Much to try to ignore

The other day, I referred to an article in The Atlantic about this man, John Ioannidis, a professor at a university in Greece.  Here is the main article he has published on the subject of most medical research results being wrong, especially the highly publicized ones.  The end of the Atlantic article has a nice summary of the reasons for so much error in studies and contradiction between them.  The most outstanding factors seem to be

• the competition for headlines,
• attempts to get exciting publishable results,
• the attempt to conduct research that will be funded and re-funded, and
• commercial support for research that pushes a given product or medicine.

It is definitely possible to run a large study which finds a "statistically significant" result that is almost certainly not mere chance but which is actually so small as to be insignificant in any practical sense.  Sort of like winning the lottery and then finding the pot is only 1 cent.