PRECEDENTIAL

         UNITED STATES COURT OF APPEALS
              FOR THE THIRD CIRCUIT
                   _____________

                       No. 15-3888
                      _____________

       JOHN HARNISH; JUSTIN SCHLUTH;
       ROBERT KLEIN; GREGORY EMOND;
 AYLA O’BRIEN KRAVITZ; CHRISTINA MARINAKIS,
                                  Appellants

                             v.

      WIDENER UNIVERSITY SCHOOL OF LAW
                _____________

   On Appeal from the United States District Court for the
                  District of New Jersey
              (D.C. Civil No. 2-12-cv-00608)
         District Judge: Hon. William H. Walls
                     _____________

                    Argued June 6, 2016

   Before: CHAGARES, KRAUSE, and BARRY, Circuit
                     Judges.

                  (Filed: August 16, 2016)

Danielle F. Moriber, Esq.
Rachel E. Simon, Esq.
David S. Stone, Esq. (Argued)
Stone & Magnanini
100 Connell Drive, Suite 2200
Berkeley Heights, NJ 07922
      Attorneys for Appellants

Suna Lee, I, Esq.
Thomas F. Quinn, Esq. (Argued)
Wilson, Elser, Moskowitz, Edelman & Dicker
200 Campus Drive, 4th Floor
Florham Park, NJ 07932

Dennis J. Drasco, Esq.
Lum Drasco & Positan
103 Eisenhower Parkway
Roseland, NJ 07068
      Attorneys for Appellee
                     _______________

                          OPINION
                      _______________

CHAGARES, Circuit Judge.

        This is an interlocutory appeal of a denial of class
certification in a suit alleging that Widener University School
of Law defrauded a putative class of law students by
publishing misleading statistics about its graduates’
employment, which caused the students to pay “inflated”
tuition. The District Court found, among other things, that
the plaintiffs failed to meet the requirement in Rule 23(b)(3)
of the Federal Rules of Civil Procedure that common
questions predominate over individual questions in order for a
class to be certified. We conclude that, although the District
Court labored under a few misconceptions about the
plaintiffs’ theory of the case, the errors were harmless and the
court ultimately reached the correct result. Even when
properly characterized, the plaintiffs’ theory is insufficiently
supported by class-wide evidence, and therefore the plaintiffs
have not established that common questions will
predominate. For that reason, we will affirm.

                               I.

       Named plaintiffs John Harnish, Justin Schluth, Robert
Klein, Gregory Emond, Ayla O’Brien Kravitz, and Christina
Marinakis are graduates of Widener University School of
Law (“Widener”), a private law school with campuses in
Harrisburg, Pennsylvania, and Wilmington, Delaware, who
graduated from Widener between 2008 and 2011. In a
complaint filed in the United States District Court for the

                               2
District of New Jersey on February 1, 2012, and amended on
April 27, 2012, they claim that Widener violated the New
Jersey Consumer Fraud Act (“NJCFA”) and the Delaware
Consumer Fraud Act (“DCFA”) by intentionally publishing
and marketing misleading statistics about the employment of
its graduates.

        Specifically, they allege the following. Between 2005
and 2011, Widener reported that 90-97% of its students were
employed after graduation. These numbers were widely and
deliberately advertised in print and online publications, along
with oral presentations, targeting prospective students. But in
reality, only 50-70% of Widener graduates ended up in full-
time legal positions, which Widener knew. The school was
including non-legal and part-time positions in its published
statistics without reporting the breakdown. When Widener
did provide a breakdown in its materials, it was a breakdown
by employer type (private firm, business and industry, etc.)
within the category of full-time legal employment, further
misleading prospective students into believing that the 90-
97% number represented full-time legal employment.
Beginning in 2011, Widener improved its reporting
somewhat, by including a breakdown that distinguished
between full-time legal positions and other jobs. But,
according to the plaintiffs, Widener continued to gather
information about its graduates in a manner that distorted the
statistics by, for example, crediting unreliable secondhand
accounts of graduates’ employment and avoiding responses
from unemployed graduates.

        The plaintiffs claim that publishing misleading
employment statistics enabled Widener to charge its students
“inflated” tuition — that is, higher tuition than what Widener
would have received if full and accurate statistics were
published instead. Joint Appendix (“J.A.”) 90 (Amended
Compl. ¶ 1). And they seek damages equal to the amount of
tuition that students allegedly overpaid. Widener moved to
dismiss the case, but the motion was denied on March 20,
2013. The parties then engaged in discovery related to class
certification.

