The root cause is their need to reach for conclusions, probably from the outset of the study in order to justify their so-designed "experiment". Statistical studies should come without any pretense of a conclusion from the researchers. There are always multiple conclusions to be had, and students are naive. If you have to start with a statistical conclusion, you are already lost.
Not sure I’m understanding this. The post mentions that pre-registering was very helpful for the precise reason that it prevents researchers from altering their statistical criteria in order to suit their desired conclusions.
The article is about measurement, which the author doesn't define one bit. Measurement reform is too late in the process to mitigate conclusion chasing. This is a problem of academics and politics, not scientific methods.