Uncovering the Patterns of the US Geography of Immigration by an Analysis of Spatial

US Geography of ImmigrationJ. Hasman, J. Novotný

Uncovering the Patterns of the US Geography of Immigration by an Analysis of Spatial Relatedness between Immigrant Groups

JiříHasman, Josef Novotný

Supplementary Material: Analyses with the population groups defined by the world region of origin

As there should be some risk that the results of our analyses can be affected by the matter of immigrant groups’ delimitation (e.g. by the number of groups), we show in this supplementary results for some tables from the main part of the paper with the data based on the much broader groups delimited by the world region of immigrants’ birth instead of their country of birth. We aggregated countries of birth into 14 regions, whose delimitation is shown in the map in the upper left part of Figure 3.

At first, Table S1 shows the same results as Table 2 in the main part of the paper. As we can see, the maximal values of Di,jare very similar in both definition of population groups, however average and median values are much higher here as with much larger groups, there is a higher chance that each pair of groups have at least some co-occurrences, so the values in the left part of the distributions of Di,jare much higher.

Table S1 Basic descriptive statistics for particular sets of spatial relatedness observations (Di,j); population groups aggregated by the world region of birth

Note: As we analysed 14 population groups, the number of Di,j observations is 91 in each of the spatial systems.

The Table S2 corresponds to the Table 6 in the main part and it shows correlation between the sets of Di,j results obtained for all data sets. As the aggregation by the region of birth is much more rough, Di,j values cannot capture more specific features of individual data sets, so the sets of Di,j are mutually much more similar than when using the country of birth data. The base results however remains the same: the highest correlation (0.8 – 0.9) can be found between data on the national level, somewhat lower between sets of Di,jfor individual MSAs (0.5 – 0.8) and relatively lowest between sets of Di,j for metropolitan and national systems (0.3 – 0.7).

Table S2 Pearson correlations between the sets of Di,j results obtained for different data sets; population groups aggregated by the world region of birth

Notes: M/MSAs denotes Metropolitan and Micropolitan Statistical Areas. All coefficients are significant at 99% level according to Mantel correlation tests for relatedness matrix data.

Then, Tables S3 and S4 show the pairs with the highest Di,jin the six MSAs and in the whole USA respectively. When we used the country-birth data, the strongest links were between pair of groups within the same region (Tables 7 and 8). Similarly, we can see, that proximity of groups matter as the strongest relations can be found between groups from the same continent now; for example, 4 out of 5 pairs with the highest Di,jin Table S3 consist of two Asian groups. Like in Tables7 and 8, the tops of the distributions at individual systems do not overlap:we can see that the Asian pairs are dominating in Atlanta, Houston and Chicago, while pairs between European groups (and North America) are the strongest in Los Angeles and Miami. In New York and on the whole US level, the strongest spatial relatedness was measured between Carribean and South America.

Table S3 Top 20 pairs of population groups with the highest average spatial relatedness in the six MSAs (average Di,j); population groups aggregated by the world region of birth

Notes: CEE denotes Central and Eastern Europe, MENA Middle East and North Africa, SEA South-Eastern Asia.

Table S4 Top 20 pairs of population groups with the highest average spatial relatedness at the whole US level; population groups aggregated by the world region of birth

Note: *Average of the Di,jobservations normalized by the respective averages of particular Di,j distributions obtained for the three levels of spatial disaggregation at the whole US level. CEE denotes Central and Eastern Europe, MENA Middle East and North Africa.