Tag Archives: correlation

6. How to Think Like a Scientist – The Custard Effect

How to think like a scientist for kidsCustard makes me puke. Violently. Specifically, chocolate custard. It’s my working hypothesis. I only have two data points, but that’s enough for me. Talking like a scientist, chocolate custard causes me to puke. The effect of me eating chocolate custard is: me puking. The arrow of cause ==> effect is one way. My need to puke did not cause me to eat chocolate custard.

Cause and effect. Every scientist studies causes and effects. The cause happens first, and then later there is an effect. Always. It’s so obvious that at times we forget. Cutting onions make me cry. Onions cause me to cry. The effect of cutting up onions is crying. And crying does not cause me to cut onions.

But, when I cut onions, I also always listen to the radio. These two things happen in pairs, but are completely unrelated.

This leads to one of the most important parts about thinking like a scientist: just because things happen together, does not mean that one necessarily caused the other, or that they are even related. Talking like a scientist, if things happen together it is called “correlation”. The biggest mistake people make is when they assume correlation also means causation. There is a correlation between the onions and listening to the radio, but there is no causation.

My favorite football team wins when I wear my lucky T-shirt. Did me wearing my “lucky T-shirt” cause (or even help) my team to win? Probably not. It’s more likely that I often wear that shirt on game days, and my team wins a lot. Correlation does not mean causation,

But sometimes it’s hard to know what’s going on. Thinking like a scientist, I observe:

Every night my dog barks at midnight.

How does my dog know its midnight? She even adjusts for daylight savings time! Is it magic? Does the clock cause my dog to bark?

No! After further investigation, it turns out that she barks every night at midnight because that’s when my neighbor lets his dog outside. Midnight and my dog barking have a correlation. But the cause of the barking is my neighbor’s dog. The time of night had nothing to do with it.

But sometimes it’s hard to know what causes what.

Suppose you find that all the kids who wear glasses also sit in the front row of your class. Does sitting up close cause their eye problems? Or are they sitting up close because they have eye problems?

The great thing about noticing correlations is that it leads to ideas for new hypotheses and experiments. The experiments will provide evidence either supporting causation or falsify the idea.

For instance, I could ask my neighbor to let his dog out at 10pm to see if my dog will bark early. Or I could do custard flavor experiments. Maybe I only puke on chocolate custard and I will love strawberry custard…

Fun Phineas Effect Facts
One of the most common grammar mistakes in the English language is confusing the words ‘effect’ and ‘affect’. Wow your English teachers by never mixing them up.

Effect with an E is a noun, meaning ” the result of”. What was the effect of me eating chocolate custard? Puke.

Affect with an A is a verb meaning “to influence or change”. Verbs are action words. Remember: A for action. Bad weather affects my mood. So does puking.