Collections > Electronic Theses and Dissertations > A test for detecting space-time clustering and a comparison with some existing methods

This research introduces a new statistical test for evaluating space-time clustering in data where exact location and time information are available for the points of interest (cases). The test statistic, DP, is defined as the length of the path from X[1] to X[n] when the n cases are ordered by time of occurrence. Significance of the test is most appropriately determined by comparing the directed path length of the data to the empirical distribution of lengths obtained from all possible orderings of the n cases, or a random subset of those orderings when n is large. The first three moments of DP are developed, and its properties are investigated using simulation on clustered and unclustered data. DP is then compared with Knox's test and Mantel's Generalized Regression using simulation on clustered and unclustered data. Two data sets from the literature (fifteen years of Burkitt's Lymphoma data from Uganda, and three years of birth defect data from California) are then used to compare the performance of these tests on actual data. This work was largely done as this author's dissertation under the competent, inspirational, and greatly appreciated advisement of Professor Emeritus Dana Quade, Department of Biostatistics, UNC-Chapel Hill.