Summary
finding between 20% and 33% of consumers being deceived sufficient to warrant preliminary injunctive relief
Summary of this case from Sanderson Farms, Inc. v. Tyson Foods, Inc.Opinion
No. 80 Civ. 4197 (RWS).
October 24, 1980.
Cooper, Dunham, Clark, Griffin Moran, New York City, for plaintiff; Gerald W. Griffin, R. Bradlee Boal, Norman H. Zivin, Milton Springut, New York City, of counsel.
Weil, Gotshal Manges, New York City, for defendant; Robert G. Sugarman, Helene D. Jaffe, Jerome C. Pontillo, Surie Rudoff, New York City, of counsel.
OPINION
Plaintiff R.J. Reynolds Tobacco Co. ("Reynolds") filed an action in this court alleging a violation of Section 43(a) of the Lanham Act, 15 U.S.C. § 1125(a), as well as a state law tort committed by defendant Loew's Theatres, Inc., doing business as Lorillard ("Lorillard"). Reynolds' complaint charges that Lorillard has caused to be published, in various print media, comparative advertising for its TRIUMPH brand that is false and deceptive with respect to four Reynolds brands. Testimony was taken and other evidence submitted on August 18-20 at a hearing on plaintiff's motion for a preliminary injunction. The motion projects the court into the midst of hotly contested comparative advertising and the methodology of public opinion studies upon which it is based. For the reasons set forth, Reynolds' motion will be granted in part and denied in part. This opinion constitutes the court's findings of fact and conclusions of law.
That section of the Lanham Act reads, in pertinent part, as follows:
Any person who shall . . . use in connection with any goods . . . any false description or representation, including words or other symbols tending falsely to describe or represent the same, and shall cause such goods . . . to enter into commerce . . . shall be liable to a civil action . . . by any person who believes that he is or is likely to be damaged by the use of such false description or representation.
Plaintiff initially sought leave to participate in the preliminary injunction hearing in a previously filed related action, Philip Morris, Inc. v. Loew's Theatres, Inc., D.C., 511 F. Supp. 855, No. 80-4082, but elected not to participate in the limited fashion then prescribed by the court. Plaintiff thereafter moved to consolidate the actions, but later voluntarily withdrew the motion without prejudice.
This court's preliminary injunction decision in Philip Morris, Inc. v. Loew's Theatres, Inc., supra, (" Philip Morris"), filed July 26, 1980, outlines the recent history of competition among manufacturers in the low tar cigarette market, and of Lorillard's introduction of the TRIUMPH brand. Reynolds, for its part, manufactures the following brands relevant to the litigation: SALEM, currently holding 3.1% of the total cigarette market, with 16 mg. of tar; SALEM LIGHTS, 2.2% market share, 11 mg. of tar; WINSTON LIGHTS, 2.2% market share, 14 mg. of tar; and VANTAGE, 2% market share, 9 mg. of tar. Lorillard's TRIUMPH currently has a .4-.5% share of the market, and contains 3 mg. of tar.
In the industry, cigarettes with 16 mg. of tar or more are called "full flavor." Those with 15 mg. or less are termed "low tar," and with less than 5 mg. — "ultra low tar."
The VANTAGE tar content recently was reduced from 11 mg. to 9 mg.
The expenditure of energy and resources in litigating the issues here and in the Philip Morris case is made understandable in the context of the many millions of dollars regularly spent by the leading tobacco companies on the development, preparation and testing of new brands, and on advertising those brands. Advertising is viewed as critical generally in the tobacco industry, and especially so in the fierce competition to gain acceptance for new brands. For the market consisting of low tar smokers, the surveys, discussed below, and the testimony established two primary considerations which outweigh all others — tar content and taste. The stakes are high. One percent of the total cigarette market is worth approximately $100 million in volume of sales. Reynolds spent a total of $96 million in 1979 in advertising its four brands involved in this litigation. Lorillard has not revealed the dollar amount for the present TRIUMPH campaign, although there was testimony that the company is spending six and one-half times the amount it normally commits to the support of established brands.
This motion concerns the current series of print advertisements favorably comparing TRIUMPH to various competing brands, based on the survey commissioned by Lorillard and conducted by SE Surveys, Inc., (the "Smoker Study") which was considered in the Philip Morris litigation. With those survey results in hand, Lorillard embarked on a campaign advertising TRIUMPH as a "National Taste Test Winner," stating in the ads that TRIUMPH "beats" the other brands tested, and that specific percentages of smokers questioned picked TRIUMPH over the competitors in "overall preference" and in taste. In Philip Morris this court, examining Lorillard's own reported results, enjoined as false or misleading in the Lorillard ads the use of the phrase "National Taste Test Winner," as well as the statistical claims of taste parity or superiority, with respect to Philip Morris's MERIT and MARLBORO LIGHTS brands. In response to that injunction, Lorillard has altered the offending advertisements to the format described below. Apparently because of the lead time necessary to place or make changes in print media ads, however, as of late August "National Taste Test Winner" ads were still appearing in some national magazines. Point of purchase, billboard, and transit ads proclaiming a taste victory are also undergoing appropriate changes although the process still is not complete.
In evidence in Philip Morris was a copy of the report entitled "A National Taste Test: Triumph Non-Menthol vs. Winston Lights, Marlboro Lights, Vantage and Merit Non-Menthols," which readers of the Lorillard ads were invited to write away for. In addition to that document, in the instant case the companion report entitled "A National Taste Test: Triumph Menthol vs. Salem, Kool, Salem Lights and Merit Menthol" was submitted. Both reports are made exhibits to this opinion.
The Lorillard survey upon which the ads were based included tests pitting TRIUMPH against Reynolds' WINSTON LIGHTS and VANTAGE, and TRIUMPH MENTHOL against Reynolds' SALEM and SALEM LIGHTS. TRIUMPH "National Taste Test Winner" ads have appeared with respect to all those brands except SALEM. Although in the Lorillard tests TRIUMPH did register "statistically significant" taste wins over WINSTON LIGHTS and SALEM in addition to "overall preference" wins over all four Reynolds' brands, counsel for Lorillard has represented to the court that, pursuant to the Philip Morris decision, the claim "National Taste Test Winner" as well as the statistical taste preference claims, based on the research at issue, will no longer appear on any of the subject TRIUMPH ads — including those addressed to the Reynolds brands.
