Errors built into thinking

The brain does not merely err from time to time. It errs in predictable ways.

The 1974 paper by Amos Tversky and Daniel Kahneman did not begin by accusing people of stupidity. The authors showed how fast heuristics — availability, representativeness, anchoring on an initial figure — let us judge under uncertainty and at the same time produce systematic errors.

In one experiment a randomly generated number became an anchor for a subsequent estimate, even though the participants understood it was random. The brain did not believe the number; it simply started adjusting from where the number had put it. Knowing the name of the error does not make a person immune to it. It only gives a reason to check where the calculation started.

We notice confirmations of what we already believe. We fear losses more than we enjoy equivalent gains. We treat what comes to mind easily as more probable. We explain our own actions by circumstances and other people’s by character.

Knowing the names of these errors confers no immunity. A person can read a list of cognitive biases and immediately use it to diagnose everyone around him.

It is more useful to build procedures that stop an error from passing unnoticed:

• look for data against your own version;
• write the forecast down before the outcome;
• ask what would change your mind;
• separate fact from interpretation;
• bring in someone who holds a different position;
• check the decision after a pause;
• use a checklist where the same error recurs.

The hardest bias to notice is one that protects a pleasant conclusion. We readily check information that gets in the way of our wishes, and accept information that supports them remarkably fast.

So the question should not only be:

Where might I be wrong?

But also:

Which error is it especially convenient for me not to notice right now?

The mind stays imperfect after you have studied psychology. Some of its tricks simply stop being entirely new.

An anchor is easy to catch in a builder’s estimate. The first contractor quotes four hundred thousand for a bathroom refit. The second quotes three hundred and twenty and instantly looks like a bargain, although the job is worth two hundred and the figure of four hundred came from a man who simply wanted a lot. From then on every calculation runs downwards from four hundred, and a negotiated discount of thirty thousand feels like a victory. The first contractor is not necessarily a crook: he may have named a price off the top of his head, for himself, so as not to undersell — and still set the frame of the conversation for a month ahead.

The defence is not in resisting: whoever knows the word “anchor” will yield too. The defence is in the order of operations — your own estimate of the scope and the price first, other people’s numbers second. Order is cheaper than willpower.