A Simulation Study on Methods of Correcting for the Effects of Extreme Response Style

摘要: The impact of response styles such as extreme style (ERS) on trait estimation has long been a matter concern to researchers and practitioners. This simulation study investigated three methods that have proposed for the correction estimates ERS effects: (a) mixed Rasch models, (b) multidimensional item (c) regression residuals. were compared with respect their ability recovering true latent levels. Data generated according unidimensional model only one trait, two populations non-ERS, two-dimensional incorporating an dimension. data analyzed using same models well linear where estimate is regressed score resulting residual considered corrected estimate. Over all conditions, achieved best recovery, though difference ...