Skip to main content
v2026.11,986 entries · CC-BY 4.0

Simple Regression in SPSS: Steps, Output, and Reporting

Run a simple regression in SPSS with Analyze > Regression > Linear, then read R square, the ANOVA and the slope using two published worked examples.

Written and maintained by CASRAI Editorial Board

Published

On this page

To run a simple regression in SPSS, choose Analyze > Regression > Linear, put the outcome variable in the Dependent box and the single predictor in the Independent(s) box, and click OK. SPSS returns a Model Summary (R, R square, adjusted R square), an ANOVA table and a Coefficients table, and the unstandardized B for your predictor is the slope of the fitted line. This guide covers the one-predictor case: the dialog options that matter, the equivalent REGRESSION syntax, how to read each table using two published worked examples, and what to do when the result looks wrong.

What SPSS is used for in research

IBM SPSS Statistics is a commercial statistical analysis package that researchers use for data management, descriptive statistics, hypothesis tests and regression through menus or command syntax. Simple linear regression is one of its standard procedures: it models one continuous outcome as a straight-line function of one predictor, which is the right tool when the research question is how much the outcome changes per unit of the predictor, or how much of its variance that predictor accounts for.

The Linear Regression procedure sits under Analyze > Regression and requires the Statistics Base module, according to IBM’s documentation of the Save dialog. The same dialog is used for multiple regression; with a single predictor, several of its options (partial plots, collinearity diagnostics, entry methods) have little to do. For the multi-predictor case see multiple regression in SPSS.

How to install or access SPSS

IBM’s SPSS Statistics product page offers a free trial through an IBM account, purchase through IBM.com or AWS Marketplace, and several licence types: subscription plans per authorized user, perpetual licences, campus editions for educational institutions, and single-user student licences sold through IBM-authorized vendors. Many universities supply SPSS through a campus licence, so check with your institution’s research computing or library service before buying. A comparison of the plan structure is in our guide to the IBM SPSS Statistics subscription. Prices are set by IBM and are not repeated here.

The Linear Regression procedure belongs to the Statistics Base edition, so it is available in the base package without an add-on module.

How to run a simple regression in SPSS step by step

Before you start: look at the scatterplot

A simple regression fits a straight line, so first check that a straight line is a sensible summary. SPSS-Tutorials draws the scatterplot through Graphs > Legacy Dialogs > Scatter/Dot and adds a fit line in the Chart Editor. The Chart Builder route is described in scatterplots in SPSS Chart Builder. A curved cloud of points is a warning that a linear model will misdescribe the data.

  1. Open your data file and choose Analyze > Regression > Linear.
  2. Move the outcome variable into the Dependent box.
  3. Move the single predictor into the Independent(s) box. With one predictor the method is Enter, which is what the generated syntax uses (/METHOD=ENTER).
  4. Click Statistics. IBM documents these options: Regression Coefficients (B, standard error, Beta, t and two-tailed significance, with optional confidence intervals), Model fit (variables entered and removed, R, R square, adjusted R square, standard error of the estimate and the ANOVA table), R squared change, Descriptives, Part and partial correlations, Collinearity diagnostics, and under Residuals the Durbin-Watson statistic and casewise diagnostics. Tick the confidence intervals option if you want to report an interval for the slope.
  5. Click Plots and request a plot of the standardized residuals (*ZRESID) on the Y axis against the standardized predicted values (*ZPRED) on the X axis, plus a histogram and normal probability plot of the residuals. IBM notes that ZRESID against ZPRED helps check linearity and equal variances.
  6. Optionally click Save to write new variables to the active data file. IBM lists predicted values, distances (Mahalanobis, Cook’s D, leverage values), prediction intervals (mean or individual, with a confidence level from 1 to 99.99), residuals (unstandardized, standardized, studentized, deleted, studentized deleted) and influence statistics (DfBeta, DfFit, covariance ratio). Each selection adds one or more variables to the data file.
  7. Click OK. Output appears in the Output Viewer.

The equivalent syntax

Pasting the dialog into a syntax window gives a REGRESSION command. UCLA’s annotated example for predicting api00 from enroll uses the shortest form:

REGRESSION /DEPENDENT api00 /METHOD=ENTER enroll.

SPSS-Tutorials shows a fuller version that adds coefficient confidence intervals, a residual scatterplot and a residual histogram:

REGRESSION /MISSING LISTWISE /STATISTICS COEFF OUTS CI(95) R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT performance /METHOD=ENTER iq /SCATTERPLOT=(*ZRESID ,*ZPRED) /RESIDUALS HISTOGRAM(ZRESID).

Saving the syntax with your data is the practical way to make the analysis reproducible; see reading SPSS output for working with the Output Viewer.

How to read the output

The data table below lists the dialog options and what IBM documents for each. This section uses two published worked examples to show how the numbers fit together.

Quantity UCLA OARC example (api00 predicted from enroll) SPSS-Tutorials example (performance predicted from IQ, N = 10)
Regression equation api00 = 744.25 – 0.20 x enroll performance = 34.26 + 0.64 x IQ
R square .101 .403
Significance of the model or slope F = 44.829, Sig. .000 p = .049
95% CI for the slope B not shown in the annotated output .004 to 1.281

Model Summary

R is the correlation between the observed and predicted values of the outcome, and R square is the proportion of variance in the outcome associated with the predictor. In UCLA’s example R is .318 and R square .101, so enrollment accounts for about 10 percent of the variance in api00; adjusted R square is .099 and the standard error of the estimate is 135.026. Report R square, and read adjusted R square as the version corrected for model size.

ANOVA table

The ANOVA table tests whether the model as a whole explains more variance than a model with no predictor. UCLA reports F = 44.829 with Sig. .000. With a single predictor this test addresses the same question as the test of the slope in the Coefficients table.

