The trouble with "Research shows"

Post Reply New Topic

Very interesting piece of commentary which looks at the parallels between the current fashions in science taking off as fashions in investments do, from time to time, creating "bubbles" which often prove to be unfounded and unsustainable.

It's a fresh way of looking at the research issues. I hadn't thought of it before.

http://theconversation.com/scientific-r ... -one-33797

"In neuroscience, a growing number of scientists, psychologists and philosophers have begun to raise concerns about the explanatory ambitions and research designs. A recent study adds to these woes, suggesting that a wide range of neuroscience studies lack statistical power to back their findings. In a paper published in Nature Reviews Neuroscience, Kate Button and her colleagues reviewed the statistical power of studies – defined as the function of sample size, significance of effects, and positive predictive value – published in renowned neuroscience journals. The investigators found that, on average, studies detect a genuine effect only about 20 per cent of the time. The review concluded that many of the published findings might not be reliable in several important respects, leading to overestimating effect size and explanatory significance"

Paragraph quoted from a blog by David Budz Petersen, London School of Economics