When a Hypothesis is Not an Educated Guess
Mon, Sep 22, 2014 3:14 PM
Partly right
This article does a decent job identifying the relationship between hypothesis and prediction, but it muddies the water in some respects. For example, it is just not true that "For a hypothesis to be proven, it must be true in all instances." For example, in their example of the hypothesis that crickets hide to avoid predators, one might observe that they do this only 80% of the time when a predator is present. This begs the question of why crickets don't hide 20% of the time. Thi might be the result of confounding variables like hunger, but point is that the failure of 100% of the crickets to hide doesn't mean the hypothesis hasn't been "proven." Also, showing that the crickets are "really" avoiding light and not avoiding actual predators doesn't mean that the hypothesis has been proven wrong: it is probably easier for a cricket to see the light than to see a (camouflaged, stealthy) predator, so it may be that avoiding light is a good way to avoid predators. I also agree with the previous point about the "law" vs "theory." Laws are DESCRIPTIONS of nature, not causal explanations. Theories, on the other hand, are (as the authors point out) collections of well validated causal hypotheses.