Smith et al v. City of Boston

Filing 245

Judge William G. Young: ORDER entered. MEMORANDUM AND ORDER(Sonnenberg, Elizabeth)

UNITED STATES DISTRICT COURT
DISTRICT OF MASSACHUSETTS

BRUCE SMITH, PAUL JOSEPH, JOHN M.
JOHNSON, ROBERT TINKER, MARTIN
JOSEPH, KIM GADDY, BRIAN KEITH
LATSON, LEIGHTON FACEY, MARWAN
MOSS, and LATEISHA ADAMS,
Plaintiffs,
v.
CITY OF BOSTON,
Defendant.

)
)
)
)
)
)
)
)
)
)
)
)
)
)
)

YOUNG, D.J.

CIVIL ACTION
NO. 12-10291-WGY

July 26, 2017
MEMORANDUM & ORDER

I.

INTRODUCTION
To understand why the Court here revisits and reconsiders

rulings it made earlier in Smith v. City of Boston, 144 F. Supp.
3d 177 (D. Mass. 2015), it is necessary to understand the timing
of my decision in Smith and how that decision may or may not
conform to two other related yet distinct decisions -- Judge
O’Toole’s thorough opinion in Lopez v. City of Lawrence (Lopez
I), No. 07-11693-GAO, 2014 U.S. Dist. LEXIS 124139 (D. Mass.
Sept. 5, 2014) (O’Toole, J.), and its affirmance by the First
Circuit, Lopez v. City of Lawrence (Lopez II), 823 F.3d 102 (1st

Cir. 2016), cert. denied, 137 S. Ct. 1088 (2017).

The latter

decision, of course, controls this Court’s analysis.
All three decisions seek accurately to apply the law of
disparate impact.

At the most superficial level, the

jurisprudence of disparate impact seeks fairly to ensure that
employment decisions are made on genuine merit.
Recognizing that all employment tests are, by their very
nature, discriminatory (after all, that’s the whole purpose of
testing -- to choose the few from the many), the plaintiffs must
(first prong) prove that the test reveals a significantly
disparate impact upon a lawfully protected minority -significantly disparate impact because we don’t want federal
judges messing around with every employment test.
If the plaintiffs prove the first prong, the employer has
the chance (second prong) to prove that the test vindicates
itself through the business necessity of choosing on the basis
of merit the best persons for the job.
Even if the employer prevails on the second prong, the
plaintiffs get one last chance (third prong) -- to prove that
there existed a test equal or better at identifying the best
person for the job thus satisfying the employer’s business
necessity, which test was available to the employer and which
test had a less disparate impact.

[2]

This is an elegant and nuanced matrix.

The devil, of

course, is in the details.
Lopez I, the first of these three related cases, commenced
on September 11, 2007, with the filing of a complaint by a
number of black and Hispanic patrolmen from various
municipalities (including the City of Boston (“Boston”))
challenging the civil service examination procedures for
promotion to the rank of sergeant (“2008 sergeants’ exam”).
Drawn to Judge George O’Toole, this case came on for an
eighteen-day bench trial commencing on July 12, 2010.

When the

trial concluded, Judge O’Toole took the case under advisement.
In February 2012, ten black police sergeants (the
“Plaintiffs”) in Boston commenced a substantially similar case
before Judge Joseph Tauro.

This case, the Smith case,

challenged the police promotional exam from sergeant to
lieutenant.

When Judge Tauro took senior status, the case was

transferred to this session on December 26, 2013.
In the meantime with Lopez I under advisement and Smith
pending, Boston substantially revamped its police promotional
testing procedures, adopting -- at significant expense -- many
of the improvements for which both the Lopez I and Smith
plaintiffs were contending.
On September 5, 2014, Judge O’Toole issued his full written
opinion in Lopez I, finding that the 2008 sergeants’ exam
[3]

imposed a significantly disparate impact on minority applicants,
2014 U.S. Dist. LEXIS 124139, at *48, and that the written
portion of that exam could not alone support its validity
“because it could not measure some skills and abilities (as
distinguished from knowledge) essential to the position, such as
leadership, decision making, interpersonal relations, and the
like,” id. at *60-61.

Judge O’Toole went on to find that the

Education and Experience portion of the examination saved it,
albeit just barely.

Id.

The plaintiffs promptly appealed.

In Smith, the Plaintiffs alleged that the multiple-choice
examination used by the Boston Police Department in 2008 to
select and rank candidates for promotion from the rank of
sergeant to lieutenant (“2008 lieutenants’ exam”) had a
disparate impact on racial minorities and was invalid under
Title VII of the Civil Rights Act of 1964.
3d at 181.

Smith, 144 F. Supp.

Boston responded that the exam did not have a

disparate impact and, even if it did, was sufficiently jobrelated to be held valid.

Id. at 180.

On December 15, 2014, at the outset of what proved to be a
ten-day bench trial, the parties commendably moved into evidence
the full trial record and exhibits from Lopez I.

Then, for ten

days, the Court heard lay and expert witnesses proffered by both
sides, some of whom had not testified in Lopez I.
181.

See id. at

On November 26, 2015, this Court issued its opinion
[4]

concluding that the 2008 lieutenants’ exam had a racially
disparate impact and was insufficiently job-related to survive
the Plaintiffs’ challenge.

