Background on Hypothesis Testing


One day I came across an interesting story concerning citron. Apparently, sometime during the first century, an Egyptian magistrate sentenced a group of convicted criminals to be executed by means of poisonous snake (yikes!). The group would be paraded through the city to an arena filled with deadly venomous snakes. The sentence was carried out; however, the magistrate was informed that none of the prisoners died. He was forced to start an investigation at once, and learned that while the criminals were paraded through the streets, people had given them citron to eat.

This intrigued the magistrate, and the very next day he ordered the criminals divided up into pairs. One of the pairs was given citrus and the other was not. One by one, the pairs were exposed to the snakes yet again. You can probably guess the outcome. Those that were given the citron were fine, those that were not died instantly.

Thus we learn that citron is an antidote for poison (this is an old story, so please don't test this with lemons and rattlesnakes!).

Whether true or not, this is an excellent example of a very simple hypothesis test. Upon hearing that his criminals did not die like they were supposed to, the first century Egyptian magistrate formed the hypothesis that citron had spared their lives, and tested that hypothesis to see if he was correct.

But what else happened here? Maybe this situation is not as simple as we think.