Fig. 11
From: Large scale analysis of violent death count in daily newspapers to quantify bias and censorship
Relative occurrence of least significant digit for \(10<k\le 100\). Top: domestic, Bottom: foreign events. Left: full dataset, Right: period of Fascist government in Italy. The dashed line shows the expected value from a random distribution (0.1). The values of ‘1’ and ‘2’ are closer to this value, with some defect in Italian newspapers and excess in NYT. ‘3’ and ‘4’ are below expectations and their defect is close to the excess of ‘5’ events. The least probable digits are ‘6’ to ‘9’, usually rounded in excess to ‘0’. Only domestic events during Fascist government in Italy are not incompatible with equiprobable distribution