In what contexts is this true and in what contexts can this be considered true and vice versa? Personally, I think it has a lot to do with confirmation bias, especially in natural sciences where for example certain researchers may develop a relationship that they deem true simply due to the data they collected. The best example I can think of is N-Rays, where the guy who supposedly discovered them, Prosper-René Blondlot, simply suffered from the experimenter's bias. Are there any other examples of this in human sciences? Or even history or Art?
Recency bias is a well-known and ubiquetous cognitive bias. It is mostly unrelated to the N-ray fiasco, which was a case of confirmation bias and priming in the context of bad experiment design. The corrective measure for these and most other cognitive biases in science is to use properly blinded experimenters or data-analysts and properly controlled experiments.