A wrongful conviction rarely begins with one spectacular blunder. More often, it develops when an early police theory hardens, ambiguous material is treated as proof, and later information is assessed through the same narrow lens. The best wrongful conviction case studies are therefore not simply stories of people cleared after prison. They are working examples of where an investigation, prosecution or trial process failed to properly test the evidence.

For Australians interested in criminal justice, these cases should be read carefully rather than consumed as true-crime spectacle. Each turns on a different point of failure: flawed forensic opinion, unreliable disclosure practices, tunnel vision, misunderstood science, or the danger of treating a disputed account as a confession. They also show why a conviction is not the end of an investigative obligation when credible doubt remains.

What makes a wrongful conviction case study useful?

The most useful cases have an accessible record: trial evidence, appeal reasons, inquiry findings, expert material and a clear timeline. That record allows the public to separate what was alleged at trial from what was later established.

“Best” does not mean most dramatic. It means most instructive. A case may expose a clear factual error, while another may reveal a process that was incapable of fairly testing a contested issue. Not every overturned conviction proves who committed the crime. But an acquittal, a quashed conviction or a pardon can still reveal serious defects in how the case was built.

1. Lindy Chamberlain-Creighton and the danger of a fixed theory

The death of nine-week-old Azaria Chamberlain at Uluru in 1980 remains Australia’s most recognised miscarriage of justice. Lindy Chamberlain-Creighton said a dingo took her baby from a tent. The prosecution case instead advanced murder, relying heavily on interpretations of forensic material and the proposition that the dingo account was implausible.

Lindy was convicted of murder in 1982 and Michael Chamberlain was convicted as an accessory after the fact. The case later shifted after the discovery of Azaria’s missing matinee jacket in an area near the site where a tourist had fallen. It was material consistent with the account that Azaria had been taken by a dingo. Lindy was released and pardoned in 1987, and the convictions were quashed in 1988. A fourth coronial inquest in 2012 found that Azaria died as a result of being attacked and taken by a dingo.

The lesson is not that investigators should accept every initial account without scrutiny. They should not. The lesson is that an account cannot be rejected merely because it conflicts with prevailing assumptions. Once police and experts treated the dingo explanation as effectively impossible, evidence was interpreted to support an alternative theory. Good investigation keeps competing hypotheses alive until the physical evidence properly eliminates them.

2. Andrew Mallard and the risk of manufactured knowledge

Andrew Mallard was convicted in 1995 of the 1994 murder of Perth jeweller Pamela Lawrence. His conviction was later overturned after serious concerns emerged about the way police interviews and alleged admissions had been handled, together with material that had not been disclosed to the defence.

A central issue was whether details attributed to Mallard were genuinely known only to the offender, or whether they had been introduced, shaped or tested during the investigation. This distinction matters. A confession is powerful evidence only when it is voluntary, accurately recorded and contains information that can be independently verified as originating from the suspect.

Mallard was acquitted in 2005 after the High Court found a substantial miscarriage of justice. The case demonstrates why interview records, original notes, unused material and disclosure schedules matter as much as the polished version of events presented in court. An investigator who believes a suspect is guilty may unintentionally feed information into an interview. A later reader then sees apparent insider knowledge where none exists.

3. Farah Jama and the limits of DNA evidence

DNA is often described as the gold standard of forensic evidence. The Farah Jama case is a necessary correction to that shorthand. Jama was convicted in Victoria in 2008 of rape after his DNA was detected on a vaginal swab taken from the complainant. He maintained his innocence.

The conviction was quashed in 2009. The inquiry that followed identified major problems, including the handling and interpretation of the DNA evidence. Crucially, the sample had been collected before the alleged assault. The presence of DNA, without a reliable explanation of when and how it was deposited, could not establish that Jama committed the offence alleged.

This is one of the best wrongful conviction case studies because it exposes a persistent error in public discussion: DNA does not interpret itself. A profile may identify biological material, but the surrounding facts determine its meaning. Was the sample correctly labelled? Was it taken at the relevant time? Is transfer or contamination possible? Does it fit the chronology and the other evidence? Those questions should be answered before DNA is treated as proof of guilt.

4. Henry Keogh and untested expert opinion

Henry Keogh was convicted in South Australia in 1995 of murdering his fiancée, Anna-Jane Habel, who was found dead in a bath in 1991. The prosecution case relied substantially on pathology evidence concerning injuries to her body. For years, concerns were raised about whether the pathology conclusion was sound and whether relevant material had been properly disclosed.

In 2014, the South Australian Court of Criminal Appeal quashed Keogh’s conviction. The court found that the non-disclosure of material had denied him a fair trial. A retrial was not pursued, and Keogh was released.

The broader lesson is that expert evidence must be tested as evidence, not accepted because it comes from an expert. Pathology, blood-pattern interpretation, injury assessment and other forensic disciplines involve judgement. Defence access to working notes, photographs, differing opinions and underlying data is essential. Courts need to know not only an expert’s conclusion but also the limits, assumptions and uncertainty behind it.

5. Kathleen Folbigg and science that moved beyond the trial

Kathleen Folbigg was convicted in 2003 over the deaths of her four children between 1989 and 1999. The prosecution relied on circumstantial evidence, including diary entries interpreted as admissions, and on the argument that multiple unexplained infant deaths within one family pointed to deliberate harm.

A later inquiry considered developments in genetic science and evidence bearing on possible natural causes of death. In 2023, Folbigg was pardoned following the inquiry, and later that year the New South Wales Court of Criminal Appeal quashed her convictions and entered acquittals.

This case raises a difficult but important point. Scientific knowledge changes. A trial can be conducted according to the accepted understanding of its time and still produce an unsafe result when later evidence substantially alters the picture. That is why post-conviction review must remain open to credible new science, particularly where the original case depended on disputed medical interpretation rather than direct proof.

6. David Eastman and the importance of reviewing the whole investigation

David Eastman was convicted in 1995 of murdering Australian Federal Police Assistant Commissioner Colin Winchester in Canberra in 1989. Following a board of inquiry, Eastman’s conviction was quashed in 2014. At a retrial in 2018, he was acquitted.

The Eastman matter is a reminder that a review cannot be limited to one disputed item of evidence. Major cases are built from decisions made across months or years: witness management, forensic testing, intelligence assessment, disclosure, suspect selection and the treatment of alternative lines of inquiry. Weaknesses may become visible only when the entire investigation is reconstructed in sequence.

It is also a case requiring precision. An acquittal does not, by itself, answer every question about the underlying crime. It does establish that the prosecution could not prove guilt beyond reasonable doubt at the retrial. That distinction matters when discussing any contested case publicly.

What these cases require from investigators and the public

Across these matters, the recurring warning signs are familiar: a theory adopted too early, evidence interpreted without sufficient context, material not disclosed, expertise treated as infallible, and alternatives not pursued with equal rigour. None of these errors is corrected by good intentions. They are corrected by disciplined case review.

A proper review starts with the chronology. What was known at each point? What was tested, and what was assumed? Which exhibits or notes were available but not used? Did a witness account change after contact with investigators? Did an expert receive all relevant facts? Were reasonable alternative explanations genuinely examined?

For anyone studying a disputed conviction, the most productive habit is to resist choosing a side before reading the record. Build the timeline first. Identify the source of every major claim. Then ask what evidence would have been expected if the prosecution theory were true, and whether it was actually found. That is where a case moves beyond narrative and towards accountability.