A Study of Statistical Efficiency on the Effects of Non-Complaint Reporting from the Indirect Questioning Techniques: Random Response and Non-Random Response Models

• Activity Number: 417
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Sports
• Author(s): Jay Schaffer*+ and Caroline Emsermann
• Companies: University of Northern Colorado and University of Northern Colorado
• Keywords: Indirect Questioning ; Random Response Models ; Non-random Response Models ; Non-complaint

Abstract

Estimating rates of sensitive behaviors using self-report result in biased estimates when direct questioning techniques are utilized. Random Response (RR) and Non-random Response (NRR) models were developed to improve honest or compliant reporting, but estimates from these techniques are often biased. A large simulation study that included a wide range of sensitive behavior rates and sample sizes was performed to determine if the NRR techniques, the item count technique (ICT), double item count technique (DICT) and the single sample count technique (SSC) were more efficient in the presence of non-compliance, as measured by their mean squared error (MSE) compared to the MSE of the RR techniques, the unrelated question technique (UQT) and forced-choice technique (FCT). Results of the study indicated the UQT optimal model is the most efficient in the presence of equivalent non-compliance rates. If the DICT optimal 5-item model can improve compliance, the model is more efficient for moderate and large sensitive behavior rates. As a result, the study demonstrated that in certain situations, the NRR DICT optimal model is at least as efficient compared to the RR UQT optimal model.