Some of the other ads, however, have included reference to TRIUMPH's preference win over SALEM.
There has also been a representation that since the VANTAGE tar content has been reduced from 11 mg. to 9 mg., Lorillard has discontinued all ad comparisons of TRIUMPH to that brand.
Therefore, Lorillard suggests that the court should focus rather on the new TRIUMPH ad, which is based on the same survey but headlined "National Smoker Study Winner." The new ads, which have run already with respect to several brands including WINSTON LIGHTS and SALEM LIGHTS, indeed contain no explicit taste superiority claim, but still feature the phrase "TRIUMPH beats [Winston Lights]," and then "Triumph, at less than [one-fourth] the tar, preferred over [Winston Lights]." Directly underneath, the body copy in the earlier "National Taste Test Winner" ads read:
In rating overall product preference, more [than twice as many] smokers independently chose Triumph over [Winston Lights]. In fact, an amazing [66%] said 3 mg. Triumph tastes as good or better than [14 mg. Winston Lights].
The "National Smoker Study Winner" ads now read:
When tar levels were revealed, more smokers independently chose Triumph over [Winston Lights] in rating overall product preference. Of those expressing a preference, over [69%] preferred 3 mg. Triumph to [14 mg. Winston Lights].
The new ads carry over from the old the invitation to "test for yourself. . . . You'll taste why we named it Triumph," as well as the close-up picture of a young, attractive person eyeing with great satisfaction the cigarette in his hand, over the caption "Taste the UMPH! in Triumph at only 3 mg. tar."
The ads still invite readers to write away for the results of the survey, which has been relabeled "National Smoker Study."
The Smoker Study recorded results from so-called "unidentified paired comparison tests" of TRIUMPH against the other brands. In these tests individuals in twenty-five shopping malls across the country were stopped and screened to determine membership in the desired class of low tar smokers. Those meeting the criteria were invited to smoke a masked TRIUMPH, drink some water to remove the taste, and then smoke one masked competing cigarette. The participant was then told the tar content of the two cigarettes and asked: 1) "[The code representing TRIUMPH] has 3 mg. of tar while [the code representing the competitor] has [x] mg. of tar. Taking this into consideration, which would you prefer to smoke?"; and 2) "Comparing the taste of the two cigarettes, how would you say the taste of the [TRIUMPH] you tried compares to [the competitor]?" While the questions were phrased so that the number representing TRIUMPH was always mentioned first, the order in which individual participants smoked the two cigarettes was rotated.
In fact, as reflected in the complete Smoker Study, four additional substantive questions were asked regarding taste and satisfaction. Two were comparative questions on "amount of taste" and "satisfying taste," in which TRIUMPH scored uniformly fairly poorly. The portion of the Smoker Study made available to the public by Lorillard gave the results only of the first two questions, and did not intimate that any more were asked. Reynolds argues that this is one factor rendering the Lorillard ads deceptive.
It was contended in the Philip Morris hearing that the disclosure of tar content before Question 1 (the "preference question") in the test unfairly biased the answer in favor of lower tar TRIUMPH. However, the court there found that too little evidence had been adduced on that matter, and held only that Lorillard's own test results from Question 2 (the "taste question") did not support the advertising claims of taste parity or superiority vis-a-vis the Philip Morris brands. Id., at 856-857. Here Reynolds mounts a broadside attack upon the test methodology itself, in addition to presenting the same type of claim and evidence which prevailed in Philip Morris addressed to the ads and the message they communicate.
The record constitutes, principally, expert testimony and contains no less than eleven consumer surveys in addition to the challenged Lorillard Smoker Study. Specifically, Reynolds attacks the methodology of the study on the basis of the timing and use of the tar disclosure, the inadequacy of participant exposure to judge the cigarettes, an unrepresentative and biased sample, the probability of interviewer bias, the "order bias" resulting from the consistent mention of the TRIUMPH code first in the questions to the participants, and suppression of important test results.
See note 8.
To demonstrate that in a more thorough, objective test than the Smoker Study TRIUMPH does not rate nearly so well in general terms, Reynolds introduces its own "monadic test" designed to compare VANTAGE and WINSTON LIGHTS with TRIUMPH, and the "EFI paired product test" which Lorillard ran two years ago in the process of developing the TRIUMPH brand. The latter probed smokers' reactions to experimental variants of TRIUMPH non-menthol and menthol compared to other, higher tar brands. In neither the Reynolds monadic test nor the Lorillard EFI test were tar levels revealed, and in both the participants were supplied with several packs of the test cigarettes to smoke and evaluate at home.
As explained by a Reynolds witness, in a monadic test consumers are asked to evaluate a product, in this case a cigarette brand, in and of itself, rating that product on various attributes. Comparative results are obtained simply by placing together the individual evaluations of the various brands.
Reynolds also presented two tests which substantially replicated the Smoker Study, except that in the first Reynolds variant the tar levels were not revealed before the questions were asked (the "no tar test") and in the second the true tar levels were disclosed in reverse (the "reversed tar test"). Reynolds claims on the basis of these tests that the "up front" tar disclosure in the Smoker Study markedly biased the answers to the preference question in favor of lower tar TRIUMPH.
In addition to expert testimony, five "communication" or "consumer perception" tests were submitted on the issue of whether the TRIUMPH ads in fact convey a taste message to the public. The first survey, commissioned by the plaintiff in the Philip Morris litigation and introduced again here by Reynolds, measures consumer reaction to the "National Taste Test Winner" ads. Then there is a group of three perception studies run for Reynolds by one of its advertising agencies, Batten, Barton, Durstine Osborn, Inc. (the "BBDO tests"). These are the "portfolio test" (in which participants were exposed to a representative TRIUMPH ad among several others), the "single magazine ad exposure test," and the "newspaper ad exposure test" — all addressed to the new "National Smoker Study Winner" ads. A total of 750 filter smokers took part in the three BBDO tests. Lorillard submits one consumer perception study conducted by SE Surveys, Inc. (the "SE communication test") involving 1900 smokers, which it claims negates Reynolds' allegation that consumers are being misled.
