By Malcolm S. Johnson, Vanessa M. Adams, and Jason Byrne.
When running an online survey, you notice an influx of hundreds of suspicious responses occurring around 1:00 am. At a closer inspection, you notice several with identical answers, many with nonsensical responses, and a few that don’t immediately look off. In the same survey, you get responses finishing in an instant and dozens of surveys with answers that don’t make sense. Despite your best efforts to prevent it, you are experiencing a typical amount of online survey fraud.
The use of the Internet to measure human dimensions of socio-ecological, climate change, and environmental justice research has become increasingly common in the last decade. However, the threat of fraudulent responses in online surveying is simultaneously expanding. While there are a range of tools to deter, prevent, detect, and remove fraud from research, they are neither widely used nor commonly reported. Our paper highlights the threat of fradulent responses, or fraudsters, to online surveys and offers potential approaches to address fraud.
During an online participatory mapping study in Tasmania, we encountered fraudulent responses at an unprecedented level. While there is significant disagreement as to the prevalence of fraud in surveys, to protect our data integrity, we had to thoroughly investigate that every response was valid. Our study demonstrates both a first-hand example of how to identify fraud even when you are not expecting it and a possible step-by-step approach to guaranteeing response validity.
In our paper, we begin by defining and describing fraudulent behavior and its relative absence in socio-ecological and environmental disciplines. We then introduce a range of existing tools to address fraud and describe their benefits and limitations. Lastly, by demonstrating our personal experience addressing fraud, we aim to shed light on the potential threat of online survey fraud more broadly as well as encourage researchers to investigate their data extensively and report the extent of fraud in their findings to help highlight the issue across disciplines.
Relational Thinking 