Michael A. Nemir, M.d. v. Mitsubishi Motors Corporation; Chrysler Corporation

*The Honorable William W Schwarzer, Senior United States District
Judge for the Northern District of California, sitting by designation.
1
RECOMMENDED FOR FULL-TEXT PUBLICATION
Pursuant to Sixth Circuit Rule 206
ELECTRONIC CITATION: 2004 FED App. 0273P (6th Cir.)
File Name: 04a0273p.06
UNITED STATES COURT OF APPEALS
FOR THE SIXTH CIRCUIT
_________________
MICHAEL A. NEMIR, M.D.,
Plaintiff-Appellant,
v.
MITSUBISHI MOTORS
CORPORATION; CHRYSLER
CORPORATION,
Defendants-Appellees.
X
-
-
-
->
,
-
-
-
-
-
N
Nos. 02-1780;
03-1228
Appeal from the United States District Court
for the Eastern District of Michigan at Detroit.
No. 96-75380—John Feikens, District Judge.
Argued: March 18, 2004
Decided and Filed: August 20, 2004
Before: COLE and GILMAN, Circuit Judges;
SCHWARZER, Senior District Judge.*
2 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
_________________
COUNSEL
ARGUED: Mark R. Bendure, BENDURE & THOMAS,
Detroit, Michigan, for Appellant. David R. Kelly,
BOWMAN & BROOKE, Minneapolis, Minnesota, for
Appellees. ON BRIEF: Mark R. Bendure, BENDURE &
THOMAS, Detroit, Michigan, Craig E. Hilborn, HILBORN
& HILBORN, Birmingham, Michigan, for Appellant. David
R. Kelly, BOWMAN & BROOKE, Minneapolis, Minnesota,
Fred J. Fresard, Gretchen A. Colter, BOWMAN & BROOKE,
Troy, Michigan, for Appellees.
_________________
OPINION
_________________
R. GUY COLE, JR., Circuit Judge. This case reaches us
for the fourth time. See Nemir v. Mitsubishi Motor Sales of
Am., Inc., 6 Fed. Appx. 266 (6th Cir. 2001) (unpublished)
(per curiam) (“Nemir I”); In re. Michael A. Nemir, No. 01-
2260 (6th Cir. Oct. 12, 2001) (“Nemir II”); Nemir v.
Mitsubishi Motor Sales of Am., Inc., No-02-1780 (6th Cir.
Feb. 3, 2003) (“Nemir III”). Plaintiff-Appellant Michael A.
Nemir brought a diversity products liability suit against
Defendants-Appellees Mitsubishi Motors Corporation and its
parent company, Chrysler Corporation, alleging that his
seatbelt’s failure to latch caused injuries that left him brain-
damaged and paralyzed following a car accident. After the
jury rejected Nemir’s claims, the district court entered
judgment for, and awarded costs to, Mitsubishi, and Nemir
appealed both decisions. For the following reasons: (1) the
district court’s judgment for Mitsubishi is REVERSED and
the case is REMANDED for a new trial; (2) the district
court’s award of fees to Mitsubishi is VACATED; and (3) the
case is to be assigned to a different district judge.

-- 1 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
3
I. BACKGROUND
Given this case’s dense history and its relevance to today’s
appeal, we will narrate the background in some detail. On
December 14, 1993, near his home in Maryland, Nemir drove
his car into a tree on the side of the road, causing his head to
strike a pillar on the car’s passenger side. He was found lying
in the car’s backseat, and the driver’s side seatbelt was
unlatched. The accident damaged Nemir’s brain stem,
leaving him wheelchair bound, cortically blind, and unable to
continue practicing medicine. On November 22, 1996, Nemir
filed suit against Mitsubishi, the maker of the 1991 Dodge
Stealth that Nemir was driving, and Chrysler, of which
Mitsubishi is a subsidiary. (For simplicity’s sake, we will
refer to the defendants collectively as “Mitsubishi.”)
Nemir’s complaint alleged that at the time of the accident,
he had been wearing his seatbelt— the Takata 52—but that it
had only “partially latched,” such that it appeared to be
properly fastened when in fact it was still prone to unlatching.
Seeking both compensatory and punitive damages, Nemir
asserted that the seatbelt was defectively designed because of
its potential to unlatch, and because of the vehicle’s failure to
alert the driver that the seatbelt was not fully latched. The
failure to warn and punitive damages claims were
subsequently dismissed, and we affirmed those dismissals in
Nemir I, 6 Fed.Appx. at 277.
Prior to trial, Nemir retained an expert witness, Dr. Thomas
Horton, the former Director of Engineering for Takata, Inc.,
the seatbelt’s manufacturer. Horton planned to testify that
Nemir’s seatbelt suffered from a design defect, known as
“partial-latching,” which caused the seatbelt to unlatch during
the accident. However, the district court ruled that Horton
was unqualified to testify as an expert witness, due to
purported deficiencies in his methodology.
4 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
In December 1994, based on complaints it had received, the
National Highway Traffic Safety Administration (NHTSA)
opened an investigation into “[a]lleged defect[s] of Takata
seat belt buckles includ[ing] failure to latch, unlatch, or
remain latched.” NHTSA requested information from
Chrysler, including complaints received from its customers,
pertaining to “all model years 1986 through 1991 Dodge
Colt[s] and any other passenger motor vehicles equipped with
Takata [52] series.” (In 1996, owing to a manufacturing
defect, some Takata 52 seatbelts—but not the seatbelt at issue
in our case—were recalled.)
Aware of this investigation, Nemir sought discovery from
Mitsubishi of “[a]ll documents containing, referring to, or
relating to any claims, assertions, or reports by an individual
or entity that any Dodge vehicle using the same or similar
seatbelt system wherein a front seat belted passenger received
a serious or fatal injuries.” On October 31, 1997, Mitsubishi
responded that it “has no documents responsive to this
request.” On September 11, 1998, Mitsubishi reaffirmed that
it “currently is unaware of any materials responsive to this
request.” When Nemir further pressed the issue at a hearing
held on November 17, 1998, Mitsubishi insisted that “[w]ith
respect to all of the requests that are contained in the order,
we produced what documents exist and what documents were
able to be located . . . we searched, as we did for all of the
other requests, and found nothing that was responsive to the
request.”
Nemir eventually filed a Freedom of Information Act
request with NHTSA, and subsequently obtained 200 to 300
complaints of partial latching in Mitsubishi and Chrysler
automobiles. The documents included evidence of
“[Mitsubishi’s] own employees going out and confirming
partial latch in the field.” When asked by the district court
whether Mitsubishi had any additional documents similar to
the ones that Nemir obtained from the NHTSA, Mitsubishi’s

