Dark Numbers. Modeling the historical vulnerability to arrest in Brussels (1879-1880) using demographic predictors

Recently it has been demonstrated that unseen species models from ecology can be meaningfully applied to historic datasets that are censored because of an imperfect observation process (e.g. loss of sources). Following the lead of researchers in statistical sociology, we apply these methods to a dataset from the field of historical criminology: the police registers of the Amigo prison (1879-1880) in Brussels, which record the names and demographic characteristics of individuals who forcefully entered the prison after being arrested. Our results are in line with prior work in this area that reported the presence of intersectional biases in this dataset: we observe that elder, local citizens of the female gender were generally more vulnerable to arrest (than e.g. younger males or immigrants). Using the Generalized Chao method, we can capitalize on these insights to estimate the “dark number“ of unapprehended perpetrators in this period and reconstruct the demographic composition of the unobserved share of the population.