Introduction to Online Survey Studies
This system includes a rather extensive online survey feature (if enabled on your system). It allows you to set up an online survey as a study within the system, so participants who sign up for the study will be immediately asked to complete the survey. Upon completion of the survey, they will be granted credit automatically by the system. You may then analyze their survey responses on an individual basis, or download the raw data across all participants who completed the survey for further analysis. There is a chance that you may notice a discrepancy in the number of responses when analyzing a single question compared to downloading the entire set of responses. This can occur if a participant is currently taking the survey, but has not completed it. Their data is included in the single-question analysis (when available) but not in the full download of responses, as there is not a full set of data for an in-progress survey. The system will prevent participants from completing the survey more than once, removing the risk of duplicate entries for participants.
In compliance with research ethics guidelines, the participant is given the opportunity to withdraw from the survey at any time. See Participant Withdrawal from a Survey for more information.
There is space to provide closing text for the survey, which is displayed to participants after they successfully complete the survey. This means after they have saved all their responses and cannot go back and edit. This is an ideal place to include any relevant debriefing information.
The online survey feature is complex due to the many features it contains. In addition, your ability to modify the survey after participants have started to take the survey is limited. Because of this, you should plan out your survey well in advance to make sure it is finalized before you make it available to participants. You may find it helpful to plan out the survey on paper before entering it into the system.
A survey may have an unlimited number of sections and an unlimited number of questions per section. We recommend limiting surveys to fewer than 300 questions total, 15 sections total, and 15 questions per section. This is to reduce participant fatigue and to ensure proper system performance. Participant fatigue can be an issue, as they must complete the entire survey all at once. In addition, you should make participants aware that their session may time out because of inactivity. Usually, this means they will be logged out after 15 minutes of inactivity. Inactivity would happen if they load up a section of questions but do not complete and save their responses for that section within that period of time. It is a good idea to structure sections accordingly, especially if a section contains lengthy text or complex questions. The system will warn participants a few minutes prior to this inactivity logout, and provide them with the opportunity to extend their session.
Questions may be free-entry (requiring the participant to type in an answer) or multiple-choice (pick only one or pick many from the list of choices). You may also specify that the system compute participant’s results for an entire section, as either a section sum or average score. This computed sum or average can be computed only for numeric, multiple-choice (pick only one) questions in the section. Such computing is often useful when a participant’s aggregate score is more important, such as with a depression battery.
For free-entry questions where the participant is asked to enter a free-form text answer, the size of the entry field cannot be resized in terms of dimensions. It can only be resized in length. Participants may enter up to 255 characters in their response.
Sections can be displayed in any order, whether that is specified, random or combination of both. Each section may contain introductory text introducing the section. The survey itself may also have introductory text introducing the survey, as well as closing text that is displayed upon completion of the survey. You may specify that questions within a section be displayed in a random order, or in the order you specify. Multiple-choice questions can have their choices displayed in the entered order or random order. This is specified on a per-question basis. Multiple-choice question choices can be displayed horizontally (across the page) or vertically (down the page) or dropdown box.
It is important to note that while the survey feature is quite extensive, there are some things it does not support. It cannot do question timing, where the speed of a participant’s response (typically in milliseconds) to each question needs to be measured. It also cannot do branching, where the questions being presented to a participant will vary depending on their answer to other questions within that same survey. For example, asking a question like “Do you smoke cigarettes?” and then presenting a different set of questions based on their response. The reason that advanced features like this are not currently supported is that our software is primarily a participant pool software product that happens to have an online survey feature within it. We asked our customers which features in an online survey were most important to them, and those features exist within the product. There are numerous products on the market specifically for doing online surveys (with no participant pool component), which contain more advanced features. These products can easily be linked from the Sona Systems product by setting up your study as an online study conducted outside the system.