Short answer is yes. You can use statistical modeling techniques on an input dataset to generate numeric scores that estimate the degree to which someone is gullible.

This, of course, assumes that the input data has a strong enough signal to predict (or describe) gullibility, which may or may not be true depending on the data you have available.

Data scientists model personality characteristics like gullibility in domains such as message testing (advertising) and persuasion modeling (political elections).