Take 97 psychological effects from top journals which are claimed to be robust. How many will replicate? Brian Nosek and his huge team tried it out and the results were sobering, to say the least. How did we get here? The data give some clues.
Sometimes the title of a paper just sounds incredible. Estimating the reproducibility of psychological science. No one had ever systematically, empirically investigated this for any science. Doing so would require huge resources. The countless authors on this paper which appeared in Science last week went to great lengths to try anyway and their findings are worrying.
When they tried to replicate 97 statistically significant effects with 92% power (i.e. a nominal 92% chance of finding the effect should it exist as claimed by the original discoverers), 89 stastically significant effect should pop up. Only 35 did. Why weren’t 54 more studies replicated?
The team behind…
View original post 651 more words
Utopia, you are standing in it! 
Recent Comments