Wrongful Conviction: Autopsies of Investigations that Failed

Henry Waldock
Last updated: 2021.10.22


"Those who cannot remember the past are condemned to repeat it." George Santayana 1906.

Investigators need to answer the questions "whodunnit?" "what happened?" and "why?" Good investigators always look for ways to find answers. Some techniques work better than others.

Busting the wrong person means:

In the early 1990's a window of opportunity opened. The arrival of DNA technology permitted interested parties to review whether the justice system convicted only guilty people. Old exhibits were re-examined using the new technology. To the surprise of many, it demonstrated that courts convicted some innocent people of serious offences.

Shocked people asked "how did this happen?" They found answers which tend to distinguish reliable investigative techniques from unreliable ones.

That window closed. Now, investigators use DNA technology as part of their investigations. We no longer have a magic new technology with which to double-check old convictions.

But the answers remain relevant today. We still use many of the investigative techniques which led to those convictions. We need to know their flaws.

Common failures of failed investigations include:

Issue Cases
Tunnel vision Failure to investigate properly alternate suspects or theories Sophonow, Johnson, Milgaard, Marshall
Jailhouse informants Some cons will say whatever you want to hear, convincingly. Sometimes, they will lie for no apparent reason. Sophonow, Morin,
Bad science Some scientific techniques turned out to be unreliable, for example, hair and fibre comparison. Morin
Skipping the good science Some scientific techniques work well, if you bother to use them. Sophonow
Overzealous interviewing Some people will believe or repeat what you say, if you tell them forcefully enough. Druken, Sophonow, Milgaard(?), Marshall
Gaps in disclosure If defence doesn't know what will reveal innocence, they can't use it. Sophonow,
Disorganized investigation Losing evidence and leads results in incomplete investigation and incomplete disclosure Sophonow

  1. Jailhouse informants: reliance upon people who have motives to lie.
  2. Inadequate forensic examination, and overestimating the strength of scientific conclusions.
  3. Pushing witnesses to believe in the suspect’s guilt.
  4. Inadequate consideration of time lines.
  5. Failure to disclose relevant information.

Lessons to be drawn include:

  1. Investigate alternate theories of guilt.
  2. Investigate the reliability of your sources of information, especially jailhouse informants.
  3. Get objective forensic evidence.
  4. Take care to understand the strength of scientific conclusions.
  5. Prevent contamination of exhibits.
  6. Prevent contamination of witnesses.
  7. Figure out the time-lines.
  8. Beware of eyewitness identification.

Interesting Research

Early research into wrongful conviction tended to rely on case studies.  Recent American research has begun to apply statistics to the many proven wrongful convictions.  This study confirms some of the case studies, and identifies what differences arose between the wrongfully convicted and similar situations in which the defendant was acquitted or discharged.  It found the following factors between similar cases:

Likelihood of Wrongful Conviction
Higher Lower
State still imposes death penalty
Prior convictions
Honest mis-identification of accused
Failure to test forensically, or consider forensic results
Incomplete disclosure - particularly of "minor" evidence
Weak case
Underfunded defence
Inexperienced counsel
Full disclosure
Use of Forensics
Deliberate mis-identification of accused (proven liars)
Well-funded District Attorney's office
Well-funded defence counsel

Sophonow Inquiry

An archive of this report may be found here in the WayBack machine

Sophonow Facts

In Winnipeg, somebody murdered the girl who worked in the donut shop by strangling her with twine. Several eyewitnesses saw him walking away. One confronted him on a bridge. He threw the twine over the bridge. Some preliminary results suggested that the twine came from British Columbia.

Sophonow came from Vancouver just before the murder. He resembled the descriptions given by the eyewitnesses. At first, the eyewitnesses gave tentative identifications of Sophonow, in live and photo lineups, but those identifications firmed up with repeated interviews.

A couple of officers familiar with the case interviewed Sophonow in Vancouver. Accidentally or deliberately, they conveyed to him how the door of the Donut shop was locked. When this same information emanated from Sophonow later in the investigation, it made him look guilty.

At the preliminary hearing, defence offered an alibi. It seemed contrived.

The Taking of Statements

Eyewitness Identification

Tunnel Vision - Terry Arnold

None of this was pursued. Therefore:

-> structure your investigational team to catch such suspects
-> proof of the positive is good; but disproof of the negative is harder. In a serious case, that disproof is necessary.

