EPID 684
Spatial Epidemiology
4/7/2022
Jon Zelner
[email protected]
epibayes.io
Defining the MAUP 😱 and the challenges it presents.
Implications of looking at the same phenomena over different scales.
Return of smoothing (if we have time…)
Aggregating up from low- to high-level spatial units results in biased estimates of measles outbreak risk.(1)
Different neighborhood boundaries at the same scale result in different estimates of local N02 exposure. (2)
Case rates of measles between Texarkana, TX and Texarkana, AR reflected state-level policy differences (3)
Variation in N02 exposure in Ottawa, Ontario (2)
Take 5-10m to identify using this miro board:
What are potential zonation effects for your project topic?
How about scale effects?
Not all exposures happen at a single scale (From (4))
Use individual-level data from Australian cities to examine correspondence between individual and neighborhood avg. wealth.
Build bespoke or egocentric neighborhoods around each location of increasing size.
Examine how the relationship between neighborhood average and individual incomes change with increasing size.
Not looking for the right scale, but instead for the implications of using each.
Higher-order structures constrain possibilities within lower-level ones.
The timescale of change at higher (e.g. city or region) levels is slower than at more-local levels.
What advantages do you see in taking a multi-scale approach to your problem?
What are the practical and theoretical challenges to doing so?
Examining segregation as a multi-scalar, multi-dimensional phenomenon