Direction test
|
|
|
|
Indications/Recommendations for use: The spread of cases
sometimes has a directional component. Examples include an epidemic
front sweeping through a population, the spread of cases mediated by
vectors that themselves exhibit a directional preference and the
outward spread of cases from a pollution source. When directional
components exist the chain of infection from one case to the next will
proceed in a common direction, and the test will be significant
because the concentration will be small (more).
|
|
Description:
A test for space-time interaction that also calculates the average direction of the spread
of cases.
|
|
Test statistic:
Two statistics are calculated, the average direction of the spread of cases, and the
concentration (or variance) in the vectors formed by linking each case to the next case in time.
|
|
|
Null Hypothesis:
No association between the times at which two cases occur and the direction of the
vector formed by connecting the spatial location of the case that occurred first to the other case.
|
|
Alternative Hypothesis:
The direction from one case to the next is similar for cases that occur at about
the same time.
|
|
GeoMed Inputs:
Spatial locations of cases and controls, and the way in which cases
are `connected' to create vectors. The Direction test requires the
name of a point file and the method to use when calculating the
elements of the directed time matrix.
|
GeoMed Outputs:
- Results table
- average angle
- angular concentration
- significance
- Plot of frequency distribution of the angular concentration
- Map of chain of infection
- the average angle is shown on the map as a vector with the first case at its tail, and pointing in the average direction of advance
|
|
Example Analysis
|
Reference:
|