       On February 2, 2015, the plaintiffs moved to certify a
class of “[a]ll persons who enrolled in Widener University

                              3
School of Law and were charged full or part-time tuition
within the statutory period for the six-year period prior to the
date the Complaint in this action was filed through the date
that this Class is certified.” J.A. 210. A disputed issue
regarding class certification was whether the plaintiffs could
prove in class-wide fashion that all the class members
suffered damages as a result of Widener’s actions. In
addressing this issue, the plaintiffs introduced a report of
economics expert Dr. Donald Martin. Dr. Martin attested that
he would be able to estimate the extent to which Widener’s
misleading statistics inflated the tuition, which could serve as
a class-wide estimate of every class member’s damages,
insofar as every class member, by definition, paid tuition. In
order to arrive at his estimate, he would perform a regression
analysis of 64 private law schools’ published tuition and
employment statistics and, by controlling for other variables,
compute how much lower Widener’s tuition would be
expected to be if full and accurate employment statistics were
published instead.       Noting that further discovery was
forthcoming and complete data was unavailable, Dr. Martin
did not provide a final estimate of class-wide damages. He
did, however, conclude that there was a statistically
significant relationship between employment rates and tuition
prices across the 64 schools and that his regression
methodology would be a reliable means of arriving at a final
estimate of class-wide damages.

        On July 1, 2015, the District Court denied class
certification on two grounds. First, it found that the plaintiffs
could not meet Federal Rule of Civil Procedure 23(b)(3)’s
requirement that common questions “predominate” over
individual questions because they had “not shown that they
c[ould] prove the proposed class members’ damages by
common evidence.” J.A. 13. The District Court rejected Dr.
Martin’s proposed class-wide method of proving damages,
pointing to the variation in class members’ employment
outcomes: some Widener graduates did obtain full-time legal
jobs, and so their damages, if any, would be different from
those of graduates who did not. The District Court also
concluded that the proposed class-wide theory of damages
relied on a “fraud-on-the-market” theory, which New Jersey
courts had rejected outside the federal securities fraud
context.

                               4
        Second, the District Court found that the plaintiffs
could not meet Rule 23(a)(3)’s requirement that the named
plaintiffs’ claims be “typical” of the claims of the proposed
class. Because the plaintiffs sought to certify a class of
students enrolled “through the date this Class is certified,” the
class would include students who enrolled in 2012 and
beyond, after Widener had improved its reporting. This,
according to the District Court, would render the named
plaintiffs atypical in relation to large portions of the class
because there would be different factual circumstances for the
post-2011 enrollees. It also found that some class members
might even have different interests than the named plaintiffs,
insofar as those pursuing legal careers might prefer not to
have Widener’s reputation tarnished by the lawsuit.

       The plaintiffs filed a timely petition for interlocutory
review under Federal Rule of Civil Procedure 23(f), which we
granted.

                               II.

       The District Court exercised jurisdiction under 28
U.S.C. § 1332. We have appellate jurisdiction under 28
U.S.C. § 1292(e) and Federal Rule of Civil Procedure 23(f).
“We review a class certification order for abuse of discretion,
which occurs if the district court’s decision rests upon a
clearly erroneous finding of fact, an errant conclusion of law
or an improper application of law to fact.” Neale v. Volvo
Cars of N. Am., LLC, 

, 358 (3d Cir. 2015)
(quotation marks omitted).

                              III.

       The plaintiffs raise three challenges to the District
Court’s finding that Rule 23(b)(3)’s predominance
requirement was not met. We will address each challenge in
turn and, for the reasons set forth below, affirm the District
Court’s denial of class certification.

                               A.



                               5
       The plaintiffs’ first argument is that the District Court
applied an improperly burdensome legal standard under Rule
23(b)(3) by scrutinizing their class-wide evidence prior to full
merits discovery and demanding that they “conclusively
prove class-wide damages.” Pls.’ Br. 36. They contend that
the predominance inquiry should be “entirely divorced from
the validity of [their] claims” and that the District Court was
limited to assessing the “viab[ility]” of their theory, not its
“valid[ity].” 

We disagree.

       A plaintiff “may not merely propose a method of
[meeting Rule 23’s requirements] without any evidentiary
support.” Carrera v. Bayer Corp., 

, 306 (3d Cir.
2013). And trial courts “must engage in a rigorous analysis
and find each of Rule 23[ ]’s requirements met by a
preponderance of the evidence before granting certification.”
Hayes v. Wal-Mart Stores, Inc., 

, 358 (3d Cir.
2013) (citations and quotation marks omitted). They must do
so even if it involves judging credibility, weighing evidence,
or deciding issues that overlap with the merits of a plaintiff’s
claims. In re Hydrogen Peroxide Antitrust Litig., 

, 316-25 (3d Cir. 2008).

       We have observed that “[t]he predominance inquiry is
especially dependent upon the merits of a plaintiff’s claim,
since the nature of the evidence that will suffice to resolve a
question determines whether the question is common or
individual.” In re Constar Int’l Inc. Sec. Litig., 

,
780 (3d Cir. 2009) (quotation marks omitted). “[B]efore a
class is certified under [Rule 23(b)(3)], a district court must
find that ‘questions of law or fact common to class members
predominate over any questions affecting only individual
members.’” Tyson Foods, Inc. v. Bouaphakeo, 

, 1045 (2016) (quoting Fed. R. Civ. P. 23(b)(3)). As
defined by the Supreme Court, “[a]n individual question is
one where members of a proposed class will need to present
evidence that varies from member to member, while a
common question is one where the same evidence will suffice
for each member to make a prima facie showing [or] the issue
is susceptible to generalized, class-wide proof.”            

omitted). Rule 23(b)(3) requires that “the
common, aggregation-enabling, issues in the case [be] more


                               6
prevalent or important than the non-common, aggregation-
defeating, individual issues.” 

omitted).

