We like to believe that ‘intelligence is based on objective analysis of hard data, albeit gleaned from secret sources and that the final product of that analysis is a factual account of what is actually happening out there in the real world. And that real world can be a potentially hostile nation (sometimes it can even be a friendly one) or a terrorist organisation. Once the analysis has been completed we are now furnished with accurate, reliable timely information that we can base actions on.

These actions can range from economic responses through diplomatic actions all the way up to preparations for war to actual armed conflict. Intelligence allows us to, if necessary, undertake the ultimate act of going to war or take the offensive against a terrorist group; in either case not blind or benighted by the fog of war, but armed with the most up-to-date information available.

Well that’s the theory. The reality can be radically different. Because intelligence services and the people who staff them are subject to the same predispositions and biases as the rest of us. And one of the most persistent of all is that of Confirmation Bias.

Confirmation bias is where we interpret the world according to our pre-existing   beliefs and notions. In other words, we don’t look at the world objectively but seek to fit aspects of it or patterns into a narrative that confirms to what we already believe to be true.

To illustrate let’s take a topical example. Two people, one who voted to remain in the EU, the other who voted to leave, are reading a textbook on the EU. The book is just that, a factual account that outlines the roles and purpose of the various EU institutions, including economic data. The book comes to no particular view of the EU: good or bad, pro or anti; just an overview of the facts.

From this impartial book our person who voted to remain will find ample grounds for concluding that Brexit is fundamentally wrong and is economic and political madness. Equally, our leave voter will see lots of grounds for why Brexit is the right decision for Britain. Both will be reading the same information, reviewing the same facts.

But for one the facts in the book will confirm that the EU leads to economic growth, cultural integration, free trade and free movement of people. For the other, the book will reinforce the belief that the EU is an over-regulated bureaucracy that swamps the independence and freedom of nations. Each will sift and filter the information contained in the book to arrive at the conclusion that best fits their views prior to reading the book.

Confirmation bias is pervasive throughout human conduct. It affects (or infects) almost everybody in all walks of life. From juries to doctors, from TV commentators to academics to you and me; all of us are partial to it.

Consider that most prosaic of examples: a football referee red-carding a player and sending him off. The fans will almost certainly be equally divided on the merits of this with fans of the team whose player was sent off loudly vociferous in their belief that the decision was utterly wrong while the opposing fans will be equally vocally adamant in their belief that the referee’s action was completely right. Both fans witnessed the same event, but the conclusion they took from it was utterly polarised.

Now, as mentioned, we’d like to think that people working in intelligence, must of necessity be relatively free of confirmation bias. Biased, prejudiced intelligence can be very harmful for the side believing in it and acting upon it. However, the history of intelligence is littered with intelligence catastrophes and failures that can trace their origins back to confirmation biases. And, when you think of it, why should it be otherwise? Intelligence services consist of people. And people, irrespective of their qualifications and talents, are always susceptible to confirmation bias.

The most recent prominent example where confirmation bias had momentous consequences was the invasion of Iraq in 2003. Of course, the justification for the invasion was Saddam Hussein’s alleged possession of ‘weapons of mass destruction’ (WMD), which in one notorious ‘dodgy dossier’ became the capability to attack the UK “in forty-five minutes.”

Post-invasion it was revealed that hordes of Iraqi WMD poised to strike the West were in that apposite Scottish expression a pile of mince: there were no WMD in Iraq.

But the belief that they were and the wholesale succumbing to confirmation bias by large sections of the US and UK intelligence and military community was brought about by three factors:

  • Saddam’s need to project an image that he still had such weapons which in his own confirmation bias might act as a deterrence to an invasion
  • US and UK politician’s need to believe in WMD as a rationale and legitimation for a risky invasion strategy in the post 9/11 situation where spurious linkages and associations were being made (such as that between al-Qaeda and the Iraqi regime which were very tenuous).
  • Not least intelligence agents and assets were playing up to what they thought, in this case, rightly, their paymasters in MI6 and the CIA expected to hear by exaggerating or in some cases completely making up false intelligence, most notoriously the agent known as Curvelball. To be sure there were opposing voices in these agencies, but they were drowned out and dismissed by those who wanted ‘intelligence’ to reflect and endorse what the politicians wanted: i.e. an invasion.

It was in this cauldron of mutually reinforcing confirmation biases and group think on all sides that Iraq was invaded with consequences for the Middle-East and ourselves which we are now too well aware of.

How then would it be possible to at least try to minimise confirmation bias when it comes to analysing, disseminating and acting on intelligence? That will be the subject of a later blog in the next few weeks.