Exploratory factor analysis (EFA) is a statistical tool
used for exploring the underlying structure of data. It was originally
developed in the early 1900s during the attempt to determine whether
intelligence is a unitary or multidimensional construct (Spearman,
1904). It has since served as a general-purpose dimension reduction
tool with many applications. In the modern social sciences it is often
used to explore the psychometric properties of an instrument or scale.
Exploratory factor analysis examines all the pairwise relationships
between individual variables (e.g., items on a scale) and seeks to
extract latent factors from the measured variables. During the 110
years since Spearman’s seminal work in this area, few statistical
techniques have been so widely used (or, unfortunately, misused).
The goal of this book
is to explore best practices in applying EFA using SAS. We will review
each of the major EFA steps (e.g., extraction, rotation), some associated
practices (estimation of factor scores and higher-order factors),
and some less common analyses that can inform the generalizability
of EFA results (replication analyses and bootstrap analyses). We will
review the SAS syntax for each task and highlight best practices according
to research and practice. We will also demonstrate the procedures
and analyses discussed throughout the book using real data, and we
will occasionally survey some poor practices as a learning tool.
To get started in our exploration of EFA, we will first
discuss the similarities and differences between EFA and principal
components analysis (PCA), another technique that is commonly used
for the same goal as EFA. We will then briefly summarize the steps
to follow when conducting an EFA and conclude with a quick introduction
to EFA in SAS.