-- 2 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
5
counsel told the court that “I don’t know of any other
documents, Your Honor.”
The district court ordered Mitsubishi to search again for
additional complaints in its possession. On April 13, 1999,
Mitsubishi admitted that it possessed, but had not previously
disclosed, approximately 4,000 customer contact reports
documenting complaints, along with approximately 25,000
warranty claims about the partial latching of the seatbelts.
Mitsubishi—who claimed to have previously withheld these
documents because “they do not show claims that partial
latching occurred in the manner alleged to have occurred in
this case”—continued to maintain that the documents should
not be disclosed to Nemir.
Although the district court disagreed, it also prohibited
Nemir from “conduct[ing] further discovery by attempting to
interview or contact complainants listed in these documents.”
Moreover, the district court scheduled argument on
Mitsubishi’s motion for summary judgment for May 21,
1999—just seventeen days after the 29,000 documents were
ordered disclosed. On July 30, 1999, the district court
granted summary judgment to Mitsubishi, holding that Nemir
had failed to present evidence by which a reasonable jury
could find (under Maryland law, which the parties agree
controls) that the buckle suffered from a design defect. Nemir
v. Mitsubishi Motor Sales of America, 60 F.Supp.2d 660, 673-
77 (E.D.Mich. 1999).
In Nemir I, we reversed. First, we noted that the district
court had erroneously stated that Horton had not examined the
actual buckle in Nemir’s seatbelt—an omission that the
district court had dubbed “inexplicable”—when in fact the
record revealed that Horton had testified that he had examined
the actual buckle and that it had partially latched twice out of
twenty attempts. Id. at 274. Second, we held that the district
court erroneously excluded the entirety of Horton’s
testimony, and specifically held that “Horton could use [the
6 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
abovementioned testing] to testify that the partial latch of Dr.
Nemir’s seatbelt caused the damage in question.” Id. at 275.
Third, we held that Nemir had produced sufficient evidence
for a jury to conclude that a design defect rendered the buckle
unreasonably dangerous. Id. at 277. We remanded for trial.
We also ordered the district court to afford Nemir sufficient
time and opportunity to investigate the consumer complaints
that Mitsubishi had turned over at “the eleventh hour.” Id. at
275. Although Mitsubishi argued that none of the consumer
complaints contained the words “partial latch,” and thus were
not discoverable, we rejoined that “the term ‘partial latch
phenomenon’ is a term of art employed by engineers, seatbelt
designers, and lawyers. The layman with a concern that his
seatbelt will not remain fastened will not use this term.” Id.
at 276. After reviewing the sample of 188 complaints that
Nemir had included in the appellate record, we noted that “
[a] complete review of the record indicates that phraseology
analogous to ‘intermittent partial latching’ occurred 188 times
in the customer complaints submitted as evidence. To
summarily dismiss this evidence without affording the
plaintiff time to investigate and further elicit information
regarding the exact nature of these incidents is clear error.”
Id.
On remand, the district court issued a new protective order.
The order limited Nemir’s discovery to “the 188 complainants
referred to by” our decision in Nemir I. Although Nemir
argued to the district court that Nemir I singled out 188
complaints only because that was the number of complaints
that had been included in the appellate record—in other
words, that we had not manually sifted through each and
every of the 29,000 complaints and determined that only 188
were relevant—the district court nonetheless refused to allow
Nemir to investigate the other 28,000-plus complaints,
insisting that “I’m going to stay literally with what the panel
said.”

-- 3 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
7
In addition, the order prohibited Nemir from directly
contacting any of Mitsubishi’s consumers outside of defense
counsel’s presence. Of the 188 customers that Nemir was
permitted to contact, he was able to locate only thirteen. Prior
to the depositions of those witnesses, Nemir asked the district
court for permission to contact additional complainants
beyond the original 188; the district court told Nemir that he
should conduct the first set of depositions, and then “if you
need to expand this discovery further, come back to me on it.”
Nemir never did so.
Nemir did ask the district court to sanction Mitsubishi for
its refusal to disclose the 29,000 documents. He appended to
his motion evidence that the improper withholding of relevant
discovery was not an uncommon practice of both Mitsubishi
and its lead counsel. Nemir attached an affidavit from a
plaintiff’s attorney, who had litigated against Mitsubishi in a
prior case and averred that “Mitsubishi [had] repeatedly
refused to turn over properly requested information and
respond to interrogatories,” and “did not produce certain
materials until it was threatened with daily fines.” Also
submitted by Nemir was a copy of the decision in Baker v.
General Motors Corp.,159 F.R.D. 519 (W.D. Mo. 1994),
reversed on other grounds, Baker v. General Motors Corp.,
86 F.3d 811, 817 (8th Cir. 1996), in which Mitsubishi’s lead
counsel had defended General Motors against lawsuits
arising from electrical fires in certain vehicles. There, the
court noted that “[e]very time a request was made for a
document, [counsel] took it upon [him]self to determine what
scope [he] thought was relevant,” id. at 523, the court found
numerous examples “of the types of critical documents that
were ‘dumped’ on plaintiffs the week before trial,” id. at 522,
and the court concluded that there was a “deliberate, willful
policy on the part of [counsel] to stonewall discovery as much
and as long as the patience of the Court would tolerate.” Id.
at 524.
8 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
When Nemir pressed for sanctions at a hearing conducted
on March 16, 2001, the district court refused to “reach the
motion for sanctions at this point,” and complained that “I
just don’t want to be bothered with sanctions.” Instead, the
district court decided to “stay that matter and return to it when
the time is appropriate.” The district court never revisited the
motion, and also prohibited Nemir from informing the jury
that Mitsubishi had initially withheld the consumer
complaints.
In addition, despite our holding in Nemir I that its
namesake had “demonstrate[d] a triable issue of fact with
regard to the design defect condition and the unreasonably
dangerous condition of the product,” the district court again
ruled that “[n]othing in the record before [us] indicates that
Dr. Nemir’s seatbelt caused an unreasonably dangerous
condition.” Convinced that the district court had ignored our
prior decision to the contrary, Nemir sought from us a writ of
mandamus compelling the district court to “submit to the jury
the issue of ‘inherently unreasonable risk’ in accord with
[Nemir I].” The district court responded to the petition for
mandamus by assuring us that its ruling was “subject to
change based on the proofs that are presented.” Nemir II, slip
op at 5. Relying on the district court’s assurances, we denied
the petition for mandamus, concluding that “[t]he case is
proceeding to trial in accordance with our determination that
the plaintiff has presented sufficient proof to create a question
for the jury regarding his design defect claim.” Nemir II, slip
op. at 2.
Notwithstanding our instructions in Nemir II, the district
court, pursuant to Federal Rule of Evidence 706, appointed
Eddie P. Cooper and Lindley Manning as experts to help the
court “determine whether the seatbelt buckle presented an
inherent unreasonable risk of danger to the plaintiff.”
Cooper—who was employed by a consulting company that
had performed significant work for Mitsubishi and
Takata—concluded that the seatbelt was not inherently

