DailyDirt: Bad Science Is Coming to Get Us
from the urls-we-dig-up dept
Scientific publishing has been a lucrative industry in recent years, even though scientists have faced increasing competition over limited funding. The publish-or-perish academic model may be contributing to an increase in scientific fraud, but maybe the increased accessibility of digital journals is simply making it easier for honest mistakes to be caught. The scientific method is supposed to weed out incorrect conclusions, but there may be a lot of wasted effort as scientists try to replicate experiments that are just completely fictitious. It gets harder and harder to make decisions based on evidence -- if there is growing uncertainty that any evidence can be trusted....
- The number of retractions from scientific journals has increased tenfold over the past decade. But it's not clear how much is misconduct and how much is honest scientific mistake... [url]
- Data detective Uri Simonsohn has published his statistical methods for exposing the suspicious data of social psychologists. Lies, damn lies and statistics... but at least statistics can be used to ferret out the lies. [url]
- Apparently, the UK is notorious for its bad science journalism. We're talking "labvertisements" -- industry/product-funded science stories about (possibly fake) studies conducted by questionable scientists with dubious methods. But at least they're honest about it and take their research with a huge grain of salt. The US just re-packages many of these reports as serious news. [url]