Information Systems Research, 2010, Volume 21,
Issue 4, Page 941-959.
Research on software piracy often relies on self-reports by individual users and thus suffers from possible response distortion attributable to a variety of human motivations. Conclusions drawn directly from distorted self-reports may misguide managerial and policy decisions. The randomized response technique (RRT) was proposed as a remedy to response distortion. In this paper, a model based on RRT was used to illustrate how truthful responses to sensitive questions can be empirically estimated. The model was tested in two empirical studies on software piracy. Consistent with our expectations, respondents responding to RRT were more willing to disclose sensitive information about their attitudes, intentions, and behaviors on software piracy. Nontrivial distortions were demonstrated in causal relationships involving sensitive and nonsensitive variables. The study extends RRT to multivariate analysis and illustrates the feasibility and usefulness of the method in studying sensitive behavioral issues in the information systems (IS) domain.
Keywords: method of moments; randomized response technique; response distortion; socially desirable responding; software piracy; structural equation modeling; unrelated question design