Id. at 180-81.

imposed liability on Boston.

The Court thus

Id. at 181.

Before engaging in extensive hearings concerning remedy,
all parties sought time to explore settlement.

After all, the

challenged 2008 lieutenants’ exam had long been out of use and
the real nub of contention appeared to be the attorneys’ fees
due the Plaintiffs’ counsel as prevailing parties.
Then, in a comprehensive opinion issued on May 18, 2016,
the First Circuit affirmed Lopez I.

Lopez II, 823 F.3d 102.

As

that court itself summarized: “[f]inding that the district court
applied the correct rules of law and that its factual findings
were not clearly erroneous, we affirm.”

Id. at 107.

Naturally, I read Lopez II with great interest.

I was

gratified to see that the First Circuit had unanimously
concluded, as did Judge O’Toole -- and as had I with respect to
the 2008 lieutenants’ exam -- that the 2008 sergeants’ exam had
a significantly disparate impact on racial minorities.
111.

Id. at

On the sole issue where I had parted company with Judge

O’Toole -- finding on different and additional evidence that
business necessity could not justify use of the 2008
lieutenants’ examination for the rank ordering of candidates for
promotion -- the Court of Appeals had split 2-1 in reviewing
[5]

Judge O’Toole’s findings as to the 2008 sergeants’ exam.
122 (Torruella, J., concurring and dissenting).

Id. at

Most important,

I detected no shift in the governing law in Lopez II from that I
had applied to the facts I found in Smith.
be expected.

Nor would any shift

Absent intervening Supreme Court precedent or

legislative change, it is the practice in the First Circuit
faithfully to adhere to the decisions of earlier panels of that
court.

See, e.g., Peralta v. Holder, 567 F.3d 31, 35 (1st Cir.

2009) (“‘We have held, time and again, that in a multi-panel
circuit, prior panel decisions are binding upon newly
constituted panels in the absence of supervening authority
sufficient to warrant disregard of established precedent.’”
(quoting Muskat v. United States, 554 F.3d 183, 189 (1st Cir.
2009))).
Whatever the potential legal effect of Lopez II on Smith,
its practical effect was immediate.
ceased.

Settlement negotiations

Now the parties sought an interlocutory appeal to

settle once and for all the propriety of this Court’s ruling on
prong 2.

This Court readily acceded to their wishes.

On October 11, 2016, the Court of Appeals gently but firmly
rebuffed this gambit:
The district court issued its findings on liability in
this case without the benefit of our subsequently-issued
opinion in the case of Lopez v. City of Lawrence, 823 F.3d
102 (1st Cir. 2016). Since then, the district court has
not yet purported to apply Lopez to the facts of this case.
[6]

For example, it has not stated whether and how its
assessment of validity has taken into consideration the
guidance we provided. Id. at 116-17. We therefore deny
the petition without prejudice to renewal, if otherwise
appropriate, after the district court has itself applied
Lopez to this case.
J. United States Ct. Appeals 1, ECF No. 229.
In one sense, this order is both generous and courteous.
It gives me first crack at applying Lopez II to my earlier legal
analysis in Smith and making such analytic adjustments as may be
necessary.

Its tenor, however, suggests I may have missed

something.

Boston certainly thinks so.

In light of the First Circuit’s order, this Court promptly
held a status conference with the parties, Electronic Clerk’s
Notes, ECF No. 232, who subsequently briefed their positions on
the effect of Lopez II on this Court’s previous ruling in Smith,
Pls.’ Br. Ct. Appeal J., ECF No. 235; Pls.’ Reply Br. Regarding
Ct. Appeals J., ECF No 241; City of Boston’s Br. Affect Lopez
Ct.’s Liability Decision (“Def.’s Br.”), ECF No. 236; City of
Boston’s Reply Br. Lopez’s Affect Ct.’s Liability Decision
(“Def.’s Reply”), ECF No. 242.

Boston argues that Lopez II

requires this Court to change its previous ruling by: (1)
applying different legal standards to its prong 2 analysis, thus
necessitating a different outcome, and (2) reaching prong 3 of
the disparate impact inquiry.

[7]

II.

THIS COURT’S PREVIOUS RULING
In these circumstances, this Court first conducts a brief,

albeit rigorous and reflective review of what it has already
done.

In Smith, this Court examined the Plaintiffs’ challenges

to Boston’s use of the 2008 lieutenants’ exam to select and rank
candidates for promotion from the rank of sergeant to
lieutenant.

144 F. Supp. 3d at 180.

The Court imposed

liability on Boston, after concluding that the 2008 lieutenants’
exam had a racially disparate impact and was insufficiently jobrelated to withstand the Court’s disparate impact inquiry.

