Thinking About Kids

Parents, kids, and the way we live together

Pirates, Piaget, and the Null Hypothesis

Did pirates wear eye patches to help see in the dark?

Did you ever wonder why 19th century pirates are often represented wearing eye patches?

I never did, because I assumed that it was for the obvious reason: they had lost an eye.  Pirates, like many seamen, had a dangerous job.  Not only was their primary job going into battle, where people were trying to hit them cannonballs, shoot them with pistols, or pierce them with sword or knife.  Just daily work was dangerous.  Climbing in the rigging, working with huge pullies, tremendous pieces of wood, and heavy canvas made crushing and puncture injuries a dime a dozen.  My father's family worked for generations in shipyards.  My generation is the first to make it to adulthood with all our fingers and limbs intact. 

And need I mention that sterile technique was non-existent?  Louis Pasteur, the proponent of germ theory, lived from 1822-1895.  'Surgeons' (who often were trained only in emergency first aid and had no formal medical training) not only had no access to anesthesia, they also had no idea that sterilization or alcohol would reduce infection.  It's not surprising that so many eyes, limbs and lives were lost.  

A Null Hypothesis: Dark Adaptation

My son was watching Myth Busters the other day - a tv show where a group of wildly enthusiastic scientist/engineer/geeks empirically test common myths and curious factoids.  It was one of their Pirate Specials.  And one hypothesis being tested: pirates wore eye patches to help them see in the dark.

It was a perfect example of how the null hypothesis works.

Psychology, like other sciences, tends to move forward based on Baconian principles.  In order to minimize the effect of cognitive bias and our tendency to love our own theories and inadvertantly confirm them, you need to work systematically.  One of the hallmarks of Baconian science is the null hypothesis.

The null hypothesis is always that there is no relationship, no effect, or no difference.  If you've taken statistics, you know that traditional statistics work by testing the null hypothesis.

My students HATE the null hypothesis, because it is a completely counter-intuitive idea. Here's how it works:

  1. Propose a research question and develop a hypothesis about what is causing the phenomenon you're interested in.
  2. Find evidence consistent with this hypothesis. 
  3. Systematically and exhaustively design a series of studies that provide alternative hypotheses to explain the phenomenon.  In other words, try as hard as you can to disprove your own theory.
  4. If none of those other alternative hypotheses pans out, your hypothesis is deemed to be plausible.  Or, in non-scientific shortcut talk: you're correct and get to dance the statistics 'happy dance'.

In other words, current scientific truth is the last theory standing after a systematic assault.

A simple example:

  1. Piaget and affilaited researches accumulated a huge body of research showing that when kids in the pre-operational stage (under 5) could not conserve number.  For example, if you asked them if two lines of candy had the same number in them when both were lined up 1:1, they would say yes.  (See diagram A below).  However, when the research spread them out and asked the child which had more (line B), they would say that the longer line had more candy.  Piaget argued that they could not decenterThey could focus either on length of the line or on the number of pieces of candy, but not on  both simultaneously. 
  2. Donaldson saw the same phenomenon that Piaget did.  But her explanation was different.  She tested the alternative hypothesis that what developed was the ability to understand the word 'more' and the ability to decontextualize.  In other words, language and the child's belief that if something had obviously changed and an adult asked them if something had changed, young children assumed the answer was always 'YES!'.  

Donaldson tested this hypothesis by having a 'naughty teddy bear' interrupt her experiment and mess up the candies so the lines were different lengths.  When the adult asked if there were stil the number of candies in each line (B), the young children were better able to say 'no'.

This is an example of Baconian science in action because a) Piaget had data consistent with his theory., b) Donaldson proposed a different explanation of the same phenomenon and provided data that was INCONSISTENT with Piaget but CONSISTENT with her theory. 

 

Then the scientific argument continued because Piagetian's argued that different phenomenon were assessed in the two experiments.

A critical point about how science works:

  • Just because Piaget had data that was consistent with his theory, did not mean this theory was correct.
  • Just because Donaldson came up with a different theory that was inconsistent with Piaget, doesn't mean her new theory is correct.  At the end of her experiment, it had simply not been disproved.

Which gets us back to pirates

The Mythbusters developed an alternative hypothesis to my naive notion that pirates wore eye patches to cover their empty sockets.  They tested the hypothesis that pirates wore eye patches to enhance night vision.

Think about their situation:

  • When on deck, they're out in the (often) blazing sunlight
  • When below deck, it is DARK.  I mean REALLY dark.  Solid wood above and below.  Any light provided by a small open hatch (at the higher levels) or a lantern.
  • Lanterns  tend to blind you, in that they are very bright and anything outside what they are shining on gets cast in shadow.
  • Lanterns were usually made of horn - peeled cow horn, for example - or mica, so they couldn't break.  Nothing is worse than a fire at sea in a ship filled with gunpowder.  Except a really dim lantern.

Eyes take time to adjust to the dark.  The Mythbusters' hypothesis was that if pirates wore eyepatches, they could speed this process.  With a patch over one eye, it is always dark adapted.  When they went into the hold, they could flip up the patch and see well with that one eye.  Mythbusters tested this empirically by designing an experiment where they performed tasks in dark rooms under two conditions: with and without the eye patch light adoption. 

Using the eye patch to keep one eye light adopted, they performed much better on a series of tasks in the dark.  (Mythbusters is available on Netflix as a download, so you can watch the experiment online if you've got a subscription.)

So Mythbusters provided evidence that their hypothesis is PLAUSIBLE.  In other words, there is data consistent with the idea that it could have happened.

Did it happen? 

Who knows.  There is no historical evidence that they found saying anyone ever did it.  Even if it would have worked.

And that's the thing about the null hypothesis.  Even if you provide evidence consistent with your theory and inconsistent with the null hypothesis (here the null hypothesis was that it wouldn't work), doesn't mean you're right.

It just means you're not wrong.

And that's how science works.

 

© 2011 Nancy Darling. All Rights Reserved

Follow me on Twitter! @pt_think_kids

Nancy Darling, Ph.D., is a Professor at Oberlin College.

more...

Subscribe to Thinking About Kids

Current Issue

Dreams of Glory

Daydreaming: How the best ideas emerge from the ether.