This is a permissible interlocutory appeal from an order certifying a securities-fraud class action. Plaintiffs allege violations of section 10(b) and 20(a) of the Securities Exchange Act of 1934 and Rule 10b-5 of the Securities Exchange Commission. Relying on the fraud-on-the-market theory, the district court certified the class. We vacate the certification order and remand, persuaded that the class certified fails for wont of any showing that the market reacted to the corrective disclosure. Given the lethal force of certifying a class of purchasers of securities enabled by the fraud-on-the-market doctrine, we now in fairness insist that such a certification be supported by a showing of loss causation that targets the corrective disclosure appearing among other negative disclosures made at the same time.
I
The class included all investors who purchased the common stock of Allegiance Telecom between April 24, 2001 and February 19, 2002. Three investors bring this suit, Oscar Private Equity Investments, its managing partner, Brett Messing, and his wife, Marla Messing. They sue Royce Holland, former chairman and CEO of Allegiance, and Anthony Parella, former executive vice president for sales. Allegiance Telecom was named in the suit, but filed for bankruptcy and is not now a party.
Allegiance was a national telecommunications provider based in Dallas, Texas. It sold local telephone service, long distance, broadband access, web hosting, and telecom equipment with maintenance to small and medium sized businesses. Founded in 1997, by February 2002 it was providing service in thirty-six U.S. markets. At the beginning of the class period, *263April 24, 2001, there were over 112 million common shares of Allegiance stock trading on the NASDAQ. Institutional investors held approximately 68 percent of Allegiance’s stock and over fifty active market makers traded it.
Plaintiffs allege that Holland and Par-ella fraudulently misrepresented Allegiance’s line-installation count in the company’s first three quarterly announcements of 2001, and that Allegiance’s stock dropped after Holland and Parella ultimately restated the count in the 4Q01 announcement. Defendants explain that the restatement occurred because Allegiance installed a new billing system in 2001 and reported line-count information from the new billing system instead of from the order management system which it replaced. Defendants further argue that the 4Q01 restatement did not cause the stock price to drop.
The relevant announcement history is as follows. Allegiance’s stock, like that of the rest of the telecom industry, was plunging during what is now the class period, losing nearly 90% of its value during 2001. On April 24, 2001, the first day of the class period, Allegiance announced its 1Q01 results, including (1) 126,200 new lines installed; (2) revenues of $105.9 million, an 11% increase over 4Q00; (3) positive sales force growth; and (4) improved gross margin. The following trading day Allegiance’s stock rose 9%, from $14.90 to $16.20, but soon declined again.
On July 24, 2001, Allegiance announced its 2Q01 results, including (1) 135,800 new lines installed; (2) revenues of $124.1 million; (3) an earnings loss of $0.92 per share, $0.03 better than the analysts’ consensus estimate; and (4) positive EBIT-DA1 results in thirteen markets. The following trading day Allegiance’s stock rose 20%, from $10.90 to $13.08 per share, but soon declined again.
On October 23, 2001, Allegiance announced its 3Q01 results, including (1) the installation of its one-millionth line; (2) revenues of $135 million; and (3) an earnings loss of $0.94 per share, $0.03 better than the analysts’ consensus estimate. The next trading day Allegiance’s stock rose 29%, from $5.21 to $6.74 per share, but remained volatile, falling to $3.70 per share by February 18, 2002, the day before the curative statements of the 4Q01 announcement.
On February 19, 2002, Allegiance announced its 4Q01 results, including (1) a restatement of the total installed-line count from 1,140,000 to 1,015,000, a difference of 125,000; (2) missed analysts’ expectations on 4Q01 and 2001 earnings per share; (3) greater EBITDA loss than some analysts expected; and (4) a very thin margin of error for meeting revenue covenants for 2002. The next trading day Allegiance’s stock continued its downward move, falling 28%, from $3.70 to $2.65 per share. Less than 90 days later, Allegiance missed its covenants putting its credit lines in default and on May 14, 2003, filed for bankruptcy.
Six months after Allegiance’s bankruptcy, plaintiffs filed this class action, alleging that Allegiance’s officers misrepresented the number of installed lines in their 1Q01, 2Q01, and 3Q01 announcements. Plaintiffs moved for class certification, relying on the fraud-on-the-market presumption for evidence of class-wide reliance. The district court certified the class,2 and we granted interlocutory review.
