For almost any paper I read, before I pick it up, I know there were statistically significant findings. Those results were not predefined. Instead they are part

For almost any paper I read, before I pick it up, I know there were statistically significant findings. Those results were not predefined. Instead they are part of the academic review process that filters out null results. That same process (depending on the field) also rewards novelty or all sorts of other things. It's very much a human process.