Using the Beta-Binomial Distribution to Describe Aggregated Patterns of Disease Incidence

Area of Science:

• Statistics
• Plant Pathology
• Ecology

Background:

• Plant disease incidence data often exhibits aggregation, meaning diseased plants are clustered.
• The standard binomial distribution assumes independence, which may not accurately reflect aggregated disease patterns.
• The beta-binomial distribution offers a flexible alternative for modeling overdispersed count data.

Purpose of the Study:

• To evaluate the utility of the beta-binomial distribution for describing aggregated plant disease incidence data.
• To compare the goodness-of-fit of the beta-binomial distribution against the binomial distribution for real-world disease datasets.
• To explore the relationship between beta-binomial parameters and disease aggregation.

Main Methods:

• The beta-binomial distribution was applied to four published plant disease incidence datasets.
• Maximum likelihood estimation was used to determine the parameters of the beta-binomial distribution (p and θ).
• Goodness-of-fit tests were conducted to assess model performance compared to the binomial distribution.

Main Results:

• The beta-binomial distribution provided a good description of all four datasets.
• In all cases, the beta-binomial distribution yielded a significantly better fit than the binomial distribution.
• The estimated parameters (p and θ) quantified mean disease incidence and the degree of aggregation.

Conclusions:

• The beta-binomial distribution is a suitable and advantageous statistical tool for analyzing aggregated plant disease incidence data.
• This distribution accounts for the overdispersion commonly observed in such ecological and agricultural datasets.
• The findings support the use of the beta-binomial model for more accurate disease modeling and risk assessment.