*264II
The class certification determination rests within the sound discretion of the trial court, exercised within the constraints of Rule 23.3 A district court that premises its legal analysis on an erroneous understanding of the governing law has abused its discretion.4
III
This dispute turns on whether the certification order properly relied upon the fraud-on-the-market theory. This theory permits a trial court to presume that each class member has satisfied the reliance element of their 10b-5 claim.5 Without this presumption, questions of individual reliance would predominate, and the proposed class would fail.6
The Supreme Court in Basic adopted this presumption of reliance with respect to materially misleading statements or omissions concerning companies whose shares are traded in an efficient market.7 Reliance is presumed if the plaintiffs can show that “(1) the defendant made public material misrepresentations, (2) the defendant’s shares were traded in an efficient market, and (3) the plaintiffs traded shares between the time the misrepresentations were made and the time the truth was revealed.”8
We have observed that Basic “allows each of the circuits room to develop its own fraud-on-the-market rules.”9 This court has used this room — in Finkel,10 Abell,11 Nathenson,12 and Greenberg13 — to *265tighten the requirements for plaintiffs seeking a presumption of reliance. We now require more than proof of a material misstatement; we require proof that the misstatement actually moved the market.14 That is, “the plaintiff [may] recover under the fraud on the market theory if he [can] prove that the defendant’s non-disclosure materially affected the market price of the security.”15 Essentially, we require plaintiffs to establish loss causation in order to trigger the fraud-on-the-market presumption.16 Our most recent statement of this rule was in Greenberg, which held that “to trigger the presumption [of reliance] plaintiffs must demonstrate that ... the cause of the decline in price is due to the revelation of the truth and not the release of the unrelated negative information.”17
This requirement was not plucked from the air. Basic plainly states that the presumption of reliance may be rebutted by “[a]ny showing that severs the link between the alleged misrepresentation and ... the price received (or paid) by the plaintiff.”18 This would include “a showing that the market price would not have been affected by the alleged misrepresentations, as in such a case the basis for finding that the fraud had been transmitted through the market price would be gone.”19
Quoting this very language, plaintiffs argue that our requirement improperly shifts the burden, from a defendant’s right of rebuttal to a plaintiffs burden of proof. We disagree. As a matter of practice, the oft-chosen defensive move is to make “any showing that severs the link” between the misrepresentation and the plaintiffs loss; to do so rebuts on arrival the plaintiffs fraud-on-the-market theory. In Nathen-son, the link was severed by publicly available information that the misrepresentation didn’t move the stock price.20 In Greenberg, it was severed by publicly available evidence that the corrective disclosure was buried in other bad news.21 Hence, in cases like this one, we have required plaintiffs invoking the fraud on the market theory to demonstrate loss causation.22
*266The contours of this requirement — that the fraud affect the stock price — is the gist of this appeal. It is a requirement complicated here by the fact that multiple items of positive information were released together with the alleged line-count inflation, and further complicated by the fact that multiple items of negative information were released together with the corrective disclosure. In such multi-layered loss-causation inquiries, the legal standard, at least, is well established: Greenberg requires that plaintiffs prove “(1) that the negative ‘truthful’ information causing the decrease in price is related to an allegedly false, non-confirmatory positive statement made earlier and (2) that it is more probable than not that it was this negative statement, and not other unrelated negative statements, that caused a significant amount of the decline.”23
Neither party disputes Greenberg’s relevance. Instead, this appeal raises the question of whether we ought to apply Greenberg’s loss-causation requirement at the class-certification stage, as well as the subsidiary question of the sufficiency of the evidence to establish the requirement. On the first question, defendants urge that the district court must consider all evidence, both for and against loss causation, at the class certification stage. On the second question, defendants argue that the district court abused its discretion in finding that plaintiffs made a showing sufficient to establish loss causation. We agree with both contentions.
