Metabolife International, Inc. v. Wornick

Opinion by Judge HAWKINS; Partial Concurrence and Partial Dissent by Judge RYMER MICHAEL DALY HAWKINS, Circuit Judge: The heart of this case lies at the difficult three-way intersection of the news media’s desire to inform the public about the potential dangers of an over-the-counter herbal supplement, California’s public policy interest in the prompt resolution of so-called “SLAPP suits,” and the liberal policies underlying the discovery provisions of the Federal Rules of Civil Procedure. Metabolite appeals the dismissal with prejudice of state law claims against (1) an investigative reporter for local Boston television station WCVB-TV, (2) the station itself, (3) the station’s parent corporation, and (4) a Harvard Medical School professor, Dr. George Blackburn. The complaint arises from a three-part “investigative report” that aired on WCVB in May 1999 detailing dangers allegedly associated with the use of Metabolite's main product, the herbal weight loss and energy supplement “Metabolite 356.” Metabolite sought relief in the district court under California law. [837] FACTS AND PROCEDURAL HISTORY The underlying facts are not in dispute. In May 1999, a local Boston television station (WCVB-TV) aired a three-part series of “investigative reports” prepared by its reporter Susan Wornick (“Wornick”).1 These reports challenged the safety of Me-tabolife 3562 as well as the public policy influence of Metabolife founder Joseph Ellis, who a decade earlier sustained a felony conviction based on methamphetamine manufacturing. When the station would not grant a retraction, Metabolife filed suit in district court based on diversity jurisdiction.3 Me-tabolife asserted claims under California state law for: (1) defamation; (2) slander; (3) trade libel; and (4) negligent and intentional interference with prospective economic advantage. Metabolife challenged eight discrete statements from the broadcast before the district court, only four of which are at issue on appeal: 1. A statement by Harvard Medical School professor Dr. George Blackburn, an obesity specialist, that “You can die from taking this product [Metabolife 356].”4 2. A statement by Wornick that, “Every expert we asked said Metabolife [356] is not safe because of its main ingredient, ma huang.” 3. A statement by Wornick allegedly implying that Metabolife 356 had not been tested for safety.5 4. Statements by Wornick that Meta-bolife and methamphetamine share the same main ingredient, ephedrine.6 The defendants — Wornick, Dr. Blackburn, the station, and the station’s parent corporation — filed motions to strike Metabolife’s complaint pursuant to California’s “anti-SLAPP” statute, Cal.Civ.Proc. Code § 425.16.7 Defendants refused to en [838] gage in discovery pending the outcome of their motions pursuant to Cal.Civ.Proe. Code § 425.16(g). Metabolife responded by moving to compel full responses to its written discovery requests. The district court temporarily stayed discovery and asked Metabolife to itemize the discovery it needed to respond to the anti-SLAPP motions, which Meta-bolife did. The district court then reversed its field, and ordered Metabolife to respond to the anti-SLAPP motions without discovery, itemized or otherwise.8 Despite the lack of discovery, Metabol-ife’s opposition to the anti-SLAPP motions included over twenty affidavits and more than 750 pages of exhibits, including the opinions of six experts on issues relating to Metabolife 356’s safety. After receiving these materials and in preparation for its next hearing, scheduled to decide venue and perhaps the anti-SLAPP issues, the district court directed the parties to be prepared to address twenty-one questions at the hearing, some of which went to the reliability of the scientific evidence presented by Metabolife in its opposition to the anti-SLAPP motions. The district court held its motions hearing, focusing solely on the anti-SLAPP motions. After the hearing, the court ordered limited discovery on two issues: (1) Wornick’s and WCVB’s editing of Dr. Blackburn’s interview and (2) what experts Wornick had spoke'n with to back up her statement that “Every expert we spoke to said Metabolife [356] is not safe because of its main ingredient, ma huang.” However, just six days later the court rescinded this order, halting all discovery under the anti-SLAPP statute. Metabolife filed an objection, and the court responded by ordering briefing on five final questions. After receiving this post-hearing material, the district court issued its decision, granting the defendants’ anti-SLAPP motions. Metabolife Int’l Inc. v. Wornick, 72 F.Supp.2d 1160 (S.D.Cal.1999). On the statement, “You can die from taking this product,” the district court held that it did not matter whether the statement was construed literally or, as Metabolife argued it should be, as “You can die from taking this product as directed.”9 Id. at 1167, n. 4. The court held that the defendants prevailed either way because “Metabolife has not provided any admissible prima facie evidence of falsity.” Id. The district court arrived at this ruling because “Metabolife’s scientific evidence [allegedly proving safety if taken as directed] is inadmissible under Dauberb because it lacks sufficient indicia of reliability.” 72 F.Supp.2d at 1168. The district court also held, alternatively, that Dr. Blackburn’s speech was protected by the First Amendment as a rational interpretation “of the ambiguous and unresolved state of scientific knowledge regarding the safety of products like Metabolife.” Id. at 1166-67. The district court dismissed the claim based on the statement that “Every expert we asked said Metabolife is not safe because of its main ingredient, ma huang,” on the same Daubert concerns. Id. at [839] 1172-73. The court also held that the statement “does not imply a ‘consensus’ in the scientific community,” and thus could not support the defamatory implication asserted by Metabolife. Id. at 1173. Wornick’s statement allegedly implying that Metabolife 356 had not been tested for safety was dismissed because her comments regarding the Vanderbilt University study were “literally true.” Id. at 1174. The district court also held that “[e]ven assuming ... that Wornick’s literally true statements about the Vanderbilt Study support the