The opinion of the court was delivered by: LEWIS A. KAPLAN
Plaintiff, a French manufacturer of ladies' footwear, claims that defendants, a Brazilian footwear manufacturer and its U.S. distributor, have violated their rights under the Lanham Act, among other legal doctrines, by imitating the trade dress of plaintiff's so-called P-sole shoes. The matter is before the Court on defendants' motions in limine. Defendants seek to exclude plaintiff's consumer survey, including the testimony of the interviewer and eight interview respondents, on the ground that plaintiff failed to comply with discovery requirements and, insofar as the survey is concerned, on the added ground that the methodology employed in the survey renders it so unreliable that it should not be placed before the jury. They claim also that the survey respondents' testimony should be excluded on the theory that it reflects "manufactured" evidence. Finally, they seek to bar testimony as to any instances of actual confusion from a number of consumer witnesses and people connected with plaintiff and its affiliates.
Surveys frequently are used in Lanham Act and other cases in which consumer perceptions and attitudes are at issue. They usually are conducted by exposing respondents to a stimulus (for example, a product or advertisement) in circumstances different from those in which the stimulus ordinarily is encountered and purchased. The surveyor then uses statistical methods to draw inferences about the probable reaction of the relevant universe of people from the answers of the respondents. See generally Shari Seidman Diamond, Reference Guide on Survey Research, in REFERENCE MANUAL ON SCIENTIFIC EVIDENCE 221, 225-26 (Federal Judicial Center 1994). The quantitative results of such surveys, e.g., the specific proportion of respondents "confused" in the experimental setting, play an important and sometimes dispositive part in determining the outcome of litigation. 4 J. THOMAS MCCARTHY, MCCARTHY ON TRADEMARKS AND UNFAIR COMPETITION § 32.54[c] (1994)(hereinafter cited as MCCARTHY). Properly designed and executed, surveys can provide valuable insight into such central issues as the existence and extent of confusion. See generally id § 32.46. Poorly designed and poorly executed, they can be a positive detriment by causing confusion while shedding little light on the subject at hand. Defendants claim that plaintiff's survey is so clearly deficient that it should be excluded altogether.
The methodology employed here bears little resemblance to the more scientifically conducted surveys often seen in litigation such as this. Plaintiff's counsel designed a questionnaire and sent one of their employees, Deborah Goldberg, a part-time typist, drama student and actress, into Washington Square Park, which is located within blocks of one of plaintiff's retail stores. Over a two day period, she approached a number of people who, she said, looked as if they could afford plaintiff's shoes, which sell at prices considerably higher than defendants'. Ms. Goldberg, who was shod in defendants' shoes, asked the well-to-do passersby whether they could identify the shoes she was wearing. Although she had a questionnaire designed by counsel, she did not necessarily ask the questions as written. All told, she interviewed forty-six people. That she hoped to find people who identified defendants' product as Arche shoes is evident both from the selection of higher income subjects and from the questionnaire, which instructed her to try to enlist persons giving the desired "right" answer as witnesses.
Defendants correctly argue that the methodology used in this survey failed to conform to accepted survey techniques. As their expert, a former director of survey research for General Foods Corporation, points out, the sample of respondents interviewed was not representative of any reproducible group because no objective selection criteria were used. The use of only a single location, while not of itself fatal, is subject to criticism because the location selected was near one of plaintiff's stores and thus may have biased the result in plaintiff's favor. The failure to ask the questions as written makes cumulation of responses suspect. The fact that the interviewer knew the purpose of the survey and, indeed, had an interest in obtaining one response rather than another is far from desirable. Even the choice of the sample universe was skewed by the focus on apparently higher income respondents. See, e.g., Universal City Studios, Inc. v. Nintendo Co., 746 F.2d 112, 118 (2d Cir. 1984); Toys "R" Us, Inc. v. Canarsie Kiddie Shop, Inc., 559 F. Supp. 1189, 1205 (E.D.N.Y. 1983); 4 MCCARTHY § 32.50.