-- 4 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
9
dangerous. Manning, echoing Horton’s conclusions, testified
that he too had been able to partially engage the Nemir buckle
during laboratory testing. Manning also opined that “the
defect that we found is unreasonably dangerous . . . . The fact
that it would happen even once or a few times makes it
unreasonably dangerous.” On May 22, 2002, relying on the
testimony of Cooper and discounting that of Manning, the
district court again held that the seatbelt “does not present an
inherently unreasonable risk of danger,” thereby reiterating its
unwillingness to instruct the jury on Nemir’s theory of strict
liability. See Nemir v. Mitsubishi Motors Corp., 201
F.Supp.2d 779, 783 (E.D.Mich. 2002).
The trial was fueled by three disputed issues. First was
whether Nemir was wearing his seatbelt at the time of the
accident. Nemir testified generally that he almost always
wore his seatbelt while driving, and specifically that he was
wearing it at the time of the accident. In support of Nemir’s
testimony, another of his experts, Dr. Joseph L. Burton,
testified that cotton fibers resembling those from Nemir’s
jacket were embedded along the edge of the seatbelt; that the
seatbelt appeared to have been stretched; and that the buckle
had “loading marks” consistent with wear. Mitsubishi
submitted evidence indicating that Nemir’s injuries had
diminished his memory of the accident, and produced expert
witnesses who downplayed the conclusiveness of Nemir’s
fiber evidence and surmised that Nemir’s seatbelt lacked the
markings that would have appeared had it been worn during
the accident.
Second was the question of the cause of Nemir’s injuries.
Burton testified that had he been belted, Nemir’s collision
with the pillar would have been much slower, and Nemir
would have suffered milder injuries. Mitsubishi’s expert, Dr.
Charles Hatsell, concluded that the injuries to Nemir’s brain
stem resulted from head rotation that even a fully functioning
seatbelt would have failed to prevent. Through cross-
examination, Nemir’s attorney attempted to impeach Hatsell
10 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
by confronting him with a contradictory excerpt from a book
that had been co-authored by Dr. Werner Spitz, another of
Mitsubishi’s experts. The district court prohibited this
questioning, however, concluding that because Dr. Spitz
would be testifying later in the case, Nemir should simply
elicit the information from Dr. Spitz when he testified.
The next day, when Nemir’s attorney attempted to question
Dr. Spitz about his book, the district court sustained
Mitsubishi’s objection to the questioning, as illustrated by the
following exchange:
Nemir’s Counsel (questioning Dr. Spitz): [Your book]
talks about the brain stem contusions and I’m looking at
the chart that Dr. Hatsell has that we all saw . . .
Mitsubishi’s Counsel: Your Honor, I would object to
any questioning of this witness on the issue of brain
injury and injury causation. Although he is qualified,
he’s not given any testimony on direct examination on
injury causation, and we’re not offering him for that.
Nemir’s Counsel: Well, I – I specifically . . .
Judge Feikens: If he’s not being offered for that, then
why are you going into it?
Nemir’s Counsel: Because you asked me to do it, your
Honor. I asked to be able to cross-examine with his book
yesterday with Dr. Hatsell, and you said the proper place
to do it was with this witness.
Judge Feikens: On . . . on brain injury even though he’s
not being offered from that reason?
Nemir’s Counsel: Your Honor, I tried to cross-examine
to show this man had written—or shown that Hatsell’s

-- 5 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
11
opinion is impossible. I tried to cross-examine. You
said, no, I should do it with . . .
Judge Feikens: I’ll sustain the objection.
Needless to say, the impeaching passage never reached the
jury.
Third was whether the seatbelt was capable of partially
latching, such that it could malfunction during an accident.
Dr. Horton testified that the seatbelt buckle was unreasonably
dangerous because it allowed for “partial engagement.” But
he was unable to explain the basis for his conclusion because
the district court prohibited Horton from testifying that he
was able, twice out of twenty attempts, to create a partial latch
of Nemir’s seatbelt buckle, opining that Horton’s method—in
which he manipulated the buckle at varying speeds and
angles—was scientifically unsound. Nemir, 200 F.Supp.2d
at 778. The district court also prohibited Horton from
testifying about causation, reiterating its prior concern that
Horton “does not demonstrate how he arrived at his
conclusion that partial latch caused Nemir’s injuries or how
he eliminated other possibilities, chief among which is the
possibility that Nemir was not wearing a seat belt.” Nemir,
200 F.Supp.2d at 776. Moreover, although Federal Rule of
Evidence 706 provides that a court-appointed witness “may
be called to testify by the court or any party,” the district court
refused to allow Nemir to call Manning at trial.
Taking advantage of the gaps in Horton’s testimony and the
absence of Manning’s testimony, Mitsubishi argued to the
jury, among other things, that partial latching had never been
demonstrated, that Horton’s testimony was “sheer
speculation,” and that “[y]ou can’t prove your case with just
the words of a hired expert.” Moreover, although the other
court-appointed witness, Cooper, had testified prior to trial
that were he a manufacturer that had received complaints of
unlatching seatbelts, he would want to examine the buckles in
12 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
question so that he could pinpoint the cause, Mitsubishi
claimed that it had not performed any such tests. The district
court prohibited Nemir from introducing—in order to explain
why there was not more evidence of partial engagement
during actual use—evidence that Mitsubishi had failed to test
the buckles flagged by its customers. And Mitsubishi’s
expert testified that it was impossible to achieve partial
engagement in actual use.
The jury’s Verdict Form posed three questions:
1. Was Nemir wearing his seatbelt at the time of the
accident?
2. If yes, was the seatbelt defective?
3. If yes, was the defect the proximate cause of Nemir’s
injuries?
The jury answered that Nemir was not wearing his seatbelt at
the time of the accident, and therefore did not reach the other
questions. On July 4, 2002, the district court entered
judgment for Mitsubishi.
On June 12, 2002, Nemir appealed, arguing that the district
court erred in excluding portions of Horton’s testimony,
restricting his access to customers who had filed complaints,
refusing to instruct the jury on strict liability, refusing to
allow him to call Dr. Manning, and inhibiting his
impeachment of Dr. Hatsell. After Nemir submitted his
appellate brief, the district court ordered Nemir to post a bond
of $20,000 to secure the costs of the appeal.
In the meantime, Mitsubishi asked the district court to
award it $48,000 in costs from Nemir. Pursuant to 28 U.S.C.
§ 1920, the Clerk of the district court entered an order
awarding $24,000, refusing to award Mitsubishi all of its
request for court reporter and duplication fees. Indeed, the

-- 6 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
13
Bill of Costs Handbook states that the latter are “not
recoverable within the discretion of the taxation clerk unless
counsel has previously secured an order authorizing the
recovery of these costs,” which Mitsubishi had not.
Mitsubishi petitioned the district court to reinstate $21,000 of
the $24,000 in denied costs.
On February 3, 2003, we reversed the district court’s
setting the surety bond at $20,000, holding that a $1,000 bond
“will be adequate to protect the defendant’s interests.” Nemir
III, slip op. at 1. The day after we issued our opinion, the
district court vacated the Clerk’s award of costs and awarded
Mitsubishi the entirety of its original request—that is, $3,000
more than even Mitsubishi had asked to be added to the
Clerk’s initial award. Nemir separately appealed the district
court’s award of costs.
II. ANALYSIS
Nemir asserts that the district court erred in: (1) refusing to
instruct the jury on Nemir’s theory of strict liability;
(2) limiting Nemir’s discovery; (3) excluding an array of
evidence; and (4) awarding excessive costs to Mitsubishi.
A. Strict Liability
Nemir contends that the district court’s refusal to instruct
the jury about his theory of strict liability violated the
mandate that we issued in Nemir I. As we explained above,
in originally granting summary judgment to Mitsubishi, the
district court held that Nemir had failed to demonstrate that
the seatbelt suffered from an unreasonably dangerous “design
defect” that could underlie a cause of action based on strict
liability. In Nemir I, we reversed and identified numerous
items of evidence that “demonstrate[] a triable issue of fact
with regard to the design defect condition and the
unreasonably dangerous condition of the product.” Id. at 276.
14 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
Our prior holding that Nemir had created a triable issue as
to strict liability bound the district court. “[U]pon remand of
a case for further proceedings after a decision by the appellate
court, the trial court must proceed in accordance with the
mandate and the law of the case as established on appeal. The
trial court must implement both the letter and the spirit of the
mandate, taking into account the appellate court’s opinion and
the circumstances it embraces.” United States v. Moored, 38
F.3d 1419, 1421 (6th Cir. 1994). As we made clear in Nemir
I, the district court was required to allow Nemir to present his
strict liability theory to the jury.
Mitsubishi maintains, however, that our prior decision did
not reach this issue, contending that “[Nemir] is under the
mistaken belief that this Court’s opinion established the law
of the case on the issue of which Maryland law standard of
proof would apply, ‘inherently unreasonably dangerous’ strict
liability or the risk/utility . . . balancing test. However, it is
clear from the language of [Nemir I] that this Court did not
address that issue . . . .” Mitsubishi’s argument misrepresents
the repeated and unambiguous language of our prior decision.
We stated plainly, more than once, that Nemir had met his
burden on the unreasonably dangerous standard. See Nemir
I, 6 Fed.Appx. at 276 (“Plaintiff demonstrates a triable issue
of fact with regard to the design defect condition and the
unreasonably dangerous condition of the product.”); id. (“The
plaintiff’s evidence could lead a reasonable jury to conclude
that the product poses an inherently unreasonable risk.”); id.
at 277 (“[Horton’s] testimony creates a jury issue with regards
to whether or not the Takata 52 seatbelt is ‘unreasonably
dangerous’ under [Maryland law].”). The district court
ignored our clear and repeated instructions regarding the
submission to the jury of Nemir’s theory of strict liability, and
Mitsubishi’s argument to the contrary is plainly without
merit.
Nemir also argues that even if the strict liability instruction
were not compelled by Nemir I, the district court erred by

