Mitigating budget constraints on visitation volume surveys: the case of U.S. National forests
Stratified random sampling (SRS) provides a scientifically based estimate of a population comprising mutually exclusive, homogenous subgroups. In the National Visitor Use Monitoring (NVUM) program, SRS is used to estimate recreation visitation and visitor characteristics across activities on National forests. However, with rising costs and declining budgets, carrying out an annually established SRS poses challenges in how and where to reduce sampling while maintaining statistical precision. Furthermore, any reductions must produce results that validly compare to previous years’ SRS in trend analyses. Accurate estimates are necessary to manage resources for current and future demand; we explore this need through simulations and describe a methodology which can be generalized to any SRS-based process facing budgetary challenges. In this research, recent historic NVUM responses serve as groundwork for simulating various reduction scenarios. The ultimate goal is a manageable set of strategies from which to systematically assign a suitable reduction scenario for each sampling unit (i.e., forest). Using baseline historic data, we determine the sampling error variability by reduction scenario and experiment with various sets of three or four pre-defined strategies to identify viable reduction candidates. A desirable candidate is one that achieves acceptable error bounds (based on historic values) and reductions overall while meeting specified budget constraints (e.g., 20 percent overall fewer sampling days).