Practicing science in a policy context demands careful attention. This research presents backcountry campsite conditions found in Shenandoah National Park during 1992. It fuels a policy process grounded in Limits-of-Acceptable-Change (LAC). This context provides an opportunity to critique natural resource science within an LAC process and to illustrate the applicability of techniques from disciplines as diverse as educational testing and policy sciences. The condition, location, and legality of 743 backcountry campsites are reported, as revealed during a census completed using methods from recreation ecology.

To assess the consistency and precision of campsite surveyors, techniques from educational measurement are adopted, including generalizability theory and decision studies. For each indicator, these techniques reveal the consistency of raters, their ability to distinguish between unique campsites, and whether raters are systematic. These three components were found to vary considerably for different indicators, reflecting the appropriateness of each indicator for justifying management decisions.

In addition to traditional description of survey results, campsite condition indicators of 220 campsites with bare soil are reintegrated using cluster analysis. Seven distinct types of campsites with bare soil exist in the park. This approach should facilitate policy makers’ understanding of findings, and improve investigators’ comprehension of survey data.

Recommendations include adopting a multiple strategy policy for managing backcountry campsites in Shenandoah National Park, continuing efforts to adapt generalizability theory to natural resource monitoring, and applying two phase research designs that combine the strengths of both census and survey in an efficient manner guided by generalizability theory.