Id.

at 181.
In examining the evidence, this Court set forth the legal
framework:
Under First Circuit case law, the plaintiff bears the
burden of establishing a prima facie case of discrimination
which consists of identification of an employment practice
(in this case, the 2008 [lieutenants’] exam and promotions
flowing therefrom), disparate impact, and causation.
If the Plaintiff meets this burden, the employer may
either debunk the Plaintiff’s prima facie case, or
alternatively, may demonstrate that the challenged practice
is “job-related and consistent with business necessity.”
If the employer demonstrates the latter, the ball bounces
back into the plaintiff’s court to demonstrate that “some
other practice, without a similarly undesirable side
effect, was available and would have served the defendant’s
legitimate interest equally well.”
. . . .
. . . Under the first prong, the Plaintiffs must make
a significant showing of actual disparate impact upon an
identified protected minority . . . .
If the plaintiff can, however, make this showing, then
under the second prong, the employer gets a chance to

[8]

demonstrate that the test in question is both job-related
and consistent with business necessity. . . .
Even if the employer succeeds, however, the case is
not over. Under the third prong, the plaintiff gets one
more shot. If the plaintiff can demonstrate the
availability of a testing program equally determinative of
job performance, yet resulting in less disparate impact,
the Court should fashion a remedy to secure the greatest
degree of equal opportunity.
Id. at 182-83 (citations omitted).

The Court then issued its

findings of fact, id. at 185-91, thoroughly discussing the role
of a Boston Police Department Lieutenant, id. at 185, pre-2005
job analyses and validation studies, id. at 185-88, the
development and administration of the 2008 lieutenants’ exam,
id. at 188-89, the development and administration of the 2014
exam, id. at 189-90, and the results of the 2005 and 2008
lieutenants’ exams, id. at 190-91.
conclusions of law.
A.

Next the Court set forth its

Id. at 191-211.

Prong 1

In its discussion of disparate impact (prong 1), id. at
191-200, the Court addressed the relevant data to consider in
determining whether the Plaintiffs showed a significant
disparate impact, id. at 192-94, whether or not to aggregate the
data, id. at 194-95, and whether to use a one-tailed or a twotailed test of statistical significance, id. at 195-98.

The

Court concluded that it would consider promotion rates, passfail rates, average scores, and delays in promotion to assess
disparate impact, id. at 194, that it would not aggregate the
[9]

data, id. at 195, and that the Plaintiffs established disparate
impact regardless of the Court’s use of a one or two-tailed
approach, id. at 198.

The Court thus ruled that the Plaintiffs

met their prong 1 burden to raise an inference of causation and
demonstrate a prima facie case of disparate impact.
B.

Id. at 200.

Prong 2

The Court then assessed prong 2 of the disparate impact
framework: job-relatedness and consistency with business
necessity.

Id. at 200-11.

The Court noted that to prevail on

prong 2, “[Boston] must convince the Court that the 2008
[lieutenants’] exam was both ‘job related’ for the position of
Boston Police Department lieutenant and consistent with
‘business necessity.’”

Id. at 200 (quoting Jones v. City of

Boston (Jones I), 752 F.3d 38, 53 (1st Cir. 2014) (quoting 42
U.S.C. § 2000e-2(k)(1)(A)(i))).

The Court cited First Circuit

precedent, stating:
to satisfy this second prong . . . “the [defendant] must
show that its program aims to measure a characteristic that
constitutes an ‘important element of work behavior’”
. . . . [and] “that the outcomes of [its challenged
practice] are ‘predictive or significantly correlated with’
the characteristic described above.”
Id. at 201 (quoting Jones I, 752 F.3d at 54 (quoting Albemarle
Paper Co. v. Moody, 422 U.S. 405, 431 (1975))).

[10]

Noting that the

Uniform Guidelines1 “provide a sensible way of evaluating whether
a given test . . . measures an important work characteristic,
and whether the outcomes of that test are actually correlated
with the characteristic measured,” the Court “look[ed] to the
Uniform Guidelines throughout its prong 2 analysis.”

Id.

Proceeding through its inquiry, this Court first held that
the 2008 lieutenants’ exam measured characteristics that are
important elements of work behavior, id. at 201-02, because the
job analyses on which the test was based “were sufficiently
thorough and current so as to form solid ground on which to
build a valid test.”

Id. at 202.

The Court then addressed if the exam results were
predictive of or correlated with the important work behaviors.
Id. at 202-10.

The Court held that they were not, explaining

that it:
ultimately agrees with [the Plaintiffs’ expert] Dr. Wiesen
that the evidence does not support the necessary inference
that those who perform better on the exam will be better
performers on the job, primarily because the exam did not
test a sufficient range of [knowledge, skills, and
abilities (“KSAs”)], and there was no evidence that the
exam was reliable enough to justify its use for rank
ordering.
1

29 C.F.R. § 1607.4(C)(2). The Court explained that
“Chapter 29 of the Code of Federal Regulations, section 1607,
was published under the name of Uniform Guidelines on Employee
Selection Procedures in 1978 by several government agencies to
interpret how selection and testing and assessment should be
conducted in accordance with the Civil Rights Act of 1964.”
Smith, 144 F. Supp. at 186 n.8 (citing 12/19/14 Tr. 23:6-18, ECF
No. 166).
[11]

Id. at 203.

Boston attempted to show job-relatedness through

content validity, “an attempt to link the important KSAs of the
job with the selection procedure.”

Id.

The Court noted that

the 2008 lieutenants’ exam had two components: a multiple choice
component, worth eighty percent of an examinee’s score, and an
Education and Experience score (“E&E”), worth twenty percent.
Id. at 204.

