Coefficientsvarying in value that are modeled as the variance in outcomes due tounexplained differences between higher-level sampling units, between groups orclasses. These may include deviations in intercepts or slopes: differencesbetween the overall average outcome and the group average; or differencesbetween the average slope and the group slope. The deviations by group are saidto be random because they are due to the particular sample obtained (and may bedifferent for other samples) and because they are independent of anydifferences in outcomes due to measured variables.
See Crossed random effects, Fixed effects