Choice questions
| Type | Pick | Use it for |
|---|---|---|
| Multiple Choice | One | Mutually exclusive options — one answer is genuinely correct or preferred. |
| Checkboxes | Many | “Select all that apply”. Reasons, features used, channels seen. |
| Dropdown | One | Long lists where radio buttons would swamp the page — country, department. |
| Image Choice | One or many | Concept, packaging, and creative tests where the image is the question. |
Scale questions
| Type | Use it for |
|---|---|
| Rating Scale | Quick satisfaction reads with stars. Familiar and fast on mobile. |
| Linear Scale | Numbered scales with your own end labels. The general-purpose workhorse. |
| Net Promoter Score | The standard 0–10 recommend question. Charts as Promoters / Passives / Detractors automatically. |
| Matrix / Grid | Several statements sharing one scale. Compact, and the basis for looped follow-ups. |
| Slider | A continuous value where the exact number matters more than a band. |
| Constant Sum | Splitting a fixed total — allocate 100 points across these five things. |
| MaxDiff (Best/Worst) | Forcing real priorities out of a long list. Far better than rating everything 5/5. |
Input questions
| Type | Use it for |
|---|---|
| Short Text | Names, codes, short factual answers. |
| Long Text | Open-ended “why” questions. Pair with a loop to keep answers attributable. |
| Multiple Textboxes | Several short fields in one question — address lines, three suggestions. |
| Date / Time | Dates, with or without a time component. |
| Ranking | Putting a short list in order. Keep it under about seven items. |
| File Upload | Receipts, screenshots, documents from respondents. |
| Video Response | Recorded video answers. Research plan. |
Multiple Choice → Checkboxes whenever more than one answer could honestly be true. Forcing one pick throws away most of the signal.
Rating everything → MaxDiff when you have a long list and need to know what actually matters. People rate nearly everything highly; MaxDiff makes them choose.
Several separate scale questions → one Matrix when they share a scale. Shorter to answer and much easier to read on the dashboard.
Writing rows for a Matrix
On an agree/disagree grid, every row should be a first-person statement covering one idea — “I feel supported by my team”, not “Do you feel supported by your team?”. You cannot agree or disagree with a question, so a question-shaped row forces respondents to translate it themselves, and different people translate differently.
It also matters downstream: looped follow-ups pipe the row's own wording into the follow-up question, and a question-shaped row reads badly there.