SPSS is a statistical analysis software package, short for Statistical Package for the Social Sciences, used to enter, clean, and run statistical tests such as t-tests, ANOVA, and regression on quantitative research data. First released in 1968, it has been sold as IBM SPSS Statistics since IBM’s 2009 acquisition.
SPSS is a statistical analysis software package used to enter, clean, and run statistical tests on quantitative data. Its name is an acronym for Statistical Package for the Social Sciences, reflecting its origin in survey and social-science research, though its actual use today extends well beyond that field. The software was first released in 1968, developed by Norman H. Nie, C. Hadlai Hull, and Dale H. Bent, and the company behind it, SPSS Inc., was acquired by IBM in 2009; releases since roughly 2015 carry the product name IBM SPSS Statistics. This page explains what SPSS actually does as a research tool, who uses it, and how it fits into a broader research data management workflow — a research-methods and data-stewardship orientation, not a product pitch.
What SPSS does
SPSS is built around a spreadsheet-like Data Editor and a menu-driven interface for running statistical procedures without writing code, alongside an equivalent command-syntax language (.sps files) for the same operations. The main things researchers use it for:
Data entry, import, and labeling
SPSS reads data from CSV, Excel, database connections, and other statistical packages’ formats, and stores its own working files in a proprietary .sav format. A defining feature of the Data Editor is the separation between a variable’s underlying numeric or string value and its label — for example, storing survey responses as 1/2/3 while displaying and exporting them as “Disagree/Neutral/Agree.” This variable-and-value-label structure functions as an informal, machine-readable codebook attached directly to the dataset.
Descriptive and inferential statistics
Base SPSS covers frequencies, cross-tabulations, and descriptive statistics; t-tests, one-way and factorial ANOVA, chi-square tests, correlation, and linear and logistic regression; and a range of nonparametric tests. IBM sells additional modules on top of the base package — Complex Samples (survey-weighted analysis), Missing Values (multiple imputation), Bootstrapping, Advanced Statistics (mixed models, generalized linear models), and SPSS Amos (structural equation modeling) — that extend it toward more specialized quantitative methods.
Syntax versus point-and-click
Every menu-driven procedure in SPSS can be run, and is often better run, from syntax: either by pasting the syntax SPSS generates behind each dialog box, or by writing it directly. This distinction matters for research data management more than it might first appear — see the workflow section below.
Output management
Results are written to the Output Viewer as a navigable, exportable log (tables, charts, and syntax echoes) that can be saved as an .spv file or exported to Word, PDF, or HTML for inclusion in a manuscript or report.
Who uses SPSS
SPSS’s core user base remains close to its name: psychology, sociology, political science, education research, and survey methodology, where its menu-driven interface lowers the barrier for researchers who are statistically literate but not necessarily programmers. It is also widely used in health, clinical, and public-health research — epidemiology, nursing research, and health-services research routinely rely on it for analyzing survey and clinical outcome data — and in market research and government/official statistics work. University site licenses mean SPSS is often the first statistical package graduate students in these fields are taught, which reinforces its continued use even where a lab could equally use an open-source alternative.
It sits in a different niche from a qualitative data analysis tool like NVivo: SPSS analyzes numeric, structured data (survey responses, experimental measures, clinical variables), while NVivo organizes and codes unstructured data such as interview transcripts. Many mixed-methods projects use both, on different halves of the same study.
How SPSS fits into a research data management workflow
SPSS sits in the middle of the data lifecycle — after data collection and before publication — and several of its features have direct research data management consequences worth planning for explicitly, ideally at the data management plan stage rather than after analysis is already underway.
Syntax as a reproducibility record
A point-and-click SPSS session, run entirely through menus with no saved syntax, leaves no re-runnable record of what was actually done to the data — the opposite of the auditable workflow that computational reproducibility requires. Saving (or pasting and keeping) the underlying .sps syntax turns an SPSS analysis into something closer to a script: a plain-text, versionable, re-runnable record of every recode, filter, and test applied to the data, which can be shared alongside a dataset or included as supplementary material.
Variable and value labels as informal metadata
The variable/value-label structure described above is a real, if informal, form of dataset documentation — comparable in function, though not in standardization, to a formal codebook such as the DDI Codebook standard used by social-science data archives. It is worth exporting or replicating this labeling into a standalone codebook when a dataset is deposited somewhere that won’t preserve SPSS’s internal metadata.