Relying on Dr. Wiesen’s report, which correlated

candidates’ scores on the multiple choice section of the exam
almost perfectly with their final score, the Court concluded
that the E&E “had virtually ‘no impact on the final exam
scores,’” and thus excluded the E&E from the remainder of the
Court’s validity analysis.

Id. (quoting 12/15/14 Tr. 58:19-21,

ECF No. 151).
The Court then looked to test construction, id. at 204-06,
and “f[ound] that the test construction process was inadequate
to support the heightened validity requirement necessary to rank
candidates,” id. at 206.

In reaching this conclusion, the Court

noted that the test outlines showed that only two abilities
appeared on the 2008 lieutenants’ exam; Boston had sufficiently
evaluated the test questions and their linkage to KSAs, but had
“failed to conduct statistical analyses to ensure the quality of
the test scores for the 2008 [lieutenants’] exam”; and that the
record failed to address whether the test developer properly

[12]

recommended cut-off scores, rankings, bands, and weighting.

Id.

at 205.
Next, the Court addressed the 2008 lieutenants’ exam’s
content.

Id. at 206-08.

Using the Uniform Guidelines’

representative sample test, the Court found that the 2008
lieutenants’ exam tested a sufficient range of the critical
knowledge areas, but that the exam’s near exclusion of any
critical skills and abilities meant that “a high score on the
2008 [lieutenants’] exam simply was not a good indicator that a
candidate would be a good lieutenant.”

Id. at 207-08.

Lastly, the Court discussed the use of the exam results to
rank candidates.

Id. at 208-10.

The Court relied on the

Guidelines’ statements that “‘evidence of both the validity and
utility of a selection procedure should support the method the
user chooses for operational use of the procedure, if that
method of use has a greater adverse impact than another method
of use,’” and that “‘[e]vidence which may be sufficient to
support the use of a selection procedure on a pass/fail
(screening) basis may be insufficient to support the use of the
same procedure on a ranking basis.’”
C.F.R. § 1607.5(G)).

Id. at 208 (quoting 29

Noting that the Uniform Guidelines allow

employers to validate rank-order exams with content validity, by
establishing “‘that a higher score on a content valid selection
procedure is likely to result in better job performance,’” id.
[13]

(quoting 29 C.F.R. § 1607.14(C)(9)), the Court found that Boston
failed to meet this standard because it neither tested a
sufficient range of critical KSAs nor convinced the Court that
the exam was valid, id. at 208-09.

Accordingly, this Court

concluded that Boston failed to show that a higher score on the
2008 lieutenants’ exam would likely result in better job
performance, id. at 210, and “h[eld] that even were the 2008
[lieutenants’] exam valid enough to be used as a screening tool,
[Boston] failed to meet its burden of showing that the 2008
[lieutenants’] exam was sufficiently valid to be used as a basis
for ranking candidates,” id. at 211.

Thus, the Court held that

Boston did not meet its burden on prong 2 of the disparate
impact inquiry, and the Plaintiffs won the case.

Id.

III. ANALYSIS
Boston raises several challenges to this Court’s prong 2
finding, essentially arguing that Lopez II effected changes in
disparate impact law that mandate a reconsideration of this
Court’s previous decision.

Boston also challenges this Court’s

failure to reach prong 3, arguing that the Court could not
reject the exam without the Plaintiffs showing an equally valid,
less discriminatory alternative.

Id. at 12-18.

The Court

concludes that none of Boston’s arguments have merit and upholds
the original decision in Smith.

[14]

A.

2008 Lieutenants’ Exam Validity (Prong 2 Challenges)

Boston argues that this Court ought reconsider its prong 2
analysis because Lopez II: (1) disposes of the Guidelines’
representative sample inquiry in favor of a better than random
selection test, (2) mandates that this Court include the E&E
component in its assessment, and (3) holds that there is not a
heightened validity requirement for Boston to use rank ordering.
Def.’s Reply 5-11.

These arguments are not convincing.

Further, even were this Court to follow Boston’s proposed test,
Smith’s conclusions would lead to the same result.
1.

Reliance on the Guidelines

Boston argues that this Court used the wrong legal standard
by relying on the Guidelines’ representativeness inquiry, rather
than Lopez II’s purported “better than random selection”
standard to determine the exam’s validity.
Reply 4.2

Def.’s Br. 6; Def.’s

The First Circuit certainly did not ban the use of the

2

In advancing this argument, Boston comes close willfully
to misreading Lopez II. Standing alone, the written multiplechoice lieutenants’ examination is pretty clearly better than
random selection yet the hard truth is that Boston’s own experts
recognize that such an examination disfavors minority applicants
and Boston knows it. See Smith, 144 F. Supp. at 197 (citing
12/15/14 Tr. 113-14, ECF No. 161; 12/18/14 Tr. 45-46, ECF No.
165; Lopez I, 07/13/10 Tr. 82-85; Lopez I, 07/14/10 Tr. 43-48,
55, 59-60; Lopez I, 07/26/10 Tr. 30; Lopez I, 09/15/10 Tr. 5859; Lopez I, 09/16/10 Tr. 110). Surely Boston is not here
arguing that such an examination, standing alone, would pass
muster. Every judge who has considered the issue has held to
the contrary. Smith, 144 F. Supp. at 208, 210-11 (Young, J.);
Lopez I, 2014 U.S. Dist. LEXIS 124139, at *60-61 (O’Toole, J.);
[15]

Uniform Guidelines’ representative sample test.