Forensic Testing – Objective Evidence

Time Lines

The Investigation Isn’t Over (ever)

Jailhouse Informants

Mr Justice Cory didn't mince words:

Jailhouse informants comprise the most deceitful and deceptive group of witnesses known to frequent the courts. The more notorious the case, the greater the number of prospective informants. They rush to testify like vultures to rotting flesh or sharks to blood. They are smooth and convincing liars. Whether they seek favours from the authorities, attention or notoriety they are in every instance completely unreliable. It will be seen how frequently they have been a major factor in the conviction of innocent people and how much they tend to corrupt the administration of justice. Usually, their presence as witnesses signals the end of any hope of providing a fair trial.

They must be recognized as a very great danger to our trial system. Steps must be taken to rid the courts of this cancerous corruption of the administration of justice. Perhaps, the greatest danger flows from their ability to testify falsely in a remarkably convincing manner. In this case, it will be seen that an experienced detective thought that Mr. Martin,a very frequent jailhouse informant with a conviction for perjury, was a credible witness. He lied in this case and he has testified in at least nine other cases, undoubtedly with the same degree of mendacity. Jailhouse informants are a festering sore. They constitute a malignant infection that renders a fair trial impossible. They should, as far as it is possible, be excised and removed from our trial process.

Jailhouse informants are a uniquely evil group. Justice Kaufman in the Morin Inquiry dealt extensively with jailhouse informants and the harm that they occasion. His thoughtful and helpful recommendations are carefully set out in his report. I will adopt them but go still further in my recommendations on this subject.”

Full Disclosure to Crown and Defence

* Crown needs to know where the holes are in advance (eg  OJ case).

Morin Inquiry


Morin Facts

The school bus dropped Christine Jessop aged 9 off near her home before her parents got home from work. When they got home she wasn’t there. 3 months later, her body was found 56 km/h away. She had been stabbed and there was semen in her underpants.

Hair and fibre found near her could have come from Morin. Hair and fibre found in his car could have come from her.

Jailhouse informants claimed Morin confessed.  They were placed together in the prison, and had ample time to compare notes.

Forensic Evidence

Mere screenings don’t pack much punch in a criminal courtroom.We need proper forensic examination and conclusion.

Police failures

The Commissioner found certain failings in the investigation by York Regional Police. These included:

  1. The failure to preserve evidence at the Jessop residence.
  2. The failure to ‘dust’ the residence (and particularly Christine Jessop’s room) for fingerprints either to preserve her fingerprints or determine if foreign fingerprints were present. The house was not fingerprinted even in the ensuing weeks following Christine’s disappearance.
  3. The failure to conduct an in-depth canvassing of the Queensville homes to meticulously document and clarify the memories of prospective witnesses at the earliest opportunity.
  4. An inadequate system for keeping track of which officers were assigned to do what and whether they had completed those tasks.
  5. An inadequate system to ensure that all reports were read and processed in a timely way. The potential for leads and follow-ups to slip through the cracks was present and, unfortunately, this did occur.
  6. The indexing system for reports was archaic. It did not permit an officer to search for all apparent sightings of Christine Jessop or vehicle sightings, without having the actual name of the author of each report or the name of the claimant for the sighting.
  7. There was a failure to systematically prioritize and follow up on ‘hot leads.’ In one case (the sighting of a man who appeared to keep a child forcibly in a car) the follow-up came 12 days after the information was passed on to the police.

Durham Regional Police also made significant errors:

