A0459
Title: Willing and able to fake: A new and flexible item response model for applicant faking
Authors: Siwei Peng - Jiangxi Normal University (China) [presenting]
Yan Cai - Jiangxi Normal University (China)
Dongbo Tu - Jiangxi Normal University (China)
Abstract: Applicant faking (AF), the intentional distortion of responses on non-cognitive assessments to present oneself with a false self-image, poses a critical concern in human resource management by increasing the risk of selecting non-ideal candidates. A new psychometric model is introduced for faking behavior, termed AF-IRT, which translates existing faking theory into a quantitative model. The AF-IRT identifies faking behavior on the item-by-respondent level and decomposes the faking process into (a) applicant's willingness to fake and (b) their perceptions of desirable response options. It helps explore person and item characteristics associated with higher prevalence of faking behavior and examine which response categories are more desirable for a specific item. The simulation studies demonstrated that the proposed AF-IRT model exhibited reasonable parameter recovery. The AF-IRT exhibited better estimation accuracy and higher reliability than the model that ignores faking. Empirical findings further supported the practical advantages of the AF-IRT. To improve clarity and accessibility, a step-by-step tutorial has been included to assist novice or non-quantitative researchers in utilizing the AF-IRT to analyze their empirical data through the R language.
