OPINION

ROBERT L. CARTER, District Judge.

Plaintiff, The Coca-Cola Company (“Coca-Cola”), filed this action alleging false description and representation of goods in violation of the Lanham Act, 15 U.S.C. § 1125(a). Coca-Cola accuses defendant, Tropicana Products, Inc. (“Tropicana”), of creating and airing a false and misleading television commercial in support of its “Premium Pack” pasteurized, ready-to-serve chilled orange juice. Claiming irreparable injury to the reputation and good will of its own Minute Maid brand chilled orange juice, plaintiff seeks a preliminary injunction barring further broadcasts of the challenged advertisement. On April 19, 1982, both parties presented witnesses who testified about the message conveyed by Tropicana’s commercial.

The goal of ready-to-serve orange juice makers is to approximate the taste and quality of unprocessed, fresh squeezed juice. Each manufacturer develops a marketing strategy designed to indicate that its product comes closer to the ideal than do competitors. At the same time, however, producers stay within the bounds of the law by mentioning, in some way, the manner in which their juice is processed. For Coca-Cola’s Minute Maid chilled juice, such qualification entails stressing that the product is reconstituted “from concentrate.” Tropicana’s ready-to-serve product, however, is pasteurized but not concentrated, and falls-within a Food & Drug Administration product category distinct from that of any other orange juice product.

*1093 The current Tropicana commercial, entitled “Only One Can Be The Best,” features Olympic athlete Bruce Jenner proclaiming his preference for “Premium Pack.” Frames 5 through 7 of the storyboard show Jenner holding a sealed Tropicana carton, then squeezing an orange and pouring its juice into an opened container of “Premium Pack.” During this sequence, Jenner endorses defendant’s product by stating:

Tropicana Premium Pack. For me, it tastes freshest. It’s pure pasteurized juice as it comes from the orange. It’s the only leading brand not made with a concentrate and water.

Plaintiff asserts that the visual depiction of squeezing fresh fruit and pouring the juice into a Tropicana carton, combined with the incongruous statement that “Premium Pack” is “pure pasteurized juice as it comes from the orange,” misleads consumers into believing that defendant is selling fresh squeezed, unprocessed juice. While Coca-Cola does not challenge Tropicana’s failure to disclose affirmatively the fact that “Premium Pack” undergoes a freezing/thawing process, it characterizes as deceptive the alleged representation that Tropicana sells fresh, unprocessed orange juice.

In this Circuit, a plaintiff seeking a preliminary injunction must establish irreparable harm and either probable success on the merits or a sufficiently serious question going to the merits as to make them a fair ground for litigation, combined with a balance of hardships in plaintiff’s favor. See Dallas Cowboys Cheerleaders, Inc. v. Pussycat Cinema, Ltd., 604 F.2d 200, 206-7 (2d Cir. 1979); Greenwald v. Whalen, 609 F.2d 665, 668 & n.2 (2d Cir. 1979); Harlequin Enterprises Limited v. Gulf & Western Corporation, 503 F.Supp. 647, 648 (S.D.N.Y. 1980) (Owen, J.), aff’d, 644 F.2d 946 (2d Cir. 1981).

Coca-Cola tried to prove the merits of its claim through expert testimony and a consumer survey, and Tropicana countered in kind. Before looking to measures of public reaction to “Only One Can Be The Best,” the court must determine whether the commercial is literally false. See American Home Products Corp. v. Johnson & Johnson, 577 F.2d 160, 164-5 (2d Cir. 1978); American Brands, Inc. v. R. J. Reynolds Tobacco Company, 413 F.Supp. 1352, 1356 (S.D.N.Y.1976) (Lasker, J.). If a merchandising statement is actually false, the court may grant relief without reference to the advertisement’s impact on the buying public. American Brands, Inc., supra at 1356.

A court engaged in this falsity inquiry must view “the ‘entire mosaic’ of the advertisement[ ] rather than ‘each tile separately’.” Vidal Sassoon, Inc. v. Bristol-Myers Company, 661 F.2d 272, 276 (2d Cir. 1981). On the whole, the Jenner commercial is not patently false. While it depicts visually the pouring of fresh squeezed juice into a Tropicana container, the commercial’s audio portion informs viewers that the product is pasteurized. Granted the phrase “pure pasteurized juice as it comes from the orange” is a nonsequitur, its inherent inconsistency, however, does not defeat completely its value as a qualifier of the visual statement. Given the clear enunciation of the word “pasteurized,” the Jenner piece is, at most, ambiguous about whether “Premium Pack” is fresh squeezed and unprocessed. Defendant therefore must survive the first level of Lanham Act review. See American Home Products Corporation v. Abbott Laboratories, 522 F.Supp. 1035, 1039-40 (S.D.N.Y. 1981) (Sofaer, J.); R. J. Reynolds Tobacco Company v. Loew’s Theatres, Inc., 511 F.Supp. 867, 874 (S.D.N.Y.1980) (Sweet, J.).

