Choose Research Survey Question Types

Choosing the right question type is easier when you start with the column you need in the final dataset. Before adding a question, decide whether the answer should become a category, a number, an ordered scale, open text, a visual coordinate, a ranked preference, or an experimental assignment-related field.

In the editor, open the survey, go to Block Builder, select the block where the measure belongs, and use the question type selector in the question card header. If you are still designing the instrument, keep a short note beside each construct that says what the exported value should look like.

• One answer from a visible list: choose Multiple Choice when respondents should compare the options on screen.
• One answer from a long controlled list: choose Dropdown for country, state, institution, occupation, or other lists that would crowd the page.
• A typed number or phrase: choose Short Answer, then add validation if the answer must be numeric, bounded, or contain required text.
• Several items on the same response scale: choose Grid Matrix or review Likert Scale Survey Questions for agreement batteries.
• Preference order: choose Ranking when the order itself is the answer, not just a top choice.
• Magnitude of a tradeoff: choose Constant Sum when the size of the allocation matters, not just which item comes first.
• Priorities across many items: choose MaxDiff to force discrimination between items instead of rating everything highly.
• Tradeoff experiment: choose Conjoint when attributes or levels need to vary across respondent tasks.
• Grouping items into categories: choose Card Sort when how respondents categorize items is the construct of interest.
• Reaction to specific phrases: choose Text Highlighter when you need to know which part of a stimulus stands out, not just an overall reaction.
• Spoken answer: choose Audio Response when tone, hesitation, or spoken language carries information text can't capture.

1. Step 1: Write the analysis purpose first. Name the construct, the expected variable, and how you will compare or model it. For example, a treatment comprehension item needs a cleaner response field than a reflection prompt.

   Step 1: Block Builder

2. Step 2: Match the response shape. Select the question type that produces the fewest cleanup decisions later: categorical lists for categories, numeric short answers for numeric entries, matrix rows for repeated scales, and ranking or conjoint tasks for preference data.

   Step 2: Question Type

3. Step 3: Check whether the answer drives routing. If the response will be used in Flow, quotas, or condition assignment, prefer stable options with clear labels and recodes.

   Step 3: Flow Canvas

4. Step 4: Add the question in Block Builder. Select the block, open the question type menu in the card header, and choose the type. Keep instructions in a nearby Content Block if they are longer than the question itself.

   Step 4: Add Question

5. Step 5: Name and code it immediately. Add a readable variable name and recodes while the construct is fresh, especially for scales, experiments, and exported eligibility fields.

   Step 5: Variable Name

6. Step 6: Test the exported shape. Use Preview to submit a realistic answer, then open Export and confirm the column names, values, and missing-data behavior.

   Step 6: Preview

• Does the respondent need to compare every option? Keep the choices visible with Multiple Choice instead of hiding them in a dropdown.
• Does the page feel dense on a phone? Review Mobile-Friendly Research Surveys before using large grids, rankings, heatmaps, or conjoint tasks.
• Could option order change the answer? Review Randomize Survey Questions and Answer Options before launch.
• Is the question sensitive? Use optional wording, inclusive labels, or a Prefer not to answer option when required answers would create ethical or data quality problems.
• Will the answer be merged with outside data? Choose a format with stable codes rather than free text when the variable must join to panel, administrative, or longitudinal records.

• Create Your First Academic Research Survey
• Survey Variable Names and Recodes
• Survey Data Quality Checks