![](http://www.sph.umich.edu/geomed/images/smallgeomed.gif)
|
Scenario:
A data set consisting of case and control locations is to be analyzed for
evidence that the case pattern is representative of a contagious disease
process.
Questions:
- Assuming that the data set contained no time of diagnosis, but was a
"snapshot" of the disease from an early point in the epidemic for the area,
you might hypothesize strong spatial clustering in the case locations.
Can you construct a gamma statistic that would show high values to indicate
strong clustering in this case, based on the spatial pattern of nearest
neighbors?
- Suppose now that you introduce time of diagnosis into the data
set. What would be your hypothesized distribution of spatial/temporal
pattern? What real-life factors would play a role in either accentuating or
obscuring your hypothesized pattern?
Return to Scenerios
|