medical variables including the site of the cancer, the

stage of the disease, and the treatment given for the

cancer.

 The investigation by Stavraky et al. (1968) illustrates

 the use of a comparison group as well as the use of the

 delineation of the patient population as a means of

 limiting variability due to disease effects such as cancer

 site and stage. The second and third investigations (Greer,

 Morris, & Pettingale, 1979; Rogentine et al.,1979) relied on

 the delineation of the population as the measure of control

 of this important part of the variability in survival.

 A third approach to the control of variability is

 illustrated by the work of Weisman and Worden (1975). This

 research used a statistical procedure to address the control

 of variability in survival due to disease and medical

 variables. This methodology was the model for the present

 study.

 In this investigation, the relationship between psycho-

social variables and survival was studied through analysis

of information on preterminal and terminal cancer patients

representing six primary cancer sites ( N = 35). The

determination of their survival as relatively long or short

was made by comparing observed survival to the survival that

was expected. The expected survival for the patients was