Bad Study Habits

Friday, November 21, 2008


Most studies published in scientific journals, it turns out, are either exaggerated or wrong. How come? According to epidemiologist John Ionnidis, editors of science journals are no different than everyone else in media. Sensationalism sells.

Comments [4]

Chris Gray from New Haven, CT

This is just another example of how the love of money has corrupted all of our institutions. We are quite proud that we are the wealthiest society on the Earth.

What has it bought us but sham, fraud, debt and disaster?

Nov. 25 2008 02:49 AM
Eleanor from Natick, MA

I'm an immunology postdoc, and it's an open secret in the biomedical science field that articles in Nature and Science, two of the highest impact journals, are notoriously irreproducible. The general feeling is that Nature and Science rush to publish first, and are focused on hot, cutting-edge topics. Since everyone in the field takes these articles with a grain of salt, waiting for the results to be reproduced in more careful, lower impact journals, the consequences are not too bad. The problem is that science journalists who don't understand the system, pick up the newest, sexiest findings and run with them.

Nov. 23 2008 03:16 PM
Kate Plummer from Vermont

This was such an important story! As a certified nurse-midwife, I struggle to practice based on evidence every day. It is harder to do than most people realize. One problem is biases like these in how research is published.

Please refer to a new report on evidence-based maternity care--I strongly encourage you to report on this:


Nov. 23 2008 10:57 AM
Robert from NYC

How sad it this! The obvious is obvious, sensationalism DOES sell and it sells everywhere and in all fields. After all we all bought into the idea that a bowl of pasta in a restaurant owned by a "celebrity" chef and 1/2 the size, or less, of what we used to get is worth $25 and even more! We love to think we're all in on the joke and in reality we are the joke.

Nov. 23 2008 10:56 AM

Leave a Comment

Email addresses are required but never displayed.