Presenter: Dr. Sarah Hession
Standard regression approaches, such as ordinary least squares (OLS), rely on an assumption of independence over space. However, many phenomena are correlated over space, as given by Tobler’s First Law of Geography. When data are spatially autocorrelated, underlying model assumptions may be violated, yielding biased results and weakened inferences about the phenomenon of interest.
Limitations in aspatial regression techniques will be discussed, and methods in spatial regression analysis presented to assess the role of predictive variables while explicitly incorporating spatial autocorrelation in parameter estimation and hypothesis testing. Important considerations such as choosing an optimal neighbor matrix, model selection using information theoretic criteria (e.g., Akaike Information Criterion [AIC]), and diagnostic testing will be discussed.
Examples and exercises will be completed in R and RStudio. This workshop does require a basic understanding of statistics at the level of a one semester statistical methods course. Familiarity or prior exposure to the R programming language and RStudio is also helpful.