        In determining whether issues that are “susceptible to
generalized, class-wide proof” are “more prevalent or
important,” 

court is called to “formulate some
prediction as to how specific issues will play out . . . in a
given case,” 

. The court cannot
rely on a mere “threshold showing” that a proposed class-
wide method of proof is “plausible in theory.” 

325. Rather, the court’s Rule 23(b)(3) finding necessarily
entails some analysis of whether the proposed class-wide
evidence will actually be sufficient for the class to prevail on
the predominant issues in the case. If class-wide evidence is
lacking, the court cannot be adequately assured that
individualized evidence will not later overwhelm the case and
render it unsuitable for class-wide adjudication. This analysis
will often resemble a merits determination, in that it relates to
plaintiffs’ ability to prove the elements of their claims.

        But the analysis is not a merits determination. First,
much like a court’s preview of the merits of a case when
imposing a preliminary injunction, “the district court’s
findings for the purpose of class certification are conclusive
on that topic” but “do not bind the fact-finder on the merits.”

& n.19. Second, a court should not address merits-
related issues “beyond what is necessary to determine
preliminarily whether certain elements will necessitate
individual or common proof.” Sullivan v. DB Invs., Inc., 

, 305 (3d Cir. 2011) (en banc). In certain situations,
it may be unnecessary to analyze the class-wide evidence as
to every issue in the case in order to reach a conclusion about
Rule 23(b)(3). For example, if the court decides that the
central, predominant issues in the case are common, then
Rule 23(b)(3) is met despite the possibility that some
subsidiary issues will require individualized evidence. 

. Further, evidence as to an issue or
element need not be produced at class certification where the
very nature of the issue or element guarantees that all class
members’ claims will “prevail or fail in unison,” and
therefore there is “no risk whatever that a failure of proof on
the common question . . . will result in individual questions


                               7
predominating.” Amgen Inc. v. Conn. Ret. Plans & Trust
Funds, 

, 1191, 1196 (2013).

       Thus, the task for the District Court was to determine
whether the plaintiffs’ proposed class-wide theories and
evidence would be sufficient to address the predominant
issues in the case. The issues in the case are defined by the
elements of a NJCFA/DCFA claim, which are: “(1) an
unlawful practice, (2) an ascertainable loss, and (3) a causal
relationship between the unlawful conduct and the
ascertainable loss.” Gonzalez v. Wilshire Credit Corp., 

, 1115 (N.J. 2011) (quotation marks omitted);
accord Teamsters Local 237 Welfare Fund v. AstraZeneca
Pharm. LP, 

, 693 (Del. 2016). 1

        The plaintiffs criticize the District Court’s focus on
“damages,” and they invoke the general rule that “individual
damages calculations do not preclude class certification under
Rule 23(b)(3).” 

. But the plaintiffs
gloss over the fact that when courts speak of “damages,” they
are often referring to two distinct concepts: the “fact of
damage” and the measure/amount of damages. The fact of
damage, often synonymous with “injury” or “impact,” is
frequently an element of liability requiring plaintiffs to prove
that they have suffered some harm traceable to the
defendant’s conduct — in other words, the “ascertainable
loss” and “causal relationship” requirements under the
NJCFA and the DCFA. See Newton v. Merrill Lynch, Pierce,
Fenner & Smith, Inc., 

, 187-89 (3d Cir. 2001); In
re Ins. Brokerage Antitrust Litig., 

, 269 (3d Cir.
2009); In re Scrap Metal Antitrust Litig., 

, 535-
36 (6th Cir. 2008); In re New Motor Vehicles Canadian Exp.
Antitrust Litig., 

, 19 n.18, 28 (1st Cir. 2008). Only
if the fact of damage is established does a court reach the
question of remedy and the exact calculation of each
plaintiff’s damages. “While obstacles to calculating damages
may not preclude class certification, the putative class must

       1
         The parties have stipulated that there are no material
differences between New Jersey and Delaware law
concerning the issues in this case. Where appropriate, we
have cited law from both states, but the majority of the
authority comes from the New Jersey courts.
                               8
first demonstrate economic loss” — that is, the fact of
damage — “on a common basis.” 

.

        We are confident that, in scrutinizing the plaintiffs’
proposed class-wide theory and evidence of damages, the
District Court was concerned not merely with the plaintiffs’
ability to calculate the precise measure of damages, but rather
with their ability to demonstrate the fact of damage —
“ascertainable loss” and a “causal relationship” — class-wide.
Ascertainable loss and a causal relationship being core
elements of liability under the NJCFA and the DCFA, it was
entirely appropriate for the District Court to examine the
plaintiffs’ theory of damages and the proof supporting it.