When challenged advertisements are not literally untrue, their tendency to mislead, confuse or deceive, and thus violate the Lanham Act, “should be tested by the reactions of the public.” American Home Products Corp. v. Johnson & Johnson, supra at 165; see American Brands, Inc., supra at 1357 (“ ‘the public’s reaction to [the] advertisement will be the starting point in any discussion of the likelihood of deception’ ”). To meet this test both parties offered the results of consumer studies and the testimony of experts who interpreted those results. In addition, each party contested the usefulness and appropriateness of the other’s survey. While the Second Circuit has endorsed *1094 the use of market research data to uncover subliminal influences in advertisements, see Vidal Sassoon, Inc., supra at 276-7; American Home Products Corp. v. Johnson & Johnson, supra at 168, it has not done so without reservation. Rather, methodological flaws are relevant to the extent that they impeach the credibility of a survey offered as an illustration of public perceptions. See American Footwear Corporation v. General Footwear Company Limited, 609 F.2d 655, 660 & n.4 (2d Cir. 1979), cert. denied, 445 U.S. 951, 100 S.Ct. 1601, 63 L.Ed.2d 787 (1980); McNeilab, Inc. v. American Home Products Corporation, 501 F.Supp. 517, 525-9 (S.D.N.Y.1980) (Lasker, J.); American Home Products Corporation v. Johnson & Johnson, 436 F.Supp. 785, 792-4 (S.D.N.Y.1977) (Stewart, J.), aff’d, 577 F.2d 160 (2d Cir. 1978).

Plaintiff commissioned ASI Market Research, Inc. (“ASI”) to perform a “copy” or “comprehension” test on the Jenner commercial to determine what it conveyed to television viewers. Interviewers were stationed in three different shopping malls, one each in New York, Baltimore and Miami. They approached women who appeared to be between 18 and 65 years of age and asked if they had purchased orange juice within the past three months and if they regularly used either bottled or car-toned chilled orange juice. Women who responded affirmatively were asked to accompany the interviewer to a private room and watch something on television about which they would be asked a brief series of questions. A total of 500 women participated in the study. Defendant’s expert, Leon Kaplan, conceded that, in general, the techniques described above, known as mall intercept and forced exposure, were standard and proper, and that 500 was a reasonable sample size.

ASI tabulated responses to three questions: What ideas was this commercial trying to get across to you about Tropicana Orange Juice? (“question 3”); What did they show in the commercial about Tropicana Orange Juice? (“question 4”); and, for those who mentioned the sequence in which an orange was squeezed, What did this portion of the commercial suggest to you? (“question 5”). Each question was followed by a series of neutral, non-directive probes such as “what else?” and “anything else?” designed to elicit complete answers. Along with the actual responses collected, ASI compiled statistics representing the number of answers falling within certain categories or “subnets”. The tabular data reveals that 31.20% of the responses to question 3, 22% of those relating to question 4, and 43.60% of the combined responses to questions 3 and 4 fell within the “it’s fresh/fresh squeezed” subnet.

William Weilbacher, plaintiff’s advertising expert, conducted an independent analysis of the verbatims collected by ASI and concluded that 43% of the respondents thought that “Premium Pack” was portrayed as fresh squeezed. In order to identify those interviewees who understood that Tropicana’s juice was fresh squeezed, Weilbacher searched for references to the squeezing of oranges which seemed connected to an articulated concept of freshness. Weilbacher lumped together responses to all three questions in seeking these concept connections. While admitting that his judgments were subjective, Weilbacher stated his belief that any other dispassionate expert would develop figures within one or two percent of his own.

Leon Kaplan criticized the methodology, execution, interpretation and design of the ASI study, and concluded that these various flaws made the survey ambiguous as to what the respondents believed the commercial said about defendant’s product. Among his many methodological critiques, two are particularly persuasive and merit detailed consideration.