Finally, results from two studies rating the relative importance to smokers of various cigarette attributes are offered, one by each party, on the question of what constitutes "preference" or "overall preference." Reynolds introduces its "attributes study" and Lorillard offers testimony as to the results of a similar study run for it by Westat, Inc. (the "Westat wants study"). Reynolds also points to the substantial list of cigarette attributes, or smokers' wants, incorporated into the questions asked in the monadic study and the Lorillard EFI test, in its attack on Lorillard's claimed "overall preference" win.
The "wants" or "attributes" evidence establishes that there are indeed many factors a person considers in choosing a cigarette, and thus several elements which may contribute to a smoker's "overall preference" for one brand over another. These include taste and tar level, as well as satisfaction, ease of draw, harshness, type of filter, and more. Also, the taste aspect is broken down into such sub-categories as sweetness, richness of taste, strength of taste, smoothness of taste and aftertaste — among others. Many of the factors overlap considerably, as variations on the same themes. The testimony of the witnesses, as well as the considerable statistical evidence, established that what are generally understood as "taste" and "tar level" are what smokers care about most. These then, are the competitive battlegrounds for Reynolds, Lorillard and the other tobacco giants — as a glance at virtually every cigarette advertisement these days will reveal. Emphasis varies depending on the category of smoker involved or the category of cigarette being advertised (full flavor, low tar, or ultra low tar) but taste and tar level are critical elements. Indeed, Reynolds contends that the claimed "overall preference" win for TRIUMPH is deceptive because the purported preference question probed "preference" for low tar level only, and not for taste.
Lorillard contends that in view of the Philip Morris injunction no further relief is now required. Experts testified that the Smoker Study was designed and conducted in accordance with commonly accepted research procedures, and that any imperfections therein do not so skew the results that it is deceptive to advertise them as it now does. Moreover, Lorillard argues, the "National Smoker Study" ad is truthful and unambiguous on its face. Even if not, it is argued, the consumer perception data before the court (notably, the Westat wants study) do not demonstrate that consumers are receiving a deceptive message.
The outstanding order respecting the taste claims is only a preliminary injunction with respect to the Philip Morris brands. Lorillard is thereby under no legal restraint from continuing or reinstituting ads containing, with respect to the Reynolds brands, the phrases and claims found to be offending in Philip Morris. Moreover, since the Smoker Study indicates a statistically significant "taste win" by TRIUMPH over Reynolds' WINSTON LIGHTS and SALEM, the reasoning of the limited holding in Philip Morris would not necessarily extend to TRIUMPH advertising concerning those brands. It is thus appropriate to consider Reynolds' application to enjoin the "Taste Test Winner" claims.
Reynolds has demonstrated that it would likely suffer irreparable harm from the Lanham Act violations Lorillard is found to have committed in advertising the taste performance of TRIUMPH. See American Home Products Corp. v. Johnson Johnson, 436 F. Supp. 785, 803 (1977), aff'd, 577 F.2d 160 (2d Cir. 1978); see also Philip Morris, supra, at 858. Further, under all the circumstances, as set forth herein, Reynolds' success on the merits of this matter is likely. Jackson Dairy, Inc. v. H.P. Hood Sons, Inc., 596 F.2d 70, 72 (2d Cir. 1979). Therefore Reynolds is entitled to an injunction barring Lorillard from including in its TRIUMPH advertisements with respect to the Reynolds brands tested, the phrase "[National] Taste Test Winner" as well as any other explicit representation as to taste parity or superiority based on the Smoker Study, even though the practice has already been discontinued. See McNeilab, Inc. v. American Home Products Corp., 501 F. Supp. 517 at 523-524 (S.D.N.Y. 1980).
The current Lorillard "National Smoker Study Winner" ad series poses more difficult questions, with regard both to the message communicated in the ads, and the Smoker Study itself. As far as the court has been able to discover, no court has enjoined comparative advertising because of the consumer survey methodology underlying the ad.
In considering a claim of this type for Lanham Act relief the court must initially determine whether the challenged advertising is false on its face. American Home Products Corp. v. Johnson Johnson, 577 F.2d 160, 165 (2d Cir. 1978); McNeilab, Inc. v. American Home Products Corp., supra, at 524-525. The TRIUMPH advertisements in question are not patently false. There was a survey conducted among smokers in twenty-five shopping malls across the nation; TRIUMPH was a "winner" in that it "beat" the competitors in the first question, addressed to preference. There are no explicit claims of taste superiority. There is testimony by market research experts to the effect that Lorillard's mall intercept Smoker Study was poorly conceived in general and flawed in several specific ways, and therefore not of the quality necessary to support a major advertising campaign. "Quick and dirty" was an epithet applied to this type of test. Be that as it may, Lorillard did conduct a test, and the TRIUMPH ads based on that test are not literally false.
This finding, of course, does not end the matter, for § 43(a) of the Lanham Act encompasses more than literal falsehoods. American Home Products Corp. v. Johnson Johnson, 577 F.2d at 165. "Were it otherwise, clever use of innuendo, indirect intimations, and ambiguous suggestions could shield the advertisement from scrutiny precisely when protection against such sophisticated deception is most needed." Id.; See American Brands, Inc. v. R.J. Reynolds Co., 413 F. Supp. 1352, 1356-57 (S.D.N.Y. 1976). The inquiry for the court then becomes whether the advertising has a tendency to mislead, confuse, or deceive. Id. The claims "national smoker study winner," "beats," and "overall preference" may be deceptive either because they do not accurately reflect the test results, because of a problem in the test itself, or both. Furthermore, it is necessary "to consider the advertisement in its entirety and not to engage in disputatious dissection. The entire mosaic should be viewed rather than each tile separately." FTC v. Sterling Drug, Inc., 317 F.2d 669, 674 (2d Cir. 1963); American Home Products v. Johnson Johnson, 577 F.2d at 165 n. 11; Country Tweeds, Inc. v. FTC, 326 F.2d 144, 148 (2d Cir. 1964). Similarly, the Code of Advertising of the Council of Better Business Bureaus states as a "basic principle" that "[a]n advertisement as a whole may be misleading although every sentence separately considered is literally true." Do's and Don'ts in Advertising Copy, section 2, Council of Better Business Bureaus, Inc., 1980.