  1. They performed an unstructured search of the site of the body, at night with a snowstorm coming on. They should have covered it with apart, and waited for the next day. Officers were smoking and discarding cigarettes at the scene; others were discovering cigarettes and maybe a lighter possibly attributable to the murderer.
  2. One of the officers kept two different notebooks with differing entries for the same events.
  3. Other officers’ evidence about the cigarettes varied as time passed.
  4. Continuity – the memory of the officer in charge of the exhibits improved over time, which suggested poor initial handling, and possible cross-contamination of exhibits.
  5. Tunnel vision – Morin was an odd guy, who had an odd way of expressing himself. Investigating officers attributed more to his remarks than in retrospect they should have. “Otherwise, I’m innocent.” “All little girls are sweet and innocent but grow up to be corrupt.” “[The body] was found across the Ravenshoe Road. Compare this with Dix, in which the investigators dismissed an alternate suspect as “just an odd guy”.
  6. Tape recorder failed for 2nd half of Morin’s interview.
  7. Time lines didn’t add up – victim’s parents’ time lines didn’t add up. This was brought forcefully to their attention, and they changed their evidence. This was done in a joint interview, with no notes taken.
  8. Interviews of sensitive witnesses done without recording or notes. Their evidence “developed” over time.
  9. Not until after the first trial did one Cst Robertson report that his dog, not a formally trained police dog, had detected the scent of the victim’s sweater in Morin’s car. He had no notes made at the time of this incident.
  10. Post-charge behaviour overblown – Morin behaved unconcerned about the victim when a police officer questioned another witness. Morin didn’t participate in the search for the child. He didn’t want to talk about it with a neighbor. He didn’t sound as though he cared. A neighbor spoke with Morin about the disappearance of the girl. Morin was holding hedge clippers to tightly that his knuckles went white. On the funeral night, someone claimed to hear a male voice say “Help me, help me, Oh God help me.” Someone recognized it to be the accused’s. (He didn’t attend the funeral.)

Jailhouse Informers

Need confirmation of their evidence with details. Need to know about informer’s character, whether he requested any benefits, and whether anyone made any promises,whether he has given reliable information in the past, whether he previously traded information for benefits, whether he previously testified, and whether he lied in that testimony, whether he made a written record, was it contemporaneous, circumstances of the statement of the informer, manner of the taking of the informer’s statement (any possible tainting or leading).

Consider checking the alibi of the jailhouse informer: maybe the informer knew these details because he waste offender.

Ongoing Post-conviction Disclosure

Kaufman recommended post-conviction disclosure of any information tending to exhonorate the accused. Eg DNA analysis.

Henry Lawsuit

Crown failed to disclose evidence which would have assisted Mr Henry v. British Columbia (Attorney General), 2015 SCC 24 from a criminal prosecution based on weak evidence. He spent 27 years in jail for sexual offences he didn't commit. After he got the conviction overturned, he sued the Crown, and won.  The court found the that state could be liable for non-disclosure when their conduct was a "marked departure from the standard of a reasonably prudent prosecutor."

Dix Lawsuit

A summary of the Judgment in Dix v. Canada (A.G.), 2002 ABQB 580



Dix had an affair, but his mistress told him it was over. Dix knew victim X. from work, and they socialized together. The night before the murders, Dix, his mistress,and victim X went to a bar together and socialized. She showed some interest in X. That night Dix tried to hire a hooker.

The next morning, October 1, 1994, X and another guy Y went to work. They were last seen alive at 9:20. One was alive at 10:00am when he spoke with family. But after 10:30, nobody at the worksite answered the phone. Their bodies were found at 9:00pm by family and friends.

Electrocution theory

The first guy to find them was Mr Adams.  He called it an electrocution.  He pointed out marks on a screwdriver consistent with this theory. They called appropriate emergency personnel, who attended.

The first police officer on the scene had 5 months experience. What the ambulance attendants told her about the injuries did not change the “electrocution” theory.

Nobody could find any obvious electrical switches in the on position. Because of the concerns about electrocution, someone turned the power off in the building. Police at the scene removed the bodies using flashlights for illumination.

The senior officer on the scene concluded in 2 minutes of investigation that the death was caused by an industrial accident.

Policy required the watch commander to attend the scene, but he didn’t go. (He was later reprimanded.)

Adams returned to the scene, asking to examine it to satisfy himself how they died. He pointed out marks on a screwdriver that X had been holding, and said they came from electricity. He left with the deceased’s truck.

The next day the medical examiner discovered 3 bullet holes each in the heads of the bodies.

RCMP appointed a lead investigator and a file coordinator. They used an organized system of tips. Unfortunately, they couldn’t get much forensic evidence at the scene because of the false start to the investigation.

RCMP interviewed the employees of the business and determined that there was no criminal gangs involved that would lead to this execution style killing. Most alibis were checked out.They relied heavily on polygraph.

