This study examines and compares four person-fit statistics (PFSs) in the framework of the "D"- scoring method (DSM): (a) van der Flier's "U3" statistic; (b) "Ud" statistic, as a modification of "U3" under the DSM; (c) "Zd" statistic, as a modification of the "Z3 (l[subscript z])" statistic under the DSM; and (d) the nonparametric statistic "H[superscript T]." The data were obtained via DSM-based simulations under a 2x3x3x3x3 design by five conditions: "type response behavior" (guessing and cheating), "sample size, test length, percent misfitting items," and "percent misfitting persons." The "Ud" and "Zd" statistics outperformed "U3" and "H[superscript T]" and demonstrated excellent performance on detecting person misfit. Recommendations for the practice are provided