Originally published in Jane’s Police Review, June 22, 2007
The role of a police protection officer is fundamentally simple: protect your principal. But how do you do that in reality, when you face a crowd of more than 2000 people, all of whom have turned out to hear your principal speak. Looking out at the sea of faces, you scan rapidly, looking for that one face. A face you have seen only once in a grainy CCTV picture. A face that your information tells you poses a credible threat. And you have to pick that one face out from a crowd of thousands.
As humans, we are instinctively talented at recognising faces. Newborn babies will automatically orientate towards an image of a face within a matter of a few hours after birth, and will spend a greater amount of time looking at an image of their mother than they will at an image of a stranger. This ability to recognise faces we are familiar with originates from neuronal networks within the brain known as face recognition units. When we are presented with a face, we automatically scan the units in search of a match – have I seen this face before? If the answer to this question is yes, then the appropriate face recognition unit is activated, triggering the retrieval of related information, such as where did I see this person before, what is their name and so on.
The more frequently the neural circuits for a particular face recognition unit becomes activated (for example, by repeatedly seeing or thinking about a person) the stronger the memory becomes, as the bonds between the separate neurons get strengthened, thus making the memory more resilient and more easily accessible. That is why it is easy for us to visualise the face of someone we love, but harder to recollect the person the person standing in front of us in the cinema queue last week.
Normally, a face recognition unit needs quite a lot of stimulation to activate it, like the sight of the person it refers to. However, if it is already warm (that is, if it has been recently activated) it becomes sensitised and so does not require as much stimulation to set it off. Therefore, if we have recently seen a picture of the person involved or have spent a lot of time talking about them, we are likely to recognise them more quickly.
It takes us only around 170 milliseconds to process sufficient visual information to identify an individual. Unsurprisingly, we are very good at recognising the people we know well, even if all we are shown is a poor quality image. However, this task becomes far more difficult when we are asked to remember the face of someone for the first time. Even with a good quality visual image, we can find it challenging.
Applying the facts
Let us consider this in terms of the role of the police protection officer. We may know that there is a potential hostile in the crowd, and that this person may pose a significant threat to you and your principal. But it is not uncommon that all you will have seen is a scratchy, distant CCTV image., and that only for a few seconds before it got passed on to the rest of the protection team. Now you have to pick that person out of the crowd. Challenging, isn’t it? However, there may be some things about the subject that will help you find them in a crowd.
One point is, are they angry? Research has shown that the people are presented with a crowd of faces, those with an angry expression will have pop-out effect. That is, people will spot them more quickly than they will if the person has a neutral or happy expression. The same is true for sad and other negative facial expressions – they will quickly capture our attention and draw our focus in towards them.
Why is this? In evolutionary terms, it was important to us to monitor the emotional status of those around us, because what had an impact on them was likely to also have an impact on us. This was particularly true if people looked scared (has he just seen a sabre tooth tiger that may eat me?), sad (has a loved one just died from a horrible disease that may kill me too?) or angry (is he about to hurt me?).
These evolutionary tools have not died out, and so, when we see a negative expression on someone’s face, our brain – particularly an area called the amygdala, whose job it is to look for danger in the environment – puts us on alert and orders us to devote a considerable amount of cognitive energy to this potential source of threat. This means that we may well find it easier to spot those threats who are dangerous because of their anger, particularly if they are mixed in with a non-angry crowd of people.
Another consideration is what do they look like? Is your subject your average Joe? If so, it is likely to be far harder to pick them out of the crowd. If, on the other hand, the subject is in some way unusual or atypical, it will generally be easier and faster to recognise them. The reason for this is that when something or someone is unusual, we tend to devote more cognitive energy to processing their visual image – in essence we think about them more. This means that atypical faces tend to lead us to form stronger and therefore more accessible memory traces, making the subject more recognisable in subsequent exposures.
Controversially, the race of the subject and observer also plays a part. A paper by Christian A Meissner and John C Brigham entitled 30 years investigating the own race bias in memory for faces: a meta-analytic review highlights how people tend to be better at remembering faces of those belonging to their own race. This is known as the cross-race identification bias. For along time, this bias was believed to be the result of restricted inter-race contact. However, this conclusion found little support in terms of evidence.
