Research Mindset

What would good look like if we had more of a research mindset? A Parade Room where it’s not about feeling you’ve got to have all the answers but more about ‘how you think’: a research mindset.

Let’s take a look at how you think is such an important aspect of research, and how some approaches can really help with the dynamic reality of the complexity of modern policing.

Let’s talk about the Reverend Thomas Bayes. His early work – 150 years ago! – was all about a pragmatic alternative to pursuing pure certainty from data precision, offering a science more based on belief and trial and error. The concept behind Bayes’ mathematical theorem is simple: in seeking to predict what happens next, we rely on our past experiences and we amend our beliefs as we go along. ‘So what?’ you may still be thinking: but just bear in mind that these principles aided the defence of the UK in WW2 when seeking to crack the Enigma Code, when all other research and scientific approaches failed. The need for a dynamic approach when the problem faced is uncertain is the very reason why the principles of Bayes are relevant to us now as policing professionals. There are, for example, huge parallels with the National Decision Making Model and critical incident decision making logs. Police Critical Decision Makers use the NDM and such decision-making approaches use a Bayesian approach – hence you probably have been doing all of this without realising it is a valid form of science.

The principles of Bayes’ theorem are all about inference and probabilities, a little like betting odds. The key point is to state a desired intention, then update and review as more evidence or information becomes available.  This makes the approach quite dynamic – just like policing! Bayes has been applied in many other fields too, from medicine through engineering to sport. So, it’s all about rational decision making where all the facts are maybe yet to be established but can be used where something needs to be done in the meantime and relies upon a transparent audit trail. Sound familiar?

This quite valid approach to science and research is maybe not quite the same stuffy white-coated and clip board carrying stereotype you perhaps were guarded against. Trial and error, charting what you’re trying to do, setting the conditions, monitoring, measuring and learning is indeed science. The big compromise with Bayes is that thinking embraces uncertainty rather than seeking to eliminate all doubt before being sure. Policing is so dynamic it is rarely that action can wait for perfectionism in data.

Some relate to this Bayesian approach as a form of logical inference. In this way it’s a bit of a middle ground between deduction (‘I have a theory and I’m going to test it’) and induction (‘I’m going to wait and see what emerges before jumping to conclusions too early’); that indeed is what abduction is all about. Thinking about Bayes as an abductive approach is fine – whichever way is better for you to understand the points and hopefully adopt them in your professional policing.  The abductive logical inference approach has that inductive balance of observing, but then compromises to use what appears to be most likely explanation at that time. The conclusions are not as fixed as in deductive thinking, but rather remain open minded for ‘what next’. It perfectly captures the essence of Bayes.

Another way of thinking about the principles of Bayes boils down to how you might think about acquiring a skill in pretty much anything. How do you learn something? Well, in simple terms you have a go, review what you’ve done and try again. Consider learning to drive a car – you learn through practice and experience, based on a body of theoretical knowledge. As you keep learning your driving improves, based on that experience, as you adjust what and how you do things. You perhaps remember the point from the Police Driving Manual Roadcraft – about the difference between someone who has years of experience compared to someone who has one day’s experience repeated over years?

National Decision Making Model

Let’s have a look at how using the NDM in a spiral loop is indeed quite Bayesian. The NDM is a risk assessment tool for pre-planned and spontaneous incidents. It was intended as an update to the conflict resolution model, and was developed through ACPO as a consistent framework for all decision making.

It breaks down into six parts:

Code of Ethics – Principles and standards of professional behaviour

Information – Gather information and intelligence. Now here we should be thinking about what do we know, what don’t we know, and how trustable any information we have, is. It helps here to also maybe ask a few more questions than we sometimes like to. What appears to be happening? Why might that be so? What might be the problem rather than the symptoms?

Assessment – Assess threat and risk and develop a working strategy

Powers and policy – Consider powers and policy

Options – Identify options and contingencies. Here we can be thinking what’s been tried before in similar circumstances, as there’s no need to re-invent wheels unnecessarily.

Action and review – Take action and review what happened. This is classically Bayes – even where things maybe don’t work out as we hoped there is huge learning for what to try next. We do not need to pretend everything works all the time, as it often doesn’t and that’s not about failure it’s about learning how to refine things and try again. You know nearly everything that has been invented in the history of the human species has been developed in spite of at first failing, often many, many times, and even appearing ridiculous. So often, only through being methodical and showing persistence do we get advancement. Not convinced? Well consider some pretty significant things that were condemned by the oh so clever naysayers at one time as stupid ideas: the telephone, the airplane, global sea navigation, train travel, television, the internet, and once upon a time, professional policing. So there.

If you’re already using the NDM, for example in a critical incident log, you’re already stating a strategic intention or objective, recording rationale and making decisions about what to try and constantly reviewing the impact and altering intervention accordingly. It’s a highly sequential yet dynamic approach, with a clear audit trail for accountability, and all geared up taking action based on what we know at the time.

You might have thought that science only comes out when it’s safe; so, you get research telling what you didn’t do right or what you might have done according to that eternal law called hindsight. Whilst some science does inform the evidence base by looking back, other approaches also measure dynamically. Bayes is a case in point that it isn’t seeking to be confirmatory – to confirm something – it is more a template to work in the dark when the variables (the information) are loose, missing or maybe conflictual.

It’s a little like having a form of structure when you’re exploring something new and dynamic. Bayes helps keep you rooted in your intention (which can also be dynamic as things change), keeps things auditable and transparent and can always be built on for more learning in new situations.

So what?

Professor Ken Pease of University College, London, points to the perceived tension over the uses of science in policing – where some approaches to introducing science to policing have been a little patronising and heavy handed. In some cases, it even comes over that a scientific approach somehow devalues professional experience – as if professional judgement doesn’t count. Bayes truly values experience as a vital and legitimate part of the dynamic process of monitoring and adjusting to real-time results: it encourages you to be in the ‘field’ and not just think about understanding as something done remotely or conceptually.

Now here is an additional point, illuminated by Terry O’Connell, a former policing practitioner from New South Wales, Australia. Terry believes pretty much all of policing is about relationships, between colleagues and the public, be they offenders or victims. It’s the relationship between people that allows connection and through meaningful conversations to get to ‘what matters’. That focus on relationships is not then just about following ‘processes’, but opening up and listening to each other more.

In having a real conversation – as opposed to going through the motions as we so often do – what happens is that people can ‘open’ up. Opening up can mean sharing their vulnerability.  Imagine in policing operational tasking where everyone is being held to account over who’s done what (or not) and should have done. It can be stressful in there, people feeling under pressure to be ‘right’, to know everything that’s going on and know all the answers. That’s good isn’t it? No, it isn’t. Why? Well, deep down we probably will admit to ourselves, even if no-one else, we don’t actually know everything. No-one does. It may come as such as a relief that everyone is in that same boat. So, you could actually stop pretending, you may feel a whole lot better.  

What we have then is potentially a room full of people who do not know it all.  This you may think must be kept a secret from the public, as the police of course always know stuff, and it could cause panic on the streets. The reality is the public do, overall, trust the police very much, but there have been many cases when the public will be aware the police, like any other bunch of professionals, have gotten it wrong. It’s what makes us human. That vulnerability then is a problem, yes? No, it’s a great opportunity.

Combine what Terry O’Connell advocates with the Bayesian science that Ken Pease argues for and what have you got? A research mindset. A powerful open-minded, thinking group of people who can operate very effectively in the dynamic complexity of modern policing, drawing upon all our emerging technology, but using the best of what makes us human.


Comments are closed.