The Smoker Study results are deceptive. The consistent disclosure of tar content just before the test participants were asked which cigarette they would prefer to smoke had the effect of eliciting responses merely on "tar preference," and not "overall preference" as Lorillard asserts. Question 1 as posed contained no reference to a taste attribute as such. This practice produced statistically significant "overall preference" wins against all four Reynolds brands; the advertisements announce that TRIUMPH is the "National Smoker Study Winner," and that "Triumph beats [the competitor]," and specifically explain in smaller print (with statistics) that TRIUMPH won an overall preference test.
The replication tests comparing TRIUMPH and WINSTON LIGHTS offered by Reynolds demonstrate this result. The Lorillard preference tally as between these two brands was 66%/29% in favor of TRIUMPH. The Reynolds preference tally from the no tar test was 54%/40% in favor of WINSTON LIGHTS. In the reversed tar test, WINSTON LIGHTS were preferred by a 70%/26% margin. The disparity between the Lorillard figures and the Reynolds no tar figures can logically be seen as an effective illustration of the influence of tar disclosure on participants. Reynolds' preference test, together with the figures from the reversed tar test, demonstrate the strong "tar reaction" bias built into Lorillard's Question 1.
It is thus deceptive to advertise TRIUMPH's Question 1 wins as representing smokers' stated "overall" preference which, as already established, is considered to be a preference based on both the dominant elements of choice — taste and tar. Because those questioned were known to have concern related to tar content, the procedure used dictated the test result. See In re Litton Industries, Inc., Dkt. No. 9123, 2 Trade Reg.Rep. (CCH) ¶ 21.710 at 21,884 (June 6, 1980). The test was therefore deceptive and advertising based upon it reflected that quality. Cf. Country Tweeds, Inc. v. FTC, supra, at 147-48; Bristol Myers v. FTC, 185 F.2d 58, 60 (4th Cir. 1950).
This finding, that the test procedures have produced a result which is deceptively claimed to be an "overall preference," is reinforced by the failure of Lorillard to disclose fully all the results of the Smoker Study — in particular, those from the questions specifically directed to taste which indicated TRIUMPH was not the winner. This finding obtains despite the results from Question 2 of the Smoker Study with respect to WINSTON LIGHTS and SALEM. TRIUMPH was reported to be better or somewhat better tasting than WINSTON LIGHTS and SALEM by percentages of 52% to 34% and 48% to 37%, respectively. However, on Questions 3 and 4, which dealt with the amount of taste and its satisfying quality, the percentages favored the Reynolds brands or in one case, TRIUMPH versus WINSTON LIGHTS as to "satisfying taste," reflected a draw. Further, Lorillard's expert failed to establish a basis upon which the answers to Questions 3 and 4 relating to those two brands could be disregarded or considered less controlling than the answers to Question 2. Based upon all the evidence, and in particular the results from Question 3 and 4, the misleading skewing effect of Question 1 with respect to overall preference is not altered for those two Reynolds brands by the answers to Question 2. Overall preference in the lexicon of advertisers and smokers of low tar cigarettes is understood to include taste, and the taste results in the Smoker Study were only partially disclosed. Failure to disclose a material aspect of the results, relating to taste, under the circumstances is misleading.
Further, I find that a mall intercept study as employed in this instance fails to produce a nationally projectable statistical percentage. The suburban character of the sample and the demonstrated effect here of an error in the weighting of the sample does not provide the statistical reliability that would be obtained by random sampling, for example. That is not to say, however, that I find a mall intercept study as such to be an unreasonable basis for comparative advertising. However, because of their "quick and dirty" nature, such studies — and ads based upon them — must receive particularly close scrutiny.
However, the finding that the ad is deceptive standing alone, is not sufficient. Once that finding is made, then a showing of the effect of the ad on the viewers is required. Professionally conducted consumer perception surveys, such as the ones introduced here, are the best such evidence. American Home Products v. Johnson Johnson, 577 F.2d at 165-67, 172; McNeilab, Inc. v. American Home Products Corp., supra, at 524-525; American Brands, Inc. v. R.J. Reynolds Co., supra, at 1356-57.
The BBDO consumer perception studies establish that while the majority of those questioned about the meaning of the phrase "TRIUMPH beats [WINSTON LIGHTS]," in the context of the whole advertisement, believed it related to tar, some 23%-33% (depending on the particular test) believed that the victory referred to taste. This conclusion is borne out by the results from the SE communication test designed, commissioned and introduced by Lorillard. This survey was in fact claimed to be a replication of the Philip Morris perception study. Lorillard emphasizes that in response to the crucial question "[w]hat do you think the statement `Triumph beats Winston Lights' means; in what way does Triumph beat [Winston Lights]?" 66% of the respondents mentioned tar alone, 24% mentioned tar and taste, and 3% mentioned only taste. Thus the SE test results are not substantially at variance with the BBDO results, for the evidence submitted establishes that taste superiority was included in the reaction of over three quarters of the 27% who mentioned taste. Thus somewhere between one-fifth to one-third of those surveyed carry away from the ad the report of a TRIUMPH taste victory.
There is no established quantitative measure of the degree to which an advertisement must mislead or deceive, for a plaintiff to prevail under the Lanham Act. In order for the court to conclude that an ad tends to mislead there must be a "qualitative showing . . . that a not insubstantial number of consumers receive a false [or] misleading impression from it." McNeilab, Inc. v. American Home Products, supra, at 528. On the evidence before the court, Reynolds has established such a "likelihood of confusion or tendency to mislead," American Brands, Inc. v. R.J. Reynolds Tobacco Co., supra, at 1356, and over 25% is not an insubstantial number of consumers.
