Standard Edition

Logistic regression with odds ratios and diagnostics Binary logistic regression with continuous and categorical predictors — odds ratios, Wald and likelihood ratio tests, leverage plots, residual diagnostics, and Cook’s D influence analysis.

Download 15-day trial Buy from US$ 155

Every feature from all five editions for 15 days, no sign-up.
Standard edition from US$ 155 a year · 30-day money-back guarantee.

Microsoft Excel with the Analyse-it tab selected, showing the binary logistic regression report for ICU patient survival: the Fit table of odds ratios with Wald 95% confidence intervals for 17 predictors, each categorical predictor against its reference level, the Effect of Model test beneath, and the Fit Model task pane open on Odds Ratios. Handwritten notes: Runs inside Excel: every analysis is on the Analyse-it tab; Fit Model: model specification, fit, odds ratios and effect tests; Odds ratio for each term with its Wald 95% CI; Likelihood ratio test of the model against the null; Terms, and the logit model; The report is an ordinary Excel worksheet: share it, archive it, open it on any PC with Excel.

Model binary outcomes with the same iterative workflow

Excel’s LINEST function and the chart trendline fit a straight line and stop. Fitted to a yes/no outcome — survived or not, responded or not, defective or not — the line predicts probabilities below zero and above one, and there is no odds ratio, no confidence interval and no test of whether a predictor matters. A table of coefficients is not what clinicians or stakeholders act on. They need odds ratios they can interpret and confidence intervals they can report.

Which predictors change the odds of the outcome, and by how much? Does the model as a whole beat the null model? Does adding a term improve the fit enough to justify keeping it? At what dose does half of the subjects respond?

Binary logistic regression with continuous and categorical predictors

Real outcomes depend on a mix of measurements and categories, not on one variable. Binary logistic regression takes continuous and categorical predictors in the same model, with simple, crossed, polynomial and factorial terms, and codes the dummy variables for each categorical predictor automatically. The model-building workflow is the same as linear regression: add or drop a term in the task pane, Recalculate, and see how the fit changes. A scatter plot shows the data.

  • Binary logistic regression
  • Probit regression with the same model terms and outputs new in v5.50
  • Continuous and categorical predictors
  • Simple, crossed, polynomial, and factorial terms
  • Automatic dummy variable coding
  • Scatter plot
Microsoft Excel showing the ICU survival logistic regression report from the top: the report title, N 200, and the odds ratio table for 17 predictors with reference levels for each categorical predictor, and the Fit Model task pane open on Fit listing the response STA and the 17 model terms. Handwritten notes: The logit equation, one term per predictor; Terms: 17 predictors, continuous and categorical; Each categorical term against its reference level.
The ICU survival model: 17 predictors listed in the Fit Model task pane, and the odds ratio table with each categorical predictor against its reference level.

Odds ratios with Wald confidence intervals and parameter estimates

An odds ratio is the number a clinician can act on. Each predictor is reported as an odds ratio with a confidence interval — directly interpretable as the change in odds per unit increase, or against the reference level of a categorical predictor. The underlying coefficients come as parameter estimates with standard errors and Wald Z statistics, with the covariance of estimates matrix and the model equation for documentation.

  • Odds ratios with confidence intervals
  • Odds ratios for all pairs or against a reference level, including models with interactions new in v6.01
  • Parameter estimates with standard errors and Wald Z statistics
  • Covariance of estimates matrix
  • Model equation

Likelihood ratio and Wald χ² tests for the model and each term

A predictor that does not contribute should be dropped, and the evidence for each one is in the report. The effect of model test shows whether the predictors collectively improve on the null model; the effect of terms test shows whether each term contributes or can be dropped, by likelihood ratio or Wald χ² test. Log-likelihood, AIC, BIC and deviance compare nested models — add a term and see whether the fit improves enough to justify the complexity.

  • Effect of model: likelihood ratio or Wald χ² test
  • Effect of terms: likelihood ratio or Wald χ² test for each term
Microsoft Excel showing the Effect of Model and Effect of Terms sections of the ICU survival logistic regression report: the likelihood ratio chi-squared test of the model against the null model, and the likelihood ratio test for each of the 17 terms with the significant p-values flagged, and the Fit Model task pane open on Effect of Terms. Handwritten notes: Likelihood ratio test of the whole model against the null; Likelihood ratio test for each term; Test: likelihood ratio or Wald.
Effect of Model and Effect of Terms for the ICU survival model: the likelihood ratio χ² for the whole model against the null, then for each of the 17 terms, with the five significant terms flagged.

Inverse prediction: X at a specified probability, with confidence interval

Often the question is not the probability at a given X but the X at which the outcome reaches a given probability. Inverse prediction gives that value with a confidence interval. For a dose-response study, that is the dose at which 50% of subjects respond. For a diagnostic assay, it is the concentration at which detection probability reaches 95%.

  • Inverse prediction: X for a given probability, with confidence interval new in v5.65
  • Multiple inverse predictions at once new in v5.60

Example analyses

See logistic regression results in detail — odds ratios, model tests, leverage plots, and influence diagnostics.

Fit Model 3 2 pages Binary logistic regression
Intensive care unit survival, 17 predictors. 200 observations. Odds ratios with Wald 95% CIs, G² likelihood ratio test for model and each term. Age, cancer, CPR, systolic BP, and admission type significant at 5%.

Part of the Standard edition

Logistic regression is one part of a complete statistical analysis toolkit. The Standard edition also includes simple and multiple regression, ANOVA and ANCOVA, PCA and factor analysis, descriptive statistics, hypothesis testing, correlation, and categorical data analysis. See everything in the Standard edition →

New to the model? Read logistic regression and odds ratios explained, including why a constant odds ratio does not mean a constant change in probability.

Validated, reliable, trusted for nearly 30 years

Validated calculations Every statistic tested against the NIST Statistical Reference Datasets, published datasets and thousands of internal test cases. No reliance on Excel’s built-in functions. See how we develop and validate Analyse-it →
Data stays on your PC No cloud processing, no uploads, no third-party access. Your data never leaves your computer — essential when working with sensitive, confidential, or patient-identifiable data.
Standard Excel workbooks Analyses are ordinary Excel workbooks that you can share with colleagues, archive for audit, and open on any machine with Excel — no Analyse-it licence required.
No formulas to break Results contain no formulas, so they cannot be accidentally edited or corrupted. The results you reported will be exactly what you find when you reopen the workbook.

Free trial and pricing

Try it on your own data first. The 15-day trial is every feature from all five editions, with no sign-up and no licence key — install it and start straight away.

Download 15-day trial See pricing or get a quote →

Standard edition from US$ 155 per year, or US$ 395 for a perpetual licence. Every purchase carries a 30-day money-back guarantee. Need a quote for purchasing? Add the licence to the cart and save it as a PDF quote.