Science is Only Right 50% of the Time, but There’s Only a 10% Chance of That
A new study gives all doubters hope. Or less hope. We’re not sure.
According to the report, about half of all peer reviewed, scientific research papers published around the world aren’t true. (No, we aren’t going to get into the self-referential questions of whether the paper itself is true, so shut up.) A variety of problems contribute to the errors, mostly; poor study design, small sample sizes, selective reporting and others.
And while no one factor is causally identified, the combination of errors means that those studies which have found statistically significant results (results that can be shown to be large enough to not be the result of random factors) are, half of the time, not significant.
Studies in ‘hot’ fields–where multiple teams are competing against one another–tend to be even less reliable, and studies with small sample sizes are even worse.
We know what you’re thinking: you knew it all along. Scientists don’t have any answers, they just pretend like they do. After all, how many times do you hear about a study that says one thing while another a few months later says the opposite?
But, we have to ask, isn’t that kind of the point? Isn’t research a venture into the unknown? And aren’t wrong conclusions sort of expected?