Dishonesty requires intent… So the best way to identify or predict it is to determine:
- whether or not a particular profile item (height, salary, eye color) is inaccurate, and
- whether this inaccuracy can be proven intentional on the part of the profile's author, or simply the result of miscommunication, or faulty perception on the part of the author.
Proving either or both points, on a traditional online dating site, would be incredibly difficult for a large number of parameters, without some form of dedicated profile verification or background checking service.
That said, if one is clever, it's possible to imagine a variety of self-correcting systems that could be created to ensure that profiles are accurate within a degree of variance, or at the very least generated by a well-meaning, if sometimes self-misreporting, human being.
I wish I could be more helpful in terms of actual methodologies, but we're currently looking at patents for these very processes over at navidate.com, so…
How can dishonesty be identified or predicted based on online profiles?
deano.com 