What Are Generalized Linear Models?
Generalized Linear Models (GLMs) are a flexible class of statistical models that relate one or more independent variables to a single dependent variable. They extend ordinary linear regression to handle response variables that have non-normal error distributions.
GraphPad Prism supports GLMs through the multiple variables data table, allowing you to model continuous, count, and binary outcomes within a single interface.
Supported GLM Types in Prism
Multiple Linear Regression
When Y is continuous
e.g. blood pressure, enzyme activity, absorbance
Use multiple linear regression when your dependent variable (Y) is a continuous measurement. Prism fits a model of the form Y = b₀ + b₁X₁ + b₂X₂ + … and reports coefficients, confidence intervals, and goodness-of-fit statistics.
Poisson Regression
When Y is counts
e.g. 0, 1, 2, 3 — non-negative integers
Use Poisson regression when your dependent variable represents counts of events. It models the log of the expected count as a linear combination of the independent variables, making it ideal for count data such as the number of mutations, bacterial colonies, or rare events.
Logistic Regression
When Y is binary
e.g. yes/no, pass/fail, alive/dead
Use logistic regression when your dependent variable has only two possible outcomes. It models the log-odds of the outcome as a linear combination of predictors. Prism reports odds ratios, confidence intervals, and significance for each predictor.
How to Access GLMs in GraphPad Prism
- 1Create a new Multiple Variables data table in Prism.
- 2Enter your independent variables (X columns) and dependent variable (Y column).
- 3Click Analyze and select Multiple Variables analyses.
- 4Choose your GLM type: Linear, Poisson, or Logistic regression.
- 5Review coefficients, confidence intervals, and goodness-of-fit statistics in the results sheet.
Quick Comparison
| Model | Y variable type | Link function | Use case |
|---|---|---|---|
| Multiple linear regression | Continuous | Identity | Predict a measured value |
| Poisson regression | Count (0, 1, 2, …) | Log | Model event counts |
| Logistic regression | Binary (yes/no) | Logit | Predict probability of outcome |
Analyze Your Data with GraphPad Prism
GraphPad Prism makes it easy to run GLMs without writing code. Download a free trial or request a quote to get started.