Estimation methods for simple linear regression with measurement error: a real data application

Authors : E.rukiye DAĞALP, İhsan KARABULUT, Fikri ÖZTÜRK

Pages : 311-322

Doi:10.1501/Commua1_0000000821

View : 13 | Download : 9

Publication Date : 2017-08-01

Article Type : Research Paper

Abstract :The classical measurement error model is discussed in the context of parameter estimation of the simple linear regression. The attenuationeğect of measurement error on the parameter estimation is eliminated usingthe regression calibration and simulation extrapolation methods. The massdensity of pebbles population is investigated as a real data application. Themass and volume of a pebble are regarded an error-free and error-prone variables, respectively. The population mass density is considered to be the slopeparameter of the simple linear regression without intercept
Keywords : Classical measurement error model, consistent estimator, error invariables, linear regression, mass density, regression calibration, SIMEX method, M estimation