The difference between a good survey from a bad survey is that a good survey evokes the truth, it will ask for a one-dimensional answer, flows well with the previous question, does not imply a desired answer, and does not use vaguely defined words.
A bad survey has loaded questions. The questions are written in a way that forces the respondent into an answer that does not accurately reflect their true opinion. Double negatives or jargon where you go on and on with no true objective is also part of a bad survey, and leading questions that steer the respondent to answer in a way that is not necessarily the way they would originally answer are all an example of a bad survey.
To get someone to take a survey is difficult enough to begin with so making the survey flow better, with questions properly asked and allowing for accurate responses are all a great recipe for a good survey as well as a good response rate. You do not want to ask too many repetitive questions where the respondent will get tired of answering or make it too lengthy.
When I was doing drop shipping, I would occasionally create surveys for my current and past customers to pinpoint areas where I could improve but I noticed it was difficult to get people to take a survey without having to offer them an incentive to do so. I would typically offer a promo code to use on my website if they completed the survey. I tried to make the survey short with direct questions and all multiple-choice questions with one final question where they could type in anything they want to expand upon. I noticed that until I had the right type of questions on the survey, I would get a better response rate. Having a good survey will make the process easier for people to complete the survey and finish it. When a survey is too difficult, there are more chances where a respondent is not willing to participate.