-- 7 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
15
basing its strict liability decision on the testimony of Cooper,
given that his employer counted both Mitsubishi and Takata
among its past clients, and by misapplying the law governing
unreasonably dangerous products. Given that the district
court’s decision to preclude the strict liability instruction was
erroneous on a more basic level—that it violated our mandate
in Nemir I—we need not address these arguments.
B. Discovery
Nemir’s claims that the district court improperly limited
discovery also descend from our prior review of this case. In
Nemir I, as part of our reversal of the district court’s grant of
summary judgment, we noted that “the district court
perfunctorily disregarded hundreds of pages of consumer
seatbelt complaints. This evidence was released to [Nemir]
in the eleventh hour, after years of discovery disputes over
whether or not such complaints were discoverable material.”
Nemir I, 6 Fed. Appx. at 275. Indeed, we explained that
“[t]his issue potentially impacts the safety of millions of
American consumers and [Nemir] should be granted
sufficient time to fully investigate [his] case.” Id. at 276.
Nemir argues that the district court failed to do so in a number
of ways.
1. Protective Order Limiting Discovery
Nemir contends that the district court erred in prohibiting
Nemir’s counsel from unilaterally contacting any of
Mitsubishi’s complaining consumers. Federal Rule of Civil
Procedure 26(c) authorizes such a protective order only under
circumstances “which justice requires to protect a party or
person from annoyance, embarrassment, oppression, or undue
burden or expense,” the potential for which must be
illustrated with “a particular and specific demonstration of
fact, as distinguished from stereotyped and conclusory
statements.” Gulf Oil Co. v. Bernard, 452 U.S. 89, 102 n. 16
(1981). The district court asserted that the order “was
16 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
necessary in this case to protect the defendants from being
unduly burdened and unfairly prejudiced if the plaintiffs were
able to . . . interview these customers ex parte.”
Yet the district court offered no relevant explanation of how
Mitsubishi would be burdened or prejudiced by Nemir’s
interviewing these consumers, nor is one apparent.
Mitsubishi argues that the restrictions protected its business
interests by shielding its customers from damaging innuendo
about the safety of its products. For support, Mitsubishi relies
upon two district court cases: Commodity Futures Trading
Commission v. Rosenthal & Co., 74 F.R.D. 454 (N.D.Ill
1977), and Vokswagen v. Westburg, 260 F.Supp. 636 (E.D.Pa.
1966). Neither supports Mitsubishi’s argument; the plaintiffs
in each of those cases had demanded unfettered access to all
of the defendant’s customers, even those who had never
complained about any defects in the defendant’s products. In
those situations, the courts hesitated to authorize fishing
expeditions that might have tainted the defendants’
relationships with their customers.
Conversely, the customers that Nemir sought to contact had
already complained to Mitsubishi of seatbelt malfunctions.
The relevance of these customers was already clear, and
interviews by plaintiff’s counsel would likely have served as
little more than a reminder of the seatbelt problems that the
consumers themselves had perceived and articulated.
Mitsubishi’s concerns, then, are speculative, and mere
speculation is insufficient to warrant such stringent limits on
pretrial investigation. See, e.g., Gottstein v. National
Association for the Self Employed, 186 F.R.D. 654, 659
(D.Kan. 1999) (“Defendants also suggest that they have a
legitimate business interest in preventing unrestricted contact
with their customers . . . . They present nothing but
supposition and conjecture, however, to support finding
future communications of counsel for plaintiffs will harm
their reputation or profits.”); Multi-Tech Systems v. Hayes
Microcomputer Products, Inc., 800 F.Supp. 825 (D.Minn.

-- 8 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
17
1992) (“Granting [plaintiff] access to the customers of
[defendant] that it has already discovered during depositions
. . . will level the parties playing field . . . . The court declines
to adopt [guidelines that the parties could follow in contacting
each others customers] because guidelines might place too
great a burden on the parties’ right to garner evidence in
support of the prosecution or defense of the claims in this
case.”); cf. Williams v. Chartwell Financial Services, Ltd.,
204 F.3d 748 (7th Cir. 2000) (“The plaintiffs have a right to
contact members of the putative class . . . . [t]he district
court’s decision as to [a] protective order must involve a
careful balancing of the potential for abuse created by the
class action and the right of the plaintiffs to contact potential
class members.”).
Moreover, the burden this protective order imposed on
Nemir’s investigation was considerable. Rather than being
able to conduct informal interviews, Nemir was required to
coordinate each and every witness interview with Mitsubishi,
as well as bear the cost, in both hours and dollars, of formal
depositions. Perhaps even more importantly, Nemir was
robbed of the privacy of his attorney’s trial preparation. As
the Supreme Court declared when refusing to compel the
discovery of counsel’s notes stemming from his ex parte
interview of witnesses, “the general policy against invading
the privacy of an attorney’s course of preparation is so well
recognized and so essential to an orderly working of our
system of legal procedure that a burden rests on the one who
would invade that privacy.” Hickman v. Taylor, 329 U.S.
495, 512 (1947). If Mitsubishi’s unsubstantiated fears of
prejudice justified a protective order, such orders would be
justified in virtually every case—a result that would obliterate
the essential principle articulated by the Court.
2. Limitation to 188 Complainants
Nemir further argues that the district court misinterpreted
our decision in Nemir I when it limited him to investigating
18 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
188 of the consumer complaints, rather than all 29,000.
Mitsubishi argued, and the district court agreed, that our
decision in Nemir I restricted Nemir to investigating only the
188 complaints that we explicitly mentioned in our decision.
But nowhere did our decision limit Nemir to those 188
incidents. Rather, in describing the relevance of the consumer
complaints to establishing a dispute of material fact, we noted
that “[a] complete review of the record indicates that
phraseology analogous to ‘intermittent partial latching’
occurred 188 times in the customer complaints submitted as
evidence. To summarily dismiss this evidence without
affording the plaintiff time to investigate and further elicit
information regarding the exact nature of these incidents is
clear error.” Nemir, 6 Fed.Appx. at 276.
Mitsubishi fails to mention that the reason we singled out
188 complaints in Nemir I was that the 188 complaints were
the only ones in our appellate record, and served as a sample
of the larger set of documents. In other words, we cited the
188 documents as examples of the 29,000-plus documents
that we deemed clearly relevant to the case and worthy of
further investigation. Mitsubishi’s argument that “[Nemir]
has given this Court no good reason to set aside its original
ruling and expand investigation discovery even further”
attempts to concoct a holding that is not even remotely
discernible from our opinion. See, e.g., First National Bank
of Cincinnati v. Pepper, 547 F.2d 708, 714 (2d Cir. 1976)
(Friendly, J.) (“Such a literal application of the language from
this court’s previous opinion . . . is, we think, not at all what
the earlier [panel in this case] had in mind.”); United States v.
Scheipe, 474 F.2d 720, 721 (3d Cir. 1973) (“We are not
prepared to give literal efficacy to isolated excerpts from
language of that opinion which would otherwise [contradict
its holding]”); Continental Motors Corp. v. Continental
Aviation Corp., 375 F.2d 857, 861-62 (5th Cir. 1967)
(reversing decision in which district court “gave too much
uncritical and literal significance to [an earlier decision]”).