A
First we address the question of whether loss causation — a fraud on the market prerequisite — should properly be addressed at the class certification stage. The district court ruled that “the class certification stage is not the proper time for defendants to rebut lead Plaintiffs’ fraud-on-the-market presumption,” and suggested that Basic “held that the presumption of reliance was rebuttable, but only as related to a summary judgment motion.” Plaintiffs defend the court’s ruling, noting that Greenberg was a summary-judgment case and urging that proof of loss causation at this stage “improperly combines the market efficiency standard with actual proof of loss causation.”
There is widespread confusion on this point. As we will explain, the confusion arises from an outdated view that fails to accord this signal event of the case its due. Under this earlier view, class certification was to be made “as soon as practicable after the commencement of the action,” mindful that the decision was tentative. It could be tailored to facts emerging in discovery, and with subclasses built around awkward difficulties of showings that cut across only part of the class first certified. In short, class certification was a light step along the way, divorced from the merits of the claim. Whatever reality this treatment was responsive to, it is not that of a class exceeding purchasers of millions of shares in a volatile and downward-turning market over a ten-month period, claiming injury from one of several simultaneous disclosures of negative information.
The power of the fraud-on-the-market doctrine is on display here. With proof that these securities were being traded in an efficient market, the district court effec*267tively concluded that if plaintiffs can establish at trial that defendants acted with the requisite intent in counting its installations then defendants would be liable for millions of dollars in paper losses on the day following the fourth-quarter filing date, less the amount the defendant may be able to persuade a jury was caused by other circumstances — whether the purchaser held on and later sold at a higher price or rode the stock down to bankruptcy. In short, the efficient market doctrine facilitates an extraordinary aggregation of claims. We cannot ignore the in terrorem power of certification, continuing to abide the practice of withholding until “trial” a merit inquiry central to the certification decision, and failing to insist upon a greater showing of loss causation to sustain certification, at least in the instance of simultaneous disclosure of multiple pieces of negative news. Nor is there sound reason for an early “tentative” certification, which leaves loss causation for later more focused examination. It is not the need for discovery. Little discovery from defendants is demanded by the fraud-on-the-market regimen. Its “proof’ is drawn from public data and public filings, as in this case. It is largely an empirical judgment that can be made then as well as later in the litigation.
These concerns have shaped the evolution of class certification and Rule 23. Rule 23(c)(1)(A) no longer demands that the district court rule on class certification “as soon as practicable,”24 but instead insists only upon a ruling “at an early practicable time.”25 And although Rule 23 still recognizes that a class may be “altered or amended,”26 it no longer characterizes the class certification order as “conditional,”27 explaining, in the advisory committee notes, that “[a] court that is not satisfied that the requirements of Rule 23 have been met should refuse certification until they have been met.”28 These subtle changes, as well as the less-subtle PSLRA, recognize that a district court’s certification order often bestows upon plaintiffs extraordinary leverage, and its bite should dictate the process that precedes it. These changes are the product of years of study by the Advisory Committee on Civil Rules, including many open hearings and symposia. This collective wisdom must not be brushed aside. That there are “important due process concerns of both plaintiffs and defendants inherent in the certification decision,” cannot be gainsaid.29 Thus, in TJnger, a similar 10b-5 case, we held that “[t]he plain text of Rule 23 requires the court to ‘find,’ not merely assume, the facts favoring class certification.”30 And we concluded that “[bjecause Rule 23 mandates a complete analysis of fraud-on-the-market indicators, district courts must address and weigh factors both for and against market efficiency.”31 This conclusion, that courts must examine factors both for and against, applies to the *268determination of all Rule 23’s requirements.
