Errors-in-variables model

In <a href="/facts/Statistics/DNPsKGYU">statistics</a>, an errors-in-variables model or a measurement error model is a <a href="/facts/Regression_model/n6z5Tf7K">regression model</a> that accounts for <a href="/facts/Measurement_errors/q7pa7bRt">measurement errors</a> in the <a href="/facts/Independent_variables/dINfzIF2">independent variables</a>. In contrast, standard regression models assume that those regressors have been measured exactly, or observed without error; as such, those models account only for errors in the <a href="/facts/Dependent_variables/dINfzIF2">dependent variables</a>, or responses.
In the case when some regressors have been measured with errors, estimation based on the standard assumption leads to <a href="/facts/Consistent_estimator/HAKDWYEM">inconsistent</a> estimates, meaning that the parameter estimates do not tend to the true values even in very large samples. For <a href="/facts/Simple_linear_regression/SNINCEFU">simple linear regression</a> the effect is an underestimate of the coefficient, known as the <a href="/facts/Attenuation_bias/yshVFGzE">attenuation bias</a>. In <a href="/facts/Nonlinear_modelling/Ivy3lBfL">non-linear models</a> the direction of the bias is likely to be more complicated.

Errors-in-variables model open-in-new

Errors-in-variables model