Really interesting finding, thanks for sharing, and a super interesting topic more broadly. Did you consider the issue of systematic biases in quantity estimation, dependent on the unit of measurement, as a possible explanation of discrepancies between perception and reality as well? (cont.)
Comments
https://link.springer.com/article/10.3758/s13423-017-1360-2
So for example in the Herda paper you cite, the average country-level proportional overestimation is more than double the real proportion; but the *correlation* between the two is strong: 0.84. Is it that people just struggle with estimating %s?