MAPPING OF THE POPULATION AT RISK OF STATELESSNESS IN ALBANIA However, the numbers above tell us nothing regarding the relationship between residence in one district and risk of statelessness. In order to assert this correlation, the graph above has to be put in a different perspective. Figure 7 compares the proportion of individuals (from our sample) in one district with the proportion of the population living in that district. In this way, we eliminate the direct effect that the size of a district has on the number of persons at risk of statelessness in that district. In other words, we evaluate the relative prevalence of the issue of statelessness. The new graph indicates that larger cities like Tirana or Durres, which have large numbers of persons at risk of statelessness, are doing better, in relative terms compared to their populations, than smaller cities with fewer statelessness cases, such as Berat or Vlora. Figure 7: Proportion of Persons at Risk of Statelessness by District Against the Relative Size of the District Table 1 follows the discussion of the earlier paragraph. Dividing the proportion of identified persons at risk of statelessness in each district by their relative population size, one arrives at a number that represents the relationship between residence in a district and the risk of statelessness, which we label as the “coefficient of statelessness risk.” As mentioned previously, the correct scientific approach would be through multivariate binary logistic regressions (twelve in this case, one for each district) or through multivariate linear regressions (depending on the type of independent variables). Nonetheless, given the scope of this report, the following table provides a suitable approximation of the effect. UNHCR - TLAS / May 2018 49