Science Gets Better at Being Wrong | WIRED

"Science is about proving yourself wrong in order to be right. That means scientists are constantly checking their own and each others’ work, course-correcting towards captial-T Truth. Problem is, capital-S Scientists are usually too busy shoving coal into the new discovery engine to fact check the canon. University of Virginia psychologist Brian Nosek wasn’t having it. So in 2013 he founded the Center for Open Science. Its flagship product is the Reproducibility Project—100 canonical psychology experiments re-run for verification. The results came out this year, and things didn’t look great. Well over half of the experiments didn’t work out the second time around. To some, that represents millions of wasted research hours, and lost research dollars. But really, it’s good news because it provides data-driven evidence that the field’s hypercompetitive push to publish new research is hurting the science. And for snigglers who think psychology was a soft target, look out: Nosek is coming after cancer biology research next."