Reverse publication bias

See also Sesardić's conjecture, a related idea. Publication bias as normally considered is really positive publication bias, i.e., the bias is away from zero, towards finding larger than reality results.…