🧫 Sampling Bees for Viral Surveillance

About this simulation

Detecting viruses like Deformed Wing Virus (DWV) or Chronic Bee Paralysis Virus (CBPV) in a colony, or across an apiary, depends entirely on how the sample of bees is collected — not just how many bees are tested. This apiary shows six hives, each hosting a hidden population of bees where some fraction is truly infected (the slider you set). Drawing a sample and testing it produces an estimate of that true prevalence, which will always carry some sampling error.

How it works

• Sample size — more bees tested shrinks the margin of error, following the classic 1/√n relationship of a binomial proportion.
• True prevalence — the hidden ground-truth infection rate across the apiary; each hive's own rate is drawn around it, since infection naturally clusters within colonies.
• Collection method — individual foragers sampled apiary-wide behaves like simple random sampling (design effect ≈ 1). A single-hive grab sample draws every bee from one colony, so it inherits that colony's own bias — a large design effect that inflates the true margin of error even with many bees. A stratified pooled sample spreads collection evenly across every hive, which is more efficient than simple random sampling (design effect < 1).
• The effective sample size divides your raw sample size by the design effect — it's the "equivalent" simple-random sample size that would give the same statistical precision.

Controls

• Sample size — number of bees collected and tested (5–150).
• True viral prevalence — the hidden ground-truth infection rate you are trying to estimate.
• Collection method — how bees are drawn from the apiary's hives.
• ↺ Draw new sample — re-randomizes which bees are collected. Drag to rotate, scroll to zoom.

Did you know?

A grab sample from a single hive can look statistically "solid" — a big number of bees, a tight textbook confidence interval — while still being badly wrong about the apiary as a whole, because it never captures the colony-to-colony variation. Reliable surveillance protocols specify not just a minimum sample size but a spatial sampling design.