Does Internet-Based Survey Have More Stable and Unbiased Results than Paper-and-Pencil Survey?

Abstract

This study designed to compare responses from an internet-based survey to those from a paper-and-pencil survey in terms of measurement reliability, validity, and equivalence using homogeneous demographic profiles of the undergraduates studying in Taiwan. Several similarities and differences were found between two types of survey in this study. For examples, contents of the survey items (i.e., internet-related vs. behavior-related) and survey environments significantly influence the distribution of responses. The normal distribution, internal consistency, in addition to construct, and convergent validity for individual construct are quite similar. However, the homological validity evidence was demonstrated through structural equation modeling across two survey modes. Implications and future research are also discussed.