        As to the plaintiffs’ general objection to the level of
scrutiny that the District Court applied to their class-wide
evidence, we see no indication that the District Court applied
the wrong legal standard. For the reasons already elaborated,
closely scrutinizing the plaintiffs’ proposed class-wide
method of proof was the District Court’s duty under Rule 23
and did not, as the plaintiffs argue, transform the court’s
decision into a premature merits determination. Nor did the
District Court purport to be deciding the merits of the case
“conclusively,” as the plaintiffs now assert. The District
Court’s analysis of the evidence was in service of predicting
whether the class-wide proof would ultimately suffice, which
it was required to do. And whatever distinction the plaintiffs
are attempting to draw between “viable” class-wide proof and
“valid” class-wide proof, the law is clear that a class-wide
method of proof must be more than “plausible in theory” and
that a district court is to consider “all relevant evidence and
arguments” in predicting whether the class-wide proof will
suffice. 

. Whether the District
Court reached the correct conclusion after considering the
evidence and arguments is a separate issue that we will
address below.

                              B.

      Next, the plaintiffs argue that the District Court
erroneously attributed significance to the fact that some
Widener graduates do obtain full-time legal employment
(meaning that some class members suffered little, if any,

                              9
damage to their career prospects), effectively ignoring that the
plaintiffs’ theory of damages is unrelated to class members’
actual employment outcomes. In other words, they argue that
the District Court injected an individualized question
(employment outcomes) that has never been at issue because
they claim damages only in the form of overpaid tuition,
which is common to all Widener graduates regardless of their
employment outcomes. While we agree that class members’
own individual employment outcomes are not at issue in this
case, we conclude that the error was harmless because, as
discussed in a later portion of this opinion, the inflated-tuition
theory of damages that is at issue has not been adequately
supported by class-wide evidence, which precludes class
certification under Rule 23(b)(3).

       It is apparent that the plaintiffs’ proposed theory of the
case does not involve an assessment of class members’ own
individual employment outcomes — a point that Widener
appears to concede. Although the amended complaint does
request “damages” in general language, in several places it
frames the damages in terms of “inflated tuition.” J.A. 116-
22. The plaintiffs also made clear in their brief in support of
class certification that inflated tuition was their sole proposed
theory of damages. And they have done so in their briefing
before us.

        Further, in the abstract, we perceive no conceptual
problem with the plaintiffs’ proposed theory. The NJCFA
and the DCFA both contemplate so-called “out-of-pocket”
damages. Under the out-of-pocket rule, a plaintiff’s damages
are “the difference between the price paid and the actual
value of the property acquired.” Romano v. Galaxy Toyota,

, 57 (N.J. Super. Ct. App. Div. 2008); accord
Stephenson v. Capano Dev., Inc., 

, 1076 (Del.
1983). That is exactly what the plaintiffs are seeking. Pls.’
Br. 17-18 (“Plaintiffs seek the difference between the tuition
that students paid and tuition that they would have paid in the
absence of Widener’s misleading marketing . . . .”); 24-25
(“The only loss Plaintiffs seek to recover is the incremental
difference between the amount of tuition that Widener could
have charged but for its fraudulent mass-marketing scheme
and the tuition it actually charged.”).


                               10
        Neither the “price paid” nor the “actual value” depends
on class members’ own individual employment outcomes.
For the price paid, that is self-evident. For the actual value, it
is less obvious. Actual value in a fraud case is generally
“determined as of the time of the transaction.” Kaufman v.
Mellon Nat’l Bank & Trust Co., 

, 331 (3d Cir.
1966) (applying Pennsylvania law); Sands v. Forrest, 

, 124 (Pa. Super. Ct. 1981) (“[I]n an action for fraud
and deceit the measure of damages is the difference in value
between the real, or market, value of the property at the time
of the transaction and the higher, or fictitious, value which the
buyer was induced to pay for it.”). 2 In the fraud context, the
actual value of a degree program therefore does not depend
on a student’s own individual post-graduation employment
outcome, because no one knows at the time of enrollment
what that outcome will be. Some students will ultimately
obtain full-time legal jobs after graduation while others will
not, but at the time of enrollment, the Widener degree offers
them a particular probability of full-time legal employment.
It is the time-of-enrollment probability of full-time legal
employment (which the then-available aggregate employment
statistics help to predict), rather than an individual student’s
own future employment outcome, that affects the true/actual
market value of a Widener education. 3

       The plaintiffs’ theory is therefore that, irrespective of
what ultimately happened to class members after graduation,
the actual value of their Widener education depended in part

       2
          To the extent that the “market” price at the time of
the transaction might itself be inflated due to widespread
dissemination of the misrepresentation, one must ascertain the
fair price that would be paid if the broader market knew the
truth. See Restatement (Second) of Torts § 549 cmt. c.
        3
          As a loose analogy, a lottery ticket’s actual value at
sale does not retroactively plummet to zero the moment a
purchaser loses or skyrocket the moment a purchaser wins. If
the odds were honestly presented, all purchasers received
exactly what they paid for, and neither a loser nor a winner
could claim any damages.            And if the odds were
misrepresented, both players would have the same damages
arising from the fraud — namely, they were overcharged for
placing their bets.
                               11
on the probability of full-time legal employment that they
faced when enrolling. Thus, all class members ended up
paying more than Widener’s actual value because the
published employment statistics (the key indicators of the
probability of full-time legal employment) were misleadingly
optimistic in comparison to the reality of the situation.