Primary suspects

They narrowed the list of suspects to three: Dix, Adams, and another co-worker, Spencer. On the advice of counsel, Spencer refused the first invitation to the polygraph. He later consented, but logistics and a change of focus of the investigation prevented it from being done.

Adams had a romantic interest in the wife of one of the deceased.  Adams’ first polygraph was inconclusive. He failed the second one, and walked out of the interview which followed. Interviews of people who knew him revealed that he knew about firearms, he would sometimes go into“trance-like” states, and he may have known that the victims were shot before the media announcement.

However, the investigators ruled him out because:

(Two years later, a more evidence arrived showing a significant romantic interest between Adams and the X’s wife.)

Reasons to suspect Dix

Police theorized that Dix was jealous of his mistress and the interest she showed in X.

An FBI profiler was briefed about what the police knew.He thought that the murderer would have been young, male, white,physically large, and self-confident. This matched Dix.

When interviewed, Dix lied about his affair and his effort to get a prostitute. In his second interview, Dix told the police that he thought the deceased would go to work the next day (Saturday morning) alone. From an ambiguous remark in that interview, investigators inferred that Dix went to the work site on the morning of the murder. They Chartered and warned him, and arranged surveillance.

After the interview an undercover operator heard one side of a telephone conversation he had with someone. Dix said that he had been to the work site. Police assumed that he meant on the morning of the murder.

Two weeks after the murder Dix’s wife gave a statement to the police which described Dix as spending a normal day involved with work. As an alibi, it was incomplete, but did suggest it would have been difficult to commit the murders.

Police interviewed the mistress at the same time. She disagreed with the jealousy theory.

Investigating Dix

After those interviews, Dix and his mistress spent the weekend together at a cabin at a lake. They shared the cabin with another woman and a child. Dix was alone for only a short time.Potential murder weapons were later found there.

Police conducted a Mr Big operation on Dix, which included a staged shooting. Dix told the undercover operators that he would participate in a murder scenario, he bragged of his proficiency with a .22. Dix denied responsibility in the murders, but made helpful suggestions on how to cover up the shooting incident. The undercover operators gave Dix the impression that they would kill him if he reported their involvement in the shooting. Dix told his wife what he was doing in the Mr Big operation.  She later said that his motive was merely the money.

After the staged shooting, Dix never contacted the undercover operators again. During the operation, Dix appeared to consider the possibility that he was being set up.

Police prepared wiretap applications which omitted some exculpatory details, and described the mistress as a suspect even though she had an alibi. Wiretap produced no evidence.

After the murder, Dix started collecting newspaper information about murders, and on this murder particularly. His wife thought that odd, but he trial judge didn’t.

Dix’ first polygraph was inconclusive.Police told him he failed. He told his wife he failed. Police knew she was considering leaving him, and did not correct the misunderstanding with her.  She did leave him.

Separate officers from the investigation team conducted a file review. They identified some weaknesses, but generally agreed with the approach of the investigation. They became the new investigation team. They re-interviewed the mistress about her conversations with Dix on the morning of the murder. At this time, she was angry with Dix, and cooperative with police, but did not provide any inculpatory information.

Another police file review concluded that Dix had an alibi for the time of the shooting, but that Dix should not be entirely removed from suspicion, particularly if other inculpatory evidence came to light.

Buried behind the cabin at the lake,police found four revolvers. One of them was “probably” the murder weapon, but not certainly. The scientific report did not express the level of scientific confidence it was a match with the bullets recovered from the victims. Police told the prosecutor it was “definitely” the murder weapon, and the prosecutor said they had a case.

Police arrested and interrogated Dix. He made little comment. “Somehow” the media found out and scrummed him,getting some photos which made him look bad.

Police kept him at the detachment rather than sending him on to the remand centre. They interrogated him for 11 hours without food. They offered, but he declined, drink. They lied to him about some of the evidence. He asserted his right to silence 200 times. They employed a “good cop bad cop” technique, involving the bad cop going nose to nose screaming “What are you? A fucking parrot?” Crown had warned the officers in advance that this technique would not produce admissible evidence.

Police took Dix to the scene of he murders, against his consent. (Crown had told them it would be OK if they got consent.) He still denied the murders.

