Also note the striking statistic in the abstract: "...over 40% of those surveyed think apartheid was a good idea, badly executed." And this was a survey conducted in 1996.
a) surely you are familiar with error bars?
b) where did you think that 30% went?
I would be astonished if the poll said otherwise.
I don't think that it follows from this that it's unfair for Y/Z to have a lower grade. They did worse. The grade can reflect that. That doesn't mean they're a worse person / student. It means they demonstrated worse performance in the class. How you interpret that later on is a different question, and interpreting Y's lower performance in the context of the environment that caused it is fair.
Trying to change the scale leads to the measurement being meaningless, and worse, can lead to bringing back biases that were tried to be balanced for: in evaluating a student for a job who looked like Y and another who looked like X and got the same GPA, which one is the better student? If you believe that the grades were more lenient toward Y, picking applicant X as an employer just makes sense.
In another context: we report the times of runners running the 100m dash in seconds and hundredths of seconds, regardless of if they're in the 100M final of the Olympics or at a local high school track meet. Is 10.8 seconds a good time? In the Olympic 100m mens final? no. In the US collegiate championships for women? definitely.
When we actually do do that, we seem to get world-beating Kenyan and Ethiopian athletes.
Not so many affluent, well-fed Swiss ones though.