Public Release:  Child witnesses -- how to improve their performance

Study examines the ability of both children and young adults to identify faces with different racial features

University of Leicester

A study at the University of Leicester into how to improve child and young adult witnesses' evidence has looked at several issues that affect witnesses' accuracy.

The investigation, carried out in Estonia by postgraduate psychologist Kristjan Kask, looked at issues such as police officers' interviewing methods with children; young adults' skill in describing people; and the ability of both children and young adults to identify faces with different racial features.

Dr Kask looked at the real-life interview protocols conducted by Estonian police officers. He found that open-ended questions, such as 'Tell me more of what happened', resulted in more significant information that could help prove or disprove a culprit guilty. Closed questions, such as: 'Did he wear a brown jacket?' proved less effective.

However, obtaining this information presented problems of its own, as Dr Kask explained: "Conducting eyewitness research involving children is a very challenging task.

"It was difficult at times to get access to relevant organizations such as police forces. For example, when I collected the data from one department then only transcripts conducted by officers not working there any more were provided."

It is known that witnesses, especially young witnesses, can recognise people more easily than they can describe them. Dr Kask's study looked at whether young adults can use a 'standard` as an aid to help them provide more accurate information about a person they may only have seen once.

This involves comparing the memory about the 'target' with the interviewer, for instance: 'Compared to me how tall was he?' However, this was found to have no effect in improving witnesses' descriptions.

A great deal of forensic psychology research has examined the recognition of people of different ethnic groups using lineups in police settings. From this, psychologists have found that people recognise faces with their own ethnic characteristics more accurately than those of other ethnic groups.

What is less clear is how faces of different ethnic origin are recognized when they are presented together in large numbers. For instance border guards or immigration officers may see a large number of people during the day and have to decide whether some of them might be in the 'wanted` list.

Dr Kask showed young Estonian adults several 'target' faces of different ethnicities followed by a larger group of faces. They had to decide for each face whether it was one of the faces they had seen earlier.

He found that the young people taking part in his study could recognise the 'target' faces from different ethnicities equally well. However, they found it harder to reject faces that were of a different ethnic group from their own.

Dr Kask concluded: "When a multi-perpetrator crime takes place, then identifying the real culprits can be problematic and will depend on different aspects of face recognition, such as own- or other-ethnicity targets; target presence or absence in the recognition set; and/or the location of lineup fillers and target(s) in the lineup."

###

Dr Kristjan Kask received his BSc and MSc in psychology from the University of Tartu, Estonia.

He is a member of the executive committee of the Nordic Network of Research on Psychology and Law (NNPL). He has presented his work in several international psychology and law conferences and has conducted several trainings in interviewing children for Estonian Police.

Currently he is a freelance adult trainer and is consulting the Ministry of Justice in Estonia regarding interviewing child and other vulnerable witnesses. His doctoral research was supported by Archimedes Foundation in Estonia.

Notes to Editors: For more information on this please contact Dr Kristjan Kask, email kristjan.kask@ut.ee

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.