alleged defamatory implications [that Metabolife 356 was not tested for safety], Metabolife cannot prove that those defamatory implications are false” because Metabolife had put forward no admissible evidence of affirmative safety studies. Id. at 1174-75. Finally, on the statements whose alleged implication was that, through the common presence of ephedrine, Metabolife 356 and methamphetamine share the same main ingredient, the district court ruled that the statements were substantially true. Id. at 1176. The court disregarded Metabolite’s expert testimony that ephedrine and ma huang are not identical because “that Me-tabolife requires expert scientific opinion to describe the limited difference between ma huang and ephedrine10 convinces the Court that such fine distinctions would have no effect on the state of minds [sic] of the audience.... ” Id. Metabolife appealed the district court’s decision on these three issues and “all interlocutory orders that gave rise to that judgment.” We have jurisdiction under 28 U.S.C. § 1291. STANDARDS OF REVIEW The admissibility of scientific evidence under Federal Rule of Evidence 702 is reviewed for abuse of discretion. Kennedy v. Collagen Corp., 161 F.3d 1226, 1227 (9th Cir.1998). The district court’s decision not to permit additional discovery pursuant to Federal Rule of Civil Procedure 56(f) is reviewed for abuse of discretion. DeGrassi v. City of Glendora, 207 F.3d 636, 641 (9th Cir.2000). The district court’s conclusions of law are reviewed de novo. Cigna Prop. and Cas. Ins. Co. v. Polaris Pictures Corp., 159 F.3d 412, 418 (9th Cir.1998). ANALYSIS This case presents three discrete, though related, issues, the district court’s: (1) exclusion of Metabolife’s scientific evidence; (2) decision under the California anti-SLAPP statute not to allow Metabol-ife discovery; and (3) conclusion that the challenged statements are alternatively protected by the first amendment. Each issue will be dealt with individually; the analysis begins with a description of the state statute under which this case was dismissed. I. California’s Anti-SLAPP Statute The anti-SLAPP statute was enacted to allow early dismissal of meritless first amendment cases aimed at chilling expression through costly, time-consuming litigation.11 Under the statute, a civil de [840] fendant may move to strike a cause of action based on an “act in furtherance of [the] right to petition or free speech.” Cal.Civ.Proc.Code § 425.16(b). An “act in furtherance” includes “any ... oral statement ... made in a ... public forum in connection with an issue of public interest.” § 425.16(e). Metabolite concedes that “the safety of products intended for human consumption is a matter of public concern,” and agrees that the statements challenged were made in a public forum. Thus, Metabolite concedes that the anti-SLAPP statute’s first step is satisfied in this case. Once it is determined that an act in furtherance of protected expression is being challenged, the plaintiff must show a “reasonable probability” of prevailing in its claims for those claims to survive dismissal. § 425.16(b); Wilcox v. Superior Court, 27 Cal.App.4th 809, 33 Cal.Rptr.2d 446, 455 (1994). To do this, the plaintiff must demonstrate that “the complaint is legally sufficient and supported by a prima facie showing of facts to sustain a favorable judgment if the evidence submitted by the plaintiff is credited.” Wilcox, 33 Cal.Rptr.2d at 454. This burden is “much like that used in determining a motion for non-suit or directed verdict,” which mandates dismissal when “no reasonable jury” could find for the plaintiff. Id. at 455 (citing Rowe v. Superior Court, 15 Cal.App.4th 1711, 19 Cal.Rptr.2d 625, 632 (1993)). Thus, a defendant’s anti-SLAPP motion should be granted when a plaintiff presents an insufficient legal basis for the claims or “when no evidence of sufficient substantiality exists to support a judgment for the plaintiff.” Id. at 457 (citing Carson v. Facilities Dev., Co., 36 Cal.3d 830, 838-39, 206 Cal.Rptr. 136, 686 P.2d 656 (1984)). Because the defendants’ speech addressed a matter of “public concern,” Metabolite must show that the statements were false and made with “actual malice.” Milkovich v. Lorain Journal Co., 497 U.S. 1, 14, 110 S.Ct. 2695, 111 L.Ed.2d 1 (1990). As a consequence of staying all discovery, the district court held that it would “not weigh Metabolife’s evidence to determine whether it has established a prima facie case of actual malice. Rather, the [court’s analysis] address[ed] the legal defenses of [the] Defendants and whether Metabolite ... established a prima facie case of falsity.” 72 F.Supp.2d at 1166. As noted above, the district court held that Metabolite could not establish its prima facie case as to the falsity of the three statements it challenges on appeal. Id. at 1166-76. II. Exclusion of Metabolife’s Scientific Evidence Metabolite could not meet its burden on falsity below after the district court excluded all of its scientific evidence regarding the safety of Metabolite 356 when used as directed. Under the anti-SLAPP statute, a plaintiff must meet its burden of proving prima facie falsity with admissible evidence. Wilcox, 33 Cal.Rptr.2d at 459; Evans v. Unkow, 38 Cal.App.4th 1490, 45 Cal.Rptr.2d 624, 628 (1995). The district court held that Metabolife’s scientific evidence was not admissible under Daubert v. Merrell Dow Pharms., Inc., 43 F.3d 1311 (9th Cir.1995) (“Daubert II”). 72 F.Supp.2d at 1167-70. A. The Daubert Standard Scientific evidence is admitted pursuant to Federal Rule of Evidence 702. In Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 587-89, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993) (Daubert I), the Supreme Court held that Rule 702 displaced the prior “general acceptance” test. Under Daubert II, the district court acts as a “gatekeeper,” excluding “bad science” that does not carry sufficient indicia of reliabili [841] ty for admission under Rule 702. 