It would be

surprising if it had since the Guidelines come from the Equal
Employment Opportunity Commission and are due an appropriate
degree of deference.
Still, Boston argues that this Court erred by applying the
Guidelines as “binding legal standards” -- Boston asserts that
Lopez II makes clear that the representative sample test cannot
be used because the Guidelines do not have a quantitative
measure for deciding whether or not a selection procedure tests
a representative number of KSAs.
This Court disagrees.

Def.’s Br. 6-7.

Lopez II noted that although the

Guidelines do not provide a quantitative measure to draw the
line between representative and nonrepresentative samples of job
performance, the Guidelines do “point to the qualitative
understandings of these concepts generally accepted by
professionals who evaluate [selection procedures].”
112.

823 F.3d at

Accordingly, although there may not be a bright line to

which reference can be made, expert testimony can still
highlight on which side of a blurry line a selection device
falls.

In fact, Lopez II recognized that the testimony of Dr.

James Outtz did just this -- “[Outtz] opined that the exams were

see also Lopez II, 823 F.3d at 113-15 (Lynch & Kayatta, JJ.);
id. at 124-25 (Torruella, J., concurring in part and dissenting
in part).
[16]

based on job analyses that validly identified the critical
skills used by actual police sergeants and that the tests
covered a ‘representative sample’ of the content of the job.”
Id. (quoting 29 C.F.R. § 1607.14(C)(4)).

Here, in contrast, a

different expert opining on a different exam did not convince
this Court that the 2008 lieutenants’ exam measured a
representative sample of relevant KSAs.

Smith, 144 F. Supp. at

207-08.
Boston also argues that Lopez II declined to follow the
Guidelines’ technical requirements and instead established a
lessened burden for the employer: a showing that the challenged
exam is more job-related than random selection.

Def.’s Reply 3.

Boston posits that Lopez II “makes clear, ‘in the absence of any
quantitative measure of “representativeness” provided by the
law,’ the proper inquiry is not about representativeness, but
whether the exam overall is more job-related than random
selection would be.”
at 116).

Def.’s Br. 6 (quoting Lopez II, 823 F.3d

Indeed, Lopez II emphasized that “[t]he Guidelines

quite understandably provide no quantitative measure for drawing
the line between ‘representative,’ and nonrepresentative samples
of job performance and behaviors.”
C.F.R. § 1607.5(B)).

823 F.3d at 112 (quoting 29

As Boston notes, the Lopez II court went

on to reject the appellants’ arguments, stating:

[17]

None of the remaining arguments advanced by the
[appellants] seriously support any claim that the exams are
not materially better predictors of success than would be
achieved by the random selection of those officers to be
promoted to sergeant. The parties’ arguments, instead,
focus on how much better the exams were. Do they test
enough skills and knowledge? Do they weigh the answers in
an appropriate, valid way? In finding Outtz persuasive on
these points, the district court as factfinder did not
clearly err.
Id. at 116-17.
Boston interprets the First Circuit’s decision as a
rejection of the argument that a test can be invalidated if it
fails either to test a sufficient representative sample of
skills and abilities or to meet a heightened standard of
validity for rank ordering.

Def.’s Reply 3-4.

This, however,

may well not be a fair interpretation -- in the quoted passage
the First Circuit was merely deferring to the district court’s
role as the finder of fact on those points.

Further, Lopez II

upheld the district court’s use of the representative sample
test, noting:
The district court’s opinion as a whole thus makes clear
that the court trained its focus on critical and important
knowledge, skills, and abilities called for by the job, and
it did not clearly err by finding that a test that measured
a large percentage of such critical and important KSAs was
a test that was sufficiently “representative of important
aspects of performance on the job.” Our conclusion to this
effect finds further support in the absence of any
quantitative measure of “representativeness” provided by
the law. Rather, the relevant aim of the law, when a
disparate impact occurs, is to ensure that the practice
causing the impact serves an important need of the
employer, in which case it can be used unless there is
another way to meet that need with lesser disparate impact.
[18]

We cannot see how it is an error of law to find that an
exam that helps determine whether an applicant possesses a
large number of critical and necessary attributes for a job
serves an important need of the employer.
Lopez II, 823 F.3d at 115-16.

Fairly read, this passage

condones the district court’s use of the representative sample
test.

It does not go so far as either to mandate or disapprove

of that use as matter of law.
Nearly all of the circuits have utilized a representative
sample test in examining content validity.

