Chapter 8 introduces methods for the evaluation of uncertainty associated with spatial analysis. Errors may be stochastic (random) or systematic (caused by a consistent shift away from an accepted or true value). The root mean square error (RMSE) measure is commonly used to quantify errors. Statistical methods for evaluating the uncertainty of analysis can be non-spatial (mean, standard deviation, standard error, probability modelling of confidence intervals, sensitivity analysis, error propagation assessment), or spatial (residual mapping, confidence interval mapping). Spatial error can be mapped and modelled using geographically weighted regression to increase the reliability of predictive models. A validation exercise can evaluate the uncertainty associated with digital maps by systematically comparing them to an independent ground referencing dataset by means of an error (or confusion) matrix. Comparisons yield metrics such as the producer, user and overall accuracy of the map.