I can elaborate.
How does this sample account for retirees?
The problem is, you’re trying to show the decline over aging, but at some age on there, the worst samples that declined the most retired and were removed from the next year.
Simply put, this data could be removing every point that disproves your conclusion...