# Survey Design Checklist

A working checklist for writing clear, unbiased survey questions and choosing the right answer format before you build.

## Question clarity

- [ ] Each question asks about one thing only (no "and" hiding two questions in one)
- [ ] Wording avoids jargon, acronyms, or internal terminology your audience won't recognize
- [ ] Questions are written at a reading level appropriate for your audience
- [ ] Time references are absolute ("in the last 30 days"), not relative ("recently")
- [ ] Leading or loaded phrasing has been removed (avoid "How much did you enjoy...")

## Scales and answer choices

- [ ] Rating scales use a consistent direction across the whole survey (low-to-high, not mixed)
- [ ] Scale labels are balanced (equal positive and negative anchor points)
- [ ] You've decided deliberately between an odd scale (with a neutral midpoint) and an even scale (forced choice)
- [ ] Multiple-choice lists include "Other (please specify)" where the list may not be exhaustive
- [ ] Answer option order is randomized where order effects could bias results

## Question order and flow

- [ ] The survey opens with an easy, low-commitment question, not a sensitive one
- [ ] Sensitive or demographic questions are placed near the end, not the start
- [ ] Related questions are grouped into logical pages or sections
- [ ] Skip logic routes respondents around questions that don't apply to them
- [ ] The survey length matches the incentive and audience patience - shorter is almost always better

## Testing before launch

- [ ] At least one colleague outside the project has completed the survey and given feedback
- [ ] Every skip-logic branch has been walked through, not just the "happy path"
- [ ] The survey has been completed on a phone-sized screen
- [ ] Estimated completion time has been checked against what respondents were told

## Reporting alignment

- [ ] Every question has a clear owner for how its results will be analyzed
- [ ] Cross-tab and segmentation needs (by role, region, tenure) are decided before launch, not after
- [ ] Open-text questions have a plan for review (manual read-through or AI-assisted theme summary)
- [ ] The final report format (dashboard, PDF, scheduled email) is agreed with stakeholders in advance

---
Published by Sierratec Survey. Adapt freely for your own workspace and studies.