ABSTRACT
There are many different types of statistical methods useful in spatial epidemiology. Sometimes these are classified according to the nature of the data (area versus point based), or according to the nature of the statistical test (for example, distance based). One can also classify them according to the purpose of the analysis, such as: Descriptive (visual) methods for disease maps. Map smoothing techniques. Spatial regression. Spatial-epidemic modelling. Small area estimation. Tests for spatial randomness. In this chapter we survey available tests for spatial randomness. This emphasis is a reflection of what Jacquez (page 22) describes as a common situation where ‘visually striking thematic maps…are not subjected to statistical tests to determine whether an apparent pattern is statistically significant and thus merits explanation’. The same problem can occur with the modelling of spatial epidemics, and we should watch that we do not use increasingly sophisticated techniques to model random spatial noise.
