Colon et al v. National Collegiate Athletic Assoc.

Filing 128

MEMORANDUM and ORDER signed by Senior District Judge William B. Shubb on 3/10/2025 DENYING 95 Motion to Exclude Expert Testimony; GRANTING 85 Motion to Certify Class. The certified class consists of: All persons who, from 3/17/2019, to 6/30 /2023, worked for an NCAA Division I sports program other than baseball in the position of "volunteer coach," as designated by NCAA Bylaws. Plaintiffs Shannon Ray, Khala Taylor, Peter Robinson, Katherine Sebbane, and Rudy Barajas are appointed as class representatives. The law firms Gustafson Gluek, Kirby McInerney, and Fairmark Partners are appointed as co-lead class counsel. (Deputy Clerk KEZ)

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 1 of 27

1
2
3
4
5
6
7
8

UNITED STATES DISTRICT COURT

9

EASTERN DISTRICT OF CALIFORNIA

10

----oo0oo----

11
12
13
14

SHANNON RAY, KHALA TAYLOR, PETER
ROBINSON, KATHERINE SEBBANE, and
RUDY BARAJAS, Individually and
on Behalf of All Those Similarly
Situated,

17
18
19

MEMORANDUM AND ORDER RE:
PLAINTIFFS’ MOTION FOR CLASS
CERTIFICATION AND DEFENDANT’S
MOTION TO EXCLUDE EXPERT
TESTIMONY

Plaintiffs,

15
16

No. 1:23-cv-00425 WBS CSK

v.
NATIONAL COLLEGIATE ATHLETIC
ASSOCIATION, an unincorporated
association,
Defendant.

20
----oo0oo----

21
22

Plaintiffs Shannon Ray, Khala Taylor, Peter Robinson,

23

Katherine Sebbane, and Rudy Barajas brought this putative class

24

action against defendant National Collegiate Athletic Association

25

(“NCAA”), alleging violation of § 1 of the Sherman Antitrust Act,

26

15 U.S.C. § 1.

27

Plaintiffs have moved for class certification.

28

(“Class Cert. Mot.”).)

(Second Amended Compl. (Docket No. 84) (“SAC”).)
(Docket No. 85

Defendant opposes the motion (Docket No.
1

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 2 of 27

1

94) and moves to exclude plaintiff’s expert evidence (Docket No.

2

95 (“Daubert Mot.”)).

3

I.

Factual and Procedural Background

4

The NCAA is an association whose members are colleges

5

and universities competing in intercollegiate athletics.

(See

6

Pl. Ex. 7 (Docket No. 85-10); Expert Report of Orley Ashenfelter

7

(“Ashenfelter Rep.”) (Docket No. 113-2) ¶ 16; Expert Report of

8

Jee-Yeon K. Lehmann (“Lehmann Rep.”) (Docket Nos. 119-2, 122-2) ¶

9

21.)

10

schools.

The NCAA governs student athletic competition at its member

11

(See id.)
NCAA schools are divided into three divisions: Division

12

I, Division II, and Division III.

13

which are at issue in this litigation, generally “manage the

14

largest athletic budgets and offer the highest number of

15

athletics scholarships.”

16

largest athletics expense for NCAA Division I schools.

17

(Ashenfelter Rep. ¶ 19.)

18

(Id.)

(See id.)

Division I schools,

Coach compensation is the

NCAA bylaws limit the number of coaches that Division I

19

schools can hire in a given sport.

20

Ashenfelter Rep. ¶ 26.)

21

than basketball and men’s bowl-division football were permitted

22

to hire a certain number of “unrestricted coaches,” who had no

23

restrictions on compensation, plus one or two “volunteer

24

coaches.”1

25

The bylaw at issue here, NCAA Bylaw 11.01.06 (hereinafter

26
27
28

(Lehmann Rep. ¶ 24;

Prior to 2023, Division I programs other

(See Lehmann Rep. ¶ 27; Ashenfelter Rep. ¶¶ 26-28.)

Most single-gender sports programs were permitted to
hire one volunteer coach, while most combined-gender programs
were permitted to hire two volunteer coaches. (Lehmann Rep. ¶
27.)
2
1

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 3 of 27

1

“Volunteer Coach Bylaw” or “the Bylaw”), defined a “volunteer

2

coach” as “any coach who does not receive compensation or

3

remuneration” from the school’s athletics department.

4

Docket No. 85-12 at 62; Lehmann Rep. ¶ 28.)2

5

(See

Following the repeal of the Volunteer Coach Bylaw,

6

effective July 2023, the volunteer coach designation was

7

eliminated and the number of unrestricted coaches was increased,

8

typically by the number of volunteer coaches allowed under the

9

prior rule.

10

previously permitted one volunteer coach were allotted one

11

additional paid coach.

(Ashenfelter Rep. ¶ 29.)

For instance, programs

(See Lehmann Rep. ¶ 29.)

12

Plaintiffs brought this putative class action alleging

13

that the Volunteer Coach Bylaw violated § 1 of the Sherman Act.

14

The proposed class consists of “[a]ll persons who, from March 17,

15

2019, to June 30, 2023, worked for an NCAA Division I sports

16

program other than baseball3 in the position of ‘volunteer

17

coach,’ as designated by NCAA Bylaws.”

18

II.

(SAC ¶ 19.)

Plaintiffs’ Expert Report

19

Plaintiffs’ motion for class certification relies

20

primarily on an expert report authored by Dr. Orley Ashenfelter.

21

(Ashenfelter Rep.)

22

declaration from Dr. Ashenfelter that provides additional

Plaintiffs have also provided a supplemental

23
24
25
26
27
28

Volunteer coaches were allowed to receive certain
benefits from schools, for example tickets to home games, meals
during team events, and compensation for working at sports camps
and clinics. (Lehmann Rep. ¶ 28.)
2

The related case Smart v. NCAA, a parallel class action
representing baseball coaches, recently settled. (See 2:22-cv02125, Docket No. 70.)
3
3

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 4 of 27

1

explanation of his methodology and updates based on additional

2

data.

