It seems like common sense that the more eyewitnesses who believe a person guilty the more likely they are to be guilty. But that is not always true.

For his thesis, Te Herenga Waka—Victoria University of Wellington PhD candidate Robbie Taylor investigated the extent to which study participants acting as mock jurors were influenced by both biased line-ups and the number of unanimous eyewitnesses when asked to determine guilt. His research could help reduce the number of wrongful convictions.

Robbie, who graduates from the University’s Te Kura Mātai Hinengaro—School of Psychology today, has specialised in jury decision-making and police line-ups.

He says that if a line-up is biased because one person sticks out, the more eyewitnesses who identify that person can paradoxically indicate the person is less likely to be guilty.

“For example, if a line-up is not biased and 10 witnesses identified the same person, that would indicate the person is more likely to be guilty than if five witnesses identified the same person. However, if the line-up is biased, the opposite is true: five witnesses identifying the same person would indicate the person is more likely to be guilty than if 10 witnesses identified the same person.

“To what extent do jurors understand this paradox?” asks Robbie.

He cites the classic case of the wrongful imprisonment for fraud of Adolf Beck in London around the turn of the twentieth century.

Police asked 22 women victims to view several line-ups of 15 men and all identified Beck. The jury found Beck guilty and the judge said “the evidence of identity has been absolutely overwhelming”.

In 1901, after five years in prison, Beck was paroled for good behaviour. But three years later five more women identified him from a line-up as the perpetrator of similar crimes. Ten days after Beck was convicted of these crimes, when he was back in prison, a similar crime took place and the real perpetrator was found at a pawnshop trying to sell stolen jewellery.

The man confessed to the earlier crimes and Beck was released and awarded £5,000 as compensation for his wrongful convictions.

Robbie says the case shows just how wrong witnesses can be.

In the United States, about 70 percent of cases where people were wrongfully convicted and later exonerated by DNA evidence were a result of mistaken witness identification.

In Robbie’s study, the participants acting as mock jurors read about a crime and saw a police line-up that was either biased or non-biased. An example of a biased line-up in the study involved a perpetrator with a scar, with witnesses viewing a line-up containing one person with a scar and seven people without.

“Participants in the study read that varying numbers of witnesses identified one person from the line-up of eight. Participants then rated the guilt of the identified person.

“In some experiments, we then told participants there had been an error in the case details and the number of witnesses who identified that person in the line-up was either greater or fewer than they had originally been told—these participants then re-rated the guilt of the identified person.

“Participants shown a biased line-up gave lower ratings of guilt for the eyewitnesses’ chosen offender, for example the man with the scar, compared with those shown a line-up that had no obvious bias.

“If study participants were warned a line-up was biased, they gave lower guilt ratings to the eyewitnesses’ choice compared with those participants who did not receive a warning of bias.”

Robbie’s research is the first systematic investigation into how the number of witnesses influences (in this case mock) jurors’ guilt ratings for biased line-ups.

If the findings translate to real cases, jurors might be influenced by simply breaking the eyewitness evidence into multiple pieces of evidence, rather than presenting it all at once.

“My research also shows some promising early findings for the impact of warnings on jurors.

“Past research has found warning jurors about biases can often be ineffective. But in my research we found warning participants about the biased line-up made them more sceptical about the witness identification evidence.

“Ultimately, this line of research might help us reduce the number of wrongful convictions due to mistaken eyewitnesses.”