# User guide

## Pooled prevalence for variable pool size and perfect tests

This method uses generalised linear modelling to calculate maximum-likelihood estimates of prevalence and confidence limits where multiple different pool sizes are used. The method assumes 100% test sensitivity and specificity. See Williams & Moffitt (2001) for more details. Prevalence estimates can be calculated for sampling strategies with up to 10 different pool sizes used.

Required inputs for this method are:

- pool size for each pool size used,
- number of pools tested for each pool size used,
- number of pools positive for each pool size used and
- desired level of confidence in the estimate.

The desired confidence level must be a decimal number >0 and <1 (for example, 0.99 = 99% or 0.95 = 95%). Pool sizes and numbers of pools must be positive integers (>0) and the numbers of positive pools must be non-negative integers (>=0). The number of positive pools must be less than or equal to the corresponding number of pools tested.

Outputs include:

- a point estimate of animal-level prevalence and
- upper and lower asymptotic confidence limits for the estimate for the specified level of confidence.