When Science Fails Us
& We Fail Justice: A Conversation About the Tragic Case of David Gavitt
By Colin W. Maguire, Publicity Editor, Thomas M. Cooley
It is not often that the law jumps out, grabs you, and
forces you to pay attention. The story of David Gavitt is now well-documented
in Michigan and around the country.
In short, David Gavitt was convicted of setting a fire at his home which killed
his wife and two small daughters. While he lay in a hospital bed recovering
from injuries sustained trying to save them, police collected evidence that
would lead them to believe that David Gavitt was an arsonist who murdered his
own family in cold blood.
Bad science in his case, but also in the arson science
community at large, led directly to David’s wrongful conviction for three murders.
David sat in jail for over twenty years before the University of Michigan’s
Innocence Clinic took up his case. David was ultimately exonerated and allowed
to finally visit the graves of his family, after decades.
Imran Syed worked on David Gavitt’s case from the time
he was a law student at the University of Michigan. He now works as the Staff
Attorney at the University of Michigan Law School’s Innocence Clinic.
Mr. Syed graciously agreed to sit down with me for an interview on September
27, 2012. I cannot thank him enough for the work he has done, and for giving me
his valuable time.
We also owe a great deal of thanks to the Detroit Legal
News and the Ingham County Legal News for publicizing this case and bringing
the particularly outstanding work of the University of Michigan Law School’s
Innocence Clinic to our attention here at the Thomas M. Cooley Law Review.
The following is an edited transcript of a conversion
involving one of the most fascinating and emotionally wrenching cases I have
ever delved into. What happened to David Gavitt, and what Mr. Syed and his team
found, tugs at the very base of our legal system because it forces us to
question our own certainties. What will they say about us in 20 years? We can
only pray it is not what this conversation reveals about courts and arson
science from the 1980s to mid-1990s.
When did you first meet David Gavitt and at
what stage was that in his legal process?
We met David in the fall of 2010. Let me give you some background
on David. David applied for legal assistance to the clinical programs at the
University of Michigan Law School before there even was an innocence clinic
here. Our clinic opened in 2009, but he applied to the general clinical program
in 2007 or 2008. Initially, the general clinic did a lot of research on his
case in order to better understand the arson science behind David’s conviction.
This process was difficult because the general clinic did not have an arson
expert at its disposal.
After the Innocence Clinic opened, the general clinic
transferred the case to us. A group of students in the Innocence Clinic really
began getting deep into the case in late 2009. Those students accomplished an
important first step – successfully acquiring lots of documentation from the State
of Michigan. In 2010, the students passed the case over to me and my partner at
the Innocence Clinic (with whom I later wrote an academic article about this
case). We also worked with an attorney out of Atlanta who specialized in arson
cases, Mike McKenzie. Mike almost immediately told us there were many “red
flags” in the case which indicated that David’s conviction was a bad
conviction. After learning this, we brought in some arson experts to look over
the case and they confirmed Mike’s suspicions.
Initially, we did not tell David Gavitt about all of the
steps we were taking. Up to that point, we were not sure whether David had a
good case; we were still learning about arson science. It is unconscionable in
a situation like David’s, where he was already incarcerated for twenty years,
to tell him “you may have something good” and then have the whole case fall
through. We see a lot of cases that are promising, and then fizzle out. So, it
was not until the fall of 2010 that we really got through the boxes and boxes
of evidence. Plus, we had to educate ourselves on arson science. In fact, I
still have a copy of John J. Lentini’s book Scientific Protocols for Fire
my desk. This book discloses many errors which arson experts routinely
committed in the period when David was convicted.
Upon reading this book and looking at David Gavitt’s
case, we realized that David was only convicted on “science.” However, the
“science” used was not accurate methodology at all. In David’s case, there was
never any motive established. In police work, the investigator usually tries to
establish a motive, but here they only relied on “science.” In skipping
standard investigatory steps, the police and prosecution missed the most
obvious explanation – that the fire which killed David’s family, and severely
injured David, was an accident. Will we ever know 100% if the fire was an
accident? No. Yet, experts have now come to a consensus that there is
absolutely no evidence that this fire was arson, so we have to conclude it was
Once we realized David was convicted on bad “science,”
we made a visit to David in prison in October 2010. I was a student at the
time. Me, two other students, two professors who are co-Directors of the
Innocence Clinic, and Mike McKenzie all met with David. The story he told us
matched perfectly with the story he first told the day he was recovering in the
hospital from injuries sustained in the fire. David’s story never deviated over
the years. All David knew is that he did not set the fire, but he thought
somebody else did.
