What is the difference between GLM and GLZ?

GLM (general linear model) assumes a normal distribution and unifies regression with ANOVA; GLZ (generalized) additionally allows binomial, Poisson, gamma and other distributions through a link function.

When should I use Poisson regression?

To model counts of events per unit of time or space (failures, calls, accidents). When the variance is much larger than the mean, choose a negative binomial model.

Generalized linear models (GLM)

Definition

Generalized linear models extend linear regression to outcomes that are not normally distributed: binary (logistic regression), counts (Poisson regression), positive skewed values (gamma regression). They connect a linear combination of predictors with the expected value through a link function (logit, log) and estimate parameters by maximum likelihood.

Regression & models

General linear models (GLM in the narrower sense) unify regression and ANOVA: predictors may be continuous or categorical, designs unbalanced, with covariates and interactions. Generalized models (GLZ) add the choice of distribution and link function on top.

The choice of distribution follows the nature of the outcome: complaints per month → Poisson (or negative binomial under overdispersion), time to failure → gamma, yes/no → binomial. Compare quality by deviance, AIC and residual analysis.

In Statistica

Statistics → Advanced Linear/Nonlinear Models contains General Linear Models (GLM), Generalized Linear/Nonlinear Models (GLZ) and General Regression Models; the wizard offers the choice of distribution, link function, stepwise selection and type I–IV tests of effects. Output includes estimates, Wald tests, odds or incidence ratios and diagnostics.

Available from edition: Statistica Desktop

Related terms

Knowledgebase guides

FAQ

What is the difference between GLM and GLZ?: GLM (general linear model) assumes a normal distribution and unifies regression with ANOVA; GLZ (generalized) additionally allows binomial, Poisson, gamma and other distributions through a link function.
When should I use Poisson regression?: To model counts of events per unit of time or space (failures, calls, accidents). When the variance is much larger than the mean, choose a negative binomial model.

Try it on your own data

Statistica free for 30 days

Full version, no credit card. Or get a pricing estimate in a minute.

Try free for 30 days Pricing estimate

Regression & models

Linear regression Logistic regression Nonlinear regression Time series and forecasting Survival analysis All terms →

Updated: September 2026.