Archiving and format considerations
.sav is a proprietary binary format tied to a specific software vendor, which repositories and funders increasingly flag as a preservation risk for long-term deposit. Many research data repositories and funder data-sharing policies expect data to also be deposited in, or convertible to, an open, non-proprietary format — typically CSV or tab-delimited text plus a separate codebook, since a flat text file alone loses the value labels and variable metadata SPSS stores internally. Planning for this conversion before deposit, rather than after, avoids losing that labeling information.
Working with sensitive data
Because SPSS is heavily used in health and social-science research, it frequently holds data that requires de-identification or is governed by a data sharing agreement before it can be shared or deposited — SPSS itself has no built-in de-identification tooling, so this remains a step a researcher or data steward has to plan and execute separately, before export.
Licensing and access
IBM offers SPSS Statistics through both subscription tiers (Base, Standard, Professional, and Premium, the last bundling SPSS Amos) and a perpetual-license option, with feature sets and pricing that have changed over time and vary by tier and region — check IBM’s current SPSS Statistics product page directly before budgeting for it, rather than relying on a cached figure. In practice, most researchers access SPSS through an institutional or departmental site license administered by their university’s IT department rather than purchasing it individually.
SPSS versus alternatives
SPSS is one of several statistical packages a researcher might reasonably choose between, and the right choice depends on the field, the analysis, and how much programming the researcher is comfortable with:
- R and Python are free, open-source, and fully scriptable, which makes them the default choice in fields (bioinformatics, much of applied econometrics, data-science-adjacent social science) that prioritize fully reproducible, code-first workflows; the tradeoff is a steeper learning curve than SPSS’s menu-driven interface.
- Stata occupies a similar niche to SPSS — commercial, licensed, with both a GUI and a scripting language (do-files) — and is particularly common in economics and epidemiology.
- PSPP is a free, open-source program designed to be broadly compatible with SPSS syntax and file formats, offered as a no-cost alternative for a subset of SPSS’s core statistical procedures.
- NVivo, as noted above, is not a competitor to SPSS at all — it analyzes qualitative rather than quantitative data, and the two are often used together on a mixed-methods project.
None of these is a strict upgrade or downgrade of another; the comparison that usually matters in practice is qualitative versus quantitative research design itself, which determines whether a quantitative package like SPSS is the right category of tool at all.
Frequently asked questions
What does SPSS stand for?
Statistical Package for the Social Sciences. The name dates to the software’s 1968 origin in survey and social-science research; current releases are branded IBM SPSS Statistics, reflecting IBM’s 2009 acquisition of SPSS Inc., but “SPSS” remains the name almost everyone actually uses.
Is SPSS free to use?
No — SPSS is commercial software, available by subscription or perpetual license from IBM, though many researchers access it at no direct cost to themselves through a university or institutional site license. PSPP is a free, open-source alternative designed for broad compatibility with SPSS’s file format and syntax, if a no-cost option is needed.
What is SPSS used for in research?
Analyzing quantitative data: cleaning and coding survey or experimental data, running descriptive and inferential statistics (t-tests, ANOVA, regression, chi-square, and more), and producing tables and charts for publication. It is most associated with social science, psychology, education, and health research, though it is used wherever structured, numeric research data needs statistical analysis.
What file format does SPSS use?
.sav for datasets (including variable and value labels) and .sps for syntax files recording the commands run against a dataset. Both are proprietary to SPSS, which is why converting data to an open format such as CSV, alongside a separate codebook, matters before long-term archiving or repository deposit.
Is SPSS’s point-and-click interface a problem for reproducibility?
Only if the syntax isn’t saved. A menu-driven SPSS session with no retained .sps file leaves no re-runnable record of the analysis; saving or pasting the generated syntax turns the same analysis into a shareable, versionable script, comparable in function to an R or Python analysis script.
How is SPSS different from NVivo?
SPSS analyzes quantitative, numeric data with statistical tests; NVivo organizes and codes qualitative, unstructured data such as interview transcripts. They serve different halves of a mixed-methods study rather than competing for the same task.