Coefficients table

The B for the constant is the predicted outcome when the predictor is zero, and the B for the predictor is the change in predicted outcome per one-unit change in the predictor. In UCLA’s output the predicted api00 is 744.25 minus 0.20 per enrolled student; in SPSS-Tutorials the predicted performance is 34.26 plus 0.64 per IQ point. The Beta column is the standardized coefficient. In UCLA’s output Beta is -.318, the same magnitude as R of .318 with the sign carried by the slope. For a deeper treatment of what a coefficient means, see what a regression coefficient is.

Confidence intervals and sample size

The SPSS-Tutorials example has only 10 cases. Its slope is significant at p = .049, but the 95 percent interval for B runs from .004 to 1.281, which shows how little a small sample pins down the slope. Report the interval alongside the p value rather than the p value alone.

Residual checks

Use the ZRESID against ZPRED plot to judge linearity and equal variances, as IBM describes. SPSS-Tutorials notes a curvilinear pattern in its example residuals and suggests Analyze > Regression > Curve Estimation as a follow-up check. The assumptions themselves, and what to do when they fail, are covered in assumptions of simple regression.

Troubleshooting SPSS

  • Linear is missing from Analyze > Regression. IBM states that the Linear Regression procedure requires the Statistics Base edition, so check which edition and modules your licence includes.
  • A partial plot option is greyed out or unavailable. IBM states that partial plots need at least two predictors, so they cannot be produced for a simple regression.
  • Curved residual plot. A systematic curve in ZRESID against ZPRED means a straight line does not describe the data well; SPSS-Tutorials uses the Curve Estimation dialog to compare alternatives.
  • Wide confidence interval for the slope. Often a sample-size problem, as in the N = 10 example above, where the interval spans nearly zero to 1.281.
  • Cases dropped from the analysis. The REGRESSION syntax shown by SPSS-Tutorials includes /MISSING LISTWISE, which excludes any case missing the outcome or predictor.
  • Extra variables appear in your data file. They come from the Save dialog; IBM states that each selection adds new variables to the active data file.

Free alternatives to SPSS

Software What its own site says
GNU PSPP Free software under GPLv3 or later, described as a replacement for SPSS that looks very similar to it; lists linear and logistic regression among its analyses; syntax and data files are compatible with SPSS; no licence fees or limits on cases or variables.
jamovi A free and open statistical spreadsheet built on R, which shows the R syntax behind each analysis; its home page shows linear-regression results beside the data.
JASP A free, open-source, cross-platform program supported by the University of Amsterdam with a graphical interface; offers standard procedures in classical and Bayesian form.

Check each program’s own documentation for how it labels the regression output, since the table names above are SPSS’s.

Sources

Data: SPSS Linear Regression dialog options documented by IBM (SPSS Statistics 32.0.0)

Compiled from three IBM Documentation pages on the Linear Regression Statistics, Plots and Save dialogs: each option group and the items IBM lists for it, plus IBM's stated constraint where one is given.

Period covered: 2026.

SPSS Linear Regression dialog options documented by IBM (SPSS Statistics 32.0.0)
OptionDialog buttonItems IBM documentsIBM stated constraint or use
Regression CoefficientsStatisticsB, standard error, Beta, t, two-tailed significance; optional confidence intervalsNone stated
Model fitStatisticsVariables entered and removed, R, R square, adjusted R square, standard error of the estimate, ANOVA tableNone stated
Collinearity diagnosticsStatisticsCollinearity diagnosticsNone stated
Residuals (Statistics)StatisticsDurbin-Watson, casewise diagnosticsNone stated
Scatterplot variablesPlotsDEPENDNT, ZPRED, ZRESID, DRESID, ADJPRED, SRESID, SDRESIDZRESID against ZPRED helps check linearity and equal variances
Standardized residual plotsPlotsHistogram, normal probability plotNone stated
Partial plotsPlotsPartial plotsNeed at least two predictors
Predicted values and distancesSaveUnstandardized, standardized, adjusted predicted values, SE of mean predictions; Mahalanobis, Cook's D, leverageEach selection adds variables to the active data file
Prediction intervalsSaveMean and individual intervalsConfidence level 1 to 99.99
Residuals and influenceSaveUnstandardized, standardized, studentized, deleted, studentized deleted residuals; DfBeta, DfFit, covariance ratioEach selection adds variables to the active data file

Sources

Compiled by CASRAI from the sources above. Free to reuse under CC BY 4.0 with attribution.

Follow CASRAI

Research-administration guidance, standards updates and independent tool reviews.

Ask CASRAI · free to try

Ask about Simple Regression in SPSS: Steps, Output, and Reporting

2 free questions a day, no account.

An AI assistant for research: funder rules, compliance, methods and publishing. It links the sources behind its answers and labels answers that come from a web search.

Answers draw on CASRAI's guides and dictionary plus the federal and funder documents we index: Federal Register, Grants.gov, Regulations.gov and UKRI.

Works on this site and inside Claude, Cursor and the AI tools you already use.

Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays open to read, with no account and no card.

Referenced across the research world

University of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logoUniversity of Cambridge logoColumbia University logoCrossref logoUniversity of Edinburgh logoHarvard University logoUniversity of Oxford logoPrinceton University logoStanford School of Medicine logoUniversity College London logoORCID logo
  • University of Cambridge logo
  • Columbia University logo
  • Crossref logo
  • University of Edinburgh logo
  • Harvard University logo
  • University of Oxford logo
  • Princeton University logo
  • Stanford School of Medicine logo
  • University College London logo
  • ORCID logo

View CASRAI adoption →