See, e.g., Johnson

v. City of Memphis, 770 F.3d 464, 478 (6th Cir. 2014), cert.
denied sub nom. Johnson v. City of Memphis, Tenn., 136 S. Ct. 81
(2015); M.O.C.H.A. Soc’y, Inc. v. City of Buffalo, 689 F.3d 263,
281 (2d Cir. 2012); Equal Emp’t Opportunity Comm’n v. Dial
Corp., 469 F.3d 735, 742 (8th Cir. 2006); Allen v. City of
Chicago, 351 F.3d 306, 309 n.3 (7th Cir. 2003); Zottola v. City
of Oakland, 32 F. App’x 307, 311 (9th Cir. 2002); Nash v.
Consolidated City of Jacksonville, Duval Cty., Fla., 837 F.2d
1534, 1539 (11th Cir. 1988), cert. granted, judgment vacated,
490 U.S. 1103 (1989), and opinion reinstated, 905 F.2d 355 (11th
Cir. 1990).

One imagines that, had the First Circuit been

adopting a legal rule it would have acknowledged this body of
law.
Boston argues that to satisfy its prong 2 burden, it need
only establish that its test is a materially better predictor of

[19]

success than random selection.

Def.’s Reply 6.

This proposed

standard, however, is not irreconcilable with the representative
sample test.

The aim of the representative sample test is to

ensure that an exam tests for success in a specific job.
Applying the representative sample test ensures that “the
content of the selection procedure is representative of
important aspects of performance on the job for which the
candidates are to be evaluated.”

29 C.F.R. § 1607.5(B).

In

other words, a court ensures that a selection device evaluates
characteristics important to job performance, rather than random
attributes that may not correlate with success in that job.

To

be materially better than random at choosing applicants who will
excel at a job, this Court can only imagine that the selection
device would necessarily examine a large proportion of the KSAs
needed to succeed at the position.
As discussed above, Lopez II did not reject the
representative sample test as matter of law; and to assume that
the First Circuit would change disparate impact law without so
much as a comment seems somewhat at odds with reality.
Accordingly, this Court declines to reconsider its use of the
representative sample test from the Guidelines.
2.

Rejection of the E&E Component

Boston argues that because Lopez II found that the E&E was
useful for qualities important to a sergeant’s daily
[20]

responsibilities, this Court should apply the E&E as part of its
validity analysis.

Def.’s Br. 9-10.

In fact, this Court did

consider the E&E in its ruling, but held the E&E did not rescue
an otherwise invalid written exam.

Smith, 144 F. Supp. 3d at

204.
The First Circuit discussed the E&E component of the 2008
sergeant’s exam:
In Outtz’s opinion, however, the addition of the E&E
component effectively pushed the selection device as a
whole across the finish line to show validity. It did
this, according to Outtz, because the level and extent of
work and educational experience and accomplishments listed
by each applicant served as a useful, if imperfect, proxy
for the kinds of qualities that were deemed to be important
to a sergeant’s daily responsibilities, yet were
insufficiently tested by the examination’s question and
answer component. Outtz recognized that the gain in
validity from the E&E component was, on its own, only
marginal or “incremental.” As the Officers stress, many of
the attributes for which the E&E assigned points . . . were
shared by all or most applicants. . . . when weighted to
provide only 20% of the combined final score, it accounted
for a range of only about 5% to 7% of a candidate’s total
score. Nevertheless, we cannot see how a rational
factfinder could ignore the impact of the E&E, small or
not, in evaluating the exam overall.
Lopez II, 823 F.3d at 113.
The evidence presented in Smith, however, varied from that
available in Lopez I.

In reaching its decision, this Court:

(1) relied on expert testimony that the E&E component failed to
differentiate among candidates or demonstrate the KSAs necessary
in a lieutenant, Smith, 144 F. Supp. 3d at 203-04; (2) had no
evidence that incumbent lieutenants performed better on the
[21]

written exam, see generally id. at 207-10 (discussing the
evidence presented to demonstrate exam validity); and (3) had no
evidence to show that the E&E component was valid on its own,
id. at 211 n.42.
These differences are crucial.

In Lopez II, the E&E inched

the exam over the line of validity due to its measuring “the
kinds of qualities that were deemed to be important to a
sergeant’s daily responsibilities.”

823 F.3d at 113.

Here,

however, the testimony does not establish that the E&E measured
qualities important to a lieutenant’s daily responsibilities.
Further, even if the E&E assessed a lieutenant’s important KSAs,
Dr. Wiesen’s testimony that the E&E “had virtually no impact on
the final exam scores,” Smith, 144 F. Supp. 3d at 204 (internal
quotation marks omitted), persuades this Court that the E&E had
so minimal an effect that it could not uphold the 2008
lieutenants’ exam’s validity.

Accordingly, this Court declines

further to address the E&E in its analysis.
3.

Rank Ordering

Boston argues that this Court inappropriately applied a
heightened validity requirement for rank ordering and that Lopez
II holds that rank ordering furthers Boston’s interest in
eliminating patronage and intentional racism.

Def.’s Br. 11.

The First Circuit’s statement is, however, dicta.
In Lopez II, the First Circuit stated:
[22]

Rank ordering furthers the City’s interest in eliminating
patronage and intentional racism under the guise of
subjective selection criteria. Such a goal is itself a
reasonable enough business need so as to provide some
weight against a challenge that is unaccompanied by any
showing that rank order selection itself caused any
disparate impact in this case.
823 F.3d at 119.

