Fit simple and curve regression models
Fit the relationship between two variables using the model that best describes the data — linear, polynomial up to 6th order, logarithmic, exponential, power, or probit. Transform variables, test for lack of fit, and visualise the result with a scatter plot, fit line, and confidence bands.
Build complex models with multiple predictors
Combine continuous and categorical predictors in a single model — with simple, crossed, polynomial, and factorial terms. Categorical variables are coded as dummy variables automatically. Build full factorial, ANOVA, and ANCOVA models with adjusted effect means and five multiple comparison procedures including Tukey-Kramer, Dunnett, and Scheffé.
Model binary outcomes with logistic regression
Fit binary logistic or probit regression models with continuous variables, categorical factors, and interactions. Get parameter estimates, odds ratios with confidence intervals, Wald and likelihood-ratio tests, AIC/BIC, and the same leverage plots and influence diagnostics as linear regression.
Refine the model interactively
Add or remove terms, change the model structure, and click Recalculate. The results update immediately in the same workbook — fit statistics, diagnostic plots, hypothesis tests, all of it. No re-importing data, no starting a new analysis from scratch. Build the model incrementally until you're satisfied with the fit.
Assess each predictor's contribution
Use sequential (Type I) and partial (Type III) hypothesis tests to understand what each term adds to the model. Leverage plots show each predictor's effect after accounting for all others. VIF flags multicollinearity. Standardised betas let you compare predictor importance on a common scale.
Diagnose model fit
Check whether the model assumptions hold before you act on the results. Raw and standardised residual plots, sequence and lag plots, and residual distribution plots reveal structural problems. The Outliers & Influence plot using Cook's D and Studentised Residuals shows whether individual observations are driving the model.
Save computed variables and predict
Save fitted values, residuals, standardised residuals, studentised residuals, leverage, and Cook's D back to the dataset for further analysis. Predict new observations directly from the fitted model — with confidence intervals for the mean response and prediction intervals for individual observations.