The handguns turned out to be related to a set of handguns stolen in Vancouver, and associated to a Vancouver criminal gang. Investigation of this gang revealed an Alberta associate who had sold the very property where the guns were found a few years earlier. Police could find no link between Dix and this criminal gang.

Prison informants

While Dix remained in remand, police approached another inmate and asked him to keep his eyes and ears open. Simpson told the police that Dix confessed to him, and he gave a version of events which contained little hold-back detail, and got many points wrong. He received some special consideration.

Another inmate volunteered that Dix confessed to him. He provided information known to Dix, but did not provide any hold-back type information.

A third inmate volunteered with his lawyer to provide information. Police took him up on it. His contact occurred during the time Dix received his Crown disclosure. This inmate did not volunteer until after the preliminary hearing. Just before the trial, he wrote a letter to defence counsel retracting his evidence, and boasting that he made this offer in order to extract money from the police. He explained his letter to police as an effort to avoid being labelled a rat. (He played both sides of the fence.)

Crown Counsel involvements

Because one of the informers said that Dix wanted the mistress killed, police prepared a letter purportedly from another inmate offering to do the job. Dix gave this letter to defence counsel. Crown misrepresented the letter in court at the bail hearing.

One of the informants had been represented by defence counsel. Crown failed to identify the conflict in interest ingood time. Crown would not reveal the identity of the informant until shortly before trial, putting defence counsel in an impossible position.

Police told lies to various witnesses which made Dix look guilty or bad. Often this served no purpose.

At the preliminary hearing, one prosecutor withdrew because he did not believe the prison informers, and the rest of thecase looked too weak to him.

At the trial, the original prosecutor had to withdraw because of the misleading remarks he made in the bail hearing.Another prosecutor reviewed the file, and concluded that there wasn’t enough evidence to proceed further.


The judge at the civil suit rejected statement analysis as a useful technique, he pointed out weaknesses of profiling.The police and prosecutors confused “consistent with” scientific opinions with “suggestive of” or “conclusively shows” (Blood spatter and weapons testing)

The police and prosecutors chose to believe inculpatory evidence and reject exculpatory evidence.

Time line evidence shows it as unlikely that Dix had the opportunity to commit the offences.

When inculpatory evidence turned up withregard to one suspect, police abandoned the other suspect with undignifiedhaste.

Nobody investigated whether other suspects had any connection to the location where the suspicious weapons were found.When police found the one bunch of weapons, they didn’t search any further.

The trip to the crime scene in the middle of the night was a kidnapping.

Police and prosecutors have a duty to investigate the reliability of the information they receive. With jailhouse rats this is particularly important, so that the court can learn of the real relationship between the rat and his testimony.

Police shouldn’t bad-mouth the suspect,nor tell lies to witnesses about the evidence they have without goodreason. In this case they helped break up his marriage by misleading his wife about the strength of the evidence against him.

Don’t embark upon painful investigative techniques without a reason. (Interrogation was not expected to produce admissible results.)

Tactical late- or non- disclosure offends the Charter, and makes trials unfair.

General damages:
RCMP Punitive damages:
Prosecutor Punitivedamages:
False imprisonment:

Other Inquiries


Donald Marshall was a 17-year-old native kid in Nova Scotia was in the wrong place (a park) at the wrong time (just before midnight).  He and a black kid tried asking Roy Ebsary for money.  Ebsary stabbed and killed Seale.  Racial prejudices led police to believe Marshall was the killer.  The officers suggested to impressionable witnesses their theory of what occurred, and the witnesses responded with evidence that inculpated Marshall.  The officers failed to investigate the scene, and ignored and discarded evidence which contradicted their theory.

Marshall spent years in jail for a murder he didn't commit.  The killer went on to commit other crimes.


In 2007, a 5-judge bench of the Ontario Court of Appeal found that the 1959 conviction of Steven Truscott, 2007 ONCA 575 was unreliable, and overturned it.  The main problems were unreliable science, inadequate disclosure, poor interviewing techniques, and an unlikely theory of the murder.  In a somewhat unusual decision, the court acquitted him even though they thought that there was evidence on which a jury could reasonably convict.

Estimating time of death from the condition of a corpse was an inexact science.  In this case, the extent of digestion of stomach contents was a key piece of evidence; unfortunately the inferences drawn were contraversial.