3

Defendant seeks to exclude all evidence from this expert, as

4

discussed below.

(Ashenfelter Suppl. Decl. (Docket Nos. 115-2, 121-1).)4

5

Dr. Ashenfelter is an emeritus professor of economics

6

at Princeton University and has extensive experience and

7

professional qualifications in the area of labor economics.

8

App. A to Ashenfelter Rep. (Docket No. 85-4 at 49-79).)

9

support of plaintiffs’ motion for class certification, Dr.

10

Ashenfelter created a statistical model to estimate the damages

11

suffered by the members of the proposed class.

12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

(See

In

To formulate his model, Dr. Ashenfelter relied upon
The supplemental declaration was provided as an exhibit
to plaintiffs’ opposition to defendant’s Daubert motion to
exclude Dr. Ashenfelter’s testimony. Defendant filed an
evidentiary objection in which it argues that the court should
not rely upon the supplemental declaration in ruling on class
certification, instead limiting the court’s consideration of the
new material to its ruling under Daubert. (See Docket No. 104.)
Defendant argues that it would be unfair for the court to rely
upon the supplemental declaration because defendant has not been
given a chance to respond to it in its class certification
briefing, as the declaration was filed following defendant’s
filing of its opposition to class certification. Alternatively,
defendant seeks leave to file an additional brief in opposition
to the motion for class certification addressing the supplemental
declaration. (See id.)
Contrary to defendant’s objection, defendant has had a
chance to address Dr. Ashenfelter’s supplemental declaration in
its reply brief in support of its Daubert motion, and indeed has
done so at length. (See Docket No. 111.) Defendant has also
deposed Dr. Ashenfelter concerning his supplemental declaration.
(See id. at 2 n.1.) Further, the supplemental declaration does
not change the underlying methodology or reasoning plaintiff
relies upon in arguing the class certification requirements are
met. Because defendant has had a fair opportunity to respond,
the court may rely on the supplemental Ashenfelter declaration in
ruling on both the Daubert and class certification issues.
Defendant’s objection (Docket No. 104) is therefore OVERRULED.
4
4

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 5 of 27

1

wage data and other documentation from hundreds of NCAA Division

2

I schools, focusing on those that expanded their coaching staff

3

beyond the prior limits on the number of unrestricted coaches

4

following the repeal of the Volunteer Coach Bylaw.

5

Ashenfelter Rep. ¶ 61; Ashenfelter Suppl. Decl. ¶¶ 11, 21.)

6

focuses on this subset of schools because they “provide the best

7

currently-available evidence of what a competitive market will

8

look like” in the absence of the repealed Bylaw.

9

Suppl. Decl. ¶ 21.)

10

following the Bylaw repeal as a “benchmark” to estimate the “but-

11

for” compensation class members would have received.

12

Ashenfelter Rep. ¶ 40.)

13

practice in antitrust cases of calculating classwide damages

14

based on what class members’ economic position would have been

15

absent the alleged antitrust violations (i.e., in the world that

16

would have existed but for the alleged violation).

17

Corp. v. Behrend, 569 U.S. 27, 36 (2013); ABA Section of

18

Antitrust Law, Proving Antitrust Damages: Legal and Economic

19

Issues § II.4.B (2d ed. 2010).

(See
He

(Ashenfelter

The model uses actual coach salary data

(See

“But-for” analysis refers to the

See Comcast

20

Dr. Ashenfelter’s analysis proceeds in two steps.

In

21

the first step, Dr. Ashenfelter categorizes sports programs

22

according to how many unrestricted coaches each program was

23

permitted to have under NCAA rules beginning July 1, 2023 (i.e.,

24

following the repeal of the Bylaw).

25

ranks coaches within each “program” (each sport within each

26

school, broken down by gender if applicable) according to their

27

actual annual pay.

(Ashenfelter Rep. ¶ 66.)

He

(See id. ¶ 67; Ashenfelter Suppl. Decl. ¶ 22

28
5

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 6 of 27

1

n.39.)

He then employs a regression analysis5 to calculate the

2

“step-down” -- i.e., degree of difference -- in pay between the

3

lowest-paid and second-lowest-paid coaches.

4

67-68.)

5

available data that for sports with a three-coach limit (for

6

instance tennis), the lowest-paid coach received pay 45% lower

7

than that of the second-lowest-paid coach.

(Ashenfelter Rep. ¶¶

For example, the model concluded based on currently

(See id. ¶ 68.)

8

In the second step, Dr. Ashenfelter produces an

9

estimate of the compensation class members would have received in

10

the “but-for” world.

11

school, the model uses the step-down differential identified at

12

step one to calculate a salary value one or more steps lower than

13

the lowest-paid coach.

14

above, the but-for compensation of a volunteer tennis coach based

15

on one “step” down would be 45% lower than the salary of the

16

lowest-paid coach.

17

varies based on school-specific factors for a given sport.

18

id.)

19

given class member based on the step-down level and actual salary

20

data associated with the sports program that employed him or her.

21

III. Defendant’s Daubert Motion

(See id. ¶ 70.)

Within each sport at each

(See id. ¶ 71.)

So, in the example

The number of steps down that are applied
(See

The model determines the damages allegedly suffered by a

22

Defendant seeks to exclude the expert report of Dr.

23

Ashenfelter pursuant to Daubert v. Merrell Dow Pharmaceuticals,

24

Inc., 509 U.S. 579, 580 (1993).

25

inquiry focused ‘solely on principles and methodology, not on the

26
27
28

Daubert requires “a flexible

A regression analysis models the relationship between
the target dependent variable -- here, coach salary -- and one or
more independent variables. See Proving Antitrust Damages §
II.6.C.1.
6
5

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 7 of 27

1

conclusions that they generate.’”

2

F.3d 1147, 1153 (9th Cir. 2004) (quoting Daubert, 509 U.S. at

3

595).

4

junk science that does not meet Federal Rule of Evidence 702’s

5

reliability standards by making a preliminary determination that

6

the expert’s testimony is reliable.”

7

Corp., 657 F.3d 970, 982 (9th Cir. 2011) (citing Kumho Tire Co.