While methodological defects in surveys usually go to the weight rather than the admissibility of the evidence,
there comes a point where the probative value of the survey is exceeded substantially by its prejudicial effect and potential for confusion and waste of time. See FED. R. EV. 403. To whatever extent plaintiff is proposing to use this evidence for the purpose of demonstrating the extent of any confusion, which is not clear at this stage, the evidence crosses that line.
Given the lack of a proper universe and sample, the lack of reproducibility, the poor choice of location, and the potential bias introduced by the choice of interviewer, among other methodological defects, this survey in my view has no significant value in establishing the proportion of consumers, if any, likely to confuse defendants' products with plaintiff's. Accordingly, the cumulative survey results and any opinion evidence based upon them is excluded. Plaintiff will not be permitted to adduce evidence that any given proportion of the respondents approached by Ms. Goldberg identified defendants' shoes as those of the plaintiff.
The Survey Respondents and the Interviewer
The next question is whether Ms. Goldberg and eight of the people whom she interviewed may testify with respect to any actual confusion. I see no sufficient reason why they should not.
Plaintiff proposes to have the eight respondents and Ms. Goldberg testify in substance to their encounters in the park and, presumably, to the respondents having mistaken defendants' shoes for plaintiff's. Defendants argue that such evidence is not probative of actual confusion because the situation in which the respondents encountered Ms. Goldberg did not realistically reflect point-of-sale or post-sale circumstances. Treating these episodes as incidents of actual confusion, they argue, would permit plaintiffs in such cases to "manufacture" incidents of actual confusion. They rely on Judge Tenney's opinion in Reebok International Ltd. v. K-Mart Corp., 849 F. Supp. 252 (S.D.N.Y. 1994), vacated as moot, No. 92 Civ. 8871, 1994 U.S. Dist. LEXIS 19145 (S.D.N.Y. Dec. 28, 1994).
In Reebok, the plaintiff sought to introduce the live testimony of two respondents who had been interviewed in a mall intercept study, as well as written declarations from several others, in an effort to prove actual confusion. The Court excluded the evidence on the grounds that the mall intercept interviews were inherently artificial situations, that receiving such evidence would "allow plaintiffs to manufacture 'incidents' of actual confusion," and that allowing testimony of individual respondents would allow plaintiffs to prevail on the issue of actual confusion even if "the survey results failed to indicate a statistically significant likelihood of confusion . . ." Id at 274.
On the other hand, the Trademark Trial and Appeal Board has come out the other way. In Blue Cross and Blue Shield Association v. Harvard Community Health Plan Inc., 1990 TTAB LEXIS 43, 17 U.S.P.Q.2d 1075 (TTAB 1990), the Board held that statements by survey respondents were evidence of actual confusion, even if the survey methodology was so deficient as to preclude receipt of the survey itself for the purpose of establishing the degree of confusion in the population. Id. at 1078 n.7.
In resolving this question, it is useful to focus on why the issue is important. As defendants have changed their designs, the case about to be tried is for damages. Commentators frequently have noted that "the public policy and theoretical basis underlying monetary awards in trademark cases have received inadequate judicial attention and have remained confused and undefined." See 3 MCCARTHY § 30.24. While a likelihood of confusion typically suffices for injunctive relief, courts frequently are reluctant to award damages without "something more." Id. Evidence of actual confusion usually is the added ingredient that results in a damage award. Id §§ 30.24, 30.25, 30.27[b]. Indeed, in this Circuit, actual confusion is a necessary prerequisite to an award of damages, although its existence may be presumed in cases of intentionally deceptive conduct. See, e.g., George Basch Co. v. Blue Coral, Inc., 968 F.2d 1532, 1537 (2d Cir.), cert. denied, 121 L. Ed. 2d 445, 113 S. Ct. 510 (1992); Resource Developers, Inc. v. Statue of Liberty-Ellis Island Foundation, Inc., 926 F.2d 134, 139 (2d ...