Once we have piloted the test, we will likely have found some minor things to change, and perhaps even a few major ones. The great part of this is that we get to fix the test before our real audience sees those mistakes.

Depending on the testing tool, we may be able to edit the existing test, or we may need to duplicate it and edit the new copy.

If the tool lets us edit the existing test, we can make our changes right there.

Most changes can be done without consequences, but we should be careful about two things:

• Editing the tree can affect the task answers.
  If we make structural changes to the tree, we should make sure that the task answers are still correct. For example, if we move a topic in the tree, and it was an answer to one of the tasks, we should update the task answer accordingly.
  
  For more on revising the tree and your tasks, see Chapter 14 - Revising and retesting.
  
  
• Delete any pilot data you’ve collected.
  If we plan to use the same test instance for the real study, we should delete the data collected during the pilot testing. However, if the data came from representative users and we didn’t make any substantial revisions to the test after piloting, we can keep this data to join the “real” data we’ll be collecting later.
  
  
  

Some tools don’t let us edit a test that has been launched. So if we launched a test for purposing of piloting, we’ll need to duplicate it first, then edit the newly created duplicate.

This has the advantage of keeping a revision history (of sorts) in case we want to go back to a previous version to check something or to grab an older version of a task’s wording (for example).

Because we’re revising (and eventually launching) the newly created duplicate, we don’t have to worry about deleting pilot results; they’ll be attached to the previous copy, not the new one.

Next: Chapter 10 - key points