43 F.3d at 1316. This is accomplished through a preliminary determination that the proffered evidence is both relevant and reliable.12 Daubert I, 509 U.S. at 589-92, 113 S.Ct. 2786. Scientific evidence is reliable if it is based on an assertion that is grounded in methods of science — the focus is on principles and methodology, not conclusions. Id. at 595-96, 113 S.Ct. 2786. The Supreme Court listed four non-exclusive factors for consideration in the reliability analysis: (1) whether the scientific theory or technique can be (and has been) tested; (2) whether the theory or technique has been subjected to peer review and publication; (3) whether a particular technique has a known potential rate of error; and '(4) whether the theory or technique is generally accepted in the relevant scientific community. Id. at 593-94, 113 S.Ct. 2786. In Daubert II we noted that a “very significant fact to be considered is whether the experts are proposing to testify about matters growing naturally and directly out of research they have conducted independent of the litigation, or whether they have developed their opinions expressly for purposes of testifying.” 43 F.3d at 1317. If the evidence is not based upon independent research, the district court must determine whether there exists any “other objective, verifiable evidence that the testimony is based on scientifically valid principles.” Id. at 1317-18 (internal quotation marks omitted). Peer review is the chief way of satisfying this requirement, though it may also be met by precisely [explaining] how [the experts] went about reaching their conclusions and point[ing] to some objective source — a learned treatise, the policy statement of a professional association, a published article in a reputable scientific journal or the like — to show that they have followed the scientific method, as it is practiced by (at least) a recognized minority of scientists in their field. Id. at 1318-19 (citing United States v. Rincon, 28 F.3d 921, 924 (9th Cir.1994)). B. Scientific Evidence Presented Metabolite presented scientific evidence through the declarations of six experts and the results of scientific research. Five of the experts presented opinions based on “scientific risk assessment.” 13 The sixth expert, Dr. Ruth Hammel Strauss, interpreted the results of an unpublished cardiovascular risk study that she conducted at Columbia Medical Center. Besides the Columbia study, the results of two other research projects were also submitted to the district court: (1) animal toxicity tests conducted at Shanghai Medical University and National Taiwan University and (2) short-term efficacy studies at Vanderbilt University Medical Center and St. Luke’s-Roosevelt Hospital Center. After noting that, with the exception of Dr. Strauss’s, all of the opinions expressed [842] in the declarations were prepared for the purpose of litigation, the court dismissed the “scientific risk assessments” because it found the underlying sources unreliable and the explanation of methodology lacking. 72 F.Supp.2d at 1167-70. The district court also found the Asian animal studies, Dr. Strauss’s interpretation of the Columbia study, and the efficacy studies too unreliable to be admitted. Id. a. The Asian Animal Studies The district court held that, as a matter of law, animal studies are inadmissible “due to the uncertainties in extrapolating from effects on mice and rats to humans.” Id. at 1169. The district court was also troubled because the animal studies “took place outside the United States government’s regulatory supervision.” Id. The district court’s ruling was incorrect. First, Daubert II itself recognized that animal studies are not per se inadmissible and should be subjected to substantive analysis, just like other scientific evidence. 43 F.3d at 1319 (conducting substantive analysis of animal studies). The cases cited by the district court in support of its per se rule — Turpin v. Merrell Dow Pharms., Inc., 959 F.2d 1349 (6th Cir.1992), Lynch v. Merrell-Nat’l Lab., Div. of Richardson-Merrell, Inc., 830 F.2d 1190 (1st Cir.1987), and In re “Agent Orange” Prod. Liab. Lit., 611 F.Supp. 1223 (E.D.N.Y.1985)—are all pre-Daubert. 72 F.Supp.2d at 1169. Beyond that, they are inapposite. Turpin and Lynch were products of the Bendectine birth-defects litigation. These cases merely hold that, in predicting birth defects, the developmental patterns of different species are too different to allow for the presence or absence of a birth defect in one species to be reliable evidence of the likelihood of such a birth defect occurring in another species. Both cases are limited to birth defects; Turpin even notes, “No doubt there may be other animal experiments which, to cite one example, because of the extreme toxicity of the substance tested, would permit a reasonable jury to find that it is more probable than not that the substance causes a similar harm to humans.” 959 F.2d at 1359. The other case, In re “Agent Orange”, cites a study for the proposition that “[a]n-imal studies are aimed at discovering a dose-response relationship, while epidemiological studies show an association between exposure and disease” and concludes that, because of the unique facts of that case, “[t]he animal studies are not helpful in the instant case....” 611 F.Supp. at 1241. Again, the case neither creates nor applies a general rule of unreliability. None of these cases holds that animal studies will always be too unreliable to provide admissible evidence about human health issues. Notwithstanding the moral and ethical problems often surrounding animal studies, in some circumstances they provide useful data about human health.14 The district court erred in rejecting the animal studies proffered by Metabolite merely because of the species gap. [843] Also wrong is the district court’s view that experimentation outside the United States is somehow presumptively unreliable. While regulation of experimentation in the United States may bolster the reliability of results generated domestically, there is no reason to assume that experimentation abroad either would not meet those regulations or is unreliable despite deviancies. We note another of the district court’s concerns, the difficulty in extrapolating from high-dose, short-term studies, such as the Asian animal studies, to the low-dose, long-term usage that would result from continued use of Metabolite 356 as directed. 