9% might not be significant but one thing that the survey needs to mention before it comes to the conclusion is what pool of people the guy is using. For instance, education level, age, gender and I can imagine most bitcoiners are somehow younger male who are probably internet geeks with some knowledge in programming, so if it is out of these pool of people, the conclusion made from the survey might not be that accurate.
That's true. Which demographic did this particular survey target is a good question to ask.
Say they polled a bunch of com sci guys in MIT or Harvard or Stanford, and I'm sure the results would be drastically different.
That said, if it was a broad survey with several age ranges, sexes, races, and walks of life and we came up with more people still not "getting it", we all need to collectively do a better job.
At the end of the day, this is good news. It means back to drawing board, and get some real traction with the masses with a true and real launch and campaign going...