First, Kaplan questioned the survey’s overemphasis on the commercial’s visual component. While question 3 was “excellent” insofar as it elicited general impressions about the advertisement, questions 4 and 5 were leading and biased because they asked, generally and then specifically, only about what was shown. One dimensional, visually oriented questions will be literally *1095 interpreted by subjects and thus responses will not illustrate fully and accurately the message conveyed by Tropicana. At the least, similarly one dimensional, aurally oriented questions (e.g. “what did it say?”, “what did you hear?”) should have been included to provide balance and a complete statistical base.

Kaplan’s position is quite convincing. Since falsity must be determined by examining a commercial in its entirety, see Vidal Sassoon, Inc., supra at 276, audio and visual messages are of equal importance and it is their cumulative effect that is determinative. Testimony concerning only one source cannot be credited without comparable evidence as to the other, especially where the two simultaneously communicate different ideas. It is obvious that “Only One Can Be The Best” shows an orange being squeezed and its juice being poured into a “Premium Pack” carton. It even may be that this sequence suggests to viewers that “Premium Pack” cartons in local markets contain fresh squeezed, unprocessed orange juice. Such a false representation violates the Lanham Act, however, only if it is not vitiated by the rest of the commercial, particularly the audio portion corresponding to the challenged video portion. In the absence of evidence of reactions to the “pure pasteurized juice . . . ” disclaimer, the court will not consider statistics generated by questions 4 and 5 probative on the strength of plaintiff’s case. Compare McNeilab, Inc., supra at 521-9 (an ASI questionaire containing one “see and hear" question and one “see only” question relied upon by the court where there was no indication that different messages were transmitted to the two senses).

The non-directive probing utilized by ASI poses another methodological problem, according to Kaplan. Since the issue under consideration was whether consumers would view the Jenner commercial as representing that “Premium Pack” is fresh squeezed, unprocessed juice, follow-up questions designed to clarify potentially ambiguous responses should have been asked. A subject who said that the commercial tried to get across the idea that Tropicana was “fresh” or “fresh squeezed” or “made right from the orange” should have been asked what she meant by those concepts. Only by probing for meaning could the tester obtain an accurate, objective picture of consumer comprehension.

Probing for meaning would have made ASI’s study more reliable for the court’s purposes. Weilbacher admitted that “fresh” is capable of several connotations, among them “not processed,” “not made from concentrate,” “refreshing,” and “100 percent pure.” Only those consumers ascribing to the first of these definitions would have been deceived by Tropicana’s commercial. Due to this ambiguity, Weilbacher, in his independent analysis of the verbatims, based his conclusions on the concept of fresh squeezing rather than the word fresh. A subject’s responses to all three questions revealed to Weilbacher her. state of mind concerning the essential issue. While the expert’s subjective determinations are worthy of some respect, the court would find more persuasive an objective account of the subjects’ reaction to “Only One Can Be The Best.”

Kaplan testified that ASI’s survey suffered from defects in execution as well as methodology. While his commentary made a great deal of sense, much of it was based on speculation about the interviewers’ conduct. To the extent, however, that some verbatims recorded “no”s and “nothing”s at the end of each response while others did not, there is a strong inference that interviewers failed to use even the non-directive prompts or failed to record responses word-for-word. With respect to a general inquiry like question 3, such inadequate field work could lead to the omission of secondary or tertiary concepts such as “pasteurized,” and thereby slant the results.

ASI’s survey must be reinterpreted in light of the above discussion. Insofar as the statistics compiled by both ASI and Weilbacher combine responses to question 3 with those relating to the subsequent one-sided inquiries, they must be rejected. In a separate table, ASI computed the number *1096 of women who answered question 3 with language conforming to any of the phrases within the “it’s fresh/fresh squeezed” sub-net. All 156 such respondents, comprising 31.20% of the subject pool, cannot be categorized as misled consumers because several of the subnet’s components are ambiguous. For example, the 31.20% includes people who responded that the Jenner commercial communicated the idea that “Premium Pack” is fresh or is “fresh juice.” As Weilbacher agreed, such comments could as easily have meant “refreshing” or “100 percent pure” as “unprocessed.” The absence of substantive probing renders these responses useless as indicators of the falsity of Tropicana’s message.

