The OLD method measured Prothombin Time but was subject to variation (errors) due to differences in the reagent chemical characteristics.
Then the INR (International Normalized Ratio) method was developed which compares coagulation time of your blood sample to a known reagent and divides the numbers. There are NO units (since you are dividing two numbers, both with the same units so mathematically, the units cancel out. It is simply a Ratio.
I would not be surprised to see a difference of even a few tenths with two tests taken back to back on the same machine.
The danger zone is on the LOW END, near 2.0 since clotting (and possible STROKE) become more likely with a Low INR (<2.0). The risk of bleeding does not rise much until INR is >5.0 so there is a LOT of margin on the high end. Many people report NO bleeding even at 6.0 to 8.0.
As Al Lodwick says, "It's easier to replace Blood Cells than Brain Cells!"
Bottom Line: Many of us prefer to be at the high end of our range to reduce the risk of STROKE.
'AL Capshaw'