Errors in measuring an explanatory variable in a regression, which leads to biases in estimated parameters.