An SPSS alternative that opens .sav files in your browser

If your university or employer already gives you SPSS, or your analysis needs multilevel models, structural equation models, complex samples or weighted cases, use SPSS: Research has none of those. If you need t-tests, ANOVA and MANOVA, correlation, linear, logistic, multinomial and ordinal regression, mediation and moderation, survival analysis, the non-parametric tests, reliability and factor analysis on a .sav, Stata, CSV or Excel file, with APA tables and the data never leaving your computer, Research does that as part of OBSCURA Pro, and codes the interviews in the same project.

What SPSS Statistics is

IBM SPSS Statistics is a desktop program for Windows and macOS and, in some versions, Linux. IBM sells it as a subscription for each user, by the month, the quarter or the year, with add-ons for further procedures such as custom tables, advanced statistics, forecasting and complex samples, and in other editions for businesses and for campuses. There is a free trial. Its data file is the .sav, and almost everything it does can be written and run as syntax.

It covers far more statistics than this page can list, many textbooks and courses are written around its menus and output, and many universities licence it for every student. Nothing on this page argues otherwise.

Side by side

SPSS’s entries are from IBM’s product, pricing and system requirements pages, checked on 6 October 2026. Its prices depend on the plan, the add-ons and the edition, and change, so none is given here; IBM’s pricing page has them.

What Research does, in SPSS’s terms

The menus are named as SPSS names them, and where SPSS has its own definition of a statistic (skewness and kurtosis, percentiles, Levene’s test, Type III sums of squares) that is the one used.

What SPSS has that Research does not

• Much of statistics. Research has no multilevel (mixed-effects) models, no structural equation models beyond confirmatory factor analysis, no multiple imputation, no stepwise regression, no complex samples, no weighting of cases, no Kolmogorov-Smirnov test of normality, and of PROCESS’s models only 1 and 4.
• Exact tests. Research gives exact p-values for Fisher’s test, the binomial and McNemar tests, and Mann-Whitney with small samples and no ties; the other non-parametric tests use the normal or chi-square approximation, as SPSS does without its Exact Tests module.
• Every kind of .sav. Numeric and string variables, labels, value labels, missing values, measures, formats and long strings are read and written. The portable .por format, multiple response sets, custom attributes and the full range of date formats are not, and a weighting variable is named but not applied.
• A licence you may already have. If your institution pays for SPSS, it costs you nothing to use it, and your supervisor, your course and your textbook probably use it too.
• Very large data sets. Hundreds of thousands of cases are slow to edit in a browser tab.

If what you need is SPSS’s syntax and files without a subscription, GNU PSPP is a free program that reads .sav files and runs much of the same syntax on the desktop.

What Research adds

• Nothing to install. It runs in a current browser on any computer, including one where you cannot install software.
• The data stay on the device. A project is kept in the browser’s private storage, files are read and written there, and nothing is sent to a server; there is no account. A data protection officer can check this in the browser’s network panel, which matters when an ethics approval says where the data may be kept.
• Checked figures. Every statistic, degree of freedom, p-value and effect size is compared, on every run of the site’s checks, with SciPy and statsmodels on several data sets, and agrees to six significant figures.
• The interviews in the same project. Transcripts and a survey’s open answers are coded beside the data, and counts from the coding become variables you can test. The NVivo page has the detail.
• One subscription for several jobs. The same Pro plan covers OBSCURA Write for the thesis itself, PDF Studio and the other paid applications.

Moving a data set across

1. Save the data from SPSS as a .sav file, compressed or not, or as .zsav.
2. In Research, open it. Variable names, labels, value labels, missing values, measures and formats come in, and the file’s character encoding is read.
3. Check any weighting and any date variables: a weight is named and not applied, and dates in the less common formats need a look.
4. Run the analysis from the menus; the Syntax view keeps what was run. To go back, save the data as .sav, which SPSS opens.

What it costs

Research on its own is US$2 a month or US$19 a year; all of OBSCURA Pro is US$5 a month or US$48 a year. Each comes with 7 days free first. For a team, each person needs a plan of their own. The price does not change with the procedures you use. There is no free version of Research with features held back: without Pro, its page shows what it does and asks for a plan.

See it before you pay. Research’s page shows it at work on its demo project, an invented evaluation of a community garden programme, and says what it cannot do.