Modeling Evasive Response Bias in Randomized Response: Cheater Detection Versus Self-protective No-Saying

Area of Science:

• Social Sciences
• Statistics
• Survey Methodology

Background:

• Randomized response techniques (RRT) aim to reduce evasive response bias in surveys on sensitive topics.
• Existing models, like the cheater detection and self-protective no sayers models, partially address this bias but have limitations.
• A hybrid
• ever/last year
• design offers potential for improved data collection.

Purpose of the Study:

• To develop and present statistical models for the hybrid
• ever/last year
• RRT design.
• To account for both self-protective no-saying and cheating behaviors within this design.
• To illustrate the application of these models using doping use surveys.

Main Methods:

• The study establishes the correspondence between existing RRT bias models.
• New models are introduced for the
• ever/last year
• design, incorporating parameters for response bias.
• Extensions of the design with additional questions and degrees of freedom are modeled.

Main Results:

• The proposed models for the
• ever/last year
• design allow for the inclusion of response bias parameters.
• Models with multiple degrees of freedom are developed for more complex survey designs.
• The models were successfully illustrated using real-world doping use survey data.

Conclusions:

• The
• ever/last year
• RRT design, with the proposed models, offers a promising approach for sensitive data collection.
• These models enhance the ability to detect and correct for response biases.
• Further research is warranted to explore the full potential of this design and its extensions.