BOWNES, Circuit Judge.
This product disparagement case arises from an article evaluating the Bose 901 Series I stereo loudspeaker system by Consumer Reports, the monthly magazine of defendant-appellant, Consumers Union (CU). The district court found, in a bench trial limited to liability, that one statement in the article was false and was published with knowledge that it was false or with reckless disregard of its falsity. Bose Corp. v. Consumers Union of the United States, Inc., 508 F.Supp. 1249 (D.Mass.1981) (Julian, J.). In a subsequent trial, damages were [191]*191assessed in the amount of $115,296.00 plus interest and costs of $95,609.64. Bose Corp. v. Consumers Union of the United States, Inc., 529 F.Supp. 359 (D.Mass.1981) (Caffrey, J.). CU appeals both the finding of liability and the assessment of damages. We reverse on the issue of liability.
The Facts
Dr. Amar Bose founded the Bose Corporation in 1964 and is the corporation’s principal owner and chief executive officer. CU conceded that he is an expert in the field of loudspeaker design.1 Dr. Bose designed the Bose 901 in 1967, and it was first marketed late in that year or early in 1968. The shape and design of these loudspeakers were “unique and unconventional.” Bose, 508 F.Supp. at 1252. Each of the two loudspeaker cabinets in the system had the five-sided shape of a baseball home plate when viewed from above; one side in the front faced the listener and two sides in the rear faced away from the listener. The front face of each cabinet contained one driver2 and each of the two rear faces had four drivers. As a result, the listener received one-ninth of the sound directly from the loudspeakers and eight-ninths of the sound by reflection off the wall behind the speakers.
Consumer Reports provides information about consumer products to mail subscribers and newsstand customers. At the time the article in question was published the magazine enjoyed “a very favorable reputation for independence, integrity, accuracy, and freedom from bias.” Id. at 1252.3 In the article, entitled “Loudspeakers,” CU evaluated the quality and performance of twenty-four loudspeakers based on tests it had conducted. In a section boxed off from the main portion of the article and entitled “Some speakers of special interest,” seven paragraphs of the eight-page article discussed the Bose 901. One other speaker, the Harmon-Kardon HK50, was also evaluated in this section.
The article explained the novel design of the Bose 901 as an attempt to simulate the sound patterns in a live concert hall, where listeners hear most of the sound after reflection off the walls and not directly from the orchestra. After describing the various components of the Bose system, the article next explained a test in which CU engineers compared the Bose 901 to another speaker4 by using a special recording “of a cricket-like noisemaker moving from extreme left to extreme right.” CU played the recording through the speakers in a listening room and “asked a panel to judge the direction from which the sound appeared to come.” CU concluded from this test that, “at least for pinpointing sharp noises, there appeared to be no difference between the speaker systems.”
The next two paragraphs of the article described the results of a second test that CU performed on the Bose 901:
We repeated the experiment using a variety of stereo sounds. When it came to music, the panelists immediately noted a remarkable difference between the systems. The Bose 901 seemed considerably more spacious and reverberant, actually to the point of giving the impression that the wall of the listening room had dropped away. The effect was rather dramatic and was felt from any listening position.
[192]*192But after listening to a number of recordings, it became clear that the panelists could pinpoint the location of various instruments much more easily with a standard speaker than with the Bose system. Worse, individual instruments heard through the Bose system seemed to grow to gigantic proportions and tended to wander about the room. For instance, a violin appeared to be 10 feet wide and a piano stretched from wall to wall. With orchestral music, such effects seemed inconsequential. But we think they might become annoying when listening to soloists. On an impulse, we also played some monophonic records through the Bose. To our surprise, they too acquired the same spacial openness and size distortions as the stereo records.
The article then discussed the sound quality of the Bose 901 and stated that if the speaker had been tested with the main group of speakers “it would have fallen between the high- and medium-accuracy groups. The overall sound was of good quality ... [but] the speakers tended to overemphasize the middle bass, giving it a somewhat overly full, heavy sound.” The article concluded that because the Bose 901 “is so unusual ... a prospective buyer must listen to it and judge it for himself.” It noted that the Bose 901 “requires a gigantic amount of power” and recommended “an amplifier of 50 watts per channel for the deepest base response.”5
The Findings of the District Court
The court determined that under the applicable law of product disparagement,6 it was the plaintiff’s burden to prove that the statements made by CU were of a disparaging or defamatory nature and that the statements were false. After concluding that “the Article, when read as a whole is disparaging,” the court proceeded to analyze in detail each alleged factual error in the article to determine if it was false and if it was disparaging. It found that although the use of the words “panel and panelists” was misleading because there were only two persons doing the testing, it was not disparaging. Bose’s attack on the “sound quality” critique was rejected on the grounds that “it is nothing more than an opinion and, as such, it cannot be proved to be true or false,” citing to Gertz v. Robert Welch, Inc., 418 U.S. 323, 339, 94 S.Ct. 2997, 3006, 41 L.Ed.2d 789 (1974). The court found that Bose had not sustained its burden of proof that the statement as to the power requirements of the system was false.
