Best practices in questionnaire design by Godfrey Parkin
This is a good check-list of creating good questionnaires.
I would just add two comments about the mechanics of creating and delivering a questionnaire.
1. Anonymity and confidentiality need not be mutually exclusive. In the case of a value-laden questionnaires, such as promotion-related teaching evaluation, we need to ensure no tampering of survey data both by students AND lecturers involved. We must ensure that only students involved can submit and only submit once in order to avoid biasing. We also need to ensure anonymity to the students through an open process which anonymity is provided by the process without relying on any trust. (I have implemented such a system, but details have to remain with me. If you require such a system, please contact me for references and demonstration.)
2. Web-based data collection seems to be most economical way of collection survey/questionnaire data. One of the greatest problem of this is the tedious task of creating the questionnaire, especially if you are a frequent user of survey. Choose a survey data collection provider who can provide you with re-use ability of your survey.