Wednesday, November 12, 2008

Gobbledy Gook

In college you have to read a lot of research and other stuff with big words. Researchers amaze me, not because they delve into such workings of the world as we've nary dared to go, or because they come up with answers to questions that have never before been answered. It's because they use lots of big words and then arrange them in such ways as to make them look like they make perfect sense. That's a real talent. Then their papers get published and people from all around the world get to act like they understand exactly what they are reading. "Aha," they say. "I heartily agree with that team's observation."

I have before me a 10-page article explaining a study that was done linking the use of cannabis to the risk of psychotic or affective mental health outcomes. In theory, it's a great topic. In reality, it will cause me to struggle for a good hour trying to decipher things like "We pooled studies using the DerSimonian and Laird random-effect model and the metan command in Stata (9.0)" and ignoring things like "...adjusted odds ratio = 2.58, 1.08 - 6.13; heterogeneity p = 0.049; K2 = 66.9%."

In fact, I propose that this article could cause perfectly healthy people to have a psychotic episode.


3 comments:

Dave said...

I love stats! It was one of my favorite subjects.

"Correlation does not imply causation". I love saying that. And living it.

It's funny, I once modified that exact same photo in the same way, but different text!

Pamela Larkin said...

If you need someone to try the cannibus and give a true opinion you can always count on me.

wildmary said...

Aha," I say. "I heartily agree with your observation."