Relatedly, Rule 23’s requirements must be given their full weight independent of the merits. District courts often tread too lightly on Rule 23 requirements that overlap with the 10b-5 merits, out of a mistaken belief that merits questions may never be addressed at the class certification stage.32 This is a misreading of Eisen, an early class-certification decision by the Supreme Court.33 The Eisen Court stated, “We find nothing in either the language or history of Rule 23 that gives a court any authority to conduct a preliminary inquiry into the merits of a suit in order to determine whether it may be maintained as a class action.”34 As Judge Jon Newman of the Second Circuit recently explained, “This statement has led some courts to think that in determining whether any Rule 23 requirement is met, a judge may not consider any aspect of the merits, and has led other courts to think that a judge may not do so at least with respect to a prerequisite of Rule 23 that overlaps with an aspect of the merits of the case.”35
Eisen did not drain Rule 23 of all rigor. A district court still must give full and independent weight to each Rule 23 requirement, regardless of whether that requirement overlaps with the merits.36 The statement in Eisen is troublesome only if read without the light of its facts. In Eisen, the district court’s improper merits inquiry was unrelated to the Rule 23 requirements. And the same was true in our Miller decision, which was relied upon by Eisen, and which also held that a district court could not deny certification based on its view of the merits.37 Both Eisen and Miller “stand for the unremarkable proposition that the strength of a plaintiffs claim should not affect the certification decision.”38 As the Second Circuit recently concluded, a district court must
resolve[ ] factual disputes relevant to each Rule 23 requirement and find[] that whatever underlying facts are relevant to a particular Rule 23 requirement have been established.... [T]he obligation to make such determinations is not lessened by overlap between a Rule 23 requirement and a merits issue, even a merits issue that is identical with a Rule 23 requirement.39
The answer to our first question, then, lies at the intersection of Greenberg and Unger. Greenberg holds that loss causation is a fraud-on-the-market prerequisite. *269Unger mandates “a complete analysis of fraud-on-the-market indicators” at the class certification stage, insisting that district courts “find” the facts favoring class certification. We hold hence that loss causation must be established at the class certification stage by a preponderance of all admissible evidence.40
Plaintiffs respond that the question of loss causation requires only a generalized inquiry into whether the misrepresentation moved the stock, an inquiry common to all members of the class. Pressing this point at oral argument, plaintiffs urged that it was inappropriate to address loss causation at the class-certification stage because loss causation necessarily predominates, unlike individualized questions of reliance.
We might agree, if loss causation were only empirical proof of materiality, unm-oored from the question of classwide reliance. Yet we have explained that the refutation of loss causation “more appropriately relates to the element of reliance.”41 This is because loss causation speaks to the semi-strong efficient market hypothesis on which classwide reliance depends, as we will explain.
The assumption that every material misrepresentation will move a stock in an efficient market is unfounded, at least as market efficiency is presently measured. There are two additional explanations, besides immateriality, for why a misrepresentation might fail to effect the stock price, both relevant to classwide reliance. First, it might be that even though the market for the defendant’s shares has been demonstrated efficient by the usual indicia,42 the market is actually inefficient with respect to the particular type of information conveyed by the material misrepresentation, ie. analysts and market makers do poorly at digesting line-count information. Thus our approach gives effect to information-type inefficiencies, recognizing that “the market price of a security will not be uniformly efficient as to all types of information.”43 A second possible explanation for a misrepresentation’s failure to move the market is that the market was strong-form efficient with respect to that type of information, ie., due to insider trading, the restated line count was reflected by the stock price well before the 4Q01 corrective disclosure. Both explanations resist application of the semi-strong efficient-market hypothesis, the theory on which the presumption of classwide reliance depends. This court honors both theory and precedent in requiring plaintiffs to *270demonstrate loss causation before triggering the presumption of reliance. The trial court erred in ruling that the class certification stage is not the proper time for defendants to rebut lead Plaintiffs’ fraud-on-the-market presumption.
B
The legal error immediately identified, however, does not alone dictate vacatur in this case, as the trial court, out of caution perhaps, did not premise its analysis on its misunderstanding of the law. Indeed, the trial court’s memorandum opinion applies Greenberg and weighs all evidence, both for and against loss causation, in concluding that “it is more likely than not that a significant part of the stock decline causing the putative Class’s loss is attributable to the line count corrective disclosure.” We vacate because this factual conclusion is untenable. The plaintiffs expert report did not establish loss causation, and the district court abused its discretion in certifying the class.
As we explained above, when unrelated negative statements are announced contemporaneous of a corrective disclosure, the plaintiff must prove “that it is more probable than not that it was this negative statement, and not other unrelated negative statements, that caused a significant amount of the decline.”44 We will not attempt to quantify what fraction of a decline is “significant.” We note only that, under these circumstances, proof of a corrective disclosure’s significant contribution to a price decline demands a peek at the plaintiffs damages model — an empirically-based inquiry, not speculation about materiality alone. Yet plaintiffs’ evidence on this point consists chiefly of analyst commentary. For example, after the line-count restatement, James Ott at Hibernia Southcoast Capital cautioned,
Unfortunately, Bears will have additional fodder during 1Q02, as [Allegiance] scrubbed their databases and found some differences in line count between billing and order management platforms ... In light of “Enron-itis,” we believe an increasingly skeptical market will have a negative view of this adjustment ... Unfortunately, the line revision will cloud the company’s otherwise strong performance.
