How can I prevent fraud/bots filling in my survey?
We have several methods we use internally and hope to publish them for general use at a later date.
For now look for
- repeated IPs
- patterned responses
- incident rates getting better over time
- nonsense text in open text fields
- known distributions not matching data e.g. 50% gender splits
These evaluations can be performed post collection and respondents removed from the dataset.
To prevent users from reentering the survey, a cookie script can be implemented