Dalton, Parsons & Druken

In October 2006, a retired judge of the Supreme Court of Canada inquired into three miscarriages of justice in Newfoundland.  The first part deals with the 8 years it took to conduct Mr Dalton's conviction appeal.  This has little relevance for police.  However, Parsons and Druken were convicted of murders they did not commit.  The report analyzes why.  The commission report is available here.


Randy Druken came from a family of criminals.  He had a turbulent relationship with the victim.  He visited the victim shortly before she was murdered.  Her daughter heard something during the night.  It seems likely that his wicked brother Paul committed the offence, leaving his DNA on a cigarette butt at the scene.

Some lessons include:


David Milgaard spent 23 years in jail for a rape-murder he didn't commit.  Larry Fisher did it.  Before killing this victim, Fisher had sexually assaulted 3 others.  Fisher went on to rape or attempt to rape 4 more women.  He attempted to murder one of them.  DNA eventually cleared Milgaard, but not before a media circus smeared the reputations of many people.

Milgaard's supporters and the press alleged misconduct by police, prosecutors and defence counsel.  The inquiry commissioner disagreed with most of these allegations, but identified some police procedures which could be improved.

These findings are not new: previous inquiries and judicial decisions already identified their importance.  Police who continue to apply discredited investigative procedures risk censure.

Some other recommendations include:

The commissioner was plainly impressed by the lack of tunnel vision in this case.  The investigators kept open minds, even as the evidence mounted against Milgaard. Their efforts to investigate all possibilities served them well during this inquiry.


Romeo Phillion, 2009 ONCA 202 was a criminal.  So was his twin brother Donald.  In 1967, someone looking like them murdered a fireman named Leopold Roy.  Mrs Roy, who was present for the murder, saw a photo lineup and a live lineup.  In the photo lineup she thought Donald did the deed.  In the live lineup, she thought Romeo might be the one.  Because she wasn't sure, police released Romeo, and kept the file open.  Donald went to jail for an unrelated murder.

Romeo told police that Donald would confess.  Donald didn't.  Romeo told the police several different alibis.  The police checked them out, but some of their paperwork was lost.  3 years after the murder, Romeo told his teenage lover that he killed the fireman.  The lover told the police, and Romeo confessed to the police.  Within a few hours, he retracted the confession.

At trial defence presented evidence of psychologists that Romeo had a personality disorder that made him want notoriety, and he liked to cause trouble for police.  Everything he said to people in authority was suspect.  It didn't work.  The jury convicted him.

Romeo complained thereafter that he was innocent.  30 years later, someone found documents which suggested that one of Romeo's alibis might be true.  Nobody told defence at the trial.  40 years later, the Ontario Court of Appeal ordered a new trial, on the basis that a miscarriage of justice may have occurred.  Experts at that trial also said that he was the kind of guy that might confess to a crime he didn't commit.

The details he gave in his confession were almost all readily available in newspaper accounts of the murder at the time.


Dr Smith

Dr Smith was a pathologist in Ontario whose testimony seemed unassailable.  Unfortunately, his work was unreliable, and several people innocent of any offence were found guilty based on his evidence. Brant, 2011 ONCA 362 C.F., 2010 ONCA 691; C.M., 2010 ONCA 690  Sherret-Robinson, 2009 ONCA 886; Mullins-Johnson, 2007 ONCA 720 - misdiagnosis of the murders of babies.  The lesson for all of us is to investigate and prosecute with healthy skepticism.

Scottish Fingerprint Inquiry 2011

What happened when Scottish fingerprint examiners mis-identified a fingerprint?  

In Kilmarnock, Scotland, somebody murdered old Ms Ross.  Detective Constable McKie, attended the victim's house, but she claimed she never entered.

Around the same time as the murder, Mr Asbury left a suicide note, and disappeared for long enough to worry his relatives.  Mr Asbury's fingerprint turned up inside the victim's house.

McKie found a small tin box containing money at Asbury's house.  It bore DC McKie's fingerprint.  It also seemed to bear Ms Ross's fingerprint suggesting that Asbury stole it from the old lady, and killed her in the process.

At Asbury's trial, defence suggested that DC McKie planted evidence: she moved the box from the victim's house to Mr Asbury's house.  DC McKie denied ever having entered the victim's residence.