The evidence also establishes that only a minority of readers coming into normal contact with cigarette ads in the print media have any specific recall of the content or message, that a substantial percentage usually misunderstands the message intended and that all advertising — including cigarette advertising — has a carryover effect. Not only are there the various taste references still present in the ads, but the evidence establishes this carryover effect from the "National Taste Test Winner" ad series in the minds of viewers. There is also evidence of smokers' own generally "conditioned" sense that taste must be the point of reference in this type of ad announcing a "winner" or a "preference." Even though there may be factors such as these extrinsic to the ad which cause it to be understood as having a taste connotation, Lorillard, fully aware of these factors, may not employ its past practices to justify a present deception.
What is offered to the public must be accurate, and then no inquiry into the public mind is required. If, as here, the advertising is found to be deceptive and to have a deceptive effect on the public mind, that effect cannot be swept away by a claim — even if true — that the public mind is frequently in error or is affected by the history of past advertising. It could well be said that it is just that past history upon which Lorillard seeks to capitalize while sailing as close as possible to the buoy marking the bounds of advertising permitted under the Lanham Act.
I conclude therefore that "overall" in terms of preference is a deceptive description of the responses given to Question 1 of the Smoker Study — since those responses reflected substantially "tar preference" and failed fairly to test a taste preference. Furthermore, the current TRIUMPH ads contain a combination of taste references and the phrases "National Smoker Study Winner" (emphasis added) and "TRIUMPH beats [the competitor]" which misleads the public into believing that TRIUMPH's victory was wholly or in fair part a taste victory when in fact it was not. I also conclude that the percentages used, as derived from a mall intercept study, do not constitute national statistics as represented. Thus the ads misrepresent the results from the preference question, and certain of the results themselves are misleading with respect to percentages. I conclude further that a substantial number of consumers is misled by the ads. Therefore the ads violate § 43(a) of the Lanham Act.
On the basis of all the evidence, and the applicable law, see Johnson Johnson v. Carter-Wallace, Inc., 631 F.2d 186, (2d Cir. 1980); Jackson Dairy v. H.P. Hood Sons, Inc., 596 F.2d at 72; Philip Morris, supra, I conclude that Reynolds has demonstrated irreparable injury and the likelihood of success on the merits. Lorillard therefore will be preliminarily enjoined from using in its advertisements based on the Smoker Study the phrase "overall preference" or any substantial equivalent as well as any explicit reference to taste or national percentages.
IT IS SO ORDERED.
APPENDIX C A NATIONAL TASTE TEST TRIUMPH NON-MENTHOL VS. WINSTON LIGHTS, MARLBORO LIGHTS, VANTAGE AND MERIT NON-MENTHOLS
A Report Based on Consumer Research Conducted by SE Surveys, Inc. New York, New York
BACKGROUND TO THE TEST
Lorillard retained SE Surveys, Inc. (a national independent research company) to conduct an unidentified paired comparison test of Triumph Non-Menthol versus Winston Lights, Marlboro Lights, Vantage Non-Menthol and Merit Non-Menthol.
To certify the objectivity of the research, Lorillard retained Dr. Russell I. Haley, an expert in the Marketing Research field and Professor of Marketing at the University of New Hampshire. Dr. Haley assisted in the study design and verified the results of the test in accordance with generally accepted research practice. Dr. Haley's opinion of this research can be found at the end of this report.
The consumer research is described in detail, in the remainder of this booklet.
METHOD
The research method employed was an unidentified paired comparison taste test. Each participant tested Triumph and one of four other Non-Menthol brands — Winston Lights, Marlboro Lights, Vantage or Merit.
Participants were recruited in 25 geographically dispersed shopping centers using a standard intercept approach. The demographic composition was 60% male, 40% female; 60% were 21-39 years of age, 40% were 40 years of age or older. This demographic profile was utilized because it is representative of the National king size non-menthol cigarette market.
The participants were required to have all of the following qualifications:
• be 21 years of age or older • smoke king size filter non-menthol cigarettes • smoke 10 or more cigarettes a day
If all qualifications were met the participants were given one of the two test cigarettes and were then told to smoke as much of the cigarette as they wished. Participants were randomly assigned test brands; the brands other than Triumph being Winston Lights, Marlboro Lights, Vantage and Merit. After the first cigarette had been smoked the participants were asked to drink some water to remove the taste of the first cigarette. The participants were then given the second test cigarette and were told to smoke as much of the cigarette as they wished. After the participants had smoked each of the two cigarettes, they were advised of the tar level of each cigarette they smoked and asked which of the two test cigarettes they preferred to smoke.
The remainder of the interview was then administered.
All cigarettes were masked with white tipping paper so that it was not possible to identify the two test cigarettes.
Throughout the test, the cigarettes smoked were rotated on a "Tried First"/"Tried Second" basis to eliminate position bias.
Field work was conducted between April 9 and April 30, 1980.
Note: The test employed for statistical significance was:
CITIES AND SAMPLE SIZE
The cities and number of non-menthol king size cigarette smokers in each city who participated in these tests were as follows:
CITY TOTAL # OF INTERVIEWS PER CITY
Atlanta, Georgia 51 Buffalo, New York 51 Chicago, Illinois 50 Cincinnati, Ohio 50 Cleveland, Ohio 50
Dallas, Texas 51 Denver, Colorado 51 Glenshaw, Pennsylvania 51 Houston, Texas 47 Indianapolis, Indiana 51
Kansas City, Missouri 50 Los Angeles, California 47 Massapequa, New York 51 Memphis, Tennessee 51 Nashville, Tennessee 50
Natick, Massachusetts 52 New Orleans, Louisiana 49 Oklahoma City, Oklahoma 52 Phoenix, Arizona 50 Portland, Oregon 50
San Francisco, California 49 Southfield, Michigan 52 Tampa, Florida 51 Trumbull, Connecticut 50 Walpole, Massachusetts 50 ------- Total # of interviews 1,257 NUMBER OF INTERVIEWS PER PAIRED COMPARISON TEST
Triumph vs. Winston Lights — 310 Triumph vs. Marlboro Lights — 313 Triumph vs. Vantage — 309 Triumph vs. Merit — 325
RESULTS OF THE BLIND TEST (comparing Triumph and Winston Lights)
Stated below are the results of the blind test comparison of Triumph against Winston Lights. Appearing under the column labeled "Total" is the number of smokers preferring each brand. Appearing under the column labeled "Respondents" is the percentage preferring Triumph and the percentage preferring Winston Lights. In other words, 66% of the respondents preferred Triumph, 29% preferred Winston Lights and 5% had no preference. This difference is statistically significant.