       We suspect that when it noted the significance of
varying individual employment outcomes, the District Court
may have had in mind a different theory of damages known
as the “benefit-of-the-bargain” rule, which is also available
under the NJCFA and the DCFA. Under this rule, a plaintiff
can claim damages “equal to that which [the] plaintiff would
have received had the representation been true,” Finderne
Mgmt. Co. v. Barrett, 

, 957 (N.J. Super. Ct.
App. Div. 2008) (quotation marks omitted), sometimes also
referred to as the “replacement cost,” “replacement value,” or
the “diminution” or “loss in value” from the purchaser’s
“expectation interest,” Lee v. Carter-Reed Co., 

,
576 (N.J. 2010); Thiedemann v. Mercedes-Benz USA, LLC,

, 789, 792 (N.J. 2005); Furst v. Einstein
Moomjy, Inc., 

, 440-42 (N.J. 2004); accord

. The District Court noted that
some Widener students “actually got the advertised jobs” —
saying, in effect, that they got what they were promised (a
benefit-of-the-bargain perspective), rather than that they got
something worth what they paid (an out-of-pocket
perspective). J.A. 60. If full-time legal employment post-
graduation was in fact the “advertised” benefit of the bargain,
then employment outcomes would be relevant to benefit-of-
the-bargain damages. But we are skeptical of the notion that
Widener was guaranteeing any particular employment
outcome. More likely, the represented value of a Widener
education to a particular student, like its actual value, is a
function of the probabilistic career-advancing potential of the
education (as claimed, in the case of represented value, versus
as truly existing, in the case of actual value), irrespective of
whether that individual student ultimately realizes the
potential.

        In any event, we do not understand the plaintiffs to be
seeking benefit-of-the-bargain damages.        Although the
plaintiffs have not explicitly invoked the out-of-pocket rule

                              12
rather than the benefit-of-the-bargain rule and have cited
cases involving both rules, they describe the damages they
seek in out-of-pocket terms. So even if class members’ own
individual employment outcomes would be relevant to
determining the represented value of the bargain, they are not
relevant to this case.

       In this respect, the District Court appears to have
mischaracterized the plaintiffs’ theory of damages and
thereby incorporated an individualized issue that was
irrelevant to the case before it. Further, we cannot agree with
Widener that the District Court’s discussion of class
members’ differing employment outcomes was a mere aside
that had no impact on the court’s analysis. The District Court
clearly attributed some significance to the individual
employment outcomes.

       We conclude, however, that the mischaracterization
was harmless because the District Court would very likely
have reached the same decision regarding Rule 23(b)(3)’s
predominance requirement despite the mischaracterization.
Eliminating any focus on class members’ actual employment
outcomes would have removed one individualized inquiry.
But even when the plaintiffs’ inflated-tuition theory is
properly understood, the crucial overarching issue in the case
— proving that each class member suffered damages as a
result of Widener’s actions — still must be shown to be
susceptible to class-wide proof. We now turn to that issue. 4

                               C.

       Finally, the plaintiffs argue that the District Court
erred in equating their theory with the non-cognizable,

       4
          The plaintiffs contend that the District Court not only
mischaracterized their theory of damages but also, in doing
so, violated the law-of-the-case doctrine, because the court, in
its earlier decision denying Widener’s motion to dismiss, had
deemed class members’ own individual employment
outcomes to be irrelevant to the plaintiffs’ case. Having
already concluded that the District Court’s discussion of class
members’ employment outcomes was erroneous but harmless,
we have no need to address this argument.
                               13
reliance-based “fraud-on-the-market” theory, and contend
that they have sufficient class-wide evidence to support a
non-reliance-based inflated-tuition theory. The plaintiffs are
correct that “fraud on the market” is not quite the proper label
for their theory, and correct that they are free to pursue non-
reliance-based theories of damages. But among non-reliance-
based theories, the plaintiffs’ chosen inflated-tuition theory
— whatever might be said about its plausibility — belongs to
the “price-inflation” species that, like the fraud-on-the-market
theory, has been rejected by the New Jersey and Delaware
courts outside the federal securities fraud context. We
therefore reach the same conclusion as the District Court,
despite employing slightly different terminology:            the
plaintiffs fail to meet the predominance requirement of Rule
23(b)(3) because the only class-wide evidence of damages
that they offer supports a non-cognizable theory.

