JMIR Publications

ABSTRACT

Background:

Online research studies enable engagement with more Black cisgender women in health-related research. Fraudulent data collection responses in online studies raises important concerns about data integrity, particularly when incentives are involved.

Objective:

The purpose of this study was to assess the strengths and limitations of fraud deterrent and detection procedures implemented in an incentivized, cross-sectional, online study about HIV prevention and sexual health with Black, cisgender women living in Texas.

Methods:

Data for the current study came from a cross-sectional web-based survey that examined factors associated with potential pre-exposure prophylaxis (PrEP) use among a convenience sample of adult, Black cisgender women from three metropolitan areas in Texas. Each eligibility screener and associated survey entry was evaluated by 4 fraud deterrent features and 7 fraud detection benchmarks with corresponding decision rules.

Results:

A total of 5862 responses provided consent and initiated the eligibility screener, of which 2,150 entries were ineligible for not meeting the inclusion criteria and 131 completed less than 80% of survey (n=131) and were removed from further consideration. Other entries were removed for not passing level 1 fraud deterrent safeguards: duplicate entry with same IP address (n=388); same phone number (n=69); same email address (n=114); same phone number and email address (n=17). An additional 1,652 entries were removed for not passing the first two items of the fraud detection benchmarks: screeners/surveys with a latitude and longitude outside of the U.S (n=347) and survey completion time of less than 10 minutes (n=1305). After the 5 remaining fraud detection benchmarks were implemented (i.e., phone number in incorrect format (n=121), invalid email address (n=113), nonsensical / improbable survey response(s) (n=470), zip code did not match city through self-report (n=185), name/email/open text responses contained non-standard characters or symbols (n=362)), an additional 1,186 entries were removed. The final enrolled sample size in the online study included 155 respondents who provided consent, were deemed eligible, and passed all fraud deterrent features and fraud detection benchmarks. In this paper, we discuss lessons learned and provide recommendations for leveraging available features in survey software programs to help deter ‘bots’ and enhance fraud detection procedures beyond relying on survey software options.

Conclusions:

Effectively identifying fraudulent responses in web-based surveys is an ongoing challenge. The data validation approach used in the current study establishes a robust protocol for identifying genuine participants, thereby contributing to the removal of false data from study findings. By sharing experiences and implementing thorough fraud deterrent and detection protocols, researchers can maintain data validity and contribute to best practices in web-based research.