Lv. Madden et G. Hughes, PLANT-DISEASE INCIDENCE - DISTRIBUTION, HETEROGENEITY, AND TEMPORAL ANALYSIS, Annual review of phytopathology, 33, 1995, pp. 529-564
The statistical properties of disease incidence are reviewed and used
to characterize spatial patterns of diseased entities (e.g. plants), s
atisfy assumptions of statistical analyses, and quantify change in inc
idence over time. Frequency of diseased plants can be represented by t
he binomial or, more commonly, the beta-binomial distribution. Spatial
patterns of disease can be described by the aggregation parameter of
the beta-binomial distribution, index of dispersion and related statis
tics, parameters of the binary form of the power law, as well as measu
res of spatial autocorrelation between sampling units. Disease inciden
ce over time can be represented by continuous- and discrete-time nonli
near disease progress models, such as the logistic, and by autoregress
ive, integrated, moving average models. Ultimately, simultaneous spati
o-temporal analyses can be performed to understand the dynamics of dis
ease incidence in populations.