72 F.Supp.2d at 1169. A variance between experimental conditions and real world usage might indeed be problematic, but we do not read the district court’s order as relying on this issue alone, and even if it did, it would be an abuse of discretion to exclude the studies merely because “[difficulties in such extrapolation has lead to controversy concerning the admissibility of such studies.” Id. Difficulties with extrapolation might render the animal studies unreliable under Daubert; however, such a determination must be made on problems inherent to the studies themselves, not a general apprehension at inter-species and inter-dosage extrapolation. After United States v. Alatorre, 222 F.3d 1098, 1100 (9th Cir.2000), we are prohibited from ordering a district court “to conduct pretrial hearings in order to discharge [its Daubert ] gatekeeping function.” Thus, we merely hold that the district court’s analysis of the reliability of these studies constituted an abuse of discretion. While evidentiary hearings might help the district court to conduct an adequate Daubert analysis, “[t]he trial court must have the same kind of latitude in deciding how to test an expert’s reliability, and to decide whether and when special briefing or other proceedings are needed to investigate reliability, as it enjoys when it decides whether or not that expert’s relevant testimony is reliable.... ” Alatorre, 222 F.3d at 1102 (quoting Kumho Tire Co. v. Carmichael, 526 U.S. 137, 153, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999)). b. The Columbia Study The district court found the Columbia cardiovascular study unreliable because (1) it was commissioned by Metabolite, (2) it was not completely finished, and (3) the part that was finished had not been subjected to peer review. 72 F.Supp.2d at 1169-70. Metabolite argues that the relevant portion of the study was completed with the data in final form, the research was begun pre-litigation, and Dr. Strauss’s declaration “provides a detailed description of the methodology that she used.” Excluding the Columbia study was an abuse of discretion. It was plain error to hold that the Columbia study was not finished — while the overall project was ongoing, all of the relevant data had been gathered in final form, and Metabolite presented an expert interpretation of that data. We remand the issue of reliability to the district court. As explained above, when research is begun pre-litigation, it may be reliable without peer review. Daubert II, 43 F.3d at 1317. Rather than disqualify the study because of “incompleteness” or because it was commissioned by Metabolite, the district court should examine the soundness of the methodology employed. The district court also noted in dicta that the Columbia study is “of questionable relevancy” to the falsity of the statement “you can die from taking this product as directed” because the study “dealt only with particular cardiovascular effects,” and “there are more ways to die than through [844] ‘significant adverse cardiovascular events’.... ” This logic is difficult to follow given that the main health risk allegedly associated with Metabolite 356 is its effect as a stimulant on the cardiovascular system. c. The Short-Term Efficacy Studies An efficacy study is a study that determines whether Metabolite 356 actually helps people lose weight. It is not designed, at least primarily, to be a safety study. Because of this, the district court held that the efficacy studies are “not reliable research methodology for testing the safety of a supplement intended for long-term use. Safety testing is not even the purpose of the study’s research design.” 72 F.Supp.2d at 1170. While we decline to create a per se rule about what safety information can be reliably gleaned from efficacy studies, we agree with the district court that any such data generated by the studies at issue in this case lacks sufficient grounding in the scientific method to be admissible under Daubert II. d. Expert Risk Assessment Risk assessment is an accepted methodology practiced extensively throughout the medical, scientific, and regulatory communities over the past thirty years. See Bernard Goldstein & Mary Sue Henifen, Reference Guide on Toxicology in Federal Judicial Center Reference Manual on Scientific Evidence 193 (1994). Standard risk assessment involves four stages: hazard identification, dose-response assessment, exposure assessment, and risk characterization. Id. Metabolite argues that consistent with this methodology, its experts consulted “a wealth of peer-reviewed articles, Food and Drug Administration adverse incident reports,15 studies, laboratory reports, and other scientific materials to formulate their opinions.” The district court rejected the experts’ risk assessments because they did “not explain precisely how they use[d] the scientific literature to support their opinion[s]. Rather, the experts list[ed] numerous articles in scientific journals and simply state[d] that, after reviewing these articles [and other information], they [were] convinced that Metabolite 356 cannot cause serious health problems.” 72 F.Supp.2d at 1170. The district court was troubled by the titles of several articles cited, noting that it did not understand “how articles such as these support the opinions of Metabolite’s experts.... ” Id. Metabolite contends that the district court’s ruling was an abuse of discretion, arguing that since the articles referred to were in peer-reviewed journals, the experts were not required to explain specifically how each article impacted their opinion. In Metabolite's view, the explanation requirement only arises when there is an absence of peer-reviewed literature directly supporting the position held by the expert. Neither the district court’s nor Metabolite’s application of the Daubert II requirements in entirely correct. Metabolite is correct that peer-review is highly probative under Daubert II, 43 F.3d at 1318-19, but here the articles were not written by the experts who now wish to interpret them. Metabolite’s experts, through risk assessment methodology, are interpreting peer-reviewed articles written by other scientists. The district court, as gatekeeper, correctly noted that the methodology of their interpretation should be open to scrutiny. [845] However, the district court abused its discretion in its summary decision that the risk assessments were not adequately explained. 72 F.Supp.2d at 1170. In Daubert II, we said that scientific evidence, such as a risk assessment, that is prepared for litigation and not peer-reviewed itself, may be bolstered “through the testimony of ... [the] experts” who prepared the evidence. 