Documentation

Downloading data and reports

Use the Download files dialog when you need respondent-level data, a bundle of reports, respondent transcripts, or a specialist analysis file. Open the survey, select the download icon in the top-right corner, then choose the dataset and file type before starting an export.

Choose the dataset first

Use the Dataset list to choose which responses the file should contain. Select Live for collected fieldwork, Test for survey testing, or another available dataset when you are checking simulated, synthetic, or separately prepared results.

The available cards can change with the dataset. For example, exports that need completed responses are unavailable until the selected dataset has completes. Check the dataset again whenever a card you expect to use is missing.

Filter the download cards

Use File Type to narrow the dialog to:

  • Excel for report workbooks, wide or long respondent data, stacked data, and reconciliation files.
  • SPSS for standard, Q-optimized, stacked, or analysis-ready respondent data.
  • Text for CSV report bundles, raw CSV data, specialist CSV outputs, and respondent transcripts.

Each card names the output and describes what it contains. Selecting a different file type does not change the dataset. The dialog remembers your most recently selected file type while you continue using the platform, so check both controls before starting the next export.

Download files dialog with Live selected as the dataset and the Excel, SPSS, and Text file-type choices

For help choosing an SPSS structure, see SPSS Export Format. For Excel structures, see Wide Excel Format and Long Excel Format.

Numeric-coded wide Excel

Choose File Type: Excel, then Numeric-coded Data in Excel (Wide) when you need one row per respondent with SPSS-compatible numeric codes. The workbook contains Data and Data Map worksheets. Keep the map with the data: it explains the numeric code, original response value and readable label for each exported field.

Download files dialog with Live and Excel selected and the Numeric-coded Data in Excel (Wide) card highlighted

This is a separate choice from the standard wide Excel export, which uses readable reporting values, and the uncoded export, which returns captured responses. Numeric coding does not turn category codes into measured quantities; a code for an age band is not the respondent's age. See Wide Excel Format for the representation.

Select Dataset: Test and File Type: Excel for:

  • 100 Test Links in Excel — 100 unique test sessions using the survey's primary language. Assign one link per reviewer or independent pass.
  • All Testing Report — test attempts grouped by saved survey version, with each attempt's answers and timestamps. Available on the original survey, including after test responses have been cleared.

100 Test Links in Excel download card describing 100 unique primary-language test links

All Testing Report download card for test attempts preserved across survey changes

See Testing and Reviewing Your Survey for how to use the workbook and how clearing test data affects its history.

All Testing Report respondent drilldown with separate attempts, chronological answers and America/New_York timestamps

Track downloads in the background

Select a download card to start preparing the file. You can close the dialog and continue working while the export runs. The Downloads panel opens with the new job marked Preparing; collapse it when you need more screen space and use the Downloads button to reopen it.

Downloads panel showing an SPSS File and Codebook job preparing in the current session

When preparation succeeds, the browser download starts automatically and the completed job is removed from the panel. If preparation fails, the job stays in the panel with the reported error:

  1. Select Retry to run the same export again.
  2. Select the dismiss button when you no longer need the failed job.

Preparing jobs cannot be dismissed from the panel. Download jobs are held only in the current browser session: refreshing, closing, or leaving the session clears the list. The platform does not provide persistent download history, so retain the downloaded file somewhere appropriate for your project.

Generated downloads use descriptive, filesystem-safe names that identify the relevant survey, project, report or dataset and include a UTC generation timestamp. Cached transcript and insight presentation downloads keep their filename across polling and repeat downloads; rebuilding one gives it a new timestamp. Keep the full extension when moving files, including compound extensions such as .tar.gz for raw CSV archives.

Use the raw-data map

Raw Excel and CSV downloads include a data map that explains how export columns relate to the survey. Use it before renaming columns, building a script, or handing the file to another analyst.

  • Raw Excel workbooks contain Data and Data Map worksheets.
  • Raw CSV downloads are delivered as a .tar.gz archive containing respondent_data.csv and data_map.csv.