The district court’s finding of liability was based on part of one sentence in the article which reads: “Worse, individual instruments heard through the Bose system seemed to grow to gigantic proportions and tended to wander about the room.” The court, in its analysis, divided the sentence into two parts. It found that the description “seemed to grow to gigantic proportions” had not been proven false but found the statement that the instruments “tended to wander about the room” was both false and disparaging. The evidence on which this finding was based is summarized as follows. CU’s two employees who conducted the listening test “testified that the wandering sounds they heard were confined to an area within a few feet of the wall near which the Bose 901 loudspeakers were placed.” CU conceded that the words “about the room” might not have described the wandering with precise accuracy, but argued that the wandering of sound was important to consumers, not where it wandered. CU maintained that the statement was substantially true because it accurately described the important observation. The district court referred to testimony that an amount of movement is expected with all stereo speakers and concluded that the location of the sounds’ movement was as important to a consumer as the fact of move[193]*193ment. Thus, the court found the statement not to be substantially true. After rejecting CU’s argument that the statements in the article about “a violin appearing] to be 10 feet wide and a piano stretchpng] from wall to wall” modified the statements about wandering sounds to imply that the wandering occurred along the wall between the speakers, the court stated that the ordinary meaning of “about” the room was “around” the room. The court found the statement to be disparaging: “A statement that attributes such grotesque qualities as instruments wandering about the room to the plaintiff’s product could have no effect other than to harm the reputation of the product.” The use of the word “worse” to introduce the statement in the article showed that CU intended it to have a harmful effect.
Having determined that Bose had proved by a preponderance of the evidence that the statement about individual instruments tending to wander about the room was false and disparaging, the district court proceeded to analyze the impact of the first amendment on the standard of care required of CU. It cited several lower court decisions and discussed the first amendment balance between the need for an uninhibited press and the legitimate state interest in compensating victims of defamation in concluding that the actual malice standard of New York Times v. Sullivan, 376 U.S. 254, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964), applies to product disparagement cases. It then applied the analysis of Bruno & Stillman, Inc. v. Globe Newspaper Co., 633 F.2d 583 (1st Cir.1980), to conclude that “Bose is a public figure, at least with respect to the limited issues of the characteristics and quality of the Bose,” and was required to show by clear and convincing proof that CU’s false statement was published with knowledge that it was false or with reckless disregard of its truth of falsity. CU’s project engineer, Arnold Seligson, conducted the listening test and wrote the words upon which the statements in the article were based. Due in part to his demeanor at trial, the court found that Seligson’s testimony as to what the words “about the room” meant was not credible. Seligson maintained that he perceived that the wandering sounds were confined to an area near the wall behind the loudspeakers. The court found that Seligson was too intelligent to not be aware of the ordinary meaning of “about” and thus concluded that Seligson knew at the time of publication that the article did not accurately describe the effects he had perceived during the test. In the court’s view, this was clear and convincing proof that CU “published a false statement of material fact with the knowledge that it was false or with reckless disregard of its truth or falsity.”