TOTAL RESPONDENTS # %
Triumph Non-MentholVs. Winston Lights PREFER TRIUMPH NON-MENTHOL 205 66 Prefer Winston Lights 89 29 No Preference/No Answer 16 5 ------- ------- Total Respondents 310 100
Appearing below is the number of smokers stating that Triumph is . . . much better tasting, somewhat better tasting, about the same in taste, somewhat worse tasting or much worse tasting than Winston Lights. As in the previous table the second column contains the percentages of respondents expressing these opinions. 66% find the taste of Triumph as good as or better than Winston Lights. This is a significantly larger percentage than the percentage of people stating that Winston Lights taste better. Similarly 52% say Triumph is better tasting while only 34% say Winston Lights are better tasting. This again is significant.
TOTAL RESPONDENTS # %
Triumph Non-MentholVs. Winston Lights Triumph Is . . . Much Better Tasting 78 25 Somewhat Better Tasting 84 27 66 About The Same In Taste 44 14 Somewhat Worse Tasting 78 25 Much Worse Tasting 26 09 34 No Answer — — ------- ------- Total Respondents 310 100
Significant at 95% level of confidence or higher.
RESULTS OF THE BLIND TEST (comparing Triumph and Marlboro Lights)
Stated below are the results of the blind test comparison of Triumph against Marlboro Lights. Appearing under the column labeled "Total" is the number of smokers preferring each brand. Appearing under the column labeled "Respondents" is the percentage preferring Triumph and the percentage preferring Marlboro Lights. In other words 62% of the respondents preferred Triumph, 33% preferred Marlboro Lights and 5% had no preference. This difference is statistically significant.
TOTAL RESPONDENTS # %
Triumph Non-MentholVs. Marlboro Lights PREFER TRIUMPH NON-MENTHOL 194 62 Prefer Marlboro Lights 102 33 No Preference 17 5 ------- ------- Total Respondents 313 100
Appearing below is the number of smokers stating that Triumph is . . . much better tasting, somewhat better tasting, about the same in taste, somewhat worse tasting or much worse tasting than Marlboro Lights or had no opinion. As in the previous table, the second column contains the percentages of respondents expressing these opinions. 61% find the taste of Triumph as good as or better than Marlboro Lights. This is a significantly larger percentage than the percentage of people stating Marlboro Lights taste better. Similarly, although the difference is not statistically significant, 43% say Triumph is better tasting while only 39% say that Marlboro Lights are better tasting.
TOTAL RESPONDENTS # %
Triumph Non-MentholVs. Marlboro Lights Triumph Is . . . Much Better Tasting 61 19 Somewhat Better Tasting 75 24 About The Same In Taste 55 18 Somewhat Worse Tasting 78 25 Much Worse Tasting 43 18 No Answer 1 ------- -------- Total Respondents 313 100
Less than 0.5%.
Significant at 95% level of confidence or higher.
RESULTS OF THE BLIND TEST (comparing Triumph and Vantage Non-Menthol)
Stated below are the results of the blind test comparison of Triumph against Vantage. Appearing under the column labeled "Total" is the number of smokers preferring each brand. Appearing under the column labeled "Respondents" is the percentage preferring Triumph and the percentage preferring Vantage. In other words 60% of the respondents preferred Triumph. 36% preferred Vantage and 4% had no preference. This difference is statistically significant.
TOTAL RESPONDENTS # %
Triumph Non-MentholVs. Vantage Non-Menthol PREFER TRIUMPH NON-MENTHOL 187 60 Prefer Vantage Non-Menthol 111 36 No Preference 11 4 ------- ------- Total Respondents 309 100
Appearing below is the number of smokers stating that Triumph is . . . much better tasting, somewhat better tasting, about the same in taste, somewhat worse tasting or much worse tasting than Vantage. As in the previous table the second column contains the percentages of respondents expressing these opinions. 62% find the taste of Triumph as good as or better than Vantage. This is a significantly larger percentage than the percentage of people stating that Vantage tastes better. Similarly, although the difference is not statistically significant, 45% say Triumph is better tasting while only 38% say that Vantage is better tasting.
TOTAL RESPONDENTS # %
Triumph Non-MentholVs. Vantage Non-Menthol Triumph Is . . . Much Better Tasting 64 21 Somewhat Better Tasting 75 24 About The Same In Taste 53 17 Somewhat Worse Tasting 80 26 Much Worse Tasting 37 12 No Answer — — ------- ------- Total Respondents 309 100
Significant at 95% level of confidence or higher.
RESULTS OF THE BLIND TEST (comparing Triumph and Merit Non-Menthol)
Stated below are the results of the blind test comparison of Triumph against Merit. Appearing under the column labeled "Total" is the number of smokers preferring each brand. Appearing under the column labeled "Respondents" is the percentage preferring Triumph and the percentage preferring Merit. In other words 53% of the respondents preferred Triumph, 42% preferred Merit and 5% had no preference. This difference is statistically significant.
