The article is good because it points out the difficulty in designing bias-free survey questions, which is practically impossible. People think about what the question may be seeking to prove, and they will game the system. I do it, or I just refuse to participate.
Many of the surveys end up asking questions designed to prove the person who designed the survey's point. They don't really want your opinion.
I had to do some surveys while in grad school, and it is a difficult undertaking to try and find people willing to answer honestly.
So when you read the results of the surveys, take them with a grain of salt. They might be well-designed and are statistically significant, but many are not.
Take that question they first mention, "do regulations on housing increase housing price?" Of course they do. But when someone asks that question, it is difficult to just answer "yes," because you are thinking about the implications of the question. Do you want no restrictions of where, how dense, and the standards of construction? Why would someone ask such a dumb question? How do I counteract the stupidness of this question? Do I answer yes, no, refuse to answer in protest, or what?
And I completely disagree with the idea that college education does not matter. I think the more educated will think more about what is behind the question, and answer, not because they don't know that regulations increase the cost of building a house, but due to one of the other considerations listed above.
There are whole college classes for instructors in how to design test questions, which is not an easy task. Same with survey questions. It is hard to accurately gauge someone's knowledge, or even their opinions in a test or survey. They are useful, but shouldn't be relied on as gospel.