Skip to main content
Start with the section that matches the view you are using. Error text is included where it helps identify the state.

Build and test a survey

Open the problems indicator in the Create toolbar and resolve each blocking item. Saving is also unavailable while the assistant is responding or when the current questionnaire matches the last saved version. See Fix survey problems.
The control offers two actions. Choose Save changes to save directly, or Save & get feedback to save and ask the assistant for a review.
Read the placeholder in the composer. If it says Save or Undo your changes before sending a message, either save or discard the local edits. A restored-draft banner can appear after reopening a study; choose Save or Discard there as well. The builder chat also becomes read-only after publication.
Open Version history, preview the version you need, and restore it. If the editor reports that another person changed the form, compare the saved version before submitting your own edits. See Edit and organize a survey.
Open its Code tab and resolve the compiler message. Also confirm that every custom-text key is non-empty and unique. In Preview, submit a representative answer for the underlying question type. See Build a custom question interface.
The source question must appear earlier in the survey and be declared under Advanced — inputs & text → Inputs. Use the declared input through the question-kit API; code cannot read the page, browser storage, or arbitrary application state.

Translation

Source wording changed after translation. Open the language bar and run Complete translation, review the updated languages, then save. See Translate your survey.
A translation job is running. Wait for the language indicators to finish, then add another language.
Survey content and the participant interface have separate language coverage. A translated questionnaire can appear within English navigation when the interface itself is unavailable in that language.

Recruiting and publishing

Open Quotas, panels & links and configure its respondent target, screening, and quotas. It is the panel configuration for the study and should be reviewed before publishing. See Configure panels and demographics.
Business panels can require a manual quote; use Request Quote. Consumer estimates update after the audience and screening rules are available. Check that the target and required audience criteria are complete.
Min is the target for a cell. Add Max when collection must stop at a specific number. The panel continues recruiting while its overall target and uncapped cells remain open. See Manage quota progress.
A panel with attached responses is retained so collected data keeps its source. Use its available management controls rather than deleting it.
Open the publish control or problems panel to see the blocking item. Common causes include validation problems, stale translations, missing panel configuration, or insufficient credits. See Test and publish your study.

Responses and quality

Use the response-table column filters. The browser only searches rows currently rendered on screen, not the full dataset.
Test-panel sessions and some direct-invite rehearsals are not quality-rated. For a recruited response, open its detail to confirm which checks ran. See Review response quality.
Open the response’s quality detail, review its answers and checks, then choose Exclude and record the reason. Exclusion is reversible and updates the analysis base.
Open Responses and export the current table to Excel. The workbook follows the active filters and includes participant-level answer columns. See Export response data.

Reports, charts, and slides

Open Analysis plan. Generation is unavailable in Draft; a plan can still be saved. Generation needs at least one qualifying completed response. Open-question themes additionally wait until qualifying completed responses reach 80% of the combined positive panel target. Resolve missing plan references, invalid weighting, or unfinished variable calculations. See Generate and update analysis results.
Check the selected questions, per-question comparisons, output types, and response filters, then open the plan folder in Charts. Screening questions get totals only. Selected tables are generated by default. If significance-only filtering is enabled, comparisons without eligible significant cells are removed. Build the full table in Charts to inspect its bases. Regenerate after correcting the plan. See Build an analysis plan.
A generation with a different output language archives the previous analysis version and its associated chats and slides. A run in the same language, including a weighting change, preserves conversations. Also check the chat selector: each generated deck has its own conversation.
Check its unweighted base and the table footnote. Small bases can be excluded from testing; a table can also have no statistically significant comparisons. See Interpret significance.
PowerPoint includes crosstabs as editable tables. Verify that the table is inserted in the saved deck you exported, then inspect continuation slides. Export Charts to Excel for the complete grid and statistics. See Export your results.
Save the current deck first. Comments and PowerPoint export require a saved version with no pending local edits. See Build and review a slide deck.
A plan run builds the plan folder in Charts, not a deck. Verify its charts, gather the ones you need in a folder, and select Generate slides from this folder; the assistant proposes a storyline for your approval. Without a custom plan, generate one folder per survey section, review its outputs, then build slides from a folder. See Build and review a slide deck.
Regenerate in “folder name” deletes everything in the plan folder, including charts you edited or filed there, then builds the charts again. Before the next run, move the charts you want to keep to another folder or duplicate the whole folder, or choose Create a new folder in Where should the charts go? to leave the current folder as it is. See Generate and update analysis results.
Deleting a chart folder deletes all charts inside it. Each saved chart belongs to one folder. Move the charts you want to keep to another folder before deleting the original folder.
The questionnaire changed after the chart was saved. Edit the chart to use the current question or rebuild it from the current data structure.

Sharing and account access

Membership and study access are separate. Add the person through Share → People, or grant access on the containing folder. See Manage workspace members.
Collaborator accounts do not have access to those settings pages. An administrator can review or change account type in Workspace.
Sign out, select Forgot your password? on the sign-in page, and follow the email link. Password reset is not in account settings.

Participant access

They must read the consent information and select its acceptance checkbox. On a small screen, ask them to scroll above the button.
Backward navigation must be enabled for the study. If no Back control is shown, provide a new invitation when the response needs to be restarted.
Ask them to allow microphone permission for the site, select the intended input device, unmute it, and speak during the test. If the study permits typed responses, they can use that alternative. See Participant FAQ.

Contact support

If the problem remains, contact support from the affected signed-in page. Include the study name, full URL, exact error, expected result, actual result, and approximate time.

Check the result

Repeat the action after the correction and check its saved result. If the same error remains, contact support with the study, URL, exact message, expected result, and approximate time.

Customer files and campaigns

  • A customer-file import conflicts with a provider: a customer-keyed file requires personal links. Use the intended campaign route or a respondent-ID file for post-fieldwork segmentation.
  • A campaign cannot send: check publication, usable recipients, message languages, sender verification, timing, contact-spacing limits and the displayed blocker.
  • A missing open-ended result: raw answers remain in Responses and its export; themes and analysis use the open-answer threshold.
  • A privacy deadline has passed: check the recorded erasure status; overdue data requires administrator action.
See Data sources, Email campaigns and Personal data.