Re:Say Hello to Full Employment
by wavegeek on Friday July 06, 2018 @10:03PM
(#17476457)
>control for everything
The studies do not "control for everything". For example I have never seen
1. Control for the much higher levels of sick days taken by women.
2. Control for the product of hours worked and years worked. Women work fewer hours and spend more years out of the work-force. This has a multiplicative effect on "hours of work experience"
Additionally the power of "controlling for X" is vastly overstated. In reality it is only reliable when the different factors are close to orthogonal, have linear relationships and are not causally related. Typically none of the above are true.
The studies do not "control for everything". For example I have never seen
1. Control for the much higher levels of sick days taken by women.
2. Control for the product of hours worked and years worked. Women work fewer hours and spend more years out of the work-force. This has a multiplicative effect on "hours of work experience"
Additionally the power of "controlling for X" is vastly overstated. In reality it is only reliable when the different factors are close to orthogonal, have linear relationships and are not causally related. Typically none of the above are true.