When constructing questions for surveys it is important that the answer categories cover all possibilities. Respondents should always be able to give a valid answer.
Below is a screengrab of the first question of a survey I was recently invited to complete:
As I imagine was the case for the vast majority of people sent the survey I hadn't bought any products from any of the brands listed from Boots in the last 12 months. But there wasn't an answer option for me to say that I hadn't bought any of them. I didn't press on with the survey but other respondents may have got confused and selected an answer that allowed them to continue. Completed surveys may have included a significant proportion with respondents not actually meeting the qualification criteria but forced to give an answer just to continue with the survey.
This is a fairly basic but not unusual mistake. To avoid it check all your questions and make sure your answer lists cover all possible answers and include "other", "none of the above" and "not sure" answers where appropriate. Always get other people to check your questionnaires wherever possible, particularly those who aren't closely involved with the survey. A fresh pair of eyes often notices mistakes that those involved in designing the survey become blind to.