And at BB&T Capital Markets, an analyst groused,
The magnitude of this [line count] adjustment (12% of total) makes it difficult to swallow ... Given the issues surrounding accounting today, the timing of such an adjustment could not be worse.
Plaintiffs cite several other such reports— one calls the line-count restatement “a yellow flag,” and another suggests that “the Street is completely unwilling to listen to management explanations.”
Defendants respond in kind, with more analyst quotes, including one from lead plaintiff Brett Messing, who reported in a May 15, 2002 column for RealMoney.com that “Allegiance’s stocks and bonds are trading at distressed levels because of fears of a revenue covenant violation,45 a more hostile regulatory environment, and customer churn.” Notably Messing did not mention the line-count restatement and named Allegiance’s management team “the industry’s best.” Similarly, Danny Zito at Lehman Brothers was concerned not with the line-count adjustment, which he opined was troubling only because it raised concerns with Allegiance’s back office operations, but with Allegiance’s “potential revenue covenant violation risk.” *271Finally, James Ott at Hibernia, the same analyst quoted extensively by the plaintiffs, also reported that “[n]o material change ha[d] occurred fundamentally in ALGX’s business,” and explained that “the vast majority of the revisions were definitional rather than functional.”
The plaintiffs have the better of this exchange, but nonetheless, their evidence is little more than well-informed speculation. To prove loss causation, and thereby trigger the presumption of rebanee, plaintiffs must do better. The plaintiffs’ expert does detail event studies supporting a finding that Allegiance’s stock reacted to the entire bundle of negative information contained in the 4Q01 announcement, but this reaction suggests only market efficiency, not loss causation, for there is no evidence linking the culpable disclosure to the stock-price movement. When multiple negative items are announced contemporaneously, mere proximity between the announcement and the stock loss is insufficient to establish loss causation.
Plaintiffs expert, in her rebuttal, disagrees, but offers as evidence only the raw opinion of analysts, without supporting study of the market at issue — such as now common use of basic principles of econometrics. The expert’s own concluding paragraph advised that her work was incomplete: “It is possible with further anal-yses to quantify the portion of the decline caused by the restated line count. However, Counsel has advised me that the quantification of damages is not appropriate at this stage of the litigation.” So this is less of a dispute over what showing must be made, and more a dispute over when.
Something like the expert’s “further analyses” is what is missing. While counsel is correct that quantification of damages is presently unnecessary — ie. proof that some percentage of the drop was attributable to the corrective disclosure — the plaintiffs must, in order to establish loss causation at this stage, offer some empirically-based showing that the corrective disclosure was more than just present at the scene.46 And this burden cannot be discharged by opinion bereft of the analysis plaintiffs own expert conceded was necessary, albeit in her counsel’s view at a later stage. The class certification decision bears due-process concerns for both plaintiffs and defendants,47 and an empirical inquiry into loss causation better addresses these concerns than an impenetrable finding akin to a reasonable man assessment. And analyst speculation about materiality, while better informed than a layman, more closely resembles the latter. At least when multiple negative items are contemporaneously announced, we are unwilling to infer loss causation without more. In sum, only a medical doctor who has either conducted a post-mortem or reviewed the work of another who did so, may credibly opine about the cause of death. We do not insist upon event studies to establish loss causation, helpful though they may be. We hold only that the opinions of these analysts, without reference to any post-mortem data they have reviewed or conducted, is insufficient here.
Because plaintiffs have failed to trigger the presumption of rebanee provided by the fraud-on-the-market theory, the class fails and we must vacate the order of certification.
We VACATE and REMAND for further proceedings consistent with this opinion.
. EBITDA is an acronym for earnings before interest, taxes, depreciation and amortization. It is a measure of profitability.
