What are two effects geo-referenced data are subject to?

• A. Temporal autocorrelation and seasonality
• B. Spatial dependence and spatial heterogeneity
• C. Measurement error and misclassification
• D. Random sampling bias

Answer: (B)

Geo-referenced data have spatial structure that affects analyses. The two effects to recognize are spatial dependence, also known as spatial autocorrelation, where values at nearby locations tend to be more similar than those farther apart; and spatial heterogeneity, which means the processes driving the data vary across space, so relationships seen in one area may not hold in another. These traits influence model choice and inference because ignoring them can inflate type I error, produce biased estimates, or miss local patterns. Temporal autocorrelation and seasonality relate to time rather than space; measurement error and misclassification are about data quality; random sampling bias concerns how data are collected. The combination of dependence and heterogeneity is specific to spatial, geo-referenced data.