A recurrent theme in security work is the hunch: when youre sure that theres something wrong, even if you cant say quite what it is or why youre so sure.  If youre a store detective, for instance, you might have a gut feel that theres something not quite right about the harmless-looking pensioner who is a popular figure at the local supermarket.  So you follow the hunch, and you keep a close eye on the pensioner, and you discover that hes a systematic shoplifter.  Once you started watching him, the evidence was easy enough to find, but what was it that made you start looking in the first place?  Thats one of the questions that G4S wanted to answer, as part of developing their training process.  To find those answers, they turned to someone who specializes in types of knowledge that are hard to put into words.

By Gordon Rugg

There are numerous different types of knowledge which are hard to put into words.  Some of these, like “the flow” in sport, have acquired a semi-mystical status.  In fact, the underlying processes for most of them are now fairly well understood, but the detailed explanations involve a significant amount of neurophysiology, so this issue is not widely known, or widely considered in terms of practical issues such as design of recruitment criteria or training programme design.  It is only recently that it has started to receive attention outside a few specialist groups.

G4S were well aware of the importance of “gut feel” in store detection.  They already had a good training program which covered the core skills, but there were a lot of other “gut feel” skills which the best store detectives had, and which nobody was able to articulate in a way that could be used in training.  Instead, it was accepted that these skills would have to be picked up the hard way, via experience on the job.  The problem was that this method was time-consuming and unreliable -- some people picked up these skills without difficulty, but others never seemed to acquire them properly despite years of experience.

GUT FEEL SKILLS

Investigating skills of this sort require specialized techniques -- you have to leave interviews and questionnaires behind when you go into this territory.  Interviews and questionnaires are fine if you’re dealing with knowledge that people can easily put into words (explicit knowledge) and especially good with knowledge that the person has already organized (codified knowledge).  They’re of little or no use when dealing with other types of knowledge.

This distinction became apparent as the project progressed.  The expert who trained the store detectives had extensive and well-organized knowledge which she could easily put into words.  The store detectives, in contrast, often had extensive knowledge and were highly skilled, but were unable to put most of their knowledge into words.  It was very similar to the difference between a sports coach and a sports player.  There were, however, many areas of expertise where even the trainer wasn’t able to put things into words, which took the project into the areas of tacit and semi-tacit knowledge.

TACIT KNOWLEDGE

Tacit knowledge, in the strict sense, involves skills which we can perform without having any understanding of how we are doing them.  Often we don’t even notice that we’re doing them.  A classic example is the skilled driver, who can manoeuvre and change gear without having to think about it.  Tacit skills are extremely valuable in most fields because they are very fast and very efficient; they are fast and efficient because they don’t require conscious mental processing, and are performed ‘automatically’ the equivalent of an organization performing a task at branch level, rather than having to go via head office.  In the case of driving, novice drivers have to process all their actions consciously, which is slow and liable to error.  Interestingly, if you force an expert to think about what they are doing, this can interfere with tacit skills, and cause the expert’s behaviour to deteriorate drastically.

Because tacit skills don’t involve conscious reasoning, they can’t be accessed via interviews, which can only reach knowledge to which we have conscious access.  Experts often think that they know what they are doing in their tacit skills, but these accounts often bear little or no relation to reality, however strongly the expert might believe them.  To get at knowledge of this sort is notoriously difficult, but if you know what you’re looking for, and watch the expert performing the right set of tasks, you can get a lot of useful insights.

STORE DETECTIVES

In the case of the store detectives, the tasks involved viewing video clips of store interiors, which included some episodes of shoplifting.  The most striking result from this was the speed with which the store detectives could reliably spot a shoplifter before the crime had taken place less than a second, for skilled store detectives watching some clips.

The store detectives were asked to explain what was happening in each clip, and their explanations fell into two main categories.  The first involved things which the store detectives could articulate, even if they didn’t have specific names for some of them, for example, using a legitimately bought item to conceal another item which was being stolen, such as a newspaper to conceal a magazine inside it.  The second involved things which they couldn’t articulate, such as someone walking in a suspicious manner they knew that the manner was suspicious, but they couldnt explain why.

Interestingly, the store detectives differed in terms of what they were able to put into words, so some of the ‘give-away’ signs can be taught explicitly as part of the training course.  An example is a behaviour which is hardly ever used by shoppers but which is used by many shoplifters, and which one store detective called ”shark’s  eyes”.  This involves looking for CCTV cameras, to identify blind spots which the cameras don’t cover.  Most people, if they want to look up, will turn their whole head upwards, which is easily visible on CCTV.  Some shoplifters attempt to get round this by keeping their head still, but looking upwards with their eyes, rather like a shark at the moment when it strikes its prey and rolls its eyes upwards.  Once you know the specific actions, it’s easy to identify this give-away; if you haven’t had them explained to you explicitly, you’lll probably notice that there’s something odd about the person’s behaviour, but not be able to say exactly what it was, especially for such a fleeting activity.

GIVE-AWAY SIGNS

What about the give-aways that nobody could put into words?  There are various ways of handling these.  The quickest solution in the short term is to teach them via what is known as implicit learning show numerous video clips, until the trainees learn to identify the give-aways, even if they still can’t say precisely what the give-aways are.  In the longer term, it’s usually possible to unpack the specific details via carefully structured studies, using methods such as systematically concealing different parts of an image to identify the key feature.

SEMI-TACIT KNOWLEDGE

The project also identified various types of semi-tacit knowledge which will sometimes be articulated, and sometimes not, depending on the situation.  For example, when the store detectives were commenting on the video clips they mentioned numerous things which they clearly considered to be general knowledge, and which they would probably not have bothered to mention in an interview, such as “doing a freeze” putting out an alert to store security personnel over the store public announcement system, so that a shoplifter would know they had been spotted, and would abandon the attempt to shoplift.  This type of “taken for granted” knowledge is a familiar feature in many fields; it is easily picked up via tasks such as the video clip task, but is usually missed by interviews.

By the end of this project, G4S had put the hunches of skilled store detectives onto a more systematic footing, in a form which could be incorporated into their training programme.  The project also identified other areas worth investigating, such as how experts mentally organised space, for instance, various zones within a store, to identify patterns of behaviour which might otherwise be missed.  Will that find anything which changes how security providers think about problems?  They can’t be sure yet, but they have a hunch.

Dr. Gordon Rugg is the head of the Knowledge Modelling Group at Keele University, and is a Visiting Senior Research Fellow at the Open University.