ABSTRACT
Quasi-least squares (QLS) is a two-stage computational approach for estimation of
the correlation parameters that is in the framework of generalized estimating equa-
tions (GEE). QLS was developed in a series of papers. Stage one was proposed for
balanced data (ni = n ∀ i) in Chaganty (1997) and was then extended for unbalanced and unequally spaced data in Shults (1996) and Shults and Chaganty (1998). A sec-
ond stage of the QLS procedure was then provided in Chaganty and Shults (1999).
In this chapter, we provide a more detailed description of the development of QLS
than was possible in previous manuscripts, which typically had little room for more
than a brief summary of the approach and of the manuscripts on which it was based.
To streamline our descriptions, we describe how QLS evolved for one particular cor-
relation structure, the first-order autoregressive or AR(1) structure. Other correlation
structures will be presented in subsequent chapters. Our focus on the AR(1) struc-
ture here is solely for clarity of exposition as we describe the results contained in the
original manuscripts on QLS.
