What does an odds ratio of 2.5 mean?

That raising the predictor by one unit (or belonging to a given group vs. the reference) makes the odds of the outcome 2.5 times higher. A confidence interval containing 1 means a non-significant effect.

How do I judge whether the model classifies well?

By the AUC of the ROC curve (0.5 chance, above 0.8 a good model), sensitivity and specificity at the chosen threshold — not by overall accuracy alone.

Logistic regression

Definition

Logistic regression models the probability of a binary outcome — a patient responds to treatment or not, a customer churns or stays — as a function of explanatory variables. Coefficients are interpreted as odds ratios: by how much the odds of the outcome are multiplied per unit change of a predictor.

Regression & models

Unlike linear regression it predicts not a value but a probability between 0 and 1 through the logistic function. Judge model quality by the classification table, ROC curve and AUC, the Hosmer–Lemeshow test and pseudo-R²; with imbalanced classes (1 % churn) plain accuracy is misleading.

Variants: multinomial logistic regression for more than two categories, ordinal for ordered categories. It needs enough events per predictor (roughly 10 or more) and no strong multicollinearity.

In Statistica

Logistic regression is available in Statistics → Advanced Linear/Nonlinear Models → Generalized Linear/Nonlinear Models (binomial distribution, logit link) and in Nonlinear Estimation; the output includes coefficients, odds ratios with intervals, a classification matrix and the ROC curve. For large tables and automatic variable selection Data Mining also serves.

Available from edition: Statistica Desktop

Related terms

Knowledgebase guides

FAQ

What does an odds ratio of 2.5 mean?: That raising the predictor by one unit (or belonging to a given group vs. the reference) makes the odds of the outcome 2.5 times higher. A confidence interval containing 1 means a non-significant effect.
How do I judge whether the model classifies well?: By the AUC of the ROC curve (0.5 chance, above 0.8 a good model), sensitivity and specificity at the chosen threshold — not by overall accuracy alone.

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 Nonlinear regression Generalized linear models (GLM)Time series and forecasting Survival analysis All terms →

Updated: September 2026.