Some survey systems tell you, right on the page, which answers end the interview. Emazing Retailing does: its demo links label every such option with "Screen out" or "Screen out quota". SurveyTester reads those labels while you test, shows them on the options in the test case editor, and turns them into test cases with one click - a Completed test case that avoids every marked option, and one Screenout (or Quota) test case per marked option. No description to write, no AI involved, and the result is exactly as good as the survey's own routing.
The questionnaire document can say the same. When you upload and analyse the questionnaire and map its questions to the survey, every option flagged TERMINATE in the document, and every option named in a rule like "screen out if S3 = 2", becomes such a marker on the mapped survey option - for any survey system, demo link or not.
How it works
- The rules. Every option that carries a marker is remembered with its outcome. In the test case editor such options show a red Screen out or orange Quota badge next to their label.
- Two sources. A marker comes either from the survey system itself or from an analysed questionnaire document (Questionnaire documents); the badge's tooltip tells you which. When both say something about the same option, the survey system wins - the live survey is the truth, the document is the plan. A document marker disappears again when a new version of the document no longer terminates that option.
- The Completed test case. For every question with marked options, the test case gets all the other options as valid answers. A single-choice question picks one of them at random on every run. A multiple-choice question stores them in pattern mode with 50% each, so a run selects a random subset of the valid options.
- One test case per marked option. It is named after the Completed test case, the question and the option
(
Complete_Screenout_Q10117_357), carries the marked option as its answer, and - so that a random answer earlier in the survey cannot screen the run out first - the Completed choice for every other question with rules. Questions without rules stay random. - Re-runnable. Generating again only adds what is missing: existing answers of the Completed test case are left as they are, and a test case whose name already exists is skipped. You can therefore start with an empty "Completed" test case and let the rules fill in while you test the survey page by page.
Step by step
- Provide the rules. Either create the project with the survey's demo link (for Emazing:
emzrtl_source=demo) - only the demo link shows the markers, a live link carries none - or upload the questionnaire document on the project's Questionnaire page, let SurveyTester analyse it and map its questions to the detected survey questions. - Test the survey once. Start a test run with the browser extension and walk through the questions - by hand or with Auto Answer. Every page SurveyTester sees records its markers. A project that was first tested through a live link picks the markers up on its next run through the demo link. Document rules need the survey questions to exist, so test first, then map; a mapping carries its terminate rules over at once.
- Open Test Cases. The button Generate from survey rules appears next to the AI buttons as soon as the project holds at least one marked option.
- Optional: pick the Completed test case. Select a test case with the expected outcome Completed - an empty one you created yourself is fine - and it will be filled. With no such selection SurveyTester finds or creates Complete (survey rules).
- Click the button. A message summarises what happened: how many questions of the Completed test case were filled, how many Screenout/Quota test cases were created, how many already existed.
- Review and run. Open the generated test cases, adjust anything you like, then run them - with the browser extension or as SurveyBots. The expected outcome of each test case is checked against the end page the run reaches.
Good to know
- Quota is not a screenout. A "Screen out quota" option only ends the interview once that quota is full. A Quota test case therefore reports Completed while the quota is still open - that is the survey behaving correctly, not a test failure.
- Rebuilding from scratch. Because generation never changes existing answers, a rule that changed on the survey side is not applied to an existing test case. Delete the generated test cases and generate again.
- Only choice questions carry rules. Markers sit on options of single- and multiple-choice questions. Open questions, numbers and grids are not part of the rules; answer them in the test case as usual.
- Combine with AI. The rule-based test cases cover the survey's own routing. Persona-driven Completed test cases and AI screenouts from a description still make sense for everything the survey does not mark.