Was it 85% of drivers who 100% ignore any urban 20 mph limit? This study shows that in Edinburgh, mean speed fell by just 1.34 mph and median speed by 0.47 mph, measured 12 months after implementation. But somehow the casualty rate fell by 39%.
In Belfast, speeds did not drop and the casualty rate fell by just 2%. The difference is ascribed to differences in the design of the schemes - Edinburgh's is more comprehensive through the city.
What can we conclude? I can't see how the tiny speed reduction can possibly be the factor that cut Edinburgh's casualty numbers, though I suppose many of the streets the limit applies to were already slower than 30 due to traffic or restricted width due to buildings or parked cars. So previously faster roads may be over-represented in the casualty reduction.
I can't really escape the conclusion though that most drivers really don't give a fig (or a stronger word) for speed limits.
Here's the paper:
If you're new please join in and if you have questions pop them below and the forum regulars will answer as best we can.