Skip to content

Statistical Sample Quotas Using Clustering Model

The Statistical Sample Quotas Using Clustering Model estimates the sample size needed for each sub-administrative area with no positive cases to assert with high probability that CWD prevalence in the population is at or below a threshold. The model explicitly accounts for increased disease transmission through the natural clustering behavior of hosts in the population and assumes a simple random sampling design.

Geographical Scale

  • Administrative area, subdivided into a sub-administrative areas

Required Data

  • Population size or population density of hosts in each sub-administrative area

User Inputs

  • Average cluster size of hosts
  • Correlation in disease status among hosts sharing a cluster
  • Sensitivity of the diagnostic test used to declare a CWD-positive case

Outputs

  • The number of randomly selected hosts that need to be tested in each sub-administrative area to have high probability (95%) that the prevalence of CWD in the overall population is at or below 0.5%, 1%, 1.5%, 2%, 3%, 4%, or 5%.

graph of sample size vs probability of freedom from disease at a given prevalence

Example of model output showing a sample size requirement.

More Information

For more information, go to the CWD Data Warehouse User Manual: Statistical Sample Quotas Using Clustering Model.

Code and Docker Image

To view the code once deployed, go to the GitHub Repository: Statistical Sample Quotas Using Clustering Model. To view the docker image, go to the Docker Hub: cwhl/statistical-sample-size-with-clustering-model.

Citations