16 May 2010

How not to misinterpret statistics

ScienceNews has an article by Tom Siegfried, titled "Odds Are, It's Wrong". Siegfried presents a pretty good analysis of a few common errors in statistics (just be sure to skip over the overdramatic introduction). The most major of these, in my mind, is misunderstanding the meaning of statistical significance.

If you've forgotten everything you learned in your last statistics class and are a bit hazy about what the difference is between these two statements,

Experimental data yielded a p-value of 0.03.

With 97% certainty, the observed difference between the groups was real (and not due to chance).

...then this article may help.

No comments:

Post a Comment