Seeing as it's a survey about D&D done by a company attenpting at the time to redesign D&D, I think it's fair to allow them to assume a D&D-like structure.
Or data manipulation.
For example, all survey responses from people over a certain age (35, I think) were tossed*; meaning most of those who had started in the early days had no voice in the data.
This makes the results and conclusions drawn from the survey, in my view anyway, highly suspect.
* - this is noted in Dancey's report, a copy of which @Morrus
booting around somewhere in here
. EDIT stored under the "Features" dropdown at the top of this or any page, it's the second option.