After Asbury's conviction, the Crown charged DC McKie with perjury.  If she never entered Ms Ross's house, how could her fingerprint be on the door?

An Inquiry answered: it wasn't her fingerprint.  Nor did the box necessarily bear the victim's fingerprint.  The experts were mistaken.  McKie, and, eventually, Asbury, were acquitted.

The commissioner found that fingerprint experts were encouraged to give absolute opinions about matches ("I am 100% certain that this print came from this person").  He explained that fingerprint comparison doesn't necessarily result in "match" or "different".  Blurry marks can lead experts to disagree.  Biases and peer pressure can influence opinions; for example if a senior expert says "match", how can a junior expert voice dissent?  Receiving information from the investigation can cause an uncertain expert to become more certain.

Defence counsel will use this report to challenge fingerprint comparison evidence.  Fingerprint examiners who have not read this report should take the time to go through it.

For the rest of us, the Commissioner observed:

"All forensic evidence should be approached with an open and questioning mind both by those who practise the discipline and by the legal community. It should not be assumed that any forensic evidence is irrefutable or infallible, if only because of the risk of human error. Nor should it be assumed that any form of forensic evidence is necessarily routine. Fingerprint evidence is no different. There is no evidence before the Inquiry to suggest that fingerprint evidence as a class is inherently [un]reliable. On the other hand there is no basis for a claim to infallibility. It is opinion evidence and where appropriate, it should be subject to robust scrutiny and challenge." (page 600)

Motherisk 2015

Ms Hayman, 2021 ONCA 242 used cocaine heavily - by reason of a tragic life, she suffered a terrible addiction. A hospital screen detected cocaine in her young son's urine. Quite reasonably, someone investigated, and got some of his hair, and sent it to a well-known lab for analysis. The expert said that the quantity of cocaine they found in his hair established that he used cocaine regularly for months. She stoutly denied ever giving him cocaine, nor using it or leaving it in places where he might get any. A judge convicted her and sent her to jail. She lost her son to social workers. When she bore a daughter, social workers took that daughter away from her too.

The lab used scientists who didn't understand proper forensic technique. Their analysis, and testimony, was not good science.  Harmful Impacts: The Reliance on Hair Testing in Child Protection; 56 families were torn apart by their shoddy work. Report of the Motherisk Hair Analysis - Independent Review

Civil Suits

Investigate badly, and you can get sued.

Hill - negligent investigation

Police placed a photo of Mr Hill (2007 SCC 41), an aboriginal robbery suspect, in a photo-lineup with 11 caucasian faces. Police interviewed two eyewitnesses together, with a photo of Mr Hill on the desk.  Some witnesses tentatively identified him.  Police collected other weak identification evidence, and charged him. After his acquittal, Hill sued the police for "negligent investigation".

The court found that you can get sued for negligent investigation.  In this case, the officers didn't stray far from the standards of investigation that applied back then. Beware.  In my view those standards have changed.

Beware of short-cuts in investigations. Always do the lineup correctly the first time.  Don't interview witnesses together.  Consider identification evidence carefully.  Investigate alternative suspects.

Fail to investigate, and you can also get sued.

Failure to warn - Jane Doe

In Jane Doe v. Metropolitan Toronto (Municipality) Commissioners of Police (1998), 160 D.L.R. (4th) 697 (Ont. Ct. (Gen. Div.)), from December 1985 to August 1986, a series of sexual assaults took place in Toronto.  The sexual assaults shared certain characteristics:  each took place in the same downtown Toronto neighbourhood; all the female victims lived in second or third floor apartments; each apartment contained an exterior balcony; and entry to the women’s apartments had been effected via the balconies.  After the fourth incident, the Metropolitan Toronto Police Force (“MTPF”) had grounds to believe that a single individual was responsible for the sexual assaults.  However, while anticipating that additional assaults were likely to occur, the MTPF deliberately refrained from informing potential victims of the specific risk to them on the grounds that doing so would cause the offender to flee.  "Jane Doe" was his next victim.  She sued the police for failing to warn her.  The trial judge, found that “the women were being used — without their knowledge or consent — as ‘bait’ to attract a predator whose specific identity then was unknown to the police, but whose general and characteristic identity most certainly was”.  The police were held liable for failing to warn.


