Last week, Professor Ball emailed with us a pilot survey for an honors student's thesis, at the following link: http://www.surveymonkey.com/s/DFDB6C8.
This survey was very different from most surveys I've ever taken because all of the main questions were open ended. They all related to the reasons I do and do not drink bottled water, who I know that does and does not drink it, and who would or would not support me drinking bottled water.
Though this type of survey would get detailed information from the respondents, it is tough to code these kinds of responses. It is best to use open-ended questions to gather data to start off a research study because it gives you an idea of what your hypothesis should be and what types of opinions you will be dealing with. However, I hope the honors student is not planning on solely using this survey. I do not think it would probe accurate answers from the respondents which would complicate the validity and reliability of the survey.
It is also frustrating as the respondent, because it is difficult to come up with answers on the spot with no aid. As a respondent, I prefer to take surveys where they offer some answers, and provide a box to add additional information. That way, it gets me started and I will think of more things after I have some help.