Use the map to find the source question and response labels for each exported column. When one question creates several columns, the map also identifies the option or topic represented by each one.

SPSS variable names and wide Excel headers use normalized stable reporting IDs, with deterministic numbered suffixes for expanded options and topics. Readable labels are kept separately. For example, reporting ID how-old-are-you exports as how_old_are_you, with reporting label Age and original question text How old are you?.

For wide Excel, join a header to export_column in Data Map, then inspect reporting_id, reporting_label, question and generated_column_source. For SPSS, choose SPSS File and Codebook and use SPSS variable name, Reporting ID, Reporting label, Full question text and Option/topic suffix. Expanded columns share a question ID; the option/topic field explains each numbered suffix. Long data keeps the source reporting_id directly.

Compatibility: these identifiers replace previous V###_descriptive_label headers. Update scripts, saved SPSS syntax, joins and BI imports that match the previous headers or readable labels. Use a fresh data map or codebook to build the old-to-new mapping, verify expanded options, and retain that mapping with the export. Do not assume an old positional variable number is a reporting ID.

See the wide-format examples and SPSS codebook examples for actual mappings and collision handling.

Open-ended verbatim text is not repeated in the data map. Use the respondent data file when you need the actual response.

Newer raw respondent exports can include respondent_sequence, a survey-wide ordinal assigned to a respondent. It is available in raw CSV, Excel, SPSS, analysis bundles, their maps or codebooks, and respondent reconciliation. Filtering respondents preserves their original sequence numbers, so gaps are expected. An older dataset without sequence metadata retains its existing columns; a sequence value can be empty when the underlying marker is unavailable. The All Testing Report is separate and does not add this field.

Respondent quality fields

Supported raw CSV, Excel and SPSS downloads include these fields when the dataset has recorded quality metadata:

FieldMeaning
quality_suspect_scoreRecorded MX8 suspect score from 0 to 5. Missing evidence is not a score of zero.
quality_duplicate_detectedRecorded positive or negative duplicate finding. Missing evidence is not a negative result.
quality_termination_reasonSaved primary reason for a quality or duplicate exclusion. A blank or unrecorded reason does not establish why a respondent was excluded.

Wide Excel and SPSS put available quality fields with respondent metadata before survey-question fields, using the latest recorded value per respondent and field. Long Excel and raw CSV retain quality metadata as rows identified by these reporting IDs; a respondent can have more than one recorded update.

Use Data Map or the SPSS codebook for each format's value representation. Raw termination reasons use stable codes; readable or coded exports supply labels or code-to-label mappings. MX8 suspect scores retain their measured 0–5 values in numeric-coded Excel and SPSS.

Quality evidence is independent of exclusion policy. A positive finding does not mean that the respondent was terminated, and current evidence can differ from the evidence available when a saved exclusion occurred. Check status and the recorded reason alongside the score. Historical datasets without quality metadata retain their existing fields.

See the field report's Quality section for aggregate distributions and actual terminations.

Download a cross-tab report by question

Open the individual cross-tab report, use its download menu, and choose Excel (Question Tabs). The workbook arranges the report into question tabs so you can move between questions in Excel.

Cross-tab download menu offering Excel, Excel (Question Tabs) and CSV

This option is for a single cross-tab report. The standard Excel download and All Reports in Excel bundles keep their existing layouts. See Reporting with MX8 Labs for where this fits in the reporting workflow.

Download a field report

Open the field report from the survey's Respondent Sources section, then use Download → Excel. It uses the report's source scope, dates, timezone and minimum question base, and includes all sections and table pages. Eligible voice/text sources also offer Dispatch CSV in this menu; that destination snapshot is separate from the field-report controls.

Choose the right report download

Use All Reports in Excel when you want one workbook containing every survey report. Use All Reports in CSV when you want the same reports as CSV files in a ZIP archive. Multi-sheet Excel workbooks include a Contents sheet and return links to help you move through long files.

Project report downloads work the same way from the project's Reports tab. For the full reporting workflow, see Reporting with MX8 Labs.