-- 9 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
19
One caveat remains. When the district court initially
limited Nemir to discovery of the 188 incidents, it gave
Nemir the option of coming back to the district court to
request that it be allowed to investigate additional complaints.
Nemir never did so, and normally this would mean that he
waived his opportunity to interview the other complainants.
However, because we are already remanding for new trial,
and because our decision has overturned a separate discovery
restriction imposed by the district court, there is no reason to
deny Nemir the opportunity to investigate further.
3. Refusal to Sanction Mitsubishi
Nemir argues that the district court erroneously failed to
sanction Mitsubishi for obstructing discovery. Federal Rule
of Civil Procedure 37(b)(2) allows a district court to impose
sanctions against a party that “fails to obey an order to
provide or permit discovery.” Nemir argued that his
discovery requests included the 29,000 consumer complaints,
but that Mitsubishi persistently maintained that it had no
documents corresponding to his request. Nemir also asserts
that he was prejudiced by the delayed production, arguing that
the opportunity to contact many of these potential witnesses
evaporated during the delay, although he is noticeably vague
about the reasons for this: did most of the potential witnesses
move in the interim? But see Computer Task Group, Inc. v.
Brotby, 364 F.3d 1112, 1116 (9th Cir. 2004) (per curiam)
(“[F]ailure to produce documents as ordered is considered
sufficient prejudice.” (internal quotations and alterations
omitted)).
Although Mitsubishi contends that the district court
expressly denied Nemir’s motion for sanctions, the record
clearly reveals that the district court stayed consideration of
the motion, never to return to it. The district court’s failure to
rule on the motion requires that we remand for consideration
anew. See Bacou Dalloz USA, Inc. v. Continental Polymers,
Inc., 344 F.3d 22, 31 n.6 (1st Cir. 2003) (“Whether to impose
20 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
sanctions . . . is a question for the district court in the first
instance. We are not equipped to make a fact finding about
[the defendant’s] alleged bad faith and failure to comply with
discovery obligations.” (internal citation omitted)); cf. Mutual
Service Insurance Co. v. Frit Industries, Inc., 358 F.3d 1312,
1326-27 (11th Cir. 2004) (“Without explanation, it is
impossible for us to determine whether or not the district
court acted within the bounds of its discretion.”).
Mitsubishi also argues that our decision in Nemir I held that
no sanctionable discovery violation occurred, such that a
contrary determination is now precluded by res judicata. In
support, Mitsubishi points to the following passage from
Nemir I: “[T]he district court perfunctorily disregarded
hundreds of pages of consumer seatbelt complaints. This
evidence was released to the plaintiff in the eleventh hour,
after years of discovery disputes over whether or not such
complaints were discoverable material.” Nemir I,
6 Fed.Appx. at 275. According to Mitsubishi, our references
to “years of discovery disputes” constituted a “holding” that
the withholding of the documents was the result of a good-
faith disagreement, rather than misconduct.
This argument is baseless. Our decision in Nemir I did not
purport to address any of the issues regarding sanctions or
discovery misconduct. That we characterized, in passing, the
belated disclosure as coming “after years of discovery
disputes” does not mean that Mitsubishi’s hands were clean:
the fact that there were disputes does not automatically mean
that both sides in the dispute were playing fairly. (If
anything, the tenor of our opinion was critical of Mitsubishi,
noting that Mitsubishi had failed to disclose the documents’
existence until the “eleventh hour” and suggesting that
Mitsubishi withheld otherwise discoverable material by
excluding any documents that did not contain a set of
legalistic magic words. See id. at 275-76.)

-- 10 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
21
As for whether evidence of any discovery misconduct
should have been admitted at trial, we cannot evaluate that
claim without first knowing whether any discovery
misconduct occurred, and whether and to what extent any
misconduct prejudiced Nemir. To give one example, it may
be the case that if Nemir can show that Mitsubishi’s delays
prevented him from contacting a large number of relevant
witnesses who could testify to the partial-latching
phenomenon in their own seatbelts, then evidence of such
delay would be a proper rebuttal to Mitsubishi’s arguments
that Nemir has been unable to demonstrate that partial-
latching actually occurs in real life. There may (or may not)
also be other examples of prejudice, or other possible
remedies. All of this is for the district court to consider in the
first instance in conjunction with Nemir’s motion for
discovery sanctions.
C. Evidentiary Restrictions
Nemir highlights an array of allegedly erroneous
evidentiary exclusions, contending that the district court:
(1) improperly restricted the testimony of Horton;
(2) unlawfully prevented Manning from testifying at trial;
(3) mistakenly prohibited evidence about Mitsubishi’s failure
to investigate other similar consumer complaints; and
(4) hamstrung Nemir’s ability to impeach one of Mitsubishi’s
expert witnesses.
1. Horton Testimony
Nemir argues that the district court erred in prohibiting
Horton from testifying about: (1) the cause of Nemir’s
injuries; and (2) his tests on Nemir’s buckle. As to the
former, Nemir submits that the district court’s barring Horton
from testifying about the cause of Nemir’s injuries
contravened Nemir I. After noting that “Horton tested
Nemir’s belt buckle and determined that it could be placed
into a state of partial engagement by slightly depressing the
22 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
release button, and by improper or incomplete insertion,” id.
at 274, we held that
an expert witness’ conclusion regarding all admissible
evidence need not eliminate all other possible causes of
injury in order to be admissible on the issue of causation.
Accordingly, Horton could use this evidence to testify
that the partial latch of Dr. Nemir’s seatbelt caused the
damage in question. While other possibilities exist with
regards to causation, the fact that several possible causes
might remain uneliminated only goes to the accuracy of
the conclusion, not the soundness of the
methodology. . . . The district court, by requiring
“specific knowledge of the precise physiological cause”
of the accident held the expert up to entirely too strict a
standard when considering the admissibility of their
testimony. The testimony of Horton, establishing
causation based upon the above referenced facts, would
adequately shoulder the Michigan burden of proof that
plaintiff show a reasonable likelihood . . . that his
explanation of the injury is correct.
Id. at 275 (emphasis added, internal quotations and citations
omitted).
At various points in the opinion, moreover, we identified
the evidence that would permit Horton to reach his
conclusions. We noted, for instance, that
[d]uring his deposition, Horton testified that he had
inspected the Nemir seatbelt, and while he did not
“measure any loads,” he did “check it for partial
latching.” . . . This document clearly presents evidence
which would allow Horton to testify that Nemir’s
seatbelt could be put into a position of partial latch, and
that as such, the seatbelt was defectively designed.