Our Review
As the parties acknowledge, the first amendment permeates' our review. The question of the truth or falsity of the statement that individual instruments tended to wander about the room is intertwined with the question of whether that statement is one of opinion or fact; both questions are difficult to answer. CU argues that if the statement is considered in its full context it becomes clear that the statement is merely the opinion of the panelists who conducted the listening test. This full context includes the tentative language that “instruments ... seemed to grow” and “tended to wander,” that “[w]ith orchestral music, such effects seemed inconsequential [although] we think they might become annoying when listening to soloists,” and finally that “the Bose system is so unusual that a prospective buyer must listen to it and judge for himself.” CU cites numerous cases to support the proposition that a reviewer’s published description of what he or she observed in a public performance, book, or restaurant is protected by the first amendment. Also, the Supreme Court has stated that “[u]nder the First Amendment there is no such thing as a false idea. However pernicious an opinion may seem, we depend for its correction not on the conscience of judges and juries but on the competition of other ideas.” Gertz v. Robert Welch, Inc., 418 U.S. 323, 339-40, 94 S.Ct. 2997, 3006-07, 41 L.Ed.2d 789 (1974) (dictum) (footnote omitted). This statement of the Court im[194]*194plies that an opinion can be neither true nor false as a matter of constitutional law. The proposition that an opinion can be neither true nor false also is reasonable as a matter of common sense.7 The determination of whether a statement is one of opinion or fact, however, is difficult to make and perhaps unreliable as a basis for decision. Cf. W. Prosser, Law of Torts 820 (4th ed. 1971) (discussing the “unsatisfactory and unreliable” opinion limitation of common law and citing Titus, Statement of Fact Versus Statement of Opinion — Dispute in Fair Comment, 15 Vand.L.Rev. 1203 (1962); Note, 62 Harv.L.Rev. 1207 (1949)). Although CU’s argument that the statement is an opinion is plausible, the seeming scientific nature of the article — indicated by quantitative ratings, a description in the beginning of the article of the laboratory testing performed, and the use of such terms as “panelists” and “engineers” to describe the CU employees who performed the tests — would support the position that the statements are factual. See Hotchner v. Castillo-Puche, 551 F.2d 910, 913 (2nd Cir.), cert. denied sub nom. Hotchner v. Doubleday, 434 U.S. 834, 98 S.Ct. 120, 54 L.Ed.2d 95 (1977) (An expression of an opinion may become as damaging as an assertion of fact, and liability for libel thus attach, if the writer indicates that “he has private, firsthand knowledge which substantiates the opinions he expresses .... ”). Similarly, and stemming at least in part from the uncertain nature of the statement as one of fact or opinion, it is difficult to determine with confidence whether it is true or false. As the district court noted, at trial the CU panelists, Seligson and Lefkow, “testified that the wandering sounds that they heard were confined to an area within a few feet of the wall near which the Bose 901 loudspeakers were placed.” Given the subjective nature of a listener’s perceptions and the imprecise language employed in the CU article, we are not sure that the statement that instruments tended to wander about the room is false.
Due to our ultimate conclusion that Bose has failed to meet its burden of proof with respect to actual malice, however, we will assume that the statement was both factual and false and not explore further the intricacies of these concepts.8 Before analyzing actual malice, we should first briefly note our agreement with the district court that the statement in question disparaged Bose 901 speakers. CU concedes that the statement told readers that the “movement of individual instruments is more pronounced and localization therefore more difficult with the Bose than with conventional speakers.” The tenor of the paragraph that contained the statement was clearly that of criticism; this is highlighted by the use of the word “worse” to introduce the statement. There can be little question that the statement was disparaging.
At oral argument Bose acknowledged that it does not dispute the finding of the district court that the corporation is a public figure with respect to the subject matter of the CU article. Bose also conceded that the rule of New York Times v. Sullivan applies in this case, and thus accepted the district court’s conclusion that the actual malice standard applies to product disparagement cases. As we have indicated earlier, the district court analyzed both of these issues at length and we accept its conclusions for the purposes of this case. Cf. Long v. Arcell, 618 F.2d 1145, 1147 & n. 2 [195]*195(5th Cir.1980) (plaintiffs bound by stipulation at trial that they were public figures when the article was written).
We focus, therefore, on the district court’s holding that Bose proved by clear and convincing evidence that CU published the words “individual instruments ... tended to wander about the room” with knowledge that they were false or with reckless disregard of their truth or falsity. In performing this review we are not limited to the clearly erroneous standard of Fed. R.Civ.P. 52(a); instead, we must perform a de novo review, independently examining the record to ensure that the district court has applied properly the governing constitutional law and that the plaintiff has indeed satisfied its burden of proof. See, e.g., New York Times v. Sullivan, 376 U.S. at 285 & n. 26, 84 S.Ct. at 729 & n. 26; Time, Inc. v. Pape, 401 U.S. 279, 284, 91 S.Ct. 633, 636, 28 L.Ed.2d 45 (1971); Long v. Arcell, 618 F.2d at 1147; Hochner v. Castillo-Puche, 551 F.2d at 913; R. Sack, supra note 7, at 560 (citing New York Times v. Sullivan, 376 U.S. at 285, 84 S.Ct. at 728 and Rosenbloom v. Metromedia, Inc., 403 U.S. 29, 54, 91 S.Ct. 1811, 1825, 29 L.Ed.2d 296 (1971) (plurality opinion)); Oakes, Proof of Actual Malice in Defamation Actions: An Unresolved Dilemma, 7 Hofstra L.Rev. 655, 663, 701, 707-09 (1979); Restatement (Second) of Torts § 580A comment g (1977). At the same time, we recognize that we are in no position to consider the credibility of witnesses and must leave questions of demeanor to the trier of fact. See Long v. Arcell, 618 F.2d at 1147; Oakes, supra, at 711.
