The degree to which the result of a sample might not be representative of the total population.