Patterns, Superstitions and the Scientific Method

People notice patterns. Our brains evolved to do this: to detect the shadow of a predator lurking in the bushes, to find the tracks of prey on the savannah. More, we find patterns over time: Which watering hole always attracts the most antelope, which stream is always dangerous in flood time.

Our minds are always searching for patterns, and always finding them – even when no pattern is really present. As pointed out in the Ted Talk by Michael Shermer "The Pattern Behind Self-Deception", the cost of missing a pattern is much greater that the cost of finding a false pattern: The cave dweller who missed the predator hiding in the grass was removed from the gene pool.

Rorschach testPsychologists have made use of the pattern-finding abilities of the human brain for centuries. For example, consider the Rorschach test, where a person is asked what picture they see in a random, symmetrical blotch. There is no pattern, of course, but the mind searches and finds one nonetheless.

This tendency to detect patterns, even when they do not exist, leads to interesting problems. Last week, a deformed lamb was born; this week, a storm destroyed the village. Do deformed lambs herald coming storms? Last week, a woman accidentally spilled grain near a dried up watering hole; this week, the watering hole again has water. Do we need to sacrifice to the water spirits?

Let us suppose that deformed lambs and dramatic storms have nothing to do with each other. Each does occur from time to time, more or less at random. Sooner or later, these two unrelated, but unusual events will occur near the same time. Our pattern-matching minds will notice this and create a pattern.

Two-headed lamb, painted by Bartolomeo Bimbi, 1648-1730When we look back at the primitive rituals of our tribal ancestors, we wonder: How could they believe those things? We know where storms come from. We understand that springs are fed by underground rivers, which are in turn fed by rain in distant mountains. We have swept away these false patterns. But our minds remain the same: How many people have an irrational fear of the number 13? How many people use some personal, secret method to choose numbers on a lottery ticket? How many people believe that electromagnetic waves cause headaches?

False patterns in the information society

In today's information society, good information is critical. Superstition costs lives! How many people still believe that vaccinations cause autism? This was a false pattern, ruthlessly exploited by a few charlatans for their own financial gain, and yet it lives on. Why? Because, occasionally, a child is diagnosed with autism shortly after a vaccination. Unrelated events do occasionally happen together.

To make things even more complicated, there is another possibility: correlation. Correlation means that two things happen together, perhaps because of some third factor we failed to consider. For example, a famous study from the University of Pennsylvania noted that small children who slept with the light on were more often near-sighted than children who slept in dark rooms. The researchers proved the correlation beyond a doubt, and therefore concluded that night lights caused near-sightedness.

It took further study to reveal the true reason behind this correlation: Near-sighted parents often have near-sighted children. Since the parents could not see well in the dark, they often left the light on, in case the child needed something at night. Hence, even though children with poor eyesight really did often sleep with the light on, these two things had nothing to do with each other. Both were caused by a third factor: the parents' poor eyesight.

Detecting false patterns

How can we detect and eliminate false patterns? We know how! We have known how for centuries. Alchemy is dead, magic discredited, both eliminated as collections of false patterns. How? The scientific method.

The scientific method is simple: Before we believe a pattern, we must test it. The test must have a clear yes/no result, and it must be possible for other people to perform the same test. If the pattern is true, we will all get the "yes" result. If results are mixed or negative, the pattern is false.

The scientific method is not only for scientists – it is for everyone in today's information society. Today's predators are people with false information. A lost Nigerian bank account? A Rolex for 20 dollars? An investment guaranteed to make you rich? A magical herb that will cure...everything?

What do you believe? How can you test? If you cannot test something yourself, what test could someone else do that you would believe? The Internet lets us share information. When you are presented with information, you can test it yourself and share your results. You can read other people's results. Together we can discredit the charlatans, and eliminate false patterns.

