One in seven science papers is not to be trusted, says new science paper

Such a high level of fraud presents an “existential threat to the scientific enterprise”, the paper says.

featured-image

The young woman strolls along the cobblestone street. A gentle breeze blows in from the coast, ruffling her shoulder-length hair. She draws her hands from her jacket pockets – and a glove, unnoticed, flutters to the ground.

It is this saccharine scenario, a decade ago on France’s sunny southern coast, that started James Heathers’ journey into the dark underworld of science. A woman drops something. What would you do? Credit: Getty Images That journey concluded last month with a dramatic claim: one in seven scientific papers is fake.



The woman is a scientific experiment. She has deliberately dropped the glove to see how passersby respond – and if the way she wears her hair changes that outcome. Sometimes she wears it in a pony or a bun.

Some men and women bend down to grab the glove. Others don’t. The study , published in 2015, reported something extraordinary.

If her hair was down, men responded to the glove “like a plate of lamb chops to a starving lumberjack”, Heathers writes in an analysis of the study. Simply putting her hair down made them desperate to help. The effect size reported in the paper was “gargantuan” – so big, Heathers couldn’t believe it.

Digging through the underlying data, he was able to show the “natural hair effect” was apparently one of the most powerful psychological phenomena ever reported. But the underlying data was potentially problematic. The Scandinavian Journal of Psychology eventually published an expression of concern, and Heathers started on his journey to hunt out fabrication and falsification within the scientific literature.

.