Measures Of Association
Pearson’s correlation coefficient, ρ, introduced in Chapter 6, is ubiquitous in applied research. It is employed even more than the least-squares estimate of the slope of a regression line and often is the lone tool used to detect and describe an association between two variables. As is evident from results covered in previous chapters, the estimate of ρ, r, can miss important associ- ations. And even if the value of ρ could be determined exactly, its value can be relatively unrevealing. The goal in this chapter is to introduce some new tools for detecting and describing the association between variables that deal with some of the practical problems associated with r. But first we look more closely at how r and ρ are interpreted and the features of data that influence their values.
KeywordsBibliographic Note Minimum Covariance Determinant Minimum Volume Ellipsoid Unusual Point Minimum Covariance Determinant Estimator
Unable to display preview. Download preview PDF.