The impact of faking on Cronbach’s alpha for dichotomous and ordered rating scores

摘要: In many psychological inventories (i.e., personnel selection surveys and diagnostic tests) the collected samples often include fraudulent records. This confronts researcher with crucial problem of biases yielded by usage standard statistical models. this paper we applied a recent probabilistic perturbation procedure, called sample generation replacement (SGR)—(Lombardi Pastore, Multivar. Behav. Res 47:519–546, 2012), to study sensitivity Cronbach’s alpha index fake perturbations in dichotomous ordered data, respectively. We used SGR perform two distinct simulation studies involving size conditions, three item set sizes, twenty levels faking perturbations. Moreover, second also evaluated an additional factor, type model, reliability under different modulations graded (uniform faking, average slight extreme faking). To simulate these more complex models proposed novel extension procedure based on discrete version generalized beta density distribution. new real behavioral data emotional instability.