-- 11 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
23
Id. at 271-72 (internal quotations and citations omitted). See
also id. at 274 (“Horton’s examination of the buckle and the
load marks on the jacket could lead a reasonable jury to
conclude that the seatbelt remained attached until impact,
when it released and retracted, causing damage to the
jacket.”). On remand, we instructed the district court that
Horton’s testimony regarding causation was admissible.
Despite our mandate, the district court again ruled that
Horton “cannot testify that partial latching [was] the sole
cause of Nemir’s injuries [because he] does not demonstrate
how he arrived at his conclusion . . . or how he eliminated
other possibilities, chief among which is the possibility that
Nemir was not wearing a seat belt.” Nemir, 200 F.Supp.2d at
776. Because we plainly held in Nemir I that the causation
testimony was admissible, the district court’s refusal to admit
this testimony contravened our explicit mandate.
Similarly, the district court prohibited Horton from
testifying about his testing of the buckle. Horton’s
testimony—that twice out of twenty attempts, using
purposeful manipulation, he was able to cause Nemir’s buckle
to unlatch—was a critical basis for our decision permitting
Horton to present his conclusions about causation. Although
we did not explicitly state that “Horton’s description of his
testing of Nemir’s buckle is admissible at trial,” we could not
have permitted Horton’s causation testimony without
permitting him to present the bases for his conclusions. See
Jones v. Lewis, 957 F.2d 260, 262 (6th Cir. 1992) (“[T]he trial
court may consider [only] those issues not decided expressly
or impliedly by the appellate court or a previous trial court.”)
(emphasis added).
Even if Nemir I did not directly foreclose the district court’s
decision to exclude Horton’s testing of the buckle, the
exclusion was nonetheless clearly erroneous. Federal Rule of
Evidence 702 provides that “[i]f scientific, technical, or other
specialized knowledge will assist the trier of fact to
24 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
understand the evidence or to demonstrate a fact in issue, a
witness qualified as an expert by knowledge, skill,
experience, training, or education, may testify thereto in the
form of an opinion or otherwise.” In Daubert v. Merrell Dow
Pharmaceuticals, Inc., 509 U.S. 579, 597 (1993), the
Supreme Court held that before admitting expert scientific
testimony, the district court must determine that the testimony
“rests on a reliable foundation and is relevant to the task at
hand.” Daubert applies “not just to scientific evidence but to
all expert testimony, including testimony based on technical
and other specialized knowledge.” Clark v. Chrysler Corp.,
310 F.3d 461, 466 (6th Cir. 2002), vacated on other grounds,
Chrysler Corp v. Clark, 124 S.Ct. 102 (2003). Because we
review the screening of expert evidence for an abuse of
discretion, “the district court has wide latitude to determine
the admissibility of expert testimony under Daubert”—but
“its discretion is not unbridled.” Hardyman v. Norfolk &
Western Railway Co., 243 F.3d 255, 267 (6th Cir. 2001).
The district court found that Horton’s method—in which he
manipulated the buckle at varying speeds and angles—was
scientifically unsound because “[n]o reasonable driver
purposefully manipulates the buckle at different speeds and
different angles to achieve a state of partial latch as Horton
does.” Yet the point of the purposeful manipulation was to
show that partial latching could occur under certain
circumstances, not to show directly that Nemir’s buckle
partially latched during the accident—given the infinite
possible variations, it would have been virtually impossible
to determine the velocity and angle with which Nemir had
actually buckled his belt on the day in question. Cf. Clark,
310 F.3d at 473 (“[I]t is not necessary to prove, as Chrysler
argues, that the prior accidents involved a vehicle identical to
the one driven by [plaintiff] or that all of the circumstances of
the accident are identical.”). That most of the combinations
did not produce partial latching might affect how heavily the
jury weighs the evidence, but not whether it should be
admitted.

-- 12 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
25
This purpose was narrow, but vital to Nemir’s case.
Although Horton was permitted to testify more generally as
to his conclusion that the buckle was capable of partial
latching, Mitsubishi’s central refrain to the jury when
disputing the buckle’s propensity to partially latch was that
Horton was just spouting words that were uncorroborated by
any actual test or demonstration. Thus, before Nemir could
convince the jury that partial latching occurred during the
accident, he had to show the jury that partial latching was
something that could have happened during the accident: in
other words, that his buckle was capable of partially latching.
Merely being allowed to announce his conclusion, sans
support, was unlikely very helpful to Horton’s testimony or
Nemir’s case. See Rosen v. Ciba-Geigy Corp., 78 F.3d 316,
319 (7th Cir. 1996) (“[A]n expert who supplies nothing but a
bottom line supplies nothing of value to the judicial process
. . . .”). And because an individual driver will inevitably
buckle his seatbelt at varying angles and with varying speeds,
purposeful manipulation would have been the only way to
demonstrate partial latching in a laboratory setting. See
Hardyman, 243 F.3d at 262 (finding abuse of discretion when
district court demanded a method of evaluation that was
scientifically impossible); Guild v. General Motors Corp., 53
F.Supp.2d 363, 366 (W.D.N.Y. 1999) (“[P]erfect identity
between experimental and actual conditions is neither
attainable nor required.”). Indeed, both of the experts that
were appointed by the district court used the same technique
as did Horton to test the buckle, a technique that was
appropriate for the conclusion that Horton was asserting.
The exclusion of these tests was exploited by Mitsubishi.
During closing argument, Mitsubishi argued that “they
haven’t shown partial latching to you except in a drawing.
They haven’t shown partial latching of an old buckle, a new
buckle, a used buckle, Dr. Nemir’s buckle, or any TK 52
buckle.” Mitsubishi went on to argue that “[i]t’s never been
demonstrated. It’s never been demonstrated. All you have is
Horton’s words of this. He has no basis to conclude the TK
26 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
52 buckle — can partial latch in a manner that’s
representative of actual use beyond his sheer speculation.”
Reinforcing the harm to Nemir’s case caused by the
prohibition, Mitsubishi argued that “[t]his is serious case.
You can’t prove your case with just the words of a hired
expert . . . Another federal court, we brought out, rejected
Mr. Horton’s testimony for something called, by the court,
parlor tricks. I submit that surely that is why he didn’t
attempt to demonstrate anything here.” Of course, Horton had
attempted to demonstrate something, and the district court’s
exclusion of this demonstration was error.
2. Manning Testimony
Nemir also sought to present testimony from court
appointed expert Lindley Manning, who, like Horton, was
able to cause Nemir’s buckle to partially latch. Federal Rule
of Evidence 706 provides that a court appointed witness
“shall advise the parties of the witness’ findings, if any; the
witness’ deposition may be taken by any party; and the
witness may be called to testify by the court or any party.”
(emphasis added) Yet the district court, without explanation,
prohibited Nemir from introducing Manning’s testimony.
Mitsubishi nevertheless argues that the “plain language” of
Rule 706 supports the district court’s decision. Yet the plain
language of the rule could hardly be more clear in its support
of Nemir’s position: either the court or any party “may” call
the witness to testify. The rule simply does not condition a
party’s doing so on the district court’s approval, nor does it
maintain merely that the district court “may allow either party
to call the witness.”
Manning’s testimony would have been important to Nemir,
for several reasons. First, it would have corroborated
Horton’s conclusions that the Nemir buckle was capable of
partially latching and had been demonstrated to partially
latch. Given that Manning was court-appointed, rather than
hired by Nemir, a jury would likely have considered his

