Witch burnings and the availability heuristic.

Here’s a quick quiz. No googling.

  1. In the average year, what kills more people globally, shark attacks or people intentionally punching themselves in the face to death?
  2. Foxes, or drinks machines?
  3. All accidental deaths combined, or strokes?
  4. What is a greater danger to the lives of the average Israeli citizen, terrorism or bad driving?
  5. In the aftermath of the twin tower attacks, where did the terrorists claim the most kills, in office buildings, in airliners or on highways?
  6. Do NRL players commit more or less criminal acts than the population average?
  7. Which has the best safety record in terms of deaths per unit of electrical output, solar power, coal power or nuclear power?
  8. Who has a better goal scoring average, Cantona for Man Utd or Lampard for Chelsea?
  9. In 2011, did the Islamic citizens of America account for 1,000 murders, a 100 murders, or 0 murders.

Now try this bonus question, which is slightly less relevant.

10. Lily is a shy, introverted girl who cherishes time alone to read and think. She has a meticulous nature and enjoys collecting and sorting objects. She prefers peace and quiet over society and conversation, being indoors over being outdoors, and the past over the future. Is Lily more likely to be employed as a librarian, or a school teacher?


The correct answers, in order are:

  1. Suicide self-punching
  2. Drinks machines
  3. Strokes
  4. Bad driving
  5. The highways
  6. NRL players commit less
  7. Nuclear power
  8. Frank Lampard
  9. Zero.

And by the way, Lily is more likely to be a teacher.


Our brains are constantly scanning our memories, searching for patterns, producing conclusions. Our minds, lazy as they are, deploy a veritable swiss army knife of shortcuts to assist them in their day to day work of analysing the world. These shortcuts save time and energy, and at stages in our evolutionary past they have no doubt been a lifesaver.

Imagine for a second a nomadic Neolithic lady that we shall name Boris. Now imagine Boris going about her daily existence, eking out a living by scavenging juniper berries around the savanna’s of Tunisia. Boris sees a moving orange blur dashing across the grass. She has to make a choice, stand up or hide. Her mind retrieves for her every image that it can find that fits the criteria of ‘moving orange blur,’ and disseminates its conclusions to her almost instantly in the form of adrenalin, fear and a possible explanation of the blur. She suddenly remembers a fireside memory of a cousin talking about a friend who was mauled to death by a large cat. Our lady concludes that this is probably just such a cat, and she hides, and survives.

She was saved by an association shortcut. ‘Orange blur’ was almost instantly translated by her subconscious into ‘threat,’ and that kept her alive.

The availability heuristic is one of the most essential shortcuts we use, and one of the most misleading and destructive, if not carefully watched. People, when attempting to work out the probability of an event, often base their estimations on how easy it is for them to bring to mind examples of that event.

This clearly worked well for our lady in the savanna. It does not always work so well in the modern media dominated world. Neolithic chick had a relatively relevant, independent sample that she used to form her opinion. She brought to mind her own experiences, and the experiences of other people much like her. She had quite a good, if small, data set, and from it she correctly concluded that an orange blur was a lion.

But for us, in this modern world, much of the ‘data’ that our brains run off is not relevant to our context, and it is certainly not a fair sample of reality. Accordingly, our snap associations often lead us astray, causing us to hide from kittens and causing us to stand up for lions.

The media reports a disproportionate number of extreme events, events that are so rare and incredible that they are newsworthy. You are also far more likely to read about events that interrelate with other newsworthy issues, that involve celebrities, or that fit the dominant meta-narrative of the time. It is pretty easy to chart whether a given issue is likely to be over or under reported. Here’s a few examples relevant to our questions above, and my opinion as to whether said issue is likely to be under or over reported relative to the issue’s importance:

Under Reported Over Reported
Ordinary, unglamorous deaths Dramatic accidental deaths
Nuclear disasters
Celebrities behaving well Celebrity misdemeanors
Flashy sporting heroes

Here is an indisputable fact, we are very bad at certain intuitive statistical calculations. The average person does not have a great sense of conditional probabilities, of the importance of sample sizes or of the power of correlations. That would be a good, if short, summary of the life’s work of the Nobel Prize winning Daniel Kahneman and Amos Tversky (two of the very best psychologists of the 20th century).

So, if the media is at work pouring information into our mental databases, and our mental databases are the resources that fuel many of our snap decisions, then in an era of skewed reporting we have a problem. If our data banks are overflowing with distorted images of reality, we need to be more skeptical of our availability heuristics. We are over exposed to extreme events, and accordingly we are likely to think that terrorist strikes, murders, break and enters etc (anything Hollywood worthy) are more likely to occur than they really are.