48 F.3d at 1319. “For such a showing to be sufficient, the experts must explain precisely how they went about reaching their conclusions and point to some objective source ... to show that they have followed the scientific method....” Id. Here, Metabolite’s experts explained the process of risk assessment and pointed to objective sources, but did not, in the district court’s view, adequately explain how those objective sources related to their methodologies and eventual conclusions. We agree with the district court that the risk assessment evidence is complex, but complexity is not an adequate ground for exclusion. Examining the declarations of the scientists who prepared the risk assessments, it is clear that they have facially complied with Daubert II’s verification requirement for evidence prepared in anticipation of litigation — the declarations explain the methodology of risk assessment and how the data found in peer-reviewed articles and adverse incident reports was used. See 43 F.3d at 1318-19. However, due to the complexity of this evidence .and our deferential role in reviewing the admissibility of scientific evidence, we are not prepared to override the district court’s role as gatekeeper and hold that the risk assessment evidence is admissible. Rather, we simply hold that the wholly conelusory grounds for exclusion listed by the district court constitute an abuse of discretion. If on remand the district court wishes to plumb the depths of the precise relationship between the materials cited and the conclusions drawn, that is entirely within its province as a Daubert II gatekeeper. III. Metabolife’s Right to Discovery Metabolite contends that, even if it did not establish a prima facie case of falsity through its submissions to the district court, the district court erred in not allowing it discovery because the discovery-limiting aspects of the anti-SLAPP statute conflict with Federal Rule of Civil Procedure 56. Procedural state laws are not used in federal court if to do so would result in a “direct collision” with a Federal Rule of Civil Procedure. Walker v. Armco Steel Corp., 446 U.S. 740, 749-50, 100 S.Ct. 1978, 64 L.Ed.2d 659 (1980). In the absence of a “direct collision,” the court must make the “typical, relatively unguided Erie Choice.” Hanna v. Plumer, 380 U.S. 460, 471, 85 S.Ct. 1136, 14 L.Ed.2d 8 (1965) (citing Erie R.R. Co. v. Tompkins, 304 U.S. 64, 58 S.Ct. 817, 82 L.Ed. 1188 (1938)). This choice should be made by balancing the state interest in its procedural rule with the twin purposes of the Erie doctrine, “discouragement of forum-shopping and avoidance of inequitable administration of the laws.” Id. at 468, 85 S.Ct. 1136. In United States v. Lockheed Missiles & Space Co., Inc., 190 F.3d 963, 970-73 (9th Cir.1999), we considered whether two subsections of the anti-SLAPP statute may properly be invoked in federal court. The subsections in question were the special motion to strike, § 425.16(b), and the availability of fees and costs, § 425.16(c). Id. We held that there is no direct conflict between these two subsections and the Federal Rules, and that the purposes of Erie are advanced by adopting the California procedural rules. Id. However, the [846] court did not address other subsections of the anti-SLAPP statute, such as §§ 425.16(f) and (g). Subsection 425.16(f) provides that the anti-SLAPP motion may be filed within sixty days of the filing of the complaint or, at the court’s discretion, at any later date. Subsection 425.16(g) provides that the filing of an anti-SLAPP motion automatically stays all further discovery until the court rules on the motion. However, “[t]he court, on noticed motion and for good cause shown, may order that specified discovery be conducted notwithstanding this subdivision.” § 425.16(g). Together, these two subsections “create a default rule that allows the defendant served with a complaint to immediately put the plaintiff to his or her proof before the plaintiff can conduct discovery.” Rogers v. Home Shopping Network, Inc., 57 F.Supp.2d 973, 980 (C.D.Cal.1999). We have not previously considered whether subsections 425.16(f) and (g) “directly collide” with the Federal Rules or are contrary to Erie’s purposes. However, a district court in our circuit addressed exactly this issue in Rogers, holding that “[i]f this expedited procedure were used in federal court to test the plaintiffs evidence before the plaintiff has completed discovery, it would collide with Federal Rule of Civil Procedure 56.” 57 F.Supp.2d at 980. Although Rule 56(f) facially gives judges the discretion to disallow discovery when the non-moving party cannot yet submit evidence supporting its opposition, the Supreme Court has restated the rule as requiring, rather than merely permitting, discovery “where the nonmoving party has not had the opportunity to discover information that is essential to its opposition.” Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 250 n. 5, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986). Taking note of this, the district court in Rogers held: Section 425.16 limits discovery and makes further discovery an exception, rather than the rule. Rule 56 does not limit discovery. On the contrary, it ensures that adequate discovery will occur before summary judgment is considered. Because the discovery-limiting aspects of § 425.16(f) and (g) collide with the discovery-allowing aspects of Rule 56, these aspects of subsections (f) and (g) cannot apply in federal court. 57 F.Supp.2d at 982. We agree. In this case, the district court also adopted the Rogers analysis, but failed to implement it properly. Recognizing that it “should not scrutinize Plaintiffs evidence of facts uniquely within the Defendants’ control before ordering discovery to enable Plaintiff to meet its burden of opposing Defendants’ anti-SLAPP motions,” the district court decided not to rule on the prima facie case of actual malice. 72 F.Supp.2d at 1166. However, the district court reached the issue of falsity regarding the statement that “Every expert we asked said Metabolife [356] is not safe because of its main ingredient, ma huang.” Id. at 1172-73. The district court found against Meta-bolife on this issue because it felt that Metabolife had not established that its product is safe when used as directed. Id. Since we are remanding for further Dau-bert II analysis of the proffered scientific evidence, for reasons discussed infra we also order the district court to allow discovery as to which experts Wornick consulted as the basis for this statement.16 [847] This information is in the defendants’ exclusive control, and may be highly probative to Metabolite’s burden of showing falsity. IV. Alternative First Amendment Rulings A. “You can die from taking this product.” The district court held, sua sponte, that “[e]ven if Metabolife proved its prima facie case of falsity” about the statement that “You can die from taking this product,” the “statement is protected as a ‘rational interpretation’ of the ambiguous and unresolved state of scientific knowledge regarding the safety of products like Metabolife.” 72 F.Supp.2d at 1170. The court concluded that, “Given the controversy surrounding the safety of Metabolife 356, Defendant Blackburn’s statement is incapable of supporting a finding of actual malice.” Id. at 1171. At this point, the context of Dr. Blackburn’s statement should be discussed. Only a small portion of Dr. Blackburn’s interview with Wornick was broadcast during the “investigative report” — the rest was edited out. The full text of the relevant portion of the interview shows that Dr. Blackburn actually said: The documents from the FDA hearings remained on the Internet in 1999, when we did this work years ago, and they know, even today as I know, there are people taking similar types of these products who are getting heart attacks, and of course the abuse [sic] can lead to death. But I mean, you can die from taking this product. All the audience heard from this portion of the interview was “You can die from taking this product.” On our facts, Dr. Blackburn is not responsible for the subsequent editing of his interview — he is only responsible for his comments in their full and complete form, not the sound bites they became. When viewed in its entirety, his statement makes two assertions: (1) people using products similar to Metabolife 356 have had incidents of heart attacks, and (2) abuse of Metabolife 356 can cause death. Metabol-ife does not dispute the validity of these statements anywhere in its pleadings. Accordingly, the dismissal of the causes of action against Dr. Blackburn is affirmed. As for the rational interpretation doctrine, it has no bearing on Dr. Blackburn’s statement as edited by Wornick and her co-defendants. “The protection for rational interpretation serves First Amendment principles by allowing an author the interpretive license that is necessary when relying upon ambiguous sources.” Masson v. Netv Yorker Magazine, Inc., 501 U.S. 496, 519, 111 S.Ct. 2419, 115 L.Ed.2d 447 (1991). The district court applied the doctrine to Dr. Blackburn’s statement because “[i]f the First Amendment provides heightened protection for rational comment on stereo speakers,17 it should also protect scientific comment on issues as [848] important as public health.” 72 F.Supp.2d at 1172, fn. 15. It is clear that the defendants’ editing of Dr. Blackburn’s statement changed its meaning. Even if the complete statement is subject to protection under the rational interpretation doctrine, Wornick and her codefendants cannot piggy-back on that protection after they changed its meaning by cutting out the crucial qualification that “abuse” can lead to death. The sound bite presented by Wornick and the station finds no protection in the rational interpretation doctrine. B. “Every expert.” Due to the exclusion of the scientific evidence, the district court held that Metabolite did not present a prima facie case proving the safety of Metabolite 356 when used as directed. 72 F.Supp. at 1172. The district court also ruled in the alternative that “Metabolite cannot proceed on the alleged defamatory implication” that there is some consensus in the scientific community as to its product’s unsafe nature because “Wornick’s ‘every expert’ statement is not capable of supporting” that implication. Id. at 1173. In order to prevail on the alleged implication of scientific consensus as to Metabolite 356’s unsafety, Metabolite must show “that the words ... uttered were reasonably capable of sustaining that meaning” and “that a jury could reasonably find by clear and convincing evidence that [the defendants] ‘intended to convey the defamatory impression.’ ”18 Dodds v. Am. Broad. Co., 145 F.3d 1053, 1063-64 (9th Cir.1998) (quoting Newton v. Nat’l Broad. Co., Inc., 930 F.2d 662, 681 (9th Cir.1990)). The statement “Every expert we asked said Metabolite [356] is not safe because of its main ingredient, ma huang” is at least susceptible to the implication alleged by Metabolite — that it implies consensus in the scientific community. As the district court noted, 72 F.Supp.2d at 1166, the issue of “actual malice” (or, to put it another way, intent to convey the defamatory impression) cannot be properly disposed of by a motion to dismiss in this case, where there has been no discovery. Therefore, we reverse the district court’s alternative holding that the statement was incapable of supporting the implication alleged. The defendants may challenge whether the asserted implication was made with “actual malice” at summary judgment, should the case proceed that far. The identity of the experts consulted by Wornick is also critical to Metabolite’s burden — without the ability to depose the experts relied on, Metabolite will be faced with a moving target its attempt to disprove consensus. Therefore, limited discovery should be allowed on this issue, as discussed supra. C. The Statement Regarding the Vanderbilt Study Metabolite argues that by discounting the Vanderbilt University study, the broadcast implied that Metabolite 356 had not been tested for safety. The argument is largely based on the introductory comment: “Remember that ad calling Metabolite clinically tested for safety?” The logic goes that by then discounting the Vanderbilt study,19 without mentioning al [849] ternative studies of which Wornick was allegedly aware, the broadcast implied that no valid safety testing had been conducted. The district court disagreed, holding that the statement regarding the Vanderbilt study was literally true, and that even if the statement supported the defamatory implication, Metabolife could not prove the implication false because it had submitted no valid affirmative safety studies: The remaining propositions, which focus on the alleged implication that no scientific studies support the safety of Meta-bolife 356, are protected as substantially true. At the time of the broadcasts, the Chinese animal studies were the only studies touting the safety of Metabolife 356. For the reasons that the court describes above, those studies are so insubstantial as to be essentially the same as “no studies” for purposes of the “gist” if Defendants’ public concern speech. 