The L.A. Times notes "For comparison purposes, the survey also asked some questions about general knowledge, which yielded the scariest finding: 4% of Americans believe that Stephen King, not Herman Melville, wrote "Moby Dick."

This is a good example of how survey results can be misstated. The author is confusing ignorance about a thing with belief in an alternative.

I doubt anyone actually believes Stephen King wrote "Moby Dick". All that can be stated with certainty is that when given the question "Who wrote 'Moby Dick'?" and possible answers of A) Herman Mellville or B) Stephen King, 4% of respondents didn't know the correct answer and guessed Stephen King incorrectly. You can safely assume another 4% also didn't know the correct answer but happened to guess correctly, which means at least 8% of Americans don't know Herman Melville wrote "Moby Dick".

A wrong answer on a multiple choice exam doesn't prove that the test taker believes his wrong answer is correct, only that he doesn't know the right answer and has guessed incorrectly. For that matter, a right answer doesn't always prove that the test taker actually knows his answer to be correct. He could have just guessed correctly.