnation Alsoy, Gerber Soy, Other, and “None, do not have a ‘1st Choice.’ ” Ex. 61 at A045025. Glickman reported the following results to Abbott for the soy-based products: 1995 1996 Isomil 58% 56% ProSoBee 18% 29% (write-in) Nursoy 8% 3% Soyalac 1% (N/A) Soyalac 2% (N/A) Carnation Alsoy (N/A) — Gerber Soy — — Other — 1% None 22% 18% Ex. 54 at A044975; Ex. 61 at A045020. Because the 1995 questionnaire did not even include Mead Johnson’s ProSoBee brand as a listed choice, those results have no probative value here. The Glickman studies used exactly the same phrase — “first choice” — that Abbott uses in its advertising and that was being tested for purposes of claims substantiation. Each of the four Glickman surveys reflected responses of between 1734 and 1859 pediatricians. There are several problems with Abbott’s reliance on the Glickman studies to support the “1st Choice of Doctors” claim. First, when a pediatrician marked more than one brand as his or her “first choice,” Glickman counted that response as a “first choice” for each brand marked. That approach is obviously inconsistent with an attempt to identify a single “first choice,” whatever that claim might mean. A doctor who marked more than one brand as “first choice” is essentially saying she does not have a single first choice. The effect of Glickman’s approach was to inflate brand preferences and to reduce the “none” or “no first choice” category. When the results of the Glickman surveys are adjusted to treat the multiple responses as responses of “none,” the results for milk-based feedings were as follows: 1995 1996 Similac • 44% 45% Enfa.mil 22% 27% None or multiple 27% 24% See Ex. 76 at 5; II Tr. 71 (Ross). That is, after a fair adjustment for multiple responses, the majority for Similac dropped down to a mere plurality. The results for the 1996 soy-based survey adjusted in the same appropriate way left Isomil with exactly 50 percent: Isomil 50% ProSoBee 23% None or multiple 23% Ex. 76 at 5; II Tr. 71 (Ross). Even after these adjustments are made, the Gliekman surveys pose problems. For example, Gliekman undertook no effort to determine whether the questionnaires were actually filled out by the pediatricians instead of by other personnel in their offices. See II Tr. 65 (Ross). More fundamental is the disconnect between on one hand the almost casual character of the question asked and the manner of its asking — on the back of a check of nominal value — suggesting the question should be answered off the top of the doctor’s head, and on the other hand the weight that patients/consumers give to the physicians’ expressions of preference. Dr. Ross described the answer to the $1.00 check survey as a “throw away exercise.” II Tr. 239. The court agrees. To illustrate the importance, consider how seriously a doctor would consider a patient’s question: “Doctor, which brand do you recommend I feed my baby?” and how seriously the doctor would treat a question on the back of a $1.00 check that is part of a mass mailing. The difference was underscored by Glickman’s use of the phrase “first choice recommendation” to describe the claim in reporting the survey results, despite the fact that its studies did not attempt to ask about actual professional recommendations. See Ex. 54 at A044972; Ex. 61 at A045016. In contrast, many other surveys discussed below asked about recommendation practices and generated results showing much larger groups of doctors who did not have a preference between these competitors. Those surveys asked questions more closely in tune with consumers’ actual perceptions of doctors’ preferences and the value that consumers place upon those preferences. There are still other problems with the Gliekman surveys as support for the “1st Choice of Doctors” claim. The Gliekman surveys use the phrase “first choice” that has been the subject of extensive advertising and promotion by Abbott. In a debate that echoes Dr. McDonald’s criticism of Dr. Ross’s survey for using the term “preference” instead of “first choice,” Dr. Ross criticized the Gliekman surveys because the use of the identical “first choice” phrase created a risk that the responses would reflect exposure to the phrase in the previous advertising rather than an actual “choice” by the doctor. II Tr. 62-63. Dr. McDonald’s rebuttal, that doctors are not so easily swayed by advertising, has some apparent validity, but it is not consistent with Abbott’s actual commercial behavior in the marketplace. As discussed above, Abbott makes a point of using its “1st Choice of Doctors” slogan in advertising directed at doctors, II Tr. 269, presumably because it expects that phrase to encourage other doctors to go along with the professional crowd. Also, there is no perfect solution to the phrasing problem. Influence on even a small proportion of doctors, however, would have a significant effect on the validity of the Glickman surveys as support for Abbott’s “1st Choice of Doctors” claim. The parties focused much of their attention on whether the Glickman surveys should have used a stronger “filter” question to determine whether a responding physician actually had any preference between formula brands. Dr. Ross used the strong filter question: “First thinking only about two brands of non-specialized milk-based infant formula with iron — Similac With Iron and Enfamil With Iron — do you or don’t you have a preference between them?” Ex. 2, Tab A at 1. That strong filter produced a response of “no preference” from 64 percent of the doctors responding. Ex. 2, Tab E at Table l. The Glickman surveys used instead what Dr. McDonald called a “quasi-filter,” III Tr. 57-58, by including the phrase “if any” in the question asking about “first choice,” and by including a “none” line among the available choices. The result of this weaker filter was a much lower rate of “no preferences.” Abbott and' Dr. McDonald point out that