        As discussed above, the plaintiffs appear to be
pursuing out-of-pocket damages, claiming that Widener’s
misrepresentations caused them to pay more for their
education than it was truly worth. Under this theory, we look
to the injuries that resulted from the defendant’s having made
the misrepresentation in the first place, and the goal is to
return the plaintiffs to the position that they would have
occupied if the misrepresentation had never been made. See,
e.g., 

; Restatement (Second) of
Torts § 549 cmt. g (1977); Michael B. Kelly, The Phantom
Reliance Interest in Tort Damages, 38 SAN DIEGO L. REV.
169, 171-76 (2001). 5 In an ordinary fraud case, this would

       5
           A benefit-of-the-bargain claim, by contrast, is
contract-like. We look to the injuries that resulted from the
defendant’s having not lived up to the misrepresentation, and
the goal is to place the plaintiffs in the position that they
would occupy if the misrepresentation were true. See 

; 

; Smajlaj
v. Campbell Soup Co., 

, 101-02 (D.N.J.
2011); 

. We do not interpret the
plaintiffs’ argument as a benefit-of-the-bargain claim because
the success of such a claim would not depend on proving that
the misrepresentation caused higher purchase prices. A
benefit-of-the-bargain class action logically does not entail
proving that all class members were induced to pay extra (a
                              14
require the plaintiffs to prove that the misrepresentation
entered their decision-making and induced them to pay more
for something than they would have otherwise — in other
words, prove reliance. But the plaintiffs do not purport to
have class-wide proof of reliance, in the traditional sense, on
the part of every single class member. Nor could they;
reliance is nearly always an individualized question, requiring
case-by-case determinations of what effect, if any, the
misrepresentation had on plaintiffs’ decision-making. 6

reliance-based theory) or even that the defendant was
empowered to charge them all extra (the price-inflation
theory that the plaintiffs have unsuccessfully pursued in this
case, discussed infra). Instead, it entails proving that class
members all reasonably expected more from the bargain than
what they received. Marcus v. BMW of N. Am., LLC, 

, 607-08 (3d Cir. 2012); 

.
        6
          For example, where “a plaintiff cannot invoke the
fraud-on-the-market presumption” in a securities fraud case,
“[s]he can . . . attempt to establish reliance through the
traditional mode of demonstrating that she was personally
aware of [the defendant’s] statement and engaged in a
relevant transaction . . . based on that specific
misrepresentation,” but “[i]ndividualized reliance issues
would predominate in such a lawsuit.” 

(quotation marks omitted).
        Without the aid of the broad presumption afforded by
the fraud-on-the-market theory, plaintiffs will rarely be able
to prove in class-wide fashion that all class members relied on
misrepresentations and were induced to pay more for
something than they would have otherwise. In Lee, the New
Jersey Supreme Court recognized that, where there are
extraordinary facts involving a “worthless product” about
which “all the representations . . . are baseless,” “a trier of
fact may fairly infer that a[ll] consumer[s] purchasing the
product w[ere] influenced, in some way or other, by the false-
marketing scheme.” 

(emphasis added).
But the court also recognized that, in the absence of such a
situation, determining whether each “purchaser bought the
product based on a fictional or real benefit” — here, the
misleading employment statistics versus Widener’s many
other real attributes — remains “a perplexing problem, the
                              15
        The District Court appears to have believed that the
plaintiffs were attempting to prove reliance class-wide by
way of the “fraud-on-the-market” presumption. It likely
believed so because the goal of Dr. Martin’s analysis was to
prove that the market that determines law school tuition
prices is an “efficient” market, meaning a market in which
price responds to publicly available information about the
value of the product. But an “efficient-market” theory and a
“fraud-on-the-market” theory are not the same, even though,
as shorthand, courts sometimes use the terms interchangeably.

        The connection between the two terms comes from the
federal securities fraud context, where courts will often find
an efficient market to exist, in which “information important
to reasonable investors . . . is immediately incorporated into
stock prices.” In re Burlington Coat Factory Sec. Litig., 

, 1425 (3d Cir. 1997). Once a securities market is
found to be efficient, the fraud-on-the-market theory employs
the efficient-market finding as the “intellectual underpinning”
for why individualized proof of reliance is not required.
Kaufman v. i-Stat Corp., 

, 1198 (N.J. 2000).
The theory “‘is based on the hypothesis that, in an open and
developed securities market, . . . [m]isleading statements will
. . . defraud purchasers of stock even if the purchasers do not
directly rely on the misstatements.’” Basic Inc. v. Levinson,

, 241-42 (1988) (quoting Peil v. Speiser, 

, 1160 (3d Cir. 1986)). As we have explained in
greater detail:

      The fraud-on-the-market theory creates a
      threefold presumption of indirect reliance.
      First, [per an efficient-market finding,] this
      court presumes that the misrepresentation
      affected the market price. Second, it presumes
      that a purchaser did in fact rely on the price of
      the stock as indicative of its value. Third, it
      presumes the reasonableness of that reliance.
      All of these presumptions are necessary to
      establish actual reliance. The first presumption
      is necessary to find that a misrepresentation was

resolution of which would depend on a number of individual
inquiries.” 

