For the polynomial regression model in which the explanatory variable contains measurement error, the correction methods for attenuation estimators of the parameters are considered. The corrected least squares method, an extension of the corrected score function method and a modification of regression calibration method are derived to calculate the corrected estimators of the model parameters. The adjusted data and simulation-extrapolation methods are also presented to compute the corrected estimators of the parameters. A comparative study between the corrected estimators, of the parameters of the quadratic regression model, for these correction methods is performed using real data in medical application and also using a simulation study. The results reveal the superiority of the corrected estimators of adjusted data and modified regression calibration methods over the corrected estimators of the other methods |