A more likely explanation for this effect is that when we see someone of another race, we tend to process the categorical distinction of their racial identity more strongly than we process those facial features that make that person unique. Conversely, when we see someone of our own race, that categorical distinction is not present and so we have more cognitive energy available to devote to processing facial features. Similarly, evidence has suggested a cross-gender identification bias, in that we tend to be better at recognising those of the same gender as ourselves.
Another aspect is: what do you know about the person you are trying to find? Do they have any hobbies? Any pets? What is their favourite colour? Surprisingly, this seemingly irrelevant information can also help you to pick a face out of a crowd. Why? Well, when we are told that this person likes to fish, has a cat called Bob and that his favourite colour is orange, then we are able to form a far more elaborate memory trace than we would based on his picture alone, as our memory would also include all of this information about him.
Memories tend to be stronger when they can latch onto something else and when we can build connections between the new information and already stored knowledge (for instance, I live near a fishing store, that guy likes fishing, that guy has brown hair). This means that such memories tend to be more resilient and far more easily accessible when we need them.
When we scan a face, we tend to focus on external features – such as hair and face shape -to help us recognise the individual. This is especially true when the person we are trying to recognise is a stranger, and when we are using a degraded visual image (a poor photo or in low light conditions). The problem with this is that external features are relatively easy to change, thereby severely compromising our ability to make a positive identification.
A team of Australian psychologist have show that when people have a limited opportunity to view a photograph and are subsequently asked to identify the subject, those who focused on the internal features (eyes, nose and mouth) tended to have more success than did those who focused on the easily changeable external features.
It is also true to say that how we are feeling will have an impact on how easily we can spot a given face in a crowd. So, for example, if the observer is feeling anxious they will be faster at spotting angry faces, while those who are suffering from depression will find happy faces far harder to pick out.
The role of the protection officer is a challenging one, and this, seemingly minor, task of picking a face out of a crowd may in some circumstances be the hinge on which the success or failure of an entire operation may depend. It is worth bearing in mind that we are drawn to faces wearing negative expressions – a useful attentional draw if your threat is truly the angry man shouting slogans. Less useful if your subject is showing little outward expression of their violent intent (we often find evidence of limited facial affect in depression and psychopathy), as your attention is likely to be pulled towards that angry but ultimately less dangerous individual elsewhere in the crowd.
On a similar note, it is worth the protection officer taking note of how they are feeling, as this too may influence their ability to spot that face. Research conducted by Andrew Yip and Pawan Sinha of the Massachusetts Institute of Technology in 2002 shows that it can be valuable to show people colour rather than black and white pictures, particularly if the image is blurry as the colour can help us identify features such as skin tone and hairline.
As we mentioned earlier, the more information that you have on a subject, the better your chance of identifying him or her. Even if the information you are being given sounds completely irrelevant, it can help you to form stronger memories and so make identification easier.
That said, it is important to avoid relying on stereotypes to support any visual information available – particularly when those stereotypes will influence the officers’ visual search patterns. So for example, a briefing note states that: ‘Our subject is John, a university professor, so look out for someone smoking a pipe and wearing leather elbow pads’. This is a cliche, but there is a real danger that encouraging officers to think in stereotypical terms will lead them to inappropriately discard people when they are scanning a crowd.
We also need to remember that it is our natural tendency to look to the external features when we are shown a picture of someone we do not know. The problem here is that these features are quite easily changed, and so it may be valuable to make a conscious effort to take in the structure of the more immutable internal features.
There is one final thing to remember, there is some evidence of what is known as the verbal overshadowing effect, by which verbally describing a subject following a visual exposure can make it harder to pick them out of a line-up. The jury is still out on the actual cause of the effect, but essentially it seems to be that giving a verbal description interferes with people’s ability to retain the visual image of what the subject looks like, thereby making it far harder to spot them in a crowd.
Where possible then it may be worth providing each team member with photographs to reduce the need to rely on shared verbal descriptions and to allow greater opportunity to study the faces of the subject or subjects, thus enabling greater depth of processing and so, in a likelihood, easier recall.
Dr Emma Kavanagh is a consultant providing training to police firearms units, hostage negotiators and other specialist police teams throughout the UK.