. The district court certified the following class: "All persons, without geographical limitation, who purchased Allegiance common stock in the open market during the period *264from April 24, 2001 through February 19, 2002, inclusive, and who were damaged by defendants’ alleged violations of Section 10(b) and/or 20(a) of the Securities Exchange Act of 1934 and Rule 10b-5 promulgated thereunder. Excluded from the Class are Defendants, their legal representatives, heirs, successors and predecessors in interest, affiliates, assigns, and any entities in which the Defendants (or any of them) had a controlling interest in during the Class Period.”
. Gulf Oil Co. v. Bernard, 452 U.S. 89, 100, 101 S.Ct. 2193, 68 L.Ed.2d 693 (1981).
. Feder v. Electronic Data Systems Corp., 429 F.3d 125, 129 (5th Cir.2005).
. The elements of a 10b-5 action include:
(1) a material misrepresentation (or omission);
(2) scienter, i.e., a wrongful state of mind;
(3) a connection with the purchase or sale of a security;
(4) reliance, often referred to in cases involving public securities markets (fraud-on-the-market cases) as "transaction causation,”;
(5) economic loss; and
(6) “loss causation,” i.e., a causal connection between the material misrepresentation and the loss. Dura Pharms., Inc. v. Broudo, 544 U.S. 336, 125 S.Ct. 1627, 161 L.Ed.2d 577 (2005).
. Fed.R.Civ.Pro. 23(b)(3).
. In re LTV Securities Litigation, 88 F.R.D. 134, 143 (N.D.Tex.1980); Basic Inc. v. Levinson, 485 U.S. 224, 244, 108 S.Ct. 978, 99 L.Ed.2d 194 (1988).
. Greenberg v. Crossroads Systems, Inc., 364 F.3d 657, 661 (5th Cir.2004).
. Abell v. Potomac Ins. Co., 858 F.2d 1104, 1117-18 (5th Cir.1988), vacated on other grounds sub. nom. Fryar v. Abell, 492 U.S. 914, 109 S.Ct. 3236, 106 L.Ed.2d 584 (1989); Nathenson v. Zonagen Inc., 267 F.3d 400, 414 (5th Cir.2001).
. Finkel v. Docutel/Olivetti Corp., 817 F.2d 356, 364 (5th Cir.1987), cert. denied, 485 U.S. 959, 108 S.Ct. 1220, 99 L.Ed.2d 421 (1988).
. Abell, 858 F.2d at 1120-21 (stating in dicta that a plaintiff may recover under the fraud-on-the-market theory “if he could prove that the defendant’s non-disclosures materially affected the market price of the security”).
. Nathenson, 267 F.3d at 414 ("It is clear that a fraud-on-the-market theory may not be the basis for recovery in respect to an alleged misrepresentation which does not affect the market price of the security in question.”).
. Greenberg, 364 F.3d at 662, 665-66.
. Cf. In re Burlington Coat Factory Securities Litigation, 114 F.3d 1410 (3d Cir.1997). ("In the context of an 'efficient' market, the concept of materiality translates into information that alters the price of the firm’s stock.”).
. Nathenson, 267 F.3d at 414 (quoting Abell, 858 F.2d 1104, 1120-21).
. Our approach is unaffected by the Supreme Court's recent and very narrow decision in Dura Pharms., 125 S.Ct. at 1627.
. Greenberg, 364 F.3d at 665.
. 485 U.S. at 245, 108 S.Ct. 978. The Basic Court continues, "For example, if [defendants] could show that the 'market makers’ were privy to the truth about the merger discusses here with Combustion, and thus that the market price would not have been affected by their misrepresentations, the causal connection could be broken.” Id. (emphasis added). Drawing on Abell and Nathenson, the Green-berg court added a showing to this list of "examples.”
. Nathenson, 267 F.3d at 414.
. Id. at 414.
. Greenberg, 364 F.3d at 665.
. Our able brother frames our differences well, but is a bit enthusiastic in our holding. We address here only the simultaneous disclosure of multiple negatives, not all of which are alleged culpable. Indeed, applying the fraud-on-the-market theory to such complex circumstances by rote would yield a victory of habit over reason. With multiple negatives, our usual approach to gauging efficiency and presuming reliance fails because we cannot know that the culpable information was priced, even if objectively material. Proof that the culpable disclosure moved the market addresses this failure. The dissent is troubled that we have not suggested what form such *266proof might take. We have mentioned one form, event studies, for the sake of exposition only. As we explain below, the plaintiff's own expert stated that such proof was well within her grasp. Our further silence is an effort to leave open options, subject to scrutiny in the first instance by opposing experts and the district courts.