TOTAL RESPONDENTS # %
Triumph Non-MentholVs. Merit Non-Menthol PREFER TRIUMPH NON-MENTHOL 173 53 Prefer Merit Non-Menthol 138 42 No Preference 14 5 ------- ------- Total Respondents 325 100
Appearing below is the number of smokers stating that Triumph is . . . much better tasting, somewhat better tasting, about the same in taste, somewhat worse tasting or much worse tasting than Merit or had no opinion. As in the previous table, the second column contains the percentages expressing these opinions. 60% find the taste of Triumph as good as or better than Merit. This is a significantly larger percentage than the percentage of people stating that Merit tastes better. The difference between those stating Triumph is better tasting and those stating Merit is better tasting is not statistically significant.
TOTAL RESPONDENTS # %
Triumph Non-MentholVs. Merit Non-Menthol Triumph Is . . . Much Better Tasting 45 14 Somewhat Better Tasting 73 22 About The Same In Taste 77 24 Somewhat Worse Tasting 93 29 Much Worse Tasting 36 11 No Answer 1 ------- ------- Total Respondents 325 100
Less than 0.5%.
Significant at 95% level of confidence or higher.
May 22, 1980
Mr. J.R. Ave Executive Vice President — Marketing Lorillard 666 Fifth Avenue New York, New York
Dear Mr. Ave:
I have reviewed the research project entitled "A National Taste Test: Triumph Non-Menthol vs. Winston Lights, Marlboro Lights, Vantage, and Merit Non-Menthols" and conducted on behalf of Lorillard by SE Surveys, Inc. in April 1980. I have examined it from the standpoint of the research design employed, sampling procedures, the questionnaire, and the analysis of the results obtained.
I find the study to have been professionally done, the procedures used to be in accordance with generally accepted research practice, and the results accurately stated in the Lorillard test result booklet.
Sincerely yours,
Russell I. Haley President
RIH/sb
APPENDIX D A NATIONAL TASTE TEST TRIUMPH MENTHOL VS. SALEM, KOOL SALEM LIGHTS AND MERIT MENTHOL
A Report Based on Consumer Research Conducted by SE Surveys, Inc. New York, New York
BACKGROUND TO THE TEST
Lorillard retained SE Surveys, Inc. (a national independent research company) to conduct an unidentified paired comparison test of Triumph Menthol versus Salem, Kool, Salem Lights and Merit Menthol.
To certify the objectivity of the research, Lorillard retained Dr. Russell I. Haley, an expert in the Marketing Research field and Professor of Marketing at the University of New Hampshire. Dr. Haley assisted in the study design and verified the results of the test in accordance with generally accepted research practice. Dr. Haley's opinion of this research can be found at the end of this report.
The consumer research is described in detail, in the remainder of this booklet.
METHOD
The research method employed was an unidentified paired comparison taste test. Each participant tested Triumph and one of four other Menthol brands — Salem, Kool, Salem Lights or Merit.
Participants were recruited in 25 geographically dispersed shopping centers using a standard intercept approach. The demographic composition was 60% male, 40% female; 60% were 21-39 years of age, 40% were 40 years of age or older. This demographic profile was utilized because it is representative of the National king size menthol cigarette market.
The participants were required to have all of the following qualifications:
• be 21 years of age or older • smoke king size filter menthol cigarettes • smoke 10 or more cigarettes a day
If all qualifications were met the participants were given one of the two test cigarettes and were then told to smoke as much of the cigarette as they wished. Participants were randomly assigned test brands; the brands other than Triumph being Salem, Kool, Salem Lights and Merit. After the first cigarette had been smoked the participants were asked to drink some water to remove the taste of the first cigarette. The participants were then given the second test cigarette and were told to smoke as much of the cigarette as they wished. After the participants had smoked each of the two cigarettes, they were advised of the tar level of each cigarette they smoked and asked which of the two test cigarettes they preferred to smoke.
The remainder of the interview was then administered.
All cigarettes were masked with white tipping paper so that it was not possible to identify the two test cigarettes.
Throughout the test, the cigarettes smoked were rotated on a "Tried First"/"Tried Second" basis to eliminate position bias.
Field work was conducted between April 9 and April 30, 1980.
Note: The test employed for statistical significance was:
CITIES AND SAMPLE SIZE
The cities and number of menthol king size cigarette smokers in each city who participated in these tests were as follows:CITY TOTAL # OF INTERVIEWS PER CITY
Atlanta, Georgia 50 Buffalo, New York 51 Chicago, Illinois 50 Cincinnati, Ohio 50 Cleveland, Ohio 50
Dallas, Texas 45 Denver, Colorado 46 Glenshaw, Pennsylvania 50 Houston, Texas 49 Indianapolis, Indiana 50
Kansas City, Missouri 50 Los Angeles, California 48 Massapequa, New York 50 Memphis, Tennessee 50 Nashville, Tennessee 50
Natick, Massachusetts 50 New Orleans, Louisiana 50 Oklahoma City, Oklahoma 50 Phoenix, Arizona 50 Portland, Oregon 49
San Francisco, California 45 Southfield, Michigan 51 Tampa, Florida 51 Trumbull, Connecticut 49 Walpole, Massachusetts 50 ------- Total # of interviews 1,234NUMBER OF INTERVIEWS PER PAIRED COMPARISON TEST
Triumph vs. Salem — 311 Triumph vs. Kool — 305 Triumph vs. Salem Lights — 308 Triumph vs. Merit — 310
RESULTS OF THE BLIND TEST (comparing Triumph and Salem)
Stated below are the results of the blind test comparison of Triumph against Salem. Appearing under the column labeled "Total" is the number of smokers preferring each brand. Appearing under the column labeled "Respondents" is the percentage preferring Triumph and the percentage preferring Salem. In other words, 62% of the respondents preferred Triumph, 33% preferred Salem and 5% had no preference. This difference is statistically significant.
TOTAL RESPONDENTS # %
Triumph Menthol Vs. Salem PREFER TRIUMPH MENTHOL 193 62 Prefer Salem 103 33 No Preference/No Answer 15 5 ------- ------- Total Respondents 311 100
Appearing below is the number of smokers stating that Triumph is . . . much better tasting, somewhat better tasting, about the same in taste, somewhat worse tasting or much worse tasting than Salem or had no opinion. As in the previous table the second column contains the percentages of respondents expressing these opinions. 63% find the taste of Triumph as good as or better than Salem. This is a significantly larger percentage than the percentage of people stating that Salem tastes better. Similarly 48% say Triumph is better tasting while only 37% say Salem is better tasting. This again is significant.
