Chapter 4 Building Hierarchical Linear Models

Data often contain information that has a hierarchical (multilevel) structure. For example, a business may have many stores, and customer purchasing behavior at each store may be influenced not only by customer-level characteristics (for example, income, age, and so on), but also by characteristics of where the store is located (for example, a large urban center), as well as characteristics of the store itself (for example, store size, store condition, number of employees). In this case, we could develop a model using variables from all levels of this hierarchy (customer and store characteristics); however, if we just incorporate all these variables in a linear regression model, for example, this ignores the influences of store and location variables on customer-level variables, as well as the correlations that exist among the observations within levels (customers within a location). As another example, if we were investigating the effects of a new teaching method on math scores for a school district, we would need to take into account not only student characteristics (IQ, gender, and so on), but also the characteristics of the classrooms (for example, class size) and schools (for example SES).

In this chapter we discuss hierarchical linear mixed models. We start with an overview of the technique and then go through an example that illustrates where you can perform this technique within SPSS Statistics. We also discuss the options ...

Get SPSS Statistics for Data Analysis and Visualization now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.