01 February 2011

What if there was a significant difference?

An Anatolian folk hero named Nasrettin Hodja* was once seen spooning yogurt culture into a large lake. When the onlookers reminded him that lake water wouldn't turn into yogurt, his response was "What if it did!"

Nasreddin was testing a hypothesis that was highly unlikely to be true. Yogurt forms when certain types of bacteria grow in milk and the acid they produce causes the milk proteins to precipitate. Compared to milk, lake water has an insignificant concentration of proteins. Thus, a lake cannot be made to turn into anything comparable to yogurt or cheese. Had Nasreddin known the chemistry and microbiology behind the formation of yogurt, he might not have bothered to carry out his test. But that's besides the point. The moral of Nasreddin's story is that we should not be afraid to test unlikely hypotheses. Not all the time, of course, but sometimes.

A couple of weeks ago, I thought of an experiment involving oppositely coiled, i.e., dextral and sinistral, snail shells. All I can reveal now is that by doing this experiment I am hoping to find an explanation for why most snail species have dextral shells. However, I don't exactly have an explanatory hypothesis, but only a vague notion. If at the end of the experiment I see a significant difference in the outcome between oppositely coiled shells, I will then try to come up with an explanatory hypothesis and then carry out further tests.

I ran the idea by my friend and colleague Tim Pearce. Tim objected to the experiment, which he called a "what if" experiment, on the grounds that such experiments have one interesting and one non-interesting outcome and that the latter is usually the more likely result. In other words, in my planned experiment a positive outcome, that is, a significant difference between oppositely coiled shells, is highly unlikely and, therefore, the experiment would be a wasteful undertaking.

Another concern Tim had is that negative results that are the more likely outcome of such experiments rarely get published. As a result, the same "what if" hypotheses may get tested over and over again, because we don't have a way of knowing that they have already been tested.

My attitude is that "what if" experiments should be done if they can be done without the expenditure of too many resources, money included, and effort and, of course, provided that they are ethical. So, if I can obtain enough number of dextral and sinistral snail shells, which is the largest hurdle in this case, I intend to perform the experiment. I did promise Tim that I will do my best to get the negative results published. In return, he will try to get me the shells I need.

And if I end up getting the unlikely positive outcome, so much the better.


Is it yogurt or what? Nasrettin Hodja may have been onto something.


*Nasrettin Hodja was apparently a real person who lived in the 13th century. But many of his tales were probably posthumous attributions.

No comments: