Estimating Plant Disease by Sampling

Sampling provides the data that link epidemiological theory to the solution of problems involving plant disease in fields, forests, and elsewhere. Without sampling, there are no data, and without data, our theories can only provide us with qualitative pictures of the dynamics of plant disease in time and space. Such qualitative pictures are useful in providing a strategic overview, but the extent to which this can contribute to practical problem-solving is limited if it is not eventually accompanied by quantitative data. At the outset we need to specify whether we are sampling to estimate the value of a population parameter, such as mean disease incidence, or to classify a population relative to a threshold level of some epidemiological significance. This chapter is concerned with various aspects of sampling for estimation. Sampling for classification in the context of crop protection decision making has been advanced by recent research in economic entomology.

Chapter topics include:

Why We Sample for Epidemiological Data

Sampling Preliminaries

Simple Random Sampling for Disease Incidence Data

Simple Random Sampling for Count Data

Cluster Sampling for Disease Incidence Data

Regression Analysis of Disease Incidence Data

Regression Analysis of Count Data

Group Testing with Incidence Data

Binomial Sampling for Count Data

Estimation of Disease Severity

Inverse Sampling for Disease Incidence

Sequential Estimation of Disease

Conclusions