It has already been discussed that there are two regression lines and they show mutual relationship between two variable . The regression line Yon X gives the most probable value of y of given value of x whereas the regression line x on y gives the most probable values of y
Why there are two Regression Lines:
First Reason: For two mutually related series there are two regression lines. First line of regression is X on Y and second line of regression is X on Y.
While constructing line of regression of X on Y, Y is treated as independent variable whereas X is treated as dependent variable. This line gives most probable values of X for given values of X for given values of Y. In the same way line of regression of Y on variable. This line gives the most probable values of Y for given values of X. Practically X and Y both variables may be required to be estimated, hence there is necessity of two regression lines. One for best estimation of X and other for Y
Second Reason: The regression lines are those best fit lines which are drawn on least squares assumption. Under least square method the line which are to be drawn should be in that manner so that the total of the squares of the deviations of the various points is minimum. The deviation of the various points of actual values up to the regression online can be measured by two ways (a) Horizontally i.e. parallel to X axis and (b) Vertically i.e. parallel to Y axis .Hence for minimising the total of squares separately there should be two regression lines.
The regression line Y and X is drawn in such a way that it minimises total of squares of the vertical deviations. In the same way regression line X on Y is drawn in such a way that it minimises the total squares of the horizontal deviations. Hence it is essential to have two regressio line under the assumptions of least square method.