It's well-known that how you ask the question in a survey can drastically skew the response, and so we have to interpret these results based on the specific questions they asked.

We know from sale prices that people actually covet walkable areas, so much so that the accusation of "rich elitist" gets tossed at proponents of walkable cities. Those places are so much more expensive. So maybe people are thinking of "houses that I can afford" when they answer this survey? Or, they're answering it from the perspective of already needing a car, so a little extra driving is no big thing.

What would the results be if they asked things like, "Do you prefer neighborhoods where kids can safely play outdoors, or neighborhoods where there is too much traffic danger?" Or, if that's too biased, "where children can walk to school versus taking a bus or being driven?" Maybe break up the question, "Do you prefer to have stores located near where you live, or do you want them farther away?"

There are lots of different ways to ask, and the different results would be informative.

(Also, this survey relies on self-reported urban/rural distinctions, and those answers are wildly inaccurate, to say the least.)