Lab6: LME Three-level Model for Clustered Data
Hierarchical Model Specification (step-up strategy)


Getting Started

To Start R: Click on the R icon

If no R icon, go to Start and then All Programs to find R


Example: Chapter 4 Dataset class

Here is a description of the data: classroom_des.txt

What does the data look like and what is read.csv?
Here is the data classroom.csv
Read the data into the class object and also,

> library(nlme)
> help(lme)

What is the default method?

We will use the textbook web page , Click on Chapter 4 and then the R Syntax for Descriptive Statistics

Let's make boxplots for (Figure 4.2).

What do you notice?

Recall, Table 4.1 has Level of Data: Level 1 is Student, Level 2 is Classroom, Level 3 is School

Let's look at the Model Analysis R Syntax:

Step 1: Model 4.1 (fixed intercept, random intercept for schools (Level 3) and classrooms (Level 2) within schools)
Test for random intercepts for classrooms within schools, what do you conclude?

Step 2: Fixed effects of student-level covariates added to Level 1 Model (Model 4.2)
Test for these fixed effects, what do you conclude? Remember tests with REML are only for same fixed effects!

Step 3: Fixed effects of classroom-level covariates added to Level 2 Model (Model 4.3)
Test for these, what do you conclude?

Step 4: Fixed effects of school-level covariate added to Level 3 Model (Model 4.4)
Test for these, what do you conclude?

What is your final model?

Model diagnostics are in the R Syntax for Final Model Diagnostics!