Software solves murder

12 March 2010

New software can prevent mistakes in detective work. At present, the police still make unnecessary mistakes when collecting and analysing evidence, which sometimes results in miscarriages of justice. Dutch researcher Susan van den Braak has developed a computer program that can help to prevent such blunders.

It is a well-known phenomenon: you’re watching a detective series, a real whodunit, and after five minutes you know for certain who the murderer is. The rest of the programme only confirms your hunch. Right up until the last scene: and it was still somebody else. Never mind. TV off, and into bed. But what if a real detective had made this mistake? Professionals also get it wrong sometimes, certainly in complex cases where lots of information has to be brought together. Tunnel vision, group thinking, the detective is not immune to these. Susan van den Braak has found the solution: use something non-human. With the help of her software, crime analysts can depict various scenarios. They can then clearly see where the holes in the case are.

During their work, detectives are supported by crime analysts. Their task is to critically follow the investigation. On the basis of all the evidence they attempt to produce plausible scenarios and to assess the likelihood of these.

An example: two men have an argument. One of them has a pistol. The other is shot. Result: one dead man. Various events could have preceded this. One of the men could have shot the other, but perhaps there was an unknown woman with a pistol in the vicinity and she was the one who fired the fatal shot. Evidence is needed to establish what happened. However, the cases that crime analysts have to deal with are complex, such as murder or sexual offences. Then it is easy to become fixed on a single scenario – one man shot the other – and to subsequently seek every shred of evidence that supports this.

Van den Braak’s software ensures that crime analysts can see at a glance what is missing in their scenarios. In addition to this, the software makes it easier to develop multiple scenarios and to critically assess these.

Example

Van den Braak's software seems to be child’s play. First you enter the events in the green boxes. For example, 'a man verbally abuses a woman', 'two men have an argument' and 'a man collapses'. Then you can fill in the evidence in blue boxes. For example, 'traces of gunpowder were found on the hands of man X' or 'witness A saw two men having an argument'. With a line you can then indicate whether the evidence supports an event. This line represents an argument. With the help of such an evidence-based argument, you can indicate whether an event is supported by a certain piece of evidence. With this approach, it is relatively simple to describe various scenarios and to quickly spot the missing link. Pieces of evidence can support or refute the events in a scenario. An event that is not supported by evidence is assigned a dotted border: something is not right there.

Van den Braak tested her software on potential users. She confirmed the utility of the software for assessing whether evidence is missing or whether the events properly explain what happened. At present, many crime analysts still only work with pen and paper or simple software. However, with the existing software they cannot represent any scenarios or indicate how these are related to the evidence in a case. A lot of software for analysts concentrates on ordering and structuring large quantities of evidence in the case. Van den Braak’s experiments have demonstrated that analysts can make better analyses with her software. And although it is still under development, the initial responses, including those from the police, are enthusiastic.

ToKeN

Susan van den Braak carried out her research within the NWO programme ToKeN (Access To Knowledge and its enhancement Netherlands). ToKeN is an interdisciplinary research programme focused on the interaction between a user and knowledge and information systems. Access refers to the extent to which individual citizens can acquire knowledge and information which is important for them. Knowledge disclosure refers to the methods that obtain knowledge from data and information for the user that would otherwise not be available. On 14 January 2010 during the annual symposium, a report containing an overview of all projects was published. This publication entitled 'ToKeN Highlights' can be ordered free of charge via www.nwo.nl/token.

About NWO

The Netherlands Organisation for Scientific Research (NWO) is the principal Dutch science funding body and its mission is to facilitate excellent scientific research in the Netherlands by means of national competition. Each year NWO spends more than 700 million euro on grants for top researchers, on innovative instruments and equipment, and on institutes where top research is performed. NWO funds the research of more than 5300 talented researchers at universities and institutes. Independent experts select proposals by means of a peer review system. NWO facilitates the transfer of knowledge to society.

..............................

For further information please contact:

• Susan van den Braak (Utrecht University)
• t.: +31 (0)6 45 19 25 34
• susan.vandenbraak@gmail.com
• The doctoral thesis was defended on 15 March 2010
• Supervisors: Prof. J.J.Ch. Meyer and Prof. H. Prakken