Search
Browse By Day
Browse By Time
Browse By Person
Browse By Area
Browse By Session Type
Search Tips
ASC Home
Sign In
Concerns about fraudulent survey data have existed as long as social scientists have been administering surveys. Although receiving less consistent attention in the criminological literature, methods for preventing and detecting fraudulent survey responses are critical for ensuring the integrity of survey-based research in criminology. We evaluate and extend Kuriakose and Robbins’ (2016) maximum percent match metric by introducing a framework that combines it with two information-theoretic measures: surprisal, which quantifies how unexpected each of a respondent’s answers is relative to the sample distribution for that item, and Shannon’s entropy, which captures the variability of an individual’s response sequence. By applying this framework to data simulated from real-world surveys thought to be free from fraud, we demonstrate that our approach substantially reduces false positive rates compared to methods that rely only on the proportion match metric. We further apply our framework to real-world criminological data and provide practical diagnostic tools implemented in R for incorporating these methods into routine data quality checks.