8

v. Carmichael, 526 U.S. 137, 145, 147–49 (1999)).

9

not require a court to admit or to exclude evidence based on its

10

persuasiveness; rather it requires a court to admit or exclude

11

evidence based on its scientific reliability and relevance.”

12

Id.; see also Primiano v. Cook, 598 F.3d 558, 564 (9th Cir. 2010)

13

(“Shaky but admissible evidence is to be attacked by cross

14

examination, contrary evidence, and attention to the burden of

15

proof, not exclusion.”).

16

United States v. Prime, 431

“[T]he trial court must act as a ‘gatekeeper’ to exclude

Ellis v. Costco Wholesale

“Daubert does

“The manner and extent to which the Daubert framework

17

applies at the class certification stage is an unsettled

18

question.”

19

(9th Cir. 2024) (collecting cases).

20

explained in Lytle that at class certification, where the

21

plaintiff’s expert is relied upon for purposes of the

22

predominance inquiry under Rule 23, “such Daubert factors as peer

23

review of the proffered model may be highly relevant, while

24

others, such as known error rate, may be more applicable to the

25

later-executed results of the test.”

26

‘full’ or ‘limited’ Daubert analysis should be applied may depend

27

on the timing of the class certification decision.”

28

“If discovery has closed and an expert’s analysis is complete and

Lytle v. Nutramax Lab’ys, Inc., 114 F.4th 1011, 1030

7

However, the Ninth Circuit

Id.

Further, “whether a

Id. at 1031.

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 8 of 27

1

her tests fully executed, there may be no reason for a district

2

court to delay its assessment of ultimate admissibility at

3

trial.”

4

Id.
But where “an expert’s model has yet to be fully

5

developed, a district court is limited at class certification to

6

making a predictive judgment about how likely it is the expert’s

7

analysis will eventually bear fruit,” and therefore a “full-blown

8

Daubert assessment of the results of the application of the model

9

would be premature.”

10

ongoing and Dr. Ashenfelter is still receiving new data and

11

updating his analysis, which indicates that a full Daubert

12

analysis is “premature” at this stage of the proceedings.

13

id.

14

Id.

In the instant case, discovery is

See

It is undisputed that Dr. Ashenfelter possesses

15

extensive experience and qualifications in the field of labor

16

economics and that he based his analysis on the review of

17

reliable documentation produced by NCAA Division I member

18

schools.

19

“yardstick,” like that employed by Dr. Ashenfelter, is a well-

20

established method of calculating class-wide antitrust impact.

21

See Proving Antitrust Damages § II.4.C.

22

represents that a similar methodology to the one applied here has

23

previously been used to evaluate the class-wide antitrust impact

24

of NCAA coach compensation restrictions.

25

Decl. ¶ 23 n.41 (discussing expert method relied upon in Law v.

26

Nat’l Collegiate Athletic Ass’n, 5 F. Supp. 2d 921 (D. Kan.

27

1998)).)

28

similar statistical analysis in antitrust cases.

Regression analysis based on a “benchmark” or

Dr. Ashenfelter

(See Ashenfelter Suppl.

Further, Dr. Ashenfelter has previously performed

8

See, e.g.,

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 9 of 27

1

Cason-Merenda v. Detroit Med. Ctr., No. 06-15601, 2013 WL

2

1721651, at *1 (E.D. Mich. Apr. 22, 2013) (denying Daubert motion

3

to exclude Dr. Ashenfelter’s “benchmark” analysis of but-for

4

wages in alleged wage-fixing conspiracy).

5

that his evidence is sufficiently reliable at this stage.

6

Lytle, 114 F.4th at 1031 (expert’s “unchallenged credentials,”

7

“review of documentary evidence and . . . data,” use of a “well-

8

established” methodology, and the fact expert had “successfully

9

performed” similar analyses in prior cases established that

10

expert evidence was admissible under Daubert at class

11

certification).

12

These factors indicate
See

Defendant argues that Dr. Ashenfelter’s report is

13

nonetheless inadmissible because it fails to account for several

14

key factors.

15

model fails to control for the experience and skill level of

16

coaches because (1) his calculations did not incorporate

17

experience level as a variable, and (2) he did not address

18

potential selection bias in the sample of additional paid coaches

19

hired after the bylaw repeal, who could have higher experience

20

levels and therefore warrant higher wages.

21

factually unfounded, as Dr. Ashenfelter’s analysis does account

22

for experience using both pay ranking within the coaching

23

hierarchy and age as proxies for experience.

24

Rep. ¶ 71; Ashenfelter Suppl. Decl. ¶¶ 32-35).

25

First, defendant contends that Dr. Ashenfelter’s

These arguments are

(See Ashenfelter

Second, defendant argues that Dr. Ashenfelter “excluded

26

evidence from schools that did not add paid coaching positions

27

after the bylaws were amended.”

28

this argument is unfounded.

(Daubert Mot. at 21.)

Again,

(See Ashenfelter Rep. ¶ 71 (“If
9

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 10 of 27

1

. . . a program reports an unrestricted coach who earns no

2

compensation, then the volunteer coach is estimated to also earn

3

no compensation [under the but-for analysis].

4

is rare: according to my analysis of the schools’ data, more than

5

99% of unrestricted coaches are paid.”).)

However, this case

6

Finally, defendant argues that Dr. Ashenfelter’s

7

analysis is based around groupings of dissimilar sports and

8

“tries to estimate market rates of pay for coaches in one sport

9

by using salaries for coaching in other sports that are

10

determined by different supply and demand conditions.”

11

Mot. at 29.)

12

analysis.

13

step one uses groupings of sports based on how many coaches the

14

NCAA permits a school to hire, the damage calculation at step two

15

uses actual salary data from each sports program at each school

16

and therefore accounts for differences across sports.

17

Ashenfelter Rep. ¶ 71.)

18

(Daubert

This argument mischaracterizes Dr. Ashenfelter’s

While the calculation of the step-down differential at

(See

To the extent that defendant thinks Dr. Ashenfelter’s

19

analysis inadequately accounts for the variables discussed above,

20

that is not a basis for exclusion under Daubert, but rather goes

21

to the weight of the evidence.