72 F.Supp.2d at 1174-75 (emphasis in original) (footnote omitted). Because we have reversed the district court’s exclusion of the Asian animal studies and remanded for further Daubert consideration, supra, we reverse this holding and remand it as well. If on remand the Asian animal studies are again excluded, then the district court’s analysis would ring true. D. The Statements that Metabolife 356 and Methamphetamine Share the Same Main Ingredient Metabolife sought relief for statements that Metabolife 356 and methamphetamine share the same main ingredient, “the controlled substance” ephedrine, “a powerful heart stimulant.” To prove falsity, Metabolife offers expert testimony that synthetic ephedrine, the active ingredient in methamphetamine, is distinct from ma huang, or naturally-occurring ephedrine, the active ingredient in Metabolife 356. The expert declared that although synthetic and naturally-occurring ephedrine have similar effects, synthetic ephedrine is several times more potent than its naturally-occurring counterpart. The district court rejected Metabolite’s argument because “[t]he fact that Metabolife requires expert scientific opinion to describe the limited factual differences between ma huang and [synthetic] ephedrine convinces the court that such fine distinctions would have no effect on the state of minds [sic] of the audience had they been raised by Defendant Wornick.” 72 F.Supp.2d at 1176. The district court’s holding was legally erroneous. A statement is not “substantially true” if it “would have a different effect on the mind of the reader [or viewer] from that which the pleaded truth would have produced.” Masson, 501 U.S. at 517, 111 S.Ct. 2419 (quoting R. Sack, Libel, Slander, and Related Problems 139 (1980)). Here, Metabolife introduced evidence that the synthetic ephedrine used in methamphetamine has significantly different potency and absorption rates than naturally-occurring ephedrine. The district court rejected this argument out-of-hand, holding essentially that the argument was too technical for viewers to grasp. 72 F.Supp.2d at 1176. This holding was in error. We do not agree with the district court that distinctions between natural and synthetic forms of substances are beyond a reasonable viewer’s comprehension, especially when the distinction comes with substantial differences in potency and absorption rates. Anyone who knows the difference between a double espresso and a regular cup of coffee should understand that a reasonable jury could, have found the differences in potency and absorption significant. The dissent may be correct that synthetic and naturally-occurring ephedrine are technically the same sub [850] stance, but we are unprepared to make that holding on this record.20 Deciding only that which was decided below, we reverse and remand the district court’s ruling that the statement that Metabolite 356 and methamphetamine share the same ingredient is “substantially true” as a matter of law. On remand, the district court should consider whether it is substantially true that, when considering Metabolite’s proffered evidence on potency and absorption (so far as its reliability and relevance extends), synthetic and naturally-occurring ephedrine are the “same main ingredient.” CONCLUSION The district court’s exclusion of Metabolite’s scientific evidence regarding the risk assessments, Asian animal studies, and Columbia study on the grounds explicated constituted an abuse of discretion and is REVERSED AND REMANDED. The risk assessments facially satisfy Daubert II’s requirements by explaining then-methodology and citation to published, peer-reviewed sources. The Asian animal studies are not unreliable simply because they involve the transposition of data across the species gap and were conducted in China and Taiwan. The Columbia study was completed, prepared independent of litigation, and its methodology appears to have been adequately explained. We do not override the district court’s role as gatekeeper and hold that this evidence is admissible. Rather, we simply hold that it was an abuse of discretion to exclude it for the reasons cited. Additionally, the district court’s exclusion of the efficacy studies is AFFIRMED. The district court’s decision not to allow Metabolite discovery on falsity issues under Federal Rule of Civil Procedure 56(f) is REVERSED because Metabolite identified and requested discovery of probative information solely available from the defendants. The district court’s alternative free speech rulings are REVERSED as to all defendants except Dr. Blackburn. The dismissal of the causes of action against Dr. Blackburn is AFFIRMED in light of his complete statement. The other defendants cannot use the rational interpretation doctrine to justify his statement because they materially altered it through editing. Wornick’s “every expert” statement is at least legally susceptible to the defamatory implication of scientific consensus. Finally, a reasonable jury could find that it is not “substantially true” that Metabolite 356 and methamphetamine share the same main ingredient. The district court’s decision to grant Dr. Blackburn’s anti-SLAPP motion is AFFIRMED. The district court’s decision to grant the other defendants’ anti-SLAPP motions is REVERSED. The case is REMANDED to the district court for further analysis of the admissibility of the proffered scientific evidence, discovery as specified, and subsequent reassessment of the other defendants’ anti-SLAPP motions. Costs on appeal to appellant and Dr. Blackburn. Footnotes [1] We grant both Metabolife’s and the station’s motions to present videotapes of the reports under Federal Rule of Appellate