I don't think behavioural science can currently assist in the identification of dangerous individuals before they kill - that is too much of a task.

However a more realistic role for behavioural science could involve helping organisations and public bodies prepare and assess their readiness in the event of a spree attack, as well as informing the media what NOT to do in the immediacy of a killing spree.

It is a tragic fact that spree killings are not rare at all around the world - even though in the UK there have only been four recorded occasions (Michael Ryan in Hungerford in 1987, Thomas Hamilton in Dunblane in 1996, Derrick Bird in Cumbria in 2010, and less well-known was Robert Sartin in Monkseaton in 1989).

Over 100 spree killing events have occurred in the USA since 1950 - with over 90% of those occurring since the 1980s.

Some social and political scientists saw such spree killings as a form of caucasian social protest against Reaganomics - with some making comparisons to slave-revolts from earlier years in USA history.

Other potential causes that often get highlighted in the aftermath of sprees include violent video games, violent pornography, or excessive use of the Internet.

However, many research studies and spree killer histories have shown such obvious influences are not always present.

We must be clear here; spree killers are not out of control or crazed individuals - in conducting their sprees they often show meticulous levels of planning, a calm ordered approach, and are frequently described by surviving witnesses as being focused.

The popular idea of a crazed gunman shooting randomly is a myth; victim selection is not random, and statements from many spree survivors describe how something about them, their looks, their history with the killer, or even how they pleaded for their life was able to make the killer's attention shift from them.

Secondly, spree killings are rarely purely political - and there will often be the elements of high levels of personal frustration and feelings of being thwarted that individuals then project onto organisations, authority, and the establishment.

This seems to have been the case with the recent Oslo tragedy, but certainly was also true of the Dunblane and Cumbria massacres, as well as the spree-that-nearly-was with Raoul Thomas moat in 2010.

The Raoul Moat incident also provides a vital clue about spree killings. Moat warned a friend he would go on a killing spree of police officers, and in that letter, Moat made references to the Cumbrian killing spree a month earlier by Derrick Bird.

Clearly Moat had been influenced with a way to settle his ongoing grievances with "the system".

Ryan's spree killings in Hungerford occurred just ten days after the "Hoddle Street" spree killings" in Melbourne, Australia, featured in the UK news at the time.

Similarly, when arrested after the spree killing of 35 people, and injury of 35 more in Tasmania in 1997, Martin Bryant claimed that he went on a killing rampage because he "wanted to beat the Dunblane record" that occurred a month before.

There is clear evidence that many spree killings are performed by people who have been influenced by other recent spree killings.

In the Virginia Tech massacre of 32 of his fellow students by Senghui Choi, the killer posted videos of himself brandishing weapons, and a copy of his "manifesto" to NBC and ABC tv networks in the USA shortly before he killed himself.

These video clips were used heavily in news coverage on the day following the massacre, and the grave concern by psychologists and psychiatrists at the time was that in repeatedly broadcasting the clips, the news networks where giving Choi the infamy he had craved but had only achieved through mass murder.

An open letter was sent to the news networks by concerned professionals, advising them to stop showing the clips for fear that they could instigate another killer. The networks complied, and to my knowledge, no additional spree killings occurred soon after the Virginia Tech tragedy.

Behavioural science could also help to identify why so many killing sprees involve children, or young people.

Partially because a frustrated vengeful killer will know that murdering children is the biggest "crime" to be committed against the system, and also because secondary schools, colleges and universities are relatively poor in terms of security.

In addition they also provide an effective offending space - many potential victims in a single location, easily compliant with instruction, unlikely to be able to physically overpower the killer, with a complex array of buildings from which to maintain a siege position with a tactical advantage.

This has been seen repeatedly since the first true spree killing involving school children by Andrew Kehoe in 1927 in Bath, Michigan, and perhaps more infamously in the Texas University massacre by Charles Whitman in 1963.

The names of Columbine High School, Dunblane Primary, Virginia Tech, Winnenden, Montreal College, and Cologne High School are but a few of many campus massacres to choose from.

Even Michael Ryan returned to his alma mater after his killings and held a final siege there in familiar terrain for a number of hours.

Because schools and workplaces provide a ready source of potential victims for spree killers, an "offending space" if you like, it is therefore of little surprise that most spree killings occur on weekdays between 9-5 as it guarantees the killer the highest impact and greater infamy.

Workplaces, particularly campuses catering for older children or students (rather than those for younger children who already have better defensive measures in place) must take heed of behavioural science in order to make themselves safer should they ever be targeted.

I said earlier that psychology may be limited in trying to predict who may fit the "profile" of a spree killer (and I use that word loosely) but that does not mean that psychology should not at least try.

I have placed UK spree killers / attempted spree killers on a "personality-circle" based on their case histories, to provide the best estimate of their personalities that I possibly can.

The most telling thing to see is that at best, each individual has a history of emotional instability - despite being introverted or extraverted.

The personality adjectives in the more stable, sanguine and phlegmatic quarters of the circle were not characteristics commonly associated with those killers in their lives.

In looking at the case histories of not just UK spree killers, but many others who claimed a grudge was the cause of their murders in other countries, it is possible to suggest a pattern of common experiences that, if coupled with a particular "blame-others immaturity" type of personality, could help to explain why some people chose spree murder.

I have suggested some common patterns of behaviour and personality among spree killers.

Spree killers often focus on a history of failures and setbacks in their lives; they display strong tendencies to blame others for any of those failings; they tend to be narcissistic and expect that they should have achieved more than they actually did; they often have an immature coping style and cannot deal with any undue strain or pressures in their lives; when these factors are coupled with a "final straw" that proves too much for the individual to cope with, a flashpoint is reached, and if the offender has access to the offending space and the lethal weaponry required, then a spree can ensue.

Spree killings are not the act of "mad" individuals - they are aware of the consequences and the horrific nature of their impending actions - but as immature and unstable narcissistic men who can no longer cope under severe strain, they believe it is the last choice they feel they have against a "system" that they feel (correctly or not) will not listen to them or help them.

Understanding the social and emotional immaturity of spree killers could be the key to future prevention.