This first of three appendixes of the user manual for CrimeStat IV - a spatial statistics package that can analyze crime incident location data - presents methods for testing the spatial differences between two distributions.
Currently, CrimeStat does not include routines for testing the differences between two or more samples. Chapter 4 of the user manual discusses the calculation of these statistics as a single distribution. This appendix first discusses differences in the mean center of two samples. For differences between two samples in the mean center, both differences in the "X" coordinate and in the "Y" coordinate must be tested. Issues addressed in this discussion are significance levels and tests. As an illustration, the distribution of burglaries and robberies in Baltimore County (Maryland) are compared for 1996. This is followed by a major section of the appendix that addresses differences in the standard distance deviation of two samples. Burglary in Baltimore County is used as an example. A major discussion of differences in the standard deviational ellipse of two samples considers differences in the mean centers, in the angle of rotation, in the standard deviations along the transformed axes, and in the areas of the two ellipses. Significance levels are also addressed. The concluding section of the appendix considers differences in the mean direction between two groups. 2 tables, 2 figures, and 5 references