The problem with surveys is that their results can be  

for instance.   You can say you conducted a survey of the entire school, and of those who responded, 88% said A and not B.

Then, when you dig deeper, you might find only just a small handful actually responded, so trying to sell that 88% are truly representative that A was the proper response is misleading, particularly when only the handful of folks who bother to take the survey are the ones who want A in the first place. 

So, it might be 88% of those who responded, but definately not 88% of the entire school, who don't care at all!

High school statistics