In New York Times v. Sullivan, the Supreme Court held that in order to prevail in a defamation action a public official must prove “that the statement was made with ‘actual malice’ — that is, with knowledge that it was false or with reckless disregard of whether it was false or not.” 376 U.S. at 279-80, 84 S.Ct. at 725-26. In Curtis Publishing Co. v. Butts, 388 U.S. 130, 87 S.Ct 1975, 18 L.Ed.2d 1094 (1967), the Court extended this requirement of proof to public figures. The Court has formulated the actual malice standard as requiring a showing that the statement was made with a “high degree of awareness of ... probable falsity,” Garrison v. Louisiana, 379 U.S. 64, 74, 85 S.Ct. 209, 215, 13 L.Ed. 125 (1964), and in another case as requiring “sufficient evidence to permit the conclusion that the defendant in fact entertained serious doubts as to the truth of his publication,” St. Amant v. Thompson, 390 U.S. 727, 731, 88 S.Ct. 1323, 1325, 20 L.Ed.2d 262 (1968).
A plaintiff required to prove actual malice under the New York Times v. Sullivan standard must do so with “convincing clarity.” 376 U.S. at 285-86, 84 S.Ct. at 728-29. This requirement, now the “clear and convincing proof” test, Gertz v. Robert Welch, Inc., 323 U.S. at 342, 94 S.Ct. at 3008; see Bruno & Stillman, 633 F.2d at 586 n. 2, calls for the plaintiff to prove more than would be necessary under the preponderance of the evidence standard but something less than what the beyond a reasonable doubt standard requires. Yiamouyiannis v. Consumers Union of the United States, Inc., 619 F.2d 932, 940 (2d Cir.1980); R. Sack, supra note 7, at 225. Although these distinctions are fine, the standard of proof adopted reflects the value that society places on the underlying right involved. Cf. Addington v. Texas, 441 U.S. 418, 425, 99 S.Ct. 1804, 1808, 60 L.Ed.2d 323 (1979) (discussing the clear and convincing proof standard in the context of the due process clause). The right of free speech and the concern about “the practical impact upon truthful speech” — which is “the whole point of the New York Times case,” Oakes, supra, at 704 (quoting T. Emerson, The System of Federal Expression 536 (1970) — obviously are valued highly. We have noted in an earlier case “the almost decisive amplitude of ‘breathing space’ surrounding defamatory falsehood, once a plaintiff is obliged to meet the New York Times standard,” as well as one commentator’s conclusion that the rule in that case “ ‘in practical effect became a near-immunity from defamation judgments.’ ” Bruno & Stillman, 633 F.2d at 686 (quoting Eaton, The American Law of Defamation Through Gertz v. Robert Welch, Inc. and Beyond: An Analytical Primer, 61 Va.L.Rev. 1349, 1373 (1975)).
[196]*196The subjective determination of whether CU in fact entertained serious doubts as to the truth of the statement may be proved by inference, as it would be rare for a defendant to admit such doubts. See Stone v. Essex County Newspapers, Inc., 367 Mass. 849, 868, 330 N.E.2d 161, 173 (1975). A court typically will infer actual malice from objective facts. Oakes, supra, at 666-67 (quoting Washington Post Co. v. Keogh, 365 F.2d 965 (D.C.Cir.1965), cert. denied, 385 U.S. 1011, 87 S.Ct. 708, 17 L.Ed.2d 548 (1967)). These facts should provide evidence of negligence, motive, and intent such that an accumulation of the evidence and appropriate inferences supports the existence of actual malice. Such evidence is lacking in this ease.
It is helpful to compare the research and editing procedures followed by CU in publishing its article on loudspeakers with publishers’ procedures that have been examined in other cases. One court, in finding that the plaintiff — a public figure — had failed to show actual malice, reviewed CU’s work in publishing a series of articles in 1978 attacking the claims of certain organizations and individuals that fluoridization causes, among other things, cancer and birth defects. Yiamouyiannis v. Consumers Union of the United States, Inc., 619 F.2d 932. It stated:
It is clear that appellee, through its agents, made a thorough investigation of the facts. Scientific writings and authorities in the field were consulted; authoritative scientific bodies speaking for substantial segments of the medical and scientific community were investigated. The unquestioned methodology of the preparation of the article exemplifies the very highest order of responsible journalism: the entire article was checked and rechecked across a spectrum of knowledge and, where necessary, changes were made in the interests of accuracy.
