A landslide inventory, statistical analyses and a Geographic Information System (GIS) were used to analyze landslide sites and adjacent terrain in the Oregon Coast Range. The objectives were to i) evaluate the potential for locating landslide sites with topographic variables, and ii) discriminate the difference between sites where landslides have and have not occurred. A landslide inventory provided a population of known landslide locations characterized as up-slope, non-road related, and associated with 1996 storm events. Topographic variables (e.g., slope and drainage area) were measured in the field at landslide sites and derived for the entire study area using a gridded Digital Elevation Model (DEM). Eleven different Topographic Indices (TIs) were subsequently determined from the topographic variables forming five broad groups; i) slopes, ii) drainage areas, iii) ratios of slope and drainage area, iv) infinite slope models - SHALSTAB, and SINDEX, and v) functions of slope and drainage area based on statistical models. Several different algorithms were used to derive slope and drainage area variables. A logistic regression (i.e., landslide occurrence versus non-occurrence) performed on slope and drainage area variables created an optimized hazard index using the logit. We plotted cumulative landslide occurrence against a continuous domain of cumulative TI area. The maximum landslide density was determined from the slope of this curve. This provided a common metric on performance between TIs indicating the ability to discriminate the location of landslide sites. Multiple linear regressions of soil depth on slope and drainage area variables at landslide sites suggest the association was highly dependant on watershed sub-basin location. This regression model assumes a physical link between geomorphology and slope stability. Logistic regressions suggest a significant difference between landslides sites and adjacent terrain. This was expressed by a higher retrogressive probability of landslide occurrence. Further, Some topographic indices appear to discriminate better than others. However, measurement of TI performance was confounded by DEM error, and GIS techniques.