                              16
       in fact made to the purchaser.            Thus, if
       defendant rebuts this presumption by showing
       that the market did not respond to the
       misrepresentation, it does no more than show
       that the market price was not misrepresentative,
       and thus that no misrepresentation was made to
       the purchaser of the stock.          The second
       presumption is necessary for a court to find that
       the plaintiff did in fact rely on the
       misrepresentation. Thus, a defendant may rebut
       this presumption by showing that the plaintiff
       would have purchased even if he had known
       about the misrepresentation.           The final
       presumption, that reliance on the market price is
       reasonable, may be rebutted by showing that the
       plaintiff knows that a representation is false.

Zlotnick v. TIE Commc’ns, 

, 822 (3d Cir. 1988)
(emphasis added); see also 

.

        Plaintiffs who claim that a market is efficient,
accordingly, may try to invoke the fraud-on-the-market
presumption, because an efficient market is a precondition of
the fraud-on-the-market theory.           But some plaintiffs,
including the plaintiffs in this case, may have other reasons
for claiming that a market is efficient. 7 Although the District
Court was correct that the New Jersey courts have declined to
extend the fraud-on-the-market presumption beyond the
federal securities fraud context, 

-
01, that was beside the point because the plaintiffs have not


       7
        At one point in their briefing, the plaintiffs muddy
the waters by appearing to disclaim even an efficient-market
theory. They argue that “the inquiry is not whether
Widener’s tuition responded to employment data, but what
Widener’s tuition would have been absent its fraudulent
marketing scheme.” Reply Br. 17 n.8. But that sentence is
incoherent and incompatible with the thrust of their argument.
They cannot claim that tuition would have been different
“absent [a] fraudulent marketing scheme” involving
employment data unless they also believe that “tuition
respond[s] to employment data.”
                              17
quite invoked the fraud-on-the-market presumption. 8 The
plaintiffs do not and need not argue the fraud-on-the-market
theory because, as they rightly point out, the NJCFA and the
DCFA do not require reliance on a misrepresentation in order
to establish a “causal relationship” between that
misrepresentation and an “ascertainable loss,” and other non-
reliance forms of “causal relationship” (under a “proximate
cause” standard) are permissible. Marcus v. BMW of N.
Am., LLC, 

, 606 (3d Cir. 2012); 

, 580; 

. See generally
supra note 5 (discussing benefit-of-the-bargain claims); John
C. P. Goldberg et al., The Place of Reliance in Fraud, 48
ARIZ. L. REV. 1001, 1004-14 (2006) (discussing examples of
how causation can still exist when reliance is lacking). What
the plaintiffs seek to establish is the existence of an efficient
market for law school tuition — which is a prerequisite to,
rather than an element established by, a fraud-on-the-market
argument.

        The plaintiffs have placed market efficiency at issue
by citing Dr. Martin’s analysis suggesting that tuition at
Widener and elsewhere responds to public information,
including employment statistics. In that respect, the District
Court’s analysis was correct. But the plaintiffs do so not to
benefit from a presumption of reliance but rather for the
purpose of supporting their theory of “price inflation.” Given
their out-of-pocket damages claim, the plaintiffs must prove
that Widener’s publication of misleading employment
statistics worsened all class members’ positions, by causing
them to pay more for something than it was worth. The
existence of an efficient market, they argue, would permit
them to prove that all class members’ positions were
worsened by the publication of misleading employment
statistics because Widener, as an efficient-market actor
responding to those statistics, charged everyone higher
tuition, regardless of whether the statistics impacted each
individual class member’s decision-making as a consumer.

       8
         As with the District Court’s mischaracterization of
the relevance of class members’ own individual employment
outcomes, we are confident that the misapplied “fraud-on-the-
market” label was harmless and that the District Court’s
concerns about the plaintiffs’ class-wide proof were valid.
                               18
That is, rather than proving that the misrepresentations
induced each class member to pay more, they propose to
prove that the misrepresentations empowered Widener to
charge more across the entire market.

        There is some plausibility to this theory, insofar as law
schools operate in a largely fixed-price market, not an
auction-style market with individually matched asks and bids.
Widener does not ask around to determine the highest price
that it can charge each prospective student. One would
imagine that Widener guesses the wisest across-the-board
tuition to charge based on a reading of the market and a self-
assessment of how prospective students, as a whole, perceive
the school, including its employment statistics. 9 It arguably
does not matter that some prospective students might be
entirely unaware of or unconcerned by Widener’s
employment statistics (and, if they were bidders in an
auction-style market, would not change their bids in response
to different employment statistics). The point is that Widener
might be expected to anticipate greater student demand in
response to misleading employment statistics and thus to
charge all students a higher price than it would have if it had
published full and accurate statistics instead.

