Development of a program to control response biases and assessment of its usefulness in typical performance measures

Identificador:  TDX:2896

Handle: https://hdl.handle.net/20.500.11797/TDX2896

Autors:  Navarro González, David

Resum:
One of the main problems when administrating self-reports questionnaires is to what extent the participant’s response reflects accurately the latent trait of interest, or may be reflecting some systematic effects. This effects are known as response biases, being the most known ones the Social Desirability (SD) and Acquiescence (AC). This response biases produce different alterations in the structure of a questionnaire, difficulty the result interpretation and validity. Some methods for controlling the impact of SD and AC have been proposed, but only the proposed by Ferrando, Lorenzo-Seva, & Chico (2009) allows performing a factor analysis controlling the impact of both biases simultaneously. However, this method was not available in any statistical software. Hence, the first objective of this thesis is develop a software allowing the applied researchers using the aforementioned method. The second objective is applying the method in two different issues related with SD and AC. The first iisue is the impact on the factor structure of two questionnaires, and the second one, the role of response biases in the effects associated with the Personality Differentaition Hypothesis across ability levels (PDH). We developed the program Psychological Test Toolbox, distributed freely, using MATLAB. So, the first objective was accomplished, providing the applied researchers a tool to apply the method in the development of questionnaires. In regards with the second objective, we demonstrated that both response biases have a big impact on the factor structure of self-reports, especially the AC. Also, we discarded the relationship between response biases and the PDH effects, since there is no relationship between SD and ability, and the relationship between AC and ability being weak.