Publishing

To Publish a survey means to make it available to Illume‘s collector service, which will make the survey available to participants or interviewers for data collection. A survey must be approved before it can be published (see Approving a Survey). Administrators, Reviewers and Power Users may approve a survey.

To preview a survey to see how it looks and to test its validation features, use the Survey Previewer. With the Survey Previewer it is possible to take the entire survey without having to make the survey available to participants.  Surveys may also be Test Published to produce a link that can be distributed to testers, with the data from the survey stored in an entirely separate table from the Published survey.  See the Test Publish section of online help for more information on that topic.

To publish a survey, follow these steps:

1. Click the Survey Administration tab of the Illume Survey Manager.

2. If the survey does not appear in the list of surveys under Search Results, click on the Survey Search tab.

3. Click Clear All Search Criteria, then click Search. This will display a list of all available surveys.

4. Right click on the name of the survey and choose Approve and Publish from the context menu.

5. (Optional) Add comments. These comments become part of the survey history.

6. A time period must be supplied for publication. A time period is usually a short code to represent the period during which a survey is given. For example, if a survey is administered every quarter, assign the time periods “2004_Q1,” “2004_Q2,” etc. If the survey will be given only once, assign a simple period code like “1.” See Understanding Time Periods for more about time periods.

7. An Email Job Resend Warning will appear.  This warning applies to surveys that are launched over multiple time periods that have associated email jobs.  If a survey is launched at time period 1 and has a running email job associated to it, and that survey is published again under a new time period without the email job being terminated, the participants on the list associated to the email job will be emailed again.  Thus, before a survey is published under a new time period, any running email jobs associated to that survey should be terminated.

8. Click OK.

The Survey will now have a Status of Running. Once a survey has been published, it is ready for participants or interviewers to begin submitting responses.

Unpublishing a survey means reverting it to the prior version.  Doing so has two effects:

Only a user with the Administrator role can Unpublish a survey, and only surveys that have been published can be unpublished. To stop a running survey so that it is unavailable to participants, do not use the Unpublish feature.  Rather, see the article on Suspending a Published Survey.

To unpublish a survey:

1. In the Survey Administration tab of the Illume Survey Manager, right click on the name of the survey to unpublish.

2. Choose Un-publish from the context menu.

3. Read the warning carefully and click OK.

4. Read the second warning carefully and click OK.

To suspend a survey means to make a running survey unavailable to participants. Nothing about the survey changes, other than its availability to participants. You must be an Administrator or Publisher to suspend a running survey.

To suspend a running survey:

1. In the Survey Administration tab of the DatStat Illume Survey Manager, right click on the name of the published survey to suspend.

2. Choose Suspend from the context menu.

3. (Optional) Add comments. These comments become part of the survey history.

4. Click OK.

A suspended survey can be resumed at any time. See Resuming a Suspended Survey.

Resuming a suspended survey means making that survey available again for active data collection.

The Resume option will be enabled in the context menu only when the survey has been published and then suspended.

To Resume a Suspended Survey:

1. In the Survey Administration tab of the Illume Survey Manager, right click on the name of the suspended survey to resume.

2. Choose to resume the survey in either the same time period or a new time period. Time period is a label that is applied to all submissions within the time period, in the variable “DATSTAT.TIMEPERIOD”.  The Time Period concept is valuable for cohort-based data collection, wherein all participants take the survey during a defined wave (e.g. Spring Quarter Time Period), and at the end of the wave, the survey is suspended.  The survey is then resumed under a new time period (e.g. Fall Quarter Time Period) so that the same participants can take the survey again.  Without a new Time Period designated, participants associated to an authenticated survey can only take the survey once.  If, however, it is a new Time Period, then the same participant can complete the survey again at the new time period.

3. (Optional) Add comments. These comments become part of the survey history.

4. Click OK.
   

As a participant moves from page to page with the Next button, all data from the page are automatically saved and sent to the server, providing the ability for participants to resume their survey sessions right where they left off.  Alternatively, an optional Save button can be added to each page, making the save of data more explicit to the participant.  When a participant returns to a saved, partially completed survey, the participant first lands on the Resume/Restore page (configured in the Illume Survey Designer under Preferences).

If a survey participant resumes a survey that was started in one version, and the survey increased in version number prior to resuming the survey session, the participant will resume the survey in the new version, and will be able to resume the survey right where they left off  if “major” changes were NOT made. If “major” changes were made the survey participant will start over from the beginning of the survey.  Any responses previously entered will be retained, but the participant will have to page forward through those responses to reach the page on which they left the survey.

The following ARE considered major changes:

The following are NOT considered major changes:

Overview

On publishing a survey, you must choose the time period for which to publish in. If checking in, approving and publishing in a single step, you must supply a time period at that time.  A Time Period is a label that is tied to a submission that indicates when it was taken.  This is primarily applicable for surveys wherein a cohort of individuals all take a survey at one point in time, and that same cohort takes the survey again at a later point in time. For example, an employee satisfaction survey opens for a week of each quarter.  That week each quarter is the Time Period, and these Time Periods might be called ‘Quarter1’, ‘Quarter2’, etc.  Although authenticated surveys by nature are designed to prevent multiple submissions from the same respondent, the use of Time Periods makes it possible for a respondent to complete the employee satisfaction survey each quarter, because it is designated as a new Time Period.  The Time Period of a submission is stored in a variable within the survey dataset (“DATSTAT.TIMEPERIOD”).  One valuable use of this variable is as a Crosstab variable when creating a query, as it allows users to see how summary statistics change over time.  Without a new Time Period designated, participants associated to an authenticated survey can only take the survey once.  If, however, it is a new Time Period, then the same participant can complete the survey again at the new time period.

