Descriptive study

 Eight variables were treated as independent variables

in these analyses. For the purpose of this study, age, sex

and race were included as biological variables, stage and

histology were included as disease variables, and initial and

subsequent treatment were included as medical variables.

 Descriptive information was obtained on the eight

variables for each cancer site in order to compare the

important characteristics in these groups with the character-

istics of individuals interviewed for the psychosocial

analysis. Measures of central tendency and dispersion were

obtained for continuous variables and the frequency distri-

bution of values on categorical variables was studied.

Special attention was paid to the distribution of values on

each of the variables intended for use as independent

variables in the regression analysis.

 The analysis of the data for the three cancer sites

also permitted evaluation of the distribution of the variable

survival, this being the variable intended for use as the

dependent variable in the regression analysis. The survival

of individuals with cancer is known to be skewed, more so for

some sites than others. As previously noted, this was

reflected by the data for lung, breast, and recto-

colon cancer. In the final regression equations selected for

prediction, the natural logarithmic transformation of