11-20 seems like a pretty uninformative range, assuming you want to draw conclusions (maybe you don't!) from the results, as it is in this range that you experience the most noticeable increase in both maturity and knowledge.
11-15 and 16-20 would have been significantly more telling. (As an aside, this is especially strange considering ranges such as 41-45.)
Message to the future: there were ~30 people who had voted 11-20 by the time the switch was made to break it up into two sections.
To figure out the actual distribution of people, you'll have to subtract 30 from the 16-20 range and then compare that with the 11-15 range, and then add back in the 30 votes distributed correctly.
I had looked at the poll just seconds after the switch had been made; There were only 11 people who had been transferred to the 16-20 range. (So the distribution isn't going to vary as significantly.)
11-15 and 16-20 would have been significantly more telling. (As an aside, this is especially strange considering ranges such as 41-45.)