VIEW ARTICLE

Techniques

Evaluation of Field Sampling Techniques for Estimation of Disease Incidence. B. R. Delp, Former graduate research fellow, Department of Plant Pathology, University of California, Davis 95616; L. J. Stowell(2), and J. J. Marois(3). (2)(3)Former postdoctoral researcher, and assistant professor, respectively, Department of Plant Pathology, University of California, Davis 95616. Phytopathology 76:1299-1305. Accepted for publication 20 May 1986. Copyright 1986 The American Phytopathological Society. DOI: 10.1094/Phyto-76-1299.

Disease incidence and disease aggregation were varied in computer-simulated field tests to determine their effects on sampling techniques. Fields with 0.01, 0.1, 1.0, and 10% disease incidence were simulated. Five degrees of disease aggregation were simulated at each disease incidence level except at 0.01%, for which only three were simulated. Ten fields were generated for each of the 18 field types. Each field was sampled with five sampling designs (left and right diagonal, left and right W, and stratified random) at three sample sizes (20, 30, and 40 plants per sample site) and seven sample intensities (0.05, 0.1, 0.2, 0.4, 1.1, 2.2, and 4.4% of the plants sampled from the entire field). There were no significant differences between the left and right diagonal sampling designs or the left and right W sampling designs. Sample size had no apparent effect unless disease was random at the lowest disease incidence, where increasing sample size from 20 to 40 plants per sample site increased percent error. This was a result of decreased number of sample sites and sample site dispersal. In all fields, percent error of the disease incidence estimates and standard deviation of percent error were lowest with the stratified random sampling design if sample intensity was ≥ 0.2% and highest with the diagonal if sample intensity was ≥ 0.4%. Percent error for all designs decreased as sample intensity increased from 0.05 to 0.2%. When sample intensity was ≥ 0.2%, the percent error for the diagonal and W designs achieved a minimum plateau; however, percent error for the stratified random sampling design continued to decrease as sample intensity increased if disease was aggregated. Percent error was inversely related to disease incidence and directly related to disease aggregation. The stratified random sampling design required the least number of samples and the lowest sample intensity to estimate disease incidence within a 95% confidence interval for all field types.