> I would imagine this gets to the larger issue that the majority of people
> taking the survey haven't read these books.

I'm guessing few people have. But then again, the news is full of detailed
statistics on how peoples' opinions on complex economic, policy, scientific,
etc issues that hardly anyone has any expertise on. I always enjoy hearing
what needs to be done in countries that few people could point out on a map,
let alone say anything intelligent about what is there.

If the concern is pollution of the results by uninformed votes, tossing in a
few fake works here and there and then not counting any results from people
who indicate a preference on such comparisons would help.

I suspect it would also reduce the sample size by at least 90% ;-)