From vaccinations to climate change, we make decisions every day that implicate us in scientific claims. Are genetically modified crops safe to eat? Do childhood vaccinations cause autism? Is climate change an emergency? In recent years, many of these issues have become politically polarized, with people rejecting scientific evidence that misaligns with their political preferences. When Greta Thunberg, the youthful climate activist, testified in Congress last month, submitting as her testimony the IPCC 1.5° report, she was asked by one member why should we trust the science. She replied, incredulously, “because it’s science!”

For several decades, there has been an extensive and organized campaign intended to generate distrust in science, funded by regulated industries and libertarian think-tanks whose interests and ideologies are threatened by the findings of modern science. In response, scientists have tended to stress the success of science. After all, scientists have been right about most things, from the structure of the universe (the Earth does revolve around the sun, rather than the other way around) to the relativity of time and space (relativistic corrections are needed to make global positioning systems work).

That answer isn’t wrong, but for many people it’s not persuasive. After all, just because scientists more than 400 years ago were right about the structure of the solar system doesn’t prove that a different group of scientists are right about a different issue today.

An alternative answer to the question—Why trust science?—is that scientists use “the scientific method.” If you’ve got a high school science textbook lying around the house, you’ll probably find that answer in it. But this answer is wrong. But what is typically asserted to be the scientific method—develop a hypothesis, then design an experiment to test it—isn’t what scientists actually do. Historians of science have shown that scientists use many different methods, and these methods have change with time. Science is dynamic: new methods get invented, old ones get abandoned, and any particular juncture scientists can be found doing many different things. And that’s a good thing, because the so-called scientific method doesn’t work. False theories can yield true results, so even if an experiment works, it doesn’t prove that the theory it was designed to test it is true. There also might be many different theories that could yield that same experimental result. Conversely, if the experiment fails, it doesn’t prove the theory is wrong; it could be that the experiment was badly designed or there was a fault in one of the instruments.