-- 13 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
27
testimony to be more credible than that of Horton, whom
Mitsubishi derided as a “hired gun.” Second, Manning’s
testimony that even infrequent partial latching may constitute
a serious defect would have bolstered the weight afforded to
Horton’s research, which found partial engagement in two out
of twenty attempts. The erroneous exclusion of this evidence
again undercut Nemir’s case.
3. Mitsubishi’s Failure to Investigate Consumer
Complaints
Nemir argues that the district court erred in prohibiting him
from introducing evidence about Mitsubishi’s alleged failure
to properly investigate seatbelt-related complaints it had
received from other consumers. Evidence of Mitsubishi’s
failure to investigate would probably be inadmissible to prove
directly the proposition that the buckles are capable of
partially latching. And although such evidence would
probably be admissible to support a case for punitive
damages, see Clark, 310 F.3d at 481 (upholding award of
punitive damages because although aware of possible safety
problems with pickup truck, Chrysler failed to perform
relevant tests), Nemir’s punitive damages claim has been
dismissed.
But at the first trial, Mitsubishi opened the door:
Mitsubishi argued that Nemir was unable to unearth any
examples of partial-latching occurring in real life. In
response, it would have been important for the jury to know
that one of the reasons that he had been unable to deliver such
evidence was that Mitsubishi had failed to perform
appropriate investigations and/or testing that could have
yielded such evidence (with Cooper’s testimony confirming
that examination of the buckles would have been appropriate).
Cf. Beck v. Haik, — F.3d — (6th Cir. July 29, 2004) (“A
party may also be entitled to special instructions if he can
raise an issue of fact as to whether a party has failed to
preserve relevant evidence.”); cf. also General Motors Corp.
28 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
v. Blake, 515 S.E.2d 166, 169 (Ga. 1999) (“Fully aware that
the issue was why did the seatbelt not restrain plaintiff, if it
did not do so, GM apparently tactically decided not to
conduct a fuller and more complete investigation of the
apparatus. Defendant should not be permitted to avoid
learning the facts by lack of thoroughness and then claim
surprise.”); Daniel v. Indiana Mills & Manufacturing, Inc.,
103 S.W.3d 302, 315 (Mo. App. 2003) (“Plaintiff was
permitted, in rebuttal, to use [defendant’s crash test videos] to
demonstrate that initial unlatching could occur.”); Wiitala v.
Ford Motor Co., 2001 WL 1179610 at *7 (Mich. App. Oct. 5,
2001) (“We are also not persuaded by defendant’s argument
that the test documents uncovered after the violation were
irrelevant because all but one involved buckles other than that
found in plaintiff’s car . . . . given that defendant took the
position at trial that inertial unlatching is simply a ‘parlor
trick,’ we believe the evidence was relevant in that it directly
countered this assertion . . . .”).
If, during the new trial, Mitsubishi argues that the absence
of real-life incidents suggests that partial-latching is a
laboratory-only phenomenon, Nemir must be permitted to
introduce evidence that the absence of real-life incidents
might also signal only that Mitsubishi failed to investigate
such incidents.
4. Impeachment Evidence
Finally, Nemir seeks reversal based on the district court’s
interference with his impeachment of Dr. Hatsell. As we
explained above, when Nemir attempted to impeach Hatsell
on cross-examination by confronting him with an excerpt
from a book that had been co-authored by Spitz, another
expert retained by Mitsubishi, the district court instructed
Nemir to save such questioning for Spitz’s testimony. When
Nemir questioned Spitz about the book, the district court
ignored its prior instructions and prohibited such questioning
as beyond the scope of his testimony. In the end, Nemir was

-- 14 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
29
altogether prohibited from introducing this concededly
relevant impeachment evidence.
Mitsubishi’s defense of the district court’s decision is
spurious. First, Mitsubishi argues that the district court
correctly refused to allow the impeachment of Dr. Hatsell
with the book authored by Dr. Spitz because “there is no
indication from the record that plaintiff was doing anything
other than attempting to elicit an opinion from Dr. Hatsell as
to the possible opinions of Dr. Spitz regarding brain injuries
caused by rotational forces applied to the brain.” This is
technically correct but highly misleading: Dr. Spitz’s
published opinions about brain injuries were offered to
impeach the contrary conclusions of Hatsell, not to have
Hatsell comment on Spitz’s upcoming testimony. Nemir
sought to impeach Dr. Hatsell with the writings of Dr. Spitz;
his appearance on the witness list was beside the point.
Second, Mitsubishi argues that even though the district
court prohibited Nemir from impeaching Hatsell with the
writings of Spitz on the grounds that Nemir would have a
chance to elicit the impeaching information directly from
Spitz, “the district court reasonably and properly sustained an
objection to this collateral inquiry because it was not a subject
of Dr. Spitz’s testimony in this case.” Mitsubishi again
flagrantly misrepresents the record. Whether the questions
targeted Spitz’s testimony was irrelevant: the questions were
designed not to impeach Spitz, but rather to impeach Hatsell
by evincing a contrary opinion. The only reason that Nemir
was impeaching Hatsell with Spitz’s testimony—rather than
confronting Hatsell directly with the treatise written by
Spitz—was that the district court ordered Nemir to proceed
this way.
Thus, Nemir was erroneously prevented from introducing
relevant impeachment evidence. And this impeachment
evidence might have been particularly valuable to Nemir,
given that it came from Mitsubishi’s own expert.
30 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
D. Harmless Error
Mitsubishi argues that any errors made at trial were
harmless, given that the errors influenced only the jury’s
determination whether the seatbelt functioned properly,
whereas the jury found that Nemir had not been wearing his
seatbelt at all. Nemir argues, in contrast, that the jury’s
determination that he was not wearing his seatbelt was
infected by the errors regarding the effectiveness of the
seatbelt. In other words, posits Nemir, if the jury did not
believe that the seatbelt was ever capable of partially latching,
it would necessarily have attributed Nemir’s injuries to his
not having worn his seatbelt at all. When reviewing whether
a claim that an error was harmless, we must reverse “[i]f we
do not have a fair assurance that the trial’s outcome was not
altered by error.” Beck, — F.3d at —.
In arguing that Nemir was not wearing his seatbelt on the
day of the accident, Mitsubishi relied primarily on evidence
unrelated to the buckle. During closing argument, Mitsubishi
argued that Nemir’s memory of the accident was suspect,
thereby casting doubt on his claim that he was indeed wearing
the seatbelt. Moreover, Mitsubishi pointed to an array of
physical evidence, independent of the buckle, which it argued
showed that Nemir was not wearing his seatbelt, and also
argued that the evidence of seatbelt-wearing introduced by
Nemir did not show otherwise. It is clear, then, that the jury
reasonably could have found that Nemir was not wearing his
seatbelt at the time of the accident, and could have done so
even if it believed that Nemir’s buckle had the potential to
partially unlatch.
Yet the jury was not required to consider only arguments
that the parties raised during summation. It is a familiar part
of any judge’s instructions to a jury that it is bound by the
evidence, not by counsel’s closing arguments. See Federal
Jury Practice and Instructions, Civil, 3 FED. JURY PRAC. &
INSTR. § 101.02 (5th ed.) (“After the evidence is presented,