22

691, 695 (9th Cir. 2005) (“[O]bjections to a [statistical]

23

study’s completeness generally go to ‘the weight, not the

24

admissibility of the statistical evidence,’ and should be

25

addressed by rebuttal, not exclusion.”) (quoting Mangold v. Cal.

26

Pub. Utils. Comm’n, 67 F.3d 1470, 1476 (9th Cir. 1995)).

27

Defendant has failed to establish that Dr. Ashenfelter’s

28

“methodology is flawed or that there is a likelihood that he will

See Obrey v. Johnson, 400 F.3d

10

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

1

improperly apply that method to the facts.”

2

at 1031.

3

Ashenfelter’s expert report will be denied.6

4

IV.

Page 11 of 27

See Lytle, 114 F.4th

Accordingly, defendant’s motion to exclude Dr.

Class Certification

5

The proposed class consists of “[a]ll persons who, from

6

March 17, 2019, to June 30, 2023, worked for an NCAA Division I

7

sports program other than baseball in the position of ‘volunteer

8

coach,’ as designated by NCAA Bylaws.”

9

(SAC ¶ 19.)

To prevail on class certification, plaintiffs must

10

establish “by a preponderance of the evidence” that the proposed

11

class satisfies the requirements of Federal Rules of Civil

12

Procedure 23(a) and 23(b).

13

v. Bumble Bee Foods LLC, 31 F.4th 651, 664-65 (9th Cir. 2022).

14

Olean Wholesale Grocery Coop., Inc.

“Rule 23 does not set forth a mere pleading standard.”

15

Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 350 (2011).

16

“[C]ertification is proper only if ‘the trial court is satisfied,

17

after a rigorous analysis, that the prerequisites of Rule 23(a)

18

have been satisfied.’”

19

Sw. v. Falcon, 457 U.S. 147, 161 (1982)).

20

be considered to the extent -- but only to the extent -- that

21

they are relevant to determining whether the Rule 23

22

prerequisites for class certification are satisfied.”

23

v. Conn. Ret. Plans & Tr. Funds, 568 U.S. 455, 466 (2013).

24

A.

25

Id. at 350-51 (quoting Gen. Tel. Co. of
“Merits questions may

Amgen Inc.

Rule 23(a)
Rule 23(a) restricts class actions to cases where: “(1)

26
27
28

The court expresses no opinion at this time as to
whether any evidence would be admissible or inadmissible at
trial.
11
6

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 12 of 27

1

the class is so numerous that joinder of all members is

2

impracticable [numerosity]; (2) there are questions of law or

3

fact common to the class [commonality]; (3) the claims or

4

defenses of the representative parties are typical of the claims

5

or defenses of the class [typicality]; and (4) the representative

6

parties will fairly and adequately protect the interests of the

7

class [adequacy of representation].”

8

See Fed. R. Civ. P. 23(a).

Defendant appears to concede that the numerosity,

9

commonality, and typicality requirements are satisfied, as its

10

brief does not address them.

11

factors as part of its “rigorous” analysis.

12

U.S. at 350-51.

The court nonetheless addresses all
See Wal-Mart, 564

13

1.

Numerosity

14

“Although ‘no specific minimum number of plaintiffs

15

asserted’ is required to obtain class certification, ‘a proposed

16

class of at least forty members presumptively satisfies the

17

numerosity requirement.’”

18

422 (E.D. Cal. 2020) (England, J.) (quoting Nguyen v. Radient

19

Pharmaceuticals Corp., 287 F.R.D. 563, 569 (C.D. Cal. 2012)).

20

Here, plaintiffs present evidence that the putative

Alger v. FCA US LLC, 334 F.R.D. 415,

21

class has thousands of members (see Ashenfelter Rep. ¶ 63), which

22

defendant does not dispute.

23

satisfies the numerosity requirement.

The proposed class therefore

24

2.

Commonality

25

Commonality requires that the class members’ claims

26

“depend upon a common contention” that is “capable of classwide

27

resolution -- which means that determination of its truth or

28

falsity will resolve an issue that is central to the validity of
12

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 13 of 27

1

each one of the claims in one stroke.”

2

350.

3

satisfy the rule,” and the “existence of shared legal issues with

4

divergent factual predicates is sufficient, as is a common core

5

of salient facts coupled with disparate legal remedies within the

6

class.”

Hanlon v. Chrysler Corp., 150 F.3d 1011, 1019 (9th Cir.

7

1998).

“So long as there is even a single common question, a

8

would-be class can satisfy the commonality requirement of Rule

9

23(a)(2).”

10

(9th Cir. 2013) (internal citation and quotation marks omitted).

11

Wal-Mart, 564 U.S. at

“[A]ll questions of fact and law need not be common to

Wang v. Chinese Daily News, Inc., 737 F.3d 538, 544

The question of whether the Volunteer Coach Bylaw

12

violated antitrust law is common to the entire class.

“Antitrust

13

liability alone constitutes a common question that will resolve

14

an issue that is central to the validity of each class member’s

15

claim in one stroke, because proof of an alleged conspiracy will

16

focus on defendants’ conduct and not on the conduct of individual

17

class members.”

18

Supp. 2d 1167, 1180 (N.D. Cal. 2013) (internal quotation marks

19

and citations omitted).

20

has been alleged, courts have consistently held that ‘the very

21

nature of a conspiracy antitrust action compels a finding that

22

common questions of law and fact exist.’”

23

(quoting In re TFT–LCD (Flat Panel) Antitrust Litig., 267 F.R.D.

24

583, 593 (N.D. Cal. 2010)).

25

common question applicable to the whole class, they have

26

satisfied the commonality requirement.

In re High-Tech Emp. Antitrust Litig., 985 F.

Thus, “[w]here an antitrust conspiracy

See id. at 1181

Because plaintiffs have identified a

27

3.

Typicality

28

Typicality requires that named plaintiffs have claims
13

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 14 of 27

1

“reasonably coextensive with those of absent class members,” but

2

their claims do not have to be “substantially identical.”

