Humans tend to be superstitious creatures, meaning that we sometimes believe in magical causation – if I wear my lucky sweater my favorite team will win. Psychologists have been examining this strange phenomenon for decades, with some interesting results.

A recent study ads one more piece of information to the emerging picture of what drives superstitious beliefs and behaviors, but let’s first give some background.

Superstitious beliefs are primarily about the illusion of control – the feeling that we have some direct or indirect control over the outcome of processes over which we objectively have zero control. Gambling is a common everyday example – gamblers tend to develop all sorts of behaviors they believe will give them a better chance of winning at games which are random. Psychologists divide the notion of control into primary and secondary. Primary control is direct control – if I throw the dice with my left hand I will get a better result. Secondary control is an attempt to harness or align with an outside force, such as luck.

It is always difficult to tease apart the complex causes of human decision-making and behavior, and studies necessarily rely on artificial situations or markers of the behavior in question. But a few fairly clear signals have emerged from the research.

One apparent driver of gambling superstition is simple innumeracy – a poor grasp of math and probability. Outcomes which are random may not seem random if we rely on a naive sense of probability. If an outcome is not truly random, then there must be some factor at work, or at the very least there may be a pattern. Figure out the pattern, and you can increase your odds of winning.

But even when gamblers know an outcome is supposed to be random, they may still see patterns in the noise, and be convinced that the patterns are real. So pattern recognition is also an important driver of superstitious beliefs.

Further, there is a motivated factor – the desire for control. The flip side of this is that lack of control makes us feel anxious, and superstitious rituals (regardless of our level of true belief) may help alleviate that anxiety. Many studies show that lack of control drives superstitious beliefs. This is likely why professional sports players and high-stakes gamblers tend to be superstitious – they have high stakes in an outcome over which they have incomplete, or even no, control.

Tying the concepts together, there are studies that show that having a feeling of lack of control increases pattern recognition itself. So it seems that when we feel lack of control, we search harder for patterns that might help us gain some control. Since we generally have hyperactive pattern recognition, this will drive the perception of illusory patterns, leading to a sense of illusory control.

Further, humans have hyperactive agency detection – we tend to assume that events are the result of deliberate design, rather than random chance, and therefore there is an agent at work. So once we perceive a pattern, we then weave a narrative involving a hidden agent to explain the illusory pattern.

The recent study attempts to tie superstitious belief and the illusion of control directly together. The researchers used a standard laboratory design for the illusion of control – they had subjects press a button, while a light randomly turned on and off. There was no relationship between the button pushing and the light. They then questioned the subjects regarding their beliefs regarding whether or not the button controlled the light – believing the button controls the light when it doesn’t is an illusion of control.

But they also did something that they say no previous study has done – they also gave the subjects a test for their superstitious beliefs. They found:

This study found that the magnitude of this illusion was predicted by people’s level of endorsement of common superstitious beliefs (measured using a novel Superstitious Beliefs Questionnaire), but was not associated with mood variables or their self‐rated locus of control.

This supports the hypothesis that an illusion of control is related to superstitious beliefs. Exact lines of cause and effect may not be straightforward, however. It makes sense that having a greater tendency to have an illusion of control would increase superstitious beliefs, but it may be the other way around, or both may be driven by some other factor or factors. Either way, lack of control, the illusion of control, pattern recognition, and superstitious beliefs are all riding the same ship.

Why would this be? There does not necessarily have to be an evolutionary advantage to the superstition phenomenon. It could just be a consequence of abilities we evolved for other reasons. For example, the advantages of pattern recognition may be so great that evolutionary pressure favor hyperactive pattern recognition.

There is some preliminary evidence to support this hypothesis. The question is – what is the the relative harm of a false positive vs false negative in terms of detecting patterns, or the benefits of a true positive vs true negative? What the models found in the linked study is that the benefits of a true positive may be so high that they outweigh the negatives of many false positives.

In other words – detecting a real pattern may convey a significant survival advantage, while detecting illusory patterns may only have a small survival disadvantage. This would push the pattern-recognition system toward the hyperactive end of the spectrum, being very sensitive so as not to miss any real patterns, even though this results in many false positives.

I have written before about the fact that our pattern recognition system is essentially a screening system, with high sensitivity. We then follow this up with reality testing, which provides the specificity.

Of course we can use metacognition to transcend our natural tendencies. We can use logic and critical thinking to better understand the world.