Import Research Surveys

Importing helps when a research team already has a survey instrument in another format and wants to rebuild, pilot, or revise it in Domandata. Supported import paths include Domandata JSON, Qualtrics QSF, document-based survey extraction from Word, PDF, HTML, TeX, and text, a public Google Forms link or embed code, and a public Microsoft Forms link or embed code.

Think of import as a draft builder. It can save time, but the imported survey still needs a full review of question types, Flow, page breaks, variables, recodes, consent text, and export shape before it becomes a research instrument. Domandata accepts JSON exports saved by common desktop editors as well as files downloaded directly from the app.

1. Step 1: Open Surveys. Start from the dashboard and use the Import button near New Survey.

   Step 1: Open Surveys

2. Step 2: Select the source. In the import dialog, choose a supported file such as JSON, QSF, DOCX, PDF, HTML, TeX, or text (survey files — JSON, QSF, and Domandata packages — can be up to 20 MB, including any embedded images; documents up to 15 MB) — or switch to the Google Forms or Microsoft Forms tab and paste the form's share link or <iframe> embed code. A Google Form must be shared as "Anyone with the link"; a Microsoft Form must be shared without requiring sign-in to respond (the form owner controls this under Collect responses → Who can fill out this form). Pasting a Google Forms or Microsoft Forms link or embed code anywhere on the Surveys dashboard imports it immediately, without opening this dialog. Text files saved as UTF-16 ("Unicode" in Windows Notepad) import like any other; an empty or blank file is rejected. A document without Domandata's own layout imports up to 2,000 questions, and the import summary says how many were left out if it has more.

   Step 2: Import Dialog

3. Step 3: Name the draft. Use the Survey name field if the file name is not a clear study title.

   Step 3: Name Draft

4. Step 4: Open the imported survey. After import, open the new draft in the editor.

   Step 4: Open Draft

5. Step 5: Review Block Builder. Check blocks, question types, prompts, answer options, variable names, and recodes.

   Step 5: Block Builder

6. Step 6: Review Flow. Open Flow and inspect branching, skip logic, condition sets, and end screens.

   Step 6: Flow Canvas

7. Step 7: Preview and pilot. Use Preview and a small pilot before using the imported survey for live research.

   Step 7: Preview

A survey exported from Export as Domandata JSON (or as a .domandata package, which also bundles the images), or fetched through the MCP, imports back as an exact copy of the instrument: every block, question, answer option, variable name and recode, custom HTML and JavaScript, skip rule, Flow layout and branching, condition set, value, quota, loop, randomization setting, theme, translation, and response quality rule. Survey settings from Deploy come along too: what happens to unfinished responses (submit when the tab closes, auto-submit after a number of days), what the link shows after the survey closes and when that link retires, how long responses are kept, and the saved export privacy choices. Blocks and questions get new ids in the copy, and every setting that refers to them is updated to match.

A few things are deliberately not carried over:

• API keys for Text Chat and Audio AI questions are blanked; saved credentials from your account are referenced, not copied, so reconnect them if the copy lives in another account.
• Live condition-assignment counts start fresh, so balancing in the copy begins from zero.
• The closing date is left unset, since a date that has already passed would close the copy as soon as it was created. Set a new one in Deploy if the copy needs it.
• Things that belong to the running study rather than the instrument — publish state, the public link name, the field-administration link, collaborators, responses, and repository deposits — stay with the original survey.
• Settings that fail validation (for example, a hand-edited file with a retention period Deploy doesn't offer) are skipped, and the copy uses the default for that setting.

After importing from another platform or a document, the new survey opens with a summary of anything that was approximated or couldn't be brought over. Nothing is dropped without being listed there.

• Qualtrics QSF: questions, choices, recode values, required and request-response settings, text-entry box size and validation (number, email, phone, zip code, date, length), matrices, multi-statement sliders (as a slider grid), constant sum, rank order, page breaks, and display logic within a block (as follow-up questions). Survey Flow randomizers that show one element become a random condition set in Flow, one arm per element, with balanced arm sizes when Evenly Present was on; randomizers that show several plain blocks, or shuffle them, become one Flow path per possible order when there are six or fewer. Embedded data set in the Survey Flow becomes Flow values: a fixed value carries over, and a field with no value ("set from panel or URL") reads the URL parameter with the same name — check it if the data came from a Qualtrics contact list — so piped text and branches on those fields keep working. Listed in the summary instead: larger randomizers (imported in a fixed order), embedded data set inside a branch or randomizer, set more than once, or computed from piped text, display logic across blocks, skip logic, choice-level display logic, carry-forward choices, custom validation, and matrix variants that are approximated.
• Google Forms: sections, question types, required, character limits, email and checkbox-count validation, and one-line Short Answer boxes. Number validation keeps "must be a number" but not its bounds, and question descriptions are listed rather than kept.
• Microsoft Forms: named sections become blocks, Dropdown questions stay dropdowns, question types, required settings, and one-line Text boxes are preserved, and answer choices that continue to another section or end the form become Flow paths. Question subtitles and any additional branching metadata that cannot be interpreted are listed in the import summary.
• Word, PDF, HTML, LaTeX, and text documents: each numbered question keeps its own answer options (including numbered lists with bulleted options, LaTeX enumerate and itemize lists, and a)/b) options), and checkbox-style options (☐ or [ ]) import as select-all-that-apply. A Domandata text or LaTeX instrument export re-imports with its dropdowns, requested questions, select-all questions, constant-sum limits, and piped text intact.

Google Forms and Microsoft Forms imports also get an end-of-survey block, ready for a closing message.

• Question type choices: imported multiple choice, dropdown, grid, and text-response fields may need to be adjusted for analysis.
• Answer option labels: check order, display text, missing options, and any Prefer not to answer choices.
• Branching and skip logic: Google Forms and Microsoft Forms sections become blocks, and supported answer-based routes to another section or the end of the form become Flow paths. Review those paths after import and rebuild only the platform-specific logic called out in the import summary.
• Variable names and recodes: rename vague imported fields and add numeric recodes before exporting.
• Experimental conditions and quotas: confirm arm names, assignment rules, and sample targets before launch.
• Consent and instructions: make sure imported document text did not merge prompts, headers, or footnotes into respondent-facing wording.

Import is a starting point, not a final quality check. Imported surveys should be reviewed like new instruments, especially if the original platform used logic, display rules, embedded data, page timing, randomization, or custom code that differs from Domandata.

• Design an Academic Survey Instrument
• Preview and Pilot Test a Survey
• Branching and Skip Logic for Research Surveys
• Duplicate and Adapt Research Surveys