. Greenberg, 364 F.3d at 666.
. Fed.R.Civ.P. 23(c)(l)(A)(2003).
. Fed.R.Civ.P. 23(c)(l)(A)(revised 2003).
. Id. at 23(c)(1)(C); see id. Advisory Committee Notes to the 2003 Amendments C'[I]t is appropriate to conduct controlled discovery into the 'merits,' limited to those aspects relevant to making the certification decision on an informed basis.”).
. This word has been demoted to the comments section.
. See Fed.R.Civ.P. 23 Advisory Committee Notes to the 2003 Amendments.
. Unger v. Amedisys Inc., 401 F.3d 316, 321 (5th Cir.2005).
. Unger, 401 F.3d at 321.
. Unger, 401 F.3d at 325.
. See, e.g., Barrie v. Intervoice-Brite, 2006 WL 2792199, *10 (N.D.Tex.,2006) (“Although Basic and Greenberg (the cases relied upon by Defendants) both held the presumption to be rebuttable at the summary judgment stage, such a finding by the court here, where the issue is class certification, would be premature, since the court cannot delve into the actual merits of Lead Plaintiffs’ claims.”).
. Bell v. Ascendant Solutions, Inc., 422 F.3d 307, 311-14 (5th Cir.2005).
. Eisen v. Carlisle & Jacquelin, 417 U.S. 156, 177-78, 94 S.Ct. 2140, 40 L.Ed.2d 732 (1974).
. Miles v. Merrill Lynch, 471 F.3d 24, 27 (2d Cir.2006).
. See General Telephone Company of the Southwest v. Falcon, 457 U.S. 147, 160, 102 S.Ct. 2364, 72 L.Ed.2d 740 (1982) (acknowledging that "class determination generally involves considerations that are enmeshed in the factual and legal issues comprising the plaintiffs’ cause of action,” and concluding that a class “may only be certified if the trial court is satisfied, after a rigorous analysis, that the prerequisites of Rule 23(a) have been satisfied.”).
. Miller v. Mackey Int’l, 452 F.2d 424, 427 (5th Cir.1971).
. Castano v. American Tobacco Co., 84 F.3d 734, 744 (5th Cir.1996).
. Miles, 471 F.3d at 41.
. This is not to say that loss causation, as an element of a 10b-5 claim, cannot be reexamined at summary judgment.
. Nathenson, 267 F.3d at 415. This relationship is foremost an artifact of the common law's influence on 10b-5 actions, yet it persists for good reason. See Schlick v. Penn-Dixie, 507 F.2d 374 (2d Cir.1974); Huddleston v. Herman & MacLean, 640 F.2d 534, 549 (5th Cir.1981), reversed in part Herman & MacLean v. Huddleston, 459 U.S. 375, 103 S.Ct. 683, 74 L.Ed.2d 548 (1983).
. These include "(1) the average weekly trading volume expressed as a percentage of total outstanding shares; (2) the number of securities analysts following and reporting on the stock; (3) the extent to which market makers and arbitrage! ]rs trade in the stock; (4) the company’s eligibility to file SEC registration Form S-3; (5) the existence of empirical facts showing a cause and effect relationship between unexpected corporate events or financial releases and an immediate response in the stock price; (6) the company’s market capitalization; (7) the bid-ask spread for stock sales; and (8) float, the stock's trading volume without counting insider-owned stock.” Bell v. Ascendant Solutions, Inc., 422 F.3d 307, 313 (5th Cir.2005).
.See, e.g., Jonathan R. Macey & Geoffrey P. Miller, Good Finance, Bad Economics: An Analysis of the Fraud-on-the-Market Theory, 42 Stan. L.Rev. 1059, 1083 (1990).
. Greenberg, 364 F.3d at 666.
. A prescient observation indeed.
. Greenberg, 364 F.3d at 666.
. Unger, 401 F.3d at 320-21.