Export Survey Data to SPSS, CSV, Excel & ASCII

The Export wizard

Every option in one view — Format, Variable content, Template, Status, Environment, Advanced. A live summary panel shows exactly what will ship before you click Export: format, variable count, respondent count, open-end handling, companion files, estimated size.

Recent exports list below the summary — one-click re-download. See Export history.

Four file formats

Four formats off one canonical model:

SPSS — .sav binary or .dat + .sps pair (fixed-width + syntax file). PSPP-verified on every build.
CSV — UTF-8 with BOM, camelCase headers. Opens cleanly in Excel, R, Python, Stata.
Excel — .xlsx with proper type hints — numbers stay numbers, dates stay dates.
ASCII — fixed-width column layout, industry-standard for legacy processing pipelines.

Numbers reconcile across formats — same rows, different packaging.

→ One canonical model, many engines

The Datamap Designer is the single source of truth for column widths, labels, and start positions across all four formats. Rename a label once; everywhere updates.

Variable content

Three presets for what goes in the file:

All — every variable in the datamap.
Closed-end — numerics, singles, multis, grids, scales.
Open-end — free text and otherSpecify columns only.

Open-end truncation toggles in Advanced (255-char default, configurable). Keeps SPSS variables under the long-string threshold.

Export templates

A saved column layout + filter + prefix-labels state. Every export off a template ships column-identical — the right unit for a tracker.

Picking a template loads the Datamap Designer side-by-side in the wizard; edit in place without leaving. Templates are created and managed in the Datamap Designer tab — see Templates →.

Companion long CSVs

The main export is always the wide file — one row per respondent, grids and lookups and card-sort and annotation folded in flat. Per-task and per-mark data ships as separate long CSVs in the same download pack:

cbc_long — one row per respondent × task × concept with choice/rating/rank response.
maxdiff_long — one row per respondent × task × item with best/worst flags.
cardsort_long — one row per card placement.
annotation_long — one row per heatmap mark, text span, or video reaction.

Conjoint and MaxDiff are long-only — no wide form, the shape doesn't fit. CBC ships with a cbc_design.csv companion (for re-running utilities in Sawtooth or R) and a Design Health HTML report in the same pack.

Export — Companion files panel with the CBC long CSV and design file card, side-by-side with the Files-in-this-export list showing the main .sav plus the cbc_long.csv, cbc_design.csv, and Design Health report rows

Partial exports

Three subset modes, stackable:

All data — the current filter set, no row-count cap.
By date range — bound the export to a response window.
By row count — first N rows of the current filter.

Plus a repeatable variable filter — where [field] [operator] [value], AND-ed together:

status   = complete
hExport  = 1
hRegion  in (1, 2, 3)
S4       >= 25

Filters run at export time against the live data — no new variable, no republish.

Partial export filter panel — All data / By date range / By row count, plus the variable filter row

→ Pattern: hExport as a quality flag

Add hExport as a hidden single-code with values 1–3, set from a script block: .set(1) include · .set(2) exclude · .set(3) hold. Filter every export to hExport = 1. When the quality logic changes, re-run the script and re-export; the next cut picks up the new calls.

Test & Live

The Export wizard reads either environment via ?env=:

Live (default) — published-link respondents from the main DB.
Test — Simulator respondents, materialized in real respondent shape so the same export engine runs unchanged.

Byte-identical for shared columns — the file you QA against sim data is the file your client downloads at close-of-field.

Every export logged

Recent exports table on the right of the wizard: format · file · respondents · created-at · size, with a download arrow on every row. Re-download any past export in one click, no regeneration.

Doubles as an audit trail (who exported what, when, against which cut). Files stored under survey_export_jobs — survive project renames, respondent recounts, template edits.

Export the way the analyst asked for.