correct, and that's why surveys need to be as scientific as possible in order to avoid various biases that skew the results. hence qualtrics and not surveymonkey for serious survey research.
I don't know, but the key is to properly account for the biases in responses of those who do answer. Substantial percentages of people do answer surveys. Don't make the mistake of thinking along the lines of "i would never answer one, therefore nobody would"