ABSTRACT
This chapter focuses on data reduction techniques. Often, researchers in the social sciences are interested in identifying underlying constructs that unite items, known as measured indicators, that overlap statistically and conceptually. For example, the underlying construct of satisfaction with work may be indicated by survey items (i.e., indicators) like “Satisfied with my pay,” “Satisfied with work-life balance,” and “The work is meaningful to me.” The chapter begins with several definitions of terms (e.g., latent variables, constructs, and observed variables) and quickly moves into definitions of the two main statistical procedures covered in the chapter: factor analysis and reliability analysis. There are many kinds of factor analysis. In this chapter, the discussion is limited to a detailed discussion of Exploratory Factor Analysis (EFA) that describes concepts like factor extraction, eigenvalues, factor rotation, and factor loadings, and includes an extended example to illustrate these concepts. Next, a brief introduction to Confirmatory Factor Analysis (CFA) is provided, including definitions of fit statistics and structural equation modeling. Next, reliability analysis is explored in some detail. Reliability analysis allows researchers to determine how well items within a construct go together, based on the strengths of the correlations among all of the items. Cronbach's alpha is discussed, and the various statistics produced in the Statistical Package for the Social Sciences (SPSS) analysis of reliability are presented and described.