The reason he thought another person set the fire was
because the State presented evidence that gasoline was found on two small
pieces of the Gavitt’s living room carpet. We obtained the evidence used to
arrive at this conclusion — gas chromatogram charts — from the State Police.
These gas chromatograms were the centerpiece of the State’s evidence because
they “proved” that a fire was purposely set in the house. Once they were
introduced into evidence, it hardly mattered to the jury that there was
absolutely no motive for David to kill his wife and two daughters.
Tragically, when we asked numerous experts to review the
gas chromatograms they all came to the same conclusion – there was no
gasoline on the carpet samples. The charts were flat-out misread. The
person who read the charts for the trial was not qualified to do so, and made
many errors in the testing and reading.
We told David that there was no gasoline on the carpets
when we met him at that meeting in October 2010. David had trouble believing
this fact because he sat in prison for over twenty years “knowing” that
somebody else set his house on fire. Needless to say, David was blown away. We
also told him the conclusion of our experts that all the science used to
convict him was essentially junk science. Still, we had to temper his
expectations and explain that, even in a best-case scenario, exoneration is a
long process. After that meeting, it took us another four months of intense
research and writing to get the motion and brief written. Those were completed
in early 2011. Around a year later, David was released.
When you started working with David Gavitt,
he had already served over twenty years in prison for a crime he did not
commit. Worse than that, he was convicted of killing his own family. Is it
possible for an attorney who has to make very formal legal arguments and think
analytically to not get wrapped up emotionally in a case like this?
To not get emotionally involved is difficult. The way I
dealt with it is this: you cannot look at the big picture until you get the
little pieces in order; otherwise, you will not actually help this person. We
took a very analytical approach to the whole situation. First, what is arson
science? Second, what are the real scientific facts of the case? Third, how
does this all work together? It was only after the October 2010 meeting that
the immense tragedy of what happened here really hit me. David’s case is unlike
any case the University of Michigan Law School’s clinical programs has ever
seen, and unlike most wrongful convictions.
Not only did David wrongfully lose twenty-seven years of
his life, but he was incarcerated on the heels of losing his wife and two
daughters. That is a tragedy that few, if any, people can ever imagine. I tell
David that even inmates who wrongfully serve twenty or thirty years are not
necessarily in his position because they often have families on the outside
that will come and see them. David lost his family, was blamed for the death of
his family, and became a pariah to many people who knew him. Still, you cannot
let the magnitude of this situation hit you as you are doing the work itself.
Unfortunately, arson cases tend to involve these types
of emotional situations. The Innocence Clinic has already seen two cases very
similar to David’s where a person is convicted of killing their family. Despite
that tragedy, we need to examine the technicalities of those cases. Of course,
not everyone convicted in an arson case is innocent. Still, those that are
innocent deserve their day in court because we know there are serious questions
about the science used to convict them (discussed below). We need to use the
legal system to correct these wrongs.
Was faulty forensic science always the focus
of David Gavitt’s appeals for exoneration?
From our standpoint, science was always to key to
reversing his conviction. However, David was not thinking the way we were early
on because David was not aware that there was no gasoline on the carpet samples
presented at his trial. Therefore, he focused on finding another suspect – who
turned out not to exist. Though, he did understand that other parts of his
trial involved junk science through his own research while incarcerated. Even
at the October 2010 meeting, he was asking about whether there was gasoline
trail or if a gas can was found.
We saw our focus on scientific evidence as an advantage.
Usually, Innocence Clinic cases rely heavily on a person’s memory from many
years earlier, on subjective motivations, or on working with an attorney you
are accusing of ineffectively representing their client. Here, we had explicit
evidence that bad science was used. Furthermore, we had six experts come to the
conclusion that the arson science used to convict David Gavitt was faulty.
