Cross-Sectional Design and Linear Statistics in Vulnerability Research

Abstract

In this chapter we examine the cross-sectional research designs and linear statistics used in community disaster vulnerability research. The foundation of vulnerability theory is based on cross-sectional data and linear statistics. Cross-sectional designs are by far the most popular type of research done on vulnerability. Similarly, linear statistics account for the vast majority of empirical results reported in the vulnerability literature. The advances in linear statistical modeling over the past several decades have made it possible to squeeze more value out of cross-sectional designs. This chapter is structured with two main sections and a summary. We begin with an overview of cross-sectional design. A brief account of the characteristics distinguishing cross-sectional design is given, and the advantages of this design for vulnerability research are discussed. We compare cross-sectional designs to experimental designs and note some of the reasons for choosing a cross-sectional design over an experimental design. We also consider the utility of cross-sectional designs for both descriptions and hypothesis testing. Next we consider linear statis-tical models. We discuss and give examples of disaster vulnerability research using linear regression, logistic regression, hierarchical regression, path analysis, and latent variable structural equation modeling (SEM). We conclude this chapter with a summary of the linear methods used in developing vulnerability theory.