Ivan Henry 2016 BCSC 1038 recently received an award of $8M. He spent 27 years in jail for rapes he probably didn't commit. This decision explores what went wrong.

In 1980-1982, police investigated around 20 similar sexual assaults. Mr Henry's ex-wife told police that she suspected Mr Henry. Police assembled a live line-up to which a group of complainants attended.  One of them gave a qualified identification of Mr Henry.

Mr Henry, who suffered mental disorder, fired the lawyers he retained. He represented himself - poorly - at trial, and the jury convicted him.

Years later, prosecutors noticed a striking similarity between his charges and and other offences committed around the same time by a guy named McRae. A review led to the conclusion that the evidence suggested Mr Henry didn't commit the crimes for which he was convicted.

Although there were problems with the police investigations too, the bulk of the blame for this wrongful conviction fell on the prosecutor, who failed to disclose investigative materials to defence.

What relevance do prosecutorial mistakes made 35 years ago have to current police practice?  More than you might first expect.

Cross-referencing similar files might have discovered Mr McRae earlier, and taken suspicion away from Mr Henry. Collecting, organizing and assessing relevant information was a problem then. It's still a problem today.

The methods of conducting lineups described in the decision may seem antique to you, but complacency about our methods today will lead you astray. Even today, some officers still tell eyewitnesses after the photo-lineup whether they got the identification "right".  Even today, we see photopacks containing images of the suspect which differ markedly from the other faces.

American Resources

In the United States, The Innocence Project has exhonorated hundreds of people convicted of serious crimes, and is investigating more.  It rescued many from death row.  Their website identifies similar problems to our own, and makes for interesting reading.

An interesting paper analyses 325 exhonorations.

Two American law schools compiled a registry of all the exhonorations that have occurred in that country, which provides a resource for research.

Innocence Canada

A similar project in Canada examines cases of wrongful conviction.  Their list of cases illustrates many of the same issues.

Heads of Prosecutions

The top prosecutors across Canada reviewed all the sources of information available, and reported on Eyewitness identification, tunnel vision, false confessions, forensic evidence and in-custody informers.  They produced two reports:

2005 Report on Miscarriages of Justice

2011 The Path to Justice: Preventing Wrongful Convictions

Other cases of Investigational Problems

In the murder case of R. v. Wood, 2006 ABCA 343 (Alta C.A.) the police forgot to mention that hold-back information was disclosed to another police force, and that the eyewitness had identified another person as the killer.  The murder occurred in 1979.  He was convicted, appealed successfully and was convicted again.  Now, more than 25 years later, he gets another trial.  He almost got a stay of proceedings.  Please remember to disclose everything in your file.

Mr Persaud (2007 ONSC) taught difficult children. One of them complained of an assault in the classroom. The school suspended Mr Persaud, and wrote a letter to parents about it after which some kids made more complaints. Police took statements from the kids that complained, but not the other children who were present at the time. The absence of corroborating evidence from other people present at the time of the alleged assaults caused the judge to doubt the allegations. It also appeared that the letter to the parents could have stirred up "me too" complaints. Defence also complained of the risk that the children colluded. When investigating complaints from a group, you should canvass with each witness:

To avoid accusations of tunnel vision, you should always canvass all reasonably available sources of information, not just the complainants.

Full disclosure matters, even when it seems obvious that the right guy is charged.  Mr Dhillon, 2014 BCCA 480 owned the place where two men raped a woman.  Police found him naked and intoxicated in the suite, only hours after the rape.  The woman said that one of the rapists fathered her child. The judge convicted Mr Dhillon of the rape. After he served his sentence, he was deported. Although police disclosed to the Crown that some DNA turned from the forensic examination of the victim, some DNA results never reached the prosecutor.After the conviction, DNA sampling of the accused went to the DNA databank.  Nobody noticed it failed match either of the male profiles from the victim.  Only when one of those two male profiles matched some other guy, years later, did police re-examine the case.  Mr Dhillon's DNA didn't match either profile, nor the baby's DNA. He could be innocent. The court granted him a new trial, and then stayed those proceedings.  He had already served his sentence, and the likelihood of conviction was low.


    Academic papers suggest that the rate of wrongful conviction may be as high as 4%, Another paper suggests it's much lower: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3276185