Mixed Models

Jointly Model the Mean and Covariance

The Mixed Models personality of the Fit Model platform is available only in JMP Pro.

In JMP Pro, the Fit Model platform’s Mixed Model personality fits a wide variety of linear models for continuous responses with complex covariance structures. These models include random coefficients, repeated measures, spatial data, and data with multiple correlated responses. Use the Mixed Model personality to specify linear mixed models and their covariance structures conveniently using an intuitive interface, and to fit these models using maximum likelihood methods.

The modeling results are supported by interactive visualization tools such as profilers, surface plots, and contour plots. You can use these tools to complement your understanding of the model.

Figure 8.1 Marginal Model Profiler for a Split Plot Experiment 

Contents

Overview of the Mixed Model Personality

Example Using Mixed Model

Launch the Mixed Model Personality

Fit Model Launch Window

Data Format

The Fit Mixed Report

Fit Statistics

Random Effects Covariance Parameter Estimates

Fixed Effects Parameter Estimates

Repeated Effects Covariance Parameter Estimates

Random Coefficients

Random Effects Predictions

Fixed Effects Tests

Sequential Tests

Multiple Comparisons

Compare Slopes

Marginal Model Inference

Actual by Predicted Plot

Residual Plots

Profilers

Conditional Model Inference

Actual by Conditional Predicted Plot

Conditional Residual Plots

Conditional Profilers

Variogram

Covariance and Correlation Matrices

Save Columns

Additional Examples of the Mixed Models Personality

Repeated Measures Example

Split Plot Example

Spatial Example: Uniformity Trial

Correlated Response Example

Statistical Details for the Mixed Models Personality

Convergence Score Test

Random Coefficient Model

Repeated Measures

Repeated Covariance Structures

Spatial and Temporal Variability

The Kackar-Harville Correction