Focusing on the five subnet phrases which arguably convey the concept of fresh squeezed juice (“freshly squeezed,” “juice made right from the orange,” “made from freshly squeezed oranges,” “fresh as it just came from the orange” and “the only one made from freshly squeezed orange juice”) reveals that the number of “deceived” subjects drops to 74. The court’s independent review of the verbatims identified 75 responses, or 15% of the total, which included references, however strained, to fresh squeezed juice.

The court, however, is unwilling to find that Tropicana’s commercial probably tends to mislead as high a percentage as 15% of its viewers. It is by no means certain that responses in the five groups identified above demonstrate that “Only One Can Be The Best” conveys a fresh squeezed, unprocessed message. After performing the type of contextual analysis done by Weilbacher for all three questions, the court is convinced that no more than half of the 75 women might have believed that Tropicana represented its product as not processed. Rather, it seems that “fresh squeezed” and its variants, like the terms describing general freshness, often were used as antonyms for concepts like “from concentrate” and as synonyms for “pure” and “no additives.” Someone who says “they squeeze it from whole oranges, it’s not concentrate” may not be stating that Tropicana comes directly from the orange to her store. To the extent that the commercial says “not from concentrate,” “pure” and maybe even “squeezed from fresh oranges” it is not stating anything false about “Premium Pack.” The survey does not prove that respondents meant anything else by their comments.

A level of consumer confusion significantly below 15% does not indicate plaintiff’s probable success on the merits. An advertisement’s tendency to mislead need not be established by any minimum percentage of confused consumers. The requisite proof is qualitative, not quantitative, and plaintiff need only show “that a not insubstantial number of consumers receive a false or misleading impression from” the commercial. McNeilab, Inc., supra at 528. The degree of confusion caused by the Jenner commercial is insubstantial. Injunctive relief has been granted in cases involving greater than 25% deception, see R. J. Reynolds Tobacco Company v. Loew’s Theatres, Inc., supra at 876 (between 25 and 33 percent); McNeilab, Inc., supra at 526 (39.5 percent); American Home Products Corporation v. Johnson & Johnson, supra at 793 (31 percent), but not, to the court’s knowledge, in any case where no more than 15% deception could be established. Although there is no established measure of the degree of deception required in a Lanham Act litigation, see R. J. Reynolds Tobacco Company v. Loew’s Theatres, Inc., supra at 876, less than 15% cannot be credited in support of a motion for preliminary relief.

On the evidence before me, I cannot conclude that consumers are likely to be misled by defendant’s advertisement. The ASI survey, as noted by Kaplan, fails to respond to the crucial question at hand. See WGBH Educational Foundation, Inc. v. Penthouse International Ltd., 453 F.Supp. 1347, 1351 (S.D.N.Y.1978) (Carter, J.), aff’d, 598 F.2d 610 (2d Cir. 1979) (imprecise language in survey question diffuses the purpose of the testing). For that reason alone, and without comment on defendant’s other criticisms of the ASI survey or the usefulness of *1097 Tropicana’s Burke day-after-recall test, Coca-Cola has failed to demonstrate likely success on the merits.

Thus, the court must consider the second branch of the preliminary injunction standard. See American Home Products Corporation v. Abbott Laboratories, supra at 1045. Plaintiff has raised significantly serious questions going to the merits as to make them a fair ground for litigation. The number of clearly deceived ASI respondents was small but not de minimus. Moreover, the court acknowledges some ambiguity in both the visual squeezing sequence and the audio “pasteurized” disclaimer. A more rigorous survey might substantiate plaintiff’s claims.

Therefore, the court must determine whether the balance of hardships tips decidedly in Coca-Cola’s favor. Coca-Cola’s alleged harm is reduced Minute Maid sales caused by the allegedly deceptive portion of the Jenner commercial. In the absence of reliable evidence concerning either the degree of deception to be expected or the amount of loss that will result, this injury cannot be accorded great weight. Tropicana’s side of the scale includes the time and expense of developing a new commercial or changing the existing one, as well as the market injury incurred in any period that defendant is left without a commercial while plaintiff and other competitors continue advertising. On the present state of the record, these prospective harms are indistinguishable. Therefore, plaintiff has not established the necessary predominance on its side of the balance.

