Abstract

A common problem in geology is the objective evaluation of points that are suspected to concentrate nonrandomly in a specified part of a map area. Evaluation can be rigorous with the binomial test, but results often are distorted by prior inspection of the only available group of points. Such distortion can be hard to avoid because concentration may not be suspected until the sample is inspected. Distortion can be evaluated by analyzing overlapping subsamples, mapping independent cells that are significantly overpopulated by points, and arguing that the sample is representative of the population from which it was drawn.

Such distortion occurs in evaluation of earthquake hazard. For example, visual inspection of epicenters in the west-central United States suggests that they have been concentrated in the Colorado Lineament for the past century. Results of distortion evaluation support that suggestion: concentration is strongly indicated, although it cannot be evaluated rigorously.

You do not currently have access to this article.