The U of S Survey Tool currently has the following limitations:
Surveys must be completed in one sitting. There is no way for participants to save a survey mid-way through and then continue completing it at a later date.
There is no "import/export" capability for surveys. Surveys can be created using other existing surveys as templates. However, there is no way to export questions to an external file, and then later import these questions into this or other survey tools.
There is no way to backup and restore individual surveys. ICT completes full system database backups nightly for disaster recovery of the entire system. However, if a survey owner accidentally deletes all the survey data or questions in the survey, the information is permanently lost. ICT does not currently have the ability to recover data for just one survey.
Surveys are completely anonymous. Even when people log in to take the survey, their identifying information is not tied to the answers they provide. Survey developers can add questions asking participants to identify themselves. However, there is no way to verify that these questions are answered truthfully. For this reason, the survey system is not suitable for online examinations. Use a learning management system like WebCT instead.
The survey system can only import a list of survey participants when using the "invitation only" access control method. There is no import feature for the "Username and Password" survey type -- use the invitation method, as this generates unique invitations for each user which function as passwords for them.
Do not use the "Username and Password" Access Control method. It is broken, and only works intermittently.
The "invitation" method of specifying users is only practical for about 1,000 invitees. It is recommended that people use PAWS groups whenever possible to target large on-campus audiences.
The question numbering scheme cannot be customized. Surveys must use the numbering scheme provided by the system.
If the survey owner grants results viewing privileges to people, the system allows these people to view both the aggregated summary data and the full detailed results. There is no way to publish only the summary data. Additionally, if survey owners grant results viewing privileges it is done for all questions; it cannot be done on a per-question basis.
Survey participants cannot use the browser "back" or "refresh" buttons when using the survey. Using these buttons can create unexpected results. This is normal for most web applications.