Curbing curbstoning: Distributional methods to detect survey data fabrication by third-parties.

Area of Science:

• Social Sciences
• Survey Methodology
• Data Integrity

Background:

• Curbstoning, the deliberate fabrication of survey responses by data collectors, poses a significant threat to research validity.
• Existing survey auditing literature indicates curbstoning is prevalent, even among professional data collectors.

Purpose of the Study:

• To propose and evaluate simple survey questions designed to detect curbstoning.
• To provide researchers with a practical method for verifying data authenticity when outsourcing collection.

Main Methods:

• Developed survey questions with statistically predictable response distributions.
• Compared observed response distributions from authentic and fabricated surveys against expected distributions.
• Assessed the effectiveness of individual and combined methods in detecting fabricated data.

Main Results:

• The proposed methods demonstrated Type I error rates at or below the alpha level of .05.
• Individual methods correctly detected false responses between 48%-90% of the time for N ≥ 50.
• Combined methods achieved high statistical power with Type I errors < 1% and were effective even for smaller sample sizes (N = 30).

Conclusions:

• The proposed survey questions offer a simple, generalizable approach to detect curbstoning.
• Combining detection methods enhances statistical power and reliability in identifying fabricated survey data.
• These methods empower researchers to ensure data accuracy when direct supervision of data collection is not feasible.