-- 15 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
31
the parties’ lawyers make closing arguments explaining what
they believe the evidence has shown. What is said in the
closing arguments is not evidence.”). Although Nemir’s
argument would certainly be stronger if Mitsubishi had more
frequently argued that the lack of seatbelt deficiency was
itself proof that Nemir was not wearing his seatbelt, that
Mitsubishi’s counsel did not stress this argument does not
mean that the jury did not consider it.
In any event, during the opening statement, Mitsubishi told
the jury that “at the end of the case, you must decide [whether
Nemir] engaged in the very difficult machinations that the
evidence will show are necessary to cause partial latching or
partial engagement, or that he simply didn’t put his belt on.”
Given that an opening statement is “an outline to help [the
jury] understand what each party expects the evidence to
show,” id., the jury could very well have viewed the evidence
at trial in light of the either-or framework presented to it by
Mitsubishi’s opening statement. And if the jury did view the
evidence this way, deficiencies in evidence about the buckle’s
malfunction could have led the jury to conclude that the
seatbelt simply could not have malfunctioned, compelling the
conclusion that Nemir was not wearing it.
Although we are not to presume that a jury disregarded the
district judge’s instructions when answering the questions in
a special verdict form, see Chute v. Roebuck and Co., 143
F.3d 629, 632 (1st Cir. 1998), we need presume no juror
malfeasance to conclude that the jury’s views about the
buckle’s reliability may have influenced its determination that
Nemir was not wearing his seatbelt. As a matter of logic, the
jury could have reasoned that: (1) The only way Nemir could
have received his injuries if he was wearing his seatbelt was
if the seatbelt was defective; (2) the seatbelt was not
defective; therefore (3) Nemir was not wearing the seatbelt.
As the Seventh Circuit put it, when reviewing a jury’s verdict
in a seatbelt case with similar facts, “[e]ngineers trying to
understand a disaster often follow causal chains (“failure
32 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
trees”) until they find one that can account for the calamity.
Sherlock Holmes observed that ‘when you have eliminated
the impossible, whatever remains, however improbable, must
be the truth.’ Courts need not disdain a method that both
engineers and detectives find useful.” Bammerlin v. Navistar
Int’l Transportation Corp., 30 F.3d 898, 902 (7th Cir. 1994)
(Easterbrook, J.) (internal citation omitted).
Our concern that the errors were not harmless is heightened
by the case of Babcock v. General Motors Corp., 299 F.3d 60,
62 (1st Cir. 2002), a case in which the plaintiff alleged that
his seatbelt—due to the same partial-latching phenomenon
alleged by Nemir—unlatched during an accident. Like our
case, the plaintiff in Babcock had to convince the jury both
that the driver had been wearing his seatbelt and that the
seatbelt was defective. As to the latter, the plaintiff was
bolstered by the testimony of an expert who had tested the
driver’s buckle. See id. at 68. Although the evidence that the
plaintiff was wearing a seatbelt was arguably weaker than it
was in our case—the plaintiff’s case was based only on
testimony about the driver’s habit of always wearing his
seatbelt, whereas Nemir also had some supportive physical
evidence—the jury found both that the driver was wearing his
seatbelt and that the seatbelt was defective. See id. at 66. Of
course, for a host of reasons, this does not compel the
conclusion that in our case, the jury’s determination whether
Nemir wore his seatbelt was affected by the lack of evidence
regarding the efficacy of the buckle. But Babcock does
reinforce our conclusion that, as a matter of logic, erroneous
evidentiary rulings pertaining to one question had the
potential to affect the jury’s resolution of the other.
Nor does Grimes v. Mazda North American Operations,
355 F.3d 566 (6th Cir. 2004), suggest otherwise. In Grimes,
a case based on similar facts, we held that any error that might
have arisen from the exclusion of evidence about prior
accidents involving defective seat belt latches was harmless,
because “[t]he jury returned a special verdict that plaintiff was

-- 16 of 17 --

Nos. 02-1780; 03-1228 Nemir v. Mitsubishi
Motor Corp., et al.
33
not wearing her seatbelt and that this, not design defect, may
have been a cause of her injury.” Id. at 573. Yet in Grimes,
the plaintiff never argued to us that one set of errors affected
the jury’s analysis of another question, such that the question
before us in today’s case was simply not before us in Grimes.
“Questions which merely lurk in the record, neither brought
to the attention of the court nor ruled upon, are not to be
considered as having been so decided as to constitute
precedents.” Webster v. Fall, 266 U.S. 507, 511 (1925). See
also United States v. Rivera-Guerrero, — F.3d — (9th Cir.
Jul. 20, 2004) (relying on quoted language). In any event, the
evidence pertaining to seatbelt wearing was far more one-
sided in the defendant’s favor in Grimes, given that the
plaintiff had told an emergency technician and two nurses that
she had not been wearing her seatbelt. Thus, even if we had
expressly considered such an argument in Grimes, the facts in
the present case much more strongly suggest that the errors
affected the trial’s outcome.
Ultimately, because we simply do not know whether the
errors pertaining to the functioning of the seatbelt affected the
jury’s determination of whether Nemir was wearing the
seatbelt, we cannot say with confidence that the errors at trial
were harmless.
E. Costs
Mitsubishi challenges the costs awarded by the district
court to Mitsubishi. Under Federal Rule of Civil Procedure
54(d)(1), “costs other than attorneys’ fees shall be allowed as
of course to the prevailing party.” Because our reversal of the
judgment for Mitsubishi strips it of its status as the
“prevailing party,” the district court’s award of costs is set
aside.
34 Nemir v. Mitsubishi
Motor Corp., et al.
Nos. 02-1780; 03-1228
III. CONCLUSION
The judgment of the district court is REVERSED, the
award of costs VACATED, and the case REMANDED for
a new trial. Finally, the case is to be assigned to a different
district judge for all further proceedings in this case. To
minimize confusion about the required course upon remand,
we summarize the specifics of our decision: (1) Nemir should
have been permitted a reasonable opportunity to contact, free
of the previous restrictions, the other consumers who filed
complaints with Mitsubishi; (2) the district court should have
ruled on Nemir’s motion for sanctions based on Mitsubishi’s
alleged violations of discovery rules; (3) Nemir should have
been permitted to present his strict liability theory to the jury;
(4) Horton should have been permitted to testify both as to
causation and to his testing of Nemir’s buckle; (5) Nemir
should have been permitted to call Drs. Cooper and/or
Manning to testify at trial; (6) Nemir should have been
permitted to impeach the testimony of Dr. Hatsell with the
writing and/or testimony of Dr. Spitz; and (7) Nemir should
have been permitted to offer, in response to Mitsubishi’s
arguments that partial-latching has never happened in the real
world, evidence that it declined to perform tests when alerted
by consumers of the potential for its buckles to unlatch.

-- 17 of 17 --