Here is yet another puzzling effect found in psychological science, and perhaps elsewhere; the decline effect, that the size of an effect decreases after repeated replications:
https://en.wikipedia.org/wiki/Decline_effect
J. W. Schooler, “Metascience Could Rescue the ‘Replication Crisis,’” Nature, vol. 515, 2014, p. 9. The decline effect was originally seen in psychological research, but it has now been found to extend to drug testing, where the effectiveness of some drug diminishes over time, and even in the behaviour of non-human animals. There are many proposed explanations of the decline effect, and no doubt specific factors may be at work in different fields of science. For example, the decline effect in the effectiveness of certain anti-psychotic drugs may be a real decline in effectiveness as people adapt to them over time. Nevertheless, it is also possible that we may be seeing here something of a generalised Heisenberg uncertainty effect, where observing the universe and measuring it, actually changes the universe: even in physics, measurements of the charge of electrons and a constant governing the decay of neutrons have exhibited the decline effect as John Horgan states:
https://blogs.scientificamerican.com/cross-check/the-truth-well-doubt-does-the-decline-effect-mean-that-all-science-is-truthy/
This is to be expected in a quantum mechanical world which is participatory rather than mechanistically passive:
https://futurism.com/john-wheelers-participatory-universe/