Evidence law spends enormous effort on what happens to an item after someone picks it up. It has almost nothing to say about how the item came to exist. Four Detroit cases, all documented in public records, show what lives in that gap.
Chain of custody establishes that the item offered at trial is the item that was collected. It verifies handling, not origin. A confession containing details only the questioners knew, a diagram attributed to the wrong hand, a kit that was collected and never tested, and a laboratory conclusion that was wrong when written all pass through the chain intact. This series examines four Detroit cases where the record was made rather than found.
Is a confession evidence of guilt or a record of an interview?
Legally it is a statement by a party. Practically, juries treat it as the strongest evidence available. The gap between those two characterizations is where the case examined in Part II sits.
Why does attribution matter so much?
Because an item’s meaning often depends entirely on who produced it. A crime scene diagram drawn by an investigator is a work product. The identical diagram drawn by a suspect is a confession. Nothing about the paper changes.
Is untested evidence the same kind of failure?
It is the inverse of the others. In the first three, a record existed and was wrong or misattributed. In the fourth, a record was collected from a person’s body and then never became evidence at all.
Are these cases connected to each other?
Not as a conspiracy and not through shared personnel in any way this series asserts. They are connected structurally, in that each shows a different point where the evidentiary record can fail before anyone contests it.
Why I Am Starting With a Doctrine
There is a phrase that appears in nearly every account of a criminal case, usually in passing, usually as reassurance. The chain of custody was maintained.
It is a real and important doctrine and I do not want to be glib about it. Evidence gets logged, sealed, initialed, transferred, and accounted for, and the point is to guarantee that the bag opened in front of a jury holds what the officer picked up off the floor. Where the chain breaks, cases fall apart, and they should.
What interests me is the starting point. The chain begins when someone collects something. Everything upstream of that moment, every decision about what to collect, what to write down, what to ask and how, and what a collected item will be said to mean, occurs before the doctrine attaches.
So the question this series asks is not whether Detroit’s evidence was handled properly after collection. In the cases examined here, mostly it was. The question is what the record was permitted to be at the moment it came into existence.
Four Places It Fails
Reading the Detroit cases against each other, the failures sort cleanly into four categories, and they are not variations on one problem. They occur at different points and would require different fixes.
A record is generated inside the investigation and then treated as though it came from outside it. A statement containing details only the questioners possessed is the clearest form.
Part II. The document is authentic. Its contents did not originate where the record implies.
A real item is credited to the wrong hand. The object is genuine, the chain is intact, and the meaning is entirely a function of who made it.
Part III. Resolved by testimony rather than by anything on the document itself.
Evidence is collected, stored, and never analyzed. It sits in custody, perfectly preserved, and never becomes information.
Part IV. The chain held. Nothing else did.
Evidence is examined by a laboratory whose conclusions are later found unreliable. The item is real, the testing occurred, and the result was wrong when it was written.
Parts V and VI. The failure is in the reading, and it propagates silently across every case the reader touched.
Note that only the fourth resembles what people usually mean by a forensic scandal. The other three involve no laboratory at all, and two of them involve no scientific question of any kind.
Evidentiary safeguards are concentrated on the part of the pipeline that is easiest to observe. Custody logs, sealing procedures, and transfer documentation all address the interval between collection and courtroom. The four failure modes above occur at the two ends, before collection and during interpretation, where no comparable documentation regime exists.
The Clutch Justice course track covers working from primary documents, reconstructing what a file shows against what testimony claimed, and identifying where a record was generated rather than gathered.
Browse the CoursesWhy Detroit
I want to be careful here, because a series titled after one city invites a reading I do not intend.
The argument is not that Detroit police were uniquely dishonest or that Detroit courts were uniquely careless. I do not have the comparative data that claim would require, and neither does anyone else, for reasons this series will eventually get to.
The argument is that Detroit’s record is unusually complete. Four separate failures, at four different points in the evidentiary pipeline, were each documented well enough by outside institutions that they can be reconstructed from public sources. A federal DNA testing effort. A state police reinvestigation. A city and county inventory of untested kits. A state police audit of a municipal laboratory.
