Stopping on the Count, Not the Clock

Allocation and Censoring Bias in Quadrat Surveys Worked Ahead of a Rising Tide

A paper from the School of Continuous Improvement treats a stratified quadrat survey of a rocky shore as a sampling design with a different deadline in each stratum, set by the flood tide. It shows that the variance-optimal allocation is the greatest convex minorant of the cumulative deadlines, that a stratum worked until the water arrives returns a mean biased downward by the quadrat left unfinished, and, in 10,000 simulated surveys at each of seven tidal windows, that a count-stopped rule cuts relative bias from as much as 2.6 per cent to 0.4 per cent while carrying the largest root mean squared error of the four designs compared, repaying its cost once about nine tides are averaged.

@misc{slop_i9mrf6,
  author       = {Mirela Hanke and Joost Nwosu and Solveig Adeyemi},
  title        = {Stopping on the Count, Not the Clock: Allocation and Censoring Bias in Quadrat Surveys Worked Ahead of a Rising Tide},
  year         = {2026},
  publisher    = {Slop University},
  doi          = {10.5555/slop.i9mrf6},
  url          = {https://slop.university/outputs/slop-paper-a-stopping-rule-i9mrf6/},
  version      = {1.0},
  note         = {Research paper},
}