Inverse regression:
When we have the variables y andx both measured with errors (the observed values being y* and x'), we consrider two regreesion equations:
1) .Regression of y' an x', which is called the "direct" regression.
2) Regression of x'on y', which is called the "inverse" or "reverse" regression.
Inverse regression bas beep frequently advocated in the case of analysis'of salary discrimination. Since the problem is one of usual errors in variables, both regressions need to be computed and whether inverse regression alone gives the correct estimates depends on the assumptions we make.
The usual model, in its simplest form, is that of two explanatory variables, one of which is measured with error
What we are interested in is the coefficient of x2. The problem is that xi is measured with error.
Then β2 > 0 implies that men are paid more than womenwth the same qualifications and thus there is gender discrimination A direct least squares estimation of eq. (9.22) with x,' substituted for x,, and β2 > 0 has been frequently used as widen& of gender discrimination.
In the inverse regression wg take
We are asking whether mep w mor<or less qualified than women ha3ing thesame salaries. The proponents of inverse regression argue that ta establish discrlimation, one has to have y2< 0 in eq. that is, among men and women receiving equal salaries, men possess lower qualifications.
The usual errors in variables results however show that we shauld nat make inferences on the basis of β2, and y2 but obtain bounds for β2, from the direct regression and the inverse regression estimates.