That completeness is itself an accident of circumstance. Each of those investigations happened because a particular institution took a particular interest at a particular moment. None of them was routine, and in most jurisdictions none of them would have occurred.
So Detroit is not the place where the most went wrong. It is the place where enough was written down afterward that the shape of the problem is legible.
A city with four documented evidentiary failures looks worse than a city with none. What it actually has is four completed outside investigations. The comparison people instinctively make is between places, and the real comparison is between what was examined and what was not.
The Four Cases
Briefly, so the shape of the series is visible from the start. Each gets a full installment.
In 1984, a sixteen-year-old girl named Michelle Jackson was raped and murdered in Detroit. Eddie Joe Lloyd, a man hospitalized and receiving treatment for mental illness, was interrogated repeatedly and produced a statement of remarkable accuracy about the crime scene. He was convicted in 1985 and excluded by DNA testing in 2002, after seventeen years.
In 2007, four people were shot on Runyon Street. Davontae Sanford, fourteen years old, was picked up in his pajamas during a neighborhood canvass and questioned across more than twenty-four hours without a parent or attorney. A crime scene diagram became a central piece of the case against him. A state police reinvestigation completed in 2016 concluded that the deputy chief who testified the boy drew it had drawn it himself.
In 2009, roughly eleven thousand untested sexual assault kits were found in a Detroit police storage facility. Each represented a person who had submitted to an invasive forensic examination on the understanding that it would be analyzed.
In 2008, the Detroit police crime laboratory was closed after a Michigan State Police audit found a ten percent error rate in reviewed firearms cases along with poor record keeping, lax evidence control, and inadequate training. Roughly thirty-one thousand firearm-related prosecutions were implicated.
In all four, the correction came from outside the institution that produced the record. A federal testing initiative and an innocence organization. A state police agency reinvestigating at a prosecutor’s request. A county prosecutor’s inventory and outside funding. A state police audit. No component of the Detroit Police Department identified any of these failures through its own internal processes.
What This Series Will Not Argue
Three clarifications, stated now so they do not have to be repeated in every installment.
This series does not argue that any person discussed here is guilty of anything not established in a legal proceeding. Where allegations were made and not adjudicated, they will be described as allegations. Where an official was investigated and not charged, that will be stated plainly and so will the reason.
It does not argue that the people who produced these records set out to produce false ones. Several of the failures are consistent with people applying their training exactly as taught. That possibility makes the structural problem worse rather than better, because training is easier to replicate than malice.
And it does not argue that the underlying crimes did not occur or did not matter. Michelle Jackson was killed. Four people died on Runyon Street. Eleven thousand people were assaulted. In each case, the evidentiary failure was also a failure to establish what happened to them.
Why This Matters
The reason to start with the doctrine rather than with a case is that the doctrine explains why these failures were survivable for so long.
A defense attorney can challenge a chain of custody, and courts take that challenge seriously, because the doctrine supplies a standard against which handling can be measured. There is a log. Either the transfers are documented or they are not.
There is no equivalent standard for how a statement’s contents came to be in it. There is no log of what a suspect was told before they said something. There is no independent record of who put pen to a diagram. There is no report generated when a collected kit is not sent for analysis, because nothing happened, and institutions do not document the absence of an action.
So the strongest safeguard in evidence law protects the stage that is easiest to observe, and the stages that produced every case in this series are governed by nothing comparable.
That is not an accident of drafting. Handling is observable and origin is not, and legal systems build rules around what they can verify. But the consequence is that a record made in a room with two people in it, one of whom is in custody, arrives in court carrying the same presumption of authenticity as a sealed evidence bag.
Part II takes up Eddie Joe Lloyd, interrogated in a psychiatric hospital and allowed to believe that a confession would help police identify the real killer. His statement described the victim’s clothing, the ligature, and the object left at the scene with striking accuracy. Every one of those details was already known to the men asking the questions.
Continue Your Investigation
If this reporting raised more questions, use the Clutch Justice ecosystem to keep going.
