Median Watch

Eyes on statistics

Testing baseline tables in trials for signs of fraud

When fraudsters make up research data, they can make mistakes. Real data is rich and complex whilst fraudsters are on a get-rich-quick scheme and make slapdash errors. One mistake they make is in randomised trials, where it’s standard to have a baseline table that compares the randomised groups. As the groups are randomised, the summary statistics should be similar. Fraudsters have no sense of ‘similar’ and so have created data where the groups are nearly identical.

Posted by Adrian Barnett Wednesday, April 23, 2025

Painting a picture of research fraud

Originally published in the Deeble and AusHSI newsletter. Vincent van Gogh only sold a few paintings during his lifetime and died a poor man. After he died, his sister-in-law, Johanna van Gogh-Bonger, cleverly built his reputation and created a thriving market for his paintings. Recent sales have been over USD $80 million. Vincent was both truly gifted and desperately unlucky. Soon after his paintings started to sell, van Gogh was the victim of fraud.

Posted by Adrian Barnett Wednesday, March 19, 2025

Aggravating acronyms (AA)

My eleven year-old uses lots of words that go over my head, such as “skibidi”, “bruh” and “floptropica”. I’m not meant to understand their conversations and that’s rad. I also struggle to understand many scientific papers. Sometimes it’s my fault, as I’m too tired or too dumb. But sometimes it’s the authors’ fault, as they’ve drowned their ideas in verbiage and acronyms. We wrote a fun paper showing that scientific papers are using more acronyms.

Posted by Adrian Barnett Monday, February 10, 2025

Funding schemes that cost more than they award: part 2

“What is the purpose of research funding?” Phew, that is a big question. If I could hear your answers, I am sure there would be a wide range of opinions. Let’s put that in the too hard basket. What about this question, “What should research funding not do?" Again, I can imagine lots of answers, but there are likely some answers that we could all agree on. How about an axiom that “A research funding scheme should never cost more than it awards.

Posted by Adrian Barnett Wednesday, February 5, 2025

90% of scientific research is crap

Reading Adrian Edmondson’s excellent autobiography, he mentioned Sturgeon’s law which is: “Ninety percent of everything is crap”. Adrian is a comedian and was applying the law to his creative work. Sturgeon was using it talk about science fiction, but I think it also applies to scientific research, and Sturgeon’s number is strikingly similar to the estimate from Chalmers and Glasziou that 87.5% of health and medical research is wasted (which they rounded down to 85%).

Posted by Adrian Barnett Tuesday, January 7, 2025