In this mini-dissertation we briefly describe the context and development of spatial statistics, spatial sampling and point patterns. Thereafter spatial homogeneity is considered in detail.
Before selecting an appropriate sampling design in the spatial context, it is important to know whether the data is first- and second-order homogeneous. Currently the method of kernel smoothing is used to construct density plots which can be used to visually and subjectively infer on first-order homogeneity. We propose the use of hypothesis tests, developed for the comparison of K Poisson intensities from independent samples, in the spatial setting as a more rigorous statistical approach to testing for first-order homogeneity. We also discuss the data assumptions required for these hypothesis tests and provide suggestions for the users.
Mini Dissertation (MSc)--University of Pretoria, 2017.