Three experts worked for us, and three worked for the State. I hope the
scientific basis of arson conviction appeals makes them easier to work on for
both sides because you are working with set data.
The scientific nature of the case also had a positive
impact on public opinion about the case and, hopefully, others. Here, there was
no “he said, she said” situation. Rather, everyone who works in this field now
realizes that grave mistakes were made in the “science” of arson investigation.
Even though not everyone convicted of arson using bad science is innocent,
everyone does deserve a reevaluation of their case.
You co-authored a fantastic article on
shifting forensic science and its effect on court decisions, with a focus in
arson analysis (Caitlin Plummer & Imran Syed, Shifted Science and
Post-Conviction Relief, 8 Stan. J. Civ. Rts. & Civ. Liberties 259 (2012)).
In it, you document how outdated arson science was. In perhaps the best
example, you show how, until recently, arson experts were unaware of the
“flashover” phenomenon in building fires – a process wherein a room becomes so
hot that the air combusts and leaves marks similar to those left from
accelerant use. Are you more fascinated or frightened that this phenomena was
unknown to investigators for decades?
This is something I have wondered about – how could this
go unnoticed for this amount of time? I have tried to talk with arson expert
John J. Lentini about it, and others, but it still blows my mind. Strangely,
there was not a lot of testing in arson science. Theories were merely based off
what sounded right and what we knew about fires that burned on smaller scales.
For instance, we “knew” that fires burned upwards, so people concluded that the
starting point of a fire is the lowest point of origin. However, science is
more than common sense, and experiments were needed. Flashover was the missing
link that was not understood in the 1980s. There is a great YouTube
video that shows a flashover simulation occurring in matter of two minutes; you
can see there that fires may burn upward initially, but very quickly consume
the whole room.
The whole situation is frightening because hundreds of
people were convicted on this junk science in the 1980s and even into the 1990s.
Still, we have to thank the core group of experts who asked the tough questions
and did the research. The flashover effect is something a lay person would not
rationally perceive. You put a smoldering cigarette on a couch, and three
minutes later windows are exploding and the structure is starting to collapse.
The average person thinks, “something crazy and malicious must have happened
here.” In fact, it could just be a cigarette on a couch.
It is really, really sad that people were convicted on
evidence that was not properly tested. True, new science has replaced this junk
science in most situations. Yet, the Innocence Clinic’s work indicates a
handful of Michigan cases in the last five to ten years where junk science was
used in an arson conviction.
Your article argues the Daubert test is a
reasonable step in the right direction but not enough. Can you summarize your
proposed test for expert evidence?
I think Daubert is a perfectly feasible and
appropriate standard for expert testimony. What we say in the article is that Daubert
is limited to the science available at a certain time. Also, Daubert
only works if it is applied correctly and we have steps like peer-reviewed
studies. Courts vary in how faithfully they apply Daubert, but there is
little doubt that Daubert has kept out lots of bad science.
Still, science evolves over time. Daubert does
not really allow a way for a review of bad science, after the fact. Therefore,
courts and state legislatures need to find other avenues by which to recognize
that a certain form of science may have been acceptable at the time of the
conviction, but is later proved to be junk. There have to be procedures in
place that allow defendants in such a situation to have their cases
reevaluated. So, in absolute cases of shifted science, like arson or
shaken-baby syndrome, you should have a right to get back into court. This is
not far-fetched because we already have a similar statute for DNA evidence.
We, as a people, cannot slam the door on those who we know were convicted using
science we now know was junk. If their incarceration is an ongoing event, then
the constitutionality of that incarceration is an ongoing question. Hopefully,
cases like David Gavitt’s can bring publicity to this issue.
If we have competing scientific theories
regarding forensic evidence, is there a risk that we turn the trier of fact
into a statistician who weighs the value of the theories; or is a single
competing theory enough to render an entire area of forensic science suspect?
I would not say that. An entire scientific field should
not fall into unreliability only because of one competing theory. The basic
premise here is that the jury deserves to see the full picture. This allows the
jury to make a proper decision.