I’ll focus on three examples that illustrate the danger of this over exposure:

On September 11 2001 a touch under three thousand people were murdered by nineteen terrorists. This was the news story of the year, perhaps even the decade. Global airline companies saw their stock prices collapse as the towers fell. For years afterwards Americans cut back on their air travel, afraid of the risk of another attack. It was as if all of America had had a data point seared in the strongest bold across their cortex. Airlines were instantly associated with risk and with death. Accordingly, long distance car trips replaced interstate flights for many people. Highway congestion increased, and so did the number of highway crashes. In fact, more people died on the roads of America *avoiding* air travel, that died directly in the terrorist attacks themselves. Long distance car trips are riskier than long distance plane trips, statistically speaking, and an awful lot of people that could have safely boarded a plane, instead died in car crashes. The majority of the victims of the September 11 attacks were on highways, not in office towers or planes, victims of the availability heuristic.

When I think of nuclear power, the first thing that comes to mind is Fukushima, and images of the (originally) fifty men who stayed at that radiated plant, breathing caesium, with geiger counters buzzing, obeying their prime minister with that famous brand of inspiring Japanese self sacrifice. The second thing I think of is a burning Chernobyl, unmasked Soviet troops, and lines of Russian army trucks clad in a dusty yellow glow.

So yeah, on first impression, nuclear power sounds pretty dangerous. But I need to be careful here. Nuclear accidents are the very definition of newsworthy, and yet I can only think of three such incidents (three mile island being the other). And just how deadly were these incidents? Fukushima produced zero direct deaths, but it is expected to result in 100 to 1,000 deaths in the next hundred years. Chernobyl is probably at around 50,000. Three mile island is probably zero. These seem like massive numbers!

But in many decades of global operation, with power having been provided to millions of homes, are the numbers so large?

Six hundred thousand people die every year in China alone as a result of coal pollution. That amounts to twelve Chernobyl’s per annum in one country from coal power. And that’s current. Chernobyl was running prehistoric equipment in a crumbling and disorganized Soviet Union twenty five years ago. Coal power is doing this in China as we speak. And Fukushima, when you think about it, was an old reactor that had not been properly maintained, lacking modern safety systems, due for refurbishment, that was hit by a once in a lifetime earthquake, and then a tidal wave, and then a fire, and the result was zero direct deaths.

By way of contrast, more people fall from their roofs and die whilst installing solar cells per annum than die of nuclear power related issues, per unit of electricity. Falling mechanics just don’t make the news however…

Now there are other arguments against nuclear power that deserve attention, such as non-proliferation and all that, but as a direct answer to the question that was asked, nuclear power is the winner.

In 1989 Meryl Streep appeared before congress to testify as to the risks of a chemical called Alar, which was used on apples. The press had reported that large amounts of the chemical, when given to rats, caused cancerous growths. As the issue continued to grow, hysteria swept the airwaves. A lady called in to a talk radio station to ask whether it was safer to pour her orange juice down the drink, or take it to a toxic waste dump. A nation of 250 million people threw out its apples.

No deaths from Alar have ever been reported, but the public health effect of millions of people not eating apples, and replacing them with other foods, was probably quite substantial.

Three examples of people overreacting to a classic ‘Hollywood’ threat (terrorism, nuclear power and chemical poisoning) in a way that is very detrimental to public health. You may disagree with me, you may be against nuclear power for instance, but as far as the death statistics go, you should see that I have at least somewhat of a point.

So why is the availability heuristic something that I am bothering to write about? Firstly, because being able to rank things by importance is essential to life in general. Secondly, because it is absolutely key to understanding politics (ever noticed that politicians only seem to care about ‘Hollywood’ issues?) Thirdly, because you cannot accurately answer the question, ‘has religion, on balance, been a force for good throughout history?’ without a healthy awareness of the availability heuristic.

I hope to come back to this in a later post, consider this the groundwork, the scaffolding, the introduction.

My contention would be that the things that religion was done poorly over the centuries are memorable, newsworthy, film-worthy (crusades, witch burnings, child sex scandals etc). By contrast the many things that religion has done well over the centuries (for Christianity – the building of hospitals and universities, improving pagan attitudes towards women, slaves, foreigners, orphans, widowers and the sick, providing the foundation for the birth of science, providing a foundation for the concept of human rights, funding most of the world’s charities etc), are less glamorous achievements, less instantaneous, without the clear punchline or movie script.

To neglect the good things that religion has done over the centuries, as has become fashionable amongst non-historians in this modern age, is to blindfold oneself to a substantial proportion of the past, confuse oneself as to the reality of the present, and delude oneself as to the prospects for the future.

I would agree with the historian David Bentley Hart that, “Many of today’s most obstreperous critics of Christianity know nothing more of Christendom’s two millenia than a few childish images of bloodthirsty crusaders and sadistic inquisitors, a few damning facts, and an even greater number of even more damning legends.”

The new atheists are slaves to the availability heuristic.

Part two coming out whenever I feel like it.


Congratulations if you worked out why Lily is a teacher.

There are forty times as many teachers in America as there are librarians. Even if she is shy and a lover of books, the odds are that she is less likely to be a librarian that a teacher.

Love the start to this song, it’s like being punched in the balls by a swarm of pitchforks: