To eliminate the influence of experimental error, analytical function is adopted to model the experimental profiles. Much attention has been paid to the analysis of the effects of such factors as Kα doublet, broadening and alignment of profiles etc. on the Fourier coefficients and deconvoluted profile obtained from the Stokes correction.The reliability of the separation of size-and strain-broadening from Fourier coefficients or deconvoluted profile is estimated. A method is proposed for the simultaneous refinement of accurate Bragg angle from the deconvolution procedure.