Development of a multi-factor social desirability bias scale

• Main Author: Tan, Houng Chien
• Format: Thesis
• Language: English
• Published: 2019
• Subjects: Social psychology; Social desirability
• Online Access: http://psasir.upm.edu.my/id/eprint/83286/1/FEP%202019%2044%20ir.pdf; http://psasir.upm.edu.my/id/eprint/83286/

Social desirability bias is one of the most common and critical survey distortions that may misleading the research findings and conclusion. Despite 60 years of research, there is still an open debate on its conceptualization and operationalization. Hence, the purpose of this study was to develop and validate a multi-dimensions scale namely, Multi-Factors Social Desirability Bias (MFSDB) scale that can be used to measure the degree of social desirability bias present in a survey. This study adopted DeVellis (2016)’s scale development guidelines and employed a mixed methodology to gather and analyses the data. In the first phase, qualitative inquiry was carried out to collection respondents’ descriptions about social desirability bias through personal interview. In particular, a total of 15 participants were interviewed to identify any potential dimensions for the construct. Six dimensions were identified in the interviews which were the Hubris State, Impression Management State, Secrecy State, Trust State, Adequacy State and Utilitarian State. The initial items pool was then generated according to the six dimensions based on the findings of the interviews and literature. In the second phase, quantitative inquiry was carried out to examine the reliability and validity of the items derived from the qualitative inquiry. The initial items pool of the MFSDB scale consisted of 100 items were tested against 688 working adults around Malaysia. The dimensions of the new measurement scale were identified through the exploratory and confirmatory factor analysis. Only 17-items grouped under six dimensions fulfilled all the criteria to form the MFSDB scale. The scale has good psychometric properties, consistently demonstrating construct validity in various tests of discriminance and convergence. Its predictability has also been demonstrated in a nomological framework with related constructs. And it meets the requirements of face validity. The MFSDB scale suggests a new composition of social desirability bias. This scale provides a more accurate measurement for researchers to identify social desirability bias.