Much has been written about junk science and how there is always another study which produces the opposite results from a previous one. p-value hacking is one way to do this. We have already mentioned metadata which is data about data.
Meta-analysis is the study of studies in which a research will look at all of the previous analyses and report what they all have to say about outcomes in totality. If your analysis agrees with other similar analyses that have been performed in the past, that would be more evidence to support your model.