3

Hanlon, 150 F.3d at 1020.

4

other members have the same or similar injury, whether the action

5

is based on conduct which is not unique to the named plaintiffs,

6

and whether other class members have been injured by the same

7

course of conduct.”

8

508 (9th Cir. 1992) (citation omitted).

The test for typicality “is whether

Hanon v. Dataproducts Corp., 976 F.2d 497,

9

Here, each class representative -- like each class

10

member -- worked as a volunteer coach at an NCAA Division I

11

school, was subject to the NCAA’s Volunteer Coach Bylaw

12

precluding them from receiving compensation, and alleges

13

antitrust injury under the Sherman Act.

14

this uniformity of class members’ injuries, claims, and legal

15

theory is typically sufficient to satisfy Rule 23(a)(3).”

16

re NCAA Student-Athlete Name & Likeness Licensing Litig. (“NCAA

17

Name & Likeness Litig.”), No. 09-cv-1967 CW, 2013 WL 5979327, at

18

*5 (N.D. Cal. Nov. 8, 2013) (finding typicality requirement

19

satisfied for class consisting of all Division I men’s football

20

and basketball players subject to an NCAA policy alleged to

21

violate antitrust law).

22

“any unique defenses which threaten to become the focus of the

23

litigation” that would cut against these similarities, see Hanon,

24

976 F.2d at 508, plaintiffs have satisfied the typicality

25

requirement.

26

4.

27

To resolve the question of adequacy, the court must

28

“In antitrust cases,

See In

Because defendant has not identified

Adequacy of Representation

consider two factors: (1) whether the named plaintiffs or their
14

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 15 of 27

1

counsel have any conflicts of interest with other class members,

2

and (2) whether the named plaintiffs and their counsel will

3

vigorously prosecute the action on behalf of the class.

4

Hyundai & Kia Fuel Econ. Litig., 926 F.3d 539, 566 (9th Cir.

5

2019).

6
7

a.

In re

Conflicts of Interest

The first portion of the adequacy inquiry “serves to

8

uncover conflicts of interest between named parties and the class

9

they seek to represent.”

10

Cir. 2023) (quoting Amchem Prods., Inc. v. Windsor, 521 U.S. 591,

11

625 (1997)).

12

interest and suffer[ed] the same [alleged] injury as the class

13

members,” indicating that their interests are “aligned.”

14

Amchem, 521 U.S. at 625–26.

15

Kim v. Allison, 87 F.4th 994, 1000 (9th

Here, the class representatives “possess the same

See

Defendant argues that each class member would need to

16

prove that a given school would have added paid positions for

17

their sport, creating a conflict with other class members who

18

coached for a different sport at the same school.

19

in greater detail below, this argument is premised on a merits-

20

based dispute between the parties’ experts about how but-for

21

damages should be calculated.

22

expert expressly reject defendant’s contention that they will

23

need to prove what hiring decisions would have been made by each

24

school, instead relying on a different method of calculating

25

antitrust injury.

26

presents only a “speculative conflict” that is not “fundamental

27

to the suit.”

28

F.3d 934, 942 (9th Cir. 2015).

As discussed

Further, plaintiffs and their

The issue identified by defendant therefore

See In re Online DVD-Rental Antitrust Litig., 779
Accordingly, there are no
15

Case 1:23-cv-00425-WBS-CSK
1
2
3

Document 128

Filed 03/11/25

Page 16 of 27

conflicts of interest precluding class certification.
b.

Vigorous Prosecution

The second portion of the adequacy inquiry examines the

4

vigor with which the named plaintiffs and their counsel have

5

pursued the class’s claims.

6

standards by which ‘vigor’ can be assayed, considerations include

7

competency of counsel.”

8

150 F.3d at 1021).

9

“Although there are no fixed

Kim, 87 F.4th at 1002 (quoting Hanlon,

Plaintiffs are represented by the firms Gustafson

10

Gluek, Kirby McInerney, and Fairmark Partners.

The extensive

11

experience and strong qualifications of plaintiffs’ counsel in

12

litigating complex antitrust cases, including litigation against

13

the NCAA concerning allegedly anticompetitive restrictions on

14

coach compensation, are undisputed.

15

(Docket No. 85-1); Decl. of Robert Gralewski, Jr. (Docket No. 85-

16

2); Decl. of Michael Lieberman (Docket No. 85-3).)

17

counsel represents that they have expended thousands of hours and

18

considerable resources in litigating this case thus far.

19

Class Cert. Mot. at 19.)

20

plaintiffs’ filings supports this conclusion.

21

no indication that the named plaintiffs will fail to vigorously

22

prosecute this case.

23

(describing named plaintiffs’ efforts to support this litigation,

24

including responding to interrogatories, searching for responsive

25

documents, sitting for depositions, and consulting with counsel

26

about case strategy and discovery).)

27

their counsel satisfy the adequacy requirement.

(See Decl. of Dennis Stewart

Plaintiffs’

(See

The court’s review of the docket and
Further, there is

(See Decl. of Michael Lieberman ¶ 8

28
16

Accordingly, plaintiffs and

Case 1:23-cv-00425-WBS-CSK
1

B.

2

Document 128

Filed 03/11/25

Page 17 of 27

Rule 23(b)
After fulfilling the threshold requirements of Rule

3

23(a), the proposed class must satisfy the requirements of one of

4

the three subdivisions of Rule 23(b).

5

Inc., 716 F.3d 510, 512 (9th Cir. 2013).

6

certification under Rule 23(b)(3), which provides that a class

7

action may be maintained only if the court finds that (1)

8

“questions of law or fact common to class members predominate

9

over questions affecting only individual members,” and (2) “a

10

class action is superior to other available methods for fairly

11

and efficiently adjudicating the controversy.”

12

23(b)(3).

13

is satisfied, but does not address superiority.

Leyva v. Medline Indus.
Plaintiffs seek

Fed. R. Civ. P.

Defendant disputes that the predominance requirement

14

1.

Predominance

15

“The predominance inquiry asks whether the common,

16

aggregation-enabling, issues in the case are more prevalent or

17

important than the non-common, aggregation-defeating, individual

18

issues.”

