The georeferencing tool described here was used in a separate project, conducted by Elizabeth J. Proctor, to measure variation in the way people interpret locality descriptions as geospatial objects; i.e., shapes on maps. This project formed the basis of Elizabeth's master's thesis in Geography at San Francisco State University. She has generously given us permission to make her complete thesis available here as a PDF file.

Biological collections curated by scientists at natural history museums are irrefutable and long-lived evidence that organisms occurred at particular places in time. Mapping these collections can aid conservation efforts. The location data associated with historical specimens, however, are usually textual descriptions like “2 mi E of Garberville,” and are unusable for spatial analysis until translated into coordinates such as latitude and longitude. This task is called retrospective georeferencing, and since many current efforts involve some amount of human interpretation, there is interest in how place descriptions are interpreted by different people. This study measures the variation among shapes drawn by test subjects interpreting a set of eleven textual localities, and to what extent guidelines reduce that variation. The results indicate that these guidelines did not make a statistically significant difference between the two groups in shape location or size. One interpretation of this is that people tend to follow reasonably similar logic when interpreting localities, with or without instructions. The guidelines appeared to have the effect of refining shape sizes and placements in small but noticeable ways in certain cases, but had little effect on interpretation of the vaguest localities.