To defend a sampling plan to a regulator concerned about safety and to a citizens' group concerned about saving taxpayer dollars requires balancing cost and risk. There must be non-biased, defensible means that sufficient samples are taken, in order to make a decision, estimate a proportion, or declare an area free of UXO with a stated level of confidence. Additionally, once samples are taken and the results processed, someone needs to apply a statistical test to actually make a decision based on the data or calculate a confidence interval. VSP incorporates all this into the code it uses to calculate a sample size and sample locations. It asks the user to enter the assumptions, acceptable risk, and costs it needs for these calculations.
VSP follows the EPA-sanctioned planning approach for data collection and decision-making called the Data Quality Objectives (DQO) process. The DQO process achieves the user's limit on acceptable risk, at a minimum cost. There are 7 steps in the DQO process. Users must complete Steps 1 through 6 in order to have the inputs VSP needs. Then, using VSP, the user can complete Step 7, Optimize the Design for Obtaining Data, because VSP can be used to try out different sampling designs and find the optimal design for the current problem.
Users familiar with the DQO process know that often a single site may have multiple sampling goals and multiple Sample Areas, each requiring its own set of DQO inputs and hence different sample requirements.
VSP allows rapid prototyping and has many features that allow the overlay of designs and comparisons across designs.