From the article: > Mayr’s observation draws attention to the fact that fashion is never merely additive; it demands we abandon certain things, and those aband

From the article: > Mayr’s observation draws attention to the fact that fashion is never merely additive; it demands we abandon certain things, and those abandonments can be, at the very least, premature. I've observed these premature abandonments in my own area (a particular subfield of fluid dynamics), but I think its cause is more complicated than "academics follow fads". I think fads are part of it, but the grant funding cycle is another part. Perhaps grants are largely driven by fads. I wouldn't really know, but I can recall calls for proposals on fad X as applied to Y. I think part of it is that people rarely do actually comprehensive literature reviews, so they don't have a good idea of what's been done and what hasn't been done, so they might assume something isn't as novel as it actually is. Or their literature review might make them think that X is solved, but a little-known paper showed that X is actually more complicated than is commonly believed and not solved. This dynamic means there's a lot of "low-hanging fruit" available that won't look like low-hanging fruit to most people. I'm happy to grab such things myself. A few years ago I made a large data compilation for a problem I'm interested in (~1200 data points from ~50 studies) and have written two papers from that already, with several more planned. Many people seem to implicitly assume this work has already been done because it seems "obvious", but I think large data complications of this variety are rarely ever done. Many things that are commonly believed are actually easily shown to be false with the right data. I'm planning to make a career of this. It's not sexy right now but that could change with time. Edit: Data compilation is a commonly missed "last step" in a review. Most reviews I've read seem to collect the conclusions of analyses of small amounts of data. It's better to instead collect all the data, filter out the bad data, and draw conclusions from the compilation instead. A collection of the conclusions of small studies will have bad data, biases, and/or not generalize as well.