Friday, June 12, 2009

Humorous Aside: The significance of scientific research?

Once a scientist's experiments are complete (measurements taken, data collected) the scientist is typically expected to show that their research is statistically significant. A previous article shows that the main statistical significance test that people use was created in a beer factory over a hundred years ago and was only used on sets of up to 10 samples (way too few by today's scientist's standards). If you collect a lot of samples, the test will almost always give you the significance you're looking for. Luckily I discovered a new statistical significance test on the internet that will allow us to compute a number to determine if our research was worth anything (thank you PhDComics.com).

(click to enlarge)