Why Pilot Testing Matters
Preview and pilot testing catch problems before a survey reaches participants. In academic survey research, a pilot should test wording, consent, eligibility, branching, randomization, quotas, data quality checks, mobile layout, and export structure.
Use Preview for quick internal checks while building. Use a small pilot when you need realistic timing, comprehension feedback, device variation, or evidence that the data export matches the analysis plan.
Preview The Respondent Experience
Step 1: Open Preview. From the survey editor, choose the Preview tab and begin the survey as a participant.
Step 1: Open Preview
Step 2: Test basic completion. Complete the default path and watch for confusing wording, missing required settings, or validation messages.
Step 2: Test basic completion
Step 3: Test mobile layout. Resize the browser or use a phone to check matrix, ranking, heatmap, slider, and conjoint tasks.
Step 3: Test mobile layout
Step 4: Test Flow paths. Return to Preview with different answers to exercise each branch, treatment arm, and end screen.
Step 4: Test Flow paths
Step 5: Test quotas if used. Review quota behavior in Deploy and confirm quota-full paths end where expected.
Step 5: Test quotas if used
Step 6: Inspect Export. Submit test responses and open Export to confirm headers, recodes, conditions, and randomized-order fields.
Step 6: Inspect Export
Pilot With Realistic Respondents
After internal preview, run a small pilot with people who resemble the target population. Ask them where wording is confusing, how long the survey takes, which page felt most burdensome, and whether any instructions or end screens were unclear. Use pilot feedback before publishing the final public link.
Keep pilot notes tied to the survey version. If a pilot response should not be analyzed later, mark it in your cleaning log or collect it through a route you can identify in Review and Clean Survey Responses.
Preview: Respondent Experience
What To Check Before Launch
- Instructions: consent, task instructions, validation messages, and end screens use participant-facing language.
- Question types: each measure matches the intended exported variable and has needed names and recodes.
- Flow: screeners, branches, experimental conditions, and end screens route correctly.
- Quotas: targets track the intended arms or eligibility groups, and quota-full paths are understandable.
- Mobile layout: dense grids, rankings, sliders, heatmaps, and conjoint tasks are usable on small screens.
- Export: response data includes fields needed for analysis, cleaning, conditions, and randomized order.
Export: Data Output