19

Bouaphakeo, 577 U.S. 442, 453 (2016)).

20

central issues in the action are common to the class and can be

21

said to predominate, the action may be considered proper under

22

Rule 23(b)(3) even though other important matters will have to be

23

tried separately, such as damages or some affirmative defenses

24

peculiar to some individual class members.”

25

U.S. at 453 (cleaned up).

26

Olean, 31 F.4th at 664 (quoting Tyson Foods, Inc. v.
“When one or more of the

Tyson Foods, 577

“‘Considering whether questions of law or fact common

27

to class members predominate begins, of course, with the elements

28

of the underlying cause of action.’”
17

Olean, 31 F.4th at 665

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 18 of 27

1

(quoting Erica P. John Fund, Inc. v. Halliburton Co., 563 U.S.

2

804, 809 (2011)) (cleaned up).

3

of the Sherman Act are “(i) the existence of an antitrust

4

violation; (ii) ‘antitrust injury’ or ‘impact’ flowing from that

5

violation (i.e., the conspiracy); and (iii) measurable damages.”

6

Id. at 666.

7

antitrust laws were intended to prevent and that flows from that

8

which makes defendants’ acts unlawful.”

9

Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 489 (1977)).

10

The elements of a claim under § 1

Antitrust impact is “injury of the type the

Id. (quoting Brunswick

Accordingly, “to prove there is a common question of

11

law or fact that relates to a central issue in an antitrust class

12

action, plaintiffs must establish that ‘essential elements of the

13

cause of action,’ such as the existence of an antitrust violation

14

or antitrust impact, are capable of being established through a

15

common body of evidence, applicable to the whole class.”

16

(quoting In re Hydrogen Peroxide Antitrust Litig., 552 F.3d 305,

17

311 (3d Cir. 2008)).

18

be “capable of answering a common question for the entire class

19

in one stroke” and of “reasonably sustain[ing] a jury verdict in

20

favor of the plaintiffs, even though a jury could still decide

21

that the evidence was not persuasive.”

22

Tyson Foods, 577 U.S. at 453; Halliburton Co. v. Erica P. John

23

Fund, Inc., 573 U.S. 258, 276 (2014)).

24

Id.

In other words, plaintiffs’ evidence must

See id. at 668 (citing

“In determining whether the ‘common question’

25

prerequisite is met, a district court is limited to resolving

26

whether the evidence establishes that a common question is

27

capable of class-wide resolution, not whether the evidence in

28

fact establishes that plaintiffs would win at trial.”
18

Olean, 31

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 19 of 27

1

F.4th at 666-67.

2

overlap with the merits of the plaintiff’s underlying claim,’ the

3

‘merits questions may be considered [only] to the extent [ ] that

4

they are relevant to determining whether the Rule 23

5

prerequisites for class certification are satisfied.’”

6

(quoting Wal-Mart, 564 U.S. at 351; Amgen, 568 U.S. at 466)

7

(alterations in original).

8

“While such an analysis may ‘entail some

Id.

It is undisputed that there are common questions

9

concerning the existence of an antitrust violation.

10

question of whether an antitrust violation under Section 1 exists

11

naturally lends itself to common proof, because that

12

determination ‘turns on defendants’ conduct and intent along with

13

the effect on the market, not on individual class members.’”

14

re Coll. Athlete NIL Litig. (“House”), No. 20-cv-03919 CW, 2023

15

WL 8372787, at *8 (N.D. Cal. Nov. 3, 2023) (quoting In re

16

Glumetza Antitrust Litig., 336 F.R.D. 468, 475 (N.D. Cal. 2020)).

17

See also Law v. Nat’l Collegiate Athletic Ass’n, No. 94-2053-KHV,

18

1998 U.S. Dist. LEXIS 6608, at *15-16 (D. Kan. Apr. 17, 1998)

19

(requirements of Rule 23(b)(3) satisfied where “the NCAA adopted

20

a scheme to fix salaries for restricted earnings coaches . . .

21

the purpose and effect of [which] was to make coaching salaries

22

unresponsive to forces that would normally prevail in a

23

competitive marketplace,” and the “plaintiff class members were

24

employed in the restrained market and . . . subjected to

25

defendant’s illegal scheme”).

26

“The

In

Defendant argues that despite the presence of common

27

questions, individual issues predominate because plaintiffs have

28

not proffered a viable form of common evidence on the issue of
19

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 20 of 27

1

antitrust impact.

2

contends that Dr. Ashenfelter’s model is incapable of providing

3

common proof because it does not address (1) whether each school

4

would have added an additional paid coaching position in the

5

absence of the Bylaw rather than choosing to provide zero pay,

6

and (2) whether each class member would have been hired for that

7

additional paid position.

8

differently, defendant argues that if the Volunteer Coach Bylaw

9

had not been in place, NCAA schools could have nonetheless chosen

10

to provide zero compensation to the additional coaches; and even

11

if they did decide to pay the additional coaches, it is not a

12

given that the proposed class members would have been hired for

13

those positions.

14

effect,” so called because other individuals could have been

15

substituted for the class members in the but-for world.

16

Defendant’s expert, Dr. Jee-Yeon Lehmann,

(See Lehmann Rep. ¶¶ 31, 33, 77.)

Put

Defendant refers to this as the “substitution

Plaintiffs contend that the “substitution effect” is

17

not grounded in accepted economic theory or binding case law and

18

instead, the proper focus in constructing the but-for world is on

19

what competitive wages would have been for plaintiffs’ coaching

20

positions absent the Bylaw.

21

analysis uses the proper framing of the but-for world and that in

22

prior wage-fixing cases he has worked on, he has never been

23

required to show that the class members would also have been

24

hired in the but-for world.

25

n.14.)7

26
27
28

Dr. Ashenfelter avers that his

(See Ashenfelter Suppl. Decl. at 6

Plaintiffs argue that this court already took a
position on the merits of the “substitution theory” in its order
denying defendant’s motion to dismiss. (See Docket No. 38.) The
court did not do so. (See Docket No. 50 (explaining that the
20
7

Case 1:23-cv-00425-WBS-CSK
1

Document 128

Filed 03/11/25

Page 21 of 27

This issue comes down to a merits-based dispute between

2

the parties’ experts concerning the appropriate method for

3

measuring impact.

4

Indeed, some authorities support plaintiffs’ position,8 while

Both positions strike the court as plausible.

5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

court’s order on the motion to dismiss “did no more nor no less
than dispose of the motion which was before the court”).)
See House, 2023 WL 8372787, at *8 (Antitrust “injury
and damages are determined by comparing, on the one hand, the
payments that each class member . . . received in the real world
with, on the other hand, the payments that that same class member
would have received in the but-for world,” and “the identity of
the class members does not change between the real world and the
but-for world . . . Accordingly, the so-called substitutions or
displacements that may or may not take place in a hypothetical
but-for world are irrelevant.”); Law v. Nat’l Collegiate Athletic
Ass’n, 185 F.R.D. 324, 330 n.6 (D. Kan. 1999) (rejecting the
merits of NCAA’s “substitution theory” argument that plaintiffs
suffered no damage because they would not have been hired at all
absent the rule at issue, which “was not anchored in established
case law”); Tawfilis v. Allergan, Inc., No. 8:15-cv-00307 JLS
JCG, 2017 WL 3084275, at *11–12 (C.D. Cal. June 26, 2017) (“[A]n
antitrust impact analysis for direct purchasers need not consider
downstream substitution effects that could have affected the
amount of the product purchased in the but-for world.”); Kamakahi
v. Am. Soc’y for Reprod. Med., 305 F.R.D. 164, 192–93 (N.D. Cal.
2015) (rejecting argument that “substitution theory” defeated
predominance and noting that “[t]o allow the specter of
substitution to defeat class certification, without evidence that
substitution would actually occur, would have wide ranging
effects on the ability to resolve antitrust claims as class
actions”).
Plaintiffs’ position also aligns with authorities
discussing the but-for analysis more generally. See Comcast, 569
U.S. at 36 (After determining “a ‘but for’ baseline -- a figure
that would show what the competitive prices would have been if
there had been no antitrust violations” -- damages are
“determined by comparing to that baseline what the actual prices
were during the charged period.”) (emphasis added); ABA Section
of Antitrust Law, Econometrics: Legal, Practical and Technical
Issues § 13.B.1.c (2d ed. 2014) (“A test of classwide impact
requires the estimation of ‘but-for prices’ (i.e., prices that
would have prevailed but for the alleged anticompetitive act).”)
(emphasis added); Proving Antitrust Damages § II.4.B (“[I]t is
21
8

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 22 of 27

1

others support defendant’s.9

It is not for the court to engage

2

in a “battle of the experts” over the merits at this juncture.

3

See In re NCAA I-A Walk-On Football Players Litig., No. C04-

4

1254C, 2006 WL 1207915, at *11 (W.D. Wash. May 3, 2006)

5

(declining to take a position on the “fundamental difference

6

between Plaintiffs’ expert and the NCAA’s expert” concerning the

7

appropriate “frame” of the but-for analysis, which was a merits

8

issue not suited for consideration at class certification).

9

also Comcast, 569 U.S. at 35 (plaintiffs’ damage model must

10

measure damages attributable to the theory advanced by

11

plaintiffs); Dolphin Tours, Inc. v. Pacifico Creative Serv.,

12

Inc., 773 F.2d 1506, 1512–13 (9th Cir. 1985) (noting

13

“deficiencies” in plaintiff’s damages model which did not

See

14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

not relevant that the defendant . . . could theoretically have
caused the same harms through lawful means,” for instance by
choosing to fix prices individually rather than as part of a
cartel.).
See NCAA Name & Likeness Litig., 2013 WL 5979327, at *8
(crediting the NCAA expert’s “substitution theory” model and
denying class certification because plaintiffs failed to
“provide[] a feasible method for determining which members of the
[proposed class] would still have played for Division I teams -and, thus, suffered the injuries alleged here -- in the absence
of the challenged restraints”); Rock v. Nat’l Collegiate Athletic
Ass’n, No. 1:12-cv-01019 TWP DKL, 2016 WL 1270087, at *14 (S.D.
Ind. Mar. 31, 2016) (denying class certification in challenge to
NCAA rule that limited athletic scholarships because “the facts
do not support [plaintiffs’ expert’s] extreme position that all
members of the [proposed class] would have received a
[scholarship] in the absence of the challenged rules”). See also
Walk-On Football Players Litig., 2006 WL 1207915, at *1 (denying
class certification because plaintiffs failed to provide method
of proving their own theory that the class members would have
received scholarships absent the NCAA rule at issue, but taking
no position on whether plaintiffs’ or the NCAA’s conception of
the but-for world was appropriate).
22
9

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 23 of 27

1

sufficiently address competitive behavior in the but-for world,

2

but reversing grant of summary judgment and allowing the issue of

3

damages to proceed to trial).

4

“Rule 23 grants courts no license to engage in [such]

5

free-ranging merits inquiries at the certification stage.”

See

6

Amgen, 568 U.S. at 466.

7

1067-68 (9th Cir. 2023) (individual issues predominated where

8

court and parties agreed that presence of individual discounts

9

defeated claim for relief, and defendants provided evidence of

10

individual discounts that would “bar recovery,” which raised “the

11

spectre of class-member-by-class-member adjudication of the

12

issue”).

Cf. Van v. LLR, Inc., 61 F.4th 1053,

13

Defendant presents a litany of other critiques of Dr.

14

Ashenfelter’s analysis -- for instance, that it does not account

15

for benefits that class members received by virtue of their

16

volunteer coach positions that could reduce their damages, and

17

does not sufficiently control for variations across different

18

sports and schools in different regions -- arguing that these

19

issues would necessitate individual damage inquiries that would

20

predominate.

21

plaintiffs’ evidence as applied to merits issues.

22

Foods, 577 U.S. at 457 (arguments that an expert study is

23

“unrepresentative or inaccurate” go to the merits and do not

24

defeat class certification).

