Exploratory Factor Analysis

In education, researchers often work with complex data sets that include a multitude of variables. One question that often arises in such contexts is whether the structure of associations that underlies the data is accounted for by a latent construct. Exploratory factor analysis is a multivariate correlational procedure that helps researchers overcome such challenges. It helps reduce large data sets into main components or identify distinct constructs that account for the pattern of correlations among observed variables. These unobservable constructs are referred to as common factors, latent variables, or internal attributes, and they exert linear influences on more than one observed variable. Although exploratory factor analysis is widely used, many applied educational researchers and practitioners are not yet familiar with this procedure and are intimidated by the technical terminology. This book provides a conceptual description of this method and includes a collection of applied research studies that illustrates the application of exploratory factor analysis in school improvement research. The first chapter provides a theoretical overview of exploratory factor analysis. It explains the purposes for which this procedure can be used, the related terminology, the distinction between key concepts, the steps that must be taken, and the criteria for making the decisions. This information can serve as a starting point for researchers who need a brief, conceptual introduction to this topic. The following chapters present a series of research studies in which exploratory factor analysis was employed either by itself or in conjunction with other statistical procedures. The studies presented in this book address a variety of research problems in the field of school improvement. They specify how the factor analytic procedure was applied, and explain the theoretical contributions and the practical applications of the factor analytic results. In most studies, results from factor analysis were used for subsequent statistical procedures, thus helping researchers address more complex research questions and enriching the results.