I suspect another reason that paper labels are questionable is because the paper changes characteristics (shrinks, expands) differently than the dvd material. However, I think this impacts readers differently depending on their features.
Comparison of 2 drives with cd / dvd paper labels.
All tests were done on the same DVD, only difference is Scan 2 and Scan 4 were done on the DVD after a paper label was attached. All the tests were done today, so there is no "age" impact.
Original Data DVD was recored on Fuji TYG02 recorded at 16x on Drive 2.
Scan 1 was original DVD, no label, QS test on drive 1.
Scan 2 was original DVD, with paper label, QS test on drive 1 :Z
This scan has more spikes present, and higher Total PIE. Note I scanned three times, spikes are in different locations, however they are present in every scan, whereas they were not present in the original 'no label' scan.
Scan 3 was original DVD, no label, QS test on drive 2
Scan 4 was original DVD, with paper label, QS test on drive 2. Suspect the reason the drive 2 has minimal effect is "Tilt - effect" and "sliding cone" features of this drive. These features appear to help but do not totally eliminate the effect of the paper label. Note the average 'jitter' is higher on the DVD with the label.
If you think about it, it makes sense that you would negatively impact the read of a DVD by adding weight (possibly uneven) to the DVD with a paper label. I suspect CD's are effected less because the data densities on the disk are much lower than on DVD's