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:
- Click the Survey Administration tab in the Illume Survey Manager.
- Right click on the name of survey to test-publish.
- Choose Test-Publish from the context menu.
- Check On under Test-Publish status.
- If the survey includes more than one translation, choose the translation to test publish(see Test Publishing Translations below for details).
- 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:
- Submit live survey as real participant. Result: Data go to live survey database table
- Submit live survey as test participant by setting the value of DATSTAT_TEST on the participant list to YES. Result: Data go to live survey database table and are marked as test data. These submissions may be deleted later from the Test Data tab of the Survey page in Data Manager. These data can be found by creating a Query, and checking the Use Test Data box in the Query Properties
- Submit test-published survey. Result: Data go to the test-published survey database table.
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.
- Once Test Publish is turned On in the Survey Administration tab of the Survey Manager.
- Go to the Enterprise Manager and navigate to the appropriate Project.
- Expand the Project and Click on Surveys.
- Click on the Test Published Surveys tab.
- Select the appropriate Survey and Click the Edit Survey icon.
- 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.