        The problem for the plaintiffs is that the state courts
have held that the ascertainable-loss and causal-relationship
elements of the NJCFA and the DCFA are not met by the
kind of price-inflation theory discussed above and advanced
by the plaintiffs. In International Union of Operating
Engineers Local No. 68 Welfare Fund v. Merck & Co., 

(N.J. 2007) (per curiam), the New Jersey Supreme
Court rejected the plaintiff’s attempt “to prove only that the
price charged for Vioxx was higher than it should have been
as a result of defendant’s fraudulent marketing campaign, and
. . . thereby to be relieved of the usual requirements [of]
prov[ing] an ascertainable loss.” 

The court’s
“rejection of the theoretical basis for that proof mechanism
remove[d] it as a potential common question entirely,” and
thus individualized questions about how the “diverse group”

       9
         Financial aid packages are, of course, a means of
customizing the price somewhat to respond to individual
students’ price thresholds.
                               19
of class members reacted to the alleged fraud predominated
instead and precluded class certification. 

1088;
see also Dabush v. Mercedes-Benz USA, LLC, 

, 1121 (N.J. Super. Ct. App. Div. 2005); N.J. Citizen
Action v. Schering-Plough Corp., 

, 178-79 (N.J.
Super. Ct. App. Div. 2003). Similarly, in AstraZeneca, the
Delaware Supreme Court rejected the plaintiffs’ attempt to
establish damages based on the mere fact that they “would
have paid a lower ‘market’ price if all [market] participants
had been fully informed.” 

.
What mattered instead was that the plaintiffs themselves were
fully informed that the advertising was misleading. 

courts, like the District Court in this case,
have emphasized that recognizing “price inflation” as a
“cause” of “ascertainable loss” is essentially the same as
extending the fraud-on-the-market presumption to all
consumer-fraud cases. The practical effect of both theories is
indeed the same, and both depend on the existence of an
efficient market. There is a conceptual difference between a
“fraud-on-the-market” theory and a “price-inflation” theory
— the former deems a price to be a representation and
presumes that buyers relied on it, whereas the latter sidesteps
reliance altogether. The distinction, however, is immaterial
because the state courts have refused to recognize either
theory outside the federal securities fraud context. 10

       10
            Even if a price-inflation theory were cognizable
under state law, the plaintiffs would still be required to do
more than propose it as an economically plausible theory;
they would need to provide proof that price inflation actually
occurred on this occasion, as a result of the specific
misrepresentation at issue. We have serious doubts about
whether they could do so. They offer no direct evidence that
Widener changed its prices in response to the employment
statistics that it published and their anticipated effect on the
overall market. It appears that the plaintiffs did at one point
seek discovery of “[a]ny information related to Widener’s
strategies, methodologies, policies, and procedures regarding
the setting of Widener’s tuition prices,” to which Widener
objected. D.C. Dkt. No. 59, at 10, 12. But it is the plaintiffs’
burden to build a record to support class certification, and
they did not pursue the discovery issue below or on appeal.
                              20
       The plaintiffs have therefore failed to propose a
cognizable theory of damages that is sufficiently supported by
class-wide evidence. And because the fact of damages (an
“ascertainable loss” having a “causal relationship” with
Widener’s conduct) is a crucial issue in the case, the inability
to resolve it in class-wide fashion will cause individual
questions to predominate over common ones, which
precludes class certification. 11




        All we have is Dr. Martin’s analysis of data from 64
private law schools and his preliminary estimate of the dollar
amount by which “average tuition costs” rose across that 64-
school sample for each “percentage point increase in reported
employment.” J.A. 341. The plaintiffs would therefore have
the fact-finder infer that because tuition prices across 64
private law schools tend to vary with employment statistics,
Widener’s tuition price was influenced by its employment
statistics on this occasion. Cf. 

,
1048 (observing that “statistical evidence . . . is a permissible
method of proving classwide liability . . . [if] each class
member could have relied on th[e] sample to establish
liability if he or she had brought an individual action” but that
statistical evidence would not suffice if the class members
“were not similarly situated”).
        According to Dr. Martin, his regression equation
represents the best possible fit for all the data from 64
schools, and we have no reason to think otherwise. But that
strikes us as a separate question from whether the equation
represents a good fit for Widener. For all we know after
consulting Dr. Martin’s rather brief analysis, Widener could
be an outlier. In any event, the rejection of the price-inflation
theory by controlling authority makes it unnecessary for us to
decide what a fact-finder could reasonably infer about
Widener’s conduct from Dr. Martin’s analysis of data from
64 private law schools.
        11
            Affirming the District Court as to Rule 23(b)(3)’s
predominance requirement makes it unnecessary for us to
consider the plaintiffs’ challenges to the District Court’s
finding that Rule 23(a)(3)’s typicality requirement was also
not met.
                               21
                           IV.

      For the foregoing reasons, the order of the District
Court will be affirmed.




                           22