Drunk witnesses to crime are no less accurate when recalling details of what they have seen than sober people, a study has suggested.

The research, from Abertay University in Dundee and London South Bank University, found people who had consumed alcohol were less confident and recalled fewer details than those who had not, but were no more likely to include errors.

It also showed that people who discuss what they have witnessed with others can provide less reliable testimony.

The researchers said the results mean police, judges and jurors should be aware drunk people’s version of events may not be less accurate than others.

They stressed, however, that their research looked at people who had only mild to moderate alcohol intoxication and were asked to recall details with little delay.

Some people who took part in the study were sober and others had drunk vodka and orange juice.

Both groups were asked to recall details from videos showing a mock theft.

Participants were tested in pairs and each person watched a different version of the crime, although they were made to believe that they watched the same video.

Before being asked to recall what they had witnessed, half of the participants were allowed to discuss what they had seen.

The study found people who engaged in discussion incorporated errors into their testimony almost seven times more often than those who recalled alone.

A total of 87.7% of these participants gave at least one piece of incorrect information that they had not actually seen themselves, but had heard about from their co-witness.

The drunk participants were no more likely to make the errors than those who were sober.

It was also found that drunk participants were not less accurate than sober ones, but their accounts were less detailed.

Drinking alcohol also had a negative effect on a participants’ confidence, meaning they thought their memory was less accurate when in fact it was not.

Our work also shows that alcohol does impact recall completeness but not accuracy

Dr Julie Gawrylowicz of Abertay University, co-leader on the study, said: “Contrary to perceptions commonly held by the general public and many professionals working within the criminal justice system, our findings suggest that mild to moderate alcohol intoxication does not make individuals more susceptible to incorporating misleading information obtained from a co-witness.

“Our work also shows that alcohol does impact recall completeness but not accuracy, so mild to moderately intoxicated witnesses may be regarded as a reliable source of information, even if questioned in an intoxicated state.

“It is important to note though, that this study tested memory at low to moderate intoxication levels, with a minimal delay before recall, and no other influencing factors.

“Those involved in the criminal justice chain, including police, judges and jurors, should therefore be aware that although mildly to moderately intoxicated witnesses might report fewer details and might be less confident in their accounts, their version of events might not be necessarily less accurate.”

She added: “The results also show there is clear potential for all witnesses, sober and intoxicated alike, who discuss a crime with a co-witness, to report information that they did not see but have just heard about from the other witness.

“This indicates it may be important to minimise co-witness discussion where a crime has been witnessed in public settings, such as bars and restaurants.”

The study was published in the journal Psychopharmacology.