Procedure 10(a)(1) (exhibits filed in the district court constitute part of the record on appeal). We have reviewed both tapes. All other pending motions are denied. [2] Under 21 U.S.C. § 321, Metabolife 356, as an herbal supplement, is a "food,” not a "drug,” and thus need not undergo the Food and Drug Administration's "new drug” testing policy found at 21 U.S.C. § 355. [3] Personal jurisdiction and venue were issues raised in the district court, but are not implicated in this appeal. [4] The actual sequence challenged was slightly longer: Dr. Blackburn: "You can die from taking this product.” Wornick: "From that product?” Dr. Blackburn: "Yes. From this product.” Wornick: “From Metabolife.” Dr. Blackburn: "Exactly.” [5] The actual statement was (Wornick speaking): "Remember that ad calling Metabolife clinically tested for safety? Metabolife was tested at Vanderbilt University, but only for two weeks and, according to their attorney, not for safety. Vanderbilt officials have ordered Metabolife to stop making that claim.” [6] The actual statements are: Wornick: "The substance has long had the attention of law enforcement, because its also the main ingredient in the illegal drug methamphetamine. On the streets they call it meth, or speed.” Wornick: "[Metabolife founder Ellis] started a vitamin company that later became Metabolife — makers of diet pills with ephedrine. Again, the same controlled substance found in methamphetamine.” Wornick: "[Interviewee] thinks she reacted to ephedrine, a powerful heart stimulant that’s the main ingredient in the illegal drug methamphetamine, known in the streets as speed.” [7] "Anti-SLAPP” stands for "Anti-Strategic Lawsuit Against Public Participation.” The purpose of the statute is to protect individuals from meritless, harassing lawsuits whose purpose is to chill protected expression. [8] The dissent argues that Metabolife conceded that it needed no discovery from the defendant on falsity issues. However, Metabolife filed repeated objections to the district court's decision not to allow discovery, arguing that it needed particular discovery as to specific falsity issues, such as which experts Wornick interviewed as the foundation for her “every expert” statement. [9] The dissent suggests that the statement cannot support this implication, an approach which would decide an issue the district court explicitly declined to reach. We prefer to leave the issue to the district court, should it be necessary to reach it on remand. [10] Metabolife argued, and argues, that ma huang is naturally-occurring ephedrine, which is distinct from the synthetic ephedrine used to make methamphetamine. [11] The California legislature passed the statute recognizing "the public interest to encourage continued participation in matters of public significance ... [finding] that this participation should not be chilled through abuse of the judicial process.” 5 Witkin, California Proc., § 962, at 422 (4th ed.1997). [12] We limit our subsequent exposition of the law and analysis of the proffered evidence to reliability because that was the sole issue addressed in the district court. Given our limited, deferential role in reviewing Daubert decisions, we think it best in this case to address the decisions made below (solely reliability), and not reach out to decide others (such as relevance) that can be dealt with in the first instance on remand, if necessary. [13] As discussed below, "scientific risk assessment” consists of reviewing incident reports and the literature on a particular topic and then forming an opinion based on the information reviewed. The expert offering the opinion has generated his opinion "in the library” — the opinion is not based on personal experimentation or direct analyses of raw experimental data. [14] The dissent contends that Metabolife's experts "recognize” that "the Chinese animal studies offer no basis for extrapolating from effects on mice, rats or beagles to humans.” (Dissent at 12288). The expert quoted in support of such "recognition” merely stated that "straight extrapolation of animal data to humans is not appropriate.” Of course, bridging the species gap may require sophisticated scientific technique and analysis (i.e., more than "straight extrapolation”) to produce reliable and relevant results. Determining whether the proffered analysis of the Asian animal studies meets the necessary threshold is precisely the district court’s role as gatekeeper on remand. [15] Through adverse incident reports, the FDA collects complaints from people who have negative experiences with "foods” like Metabolife 356. [16] The dissent makes the interesting contention that Metabolife has not "proffer[ed] sufficient facts to show that the evidence sought exists, and that it would prevent summary judgment.” (Dissent at 12290 (citing Chance v. Pac-Tel Teletrac Inc., 242 F.3d 1151, 1161 [847] n. 6 (9th Cir.2001))). Unless they do not exist, Wornick should be able to divulge which experts she spoke with while preparing her story. [17] The district court's reference to stereo speakers is a reference to Bose Corp. v. Consumers Union, 466 U.S. 485, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984), one of just three Supreme Court cases construing the rational interpretation doctrine. There, the issue was the published review of stereo speakers. The Court concluded “that the result was not an assessment of events that speak for themselves, but 'one of a number of possible interpretations of an event “that bristled with ambiguities'' and descriptive challenges for the writer.’ ” Masson, 501 U.S. at 519, 111 S.Ct. 2419 (quoting Bose Corp., 466 U.S. at 512, 104 S.Ct. 1949). [18] The second prong is the well known "actual malice” requirement dating back to New York Times Co. v. Sullivan, 376 U.S. 254, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964). See Newton v. Nat’l Broad. Co., Inc., 930 F.2d 662, 681-83 (9th Cir.1990). [19] As discussed, supra, the broadcast was correct in its assertion that the Vanderbilt efficacy study did not prove Metabolife 356’s safety. [20] We know that synthetic ephedrine is more potent and has a higher absorption rate than natural ephedrine. However, we do not know whether this comes from sheer concentration, or whether the two substances have subtle, but important, chemical differences.