A .sav file converter to Word that asks which columns first.
Drop an SPSS .sav, tick the variables you want, and watch where the page will break before you commit. The download is a genuine .docx, assembled in this browser tab.
- A Word page has a fixed width, so a .sav with 40 variables cannot become one readable table. This page makes you choose.
- 16 columns fit A4 portrait, 24 fit A4 landscape, measured from the page geometry rather than guessed.
- Value labels or numeric codes in the cells; variable names or variable labels in the heading row, chosen separately.
- Nothing is uploaded, so there is no file size ceiling and no queue.
Why the picker comes first
A sav file converter to Word carries one constraint the other formats on this site do not, and it is the reason this page is built back to front. Every other export can afford to ignore how wide your dataset is. A CSV has no page. A spreadsheet scrolls sideways forever. Word does not: a document is a stack of fixed-size sheets, and a table with more columns than the sheet can hold does not scroll, it gets clipped or squeezed into unreadable slivers.
So the order of operations is inverted here. Instead of converting everything and letting you discover the damage in Word, the tool reads the file, tells you how many columns this page size can carry, and hands you the checklist. The preview underneath is drawn from the same width arithmetic that goes into the document, so the shape you see is the shape you get.
Choosing columns is also choosing what the table means
A 60-variable survey exported whole is not a table anyone reads. Six variables, headed with the actual question text, is a table someone can look at once and understand. The picker is a formatting control on paper and an editorial one in practice.
What lands in the .docx
A title line with the file name, one line stating the row count, how many of the file's variables you kept, and whether the cells hold labels or codes, then the table. The heading row is bold, shaded, and marked to repeat at the top of every page the table spans, so page four is still readable without scrolling back to page one.
Getting an SPSS to Word table with labels resolved
A variable with value labels stores a number and displays text. In the cells you pick which of the two you want. In the heading row you pick again, separately: the short variable name that matches your syntax, or the variable label that a reader can follow. Two variables sharing a label get numbered rather than duplicated, because two identically headed columns cannot be read back against the dataset.
Where the .docx comes from
A .docx is a zip of XML parts. This one holds three: the content types, the package relationships, and word/document.xml. Microsoft's own description of the format, quoting ISO/IEC 29500, puts the floor exactly there: At a minimum, to create a valid WordprocessingML document using code, add a main document part to the document.
That is why there is no document library in the dependency list and no styles part in the file. Fewer moving pieces, and the zip is built with the browser's own compression.
Four layouts that work
Nothing here is a saved preset you can click, because the tool never sees your file and cannot store anything. These are the four combinations that come up most, with the column budget each one has.
The width limit, in inches
A4 is 8.27in across. Take 0.75in of margin off each side and 6.77in of text width is left. The narrowest column this tool will lay out is 0.42in, which holds about five characters of the 9pt body font before it wraps mid-word. Divide one by the other and A4 portrait carries 16 columns. Landscape carries 24, Letter portrait 16, Letter landscape 22.
Past that the tool stops rather than shipping a broken document, and says which two things to change. Below the ceiling but close to it, some columns sit at that 0.42in minimum and their text wraps over several lines; the widget counts those and tells you, because a table that is technically correct and four times taller than you expected is still a surprise.
Length is a judgement call, not a cap
Rows are cheaper than columns. Building 50,000 rows across 10 columns took about 2.3 seconds here and produced a 2MB file, and a stricter OOXML parser than Word reads it back with all 50,001 rows intact. The honest gap is Word itself: a table estimated at 1,100 pages is beyond what has been tested on real Office installs, so the tool says so above 5,000 rows instead of implying a number it verified. Default is 200 rows, and every document prints the range it contains.
When this sav file converter to Word is the wrong export
If the next step is analysis, take the CSV export or the Excel workbook instead. Word stores text and layout, and pulling numbers back out of a document table is tedious and lossy. If the deliverable is a printable record of the whole file rather than a table you will edit, the PDF export writes a full variable codebook alongside the data and does not ask you to drop columns.
A sav file converter to Word earns its place when the table is going to be edited: dropped into a report, given a caption, reformatted by a co-author, tracked with comments. That is the job no other format on this site does.
If the file will not open
Four unrelated problems used to arrive as one message. They are separated now, because a password-protected dataset and a game save that happens to end in .sav need opposite advice.
| Cause | What you get told |
|---|---|
| The file reads as 0 bytes | It came through empty, and the likely reason is that it lives in iCloud, OneDrive or Google Drive and has not been downloaded to the device yet. |
| SPSS password protection | The file is encrypted, so its contents cannot be read. The keyword sits eight bytes in rather than at the start, which is why so many tools call this "not a valid file" instead. The fix is a copy saved without a password. |
| A zip, .xlsx or .spv that was never extracted | It is a zip container, not a .sav. Extract it and pick the file inside. An .spv holds output tables, not the dataset. |
| A game save that happens to use .sav | It is an Unreal Engine save from a game, not an SPSS dataset. Same extension, unrelated format, and a save editor is the tool for it. |
A file that is text or HTML, a gzip that was never decompressed, and a file with no SPSS header at all each get their own message too. To look at a file's structure before exporting anything, open it in the .sav viewer, or read how to open a .sav file without SPSS.
Questions
Is the download a real .docx, or a renamed HTML file?
A real one. The download is an Open Packaging Convention zip holding [Content_Types].xml, the package relationships, and word/document.xml, which is the WordprocessingML main document part with the table written as w:tbl, w:tr and w:tc elements. The test that ships with the site builds one from the sample file, unzips it with an unrelated zip implementation, and checks the XML parses and the table rows match the data. Plenty of converters write an HTML table, name it .doc and let Word guess; that file cannot be styled, cannot repeat a heading row, and breaks differently in Google Docs.
How many variables fit in the table?
Sixteen on A4 portrait, twenty four on A4 landscape, sixteen on Letter portrait and twenty two on Letter landscape. That is arithmetic, not a guess: A4 portrait leaves 6.77in of text width between 0.75in margins, and below 0.42in a column holds roughly five characters before it wraps mid-word. The widget shows the number for whatever page you pick, preselects that many variables, and refuses to build a document wider than that instead of handing you a table with columns off the edge of the page.
Can I convert spss sav to word with the value labels instead of the codes?
Yes, and that is the default. A variable with labels defined for 1 and 2 writes "Strongly agree" rather than 1. Switch the cell values control to numeric codes if the document is going to someone re-entering the data. The column heading has its own control, separate from the cells: variable names for something you will read against a syntax file, variable labels for something a reader has to understand on its own.
How do I open an SPSS file in Word directly?
You cannot. Word has no import filter for the .sav binary format, so double-clicking a .sav with Word set as the handler gets you either an error or a page of mojibake. Something has to read the SPSS format and write a document, which is what this page does in the browser tab. The same is true of the reverse question about SPSS output: an .spv file is a zip of tables and charts, not the dataset, and it needs SPSS itself or its output viewer.
How long can the document be?
The row count is your choice, defaulting to the first 200. Building is not the constraint: 50,000 rows across 10 columns took about 2.3 seconds and came out at 2MB when measured here, and a stricter OOXML parser than Word read the result back correctly. What has not been tested is how a given copy of Word behaves on a 1,100-page table, so above 5,000 rows the widget says exactly that rather than pretending to know.
Does the file get uploaded anywhere?
No. The .sav is read in your browser and the .docx is assembled there, including the zip container, so the bytes never reach a server. No account, no email, no size limit imposed from the outside. Nothing about the file reaches analytics either, which is why the page cannot show you a gallery of other people's documents.
Other formats from the same file
Same parser, different destination.