Jump to Main Content
Algorithms for computing operating characteristic and average sample number functions for sequential sampling plans based on binomial count models and revised plans for European red mite (Acari: Tetranychidae)
- Nyrop, J.P., Binns, M.R.
- Journal of economic entomology 1992 v.85 no.4 pp. 1253-1273
- Malus domestica, Panonychus ulmi, simulation models, algorithms
- Algorithms and computer code (Fortran 77) were developed for computing operating characteristic (OC) and average sample number (ASN) functions for sequential classification sampling plans based on presence-absence counts. Two models that relate the proportion of sample units with more than T (tally threshold) organisms (p(T)) to mean density (m) were considered; an empirical model of the form ln(-ln(1-p(T))) = gamma + delta ln(m) where m is the mean density, and the negative binomial distribution (NBD). The algorithms compute OC and ASN for nominal model parameters and for model parameters used to characterize the effect of p(t) -- m model variability on the OC and ASN functions. Changes in the OC and ASN functions as a result of change in the p(T) -- m model are a measure of the robustness of the sampling procedure. Both p(T) -- m models provided good descriptions of counts of European red mite (Panonychus ulmi (Koch), on apple trees, Malus domestica (Borkh). The robustness of sampling plans that used T = 0 was poor. Expected OC and ASN functions for sampling plans based on the two p(T) -- m models were similar. Robustness of sampling plans based on NBD improved significantly with larger values of T. When sampling using thresholds of 2.5, 5.0, and 7.5 mites per leaf, recommended values of T are 4, 6, and 6, respectively.