Boston asserts that this statement binds this

Court, Def.’s Br. 11, and so Boston’s business need to rank
order supports its meeting the prong 2 burden in light of the
Plaintiff’s failure to set for evidence that rank ordering
results in disparate impact, id. at 11-12.

This Court, however,

did find that rank ordering based on the 2008 lieutenants’ exam
had a disparate impact on minority applicants.
Supp. 3d at 199-200.

Smith, 144 F.

In Smith, this Court stated, “the

Plaintiffs have presented evidence of statistically significant
disparate impact on . . . delay in promotions”; “if the
eligibility list from the 2008 [lieutenants’] exam had not been
extended due to the Lopez litigation, but instead had expired
three years after its creation, as is typical, not a single
black sergeant would have been promoted to lieutenant”; and
“[a]ll of this evidence combined is enough for this Court to
rule that the Plaintiffs have met their burden of raising an
inference of causation and demonstrating a prima facie case of
disparate impact.”

Id.

Although these statements do not use

the precise language “rank order selection caused disparate
impact,” this Court made it sufficiently clear -- through its
[23]

discussion of disparate impact in relation to delayed promotions
-- that rank ordering resulted in disparate impact.
Boston goes on to argue that a prong 2 analysis does not
depend on whether rank order selection increases disparate
impact and that neither Title VII nor the Guidelines provide a
quantitative requirement about how job-related a selection
device has to be or how much better it need be for rank
ordering.

Def.’s Reply 9-10.

“Employers should tailor a

discriminatory hiring practice to a job-related risk, making
sure to proportionally weigh the costs and benefits of
accommodating that risk.”

Jake Elijah Struebing, Note,

Reconsidering Disparate Impact Under Title VII: Business
Necessity as Risk Management, 34 Yale L. & Pol’y Rev. 499, 507
(2016).

Where a selection procedure not only has a disparate

impact on a pass-fail basis, but also compounds that effect
through use of rank ordering, each hiring decision carries an
increased risk of a discriminatory result.

Such heightened risk

merits applying a more stringent validity requirement to ensure
that the exam is sufficiently job-related to warrant the cost of
potential discrimination.

Accordingly, this Court did not err

in applying a heightened validity requirement for rank ordering.
Further, even if Boston faced a lower bar to establish
validity, it still failed to show that it met its burden.
Although, as Boston argues, Def.’s Reply 10, this Court stated
[24]

that “[w]hat the Court can conclude from the 2008 [lieutenants’]
exam is that those who excelled at the exam would exhibit
superior levels of knowledge on the job, and that the 2008
[lieutenants’] exam differentiated among levels of candidates’
knowledge levels,” Smith, 144 F. Supp. 3d at 209, the Court also
noted “that this is insufficient for predicting who will be a
good police lieutenant,” id.

In particular, this Court

emphasized that testing only knowledge, rather than including
other necessary skills and abilities, could not persuade this
Court that those who did well would be better performers on the
job.

See id.

Put another way, even were this Court to apply

the “better than random” standard Boston advocates, Def.’s Reply
10, the Court concluded that it cannot presume that testing only
knowledge will result in a better than random procedure for
selecting candidates for promotion.

As this Court stated, “the

evidence does not support the necessary inference that those who
perform better on the exam will be better performers on the job,
primarily because the exam did not test a sufficient range of
KSAs, and there was no evidence that the exam was reliable
enough to justify its use for rank ordering.”
Supp. 3d at 203.

Smith, 144 F.

Additionally, the Court later emphasized, “the

Court cannot find that [Boston] met its burden on [adequacy of
test construction]: too many skills and abilities were missing
from the 2008 test outline,” id. at 205, and “[t]he Court
[25]

concludes that the 2008 [lieutenants’] exam did not sufficiently
test for a representative sample of the critical KSAs,” id. at
207.

This goes to show that the Court held that even if Boston

did not need to meet a heightened standard for rank ordering, it
still failed to carry its burden to establish test validity.
B.

Equally Valid, Less Discriminatory Alternative (Prong
3 Challenges)

Boston last argues that this Court cannot reject Boston’s
business justification unless there is some showing that there
“exists an available alternative with less disparate impact that
serves [Boston’s] legitimate needs.”

Def.’s Br. 12.

This

Court, however, is confident in its understanding of the
shifting burdens of the disparate impact framework: if the
defendant fails to meet its burden of proof on prong 2, then the
defendant loses, regardless of the plaintiffs’ showing of an
alternative.
Boston argues that Jones v. City of Boston (Jones II), 845
F.3d 28 (1st Cir. 2016), affirms Lopez II’s lowered prong 2
standards and emphasizes the importance of prong 3.
Reply 1-2.

This contention is hardly convincing.

Def.’s

In Jones II,

a number of police department employees brought a disparate
impact challenge to the Boston Police Department’s hair drug
test, which they claimed was racially discriminatory.
at 30-31.

845 F.3d

Having held that the police department employees met

[26]

the first prong of the disparate impact inquiry in Jones I, the
First Circuit examined the district court’s grant of summary
judgment on prongs 2 and 3, upholding the former and vacating
the latter.

See id.

The First Circuit first turned to prong 2:

whether the challenged test was job-related and consistent with
business necessity.

See id. at 32.

