Chapter Bootstrapping as an aid to nonnormal data
Barbara M. Byrne · 2013
Bootstrapping as an aid to nonnormal data Two critically important assumptions associated with structural equation modeling (SEM), in the analysis of covariance and mean structures, is the requirement that the data are of a continuous scale and have a multivariate normal distribution. These underlying assumptions are linked to large-sample (i.e., asymptotic) theory within which SEM is embedded. More specically, they derive from the approach taken in the estimation of parameters using the SEM methodology. Typically, either maximum likelihood (ML) or normal theory generalized least squares (GLS) estimation is used; both demand that the data be continuous and multivariate normal. This chapter focuses on the issue of multivariate nonnormality; readers interested in the issue of noncontinuous variables are referred to Bollen (1989a); Byrne (1998); Coenders, Satorra, and Saris (1997); and West, Finch, and Curran (1995).