TOTAL RESPONDENTS # % Triumph Menthol Vs. Salem Triumph Is . . . Much Better Tasting 69 22 Somewhat Better Tasting 81 26 63 About the Same in Taste 46 15 Somewhat Worse Tasting 79 26 Much Worse Tasting 35 11 37 No Answer 1 ------- ------- Total Respondents 311 100
Significant at 95% level of confidence or higher.
RESULTS OF THE BLIND TEST (comparing Triumph and Kool)
Stated below are the results of the blind test comparison of Triumph against Kool. Appearing under the column labeled "Total" is the number of smokers preferring each brand. Appearing under the column labeled "Respondents" is the percentage preferring Triumph and the percentage preferring Kool. In other words, 60% of the respondents preferred Triumph, 35% preferred Kool and 5% had no preference. This difference is statistically significant.
TOTAL RESPONDENTS # %
Triumph Menthol Vs. Kool PREFER TRIUMPH MENTHOL 183 60 Prefer Kool 107 35 No Preference 15 5 ------- ------- Total Respondents 305 100
Appearing below is the number of smokers stating that Triumph is . . . much better tasting, somewhat better tasting, about the same in taste, somewhat worse tasting or much worse tasting than Kool. As in the previous table, the second column contains the percentages of respondents expressing these opinions. 63% find the taste of Triumph as good as or better than Kool. This is a significantly larger percentage than the percentage of people stating Kool tastes better. Similarly, although the difference is not statistically significant, 46% say Triumph is better tasting while only 37% say Kool is better tasting.
TOTAL RESPONDENTS # %
Triumph Menthol Vs. Kool Triumph Is . . . Much Better Tasting 66 22 Somewhat Better Tasting 74 24 About the Same in Taste 53 17 Somewhat Worse Tasting 75 25 Much Worse Tasting 37 12 No Answer — — ------- ------- Total Respondents 305 100
Significant at 95% level of confidence or higher.
RESULTS OF THE BLIND TEST (comparing Triumph and Salem Lights)
Stated below are the results of the blind test comparison of Triumph against Salem Lights. Appearing under the column labeled "Total" is the number of smokers preferring each brand. Appearing under the column labeled "Respondents" is the percentage preferring Triumph and the percentage preferring Salem Lights. In other words, 59% of the respondents preferred Triumph, 37% preferred Salem Lights and 4% had no preference. This difference is statistically significant.
TOTAL RESPONDENTS # %
Triumph Menthol Vs. Salem Lights PREFER TRIUMPH MENTHOL 183 59 Prefer Salem Lights 113 37 No Preference 12 4 ------- ------- Total Respondents 308 100
Appearing below is the number of smokers stating that Triumph is . . . much better tasting, somewhat better tasting, about the same in taste, somewhat worse tasting or much worse tasting than Salem Lights or had no opinion. As in the previous table the second column contains the percentages of respondents expressing these opinions. 61% find the taste of Triumph as good as or better than Salem Lights. This is a significantly larger percentage than the percentage of people stating that Salem Lights taste better. Similarly, although the difference is not statistically significant, 44% say Triumph is better tasting while only 39% say that Salem Lights are better tasting.
TOTAL RESPONDENTS # %
Triumph Menthol Vs. Salem Lights Triumph Is . . . Much Better Tasting 68 22 Somewhat Better Tasting 69 22 About the Same in Taste 52 17 Somewhat Worse Tasting 78 26 Much Worse Tasting 41 13 No Answer — — ------- ------- Total Respondents 308 100
Significant at 95% level of confidence or higher.
RESULTS OF THE BLIND TEST (comparing Triumph and Merit Menthol)
Stated below are the results of the blind test comparison of Triumph against Merit. Appearing under the column labeled "Total" is the number of smokers preferring each brand. Appearing under the column labeled "Respondents" is the percentage preferring Triumph and the percentage preferring Merit. In other words, 55% of the respondents preferred Triumph, 40% preferred Merit and 5% had no preference. This difference is statistically significant.
TOTAL RESPONDENTS # %
Triumph Menthol Vs. Merit Menthol PREFER TRIUMPH MENTHOL 170 55 Prefer Merit Menthol 123 40 No Preference 17 5 ------- ------- Total Respondents 310 100
Appearing below is the number of smokers stating that Triumph is . . . much better tasting, somewhat better tasting, about the same in taste, somewhat worse tasting or much worse tasting than Merit. As in the previous table, the second column contains the percentages expressing these opinions. 61% find the taste of Triumph as good as or better than Merit. This is a significantly larger percentage than the percentage of people stating that Merit tastes better. Similarly, although the difference is not statistically significant, 41% say Triumph is better tasting while only 39% say that Merit is better tasting.
TOTAL RESPONDENTS # %
Triumph Menthol Vs. Merit Menthol Triumph Is . . . Much Better Tasting 55 18 Somewhat Better Tasting 73 23 About the Same in Taste 62 20 Somewhat Worse Tasting 84 27 Much Worse Tasting 36 12 No Answer — — ------- ------- Total Respondents 310 100
Significant at 95% level of confidence or higher.
July 7, 1980
Mr. J.R. Ave Executive Vice President — Marketing Lorillard 666 Fifth Avenue New York, New York
Dear Mr. Ave:
I have reviewed the research project entitled "A National Taste Test: Triumph Menthol vs. Salem, Kool, Salem Lights and Merit Menthol" and conducted on behalf of Lorillard by SE Surveys, Inc. in April 1980. I have examined it from the standpoint of the research design employed, sampling procedures, the questionnaire, and the analysis of the results obtained.
I find the study to have been professionally done, the procedures used to be in accordance with generally accepted research practice, and the results accurately stated in the Lorillard test result booklet.
Sincerely yours,
Russell I. Haley President
RIH: mp