Moreover, to obtain injunctive relief, a Lanham Act plaintiff must show that it is likely to be damaged as a result of its competitor’s allegedly false advertising. Johnson & Johnson v. Carter-Wallace, Inc., 631 F.2d 186, 190 (2d Cir. 1980). The likelihood of lost sales “will not be presumed, but must be demonstrated.” Id. Coca-Cola claims to have met this burden of proof. Inasmuch as Coca-Cola and Tropicana are the two largest competitors in the chilled orange juice market, Coca-Cola suggests that the very fact of intensive Tropicana advertising will benefit “Premium Pack” sales at the expense of the Minute Maid product. Without reference to the record, plaintiff chronicled Minute Maid’s average weekly sales volume during the first four months of 1982 to show the supposed effects of the Jenner commercial. Plaintiff’s Hearing Brief at 35-6.

Under recent Lanham Act decisions in this Circuit, plaintiff’s showing seems insufficient. In Vidal Sassoon, Inc., supra, the Circuit Court was satisfied by a showing that the parties competed in the same market and that defendant’s allegedly false commercial contained “apparently effective suggestions of competitive superiority” which “if repeatedly communicated to consumers, would eventually result in loss of sales to” plaintiff. Id. at 278 (emphasis added); see also American Home Products Corporation v. Abbott Laboratories, supra at 1038. As noted above, there is no convincing evidence that Tropicana is, effectively or otherwise, conveying a false message. In both cited cases, moreover, the proof went beyond the axiom that advertising is designed to enhance sales and/or image at the competitors’ expense. Thus, significant sales increases by the challenged advertiser, see Vidal Sassoon, Inc., supra at 278, or lost market share by the aggrieved competitor, see American Home Products Corporation v. Abbott Laboratories, supra at 1038, accompanied the invocation of new marketing schemes in those cases. See also Johnson & Johnson v. Carter-Wallace, Inc., supra at 191 (plaintiff’s sales declined and consumer evidence indicated that some people thought they could do without plaintiff’s product after seeing defendant’s commercial).

As noted, Coca-Cola’s legal memorandum alleges that Minute Maid’s sales have fallen off in recent months. The figures do not, however, establish “a logical causal connection between the alleged false advertising and” the diminishing sales. Id. at 190. First of all, the sales figures for the two regions involved, New York and the Southeast, are quite dissimilar. In New York, January sales were far above the 1981 aver *1098 age, February showed a decrease of approximately one-third, March brought on a slight improvement and April sales dropped significantly. January sales in the Southeast Region were slightly below the 1981 monthly average, and were followed by an increase in February, a large downturn in March and a rebound in April. “Only One Can Be The Best” first aired in mid-February. Given Minute Maid’s erratic, though generally downward, sales history both before and after that time it is impossible to conclude that the commercial has caused plaintiff’s troubles. The advertisement cannot explain the huge sales loss in New York in February since it aired for only part of the month. In March, when the primary effect would be expected, New York sales increased and Southeastern sales dropped sharply. April brought on the opposite results.

The Minute Maid sales figures are clouded further by the testimony of Harry Robinson, Marketing Manager for Refrigerated Products for the Coca-Cola Foods Division. Robinson stated that market share figures for 1982 have not yet been published, so plaintiff does not know whether it has suffered losses since the onset of defendant’s new advertising campaign. Based on other available evidence, however, Robinson opined that overall industry sales are down relative to the same period last year. These current losses are due to a freeze and resultant high prices. To the extent that Minute Maid’s sales figures reflect these general trends, they are' irrelevant to these proceedings.

Robinson also stated that the most recent market share analysis, covering the period between December 1981 and January 1982, indicated that Minute Maid’s share was slightly below that of the immediately prior 60 day period. This may also imply market trends completely independent of the Jenner commercial. Since Minute Maid’s sales figures are inconclusive and its own representative admitted that extraneous factors may be involved, plaintiff has failed to meet its burden. Despite the loose standards applied in the most recent cases, likely irreparable injury is not established merely by the existence of a competitor’s advertising. Such a standard would, in effect reduce the bifurcated preliminary injunction test to a one-track analysis in all Lanham Act cases. The Second Circuit decisions do not anticipate removing irreparable harm from the preliminary injunction formula. See Johnson & Johnson v. Carter-Wallace, Inc., supra at 189 (“something more than a plaintiff’s mere subjective belief that he is injured or likely to be damaged is required before he will be entitled even to injunctive relief”). In each case discussed above, plaintiff submitted some substantial evidence linking the challenged advertisement to actual competitive injury.

For the reasons discussed above, plaintiff’s motion for a preliminary injunction is denied.

IT IS SO ORDERED.