In State v. Edmunds, a famous shaken-baby
syndrome case in the Wisconsin Court of Appeals,
that court took ground-breaking steps toward recognizing that scientific shifts
constitute new evidence that a court can consider. That court stated that it
was important that new mainstream theories had emerged, and that the majority
of people believed shaken-baby syndrome was not what scientists originally
thought it was. Once the new scientific shift emerged, it changed the entire
weight of the evidence in the case. At the time of Ms. Edmunds’s original
trial, something like nine out of every ten experts believed that if X, Y, and
Z signs are present, then there had to have been physical abuse of the child.
Years later, the consensus has shifted, and only about three or four out of ten
experts would say the same thing. That court decided that a jury needed to hear
about the changed consensus and the competing theories. For that reason it
threw out Ms. Edmunds’s original conviction — we now know it was based on junk
In David’s case, at trial nobody challenged the
determination that an arsonist started the fire. David’s own attorney, who was
a good, well-paid defense attorney, had no basis to challenge that finding.
David’s only defense was to blame somebody else. The jury had no other
information than it was arson. Yet, now we know that the jury did not receive
all the information; there were other explanations, and the signs that were
said to point to arson could actually be read another way as well. Not a single
expert today would say the fire was arson. Therefore, we argue that the case
should go back to court where a judge can reevaluate it and ask, “Was this
conviction proper based on what we know now?” Still, the judge has the ability
to view this issue very narrowly, and this is why our paper proposes a state
statute for older arson and shaken-baby syndrome cases that explicitly states a
right to get back into court.
Overall, we may need to present juries with competing
theories. In arson cases, we know juries weren’t told the full story in the
1980s, and that led to some wrongful convictions.
You use the hypothetical accused arsonist
“Dennis” in your article. Dennis’ timetable is different from David Gavitt, but
was David Gavitt the main inspiration for Dennis?
Yes. At one point, the character was actually named
“David.” And the hypothetical varies at different points of the article so that
we can make different points. But yes, the initial facts we present for
“Dennis” are David Gavitt’s facts.
We wanted to play with the timeline to show that bad
arson science means different things depending on when you were convicted. In
David’s case, the science was awful when he was convicted, but there was a
consensus among arson experts. Therefore, there is no way a jury could have
considered evidence that David’s case was not arson. Ironically, this puts him in
a very strong position in appealing his conviction today because we can say
there has been a clear scientific shift.
Unfortunately, somebody convicted on bad arson science
in 1995, for instance, is in a worse position because alternative arson
theories were emerging and available. If you convicted using bad arson science
in 2007, better science was definitely available. In both cases, the State
could reasonably argue that the defense should have presented the better
science at trial. This argument if accepted, could cut off a conviction
appeal. Then you have to rely on more nuanced arguments like ineffective
counsel or judicial error in failing to apply Daubert.
The 1995 example is the most challenging case because
the good science was available, as was the bad science. Yet, a judge would not
clearly violate Daubert in allowing the bad science in. Also, the jury
could have seen the good science; the defense simply never brought it to their
attention. In fact, we are working on such a case right now.
Beyond litigating these suspect convictions, the
Innocence Clinic’s main goal is to publicize cases like David Gavitt’s to
educate legal practitioners and the public, how incredibly tragic but fixable
these situations are. Though Michigan, in particular, has good court rules to
review cases using bad arson science, reviews are always subject to a judge’s
discretion, absent a controlling statute. Still, we hope judges will let these
cases back into court because when there is an injustice in these cases, the
injustice is almost unfathomable. As a legal community, we must move quickly to
correct such injustices. Thankfully, the prosecutor in David’s case worked with
us and helped expedite the process.
We won’t ever bring cases to court where there are
serious questions about the claim of innocence, even if the science used in the
case was bad (that’s not a case for an “innocence” clinic). Still, everyone
deserves a level playing field and that is a basic tenet of our judicial
system, so there should be a procedural way for everyone convicted on bad
science to get back into court. Bad arson science is an issue, like DNA, that
can be dealt with and then we can move on and not make the same mistakes again.
 Mich. Comp. Laws § 770.16 (1979), available at
 State v. Edmunds , 308 Wis.2d 374 (Wis. Ct.
App. 2008), available at