Improving the efficiency of surveys with randomized response models: A sequential approach based on curtailed...

Area of Science:

• Statistics
• Survey Methodology
• Psychological Measurement

Background:

• Randomized response models (RRMs) enhance data validity for sensitive attributes by ensuring respondent anonymity.
• The randomization inherent in RRMs increases sampling variance, necessitating larger sample sizes.
• This poses a challenge for the practical application of RRMs.

Purpose of the Study:

• To address the increased sample size requirements of RRMs.
• To integrate RRMs with curtailed sampling techniques.
• To enhance the feasibility and resource efficiency of sensitive attribute measurement.

Main Methods:

• Combined Randomized Response Models (RRMs) with a curtailed sampling design.
• Implemented a sequential sampling approach that terminates data collection upon reaching sufficient evidence for hypothesis testing.
• Defined a maximum sample size within the curtailed sampling plan for straightforward prevalence estimation.

Main Results:

• The integration of RRMs with curtailed sampling effectively reduces sample size requirements.
• This combined approach offers a more resource-efficient method for collecting sensitive data.
• Prevalence estimation is simplified due to the defined maximum sample size.

Conclusions:

• Combining RRMs with curtailed sampling offers a practical solution to mitigate increased sample size needs.
• This methodology enhances the feasibility of using RRMs for sensitive attribute measurement.
• An R Shiny web application is available to facilitate the application of these procedures.