statistics · cause
Lord's paradox
Same rise. Two ways to say who did better.
Two groups start with different averages. Both groups change by the same amount. Looking at final scores versus looking at change scores can suggest opposite conclusions about which group did better.
Both groups improved by the same amount. Which comparison do you trust?
Group A starts lower and ends lower. Group B starts higher and ends higher.
Both groups rose by exactly the same amount.
If you compare only the final scores, Group B looks better.
If you compare the improvement, the groups are equal.
The same data support different stories depending on whether you look at levels or at changes.