Is there a Variation of the Effectiveness of these Models Across Space?
Mapping the residuals of the models allows us to identify the areas in which the model underestimated, overestimated or was effective at estimating the dependent variables.
If we compare the maps of the distribution of foreign born citizens and foreign born not citizens with their respective residuals maps, we can observe that in general, places in which the model underestimates the value of the outcome variable (represented by red), have high percentages of both variables. This means that at these places, other factors explain some of the variation in both dependent variables. Spatially, these places are located along the Border with Mexico, but there are a few around densely populated areas like Washington D.C., New York City and Chicago. This strong relationship leads to an improvement of the regression model trying to account for proximity to Mexico and population density.