Two examples: Homeopathy and Global Warming

Let us take homeopathy as an example. Homeopathy makes an extraordinary claim: By drinking specially prepared water, you can cure a disease. There could be a true pattern here: drink the remedy and you are cured. But there is also a danger of a false pattern: People recover naturally from diseases. If a person recovers naturally, and also drank the remedy, our minds will try to create a pattern.

What would be a good test? What if we took two groups of sick people, gave one group plain water and gave the second group the remedy. Importantly: The groups cannot know what they are getting: this must be a "blind" test. If the pattern is true – if homeopathy works – then the group getting the remedy must recover the fastest or the most often. If the results are mixed, then the remedy is no better than water.

Of course, this test has been carried out, many times. The results are always mixed: sometimes one group recovers faster, sometimes the other. The homeopathic remedies work no better (and no worse) than plain water.

We only discover this information if we look. If we only listen to the people who believe in a pattern, we are only getting one side of the story. To find out if a pattern is true, we must search out the skeptics: people who question – and test – the idea. We must read about their tests, and decide if the tests are good. Do they have a clear yes/no result? Are they described so that anyone could repeat them? Have other people repeated them?

Correlation of global temperature and CO2Let's look at a second example, one with a clear correlation: global warming. If we look at the historical evidence, it is entirely clear that the earth is warmer than it was 200 years ago. It is also entirely clear that, through human activity, there is more CO2 in the atmosphere today than there was 200 years ago.

Is there a true pattern here? Is there causality, or only correlation? What test would we believe?

CO2 continues to rise. Assuming that this pattern is true, and that CO2 does cause dramatic global warming, scientists created models to predict how much warmer the earth would get.

Recall that a test must give a clear yes/no answer, and one that we can actually check. A prediction for the year 2100 is useless! We need tests that give an answer that we can check. Fortunately, some scientists did make clear predictions. If their predictions were right, we have good reason to believe that the pattern is true.

The IPCC predictions turn out to be wrongBack in 1991, the IPCC released their predictions for the coming two decades. This chart shows their best prediction (the middle line), as well as the top and bottom limits. As long as measuredtemperatures remain in this range, preferably near the center line, there is good reason to believe that CO2 causes global warming.

We can see that the predictions were wrong. Actual temperatures (measured two different ways) are shown in the blue and green lines. The test failed, the pattern is false. Human pride makes it difficult to admit when we are wrong; this graph was not published by the IPCC. It took a skeptic who wanted to test the idea. His test is a good one, because it uses public information that anyone can access, so anyone can recreate this graph.

Since then, scientists have come up with new models and new ideas. Are these new models correct? To find out, we only need to ask: Have they made predictions that can be tested? Has anyone tested these predictions? What are the results?


Good information is critical to today's society. We are flooded with information. We are flooded with sensational news stories. We are flooded with ideas, with proposals, with theories. Deciding which ones to believe and which ones to reject can have a dramatic impact on our lives and our society. Here are just a few examples:

  • Government must stop spending, if the economy is to recover!
  • Government must spend, if the economy is to recover!
  • Gun control will reduce violent crime!
  • If more people carried guns, there would be less crime!
  • Nuclear energy is dangerous!
  • Nuclear energy is safe!
  • Children should never talk to strangers – they might be abducted!
  • Child abduction is rare – making children scared is counterproductive!

How do we know what is right? Which patterns are true? The scientific method is not only for scientists. Everyone should understand it:

  • An idea must be tested, before we believe it
  • The test must give a clear answer
  • Anybody should be able to do the test

Since we cannot test every idea ourselves, we search out tests done by other people. The Internet makes this easy! When we do this, we should always look for people on both sides of the question: not only people who believe in the idea, but also the skeptics. Which group has the better tests? What are the results? Finally, we must remember that mixed results mean that a pattern is false; only consistent results indicate a true pattern.

Our minds seek patterns even where there are none. This was a huge advantage that helped us survive and evolve. However, in today's information society, false patterns carry a high price. We can use the scientific method to decide which patterns are false, and which are true.

Bradley Richards
February 2013