These critiques similarly speak to the weight of
See Tyson

25

Further, the Ninth Circuit has repeatedly held that

26

individualized damage calculations alone do not defeat class

27

certification.

28

no per se rule that a district court is precluded from certifying

See, e.g., Olean, 31 F.4th at 681–82 (“there is

23

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 24 of 27

1

a class if plaintiffs may have to prove individualized damages at

2

trial”) (citing Halliburton, 573 U.S. at 276); Leyva, 716 F.3d at

3

514 (“the amount of damages is invariably an individual

4

question,” and “the potential existence of individualized damage

5

assessments does not detract from the action’s suitability for

6

class certification”) (quoting Blackie v. Barrack, 524 F.2d 891,

7

905 (9th Cir. 1975); Yokoyama v. Midland Nat. Life Ins. Co., 594

8

F.3d 1087, 1089 (9th Cir. 2010)).

9

that individualized inquiries into damages would predominate over

10

the common issues already identified.

11

679–80 (“While individualized differences among the [actual

12

damages of each class member as compared to the regression

13

model’s estimates] may require a court to determine damages on an

14

individualized basis, such a task would not undermine the

15

regression model’s ability to provide evidence of common

16

impact.”).

17

Defendant has not established

See Olean, 31 F.4th at

As discussed in detail above, Dr. Ashenfelter has

18

provided a model that estimates but-for wages for each proposed

19

class member based on extensive documentation produced by NCAA

20

Division I schools.

21

a benchmark, a widely accepted form of expert evidence, and Dr.

22

Ashenfelter avers that his analysis provides “a reasonable

23

methodology by which to estimate damages using data” and employs

24

“methods that are common to the class.”

25

10.)

26

“capable of showing that the [proposed class] members suffered

27

antitrust impact on a class-wide basis, notwithstanding [Dr.

28

Lehmann’s] critique,” which is “all that [is] necessary at the

His model uses regression analysis based on

(See Ashenfelter Rep. ¶

Plaintiffs have established that Dr. Ashenfelter’s model is

24

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 25 of 27

1

certification stage.”

See Olean, 31 F.4th at 681 (emphasis

2

added); see also id. at 683 (“a regression model . . . may be

3

capable of showing class-wide antitrust impact, provided that the

4

district court considers factors that may undercut the model’s

5

reliability”).

6

common questions of law and fact predominate.

Accordingly, plaintiffs have established that

7

2.

Superiority

8

The second part of the inquiry under Rule 23(b)(3) asks

9

whether “a class action is superior to other available methods

10

for fairly and efficiently adjudicating the controversy.”

11

“Generally, the factors relevant to assessing superiority include

12

‘(A) the class members’ interests in individually controlling the

13

prosecution or defense of separate actions; (B) the extent and

14

nature of any litigation concerning the controversy already begun

15

by or against class members; (C) the desirability or

16

undesirability of concentrating the litigation of the claims in

17

the particular forum; and (D) the likely difficulties in managing

18

a class action.’”

19

F.3d 1168, 1175 (9th Cir. 2010) (quoting Fed. R. Civ. P.

20

23(b)(3)).

21

Wolin v. Jaguar Land Rover N. Am., LLC, 617

The proposed class contains thousands of individuals,

22

and the parties have not identified any competing litigation

23

involving members of the proposed class.

24

that the amount of damages each coach suffered is high enough to

25

make individual litigation an efficient method of resolving their

26

claims, especially given the complexity of antitrust litigation

27

and the presence of several common legal and factual questions.

28

“Forcing individual [class members] to litigate their cases,
25

It appears unlikely

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 26 of 27

1

particularly where common issues predominate for the proposed

2

class,” would be “an inferior method of adjudication.”

3

Wolin, 617 F.3d at 1176.

4

of ‘common issues will reduce litigation costs and promote

5

greater efficiency,’” and the superiority requirement is

6

satisfied.

7

97 F.3d 1227, 1234 (9th Cir. 1996)).

8

Accordingly, “class-wide adjudication

See id. (quoting Valentino v. Carter–Wallace, Inc.,

For the foregoing reasons, the class certification

9

requirements of Rules 23(a) and 23(b)(3) are satisfied.

10

V.

11

See

Appointment of Class Counsel
“An order that certifies a class action . . . must

12

appoint class counsel under Rule 23(g).”

13

23(c)(1)(B).

14

“(i) the work counsel has done in identifying or investigating

15

potential claims in the action; (ii) counsel’s experience in

16

handling class actions, other complex litigation, and the types

17

of claims asserted in the action; (iii) counsel’s knowledge of

18

the applicable law; and (iv) the resources that counsel will

19

commit to representing the class.”

20

discussed above, plaintiffs’ counsel has considerable knowledge

21

and experience in antitrust litigation and has dedicated

22

significant effort and resources to litigating this action.

23

Accordingly, the court will appoint Gustafson Gluek, Kirby

24

McInerney, and Fairmark Partners as co-lead class counsel.

25

Fed. R. Civ. P.

In appointing class counsel, the court considers

Fed. R. Civ. P. 23(g)(1).

As

IT IS THEREFORE ORDERED that defendant’s motion to

26

exclude expert testimony (Docket No. 95) be, and the same hereby

27

is, DENIED.

28

IT IS FURTHER ORDERED that plaintiffs’ motion for class
26

Case 1:23-cv-00425-WBS-CSK

Document 128

Filed 03/11/25

Page 27 of 27

1

certification (Docket No. 85) be, and the same hereby is,

2

GRANTED.

3

March 17, 2019, to June 30, 2023, worked for an NCAA Division I

4

sports program other than baseball in the position of “volunteer

5

coach,” as designated by NCAA Bylaws.

6

The certified class consists of: All persons who, from

Plaintiffs Shannon Ray, Khala Taylor, Peter Robinson,

7

Katherine Sebbane, and Rudy Barajas are hereby appointed as class

8

representatives.

9

and Fairmark Partners are hereby appointed as co-lead class

10

counsel.

11

Dated: March 10, 2025

The law firms Gustafson Gluek, Kirby McInerney,

12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
27