Id. at 940. Although we would refrain from describing CU’s loudspeaker article as exemplifying the very highest order of responsible journalism, CU does not have to meet such high standards to prevail. In addition, these two CU projects are distinguishable. In the fluoridization article there existed an abundance of scientific research and writing which CU merely cited in drawing its conclusions. In the instant case, CU had the much more difficult task of performing the original research.
It is important to point out that in conducting the listening test CU used experts, Seligson and Lefkow,9 who brought their expertise and experience to bear in evaluating the Bose speakers. In Reliance Insurance Co. v. Barron’s, 442 F.Supp. 1341 (S.D. N.Y.1977), the court noted the expertise of the author and granted a motion for summary judgment because there was no triable issue of fact that the article was published with actual malice. Id. at 1350-51. The Second Circuit found lack of expertise to be evidence of malice in Goldwater v. Ginzburg, 414 F.2d 324 (2d Cir.1969). The court affirmed a jury finding of actual malice in an article stating that Senator Goldwater suffered from the mental disease of paranoia. The court noted that this conclusion “was reached only upon his [the author’s] own non-expert evaluation of Senator Goldwater’s life and political career.” Id. at 331.
Some evidence of actual malice may be found “if there is a complete departure from the standards of investigation and reporting ordinarily adhered to by responsible [197]*197publishers.” Reliance Insurance, F.Supp. at 1341. CU’s editorial procedures reveal no evidence of actual malice. As in Reliance Insurance, the testimony in this case indicated that normal editorial procedures were followed; there was no evidence of CU knowingly departing from these procedures in order to publish the article regardless of its truth or falsity. After testing the loudspeakers, Seligson prepared a rough draft of the manuscript, commonly referred to as a “report to editorial,” which was reviewed by an associate technical director. The Editorial Department then reviewed this report and drafted the manuscript for publication. Among other editorial alterations, the Department changed Seligson’s words that instruments “suffered [from] a tendency to wander around the room” to the statement ultimately published that instruments “tended to wander about the room.” This manuscript was sent back to Seligson for “line by line checking” and then forwarded to the associate technical director for his review. It was then returned to the Editorial Department. These same procedures were applied to galley proofs, second galley proofs, page proofs, and second page proofs. The associate technical director testified that when he performed his reviews the words “tended to wander about the room” conjured up “the mental image ... of the sound moving about in front of the listener.” He also testified that when he approved the article for publication he never really pondered the meaning of the word “about” in the statement. The most we can conclude from this is that in reviewing the manuscript CU employees could have inquired more painstakingly into the precise language being used. 442
Even though we accord relatively little weight to CU’s claims of good faith and lack of any motivation to disparage the Bose 901, we are unable to find clear and convincing evidence that CU published the statement that individual instruments tended to wander about the room with knowledge that it was false or with reckless disregard of whether it was false or not. The evidence presented merely shows that the words in the article may not have described precisely what the two panelists heard during the listening test. CU was guilty of using imprecise language in the article— perhaps resulting from an attempt to produce a readable article for its mass audience. Certainly this does not support an inference of actual malice. See Wolston v. Reader’s Digest Ass’n, 578 F.2d 427, 434 (D.C.Cir.1978); cf. Lambert v. Providence Journal Co., 508 F.2d 656, 659 (1st Cir. 1975) (discussing “the court’s reluctance to entertain libel suits dependent upon a precise construction of a newspaper’s use of technical legal terminology”); R. Sack, supra, at 175 (“ ‘mere exaggeration, slight irony or wit, or all of those delightful touches of style that go to make an article readable, do not push beyond the limitations of fair comment’ ” (quoting Briarcliff Lodge Hotel, Inc. v. Citizen-Sentinel Publishers, Inc., 260 N.Y. 106, 118, 183 N.E. 193, 198 (1932)). To find actual malice in this case would be to interpret that concept to require little more than proof of falsity, an interpretation that Justice Goldberg expressed fears about in his concurrence in New York Times v. Sullivan, 376 U.S. at 298 n. 2, 84 S.Ct. at 736 n. 2 (Goldberg, J., concurring).
Due to our holding on the issue of liability, there is no need for us to review the district court’s findings on damages.
Reversed.