Noting that “[t]he parties

agree[d] that ‘abstention from drug use is an important element
of police behavior,’ and is thus job related . . . . [and] that
selecting police officers for retention or discharge based on
that job-related behavior is consistent with business
necessity,” the court turned its analysis to whether the drug
test was so unreliable that a reasonable juror could find that
the test “did not meaningfully further the [Boston Police]
Department’s legitimate need for a drug-abstaining police
force.”

Id.

The court emphasized that the test had a high

degree of accuracy, but that
there is no reason why a test need be anything near 100%
reliable (few tests are) to be consistent with business
necessity (keeping in mind that the presence of an
alternative method that would have had less of a disparate
impact will still be relevant under the third prong of the
inquiry).
Id. at 33.

The court then ruled that the Boston Police

Department had clearly met its prong 2 burden to establish that
the test was job-related and consistent with business necessity.
Id. at 33-34.

Turning to prong 3 of the inquiry, the court

[27]

eventually held that police department employees could
potentially succeed in showing that the Boston Police Department
refused to adopt an alternative test with less of a disparate
impact.

See id. at 38.

Accordingly, the circuit vacated the

district court’s grant of summary judgment on prong 3.

Id.

Boston latches onto the Jones II court’s focus on
reliability, arguing that it evidences the First Circuit
lessening the burden on employers in prong 2 while increasing
the importance of prong 3.

Def.’s Reply 2.

Jones II, however,

is distinguishable from the instant case on numerous grounds.
First and foremost, in Jones II, both parties had essentially
agreed that, if reliable, the drug test was job-related and
consistent with business necessity.

See 845 F.3d at 32.

Here,

in contrast, this Court specifically held that Boston failed to
show that the 2008 lieutenants’ exam was job-related and
consistent with business necessity.
211.

Smith, 144 F. Supp. 3d at

Accordingly, an emphasis on reliability would be

inappropriate.

Second, Jones II addressed the validity of a

test designed to look only at one thing -- drug abstention -rather than a complex examination designed to test for the KSAs
of promotion candidates.

For this reason, the Jones II court’s

validity examination is distinguishable from this Court’s
business necessity analysis in Smith.

[28]

Further, Lopez II itself is consistent with the traditional
burden-shifting framework of disparate impact.

In Lopez II, the

court summarizes the law of disparate impact, explicitly stating
that a plaintiff who satisfies prong 1 will prevail either by
the employer failing to meet their burden on prong 2, or by the
plaintiff meeting their burden on prong 3.
F.3d at 110-11.

See Lopez II, 823

This outright statement of law warns against

heeding Boston’s call to collapse the inquiry of prongs 2 and 3.
IV.

CONCLUSION
Both Lopez I and Smith are fact intensive cases.

In Lopez

I, Boston persuaded Judge O’Toole, albeit barely, that it ought
prevail on prong 2.

Smith is a different case, with a different

evidentiary record, involving different expert testimony about a
different and more demanding senior officer position.

In Smith,

Boston failed to convince me that it ought prevail on prong 2,
an aspect of the case on which it bears the burden of proof.
The Court of Appeals affirmed the district court’s decision in
Lopez I.
In view of this affirmance, Boston seems to be arguing that
it is now legal error to reach a contrary result in a
significantly different case.3

That’s not how it works.

3

Fact

Indeed, Boston apparently seeks to conjure up a hitherto
unrecognized species of offensive issue preclusion, cf. BlonderTongue Labs., Inc. v. University of Ill. Found., 402 U.S. 313
(1971), based on the overlapping portions of the sergeants’ and
[29]

finding is the province of the district courts.

In Lopez II,

the Court of Appeals did what it always does -- it carefully
scrutinized the evidentiary record, giving due deference to the
fact-finding role of the district judge, to see whether any of
his conclusions were clearly erroneous.
was doing.
did.

That’s what it said it

See Lopez II, 823 F.3d at 107-08.

That’s what it

Should this Smith case be appealed, it will do the same

for me.
Boston fails to offer any convincing argument as to why
this Court ought disrupt its previous ruling in Smith.
Accordingly, it declines to do so.

The Court also takes this

opportunity to note that Boston is not left without any useable
test.

Boston utilized an updated selection procedure in 2014.

Def.’s Br. 4.

Although Boston has represented to this Court

that the 2014 exam resulted in a greater disparate impact and
already faces legal challenges, id. at 4 & n.3, none of these
issues are properly before this Court.
What is clear is that this case has gone on far too long.
It is nearly a decade since the original Boston patrolmen
brought suit in Lopez I.
long out of date.

The tests that so engross us here are

In its order rejecting an interlocutory

appeal, the First Circuit indicated it was amenable to

lieutenants’ written exams and the fact that the same
plaintiffs’ counsel appears in both cases.
[30]

entertaining such an appeal once this Court had analyzed the
effect of Lopez II on its earlier decision in Smith.
done so.

It has now

Should the parties jointly move, within 14 days of the

date of this memorandum, for an order authorizing application
for a second interlocutory appeal, this Court will allow such
motion.

If not, this Court will promptly schedule hearings on

remedy, settlement talks or no.
SO ORDERED.
/s/ William G. Young
WILLIAM G. YOUNG
DISTRICT JUDGE

[31]