If publishing a survey, where there is currently a version of that survey running, it is not possible to change the time period. The survey must be suspended first, and then resumed under the new time period

Change the Name of a Time Period

To rename a time period:

1. Select a survey in the Survey Administration tab.

2. Right-click Properties

3. Select the Time Periods tab.

4. Select a specific time period.

5. Rename the time period name in the text field.

6. Click the Rename button.

Test publishing makes a survey available for testing by generating a link that can be distributed to testers. When a survey’s test-publish option is on, others can view and test the survey using only a web browser.  The data from the test published survey go into an entirely separate data table than the Live survey.

One very useful function of test publishing is to test changes made to a live published survey prior to pushing those changes to production by Publishing the survey.  When changes are made to a published survey, the survey can be checked in with test publishing turned on.  The test-publish link can be distributed to testers, feedback received, and changes made prior to clicking Publish.

To turn on test publishing for a specific survey:

1. Click the Survey Administration tab in the Illume Survey Manager.

2. Right click on the name of survey to test-publish.

3. Choose Test-Publish from the context menu.

4. Check On under Test-Publish status.

5. If the survey includes more than one translation, choose the translation to test publish(see Test Publishing Translations below for details).

6. Click Done

Testing in Test Publish Mode

One of the main differences between taking a survey in Test Publish mode vs. Live mode has to do with what happens when changes are made to the survey.  One of the benefits of test publish mode is that you can make any types of changes to the survey while it is in this mode.  You can delete questions, change data types, change scale values, and other changes that are not permitted after the survey is published.

When the Test Publish link has been distributed to testers, and testers are actively taking the survey, you may want to check out the survey to make changes, and then check it back in again.  When you check it back in again, if testers are actively entering data into the survey, the testers will get kicked out of the survey at the moment you check the survey back in.  This will NOT happen in a live survey.  In a live survey, data collection is not interrupted by making changes to the survey.  Thus, you should either refrain from making changes while a test published survey is being administered to respondents, and/or you should warn testers that they may get kicked out of the survey if you make changes (and you should also let them know that this will not happen in the live survey, that this is an artifact of test publish mode).

Test Publishing Translations

When test publishing a survey that includes more than one language, Illume test-publishes all of the survey’s translations, no matter what translation is chosen in the test-publish dialog. Choosing a translation only affects the Survey Preview URL and the Survey Test URL that appears in the test-publish dialog.

For example, in the dialog above, changing the translation to Italian (Italy) changes the end of the Test and Preview URLs from Translation=en-US to Translation=it-IT.

It is possible to get the URL for any language in a multi-lingual survey by changing the language in the test-publish dialog. Clicking on the URL directly to view the survey in a browser, or use the Copy URL button (described below) to copy the URL into an email, instant message, or other document.

Preview and Test URLs

In the Test Publish Survey there is a Survey Preview URL and a Survey Test URL. The Survey Preview URL provides access to the latest checked-in version of the survey in preview layout: the entire survey appears on a single HTML page, with show-if logic and other normally hidden information displayed for printing and review.

The Survey Test URL links to an interactive copy of the latest checked-in version of the survey. This version implements all of the survey’s built-in logic, piping, calculations, and other behaviors. Testers can complete and submit surveys from this URL, and the data will be available for querying through the Data Manager.

The Test-Publish dialog includes buttons to copy and reset these URLs. Clicking Copy URL will copy the URL into the Windows clipboard. You can then paste the URL into most applications by choosing Edit Paste from the application’s menu, or by pressing Ctrl-V.

The Reset URL button generates a new preview URL or a new Test URL. Resetting the URL effectively cancels the old URL. No one using the old Test/Preview URL will be able to access the test-published survey unless the new URL it provided.

Both the Test URL and the Preview URL are also available through the Data Manager’s survey page.

Published and Test-Published Versions

Surveys may have simultaneous Published and Test-Published versions. The published version, which is intended for actual participants, displays the last version of the survey that was explicitly published. The test-published version displays the last version of the survey that was checked in. It’s possible, therefore, for the two versions to have different content. Both the Survey Preview URL and the Survey Test URL end with long alphanumeric id strings. These are designed to prevent unauthorized individuals from guessing the URL and accessing the test version of a survey. Publishing, unpublishing, and suspending a survey has no effect on the availability of the test-published version. Turning the test-publish option on and off has no effect on the published version of the survey.

Separation of Data

Data submitted to the test published version of a survey are stored separately from data submitted to the published version of a survey.

Consider these three scenarios:

Once a test-published survey collects data, if the test-published survey is checked out to insert changes and the survey changes substantially, and then the survey is checked back in, the test-publish data will be wiped from the test publish data table.   This means that data submitted to the test-published version of the survey might be lost, if changes are made to the data type of questions from one test-published version to the next, or if questions are deleted between test-published versions. These changes will NOT affect the live survey or any data collected from any version of the live survey.

Associating a Participant List to a Test Published Survey

It is possible to test an authenticated survey in Test Publish Mode. Create a Participant list following the steps for creating participant lists. This list could be the Live list with only Test Participants added.

1. Once Test Publish is turned On in the Survey Administration tab of the Survey Manager.

2. Go to the Enterprise Manager and navigate to the appropriate Project.

3. Expand the Project and Click on Surveys.

4. Click on the Test Published Surveys tab.

5. Select the appropriate Survey and Click the Edit Survey icon.

6. Select the appropriate participant list to associate and click the Add button.

Test Participants vs. Test Publish

A Test Participant is a participant that is part of the live participant list.  These Participants data will not show in queries of live data, but can be used to test email jobs and the live survey. Test Published Participants all behave as if they were live participants but the data will never mix with the production data.