400 CHAPTER 11 SIMPLE LINEAR REGRESSION AND CORRELATION11-9 TRANSFORMATIONS TO A STRAIGHT LINEWe occasionally find that the straight-line regression model Y 0 1 xis inappropri-
ate because the true regression function is nonlinear. Sometimes nonlinearity is visually de-
termined from the scatter diagram, and sometimes, because of prior experience or underlying
theory, we know in advance that the model is nonlinear. Occasionally, a scatter diagram will
exhibit an apparent nonlinear relationship between Yand x. In some of these situations, a non-
linear function can be expressed as a straight line by using a suitable transformation. Such
nonlinear models are called intrinsically linear.
As an example of a nonlinear model that is intrinsically linear, consider the exponential
functionThis function is intrinsically linear, since it can be transformed to a straight line by a logarith-
mic transformationThis transformation requires that the transformed error terms ln are normally and independ-
ently distributed with mean 0 and variance ^2.
Another intrinsically linear function isBy using the reciprocal transformation z 1 x, the model is linearized toSometimes several transformations can be employed jointly to linearize a function. For ex-
ample, consider the functionletting , we have the linearized formFor examples of fitting these models, refer to Montgomery, Peck, and Vining (2001) or
Myers (1990).11-10 MORE ABOUT TRANSFORMATIONS (CD ONLY)11-11 CORRELATIONOur development of regression analysis has assumed that xis a mathematical variable, meas-
ured with negligible error, and that Yis a random variable. Many applications of regression
analysis involve situations in which both Xand Yare random variables. In these situations, itln Y* 0 1 xY* 1
Y
Y1
exp 1 0 1 x 2Y 0 1 zY 0 1 a1
xbln Yln 0 1 xln Y 0 e^1 xc 11 .qxd 5/20/02 1:17 PM Page 400 RK UL 6 RK UL 6:Desktop Folder:TEMP WORK:MONTGOMERY:REVISES UPLO D CH114 FIN L:Quark Files: