Studies that find nothing don't get published.

100 fair studies will have 5 that show a positive result where none exists because of the statistical threshold for publishing.

Publish or perish means that researchers are motivated to find a result.

This leads to a lot of bunk getting published in peer reviewed journals. The Reproducibility Crisis is a major problem that gets almost no mainstream press. We may be better than pre scientific method but we are far from perfect in our methods.