SPSS command: SPSS LOGLINEAR
Loglinear and logit models with a custom design
The legacy LOGLINEAR command fits loglinear and logit models to cross-tabulations with any design you specify.
Free · No account · Files stay on your device · Updated 18 October 2026
What it tells you
Crosstabs runs it on your device with deviation contrasts, goodness-of-fit tests, standardized and adjusted residuals and, for logit models, Haberman's analysis of dispersion.
- Whether a chosen set of effects reproduces the table (likelihood-ratio and Pearson chi-square).
- Deviation-contrast parameters for every effect, with standard errors and confidence intervals.
- For logit models, entropy and concentration measures of how much of the response's dispersion the factors explain.
When to use it
Use it when
- You want a specific non-hierarchical or custom design rather than K-way searches.
- You model a categorical response from categorical factors as a logit model and want dispersion measures.
Use something else when
- You want automatic backward elimination: use hierarchical loglinear analysis.
- Your predictors are continuous: use multinomial logistic regression.
How to run it in Crosstabs
- 01
Open your data
Open a .sav, CSV or Excel file in the workspace and switch to Statistics mode. The file is read in your browser and never uploaded.
- 02
Choose Loglinear and logit models (legacy)
In the General linear model group, pick Loglinear and logit models (legacy).
- 03
Pick the variables
Choose a Dependent variable (logit model) or None (general loglinear model), and tick the Factors or Variables.
- 04
Specify the design
Keep Saturated design, or untick it and type the Design effects, separated by commas, with BY or * for interactions.
- 05
Run
Click Run and read the goodness-of-fit, parameter, residual and dispersion tables.
Output tables explained
- Goodness-of-Fit Tests
- Likelihood-ratio and Pearson chi-square with df = I(J − 1) − p.
- Parameter Estimates
- Deviation-contrast estimates: each compares a category with the average over categories.
- Analysis of Dispersion
- Entropy and concentration dispersion due to the model, the residual and the total, with association measures and chi-square tests.
Assumptions and notes
- Counts are multinomial within each combination of the factors; 0.5 is added to every cell of a saturated model.
- Expected counts should mostly exceed 5 for the chi-square tests to be reliable.
How the results are validated
R glm Poisson fits with one constant per factor combination, run in Docker, with deviation codes cross-checked against contr.sum, match estimates, standard errors, fitted counts, adjusted residuals and both chi-squares within 1e-7, and Haberman's dispersion within 1e-7. Validation evidence
Questions
How does LOGLINEAR differ from GENLOG?
LOGLINEAR takes a free-form design and reports deviation contrasts; GENLOG offers saturated, two-way and main-effects designs with last-category-redundant parameters. Both report the analysis of dispersion for logit models.
Why are some effects ignored in a logit model?
Effects without the dependent variable are absorbed by the constants that fix each factor combination's total, so they carry no information about the response.
Will results match SPSS to every decimal?
Estimates are iterated to full convergence; SPSS LOGLINEAR stops at a looser default criterion, so its last digits can differ slightly.
What do the entropy and concentration measures mean?
They are the share of the response's dispersion explained by the factors, like an R² for categorical responses.
Related procedures
- General Loglinear AnalysisGENLOG
- Hierarchical Loglinear AnalysisHILOGLINEAR
- Multinomial Logistic RegressionNOMREG
- Crosstabs StatisticsCROSSTABS /STATISTICS
- Free SPSS alternative: full coverage list
Run Loglinear and Logit Models (LOGLINEAR) on your own data
Open a .sav, CSV or Excel file. Everything runs in your browser; the file never leaves your device.
SPSS is a trademark of IBM Corp. Crosstabs is not affiliated with or endorsed by IBM.