ABSTRACT
Linear regression was introduced in the previous chapter. We now provide a more detailed discussion of this topic. Our aim is to calculate potency values (such as
D
) from regression lines that are modeled on a logarithmic scale, i.e., as effect vs. log dose. These potencies are needed in isobolar combination analysis discussed in the next chapter. Here we will be especially concerned with tests of the parallelism of the regression lines of effect on log dose and the importance of this in combination drug analysis. The concepts and methods presented here will have many other applications also. When it is necessary to distinguish between an observed
y
-value,
y
, and the corresponding value on the line, the latter will be denoted with a capital letter,
Y
, as we did in the previous chapter. Thus, the regression line is written,
Y
=
a
+
bx
, and at any
x
-value,
x
, we have
Y
=
a
+
bx
.
