Seven years of British court appeals shows that evidence presented at trial—and not necessarily new evidence presented afterward—drove several hundred overturned criminal convictions, according to a new study.

The misrepresentation of evidence at trial, bad directions from judges, and errors were foundation for the majority of the valid criminal appeal arguments in the U.K. between 2010 and 2016, argues a team from University College London’s Centre for the Forensic Sciences, in a new publication in the journal Science and Justice.

“What the results (…) mostly show is that wrongful convictions are not always just an issue of flawed science or bad lawyering, but rather, flawed communication and interpretation, an issue both sides should take responsibility for,” the UCL team writes.

The survey spanned all appeals in the courts of England and Wales from 2010 to 2016. The team searched the Westlaw UK database and identified 10,859 cases.

The subset involved “criminal evidence,” as defined in the coding for the court documents. The sample group turned up 996 cases.

A fifth of those convictions (201) were ultimately overturned, according to the survey. The crimes were sexual assault, assault, murder, robbery, fraud, and drug crimes.

Of those 996 cases, 218 rulings were argued unsafe, due to 235 pieces of “misleading evidence,” as defined in the transcripts of the court proceedings.

Eighty of those 201 overturned convictions resulted in no retrial.

But the majority only involved the same evidence that had been presented at trial—and just different arguments around those same exhibits.

“In the majority (76 percent) of cases (…) the ruling was overturned given the same set of information available in the trial,” they write. “This, again, suggests that perhaps not enough focus is directed on cases where there is not necessarily new evidence.”

Causes of the “unsafe” rulings were split. Some 36 percent were due to the presentation of evidence in court being generally misleading in the proceedings, including whether it was relevant, had probative value or validity, or was hearsay. Twenty-six percent were due to the directions of the trial judge with respect to those same issues, and with standards of proof. Prosecution and police errors, absence of evidence, and jury issues accounted for another 15 percent of the cases.

Only some 24 percent involved new relevant information presented at the appeal. 

“This study provides a foundation of detailed issues related to the interpretation of forensic evidence and highlights the possibility of identifying these issues from case transcripts, fueling the important debate around improving access to case documents post-conviction,” said Ruth Morgan, director of the UCL forensic program, in comments to Forensic Magazine.

“The findings from this research support previous findings that the identified cases are only the tip of the iceberg and can no longer be attributed to simple individual ‘bad apples’ in the system,” she added.