COLIN GALLAGHER
Dept. of Math and Statistics
California State University, Chico

ABSTRACT

We will consider a measure of linear association between random variables, which like the correlation can be defined as the ratio of two expectations.  Because of its relationship to the covariation--a measure of dependence between jointly stable random variables, I call this measure the covariation. We will explore some of the properties of a simple estimator of the covariation (the sample covariation function) and some possible applications to time series modeling.    Finally, we will compare the behavior of the sample covariation and correlation functions for symmetric stable linear processes.