Online Survey Creation Tool

IRT's Online Survey Creation Tool supplies the Medical School community with a way to rapidly deploy HIPAA compliant online surveys. A survey created using this tool consists of a single web-form page with a sequence of questions and a "submit" button at the bottom. On submission of the form the user is given the option of navigating back to the form to review their responses, correct any errors and resubmit if necessary. The sample Bicycle Use on Campus survey provides a demonstration of the basic features of an online survey created with this tool.

All surveys created using this tool are automatically HIPAA compliant in regard to securing the data submitted by survey respondents. All responses are submitted using secure http (https); no respondent-supplied data is ever transmitted in the clear over the internet. The data is then housed in IRT's secure Data Center behind the IRT firewall, completely inaccessible to internet hackers. The final report which summarizes the data is viewable only by the authorized Stanford community member who commissioned the survey.

Surveys can be run in one of two modes, anonymous and authenticated. In an anonymous survey the respondent is taken directly to the survey input form; whether the respondent chooses to identify themselves (assuming the survey even asks for their identity) is up to them. This is a good deployment option when you need to permit non-Stanford community members to participate in a survey, or when running a survey where you do not require absolute certainty in identifying the respondents. This version of the sample Bicycle Use on Campus survey demonstrates this deployment option.

In an authenticated survey (also known as a WebAuthed survey) access to the survey instrument is protected with Stanford's single source authentication service for web pages, WebAuth. This version of the sample Bicycle Use on Campus survey provides a demonstration of how an authenticated survey works (you will need a valid SUNetId to actually view the survey form after following this link). In authenticated surveys the respondent's SUNetID is associated with the response in the database, available for possible later inclusion in the data report.

The person who commissioned the survey (and only this person) has access at any time to the current data report, generated on the fly by the same application that runs the survey. The current complete data tabulation and summary rollup report is delivered as a Microsoft Excel spreadsheet through a WebAuthed URL (here is an unsecured example of the format used in this report). Each row in the spreadsheet corresponds to one person's response; each column corresponds to an answer. Questions with a single answer, like free-text or radio-button multiple choice, get a single column. Questions with multiple potential answers (checkboxes) get one column per possible answer; if the respondent did not check that box, the cell is left blank. The second sheet of this report has a rollup report on the total number of responses per checkbox and radio-button category.

Additional Features

The survey is guaranteed to be available 24 hours a day, 7 days a week by virtue of being hosted on IRT's highly available high performance web servers.
Survey designers can choose to auto-number questions so they look like a numbered list
Survey designers have the option to in-line a submit button. This feature is useful in long surveys where a "no" response to the initial question disqualifies the respondent from from supplying any further responses.
Obfusticated user ids are available for HIPAA compliant reporting where the respondent's identity is used to correlate results from a sequence of related surveys (this feature is only useful when running a series of Stanford community secure surveys).
Custom Javascript can be embedded in the head of the survey form.
In a WebAuthed survey the user's full name can be added to the final report; the survey engine is able to automatically look this information up from the supplied SUNetID. Use of this feature is of course contingent on the survey containing no questions that would elicit patient-related information.

Current Limitations and Known Issues

The survey tool is currently limited to single page instruments. Support for multi-page "wizard"-style surveys is planned for a forthcoming release.
If you wish to identify your respondents by having them log in to the survey with a username and password, the respondents must all have SUNetIDs. In other words, there is currently no support for authentication of non-Stanford community respondents. You can, of course, request that your respondents voluntarily supply their name as one of the questions you ask on the survey , but there is no guarantee in this case that this information will be supplied.
In an anonymous survey all responses are anonymous, leaving open the possibility of ballot-box stuffing (one user impersonating many). However, the identity of the anonymous user is maintained for the duration of the browser's session, meaning that if the user does resubmit within the same browser session it merely updates their response. This permits anonymous users to review and potentially edit their responses, at least until they terminate their browser session. In order to actually "stuff the ballot box", they would need to first close their browser then open a new one to create a brand new anonymous user identity.

For More Information

If you would like to find out how to use the Online Survey Creation Tool to publish your survey, please contact Philip Constantinou.