I've been uber busy lately, but I was just reading this article on the use of statistics in science and wanted to share it.
It's from Science News on March 27: Odds Are, It's Wrong
It's a discussion of how reliant scientists have become on statistics even while not fully understanding how to use statistics, and how statistical significance has been misinterpreted as real world significance.
This topic appeals to me because it was something I struggled with as a graduate student. My dissertation work required statistical evaluations of genetic associations. I even had a biostatistician on my dissertation committee because my advisor was dead-set on it being done right. I met with him time and time again, trying to understand what I was testing and why. I had to limit my lessons to only what I needed for my work, fully planning to expand and learn it better afterwards. Sorry to say, I don't even remember the appropriate methods used to compared results from my own dissertation - statistics are just beyond my long-term comprehension skills (I'm also not good at calculus and physics!)
But this is what science needs - dedicated math people to work alongside bench scientists, ensuring that the results are properly compared and interpreted. There also needs to be that moment of teaching where a future scientist is told "statistical significance is not significant on its own".
That article is very good and very important, and I just wanted to share.
TIP:
Enter the amount, and click "Tip" to submit!
Toby Beth Jarman
Dr. Evan Levine
Oryoki Bowl
Fay Paxton
lorianne
Mary Ann Sorrentino
Salon.com
Comments
Statistics are so easily misread unintentionally and so easily misused intentionally that they are an important entity that must be understood fully to have any value at all.
Good post.
RATED
Rick, thanks!
Oryoki, I've had numerous arguments with people along those same lines. "the odds are against it!" "but that doesn't change the fact that it happened!" "but the odds are against it!" "but. it. happened. Something has to happen, and yes, the odds are against any particular thing on its own, but overall, something. happened. and this. was. it."
The “scientific method” of testing hypotheses by statistical analysis stands on a flimsy foundation. Statistical tests are supposed to guide scientists in judging whether an experimental result reflects some real effect or is merely a random fluke, but the standard methods mix mutually inconsistent philosophies and offer no meaningful basis for making such decisions.
I don't think this is at all the case. Sure, there are arguments in the statistical literature (e.g. between frequentists and Bayesians), but the author might as well say that because of disagreements between idealists, realists, and nominalists, science is also a "mix of mutually inconsistent philosophies and offer[s] no meaningful basis for making such decisions." I mean, we don't even have a universally agreed upon philosophical justification for induction!
thanks!