More than that, a poll about fears is likely to have a pretty high false positive rate - just considering the question is likely to bring up a significant number of instances of anything you fear at all, and if it is phrased as generically as "often" with no definition? Getting at the true numbers would require.. Uhm. No, asking people to monitor their fears would be Nigh-certain to make them much more fearful ("log thoughts of sex" has been tried. The results that came back were blatantly a case of "dont think of a pink elephant" coloring everything) and thus would be deeply unethical. I think the cleanest lift would be a large collection of extensive daily journals, or outright annotated lifelogs. That would probably make your subject pool more introspective than the general population, but it should not skew these specific numbers much. Expensive, however.