1. A quick summary of the key findings of my just released paper with @CharlottadcM on the Geography of COVID-19 in Sweden: link.springer.com/epdf/10.1007/s…
2. Our paper examines the role of 2 kinds of factors in the geography of COVID-19 in Sweden.
3. Place-based factors like density, socio-economic disparity, age levels, versus diffusion factors like proximity to harder hit areas which would be associated with the spread of COVID.
4. Sweden is a useful case study to examine the interplay these factors because it did not impose mandatory lockdowns and because there was little or no geographic variation in the pandemic restrictions or policies.
5. Our research examines COVID infections & deaths across two geographic scales: all city-regions (municipalities) and neighborhoods in Sweden's 3 largest cities: Stockholm, Gothenburg and Malmö. The data cover the initial wave of COVID infections, February-August 2020.
6. Overall, we find diffusion factors to be significantly more important than place-based factors in the spread of COVID-19 across Sweden.
7. Proximity matters: The factor that explains the most variation in our models is proximity to other places with higher rates of infection. This factor explains more than half of the variation in COVID-19 across Swedish cities.
8. Also significant is the timing of first infection. Swedish cites & neighborhoods that were hit earlier by the virus seem to have been hit harder (a finding which is in line with previous research).
9. We find that place-based factors played a far less significant role in the geographic incidence & spread of COVID-19 in Sweden.
10. We find density and population size, two factors that the media seized upon early in the pandemic, to be weakly correlated to the geographic variation in COVID-19 infections – both in municipalities and neighborhoods.
11. Our findings for the variables for socio-economic status are mixed & somewhat counter-intuitive.
12. The variables for income and education are positively related with COVID-19 cases at the municipal level, which seems counter-intuitive. However, as in other studies, the variable for income inequality is positively associated with COVID-19.
13. These findings may reflect a more general result for city size, as bigger cities tend to be more affluent and educated but also more economically unequal.
14. The factor for disadvantaged neighborhoods is modestly associated with COVID-19 cases in the neighborhood-level analysis.
15. The findings for frontline workers also generate mixed results at the municipal level but are positively and significantly related to COVID-19 cases in the neighborhood-level analysis
16. The variables for immigrants are positively and significantly associated with COVID-19. This may provide additional evidence for the association between COVID-19 & disadvantaged neighborhoods, as immigrants in Sweden are concentrated in less advantaged neighborhoods.
17. The findings for age are also counter-intuitive. COVID-19 is negatively associated with the variables for age.
18. This likely reflects the fact that younger people were free to have contact with one another because Sweden did not impose any lockdowns, while older and more vulnerable people were more likely to isolate and physically distance.
19. The most significant place-based factor is the presence of high-risk nursing homes - not surprising given what we know about COVID in Sweden.
20. When it comes to place-based factors, our findings suggest that the geographic variation of COVID-19 is only weakly related to socio-economic factors, and even fewer of these variables are significantly directly related to deaths from COVID-19.
21. Place-based factors explain only a small amount of the geographic variation of COVID-19 across Sweden—in fact, roughly only ten percent of the variation in COVID-19 cases across municipalities and just one precent across neighborhoods
22. Perhaps the biggest takeaway from this study is that the geographic variation in COVID-19 and the vulnerability of certain places to it appear to have relatively little to do with their own characteristics.
23. The spread of COVID-19 has much more to do with factors which bear on diffusion, particularly location near hard-hit regions.
24. Ultimately, it is not so much place-based characteristics that mattered to the spread of COVID-19 but rather location, randomness, and bad luck.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
