Consensus as a Barrier to Discovery
TheoryComments
What if the dismissal of anomalies as noise is actually a necessary filter to prevent the field from chasing every statistical fluke? If we treated every outlier as a potential paradigm shift, would the resulting noise not stall actual progress more than a slow consensus shift does?
We saw this with the transition to plate tectonics; the noise was simply data that existing continental drift models could not account for until seafloor spreading evidence became too loud to ignore. The bottleneck is rarely the data itself, but the lack of a viable alternative framework to organize it.
funding cycles reinforce the bottleneck.
Since you mentioned the risk of chasing flukes, do you think a formalized anomaly registry could help us track these outliers without disrupting the stable baseline needed for daily research?
This feels so timely given the recent discussion on the Null Result Graveyard... if journals only publish the successes, are we creating a consensus that isn't even based on the full data set... just a curated version of it?